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 = osutils.safe_revision_id(revision_id)
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self._repo_graph = repository.get_graph()
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
symbol_versioning.warn('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _check_root(self, ie, parent_invs, tree):
192
"""Helper for record_entry_contents.
194
:param ie: An entry being added.
195
:param parent_invs: The inventories of the parent revisions of the
197
:param tree: The tree that is being committed.
199
# In this revision format, root entries have no knit or weave When
200
# serializing out to disk and back in root.revision is always
202
ie.revision = self._new_revision_id
204
def _get_delta(self, ie, basis_inv, path):
205
"""Get a delta against the basis inventory for ie."""
206
if ie.file_id not in basis_inv:
208
return (None, path, ie.file_id, ie)
209
elif ie != basis_inv[ie.file_id]:
211
# TODO: avoid tis id2path call.
212
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
217
def record_entry_contents(self, ie, parent_invs, path, tree,
219
"""Record the content of ie from tree into the commit if needed.
221
Side effect: sets ie.revision when unchanged
223
:param ie: An inventory entry present in the commit.
224
:param parent_invs: The inventories of the parent revisions of the
226
:param path: The path the entry is at in the tree.
227
:param tree: The tree which contains this entry and should be used to
229
:param content_summary: Summary data from the tree about the paths
230
content - stat, length, exec, sha/link target. This is only
231
accessed when the entry has a revision of None - that is when it is
232
a candidate to commit.
233
:return: A tuple (change_delta, version_recorded). change_delta is
234
an inventory_delta change for this entry against the basis tree of
235
the commit, or None if no change occured against the basis tree.
236
version_recorded is True if a new version of the entry has been
237
recorded. For instance, committing a merge where a file was only
238
changed on the other side will return (delta, False).
240
if self.new_inventory.root is None:
241
if ie.parent_id is not None:
242
raise errors.RootMissing()
243
self._check_root(ie, parent_invs, tree)
244
if ie.revision is None:
245
kind = content_summary[0]
247
# ie is carried over from a prior commit
249
# XXX: repository specific check for nested tree support goes here - if
250
# the repo doesn't want nested trees we skip it ?
251
if (kind == 'tree-reference' and
252
not self.repository._format.supports_tree_reference):
253
# mismatch between commit builder logic and repository:
254
# this needs the entry creation pushed down into the builder.
255
raise NotImplementedError('Missing repository subtree support.')
256
# transitional assert only, will remove before release.
257
assert ie.kind == kind
258
self.new_inventory.add(ie)
260
# TODO: slow, take it out of the inner loop.
262
basis_inv = parent_invs[0]
264
basis_inv = Inventory(root_id=None)
266
# ie.revision is always None if the InventoryEntry is considered
267
# for committing. We may record the previous parents revision if the
268
# content is actually unchanged against a sole head.
269
if ie.revision is not None:
270
if self._versioned_root or path != '':
271
# not considered for commit
274
# repositories that do not version the root set the root's
275
# revision to the new commit even when no change occurs, and
276
# this masks when a change may have occurred against the basis,
277
# so calculate if one happened.
278
if ie.file_id not in basis_inv:
280
delta = (None, path, ie.file_id, ie)
282
basis_id = basis_inv[ie.file_id]
283
if basis_id.name != '':
285
delta = (basis_inv.id2path(ie.file_id), path,
290
# not considered for commit, OR, for non-rich-root
291
return delta, ie.revision == self._new_revision_id and (path != '' or
292
self._versioned_root)
294
# XXX: Friction: parent_candidates should return a list not a dict
295
# so that we don't have to walk the inventories again.
296
parent_candiate_entries = ie.parent_candidates(parent_invs)
297
head_set = self._repo_graph.heads(parent_candiate_entries.keys())
299
for inv in parent_invs:
300
if ie.file_id in inv:
301
old_rev = inv[ie.file_id].revision
302
if old_rev in head_set:
303
heads.append(inv[ie.file_id].revision)
304
head_set.remove(inv[ie.file_id].revision)
307
# now we check to see if we need to write a new record to the
309
# We write a new entry unless there is one head to the ancestors, and
310
# the kind-derived content is unchanged.
312
# Cheapest check first: no ancestors, or more the one head in the
313
# ancestors, we write a new node.
317
# There is a single head, look it up for comparison
318
parent_entry = parent_candiate_entries[heads[0]]
319
# if the non-content specific data has changed, we'll be writing a
321
if (parent_entry.parent_id != ie.parent_id or
322
parent_entry.name != ie.name):
324
# now we need to do content specific checks:
326
# if the kind changed the content obviously has
327
if kind != parent_entry.kind:
331
if (# if the file length changed we have to store:
332
parent_entry.text_size != content_summary[1] or
333
# if the exec bit has changed we have to store:
334
parent_entry.executable != content_summary[2]):
336
elif parent_entry.text_sha1 == content_summary[3]:
