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._heads = graph.HeadsCache(repository.get_graph()).heads
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
raise AssertionError('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
self.new_inventory.add(ie)
258
# TODO: slow, take it out of the inner loop.
260
basis_inv = parent_invs[0]
262
basis_inv = Inventory(root_id=None)
264
# ie.revision is always None if the InventoryEntry is considered
265
# for committing. We may record the previous parents revision if the
266
# content is actually unchanged against a sole head.
267
if ie.revision is not None:
268
if not self._versioned_root and path == '':
269
# repositories that do not version the root set the root's
270
# revision to the new commit even when no change occurs, and
271
# this masks when a change may have occurred against the basis,
272
# so calculate if one happened.
273
if ie.file_id in basis_inv:
274
delta = (basis_inv.id2path(ie.file_id), path,
278
delta = (None, path, ie.file_id, ie)
281
# we don't need to commit this, because the caller already
282
# determined that an existing revision of this file is
284
return None, (ie.revision == self._new_revision_id)
285
# XXX: Friction: parent_candidates should return a list not a dict
286
# so that we don't have to walk the inventories again.
287
parent_candiate_entries = ie.parent_candidates(parent_invs)
288
head_set = self._heads(parent_candiate_entries.keys())
290
for inv in parent_invs:
291
if ie.file_id in inv:
292
old_rev = inv[ie.file_id].revision
293
if old_rev in head_set:
294
heads.append(inv[ie.file_id].revision)
295
head_set.remove(inv[ie.file_id].revision)
298
# now we check to see if we need to write a new record to the
300
# We write a new entry unless there is one head to the ancestors, and
301
# the kind-derived content is unchanged.
303
# Cheapest check first: no ancestors, or more the one head in the
304
# ancestors, we write a new node.
308
# There is a single head, look it up for comparison
309
parent_entry = parent_candiate_entries[heads[0]]
310
# if the non-content specific data has changed, we'll be writing a
312
if (parent_entry.parent_id != ie.parent_id or
313
parent_entry.name != ie.name):
315
# now we need to do content specific checks:
317
# if the kind changed the content obviously has
318
if kind != parent_entry.kind:
321
assert content_summary[2] is not None, \
322
"Files must not have executable = None"
324
if (# if the file length changed we have to store:
325
parent_entry.text_size != content_summary[1] or
326
# if the exec bit has changed we have to store:
327
parent_entry.executable != content_summary[2]):
329
elif parent_entry.text_sha1 == content_summary[3]:
330
# all meta and content is unchanged (using a hash cache
331
# hit to check the sha)
332
ie.revision = parent_entry.revision
333
ie.text_size = parent_entry.text_size
334
ie.text_sha1 = parent_entry.text_sha1
335
ie.executable = parent_entry.executable
336
return self._get_delta(ie, basis_inv, path), False
338
# Either there is only a hash change(no hash cache entry,
339
# or same size content change), or there is no change on
341
# Provide the parent's hash to the store layer, so that the
342
# content is unchanged we will not store a new node.
343
nostore_sha = parent_entry.text_sha1
345
# We want to record a new node regardless of the presence or
346
# absence of a content change in the file.
348
ie.executable = content_summary[2]
349
lines = tree.get_file(ie.file_id, path).readlines()
351
ie.text_sha1, ie.text_size = self._add_text_to_weave(
352
ie.file_id, lines, heads, nostore_sha)
353
except errors.ExistingContent:
354
# Turns out that the file content was unchanged, and we were
355
# only going to store a new node if it was changed. Carry over
357
ie.revision = parent_entry.revision
358
ie.text_size = parent_entry.text_size
359
ie.text_sha1 = parent_entry.text_sha1
360
ie.executable = parent_entry.executable
361
return self._get_delta(ie, basis_inv, path), False
362
elif kind == 'directory':
364
# all data is meta here, nothing specific to directory, so
366
ie.revision = parent_entry.revision
367
return self._get_delta(ie, basis_inv, path), False
369
self._add_text_to_weave(ie.file_id, lines, heads, None)
370
elif kind == 'symlink':
371
current_link_target = content_summary[3]
373
# symlink target is not generic metadata, check if it has
375
if current_link_target != parent_entry.symlink_target:
378
# unchanged, carry over.
379
ie.revision = parent_entry.revision
380
ie.symlink_target = parent_entry.symlink_target
381
return self._get_delta(ie, basis_inv, path), False
382
ie.symlink_target = current_link_target
384
self._add_text_to_weave(ie.file_id, lines, heads, None)
385
elif kind == 'tree-reference':
387
if content_summary[3] != parent_entry.reference_revision:
390
# unchanged, carry over.
391
ie.reference_revision = parent_entry.reference_revision
392
ie.revision = parent_entry.revision
393
return self._get_delta(ie, basis_inv, path), False
394
ie.reference_revision = content_summary[3]
396
self._add_text_to_weave(ie.file_id, lines, heads, None)
398
raise NotImplementedError('unknown kind')
399
ie.revision = self._new_revision_id
400
return self._get_delta(ie, basis_inv, path), True
402
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
403
versionedfile = self.repository.weave_store.get_weave_or_empty(
404
file_id, self.repository.get_transaction())
405
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
406
# than add_lines, and allows committing when a parent is ghosted for
408
# Note: as we read the content directly from the tree, we know its not
409
# been turned into unicode or badly split - but a broken tree
410
# implementation could give us bad output from readlines() so this is
411
# not a guarantee of safety. What would be better is always checking
412
# the content during test suite execution. RBC 20070912
414
return versionedfile.add_lines_with_ghosts(
415
self._new_revision_id, parents, new_lines,
416
nostore_sha=nostore_sha, random_id=self.random_revid,
417
check_content=False)[0:2]
419
versionedfile.clear_cache()
422
class RootCommitBuilder(CommitBuilder):
423
"""This commitbuilder actually records the root id"""
425
# the root entry gets versioned properly by this builder.
426
_versioned_root = True
428
def _check_root(self, ie, parent_invs, tree):
429
"""Helper for record_entry_contents.
431
:param ie: An entry being added.
432
:param parent_invs: The inventories of the parent revisions of the
434
:param tree: The tree that is being committed.
438
######################################################################
441
class Repository(object):
442
"""Repository holding history for one or more branches.
444
The repository holds and retrieves historical information including
445
revisions and file history. It's normally accessed only by the Branch,
446
which views a particular line of development through that history.
448
The Repository builds on top of Stores and a Transport, which respectively
449
describe the disk data format and the way of accessing the (possibly
453
# What class to use for a CommitBuilder. Often its simpler to change this
454
# in a Repository class subclass rather than to override
455
# get_commit_builder.
456
_commit_builder_class = CommitBuilder
457
# The search regex used by xml based repositories to determine what things
458
# where changed in a single commit.
459
_file_ids_altered_regex = lazy_regex.lazy_compile(
460
r'file_id="(?P<file_id>[^"]+)"'
461
r'.* revision="(?P<revision_id>[^"]+)"'
464
def abort_write_group(self):
465
"""Commit the contents accrued within the current write group.
467
:seealso: start_write_group.
469
if self._write_group is not self.get_transaction():
470
# has an unlock or relock occured ?
471
raise errors.BzrError('mismatched lock context and write group.')
472
self._abort_write_group()
473
self._write_group = None
475
def _abort_write_group(self):
476
"""Template method for per-repository write group cleanup.
478
This is called during abort before the write group is considered to be
479
finished and should cleanup any internal state accrued during the write
480
group. There is no requirement that data handed to the repository be
481
*not* made available - this is not a rollback - but neither should any
482
attempt be made to ensure that data added is fully commited. Abort is
483
invoked when an error has occured so futher disk or network operations
484
may not be possible or may error and if possible should not be
489
def add_inventory(self, revision_id, inv, parents):
490
"""Add the inventory inv to the repository as revision_id.
492
:param parents: The revision ids of the parents that revision_id
493
is known to have and are in the repository already.
495
returns the sha1 of the serialized inventory.
497
_mod_revision.check_not_reserved_id(revision_id)
498
assert inv.revision_id is None or inv.revision_id == revision_id, \
499
"Mismatch between inventory revision" \
500
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
501
assert inv.root is not None
502
inv_lines = self._serialise_inventory_to_lines(inv)
503
inv_vf = self.get_inventory_weave()
504
return self._inventory_add_lines(inv_vf, revision_id, parents,
505
inv_lines, check_content=False)
507
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
509
"""Store lines in inv_vf and return the sha1 of the inventory."""
511
for parent in parents:
513
final_parents.append(parent)
514
return inv_vf.add_lines(revision_id, final_parents, lines,
515
check_content=check_content)[0]
518
def add_revision(self, revision_id, rev, inv=None, config=None):
519
"""Add rev to the revision store as revision_id.
521
:param revision_id: the revision id to use.
522
:param rev: The revision object.
