1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
44
from bzrlib.bundle import serializer
45
from bzrlib.revisiontree import RevisionTree
46
from bzrlib.store.versioned import VersionedFileStore
47
from bzrlib.store.text import TextStore
48
from bzrlib.testament import Testament
49
from bzrlib.util import bencode
52
from bzrlib.decorators import needs_read_lock, needs_write_lock
53
from bzrlib.inter import InterObject
54
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
55
from bzrlib.symbol_versioning import (
58
from bzrlib.trace import mutter, mutter_callsite, note, warning
61
# Old formats display a warning, but only once
62
_deprecation_warning_done = False
65
class CommitBuilder(object):
66
"""Provides an interface to build up a commit.
68
This allows describing a tree to be committed without needing to
69
know the internals of the format of the repository.
72
# all clients should supply tree roots.
73
record_root_entry = True
74
# the default CommitBuilder does not manage trees whose root is versioned.
75
_versioned_root = False
77
def __init__(self, repository, parents, config, timestamp=None,
78
timezone=None, committer=None, revprops=None,
80
"""Initiate a CommitBuilder.
82
:param repository: Repository to commit to.
83
:param parents: Revision ids of the parents of the new revision.
84
:param config: Configuration to use.
85
:param timestamp: Optional timestamp recorded for commit.
86
:param timezone: Optional timezone for timestamp.
87
:param committer: Optional committer to set for commit.
88
:param revprops: Optional dictionary of revision properties.
89
:param revision_id: Optional revision id.
94
self._committer = self._config.username()
96
assert isinstance(committer, basestring), type(committer)
97
self._committer = committer
99
self.new_inventory = Inventory(None)
100
self._new_revision_id = revision_id
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self._repo_graph = repository.get_graph()
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
symbol_versioning.warn('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _check_root(self, ie, parent_invs, tree):
192
"""Helper for record_entry_contents.
194
:param ie: An entry being added.
195
:param parent_invs: The inventories of the parent revisions of the
197
:param tree: The tree that is being committed.
199
# In this revision format, root entries have no knit or weave When
200
# serializing out to disk and back in root.revision is always
202
ie.revision = self._new_revision_id
204
def _get_delta(self, ie, basis_inv, path):
205
"""Get a delta against the basis inventory for ie."""
206
if ie.file_id not in basis_inv:
208
return (None, path, ie.file_id, ie)
209
elif ie != basis_inv[ie.file_id]:
211
# TODO: avoid tis id2path call.
212
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
217
def record_entry_contents(self, ie, parent_invs, path, tree,
219
"""Record the content of ie from tree into the commit if needed.
221
Side effect: sets ie.revision when unchanged
223
:param ie: An inventory entry present in the commit.
224
:param parent_invs: The inventories of the parent revisions of the
226
:param path: The path the entry is at in the tree.
227
:param tree: The tree which contains this entry and should be used to
229
:param content_summary: Summary data from the tree about the paths
230
content - stat, length, exec, sha/link target. This is only
231
accessed when the entry has a revision of None - that is when it is
232
a candidate to commit.
233
:return: A tuple (change_delta, version_recorded). change_delta is
234
an inventory_delta change for this entry against the basis tree of
235
the commit, or None if no change occured against the basis tree.
236
version_recorded is True if a new version of the entry has been
237
recorded. For instance, committing a merge where a file was only
238
changed on the other side will return (delta, False).
240
if self.new_inventory.root is None:
241
if ie.parent_id is not None:
242
raise errors.RootMissing()
243
self._check_root(ie, parent_invs, tree)
244
if ie.revision is None:
245
kind = content_summary[0]
247
# ie is carried over from a prior commit
249
# XXX: repository specific check for nested tree support goes here - if
250
# the repo doesn't want nested trees we skip it ?
251
if (kind == 'tree-reference' and
252
not self.repository._format.supports_tree_reference):
253
# mismatch between commit builder logic and repository:
254
# this needs the entry creation pushed down into the builder.
255
raise NotImplementedError('Missing repository subtree support.')
256
# transitional assert only, will remove before release.
257
assert ie.kind == kind
258
self.new_inventory.add(ie)
260
# TODO: slow, take it out of the inner loop.
262
basis_inv = parent_invs[0]
264
basis_inv = Inventory(root_id=None)
266
# ie.revision is always None if the InventoryEntry is considered
267
# for committing. We may record the previous parents revision if the
268
# content is actually unchanged against a sole head.
269
if ie.revision is not None:
270
if self._versioned_root or path != '':
271
# not considered for commit
274
# repositories that do not version the root set the root's
275
# revision to the new commit even when no change occurs, and
276
# this masks when a change may have occurred against the basis,
277
# so calculate if one happened.
278
if ie.file_id not in basis_inv:
280
delta = (None, path, ie.file_id, ie)
282
basis_id = basis_inv[ie.file_id]
283
if basis_id.name != '':
285
delta = (basis_inv.id2path(ie.file_id), path,
290
# not considered for commit, OR, for non-rich-root
291
return delta, ie.revision == self._new_revision_id and (path != '' or
292
self._versioned_root)
294
# XXX: Friction: parent_candidates should return a list not a dict
295
# so that we don't have to walk the inventories again.
296
parent_candiate_entries = ie.parent_candidates(parent_invs)
297
head_set = self._repo_graph.heads(parent_candiate_entries.keys())
299
for inv in parent_invs:
300
if ie.file_id in inv:
301
old_rev = inv[ie.file_id].revision
302
if old_rev in head_set:
303
heads.append(inv[ie.file_id].revision)
304
head_set.remove(inv[ie.file_id].revision)
307
# now we check to see if we need to write a new record to the
309
# We write a new entry unless there is one head to the ancestors, and
310
# the kind-derived content is unchanged.
312
# Cheapest check first: no ancestors, or more the one head in the
313
# ancestors, we write a new node.
317
# There is a single head, look it up for comparison
318
parent_entry = parent_candiate_entries[heads[0]]
319
# if the non-content specific data has changed, we'll be writing a
321
if (parent_entry.parent_id != ie.parent_id or
322
parent_entry.name != ie.name):
324
# now we need to do content specific checks:
326
# if the kind changed the content obviously has
327
if kind != parent_entry.kind:
330
assert content_summary[2] is not None, \
331
"Files must not have executable = None"
333
if (# if the file length changed we have to store:
334
parent_entry.text_size != content_summary[1] or
335
# if the exec bit has changed we have to store:
336
parent_entry.executable != content_summary[2]):
338
elif parent_entry.text_sha1 == content_summary[3]:
339
# all meta and content is unchanged (using a hash cache
340
# hit to check the sha)
341
ie.revision = parent_entry.revision
342
ie.text_size = parent_entry.text_size
343
ie.text_sha1 = parent_entry.text_sha1
344
ie.executable = parent_entry.executable
345
return self._get_delta(ie, basis_inv, path), False
347
# Either there is only a hash change(no hash cache entry,
348
# or same size content change), or there is no change on
350
# Provide the parent's hash to the store layer, so that the
351
# content is unchanged we will not store a new node.
352
nostore_sha = parent_entry.text_sha1
354
# We want to record a new node regardless of the presence or
355
# absence of a content change in the file.
357
ie.executable = content_summary[2]
358
lines = tree.get_file(ie.file_id, path).readlines()
360
ie.text_sha1, ie.text_size = self._add_text_to_weave(
361
ie.file_id, lines, heads, nostore_sha)
362
except errors.ExistingContent:
363
# Turns out that the file content was unchanged, and we were
364
# only going to store a new node if it was changed. Carry over
366
ie.revision = parent_entry.revision
367
ie.text_size = parent_entry.text_size
368
ie.text_sha1 = parent_entry.text_sha1
369
ie.executable = parent_entry.executable
370
return self._get_delta(ie, basis_inv, path), False
371
elif kind == 'directory':
373
# all data is meta here, nothing specific to directory, so
375
ie.revision = parent_entry.revision
376
return self._get_delta(ie, basis_inv, path), False
378
self._add_text_to_weave(ie.file_id, lines, heads, None)
379
elif kind == 'symlink':
380
current_link_target = content_summary[3]
382
# symlink target is not generic metadata, check if it has
384
if current_link_target != parent_entry.symlink_target:
387
# unchanged, carry over.
388
ie.revision = parent_entry.revision
389
ie.symlink_target = parent_entry.symlink_target
390
return self._get_delta(ie, basis_inv, path), False
391
ie.symlink_target = current_link_target
393
self._add_text_to_weave(ie.file_id, lines, heads, None)
394
elif kind == 'tree-reference':
396
if content_summary[3] != parent_entry.reference_revision:
399
# unchanged, carry over.
400
ie.reference_revision = parent_entry.reference_revision
401
ie.revision = parent_entry.revision
402
return self._get_delta(ie, basis_inv, path), False
403
ie.reference_revision = content_summary[3]
405
self._add_text_to_weave(ie.file_id, lines, heads, None)
407
raise NotImplementedError('unknown kind')
408
ie.revision = self._new_revision_id
409
return self._get_delta(ie, basis_inv, path), True
411
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
412
versionedfile = self.repository.weave_store.get_weave_or_empty(
413
file_id, self.repository.get_transaction())
414
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
415
# than add_lines, and allows committing when a parent is ghosted for
417
# Note: as we read the content directly from the tree, we know its not
418
# been turned into unicode or badly split - but a broken tree
419
# implementation could give us bad output from readlines() so this is
420
# not a guarantee of safety. What would be better is always checking
421
# the content during test suite execution. RBC 20070912
423
return versionedfile.add_lines_with_ghosts(
424
self._new_revision_id, parents, new_lines,
425
nostore_sha=nostore_sha, random_id=self.random_revid,
426
check_content=False)[0:2]
428
versionedfile.clear_cache()
431
class RootCommitBuilder(CommitBuilder):
432
"""This commitbuilder actually records the root id"""
434
# the root entry gets versioned properly by this builder.
435
_versioned_root = True
437
def _check_root(self, ie, parent_invs, tree):
438
"""Helper for record_entry_contents.