337
# all meta and content is unchanged (using a hash cache
338
# hit to check the sha)
339
ie.revision = parent_entry.revision
340
ie.text_size = parent_entry.text_size
341
ie.text_sha1 = parent_entry.text_sha1
342
ie.executable = parent_entry.executable
343
return self._get_delta(ie, basis_inv, path), False
345
# Either there is only a hash change(no hash cache entry,
346
# or same size content change), or there is no change on
348
# Provide the parent's hash to the store layer, so that the
349
# content is unchanged we will not store a new node.
350
nostore_sha = parent_entry.text_sha1
352
# We want to record a new node regardless of the presence or
353
# absence of a content change in the file.
355
ie.executable = content_summary[2]
356
lines = tree.get_file(ie.file_id, path).readlines()
358
ie.text_sha1, ie.text_size = self._add_text_to_weave(
359
ie.file_id, lines, heads, nostore_sha)
360
except errors.ExistingContent:
361
# Turns out that the file content was unchanged, and we were
362
# only going to store a new node if it was changed. Carry over
364
ie.revision = parent_entry.revision
365
ie.text_size = parent_entry.text_size
366
ie.text_sha1 = parent_entry.text_sha1
367
ie.executable = parent_entry.executable
368
return self._get_delta(ie, basis_inv, path), False
369
elif kind == 'directory':
371
# all data is meta here, nothing specific to directory, so
373
ie.revision = parent_entry.revision
374
return self._get_delta(ie, basis_inv, path), False
376
self._add_text_to_weave(ie.file_id, lines, heads, None)
377
elif kind == 'symlink':
378
current_link_target = content_summary[3]
380
# symlink target is not generic metadata, check if it has
382
if current_link_target != parent_entry.symlink_target:
385
# unchanged, carry over.
386
ie.revision = parent_entry.revision
387
ie.symlink_target = parent_entry.symlink_target
388
return self._get_delta(ie, basis_inv, path), False
389
ie.symlink_target = current_link_target
391
self._add_text_to_weave(ie.file_id, lines, heads, None)
392
elif kind == 'tree-reference':
394
if content_summary[3] != parent_entry.reference_revision:
397
# unchanged, carry over.
398
ie.reference_revision = parent_entry.reference_revision
399
ie.revision = parent_entry.revision
400
return self._get_delta(ie, basis_inv, path), False
401
ie.reference_revision = content_summary[3]
403
self._add_text_to_weave(ie.file_id, lines, heads, None)
405
raise NotImplementedError('unknown kind')
406
ie.revision = self._new_revision_id
407
return self._get_delta(ie, basis_inv, path), True
409
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
410
versionedfile = self.repository.weave_store.get_weave_or_empty(
411
file_id, self.repository.get_transaction())
412
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
413
# than add_lines, and allows committing when a parent is ghosted for
415
# Note: as we read the content directly from the tree, we know its not
416
# been turned into unicode or badly split - but a broken tree
417
# implementation could give us bad output from readlines() so this is
418
# not a guarantee of safety. What would be better is always checking
419
# the content during test suite execution. RBC 20070912
421
return versionedfile.add_lines_with_ghosts(
422
self._new_revision_id, parents, new_lines,
423
nostore_sha=nostore_sha, random_id=self.random_revid,
424
check_content=False)[0:2]
426
versionedfile.clear_cache()
429
class RootCommitBuilder(CommitBuilder):
430
"""This commitbuilder actually records the root id"""
432
# the root entry gets versioned properly by this builder.
433
_versioned_root = True
435
def _check_root(self, ie, parent_invs, tree):
436
"""Helper for record_entry_contents.
438
:param ie: An entry being added.
439
:param parent_invs: The inventories of the parent revisions of the
441
:param tree: The tree that is being committed.
445
######################################################################
448
class Repository(object):
449
"""Repository holding history for one or more branches.
451
The repository holds and retrieves historical information including
452
revisions and file history. It's normally accessed only by the Branch,
453
which views a particular line of development through that history.
455
The Repository builds on top of Stores and a Transport, which respectively
456
describe the disk data format and the way of accessing the (possibly
460
# What class to use for a CommitBuilder. Often its simpler to change this
461
# in a Repository class subclass rather than to override
462
# get_commit_builder.
463
_commit_builder_class = CommitBuilder
464
# The search regex used by xml based repositories to determine what things
465
# where changed in a single commit.
466
_file_ids_altered_regex = lazy_regex.lazy_compile(
467
r'file_id="(?P<file_id>[^"]+)"'
468
r'.* revision="(?P<revision_id>[^"]+)"'
471
def abort_write_group(self):
472
"""Commit the contents accrued within the current write group.
474
:seealso: start_write_group.
476
if self._write_group is not self.get_transaction():
477
# has an unlock or relock occured ?
478
raise errors.BzrError('mismatched lock context and write group.')
479
self._abort_write_group()
480
self._write_group = None
482
def _abort_write_group(self):
483
"""Template method for per-repository write group cleanup.
485
This is called during abort before the write group is considered to be
486
finished and should cleanup any internal state accrued during the write
487
group. There is no requirement that data handed to the repository be
488
*not* made available - this is not a rollback - but neither should any
489
attempt be made to ensure that data added is fully commited. Abort is
490
invoked when an error has occured so futher disk or network operations
491
may not be possible or may error and if possible should not be
496
def add_inventory(self, revision_id, inv, parents):
497
"""Add the inventory inv to the repository as revision_id.