523
:param inv: The inventory for the revision. if None, it will be looked
524
up in the inventory storer
525
:param config: If None no digital signature will be created.
526
If supplied its signature_needed method will be used
527
to determine if a signature should be made.
529
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
531
_mod_revision.check_not_reserved_id(revision_id)
532
if config is not None and config.signature_needed():
534
inv = self.get_inventory(revision_id)
535
plaintext = Testament(rev, inv).as_short_text()
536
self.store_revision_signature(
537
gpg.GPGStrategy(config), plaintext, revision_id)
538
if not revision_id in self.get_inventory_weave():
540
raise errors.WeaveRevisionNotPresent(revision_id,
541
self.get_inventory_weave())
543
# yes, this is not suitable for adding with ghosts.
544
self.add_inventory(revision_id, inv, rev.parent_ids)
545
self._revision_store.add_revision(rev, self.get_transaction())
547
def _add_revision_text(self, revision_id, text):
548
revision = self._revision_store._serializer.read_revision_from_string(
550
self._revision_store._add_revision(revision, StringIO(text),
551
self.get_transaction())
553
def all_revision_ids(self):
554
"""Returns a list of all the revision ids in the repository.
556
This is deprecated because code should generally work on the graph
557
reachable from a particular revision, and ignore any other revisions
558
that might be present. There is no direct replacement method.
560
if 'evil' in debug.debug_flags:
561
mutter_callsite(2, "all_revision_ids is linear with history.")
562
return self._all_revision_ids()
564
def _all_revision_ids(self):
565
"""Returns a list of all the revision ids in the repository.
567
These are in as much topological order as the underlying store can
570
raise NotImplementedError(self._all_revision_ids)
572
def break_lock(self):
573
"""Break a lock if one is present from another instance.
575
Uses the ui factory to ask for confirmation if the lock may be from
578
self.control_files.break_lock()
581
def _eliminate_revisions_not_present(self, revision_ids):
582
"""Check every revision id in revision_ids to see if we have it.
584
Returns a set of the present revisions.
587
for id in revision_ids:
588
if self.has_revision(id):
593
def create(a_bzrdir):
594
"""Construct the current default format repository in a_bzrdir."""
595
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
597
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
598
"""instantiate a Repository.
600
:param _format: The format of the repository on disk.
601
:param a_bzrdir: The BzrDir of the repository.
603
In the future we will have a single api for all stores for
604
getting file texts, inventories and revisions, then
605
this construct will accept instances of those things.
607
super(Repository, self).__init__()
608
self._format = _format
609
# the following are part of the public API for Repository:
610
self.bzrdir = a_bzrdir
611
self.control_files = control_files
612
self._revision_store = _revision_store
613
# backwards compatibility
614
self.weave_store = text_store
616
self._reconcile_does_inventory_gc = True
617
self._reconcile_fixes_text_parents = False
618
# not right yet - should be more semantically clear ?
620
self.control_store = control_store
621
self.control_weaves = control_store
622
# TODO: make sure to construct the right store classes, etc, depending
623
# on whether escaping is required.
624
self._warn_if_deprecated()
625
self._write_group = None
626
self.base = control_files._transport.base
629
return '%s(%r)' % (self.__class__.__name__,
632
def has_same_location(self, other):
633
"""Returns a boolean indicating if this repository is at the same
634
location as another repository.
636
This might return False even when two repository objects are accessing
637
the same physical repository via different URLs.
639
if self.__class__ is not other.__class__:
641
return (self.control_files._transport.base ==
642
other.control_files._transport.base)
644
def is_in_write_group(self):
645
"""Return True if there is an open write group.
647
:seealso: start_write_group.
649
return self._write_group is not None
652
return self.control_files.is_locked()
654
def is_write_locked(self):
655
"""Return True if this object is write locked."""
656
return self.is_locked() and self.control_files._lock_mode == 'w'
658
def lock_write(self, token=None):
659
"""Lock this repository for writing.
661
This causes caching within the repository obejct to start accumlating
662
data during reads, and allows a 'write_group' to be obtained. Write
663
groups must be used for actual data insertion.
665
:param token: if this is already locked, then lock_write will fail
666
unless the token matches the existing lock.
667
:returns: a token if this instance supports tokens, otherwise None.
668
:raises TokenLockingNotSupported: when a token is given but this
669
instance doesn't support using token locks.
670
:raises MismatchedToken: if the specified token doesn't match the token
671
of the existing lock.
672
:seealso: start_write_group.
674
A token should be passed in if you know that you have locked the object
675
some other way, and need to synchronise this object's state with that
678
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
680
result = self.control_files.lock_write(token=token)
685
self.control_files.lock_read()
688
def get_physical_lock_status(self):
689
return self.control_files.get_physical_lock_status()
691
def leave_lock_in_place(self):
692
"""Tell this repository not to release the physical lock when this
695
If lock_write doesn't return a token, then this method is not supported.
697
self.control_files.leave_in_place()
699
def dont_leave_lock_in_place(self):
700
"""Tell this repository to release the physical lock when this
701
object is unlocked, even if it didn't originally acquire it.
703
If lock_write doesn't return a token, then this method is not supported.
705
self.control_files.dont_leave_in_place()
708
def gather_stats(self, revid=None, committers=None):
709
"""Gather statistics from a revision id.
711
:param revid: The revision id to gather statistics from, if None, then
712
no revision specific statistics are gathered.
713
:param committers: Optional parameter controlling whether to grab
714
a count of committers from the revision specific statistics.
715
:return: A dictionary of statistics. Currently this contains:
716
committers: The number of committers if requested.
717
firstrev: A tuple with timestamp, timezone for the penultimate left
718
most ancestor of revid, if revid is not the NULL_REVISION.
719
latestrev: A tuple with timestamp, timezone for revid, if revid is
720
not the NULL_REVISION.
721
revisions: The total revision count in the repository.
722
size: An estimate disk size of the repository in bytes.
725
if revid and committers:
726
result['committers'] = 0
727
if revid and revid != _mod_revision.NULL_REVISION:
729
all_committers = set()
730
revisions = self.get_ancestry(revid)
731
# pop the leading None
733
first_revision = None
735
# ignore the revisions in the middle - just grab first and last
736
revisions = revisions[0], revisions[-1]
737
for revision in self.get_revisions(revisions):
738
if not first_revision:
739
first_revision = revision
741
all_committers.add(revision.committer)
742
last_revision = revision
744
result['committers'] = len(all_committers)
745
result['firstrev'] = (first_revision.timestamp,
746
first_revision.timezone)
747
result['latestrev'] = (last_revision.timestamp,
748
last_revision.timezone)
750
# now gather global repository information
751
if self.bzrdir.root_transport.listable():
752
c, t = self._revision_store.total_size(self.get_transaction())
753
result['revisions'] = c
757
def get_data_stream(self, revision_ids):
758
raise NotImplementedError(self.get_data_stream)
760
def insert_data_stream(self, stream):
761
for item_key, bytes in stream:
762
if item_key[0] == 'file':
763
(file_id,) = item_key[1:]
764
knit = self.weave_store.get_weave_or_empty(
765
file_id, self.get_transaction())
766
elif item_key == ('inventory',):
767
knit = self.get_inventory_weave()
768
elif item_key == ('revisions',):
769
knit = self._revision_store.get_revision_file(
770
self.get_transaction())
771
elif item_key == ('signatures',):
772
knit = self._revision_store.get_signature_file(
773
self.get_transaction())
775
raise RepositoryDataStreamError(
776
"Unrecognised data stream key '%s'" % (item_key,))
777
decoded_list = bencode.bdecode(bytes)
778
format = decoded_list.pop(0)
781
for version, options, parents, some_bytes in decoded_list:
782
data_list.append((version, options, len(some_bytes), parents))
783
knit_bytes += some_bytes
784
knit.insert_data_stream(
785
(format, data_list, StringIO(knit_bytes).read))
788
def missing_revision_ids(self, other, revision_id=None):
789
"""Return the revision ids that other has that this does not.
791
These are returned in topological order.
793
revision_id: only return revision ids included by revision_id.
795
return InterRepository.get(other, self).missing_revision_ids(revision_id)
799
"""Open the repository rooted at base.
801
For instance, if the repository is at URL/.bzr/repository,
802
Repository.open(URL) -> a Repository instance.
804
control = bzrdir.BzrDir.open(base)
805
return control.open_repository()
807
def copy_content_into(self, destination, revision_id=None):
808
"""Make a complete copy of the content in self into destination.
810
This is a destructive operation! Do not use it on existing
813
return InterRepository.get(self, destination).copy_content(revision_id)
815
def commit_write_group(self):
816
"""Commit the contents accrued within the current write group.
818
:seealso: start_write_group.
820
if self._write_group is not self.get_transaction():
821
# has an unlock or relock occured ?
822
raise errors.BzrError('mismatched lock context %r and '
824
(self.get_transaction(), self._write_group))
825
self._commit_write_group()
826
self._write_group = None
828
def _commit_write_group(self):
829
"""Template method for per-repository write group cleanup.