440
:param ie: An entry being added.
441
:param parent_invs: The inventories of the parent revisions of the
443
:param tree: The tree that is being committed.
447
######################################################################
450
class Repository(object):
451
"""Repository holding history for one or more branches.
453
The repository holds and retrieves historical information including
454
revisions and file history. It's normally accessed only by the Branch,
455
which views a particular line of development through that history.
457
The Repository builds on top of Stores and a Transport, which respectively
458
describe the disk data format and the way of accessing the (possibly
462
# What class to use for a CommitBuilder. Often its simpler to change this
463
# in a Repository class subclass rather than to override
464
# get_commit_builder.
465
_commit_builder_class = CommitBuilder
466
# The search regex used by xml based repositories to determine what things
467
# where changed in a single commit.
468
_file_ids_altered_regex = lazy_regex.lazy_compile(
469
r'file_id="(?P<file_id>[^"]+)"'
470
r'.* revision="(?P<revision_id>[^"]+)"'
473
def abort_write_group(self):
474
"""Commit the contents accrued within the current write group.
476
:seealso: start_write_group.
478
if self._write_group is not self.get_transaction():
479
# has an unlock or relock occured ?
480
raise errors.BzrError('mismatched lock context and write group.')
481
self._abort_write_group()
482
self._write_group = None
484
def _abort_write_group(self):
485
"""Template method for per-repository write group cleanup.
487
This is called during abort before the write group is considered to be
488
finished and should cleanup any internal state accrued during the write
489
group. There is no requirement that data handed to the repository be
490
*not* made available - this is not a rollback - but neither should any
491
attempt be made to ensure that data added is fully commited. Abort is
492
invoked when an error has occured so futher disk or network operations
493
may not be possible or may error and if possible should not be
498
def add_inventory(self, revision_id, inv, parents):
499
"""Add the inventory inv to the repository as revision_id.
501
:param parents: The revision ids of the parents that revision_id
502
is known to have and are in the repository already.
504
returns the sha1 of the serialized inventory.
506
_mod_revision.check_not_reserved_id(revision_id)
507
assert inv.revision_id is None or inv.revision_id == revision_id, \
508
"Mismatch between inventory revision" \
509
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
510
assert inv.root is not None
511
inv_lines = self._serialise_inventory_to_lines(inv)
512
inv_vf = self.get_inventory_weave()
513
return self._inventory_add_lines(inv_vf, revision_id, parents,
514
inv_lines, check_content=False)
516
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
518
"""Store lines in inv_vf and return the sha1 of the inventory."""
520
for parent in parents:
522
final_parents.append(parent)
523
return inv_vf.add_lines(revision_id, final_parents, lines,
524
check_content=check_content)[0]
527
def add_revision(self, revision_id, rev, inv=None, config=None):
528
"""Add rev to the revision store as revision_id.
530
:param revision_id: the revision id to use.
531
:param rev: The revision object.
532
:param inv: The inventory for the revision. if None, it will be looked
533
up in the inventory storer
534
:param config: If None no digital signature will be created.
535
If supplied its signature_needed method will be used
536
to determine if a signature should be made.
538
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
540
_mod_revision.check_not_reserved_id(revision_id)
541
if config is not None and config.signature_needed():
543
inv = self.get_inventory(revision_id)
544
plaintext = Testament(rev, inv).as_short_text()
545
self.store_revision_signature(
546
gpg.GPGStrategy(config), plaintext, revision_id)
547
if not revision_id in self.get_inventory_weave():
549
raise errors.WeaveRevisionNotPresent(revision_id,
550
self.get_inventory_weave())
552
# yes, this is not suitable for adding with ghosts.
553
self.add_inventory(revision_id, inv, rev.parent_ids)
554
self._revision_store.add_revision(rev, self.get_transaction())
556
def _add_revision_text(self, revision_id, text):
557
revision = self._revision_store._serializer.read_revision_from_string(
559
self._revision_store._add_revision(revision, StringIO(text),
560
self.get_transaction())
562
def all_revision_ids(self):
563
"""Returns a list of all the revision ids in the repository.
565
This is deprecated because code should generally work on the graph
566
reachable from a particular revision, and ignore any other revisions
567
that might be present. There is no direct replacement method.
569
if 'evil' in debug.debug_flags:
570
mutter_callsite(2, "all_revision_ids is linear with history.")
571
return self._all_revision_ids()
573
def _all_revision_ids(self):
574
"""Returns a list of all the revision ids in the repository.
576
These are in as much topological order as the underlying store can
579
raise NotImplementedError(self._all_revision_ids)
581
def break_lock(self):
582
"""Break a lock if one is present from another instance.
584
Uses the ui factory to ask for confirmation if the lock may be from
587
self.control_files.break_lock()
590
def _eliminate_revisions_not_present(self, revision_ids):
591
"""Check every revision id in revision_ids to see if we have it.
593
Returns a set of the present revisions.
596
for id in revision_ids:
597
if self.has_revision(id):
602
def create(a_bzrdir):
603
"""Construct the current default format repository in a_bzrdir."""
604
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
606
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
607
"""instantiate a Repository.
609
:param _format: The format of the repository on disk.
610
:param a_bzrdir: The BzrDir of the repository.
612
In the future we will have a single api for all stores for
613
getting file texts, inventories and revisions, then
614
this construct will accept instances of those things.
616
super(Repository, self).__init__()
617
self._format = _format
618
# the following are part of the public API for Repository:
619
self.bzrdir = a_bzrdir
620
self.control_files = control_files
621
self._revision_store = _revision_store
622
# backwards compatibility
623
self.weave_store = text_store
625
self._reconcile_does_inventory_gc = True
626
self._reconcile_fixes_text_parents = False
627
# not right yet - should be more semantically clear ?
629
self.control_store = control_store
630
self.control_weaves = control_store
631
# TODO: make sure to construct the right store classes, etc, depending
632
# on whether escaping is required.
633
self._warn_if_deprecated()
634
self._write_group = None
635
self.base = control_files._transport.base
638
return '%s(%r)' % (self.__class__.__name__,
641
def has_same_location(self, other):
642
"""Returns a boolean indicating if this repository is at the same
643
location as another repository.
645
This might return False even when two repository objects are accessing
646
the same physical repository via different URLs.
648
if self.__class__ is not other.__class__:
650
return (self.control_files._transport.base ==
651
other.control_files._transport.base)
653
def is_in_write_group(self):
654
"""Return True if there is an open write group.
656
:seealso: start_write_group.
658
return self._write_group is not None
661
return self.control_files.is_locked()
663
def is_write_locked(self):
664
"""Return True if this object is write locked."""
665
return self.is_locked() and self.control_files._lock_mode == 'w'
667
def lock_write(self, token=None):
668
"""Lock this repository for writing.
670
This causes caching within the repository obejct to start accumlating
671
data during reads, and allows a 'write_group' to be obtained. Write
672
groups must be used for actual data insertion.
674
:param token: if this is already locked, then lock_write will fail
675
unless the token matches the existing lock.
676
:returns: a token if this instance supports tokens, otherwise None.
677
:raises TokenLockingNotSupported: when a token is given but this
678
instance doesn't support using token locks.
679
:raises MismatchedToken: if the specified token doesn't match the token
680
of the existing lock.
681
:seealso: start_write_group.
683
A token should be passed in if you know that you have locked the object
684
some other way, and need to synchronise this object's state with that
687
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
689
result = self.control_files.lock_write(token=token)
694
self.control_files.lock_read()
697
def get_physical_lock_status(self):
698
return self.control_files.get_physical_lock_status()
700
def leave_lock_in_place(self):
701
"""Tell this repository not to release the physical lock when this
704
If lock_write doesn't return a token, then this method is not supported.
706
self.control_files.leave_in_place()
708
def dont_leave_lock_in_place(self):
709
"""Tell this repository to release the physical lock when this
710
object is unlocked, even if it didn't originally acquire it.
712
If lock_write doesn't return a token, then this method is not supported.
714
self.control_files.dont_leave_in_place()
717
def gather_stats(self, revid=None, committers=None):
718
"""Gather statistics from a revision id.
720
:param revid: The revision id to gather statistics from, if None, then
721
no revision specific statistics are gathered.
722
:param committers: Optional parameter controlling whether to grab
723
a count of committers from the revision specific statistics.
724
:return: A dictionary of statistics. Currently this contains:
725
committers: The number of committers if requested.
726
firstrev: A tuple with timestamp, timezone for the penultimate left
727
most ancestor of revid, if revid is not the NULL_REVISION.
728
latestrev: A tuple with timestamp, timezone for revid, if revid is
729
not the NULL_REVISION.
730
revisions: The total revision count in the repository.
731
size: An estimate disk size of the repository in bytes.