831
This is called before the write group is considered to be
832
finished and should ensure that all data handed to the repository
833
for writing during the write group is safely committed (to the
834
extent possible considering file system caching etc).
837
def fetch(self, source, revision_id=None, pb=None):
838
"""Fetch the content required to construct revision_id from source.
840
If revision_id is None all content is copied.
842
# fast path same-url fetch operations
843
if self.has_same_location(source):
844
# check that last_revision is in 'from' and then return a
846
if (revision_id is not None and
847
not _mod_revision.is_null(revision_id)):
848
self.get_revision(revision_id)
850
inter = InterRepository.get(source, self)
852
return inter.fetch(revision_id=revision_id, pb=pb)
853
except NotImplementedError:
854
raise errors.IncompatibleRepositories(source, self)
856
def create_bundle(self, target, base, fileobj, format=None):
857
return serializer.write_bundle(self, target, base, fileobj, format)
859
def get_commit_builder(self, branch, parents, config, timestamp=None,
860
timezone=None, committer=None, revprops=None,
862
"""Obtain a CommitBuilder for this repository.
864
:param branch: Branch to commit to.
865
:param parents: Revision ids of the parents of the new revision.
866
:param config: Configuration to use.
867
:param timestamp: Optional timestamp recorded for commit.
868
:param timezone: Optional timezone for timestamp.
869
:param committer: Optional committer to set for commit.
870
:param revprops: Optional dictionary of revision properties.
871
:param revision_id: Optional revision id.
873
result = self._commit_builder_class(self, parents, config,
874
timestamp, timezone, committer, revprops, revision_id)
875
self.start_write_group()
879
if (self.control_files._lock_count == 1 and
880
self.control_files._lock_mode == 'w'):
881
if self._write_group is not None:
882
raise errors.BzrError(
883
'Must end write groups before releasing write locks.')
884
self.control_files.unlock()
887
def clone(self, a_bzrdir, revision_id=None):
888
"""Clone this repository into a_bzrdir using the current format.
890
Currently no check is made that the format of this repository and
891
the bzrdir format are compatible. FIXME RBC 20060201.
893
:return: The newly created destination repository.
895
# TODO: deprecate after 0.16; cloning this with all its settings is
896
# probably not very useful -- mbp 20070423
897
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
898
self.copy_content_into(dest_repo, revision_id)
901
def start_write_group(self):
902
"""Start a write group in the repository.
904
Write groups are used by repositories which do not have a 1:1 mapping
905
between file ids and backend store to manage the insertion of data from
906
both fetch and commit operations.
908
A write lock is required around the start_write_group/commit_write_group
909
for the support of lock-requiring repository formats.
911
One can only insert data into a repository inside a write group.
915
if not self.is_write_locked():
916
raise errors.NotWriteLocked(self)
917
if self._write_group:
918
raise errors.BzrError('already in a write group')
919
self._start_write_group()
920
# so we can detect unlock/relock - the write group is now entered.
921
self._write_group = self.get_transaction()
923
def _start_write_group(self):
924
"""Template method for per-repository write group startup.
926
This is called before the write group is considered to be
931
def sprout(self, to_bzrdir, revision_id=None):
932
"""Create a descendent repository for new development.
934
Unlike clone, this does not copy the settings of the repository.
936
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
937
dest_repo.fetch(self, revision_id=revision_id)
940
def _create_sprouting_repo(self, a_bzrdir, shared):
941
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
942
# use target default format.
943
dest_repo = a_bzrdir.create_repository()
945
# Most control formats need the repository to be specifically
946
# created, but on some old all-in-one formats it's not needed
948
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
949
except errors.UninitializableFormat:
950
dest_repo = a_bzrdir.open_repository()
954
def has_revision(self, revision_id):
955
"""True if this repository has a copy of the revision."""
956
if 'evil' in debug.debug_flags:
957
mutter_callsite(3, "has_revision is a LBYL symptom.")
958
return self._revision_store.has_revision_id(revision_id,
959
self.get_transaction())
962
def get_revision(self, revision_id):
963
"""Return the Revision object for a named revision."""
964
return self.get_revisions([revision_id])[0]
967
def get_revision_reconcile(self, revision_id):
968
"""'reconcile' helper routine that allows access to a revision always.
970
This variant of get_revision does not cross check the weave graph
971
against the revision one as get_revision does: but it should only
972
be used by reconcile, or reconcile-alike commands that are correcting
973
or testing the revision graph.
975
return self._get_revisions([revision_id])[0]
978
def get_revisions(self, revision_ids):
979
"""Get many revisions at once."""
980
return self._get_revisions(revision_ids)
983
def _get_revisions(self, revision_ids):
984
"""Core work logic to get many revisions without sanity checks."""
985
for rev_id in revision_ids:
986
if not rev_id or not isinstance(rev_id, basestring):
987
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
988
revs = self._revision_store.get_revisions(revision_ids,
989
self.get_transaction())
991
assert not isinstance(rev.revision_id, unicode)
992
for parent_id in rev.parent_ids:
993
assert not isinstance(parent_id, unicode)
997
def get_revision_xml(self, revision_id):
998
# TODO: jam 20070210 This shouldn't be necessary since get_revision
999
# would have already do it.
1000
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1001
rev = self.get_revision(revision_id)
1002
rev_tmp = StringIO()
1003
# the current serializer..
1004
self._revision_store._serializer.write_revision(rev, rev_tmp)
1006
return rev_tmp.getvalue()
1009
def get_deltas_for_revisions(self, revisions):
1010
"""Produce a generator of revision deltas.
1012
Note that the input is a sequence of REVISIONS, not revision_ids.
1013
Trees will be held in memory until the generator exits.
1014
Each delta is relative to the revision's lefthand predecessor.
1016
required_trees = set()
1017
for revision in revisions:
1018
required_trees.add(revision.revision_id)
1019
required_trees.update(revision.parent_ids[:1])
1020
trees = dict((t.get_revision_id(), t) for
1021
t in self.revision_trees(required_trees))
1022
for revision in revisions:
1023
if not revision.parent_ids:
1024
old_tree = self.revision_tree(None)
1026
old_tree = trees[revision.parent_ids[0]]
1027
yield trees[revision.revision_id].changes_from(old_tree)
1030
def get_revision_delta(self, revision_id):
1031
"""Return the delta for one revision.
1033
The delta is relative to the left-hand predecessor of the
1036
r = self.get_revision(revision_id)
1037
return list(self.get_deltas_for_revisions([r]))[0]
1040
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1041
signature = gpg_strategy.sign(plaintext)
1042
self._revision_store.add_revision_signature_text(revision_id,
1044
self.get_transaction())
1046
def fileids_altered_by_revision_ids(self, revision_ids):
1047
"""Find the file ids and versions affected by revisions.
1049
:param revisions: an iterable containing revision ids.
1050
:return: a dictionary mapping altered file-ids to an iterable of
1051
revision_ids. Each altered file-ids has the exact revision_ids that
1052
altered it listed explicitly.
1054
assert self._serializer.support_altered_by_hack, \
1055
("fileids_altered_by_revision_ids only supported for branches "
1056
"which store inventory as unnested xml, not on %r" % self)
1057
selected_revision_ids = set(revision_ids)
1058
w = self.get_inventory_weave()
1061
# this code needs to read every new line in every inventory for the
1062
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1063
# not present in one of those inventories is unnecessary but not
1064
# harmful because we are filtering by the revision id marker in the
1065
# inventory lines : we only select file ids altered in one of those
1066
# revisions. We don't need to see all lines in the inventory because
1067
# only those added in an inventory in rev X can contain a revision=X
1069
unescape_revid_cache = {}
1070
unescape_fileid_cache = {}
1072
# jam 20061218 In a big fetch, this handles hundreds of thousands
1073
# of lines, so it has had a lot of inlining and optimizing done.
1074
# Sorry that it is a little bit messy.
1075
# Move several functions to be local variables, since this is a long
1077
search = self._file_ids_altered_regex.search
1078
unescape = _unescape_xml
1079
setdefault = result.setdefault
1080
pb = ui.ui_factory.nested_progress_bar()
1082
for line in w.iter_lines_added_or_present_in_versions(
1083
selected_revision_ids, pb=pb):
1084
match = search(line)
1087
# One call to match.group() returning multiple items is quite a
1088
# bit faster than 2 calls to match.group() each returning 1
1089
file_id, revision_id = match.group('file_id', 'revision_id')
1091
# Inlining the cache lookups helps a lot when you make 170,000
1092
# lines and 350k ids, versus 8.4 unique ids.
1093
# Using a cache helps in 2 ways:
1094
# 1) Avoids unnecessary decoding calls
1095
# 2) Re-uses cached strings, which helps in future set and
1097
# (2) is enough that removing encoding entirely along with
1098
# the cache (so we are using plain strings) results in no
1099
# performance improvement.
1101
revision_id = unescape_revid_cache[revision_id]
1103
unescaped = unescape(revision_id)
1104
unescape_revid_cache[revision_id] = unescaped
1105
revision_id = unescaped
1107
if revision_id in selected_revision_ids:
1109
file_id = unescape_fileid_cache[file_id]
1111
unescaped = unescape(file_id)
1112
unescape_fileid_cache[file_id] = unescaped
1114
setdefault(file_id, set()).add(revision_id)
1119
def iter_files_bytes(self, desired_files):
1120
"""Iterate through file versions.
1122
Files will not necessarily be returned in the order they occur in
1123
desired_files. No specific order is guaranteed.
1125
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1126
value supplied by the caller as part of desired_files. It should
1127
uniquely identify the file version in the caller's context. (Examples:
1128
an index number or a TreeTransform trans_id.)
1130
bytes_iterator is an iterable of bytestrings for the file. The
1131
kind of iterable and length of the bytestrings are unspecified, but for
1132
this implementation, it is a list of lines produced by
1133
VersionedFile.get_lines().
1135
:param desired_files: a list of (file_id, revision_id, identifier)
1138
transaction = self.get_transaction()
1139
for file_id, revision_id, callable_data in desired_files:
1141
weave = self.weave_store.get_weave(file_id, transaction)
1142
except errors.NoSuchFile:
1143
raise errors.NoSuchIdInRepository(self, file_id)
1144
yield callable_data, weave.get_lines(revision_id)
1146
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1147
"""Get an iterable listing the keys of all the data introduced by a set
1150
The keys will be ordered so that the corresponding items can be safely
1151
fetched and inserted in that order.
1153
:returns: An iterable producing tuples of (knit-kind, file-id,
1154
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1155
'revisions'. file-id is None unless knit-kind is 'file'.
1157
# XXX: it's a bit weird to control the inventory weave caching in this
1158
# generator. Ideally the caching would be done in fetch.py I think. Or
1159
# maybe this generator should explicitly have the contract that it
1160
# should not be iterated until the previously yielded item has been
1163
inv_w = self.get_inventory_weave()
1164
inv_w.enable_cache()
1166
# file ids that changed
1167
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1169
num_file_ids = len(file_ids)
1170
for file_id, altered_versions in file_ids.iteritems():
1171
if _files_pb is not None:
1172
_files_pb.update("fetch texts", count, num_file_ids)
1174
yield ("file", file_id, altered_versions)
1175
# We're done with the files_pb. Note that it finished by the caller,
1176
# just as it was created by the caller.
1180
yield ("inventory", None, revision_ids)
1184
revisions_with_signatures = set()
1185
for rev_id in revision_ids:
1187
self.get_signature_text(rev_id)
1188
except errors.NoSuchRevision:
1192
revisions_with_signatures.add(rev_id)
1194
yield ("signatures", None, revisions_with_signatures)
1197
yield ("revisions", None, revision_ids)
1200
def get_inventory_weave(self):
1201
return self.control_weaves.get_weave('inventory',
1202
self.get_transaction())
1205
def get_inventory(self, revision_id):
1206
"""Get Inventory object by hash."""
1207
return self.deserialise_inventory(
1208
revision_id, self.get_inventory_xml(revision_id))
1210
def deserialise_inventory(self, revision_id, xml):
1211
"""Transform the xml into an inventory object.
1213
:param revision_id: The expected revision id of the inventory.
1214
:param xml: A serialised inventory.
1216
return self._serializer.read_inventory_from_string(xml, revision_id)
1218
def serialise_inventory(self, inv):
1219
return self._serializer.write_inventory_to_string(inv)
1221
def _serialise_inventory_to_lines(self, inv):
1222
return self._serializer.write_inventory_to_lines(inv)
1224
def get_serializer_format(self):
1225
return self._serializer.format_num
1228
def get_inventory_xml(self, revision_id):
1229
"""Get inventory XML as a file object."""
1231
assert isinstance(revision_id, str), type(revision_id)
1232
iw = self.get_inventory_weave()
1233
return iw.get_text(revision_id)
1235
raise errors.HistoryMissing(self, 'inventory', revision_id)
1238
def get_inventory_sha1(self, revision_id):
1239
"""Return the sha1 hash of the inventory entry
1241
return self.get_revision(revision_id).inventory_sha1
1244
def get_revision_graph(self, revision_id=None):
1245
"""Return a dictionary containing the revision graph.
1247
NB: This method should not be used as it accesses the entire graph all
1248
at once, which is much more data than most operations should require.
1250
:param revision_id: The revision_id to get a graph from. If None, then
1251
the entire revision graph is returned. This is a deprecated mode of
1252
operation and will be removed in the future.
1253
:return: a dictionary of revision_id->revision_parents_list.
1255
raise NotImplementedError(self.get_revision_graph)
1258
def get_revision_graph_with_ghosts(self, revision_ids=None):
1259
"""Return a graph of the revisions with ghosts marked as applicable.
1261
:param revision_ids: an iterable of revisions to graph or None for all.
1262
:return: a Graph object with the graph reachable from revision_ids.
1264
if 'evil' in debug.debug_flags:
1266
"get_revision_graph_with_ghosts scales with size of history.")
1267
result = deprecated_graph.Graph()
1268
if not revision_ids:
1269
pending = set(self.all_revision_ids())
1272
pending = set(revision_ids)
1273
# special case NULL_REVISION
1274
if _mod_revision.NULL_REVISION in pending:
1275
pending.remove(_mod_revision.NULL_REVISION)
1276
required = set(pending)
1279
revision_id = pending.pop()
1281
rev = self.get_revision(revision_id)
1282
except errors.NoSuchRevision:
1283
if revision_id in required:
1286
result.add_ghost(revision_id)
1288
for parent_id in rev.parent_ids:
1289
# is this queued or done ?
1290
if (parent_id not in pending and
1291
parent_id not in done):
1293
pending.add(parent_id)
1294
result.add_node(revision_id, rev.parent_ids)
1295
done.add(revision_id)
1298
def _get_history_vf(self):
1299
"""Get a versionedfile whose history graph reflects all revisions.
1301
For weave repositories, this is the inventory weave.
1303
return self.get_inventory_weave()
1305
def iter_reverse_revision_history(self, revision_id):
1306
"""Iterate backwards through revision ids in the lefthand history
1308
:param revision_id: The revision id to start with. All its lefthand
1309
ancestors will be traversed.
1311
if revision_id in (None, _mod_revision.NULL_REVISION):
1313
next_id = revision_id
1314
versionedfile = self._get_history_vf()
1317
parents = versionedfile.get_parents(next_id)
1318
if len(parents) == 0:
1321
next_id = parents[0]
1324
def get_revision_inventory(self, revision_id):
1325
"""Return inventory of a past revision."""
1326
# TODO: Unify this with get_inventory()
1327
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1328
# must be the same as its revision, so this is trivial.
1329
if revision_id is None:
1330
# This does not make sense: if there is no revision,
1331
# then it is the current tree inventory surely ?!
1332
# and thus get_root_id() is something that looks at the last
1333
# commit on the branch, and the get_root_id is an inventory check.
1334
raise NotImplementedError
1335
# return Inventory(self.get_root_id())
1337
return self.get_inventory(revision_id)
1340
def is_shared(self):
1341
"""Return True if this repository is flagged as a shared repository."""
1342
raise NotImplementedError(self.is_shared)
1345
def reconcile(self, other=None, thorough=False):
1346
"""Reconcile this repository."""
1347
from bzrlib.reconcile import RepoReconciler
1348
reconciler = RepoReconciler(self, thorough=thorough)
1349
reconciler.reconcile()
1352
def _refresh_data(self):
1353
"""Helper called from lock_* to ensure coherency with disk.
1355
The default implementation does nothing; it is however possible
1356
for repositories to maintain loaded indices across multiple locks
1357
by checking inside their implementation of this method to see
1358
whether their indices are still valid. This depends of course on
1359
the disk format being validatable in this manner.
1363
def revision_tree(self, revision_id):
1364
"""Return Tree for a revision on this branch.
1366
`revision_id` may be None for the empty tree revision.
1368
# TODO: refactor this to use an existing revision object
1369
# so we don't need to read it in twice.
1370
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1371
return RevisionTree(self, Inventory(root_id=None),
1372
_mod_revision.NULL_REVISION)
1374
inv = self.get_revision_inventory(revision_id)
1375
return RevisionTree(self, inv, revision_id)
1378
def revision_trees(self, revision_ids):
1379
"""Return Tree for a revision on this branch.
1381
`revision_id` may not be None or 'null:'"""
1382
assert None not in revision_ids
1383
assert _mod_revision.NULL_REVISION not in revision_ids
1384
texts = self.get_inventory_weave().get_texts(revision_ids)
1385
for text, revision_id in zip(texts, revision_ids):
1386
inv = self.deserialise_inventory(revision_id, text)
1387
yield RevisionTree(self, inv, revision_id)
1390
def get_ancestry(self, revision_id, topo_sorted=True):
1391
"""Return a list of revision-ids integrated by a revision.