734
if revid and committers:
735
result['committers'] = 0
736
if revid and revid != _mod_revision.NULL_REVISION:
738
all_committers = set()
739
revisions = self.get_ancestry(revid)
740
# pop the leading None
742
first_revision = None
744
# ignore the revisions in the middle - just grab first and last
745
revisions = revisions[0], revisions[-1]
746
for revision in self.get_revisions(revisions):
747
if not first_revision:
748
first_revision = revision
750
all_committers.add(revision.committer)
751
last_revision = revision
753
result['committers'] = len(all_committers)
754
result['firstrev'] = (first_revision.timestamp,
755
first_revision.timezone)
756
result['latestrev'] = (last_revision.timestamp,
757
last_revision.timezone)
759
# now gather global repository information
760
if self.bzrdir.root_transport.listable():
761
c, t = self._revision_store.total_size(self.get_transaction())
762
result['revisions'] = c
766
def get_data_stream(self, revision_ids):
767
raise NotImplementedError(self.get_data_stream)
769
def insert_data_stream(self, stream):
770
for item_key, bytes in stream:
771
if item_key[0] == 'file':
772
(file_id,) = item_key[1:]
773
knit = self.weave_store.get_weave_or_empty(
774
file_id, self.get_transaction())
775
elif item_key == ('inventory',):
776
knit = self.get_inventory_weave()
777
elif item_key == ('revisions',):
778
knit = self._revision_store.get_revision_file(
779
self.get_transaction())
780
elif item_key == ('signatures',):
781
knit = self._revision_store.get_signature_file(
782
self.get_transaction())
784
raise RepositoryDataStreamError(
785
"Unrecognised data stream key '%s'" % (item_key,))
786
decoded_list = bencode.bdecode(bytes)
787
format = decoded_list.pop(0)
790
for version, options, parents, some_bytes in decoded_list:
791
data_list.append((version, options, len(some_bytes), parents))
792
knit_bytes += some_bytes
793
knit.insert_data_stream(
794
(format, data_list, StringIO(knit_bytes).read))
797
def missing_revision_ids(self, other, revision_id=None):
798
"""Return the revision ids that other has that this does not.
800
These are returned in topological order.
802
revision_id: only return revision ids included by revision_id.
804
return InterRepository.get(other, self).missing_revision_ids(revision_id)
808
"""Open the repository rooted at base.
810
For instance, if the repository is at URL/.bzr/repository,
811
Repository.open(URL) -> a Repository instance.
813
control = bzrdir.BzrDir.open(base)
814
return control.open_repository()
816
def copy_content_into(self, destination, revision_id=None):
817
"""Make a complete copy of the content in self into destination.
819
This is a destructive operation! Do not use it on existing
822
return InterRepository.get(self, destination).copy_content(revision_id)
824
def commit_write_group(self):
825
"""Commit the contents accrued within the current write group.
827
:seealso: start_write_group.
829
if self._write_group is not self.get_transaction():
830
# has an unlock or relock occured ?
831
raise errors.BzrError('mismatched lock context %r and '
833
(self.get_transaction(), self._write_group))
834
self._commit_write_group()
835
self._write_group = None
837
def _commit_write_group(self):
838
"""Template method for per-repository write group cleanup.
840
This is called before the write group is considered to be
841
finished and should ensure that all data handed to the repository
842
for writing during the write group is safely committed (to the
843
extent possible considering file system caching etc).
846
def fetch(self, source, revision_id=None, pb=None):
847
"""Fetch the content required to construct revision_id from source.
849
If revision_id is None all content is copied.
851
# fast path same-url fetch operations
852
if self.has_same_location(source):
853
# check that last_revision is in 'from' and then return a
855
if (revision_id is not None and
856
not _mod_revision.is_null(revision_id)):
857
self.get_revision(revision_id)
859
inter = InterRepository.get(source, self)
861
return inter.fetch(revision_id=revision_id, pb=pb)
862
except NotImplementedError:
863
raise errors.IncompatibleRepositories(source, self)
865
def create_bundle(self, target, base, fileobj, format=None):
866
return serializer.write_bundle(self, target, base, fileobj, format)
868
def get_commit_builder(self, branch, parents, config, timestamp=None,
869
timezone=None, committer=None, revprops=None,
871
"""Obtain a CommitBuilder for this repository.
873
:param branch: Branch to commit to.
874
:param parents: Revision ids of the parents of the new revision.
875
:param config: Configuration to use.
876
:param timestamp: Optional timestamp recorded for commit.
877
:param timezone: Optional timezone for timestamp.
878
:param committer: Optional committer to set for commit.
879
:param revprops: Optional dictionary of revision properties.
880
:param revision_id: Optional revision id.
882
result = self._commit_builder_class(self, parents, config,
883
timestamp, timezone, committer, revprops, revision_id)
884
self.start_write_group()
888
if (self.control_files._lock_count == 1 and
889
self.control_files._lock_mode == 'w'):
890
if self._write_group is not None:
891
raise errors.BzrError(
892
'Must end write groups before releasing write locks.')
893
self.control_files.unlock()
896
def clone(self, a_bzrdir, revision_id=None):
897
"""Clone this repository into a_bzrdir using the current format.
899
Currently no check is made that the format of this repository and
900
the bzrdir format are compatible. FIXME RBC 20060201.
902
:return: The newly created destination repository.
904
# TODO: deprecate after 0.16; cloning this with all its settings is
905
# probably not very useful -- mbp 20070423
906
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
907
self.copy_content_into(dest_repo, revision_id)
910
def start_write_group(self):
911
"""Start a write group in the repository.
913
Write groups are used by repositories which do not have a 1:1 mapping
914
between file ids and backend store to manage the insertion of data from
915
both fetch and commit operations.
917
A write lock is required around the start_write_group/commit_write_group
918
for the support of lock-requiring repository formats.
920
One can only insert data into a repository inside a write group.
924
if not self.is_write_locked():
925
raise errors.NotWriteLocked(self)
926
if self._write_group:
927
raise errors.BzrError('already in a write group')
928
self._start_write_group()
929
# so we can detect unlock/relock - the write group is now entered.
930
self._write_group = self.get_transaction()
932
def _start_write_group(self):
933
"""Template method for per-repository write group startup.
935
This is called before the write group is considered to be
940
def sprout(self, to_bzrdir, revision_id=None):
941
"""Create a descendent repository for new development.
943
Unlike clone, this does not copy the settings of the repository.
945
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
946
dest_repo.fetch(self, revision_id=revision_id)
949
def _create_sprouting_repo(self, a_bzrdir, shared):
950
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
951
# use target default format.
952
dest_repo = a_bzrdir.create_repository()
954
# Most control formats need the repository to be specifically
955
# created, but on some old all-in-one formats it's not needed
957
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
958
except errors.UninitializableFormat:
959
dest_repo = a_bzrdir.open_repository()
963
def has_revision(self, revision_id):
964
"""True if this repository has a copy of the revision."""
965
if 'evil' in debug.debug_flags:
966
mutter_callsite(3, "has_revision is a LBYL symptom.")
967
return self._revision_store.has_revision_id(revision_id,
968
self.get_transaction())
971
def get_revision(self, revision_id):
972
"""Return the Revision object for a named revision."""
973
return self.get_revisions([revision_id])[0]
976
def get_revision_reconcile(self, revision_id):
977
"""'reconcile' helper routine that allows access to a revision always.
979
This variant of get_revision does not cross check the weave graph
980
against the revision one as get_revision does: but it should only
981
be used by reconcile, or reconcile-alike commands that are correcting
982
or testing the revision graph.
984
return self._get_revisions([revision_id])[0]
987
def get_revisions(self, revision_ids):
988
"""Get many revisions at once."""
989
return self._get_revisions(revision_ids)
992
def _get_revisions(self, revision_ids):
993
"""Core work logic to get many revisions without sanity checks."""
994
for rev_id in revision_ids:
995
if not rev_id or not isinstance(rev_id, basestring):
996
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
997
revs = self._revision_store.get_revisions(revision_ids,
998
self.get_transaction())
1000
assert not isinstance(rev.revision_id, unicode)
1001
for parent_id in rev.parent_ids:
1002
assert not isinstance(parent_id, unicode)
1006
def get_revision_xml(self, revision_id):
1007
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1008
# would have already do it.
1009
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1010
rev = self.get_revision(revision_id)
1011
rev_tmp = StringIO()
1012
# the current serializer..
1013
self._revision_store._serializer.write_revision(rev, rev_tmp)
1015
return rev_tmp.getvalue()
1018
def get_deltas_for_revisions(self, revisions):
1019
"""Produce a generator of revision deltas.
1021
Note that the input is a sequence of REVISIONS, not revision_ids.
1022
Trees will be held in memory until the generator exits.
1023
Each delta is relative to the revision's lefthand predecessor.
1025
required_trees = set()
1026
for revision in revisions:
1027
required_trees.add(revision.revision_id)
1028
required_trees.update(revision.parent_ids[:1])
1029
trees = dict((t.get_revision_id(), t) for
1030
t in self.revision_trees(required_trees))
1031
for revision in revisions:
1032
if not revision.parent_ids:
1033
old_tree = self.revision_tree(None)
1035
old_tree = trees[revision.parent_ids[0]]
1036
yield trees[revision.revision_id].changes_from(old_tree)
1039
def get_revision_delta(self, revision_id):
1040
"""Return the delta for one revision.