1393
The first element of the list is always None, indicating the origin
1394
revision. This might change when we have history horizons, or
1395
perhaps we should have a new API.
1397
This is topologically sorted.
1399
if _mod_revision.is_null(revision_id):
1401
if not self.has_revision(revision_id):
1402
raise errors.NoSuchRevision(self, revision_id)
1403
w = self.get_inventory_weave()
1404
candidates = w.get_ancestry(revision_id, topo_sorted)
1405
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1408
"""Compress the data within the repository.
1410
This operation only makes sense for some repository types. For other
1411
types it should be a no-op that just returns.
1413
This stub method does not require a lock, but subclasses should use
1414
@needs_write_lock as this is a long running call its reasonable to
1415
implicitly lock for the user.
1419
def print_file(self, file, revision_id):
1420
"""Print `file` to stdout.
1422
FIXME RBC 20060125 as John Meinel points out this is a bad api
1423
- it writes to stdout, it assumes that that is valid etc. Fix
1424
by creating a new more flexible convenience function.
1426
tree = self.revision_tree(revision_id)
1427
# use inventory as it was in that revision
1428
file_id = tree.inventory.path2id(file)
1430
# TODO: jam 20060427 Write a test for this code path
1431
# it had a bug in it, and was raising the wrong
1433
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1434
tree.print_file(file_id)
1436
def get_transaction(self):
1437
return self.control_files.get_transaction()
1439
def revision_parents(self, revision_id):
1440
return self.get_inventory_weave().parent_names(revision_id)
1442
def get_parents(self, revision_ids):
1443
"""See StackedParentsProvider.get_parents"""
1445
for revision_id in revision_ids:
1446
if revision_id == _mod_revision.NULL_REVISION:
1450
parents = self.get_revision(revision_id).parent_ids
1451
except errors.NoSuchRevision:
1454
if len(parents) == 0:
1455
parents = [_mod_revision.NULL_REVISION]
1456
parents_list.append(parents)
1459
def _make_parents_provider(self):
1462
def get_graph(self, other_repository=None):
1463
"""Return the graph walker for this repository format"""
1464
parents_provider = self._make_parents_provider()
1465
if (other_repository is not None and
1466
other_repository.bzrdir.transport.base !=
1467
self.bzrdir.transport.base):
1468
parents_provider = graph._StackedParentsProvider(
1469
[parents_provider, other_repository._make_parents_provider()])
1470
return graph.Graph(parents_provider)
1472
def get_versioned_file_checker(self, revisions, revision_versions_cache):
1473
return VersionedFileChecker(revisions, revision_versions_cache, self)
1476
def set_make_working_trees(self, new_value):
1477
"""Set the policy flag for making working trees when creating branches.
1479
This only applies to branches that use this repository.
1481
The default is 'True'.
1482
:param new_value: True to restore the default, False to disable making
1485
raise NotImplementedError(self.set_make_working_trees)
1487
def make_working_trees(self):
1488
"""Returns the policy for making working trees on new branches."""
1489
raise NotImplementedError(self.make_working_trees)
1492
def sign_revision(self, revision_id, gpg_strategy):
1493
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1494
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1497
def has_signature_for_revision_id(self, revision_id):
1498
"""Query for a revision signature for revision_id in the repository."""
1499
return self._revision_store.has_signature(revision_id,
1500
self.get_transaction())
1503
def get_signature_text(self, revision_id):
1504
"""Return the text for a signature."""
1505
return self._revision_store.get_signature_text(revision_id,
1506
self.get_transaction())
1509
def check(self, revision_ids=None):
1510
"""Check consistency of all history of given revision_ids.
1512
Different repository implementations should override _check().
1514
:param revision_ids: A non-empty list of revision_ids whose ancestry
1515
will be checked. Typically the last revision_id of a branch.
1517
return self._check(revision_ids)
1519
def _check(self, revision_ids):
1520
result = check.Check(self)
1524
def _warn_if_deprecated(self):
1525
global _deprecation_warning_done
1526
if _deprecation_warning_done:
1528
_deprecation_warning_done = True
1529
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1530
% (self._format, self.bzrdir.transport.base))
1532
def supports_rich_root(self):
1533
return self._format.rich_root_data
1535
def _check_ascii_revisionid(self, revision_id, method):
1536
"""Private helper for ascii-only repositories."""
1537
# weave repositories refuse to store revisionids that are non-ascii.
1538
if revision_id is not None:
1539
# weaves require ascii revision ids.
1540
if isinstance(revision_id, unicode):
1542
revision_id.encode('ascii')
1543
except UnicodeEncodeError:
1544
raise errors.NonAsciiRevisionId(method, self)
1547
revision_id.decode('ascii')
1548
except UnicodeDecodeError:
1549
raise errors.NonAsciiRevisionId(method, self)
1551
def revision_graph_can_have_wrong_parents(self):
1552
"""Is it possible for this repository to have a revision graph with
1555
If True, then this repository must also implement
1556
_find_inconsistent_revision_parents so that check and reconcile can
1557
check for inconsistencies before proceeding with other checks that may
1558
depend on the revision index being consistent.
1560
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1562
# remove these delegates a while after bzr 0.15
1563
def __make_delegated(name, from_module):
1564
def _deprecated_repository_forwarder():
1565
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1566
% (name, from_module),
1569
m = __import__(from_module, globals(), locals(), [name])
1571
return getattr(m, name)
1572
except AttributeError:
1573
raise AttributeError('module %s has no name %s'
1575
globals()[name] = _deprecated_repository_forwarder
1578
'AllInOneRepository',
1579
'WeaveMetaDirRepository',
1580
'PreSplitOutRepositoryFormat',
1581
'RepositoryFormat4',
1582
'RepositoryFormat5',
1583
'RepositoryFormat6',
1584
'RepositoryFormat7',
1586
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1590
'RepositoryFormatKnit',
1591
'RepositoryFormatKnit1',
1593
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1596
def install_revision(repository, rev, revision_tree):
1597
"""Install all revision data into a repository."""
1598
present_parents = []
1600
for p_id in rev.parent_ids:
1601
if repository.has_revision(p_id):
1602
present_parents.append(p_id)
1603
parent_trees[p_id] = repository.revision_tree(p_id)
1605
parent_trees[p_id] = repository.revision_tree(None)
1607
inv = revision_tree.inventory
1608
entries = inv.iter_entries()
1609
# backwards compatibility hack: skip the root id.
1610
if not repository.supports_rich_root():
1611
path, root = entries.next()
1612
if root.revision != rev.revision_id:
1613
raise errors.IncompatibleRevision(repr(repository))
1614
# Add the texts that are not already present
1615
for path, ie in entries:
1616
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1617
repository.get_transaction())
1618
if ie.revision not in w:
1620
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1621
# with InventoryEntry.find_previous_heads(). if it is, then there
1622
# is a latent bug here where the parents may have ancestors of each
1624
for revision, tree in parent_trees.iteritems():
1625
if ie.file_id not in tree:
1627
parent_id = tree.inventory[ie.file_id].revision
1628
if parent_id in text_parents:
1630
text_parents.append(parent_id)
1632
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1633
repository.get_transaction())
1634
lines = revision_tree.get_file(ie.file_id).readlines()
1635
vfile.add_lines(rev.revision_id, text_parents, lines)
1637
# install the inventory
1638
repository.add_inventory(rev.revision_id, inv, present_parents)
1639
except errors.RevisionAlreadyPresent:
1641
repository.add_revision(rev.revision_id, rev, inv)
1644
class MetaDirRepository(Repository):
1645
"""Repositories in the new meta-dir layout."""
1647
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1648
super(MetaDirRepository, self).__init__(_format,
1654
dir_mode = self.control_files._dir_mode
1655
file_mode = self.control_files._file_mode
1658
def is_shared(self):
1659
"""Return True if this repository is flagged as a shared repository."""
1660
return self.control_files._transport.has('shared-storage')
1663
def set_make_working_trees(self, new_value):
1664
"""Set the policy flag for making working trees when creating branches.
1666
This only applies to branches that use this repository.
1668
The default is 'True'.
1669
:param new_value: True to restore the default, False to disable making
1674
self.control_files._transport.delete('no-working-trees')
1675
except errors.NoSuchFile:
1678
self.control_files.put_utf8('no-working-trees', '')
1680
def make_working_trees(self):
1681
"""Returns the policy for making working trees on new branches."""
1682
return not self.control_files._transport.has('no-working-trees')
1685
class RepositoryFormatRegistry(registry.Registry):
1686
"""Registry of RepositoryFormats."""
1688
def get(self, format_string):
1689
r = registry.Registry.get(self, format_string)
1695
format_registry = RepositoryFormatRegistry()
1696
"""Registry of formats, indexed by their identifying format string.