1042
The delta is relative to the left-hand predecessor of the
1045
r = self.get_revision(revision_id)
1046
return list(self.get_deltas_for_revisions([r]))[0]
1049
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1050
signature = gpg_strategy.sign(plaintext)
1051
self._revision_store.add_revision_signature_text(revision_id,
1053
self.get_transaction())
1055
def fileids_altered_by_revision_ids(self, revision_ids):
1056
"""Find the file ids and versions affected by revisions.
1058
:param revisions: an iterable containing revision ids.
1059
:return: a dictionary mapping altered file-ids to an iterable of
1060
revision_ids. Each altered file-ids has the exact revision_ids that
1061
altered it listed explicitly.
1063
assert self._serializer.support_altered_by_hack, \
1064
("fileids_altered_by_revision_ids only supported for branches "
1065
"which store inventory as unnested xml, not on %r" % self)
1066
selected_revision_ids = set(revision_ids)
1067
w = self.get_inventory_weave()
1070
# this code needs to read every new line in every inventory for the
1071
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1072
# not present in one of those inventories is unnecessary but not
1073
# harmful because we are filtering by the revision id marker in the
1074
# inventory lines : we only select file ids altered in one of those
1075
# revisions. We don't need to see all lines in the inventory because
1076
# only those added in an inventory in rev X can contain a revision=X
1078
unescape_revid_cache = {}
1079
unescape_fileid_cache = {}
1081
# jam 20061218 In a big fetch, this handles hundreds of thousands
1082
# of lines, so it has had a lot of inlining and optimizing done.
1083
# Sorry that it is a little bit messy.
1084
# Move several functions to be local variables, since this is a long
1086
search = self._file_ids_altered_regex.search
1087
unescape = _unescape_xml
1088
setdefault = result.setdefault
1089
pb = ui.ui_factory.nested_progress_bar()
1091
for line in w.iter_lines_added_or_present_in_versions(
1092
selected_revision_ids, pb=pb):
1093
match = search(line)
1096
# One call to match.group() returning multiple items is quite a
1097
# bit faster than 2 calls to match.group() each returning 1
1098
file_id, revision_id = match.group('file_id', 'revision_id')
1100
# Inlining the cache lookups helps a lot when you make 170,000
1101
# lines and 350k ids, versus 8.4 unique ids.
1102
# Using a cache helps in 2 ways:
1103
# 1) Avoids unnecessary decoding calls
1104
# 2) Re-uses cached strings, which helps in future set and
1106
# (2) is enough that removing encoding entirely along with
1107
# the cache (so we are using plain strings) results in no
1108
# performance improvement.
1110
revision_id = unescape_revid_cache[revision_id]
1112
unescaped = unescape(revision_id)
1113
unescape_revid_cache[revision_id] = unescaped
1114
revision_id = unescaped
1116
if revision_id in selected_revision_ids:
1118
file_id = unescape_fileid_cache[file_id]
1120
unescaped = unescape(file_id)
1121
unescape_fileid_cache[file_id] = unescaped
1123
setdefault(file_id, set()).add(revision_id)
1128
def iter_files_bytes(self, desired_files):
1129
"""Iterate through file versions.
1131
Files will not necessarily be returned in the order they occur in
1132
desired_files. No specific order is guaranteed.
1134
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1135
value supplied by the caller as part of desired_files. It should
1136
uniquely identify the file version in the caller's context. (Examples:
1137
an index number or a TreeTransform trans_id.)
1139
bytes_iterator is an iterable of bytestrings for the file. The
1140
kind of iterable and length of the bytestrings are unspecified, but for
1141
this implementation, it is a list of lines produced by
1142
VersionedFile.get_lines().
1144
:param desired_files: a list of (file_id, revision_id, identifier)
1147
transaction = self.get_transaction()
1148
for file_id, revision_id, callable_data in desired_files:
1150
weave = self.weave_store.get_weave(file_id, transaction)
1151
except errors.NoSuchFile:
1152
raise errors.NoSuchIdInRepository(self, file_id)
1153
yield callable_data, weave.get_lines(revision_id)
1155
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1156
"""Get an iterable listing the keys of all the data introduced by a set
1159
The keys will be ordered so that the corresponding items can be safely
1160
fetched and inserted in that order.
1162
:returns: An iterable producing tuples of (knit-kind, file-id,
1163
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1164
'revisions'. file-id is None unless knit-kind is 'file'.
1166
# XXX: it's a bit weird to control the inventory weave caching in this
1167
# generator. Ideally the caching would be done in fetch.py I think. Or
1168
# maybe this generator should explicitly have the contract that it
1169
# should not be iterated until the previously yielded item has been
1172
inv_w = self.get_inventory_weave()
1173
inv_w.enable_cache()
1175
# file ids that changed
1176
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1178
num_file_ids = len(file_ids)
1179
for file_id, altered_versions in file_ids.iteritems():
1180
if _files_pb is not None:
1181
_files_pb.update("fetch texts", count, num_file_ids)
1183
yield ("file", file_id, altered_versions)
1184
# We're done with the files_pb. Note that it finished by the caller,
1185
# just as it was created by the caller.
1189
yield ("inventory", None, revision_ids)
1193
revisions_with_signatures = set()
1194
for rev_id in revision_ids:
1196
self.get_signature_text(rev_id)
1197
except errors.NoSuchRevision:
1201
revisions_with_signatures.add(rev_id)
1203
yield ("signatures", None, revisions_with_signatures)
1206
yield ("revisions", None, revision_ids)
1209
def get_inventory_weave(self):
1210
return self.control_weaves.get_weave('inventory',
1211
self.get_transaction())
1214
def get_inventory(self, revision_id):
1215
"""Get Inventory object by hash."""
1216
return self.deserialise_inventory(
1217
revision_id, self.get_inventory_xml(revision_id))
1219
def deserialise_inventory(self, revision_id, xml):
1220
"""Transform the xml into an inventory object.
1222
:param revision_id: The expected revision id of the inventory.
1223
:param xml: A serialised inventory.
1225
return self._serializer.read_inventory_from_string(xml, revision_id)
1227
def serialise_inventory(self, inv):
1228
return self._serializer.write_inventory_to_string(inv)
1230
def _serialise_inventory_to_lines(self, inv):
1231
return self._serializer.write_inventory_to_lines(inv)
1233
def get_serializer_format(self):
1234
return self._serializer.format_num
1237
def get_inventory_xml(self, revision_id):
1238
"""Get inventory XML as a file object."""
1240
assert isinstance(revision_id, str), type(revision_id)
1241
iw = self.get_inventory_weave()
1242
return iw.get_text(revision_id)
1244
raise errors.HistoryMissing(self, 'inventory', revision_id)
1247
def get_inventory_sha1(self, revision_id):
1248
"""Return the sha1 hash of the inventory entry
1250
return self.get_revision(revision_id).inventory_sha1
1253
def get_revision_graph(self, revision_id=None):
1254
"""Return a dictionary containing the revision graph.
1256
NB: This method should not be used as it accesses the entire graph all
1257
at once, which is much more data than most operations should require.
1259
:param revision_id: The revision_id to get a graph from. If None, then
1260
the entire revision graph is returned. This is a deprecated mode of
1261
operation and will be removed in the future.
1262
:return: a dictionary of revision_id->revision_parents_list.
1264
raise NotImplementedError(self.get_revision_graph)
1267
def get_revision_graph_with_ghosts(self, revision_ids=None):
1268
"""Return a graph of the revisions with ghosts marked as applicable.
1270
:param revision_ids: an iterable of revisions to graph or None for all.
1271
:return: a Graph object with the graph reachable from revision_ids.
1273
if 'evil' in debug.debug_flags:
1275
"get_revision_graph_with_ghosts scales with size of history.")
1276
result = deprecated_graph.Graph()
1277
if not revision_ids:
1278
pending = set(self.all_revision_ids())
1281
pending = set(revision_ids)
1282
# special case NULL_REVISION
1283
if _mod_revision.NULL_REVISION in pending:
1284
pending.remove(_mod_revision.NULL_REVISION)
1285
required = set(pending)
1288
revision_id = pending.pop()
1290
rev = self.get_revision(revision_id)
1291
except errors.NoSuchRevision:
1292
if revision_id in required:
1295
result.add_ghost(revision_id)
1297
for parent_id in rev.parent_ids:
1298
# is this queued or done ?
1299
if (parent_id not in pending and
1300
parent_id not in done):
1302
pending.add(parent_id)
1303
result.add_node(revision_id, rev.parent_ids)
1304
done.add(revision_id)
1307
def _get_history_vf(self):
1308
"""Get a versionedfile whose history graph reflects all revisions.
1310
For weave repositories, this is the inventory weave.
1312
return self.get_inventory_weave()
1314
def iter_reverse_revision_history(self, revision_id):
1315
"""Iterate backwards through revision ids in the lefthand history
1317
:param revision_id: The revision id to start with. All its lefthand
1318
ancestors will be traversed.
1320
if revision_id in (None, _mod_revision.NULL_REVISION):
1322
next_id = revision_id
1323
versionedfile = self._get_history_vf()
1326
parents = versionedfile.get_parents(next_id)
1327
if len(parents) == 0:
1330
next_id = parents[0]
1333
def get_revision_inventory(self, revision_id):
1334
"""Return inventory of a past revision."""