1698
This can contain either format instances themselves, or classes/factories that
1699
can be called to obtain one.
1703
#####################################################################
1704
# Repository Formats
1706
class RepositoryFormat(object):
1707
"""A repository format.
1709
Formats provide three things:
1710
* An initialization routine to construct repository data on disk.
1711
* a format string which is used when the BzrDir supports versioned
1713
* an open routine which returns a Repository instance.
1715
There is one and only one Format subclass for each on-disk format. But
1716
there can be one Repository subclass that is used for several different
1717
formats. The _format attribute on a Repository instance can be used to
1718
determine the disk format.
1720
Formats are placed in an dict by their format string for reference
1721
during opening. These should be subclasses of RepositoryFormat
1724
Once a format is deprecated, just deprecate the initialize and open
1725
methods on the format class. Do not deprecate the object, as the
1726
object will be created every system load.
1728
Common instance attributes:
1729
_matchingbzrdir - the bzrdir format that the repository format was
1730
originally written to work with. This can be used if manually
1731
constructing a bzrdir and repository, or more commonly for test suite
1736
return "<%s>" % self.__class__.__name__
1738
def __eq__(self, other):
1739
# format objects are generally stateless
1740
return isinstance(other, self.__class__)
1742
def __ne__(self, other):
1743
return not self == other
1746
def find_format(klass, a_bzrdir):
1747
"""Return the format for the repository object in a_bzrdir.
1749
This is used by bzr native formats that have a "format" file in
1750
the repository. Other methods may be used by different types of
1754
transport = a_bzrdir.get_repository_transport(None)
1755
format_string = transport.get("format").read()
1756
return format_registry.get(format_string)
1757
except errors.NoSuchFile:
1758
raise errors.NoRepositoryPresent(a_bzrdir)
1760
raise errors.UnknownFormatError(format=format_string)
1763
def register_format(klass, format):
1764
format_registry.register(format.get_format_string(), format)
1767
def unregister_format(klass, format):
1768
format_registry.remove(format.get_format_string())
1771
def get_default_format(klass):
1772
"""Return the current default format."""
1773
from bzrlib import bzrdir
1774
return bzrdir.format_registry.make_bzrdir('default').repository_format
1776
def _get_control_store(self, repo_transport, control_files):
1777
"""Return the control store for this repository."""
1778
raise NotImplementedError(self._get_control_store)
1780
def get_format_string(self):
1781
"""Return the ASCII format string that identifies this format.
1783
Note that in pre format ?? repositories the format string is
1784
not permitted nor written to disk.
1786
raise NotImplementedError(self.get_format_string)
1788
def get_format_description(self):
1789
"""Return the short description for this format."""
1790
raise NotImplementedError(self.get_format_description)
1792
def _get_revision_store(self, repo_transport, control_files):
1793
"""Return the revision store object for this a_bzrdir."""
1794
raise NotImplementedError(self._get_revision_store)
1796
def _get_text_rev_store(self,
1803
"""Common logic for getting a revision store for a repository.
1805
see self._get_revision_store for the subclass-overridable method to
1806
get the store for a repository.
1808
from bzrlib.store.revision.text import TextRevisionStore
1809
dir_mode = control_files._dir_mode
1810
file_mode = control_files._file_mode
1811
text_store = TextStore(transport.clone(name),
1813
compressed=compressed,
1815
file_mode=file_mode)
1816
_revision_store = TextRevisionStore(text_store, serializer)
1817
return _revision_store
1819
# TODO: this shouldn't be in the base class, it's specific to things that
1820
# use weaves or knits -- mbp 20070207
1821
def _get_versioned_file_store(self,
1826
versionedfile_class=None,
1827
versionedfile_kwargs={},
1829
if versionedfile_class is None:
1830
versionedfile_class = self._versionedfile_class
1831
weave_transport = control_files._transport.clone(name)
1832
dir_mode = control_files._dir_mode
1833
file_mode = control_files._file_mode
1834
return VersionedFileStore(weave_transport, prefixed=prefixed,
1836
file_mode=file_mode,
1837
versionedfile_class=versionedfile_class,
1838
versionedfile_kwargs=versionedfile_kwargs,
1841
def initialize(self, a_bzrdir, shared=False):
1842
"""Initialize a repository of this format in a_bzrdir.
1844
:param a_bzrdir: The bzrdir to put the new repository in it.
1845
:param shared: The repository should be initialized as a sharable one.
1846
:returns: The new repository object.
1848
This may raise UninitializableFormat if shared repository are not
1849
compatible the a_bzrdir.
1851
raise NotImplementedError(self.initialize)
1853
def is_supported(self):
1854
"""Is this format supported?
1856
Supported formats must be initializable and openable.
1857
Unsupported formats may not support initialization or committing or
1858
some other features depending on the reason for not being supported.
1862
def check_conversion_target(self, target_format):
1863
raise NotImplementedError(self.check_conversion_target)
1865
def open(self, a_bzrdir, _found=False):
1866
"""Return an instance of this format for the bzrdir a_bzrdir.
1868
_found is a private parameter, do not use it.
1870
raise NotImplementedError(self.open)
1873
class MetaDirRepositoryFormat(RepositoryFormat):
1874
"""Common base class for the new repositories using the metadir layout."""
1876
rich_root_data = False
1877
supports_tree_reference = False
1878
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1881
super(MetaDirRepositoryFormat, self).__init__()
1883
def _create_control_files(self, a_bzrdir):
1884
"""Create the required files and the initial control_files object."""
1885
# FIXME: RBC 20060125 don't peek under the covers
1886
# NB: no need to escape relative paths that are url safe.
1887
repository_transport = a_bzrdir.get_repository_transport(self)
1888
control_files = lockable_files.LockableFiles(repository_transport,
1889
'lock', lockdir.LockDir)
1890
control_files.create_lock()
1891
return control_files
1893
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1894
"""Upload the initial blank content."""
1895
control_files = self._create_control_files(a_bzrdir)
1896
control_files.lock_write()
1898
control_files._transport.mkdir_multi(dirs,
1899
mode=control_files._dir_mode)
1900
for file, content in files:
1901
control_files.put(file, content)
1902
for file, content in utf8_files:
1903
control_files.put_utf8(file, content)
1905
control_files.put_utf8('shared-storage', '')
1907
control_files.unlock()
1910
# formats which have no format string are not discoverable
1911
# and not independently creatable, so are not registered. They're
1912
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1913
# needed, it's constructed directly by the BzrDir. Non-native formats where
1914
# the repository is not separately opened are similar.
1916
format_registry.register_lazy(
1917
'Bazaar-NG Repository format 7',
1918
'bzrlib.repofmt.weaverepo',
1922
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1923
# default control directory format
1924
format_registry.register_lazy(
1925
'Bazaar-NG Knit Repository Format 1',
1926
'bzrlib.repofmt.knitrepo',
1927
'RepositoryFormatKnit1',
1929
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1931
format_registry.register_lazy(
1932
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1933
'bzrlib.repofmt.knitrepo',
1934
'RepositoryFormatKnit3',
1938
class InterRepository(InterObject):
1939
"""This class represents operations taking place between two repositories.
1941
Its instances have methods like copy_content and fetch, and contain
1942
references to the source and target repositories these operations can be
1945
Often we will provide convenience methods on 'repository' which carry out
1946
operations with another repository - they will always forward to
1947
InterRepository.get(other).method_name(parameters).
1951
"""The available optimised InterRepository types."""
1953
def copy_content(self, revision_id=None):
1954
raise NotImplementedError(self.copy_content)
1956
def fetch(self, revision_id=None, pb=None):
1957
"""Fetch the content required to construct revision_id.
1959
The content is copied from self.source to self.target.
1961
:param revision_id: if None all content is copied, if NULL_REVISION no
1963
:param pb: optional progress bar to use for progress reports. If not
1964
provided a default one will be created.
1966
Returns the copied revision count and the failed revisions in a tuple:
1969
raise NotImplementedError(self.fetch)
1972
def missing_revision_ids(self, revision_id=None):
1973
"""Return the revision ids that source has that target does not.
1975
These are returned in topological order.
1977
:param revision_id: only return revision ids included by this
1980
# generic, possibly worst case, slow code path.
1981
target_ids = set(self.target.all_revision_ids())
1982
if revision_id is not None:
1983
source_ids = self.source.get_ancestry(revision_id)
1984
assert source_ids[0] is None
1987
source_ids = self.source.all_revision_ids()
1988
result_set = set(source_ids).difference(target_ids)
1989
# this may look like a no-op: its not. It preserves the ordering
1990
# other_ids had while only returning the members from other_ids
1991
# that we've decided we need.
1992
return [rev_id for rev_id in source_ids if rev_id in result_set]
1995
def _same_model(source, target):
1996
"""True if source and target have the same data representation."""