1335
# TODO: Unify this with get_inventory()
1336
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1337
# must be the same as its revision, so this is trivial.
1338
if revision_id is None:
1339
# This does not make sense: if there is no revision,
1340
# then it is the current tree inventory surely ?!
1341
# and thus get_root_id() is something that looks at the last
1342
# commit on the branch, and the get_root_id is an inventory check.
1343
raise NotImplementedError
1344
# return Inventory(self.get_root_id())
1346
return self.get_inventory(revision_id)
1349
def is_shared(self):
1350
"""Return True if this repository is flagged as a shared repository."""
1351
raise NotImplementedError(self.is_shared)
1354
def reconcile(self, other=None, thorough=False):
1355
"""Reconcile this repository."""
1356
from bzrlib.reconcile import RepoReconciler
1357
reconciler = RepoReconciler(self, thorough=thorough)
1358
reconciler.reconcile()
1361
def _refresh_data(self):
1362
"""Helper called from lock_* to ensure coherency with disk.
1364
The default implementation does nothing; it is however possible
1365
for repositories to maintain loaded indices across multiple locks
1366
by checking inside their implementation of this method to see
1367
whether their indices are still valid. This depends of course on
1368
the disk format being validatable in this manner.
1372
def revision_tree(self, revision_id):
1373
"""Return Tree for a revision on this branch.
1375
`revision_id` may be None for the empty tree revision.
1377
# TODO: refactor this to use an existing revision object
1378
# so we don't need to read it in twice.
1379
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1380
return RevisionTree(self, Inventory(root_id=None),
1381
_mod_revision.NULL_REVISION)
1383
inv = self.get_revision_inventory(revision_id)
1384
return RevisionTree(self, inv, revision_id)
1387
def revision_trees(self, revision_ids):
1388
"""Return Tree for a revision on this branch.
1390
`revision_id` may not be None or 'null:'"""
1391
assert None not in revision_ids
1392
assert _mod_revision.NULL_REVISION not in revision_ids
1393
texts = self.get_inventory_weave().get_texts(revision_ids)
1394
for text, revision_id in zip(texts, revision_ids):
1395
inv = self.deserialise_inventory(revision_id, text)
1396
yield RevisionTree(self, inv, revision_id)
1399
def get_ancestry(self, revision_id, topo_sorted=True):
1400
"""Return a list of revision-ids integrated by a revision.
1402
The first element of the list is always None, indicating the origin
1403
revision. This might change when we have history horizons, or
1404
perhaps we should have a new API.
1406
This is topologically sorted.
1408
if _mod_revision.is_null(revision_id):
1410
if not self.has_revision(revision_id):
1411
raise errors.NoSuchRevision(self, revision_id)
1412
w = self.get_inventory_weave()
1413
candidates = w.get_ancestry(revision_id, topo_sorted)
1414
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1417
"""Compress the data within the repository.
1419
This operation only makes sense for some repository types. For other
1420
types it should be a no-op that just returns.
1422
This stub method does not require a lock, but subclasses should use
1423
@needs_write_lock as this is a long running call its reasonable to
1424
implicitly lock for the user.
1428
def print_file(self, file, revision_id):
1429
"""Print `file` to stdout.
1431
FIXME RBC 20060125 as John Meinel points out this is a bad api
1432
- it writes to stdout, it assumes that that is valid etc. Fix
1433
by creating a new more flexible convenience function.
1435
tree = self.revision_tree(revision_id)
1436
# use inventory as it was in that revision
1437
file_id = tree.inventory.path2id(file)
1439
# TODO: jam 20060427 Write a test for this code path
1440
# it had a bug in it, and was raising the wrong
1442
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1443
tree.print_file(file_id)
1445
def get_transaction(self):
1446
return self.control_files.get_transaction()
1448
def revision_parents(self, revision_id):
1449
return self.get_inventory_weave().parent_names(revision_id)
1451
def get_parents(self, revision_ids):
1452
"""See StackedParentsProvider.get_parents"""
1454
for revision_id in revision_ids:
1455
if revision_id == _mod_revision.NULL_REVISION:
1459
parents = self.get_revision(revision_id).parent_ids
1460
except errors.NoSuchRevision:
1463
if len(parents) == 0:
1464
parents = [_mod_revision.NULL_REVISION]
1465
parents_list.append(parents)
1468
def _make_parents_provider(self):
1471
def get_graph(self, other_repository=None):
1472
"""Return the graph walker for this repository format"""
1473
parents_provider = self._make_parents_provider()
1474
if (other_repository is not None and
1475
other_repository.bzrdir.transport.base !=
1476
self.bzrdir.transport.base):
1477
parents_provider = graph._StackedParentsProvider(
1478
[parents_provider, other_repository._make_parents_provider()])
1479
return graph.Graph(parents_provider)
1481
def get_versioned_file_checker(self, revisions, revision_versions_cache):
1482
return VersionedFileChecker(revisions, revision_versions_cache, self)
1485
def set_make_working_trees(self, new_value):
1486
"""Set the policy flag for making working trees when creating branches.
1488
This only applies to branches that use this repository.
1490
The default is 'True'.
1491
:param new_value: True to restore the default, False to disable making
1494
raise NotImplementedError(self.set_make_working_trees)
1496
def make_working_trees(self):
1497
"""Returns the policy for making working trees on new branches."""
1498
raise NotImplementedError(self.make_working_trees)
1501
def sign_revision(self, revision_id, gpg_strategy):
1502
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1503
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1506
def has_signature_for_revision_id(self, revision_id):
1507
"""Query for a revision signature for revision_id in the repository."""
1508
return self._revision_store.has_signature(revision_id,
1509
self.get_transaction())
1512
def get_signature_text(self, revision_id):
1513
"""Return the text for a signature."""
1514
return self._revision_store.get_signature_text(revision_id,
1515
self.get_transaction())
1518
def check(self, revision_ids=None):
1519
"""Check consistency of all history of given revision_ids.
1521
Different repository implementations should override _check().
1523
:param revision_ids: A non-empty list of revision_ids whose ancestry
1524
will be checked. Typically the last revision_id of a branch.
1526
return self._check(revision_ids)
1528
def _check(self, revision_ids):
1529
result = check.Check(self)
1533
def _warn_if_deprecated(self):
1534
global _deprecation_warning_done
1535
if _deprecation_warning_done:
1537
_deprecation_warning_done = True
1538
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1539
% (self._format, self.bzrdir.transport.base))
1541
def supports_rich_root(self):
1542
return self._format.rich_root_data
1544
def _check_ascii_revisionid(self, revision_id, method):
1545
"""Private helper for ascii-only repositories."""
1546
# weave repositories refuse to store revisionids that are non-ascii.
1547
if revision_id is not None:
1548
# weaves require ascii revision ids.
1549
if isinstance(revision_id, unicode):
1551
revision_id.encode('ascii')
1552
except UnicodeEncodeError:
1553
raise errors.NonAsciiRevisionId(method, self)
1556
revision_id.decode('ascii')
1557
except UnicodeDecodeError:
1558
raise errors.NonAsciiRevisionId(method, self)
1560
def revision_graph_can_have_wrong_parents(self):
1561
"""Is it possible for this repository to have a revision graph with
1564
If True, then this repository must also implement
1565
_find_inconsistent_revision_parents so that check and reconcile can
1566
check for inconsistencies before proceeding with other checks that may
1567
depend on the revision index being consistent.
1569
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1571
# remove these delegates a while after bzr 0.15
1572
def __make_delegated(name, from_module):
1573
def _deprecated_repository_forwarder():
1574
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1575
% (name, from_module),
1578
m = __import__(from_module, globals(), locals(), [name])
1580
return getattr(m, name)
1581
except AttributeError:
1582
raise AttributeError('module %s has no name %s'
1584
globals()[name] = _deprecated_repository_forwarder
1587
'AllInOneRepository',
1588
'WeaveMetaDirRepository',
1589
'PreSplitOutRepositoryFormat',
1590
'RepositoryFormat4',
1591
'RepositoryFormat5',
1592
'RepositoryFormat6',
1593
'RepositoryFormat7',
1595
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1599
'RepositoryFormatKnit',
1600
'RepositoryFormatKnit1',
1602
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1605
def install_revision(repository, rev, revision_tree):
1606
"""Install all revision data into a repository."""
1607
present_parents = []
1609
for p_id in rev.parent_ids:
1610
if repository.has_revision(p_id):
1611
present_parents.append(p_id)
1612
parent_trees[p_id] = repository.revision_tree(p_id)
1614
parent_trees[p_id] = repository.revision_tree(None)
1616
inv = revision_tree.inventory
1617
entries = inv.iter_entries()
1618
# backwards compatibility hack: skip the root id.
1619
if not repository.supports_rich_root():
1620
path, root = entries.next()
1621
if root.revision != rev.revision_id:
1622
raise errors.IncompatibleRevision(repr(repository))
1623
# Add the texts that are not already present
1624
for path, ie in entries:
1625
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1626
repository.get_transaction())
1627
if ie.revision not in w:
1629
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1630
# with InventoryEntry.find_previous_heads(). if it is, then there
1631
# is a latent bug here where the parents may have ancestors of each
1633
for revision, tree in parent_trees.iteritems():
1634
if ie.file_id not in tree:
1636
parent_id = tree.inventory[ie.file_id].revision
1637
if parent_id in text_parents:
1639
text_parents.append(parent_id)
1641
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1642
repository.get_transaction())
1643
lines = revision_tree.get_file(ie.file_id).readlines()
1644
vfile.add_lines(rev.revision_id, text_parents, lines)
1646
# install the inventory
1647
repository.add_inventory(rev.revision_id, inv, present_parents)
1648
except errors.RevisionAlreadyPresent:
1650
repository.add_revision(rev.revision_id, rev, inv)
1653
class MetaDirRepository(Repository):
1654
"""Repositories in the new meta-dir layout."""
1656
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1657
super(MetaDirRepository, self).__init__(_format,
1663
dir_mode = self.control_files._dir_mode
1664
file_mode = self.control_files._file_mode
1667
def is_shared(self):
1668
"""Return True if this repository is flagged as a shared repository."""
1669
return self.control_files._transport.has('shared-storage')
1672
def set_make_working_trees(self, new_value):
1673
"""Set the policy flag for making working trees when creating branches.
1675
This only applies to branches that use this repository.
1677
The default is 'True'.
1678
:param new_value: True to restore the default, False to disable making
1683
self.control_files._transport.delete('no-working-trees')
1684
except errors.NoSuchFile:
1687
self.control_files.put_utf8('no-working-trees', '')
1689
def make_working_trees(self):
1690
"""Returns the policy for making working trees on new branches."""
1691
return not self.control_files._transport.has('no-working-trees')
1694
class RepositoryFormatRegistry(registry.Registry):
1695
"""Registry of RepositoryFormats."""
1697
def get(self, format_string):
1698
r = registry.Registry.get(self, format_string)
1704
format_registry = RepositoryFormatRegistry()
1705
"""Registry of formats, indexed by their identifying format string.
1707
This can contain either format instances themselves, or classes/factories that
1708
can be called to obtain one.
1712
#####################################################################
1713
# Repository Formats
1715
class RepositoryFormat(object):
1716
"""A repository format.
1718
Formats provide three things:
1719
* An initialization routine to construct repository data on disk.
1720
* a format string which is used when the BzrDir supports versioned
1722
* an open routine which returns a Repository instance.
1724
There is one and only one Format subclass for each on-disk format. But
1725
there can be one Repository subclass that is used for several different
1726
formats. The _format attribute on a Repository instance can be used to
1727
determine the disk format.
1729
Formats are placed in an dict by their format string for reference
1730
during opening. These should be subclasses of RepositoryFormat
1733
Once a format is deprecated, just deprecate the initialize and open
1734
methods on the format class. Do not deprecate the object, as the
1735
object will be created every system load.
1737
Common instance attributes:
1738
_matchingbzrdir - the bzrdir format that the repository format was
1739
originally written to work with. This can be used if manually
1740
constructing a bzrdir and repository, or more commonly for test suite
1745
return "<%s>" % self.__class__.__name__
1747
def __eq__(self, other):
1748
# format objects are generally stateless
1749
return isinstance(other, self.__class__)
1751
def __ne__(self, other):
1752
return not self == other
1755
def find_format(klass, a_bzrdir):
1756
"""Return the format for the repository object in a_bzrdir.
1758
This is used by bzr native formats that have a "format" file in
1759
the repository. Other methods may be used by different types of
1763
transport = a_bzrdir.get_repository_transport(None)
1764
format_string = transport.get("format").read()
1765
return format_registry.get(format_string)
1766
except errors.NoSuchFile:
1767
raise errors.NoRepositoryPresent(a_bzrdir)
1769
raise errors.UnknownFormatError(format=format_string)
1772
def register_format(klass, format):
1773
format_registry.register(format.get_format_string(), format)
1776
def unregister_format(klass, format):
1777
format_registry.remove(format.get_format_string())
1780
def get_default_format(klass):
1781
"""Return the current default format."""
1782
from bzrlib import bzrdir
1783
return bzrdir.format_registry.make_bzrdir('default').repository_format
1785
def _get_control_store(self, repo_transport, control_files):
1786
"""Return the control store for this repository."""
1787
raise NotImplementedError(self._get_control_store)
1789
def get_format_string(self):
1790
"""Return the ASCII format string that identifies this format.
1792
Note that in pre format ?? repositories the format string is
1793
not permitted nor written to disk.
1795
raise NotImplementedError(self.get_format_string)
1797
def get_format_description(self):
1798
"""Return the short description for this format."""
1799
raise NotImplementedError(self.get_format_description)
1801
def _get_revision_store(self, repo_transport, control_files):
1802
"""Return the revision store object for this a_bzrdir."""
1803
raise NotImplementedError(self._get_revision_store)
1805
def _get_text_rev_store(self,
1812
"""Common logic for getting a revision store for a repository.
1814
see self._get_revision_store for the subclass-overridable method to
1815
get the store for a repository.
1817
from bzrlib.store.revision.text import TextRevisionStore
1818
dir_mode = control_files._dir_mode
1819
file_mode = control_files._file_mode
1820
text_store = TextStore(transport.clone(name),
1822
compressed=compressed,
1824
file_mode=file_mode)
1825
_revision_store = TextRevisionStore(text_store, serializer)
1826
return _revision_store
1828
# TODO: this shouldn't be in the base class, it's specific to things that
1829
# use weaves or knits -- mbp 20070207
1830
def _get_versioned_file_store(self,
1835
versionedfile_class=None,
1836
versionedfile_kwargs={},
1838
if versionedfile_class is None:
1839
versionedfile_class = self._versionedfile_class
1840
weave_transport = control_files._transport.clone(name)
1841
dir_mode = control_files._dir_mode
1842
file_mode = control_files._file_mode
1843
return VersionedFileStore(weave_transport, prefixed=prefixed,
1845
file_mode=file_mode,
1846
versionedfile_class=versionedfile_class,
1847
versionedfile_kwargs=versionedfile_kwargs,
1850
def initialize(self, a_bzrdir, shared=False):
1851
"""Initialize a repository of this format in a_bzrdir.
1853
:param a_bzrdir: The bzrdir to put the new repository in it.
1854
:param shared: The repository should be initialized as a sharable one.
1855
:returns: The new repository object.
1857
This may raise UninitializableFormat if shared repository are not
1858
compatible the a_bzrdir.
1860
raise NotImplementedError(self.initialize)
1862
def is_supported(self):
1863
"""Is this format supported?
1865
Supported formats must be initializable and openable.
1866
Unsupported formats may not support initialization or committing or
1867
some other features depending on the reason for not being supported.
1871
def check_conversion_target(self, target_format):
1872
raise NotImplementedError(self.check_conversion_target)
1874
def open(self, a_bzrdir, _found=False):
1875
"""Return an instance of this format for the bzrdir a_bzrdir.
1877
_found is a private parameter, do not use it.
1879
raise NotImplementedError(self.open)
1882
class MetaDirRepositoryFormat(RepositoryFormat):
1883
"""Common base class for the new repositories using the metadir layout."""
1885
rich_root_data = False
1886
supports_tree_reference = False
1887
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1890
super(MetaDirRepositoryFormat, self).__init__()
1892
def _create_control_files(self, a_bzrdir):
1893
"""Create the required files and the initial control_files object."""
1894
# FIXME: RBC 20060125 don't peek under the covers
1895
# NB: no need to escape relative paths that are url safe.
1896
repository_transport = a_bzrdir.get_repository_transport(self)
1897
control_files = lockable_files.LockableFiles(repository_transport,
1898
'lock', lockdir.LockDir)
1899
control_files.create_lock()
1900
return control_files
1902
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1903
"""Upload the initial blank content."""
1904
control_files = self._create_control_files(a_bzrdir)
1905
control_files.lock_write()
1907
control_files._transport.mkdir_multi(dirs,
1908
mode=control_files._dir_mode)
1909
for file, content in files:
1910
control_files.put(file, content)
1911
for file, content in utf8_files:
1912
control_files.put_utf8(file, content)
1914
control_files.put_utf8('shared-storage', '')
1916
control_files.unlock()
1919
# formats which have no format string are not discoverable
1920
# and not independently creatable, so are not registered. They're
1921
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1922
# needed, it's constructed directly by the BzrDir. Non-native formats where
1923
# the repository is not separately opened are similar.
1925
format_registry.register_lazy(
1926
'Bazaar-NG Repository format 7',
1927
'bzrlib.repofmt.weaverepo',
1931
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1932
# default control directory format
1933
format_registry.register_lazy(
1934
'Bazaar-NG Knit Repository Format 1',
1935
'bzrlib.repofmt.knitrepo',
1936
'RepositoryFormatKnit1',
1938
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1940
format_registry.register_lazy(
1941
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1942
'bzrlib.repofmt.knitrepo',
1943
'RepositoryFormatKnit3',
1947
class InterRepository(InterObject):
1948
"""This class represents operations taking place between two repositories.
1950
Its instances have methods like copy_content and fetch, and contain
1951
references to the source and target repositories these operations can be
1954
Often we will provide convenience methods on 'repository' which carry out
1955
operations with another repository - they will always forward to
1956
InterRepository.get(other).method_name(parameters).
1960
"""The available optimised InterRepository types."""
1962
def copy_content(self, revision_id=None):
1963
raise NotImplementedError(self.copy_content)
1965
def fetch(self, revision_id=None, pb=None):
1966
"""Fetch the content required to construct revision_id.
1968
The content is copied from self.source to self.target.
1970
:param revision_id: if None all content is copied, if NULL_REVISION no
1972
:param pb: optional progress bar to use for progress reports. If not
1973
provided a default one will be created.