1997
if source.supports_rich_root() != target.supports_rich_root():
1999
if source._serializer != target._serializer:
2004
class InterSameDataRepository(InterRepository):
2005
"""Code for converting between repositories that represent the same data.
2007
Data format and model must match for this to work.
2011
def _get_repo_format_to_test(self):
2012
"""Repository format for testing with.
2014
InterSameData can pull from subtree to subtree and from non-subtree to
2015
non-subtree, so we test this with the richest repository format.
2017
from bzrlib.repofmt import knitrepo
2018
return knitrepo.RepositoryFormatKnit3()
2021
def is_compatible(source, target):
2022
return InterRepository._same_model(source, target)
2025
def copy_content(self, revision_id=None):
2026
"""Make a complete copy of the content in self into destination.
2028
This copies both the repository's revision data, and configuration information
2029
such as the make_working_trees setting.
2031
This is a destructive operation! Do not use it on existing
2034
:param revision_id: Only copy the content needed to construct
2035
revision_id and its parents.
2038
self.target.set_make_working_trees(self.source.make_working_trees())
2039
except NotImplementedError:
2041
# but don't bother fetching if we have the needed data now.
2042
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2043
self.target.has_revision(revision_id)):
2045
self.target.fetch(self.source, revision_id=revision_id)
2048
def fetch(self, revision_id=None, pb=None):
2049
"""See InterRepository.fetch()."""
2050
from bzrlib.fetch import GenericRepoFetcher
2051
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2052
self.source, self.source._format, self.target,
2053
self.target._format)
2054
f = GenericRepoFetcher(to_repository=self.target,
2055
from_repository=self.source,
2056
last_revision=revision_id,
2058
return f.count_copied, f.failed_revisions
2061
class InterWeaveRepo(InterSameDataRepository):
2062
"""Optimised code paths between Weave based repositories.
2064
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2065
implemented lazy inter-object optimisation.
2069
def _get_repo_format_to_test(self):
2070
from bzrlib.repofmt import weaverepo
2071
return weaverepo.RepositoryFormat7()
2074
def is_compatible(source, target):
2075
"""Be compatible with known Weave formats.
2077
We don't test for the stores being of specific types because that
2078
could lead to confusing results, and there is no need to be
2081
from bzrlib.repofmt.weaverepo import (
2087
return (isinstance(source._format, (RepositoryFormat5,
2089
RepositoryFormat7)) and
2090
isinstance(target._format, (RepositoryFormat5,
2092
RepositoryFormat7)))
2093
except AttributeError:
2097
def copy_content(self, revision_id=None):
2098
"""See InterRepository.copy_content()."""
2099
# weave specific optimised path:
2101
self.target.set_make_working_trees(self.source.make_working_trees())
2102
except NotImplementedError:
2104
# FIXME do not peek!
2105
if self.source.control_files._transport.listable():
2106
pb = ui.ui_factory.nested_progress_bar()
2108
self.target.weave_store.copy_all_ids(
2109
self.source.weave_store,
2111
from_transaction=self.source.get_transaction(),
2112
to_transaction=self.target.get_transaction())
2113
pb.update('copying inventory', 0, 1)
2114
self.target.control_weaves.copy_multi(
2115
self.source.control_weaves, ['inventory'],
2116
from_transaction=self.source.get_transaction(),
2117
to_transaction=self.target.get_transaction())
2118
self.target._revision_store.text_store.copy_all_ids(
2119
self.source._revision_store.text_store,
2124
self.target.fetch(self.source, revision_id=revision_id)
2127
def fetch(self, revision_id=None, pb=None):
2128
"""See InterRepository.fetch()."""
2129
from bzrlib.fetch import GenericRepoFetcher
2130
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2131
self.source, self.source._format, self.target, self.target._format)
2132
f = GenericRepoFetcher(to_repository=self.target,
2133
from_repository=self.source,
2134
last_revision=revision_id,
2136
return f.count_copied, f.failed_revisions
2139
def missing_revision_ids(self, revision_id=None):
2140
"""See InterRepository.missing_revision_ids()."""
2141
# we want all revisions to satisfy revision_id in source.
2142
# but we don't want to stat every file here and there.
2143
# we want then, all revisions other needs to satisfy revision_id
2144
# checked, but not those that we have locally.
2145
# so the first thing is to get a subset of the revisions to
2146
# satisfy revision_id in source, and then eliminate those that
2147
# we do already have.
2148
# this is slow on high latency connection to self, but as as this
2149
# disk format scales terribly for push anyway due to rewriting
2150
# inventory.weave, this is considered acceptable.
2152
if revision_id is not None:
2153
source_ids = self.source.get_ancestry(revision_id)
2154
assert source_ids[0] is None
2157
source_ids = self.source._all_possible_ids()
2158
source_ids_set = set(source_ids)
2159
# source_ids is the worst possible case we may need to pull.
2160
# now we want to filter source_ids against what we actually
2161
# have in target, but don't try to check for existence where we know
2162
# we do not have a revision as that would be pointless.
2163
target_ids = set(self.target._all_possible_ids())
2164
possibly_present_revisions = target_ids.intersection(source_ids_set)
2165
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2166
required_revisions = source_ids_set.difference(actually_present_revisions)
2167
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2168
if revision_id is not None:
2169
# we used get_ancestry to determine source_ids then we are assured all
2170
# revisions referenced are present as they are installed in topological order.
2171
# and the tip revision was validated by get_ancestry.
2172
return required_topo_revisions
2174
# if we just grabbed the possibly available ids, then
2175
# we only have an estimate of whats available and need to validate
2176
# that against the revision records.
2177
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2180
class InterKnitRepo(InterSameDataRepository):
2181
"""Optimised code paths between Knit based repositories."""
2184
def _get_repo_format_to_test(self):
2185
from bzrlib.repofmt import knitrepo
2186
return knitrepo.RepositoryFormatKnit1()
2189
def is_compatible(source, target):
2190
"""Be compatible with known Knit formats.
2192
We don't test for the stores being of specific types because that
2193
could lead to confusing results, and there is no need to be
2196
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2198
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2199
isinstance(target._format, RepositoryFormatKnit))
2200
except AttributeError:
2202
return are_knits and InterRepository._same_model(source, target)
2205
def fetch(self, revision_id=None, pb=None):
2206
"""See InterRepository.fetch()."""
2207
from bzrlib.fetch import KnitRepoFetcher
2208
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2209
self.source, self.source._format, self.target, self.target._format)
2210
f = KnitRepoFetcher(to_repository=self.target,
2211
from_repository=self.source,
2212
last_revision=revision_id,
2214
return f.count_copied, f.failed_revisions
2217
def missing_revision_ids(self, revision_id=None):
2218
"""See InterRepository.missing_revision_ids()."""
2219
if revision_id is not None:
2220
source_ids = self.source.get_ancestry(revision_id)
2221
assert source_ids[0] is None
2224
source_ids = self.source.all_revision_ids()
2225
source_ids_set = set(source_ids)
2226
# source_ids is the worst possible case we may need to pull.
2227
# now we want to filter source_ids against what we actually
2228
# have in target, but don't try to check for existence where we know
2229
# we do not have a revision as that would be pointless.
2230
target_ids = set(self.target.all_revision_ids())
2231
possibly_present_revisions = target_ids.intersection(source_ids_set)
2232
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2233
required_revisions = source_ids_set.difference(actually_present_revisions)
2234
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2235
if revision_id is not None:
2236
# we used get_ancestry to determine source_ids then we are assured all
2237
# revisions referenced are present as they are installed in topological order.
2238
# and the tip revision was validated by get_ancestry.
2239
return required_topo_revisions
2241
# if we just grabbed the possibly available ids, then
2242
# we only have an estimate of whats available and need to validate
2243
# that against the revision records.
2244
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2247
class InterModel1and2(InterRepository):
2250
def _get_repo_format_to_test(self):
2254
def is_compatible(source, target):
2255
if not source.supports_rich_root() and target.supports_rich_root():
2261
def fetch(self, revision_id=None, pb=None):
2262
"""See InterRepository.fetch()."""
2263
from bzrlib.fetch import Model1toKnit2Fetcher
2264
f = Model1toKnit2Fetcher(to_repository=self.target,
2265
from_repository=self.source,
2266
last_revision=revision_id,
2268
return f.count_copied, f.failed_revisions
2271
def copy_content(self, revision_id=None):
2272
"""Make a complete copy of the content in self into destination.
2274
This is a destructive operation! Do not use it on existing
2277
:param revision_id: Only copy the content needed to construct
2278
revision_id and its parents.