1975
Returns the copied revision count and the failed revisions in a tuple:
1978
raise NotImplementedError(self.fetch)
1981
def missing_revision_ids(self, revision_id=None):
1982
"""Return the revision ids that source has that target does not.
1984
These are returned in topological order.
1986
:param revision_id: only return revision ids included by this
1989
# generic, possibly worst case, slow code path.
1990
target_ids = set(self.target.all_revision_ids())
1991
if revision_id is not None:
1992
source_ids = self.source.get_ancestry(revision_id)
1993
assert source_ids[0] is None
1996
source_ids = self.source.all_revision_ids()
1997
result_set = set(source_ids).difference(target_ids)
1998
# this may look like a no-op: its not. It preserves the ordering
1999
# other_ids had while only returning the members from other_ids
2000
# that we've decided we need.
2001
return [rev_id for rev_id in source_ids if rev_id in result_set]
2004
def _same_model(source, target):
2005
"""True if source and target have the same data representation."""
2006
if source.supports_rich_root() != target.supports_rich_root():
2008
if source._serializer != target._serializer:
2013
class InterSameDataRepository(InterRepository):
2014
"""Code for converting between repositories that represent the same data.
2016
Data format and model must match for this to work.
2020
def _get_repo_format_to_test(self):
2021
"""Repository format for testing with.
2023
InterSameData can pull from subtree to subtree and from non-subtree to
2024
non-subtree, so we test this with the richest repository format.
2026
from bzrlib.repofmt import knitrepo
2027
return knitrepo.RepositoryFormatKnit3()
2030
def is_compatible(source, target):
2031
return InterRepository._same_model(source, target)
2034
def copy_content(self, revision_id=None):
2035
"""Make a complete copy of the content in self into destination.
2037
This copies both the repository's revision data, and configuration information
2038
such as the make_working_trees setting.
2040
This is a destructive operation! Do not use it on existing
2043
:param revision_id: Only copy the content needed to construct
2044
revision_id and its parents.
2047
self.target.set_make_working_trees(self.source.make_working_trees())
2048
except NotImplementedError:
2050
# but don't bother fetching if we have the needed data now.
2051
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2052
self.target.has_revision(revision_id)):
2054
self.target.fetch(self.source, revision_id=revision_id)
2057
def fetch(self, revision_id=None, pb=None):
2058
"""See InterRepository.fetch()."""
2059
from bzrlib.fetch import GenericRepoFetcher
2060
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2061
self.source, self.source._format, self.target,
2062
self.target._format)
2063
f = GenericRepoFetcher(to_repository=self.target,
2064
from_repository=self.source,
2065
last_revision=revision_id,
2067
return f.count_copied, f.failed_revisions
2070
class InterWeaveRepo(InterSameDataRepository):
2071
"""Optimised code paths between Weave based repositories.
2073
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2074
implemented lazy inter-object optimisation.
2078
def _get_repo_format_to_test(self):
2079
from bzrlib.repofmt import weaverepo
2080
return weaverepo.RepositoryFormat7()
2083
def is_compatible(source, target):
2084
"""Be compatible with known Weave formats.
2086
We don't test for the stores being of specific types because that
2087
could lead to confusing results, and there is no need to be
2090
from bzrlib.repofmt.weaverepo import (
2096
return (isinstance(source._format, (RepositoryFormat5,
2098
RepositoryFormat7)) and
2099
isinstance(target._format, (RepositoryFormat5,
2101
RepositoryFormat7)))
2102
except AttributeError:
2106
def copy_content(self, revision_id=None):
2107
"""See InterRepository.copy_content()."""
2108
# weave specific optimised path:
2110
self.target.set_make_working_trees(self.source.make_working_trees())
2111
except NotImplementedError:
2113
# FIXME do not peek!
2114
if self.source.control_files._transport.listable():
2115
pb = ui.ui_factory.nested_progress_bar()
2117
self.target.weave_store.copy_all_ids(
2118
self.source.weave_store,
2120
from_transaction=self.source.get_transaction(),
2121
to_transaction=self.target.get_transaction())
2122
pb.update('copying inventory', 0, 1)
2123
self.target.control_weaves.copy_multi(
2124
self.source.control_weaves, ['inventory'],
2125
from_transaction=self.source.get_transaction(),
2126
to_transaction=self.target.get_transaction())
2127
self.target._revision_store.text_store.copy_all_ids(
2128
self.source._revision_store.text_store,
2133
self.target.fetch(self.source, revision_id=revision_id)
2136
def fetch(self, revision_id=None, pb=None):
2137
"""See InterRepository.fetch()."""
2138
from bzrlib.fetch import GenericRepoFetcher
2139
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2140
self.source, self.source._format, self.target, self.target._format)
2141
f = GenericRepoFetcher(to_repository=self.target,
2142
from_repository=self.source,
2143
last_revision=revision_id,
2145
return f.count_copied, f.failed_revisions
2148
def missing_revision_ids(self, revision_id=None):
2149
"""See InterRepository.missing_revision_ids()."""
2150
# we want all revisions to satisfy revision_id in source.
2151
# but we don't want to stat every file here and there.
2152
# we want then, all revisions other needs to satisfy revision_id
2153
# checked, but not those that we have locally.
2154
# so the first thing is to get a subset of the revisions to
2155
# satisfy revision_id in source, and then eliminate those that
2156
# we do already have.
2157
# this is slow on high latency connection to self, but as as this
2158
# disk format scales terribly for push anyway due to rewriting
2159
# inventory.weave, this is considered acceptable.
2161
if revision_id is not None:
2162
source_ids = self.source.get_ancestry(revision_id)
2163
assert source_ids[0] is None
2166
source_ids = self.source._all_possible_ids()
2167
source_ids_set = set(source_ids)
2168
# source_ids is the worst possible case we may need to pull.
2169
# now we want to filter source_ids against what we actually
2170
# have in target, but don't try to check for existence where we know
2171
# we do not have a revision as that would be pointless.
2172
target_ids = set(self.target._all_possible_ids())
2173
possibly_present_revisions = target_ids.intersection(source_ids_set)
2174
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2175
required_revisions = source_ids_set.difference(actually_present_revisions)
2176
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2177
if revision_id is not None:
2178
# we used get_ancestry to determine source_ids then we are assured all
2179
# revisions referenced are present as they are installed in topological order.
2180
# and the tip revision was validated by get_ancestry.
2181
return required_topo_revisions
2183
# if we just grabbed the possibly available ids, then
2184
# we only have an estimate of whats available and need to validate
2185
# that against the revision records.
2186
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2189
class InterKnitRepo(InterSameDataRepository):
2190
"""Optimised code paths between Knit based repositories."""
2193
def _get_repo_format_to_test(self):
2194
from bzrlib.repofmt import knitrepo
2195
return knitrepo.RepositoryFormatKnit1()
2198
def is_compatible(source, target):
2199
"""Be compatible with known Knit formats.
2201
We don't test for the stores being of specific types because that
2202
could lead to confusing results, and there is no need to be
2205
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2207
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2208
isinstance(target._format, RepositoryFormatKnit))
2209
except AttributeError:
2211
return are_knits and InterRepository._same_model(source, target)
2214
def fetch(self, revision_id=None, pb=None):
2215
"""See InterRepository.fetch()."""
2216
from bzrlib.fetch import KnitRepoFetcher
2217
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2218
self.source, self.source._format, self.target, self.target._format)
2219
f = KnitRepoFetcher(to_repository=self.target,
2220
from_repository=self.source,
2221
last_revision=revision_id,
2223
return f.count_copied, f.failed_revisions
2226
def missing_revision_ids(self, revision_id=None):
2227
"""See InterRepository.missing_revision_ids()."""
2228
if revision_id is not None:
2229
source_ids = self.source.get_ancestry(revision_id)
2230
assert source_ids[0] is None
2233
source_ids = self.source.all_revision_ids()
2234
source_ids_set = set(source_ids)
2235
# source_ids is the worst possible case we may need to pull.
2236
# now we want to filter source_ids against what we actually
2237
# have in target, but don't try to check for existence where we know
2238
# we do not have a revision as that would be pointless.
2239
target_ids = set(self.target.all_revision_ids())
2240
possibly_present_revisions = target_ids.intersection(source_ids_set)
2241
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2242
required_revisions = source_ids_set.difference(actually_present_revisions)
2243
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2244
if revision_id is not None:
2245
# we used get_ancestry to determine source_ids then we are assured all
2246
# revisions referenced are present as they are installed in topological order.
2247
# and the tip revision was validated by get_ancestry.
2248
return required_topo_revisions
2250
# if we just grabbed the possibly available ids, then
2251
# we only have an estimate of whats available and need to validate
2252
# that against the revision records.
2253
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2256
class InterModel1and2(InterRepository):
2259
def _get_repo_format_to_test(self):
2263
def is_compatible(source, target):
2264
if not source.supports_rich_root() and target.supports_rich_root():
2270
def fetch(self, revision_id=None, pb=None):
2271
"""See InterRepository.fetch()."""
2272
from bzrlib.fetch import Model1toKnit2Fetcher
2273
f = Model1toKnit2Fetcher(to_repository=self.target,
2274
from_repository=self.source,
2275
last_revision=revision_id,
2277
return f.count_copied, f.failed_revisions
2280
def copy_content(self, revision_id=None):
2281
"""Make a complete copy of the content in self into destination.