2281
self.target.set_make_working_trees(self.source.make_working_trees())
2282
except NotImplementedError:
2284
# but don't bother fetching if we have the needed data now.
2285
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2286
self.target.has_revision(revision_id)):
2288
self.target.fetch(self.source, revision_id=revision_id)
2291
class InterKnit1and2(InterKnitRepo):
2294
def _get_repo_format_to_test(self):
2298
def is_compatible(source, target):
2299
"""Be compatible with Knit1 source and Knit3 target"""
2300
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2302
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
2303
RepositoryFormatKnit3
2304
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2305
isinstance(target._format, (RepositoryFormatKnit3)))
2306
except AttributeError:
2310
def fetch(self, revision_id=None, pb=None):
2311
"""See InterRepository.fetch()."""
2312
from bzrlib.fetch import Knit1to2Fetcher
2313
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2314
self.source, self.source._format, self.target,
2315
self.target._format)
2316
f = Knit1to2Fetcher(to_repository=self.target,
2317
from_repository=self.source,
2318
last_revision=revision_id,
2320
return f.count_copied, f.failed_revisions
2323
class InterRemoteToOther(InterRepository):
2325
def __init__(self, source, target):
2326
InterRepository.__init__(self, source, target)
2327
self._real_inter = None
2330
def is_compatible(source, target):
2331
if not isinstance(source, remote.RemoteRepository):
2333
source._ensure_real()
2334
real_source = source._real_repository
2335
# Is source's model compatible with target's model, and are they the
2336
# same format? Currently we can only optimise fetching from an
2337
# identical model & format repo.
2338
assert not isinstance(real_source, remote.RemoteRepository), (
2339
"We don't support remote repos backed by remote repos yet.")
2340
return real_source._format == target._format
2343
def fetch(self, revision_id=None, pb=None):
2344
"""See InterRepository.fetch()."""
2345
from bzrlib.fetch import RemoteToOtherFetcher
2346
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2347
self.source, self.target, self.target._format)
2348
# TODO: jam 20070210 This should be an assert, not a translate
2349
revision_id = osutils.safe_revision_id(revision_id)
2350
f = RemoteToOtherFetcher(to_repository=self.target,
2351
from_repository=self.source,
2352
last_revision=revision_id,
2354
return f.count_copied, f.failed_revisions
2357
def _get_repo_format_to_test(self):
2361
class InterOtherToRemote(InterRepository):
2363
def __init__(self, source, target):
2364
InterRepository.__init__(self, source, target)
2365
self._real_inter = None
2368
def is_compatible(source, target):
2369
if isinstance(target, remote.RemoteRepository):
2373
def _ensure_real_inter(self):
2374
if self._real_inter is None:
2375
self.target._ensure_real()
2376
real_target = self.target._real_repository
2377
self._real_inter = InterRepository.get(self.source, real_target)
2379
def copy_content(self, revision_id=None):
2380
self._ensure_real_inter()
2381
self._real_inter.copy_content(revision_id=revision_id)
2383
def fetch(self, revision_id=None, pb=None):
2384
self._ensure_real_inter()
2385
self._real_inter.fetch(revision_id=revision_id, pb=pb)
2388
def _get_repo_format_to_test(self):
2392
InterRepository.register_optimiser(InterSameDataRepository)
2393
InterRepository.register_optimiser(InterWeaveRepo)
2394
InterRepository.register_optimiser(InterKnitRepo)
2395
InterRepository.register_optimiser(InterModel1and2)
2396
InterRepository.register_optimiser(InterKnit1and2)
2397
InterRepository.register_optimiser(InterRemoteToOther)
2398
InterRepository.register_optimiser(InterOtherToRemote)
2401
class CopyConverter(object):
2402
"""A repository conversion tool which just performs a copy of the content.
2404
This is slow but quite reliable.
2407
def __init__(self, target_format):
2408
"""Create a CopyConverter.
2410
:param target_format: The format the resulting repository should be.
2412
self.target_format = target_format
2414
def convert(self, repo, pb):
2415
"""Perform the conversion of to_convert, giving feedback via pb.
2417
:param to_convert: The disk object to convert.
2418
:param pb: a progress bar to use for progress information.
2423
# this is only useful with metadir layouts - separated repo content.
2424
# trigger an assertion if not such
2425
repo._format.get_format_string()
2426
self.repo_dir = repo.bzrdir
2427
self.step('Moving repository to repository.backup')
2428
self.repo_dir.transport.move('repository', 'repository.backup')
2429
backup_transport = self.repo_dir.transport.clone('repository.backup')
2430
repo._format.check_conversion_target(self.target_format)
2431
self.source_repo = repo._format.open(self.repo_dir,
2433
_override_transport=backup_transport)
2434
self.step('Creating new repository')
2435
converted = self.target_format.initialize(self.repo_dir,
2436
self.source_repo.is_shared())
2437
converted.lock_write()
2439
self.step('Copying content into repository.')
2440
self.source_repo.copy_content_into(converted)
2443
self.step('Deleting old repository content.')
2444
self.repo_dir.transport.delete_tree('repository.backup')
2445
self.pb.note('repository converted')
2447
def step(self, message):
2448
"""Update the pb by a step."""
2450
self.pb.update(message, self.count, self.total)
2462
def _unescaper(match, _map=_unescape_map):
2463
code = match.group(1)
2467
if not code.startswith('#'):
2469
return unichr(int(code[1:])).encode('utf8')
2475
def _unescape_xml(data):
2476
"""Unescape predefined XML entities in a string of data."""
2478
if _unescape_re is None:
2479
_unescape_re = re.compile('\&([^;]*);')
2480
return _unescape_re.sub(_unescaper, data)
2483
class _RevisionTextVersionCache(object):
2484
"""A cache of the versionedfile versions for revision and file-id."""
2486
def __init__(self, repository):
2487
self.repository = repository
2488
self.revision_versions = {}
2489
self.revision_parents = {}
2490
self.repo_graph = self.repository.get_graph()
2491
# XXX: RBC: I haven't tracked down what uses this, but it would be
2492
# better to use the headscache directly I think.
2493
self.heads = graph.HeadsCache(self.repo_graph).heads
2495
def add_revision_text_versions(self, tree):
2496
"""Cache text version data from the supplied revision tree"""
2498
for path, entry in tree.iter_entries_by_dir():
2499
inv_revisions[entry.file_id] = entry.revision
2500
self.revision_versions[tree.get_revision_id()] = inv_revisions
2501
return inv_revisions
2503
def get_text_version(self, file_id, revision_id):
2504
"""Determine the text version for a given file-id and revision-id"""
2506
inv_revisions = self.revision_versions[revision_id]
2509
tree = self.repository.revision_tree(revision_id)
2510
except errors.RevisionNotPresent:
2511
self.revision_versions[revision_id] = inv_revisions = {}
2513
inv_revisions = self.add_revision_text_versions(tree)
2514
return inv_revisions.get(file_id)
2516
def prepopulate_revs(self, revision_ids):
2517
# Filter out versions that we don't have an inventory for, so that the
2518
# revision_trees() call won't fail.
2519
inv_weave = self.repository.get_inventory_weave()
2520
revs = [r for r in revision_ids if inv_weave.has_version(r)]
2521
# XXX: this loop is very similar to
2522
# bzrlib.fetch.Inter1and2Helper.iter_rev_trees.
2524
for tree in self.repository.revision_trees(revs[:100]):
2525
if tree.inventory.revision_id is None:
2526
tree.inventory.revision_id = tree.get_revision_id()
2527
self.add_revision_text_versions(tree)
2530
def get_parents(self, revision_id):
2532
return self.revision_parents[revision_id]
2534
parents = self.repository.get_parents([revision_id])[0]
2535
self.revision_parents[revision_id] = parents
2539
class VersionedFileChecker(object):
2541
def __init__(self, planned_revisions, revision_versions, repository):
2542
self.planned_revisions = planned_revisions
2543
self.revision_versions = revision_versions
2544
self.repository = repository
2546
def calculate_file_version_parents(self, revision_id, file_id):
2547
text_revision = self.revision_versions.get_text_version(
2548
file_id, revision_id)
2549
if text_revision is None:
2551
parents_of_text_revision = self.revision_versions.get_parents(
2553
parents_from_inventories = []
2554
for parent in parents_of_text_revision:
2555
if parent == _mod_revision.NULL_REVISION:
2557
introduced_in = self.revision_versions.get_text_version(file_id,
2559
if introduced_in is not None:
2560
parents_from_inventories.append(introduced_in)
2561
heads = set(self.revision_versions.heads(parents_from_inventories))
2563
for parent in parents_from_inventories:
2564
if parent in heads and parent not in new_parents:
2565
new_parents.append(parent)
2568
def check_file_version_parents(self, weave, file_id):
2570
for num, revision_id in enumerate(self.planned_revisions):
2571
correct_parents = self.calculate_file_version_parents(
2572
revision_id, file_id)
2573
if correct_parents is None:
2575
text_revision = self.revision_versions.get_text_version(
2576
file_id, revision_id)
2577
knit_parents = weave.get_parents(text_revision)
2578
if correct_parents != knit_parents:
2579
result[revision_id] = (knit_parents, correct_parents)
added
removed
bzrlib/repository.py