2283
This is a destructive operation! Do not use it on existing
2286
:param revision_id: Only copy the content needed to construct
2287
revision_id and its parents.
2290
self.target.set_make_working_trees(self.source.make_working_trees())
2291
except NotImplementedError:
2293
# but don't bother fetching if we have the needed data now.
2294
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2295
self.target.has_revision(revision_id)):
2297
self.target.fetch(self.source, revision_id=revision_id)
2300
class InterKnit1and2(InterKnitRepo):
2303
def _get_repo_format_to_test(self):
2307
def is_compatible(source, target):
2308
"""Be compatible with Knit1 source and Knit3 target"""
2309
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2311
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
2312
RepositoryFormatKnit3
2313
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2314
isinstance(target._format, (RepositoryFormatKnit3)))
2315
except AttributeError:
2319
def fetch(self, revision_id=None, pb=None):
2320
"""See InterRepository.fetch()."""
2321
from bzrlib.fetch import Knit1to2Fetcher
2322
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2323
self.source, self.source._format, self.target,
2324
self.target._format)
2325
f = Knit1to2Fetcher(to_repository=self.target,
2326
from_repository=self.source,
2327
last_revision=revision_id,
2329
return f.count_copied, f.failed_revisions
2332
class InterRemoteToOther(InterRepository):
2334
def __init__(self, source, target):
2335
InterRepository.__init__(self, source, target)
2336
self._real_inter = None
2339
def is_compatible(source, target):
2340
if not isinstance(source, remote.RemoteRepository):
2342
source._ensure_real()
2343
real_source = source._real_repository
2344
# Is source's model compatible with target's model, and are they the
2345
# same format? Currently we can only optimise fetching from an
2346
# identical model & format repo.
2347
assert not isinstance(real_source, remote.RemoteRepository), (
2348
"We don't support remote repos backed by remote repos yet.")
2349
return real_source._format == target._format
2352
def fetch(self, revision_id=None, pb=None):
2353
"""See InterRepository.fetch()."""
2354
from bzrlib.fetch import RemoteToOtherFetcher
2355
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2356
self.source, self.target, self.target._format)
2357
# TODO: jam 20070210 This should be an assert, not a translate
2358
revision_id = osutils.safe_revision_id(revision_id)
2359
f = RemoteToOtherFetcher(to_repository=self.target,
2360
from_repository=self.source,
2361
last_revision=revision_id,
2363
return f.count_copied, f.failed_revisions
2366
def _get_repo_format_to_test(self):
2370
class InterOtherToRemote(InterRepository):
2372
def __init__(self, source, target):
2373
InterRepository.__init__(self, source, target)
2374
self._real_inter = None
2377
def is_compatible(source, target):
2378
if isinstance(target, remote.RemoteRepository):
2382
def _ensure_real_inter(self):
2383
if self._real_inter is None:
2384
self.target._ensure_real()
2385
real_target = self.target._real_repository
2386
self._real_inter = InterRepository.get(self.source, real_target)
2388
def copy_content(self, revision_id=None):
2389
self._ensure_real_inter()
2390
self._real_inter.copy_content(revision_id=revision_id)
2392
def fetch(self, revision_id=None, pb=None):
2393
self._ensure_real_inter()
2394
self._real_inter.fetch(revision_id=revision_id, pb=pb)
2397
def _get_repo_format_to_test(self):
2401
InterRepository.register_optimiser(InterSameDataRepository)
2402
InterRepository.register_optimiser(InterWeaveRepo)
2403
InterRepository.register_optimiser(InterKnitRepo)
2404
InterRepository.register_optimiser(InterModel1and2)
2405
InterRepository.register_optimiser(InterKnit1and2)
2406
InterRepository.register_optimiser(InterRemoteToOther)
2407
InterRepository.register_optimiser(InterOtherToRemote)
2410
class CopyConverter(object):
2411
"""A repository conversion tool which just performs a copy of the content.
2413
This is slow but quite reliable.
2416
def __init__(self, target_format):
2417
"""Create a CopyConverter.
2419
:param target_format: The format the resulting repository should be.
2421
self.target_format = target_format
2423
def convert(self, repo, pb):
2424
"""Perform the conversion of to_convert, giving feedback via pb.
2426
:param to_convert: The disk object to convert.
2427
:param pb: a progress bar to use for progress information.
2432
# this is only useful with metadir layouts - separated repo content.
2433
# trigger an assertion if not such
2434
repo._format.get_format_string()
2435
self.repo_dir = repo.bzrdir
2436
self.step('Moving repository to repository.backup')
2437
self.repo_dir.transport.move('repository', 'repository.backup')
2438
backup_transport = self.repo_dir.transport.clone('repository.backup')
2439
repo._format.check_conversion_target(self.target_format)
2440
self.source_repo = repo._format.open(self.repo_dir,
2442
_override_transport=backup_transport)
2443
self.step('Creating new repository')
2444
converted = self.target_format.initialize(self.repo_dir,
2445
self.source_repo.is_shared())
2446
converted.lock_write()
2448
self.step('Copying content into repository.')
2449
self.source_repo.copy_content_into(converted)
2452
self.step('Deleting old repository content.')
2453
self.repo_dir.transport.delete_tree('repository.backup')
2454
self.pb.note('repository converted')
2456
def step(self, message):
2457
"""Update the pb by a step."""
2459
self.pb.update(message, self.count, self.total)
2471
def _unescaper(match, _map=_unescape_map):
2472
code = match.group(1)
2476
if not code.startswith('#'):
2478
return unichr(int(code[1:])).encode('utf8')
2484
def _unescape_xml(data):
2485
"""Unescape predefined XML entities in a string of data."""
2487
if _unescape_re is None:
2488
_unescape_re = re.compile('\&([^;]*);')
2489
return _unescape_re.sub(_unescaper, data)
2492
class _RevisionTextVersionCache(object):
2493
"""A cache of the versionedfile versions for revision and file-id."""
2495
def __init__(self, repository):
2496
self.repository = repository
2497
self.revision_versions = {}
2498
self.revision_parents = {}
2499
self.repo_graph = self.repository.get_graph()
2502
def add_revision_text_versions(self, tree):
2503
"""Cache text version data from the supplied revision tree"""
2505
for path, entry in tree.iter_entries_by_dir():
2506
inv_revisions[entry.file_id] = entry.revision
2507
self.revision_versions[tree.get_revision_id()] = inv_revisions
2508
return inv_revisions
2510
def get_text_version(self, file_id, revision_id):
2511
"""Determine the text version for a given file-id and revision-id"""
2513
inv_revisions = self.revision_versions[revision_id]
2516
tree = self.repository.revision_tree(revision_id)
2517
except errors.RevisionNotPresent:
2518
self.revision_versions[revision_id] = inv_revisions = {}
2520
inv_revisions = self.add_revision_text_versions(tree)
2521
return inv_revisions.get(file_id)
2523
def prepopulate_revs(self, revision_ids):
2524
# Filter out versions that we don't have an inventory for, so that the
2525
# revision_trees() call won't fail.
2526
inv_weave = self.repository.get_inventory_weave()
2527
revs = [r for r in revision_ids if inv_weave.has_version(r)]
2528
# XXX: this loop is very similar to
2529
# bzrlib.fetch.Inter1and2Helper.iter_rev_trees.
2531
for tree in self.repository.revision_trees(revs[:100]):
2532
if tree.inventory.revision_id is None:
2533
tree.inventory.revision_id = tree.get_revision_id()
2534
self.add_revision_text_versions(tree)
2537
def get_parents(self, revision_id):
2539
return self.revision_parents[revision_id]
2541
parents = self.repository.get_parents([revision_id])[0]
2542
self.revision_parents[revision_id] = parents
2545
def heads(self, revision_ids):
2546
revision_ids = tuple(revision_ids)
2548
return self.rev_heads[revision_ids]
2550
heads = self.repo_graph.heads(revision_ids)
2551
self.rev_heads[revision_ids] = heads
2554
class VersionedFileChecker(object):
2556
def __init__(self, planned_revisions, revision_versions, repository):
2557
self.planned_revisions = planned_revisions
2558
self.revision_versions = revision_versions
2559
self.repository = repository
2561
def calculate_file_version_parents(self, revision_id, file_id):
2562
text_revision = self.revision_versions.get_text_version(
2563
file_id, revision_id)
2564
if text_revision is None:
2566
parents_of_text_revision = self.revision_versions.get_parents(
2568
parents_from_inventories = []
2569
for parent in parents_of_text_revision:
2570
if parent == _mod_revision.NULL_REVISION:
2572
introduced_in = self.revision_versions.get_text_version(file_id,
2574
if introduced_in is not None:
2575
parents_from_inventories.append(introduced_in)
2576
heads = set(self.revision_versions.heads(parents_from_inventories))
2578
for parent in parents_from_inventories:
2579
if parent in heads and parent not in new_parents:
2580
new_parents.append(parent)
2583
def check_file_version_parents(self, weave, file_id):
2585
for num, revision_id in enumerate(self.planned_revisions):
2586
correct_parents = self.calculate_file_version_parents(
2587
revision_id, file_id)
2588
if correct_parents is None:
2590
text_revision = self.revision_versions.get_text_version(
2591
file_id, revision_id)
2592
knit_parents = weave.get_parents(text_revision)
2593
if correct_parents != knit_parents:
2594
result[revision_id] = (knit_parents, correct_parents)
added
removed
bzrlib/repository.py