499
:param parents: The revision ids of the parents that revision_id
500
is known to have and are in the repository already.
502
returns the sha1 of the serialized inventory.
504
revision_id = osutils.safe_revision_id(revision_id)
505
_mod_revision.check_not_reserved_id(revision_id)
506
assert inv.revision_id is None or inv.revision_id == revision_id, \
507
"Mismatch between inventory revision" \
508
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
509
assert inv.root is not None
510
inv_lines = self._serialise_inventory_to_lines(inv)
511
inv_vf = self.get_inventory_weave()
512
return self._inventory_add_lines(inv_vf, revision_id, parents,
513
inv_lines, check_content=False)
515
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
517
"""Store lines in inv_vf and return the sha1 of the inventory."""
519
for parent in parents:
521
final_parents.append(parent)
522
return inv_vf.add_lines(revision_id, final_parents, lines,
523
check_content=check_content)[0]
526
def add_revision(self, revision_id, rev, inv=None, config=None):
527
"""Add rev to the revision store as revision_id.
529
:param revision_id: the revision id to use.
530
:param rev: The revision object.
531
:param inv: The inventory for the revision. if None, it will be looked
532
up in the inventory storer
533
:param config: If None no digital signature will be created.
534
If supplied its signature_needed method will be used
535
to determine if a signature should be made.
537
revision_id = osutils.safe_revision_id(revision_id)
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
# not right yet - should be more semantically clear ?
628
self.control_store = control_store
629
self.control_weaves = control_store
630
# TODO: make sure to construct the right store classes, etc, depending
631
# on whether escaping is required.
632
self._warn_if_deprecated()
633
self._write_group = None
634
self.base = control_files._transport.base
637
return '%s(%r)' % (self.__class__.__name__,
640
def has_same_location(self, other):
641
"""Returns a boolean indicating if this repository is at the same
642
location as another repository.
644
This might return False even when two repository objects are accessing
645
the same physical repository via different URLs.
647
if self.__class__ is not other.__class__:
649
return (self.control_files._transport.base ==
650
other.control_files._transport.base)
652
def is_in_write_group(self):
653
"""Return True if there is an open write group.
655
:seealso: start_write_group.
657
return self._write_group is not None
660
return self.control_files.is_locked()
662
def lock_write(self, token=None):
663
"""Lock this repository for writing.
665
This causes caching within the repository obejct to start accumlating
666
data during reads, and allows a 'write_group' to be obtained. Write
667
groups must be used for actual data insertion.
669
:param token: if this is already locked, then lock_write will fail
670
unless the token matches the existing lock.
671
:returns: a token if this instance supports tokens, otherwise None.
672
:raises TokenLockingNotSupported: when a token is given but this
673
instance doesn't support using token locks.
674
:raises MismatchedToken: if the specified token doesn't match the token
675
of the existing lock.
676
:seealso: start_write_group.
678
A token should be passed in if you know that you have locked the object
679
some other way, and need to synchronise this object's state with that
682
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
684
result = self.control_files.lock_write(token=token)
689
self.control_files.lock_read()
692
def get_physical_lock_status(self):
693
return self.control_files.get_physical_lock_status()
695
def leave_lock_in_place(self):
696
"""Tell this repository not to release the physical lock when this
699
If lock_write doesn't return a token, then this method is not supported.
701
self.control_files.leave_in_place()
703
def dont_leave_lock_in_place(self):
704
"""Tell this repository to release the physical lock when this
705
object is unlocked, even if it didn't originally acquire it.
707
If lock_write doesn't return a token, then this method is not supported.
709
self.control_files.dont_leave_in_place()
712
def gather_stats(self, revid=None, committers=None):
713
"""Gather statistics from a revision id.
715
:param revid: The revision id to gather statistics from, if None, then
716
no revision specific statistics are gathered.
717
:param committers: Optional parameter controlling whether to grab
718
a count of committers from the revision specific statistics.
719
:return: A dictionary of statistics. Currently this contains:
720
committers: The number of committers if requested.
721
firstrev: A tuple with timestamp, timezone for the penultimate left
722
most ancestor of revid, if revid is not the NULL_REVISION.
723
latestrev: A tuple with timestamp, timezone for revid, if revid is
724
not the NULL_REVISION.
725
revisions: The total revision count in the repository.
726
size: An estimate disk size of the repository in bytes.
729
if revid and committers:
730
result['committers'] = 0
731
if revid and revid != _mod_revision.NULL_REVISION:
733
all_committers = set()
734
revisions = self.get_ancestry(revid)
735
# pop the leading None
737
first_revision = None
739
# ignore the revisions in the middle - just grab first and last
740
revisions = revisions[0], revisions[-1]
741
for revision in self.get_revisions(revisions):
742
if not first_revision:
743
first_revision = revision
745
all_committers.add(revision.committer)
746
last_revision = revision
748
result['committers'] = len(all_committers)
749
result['firstrev'] = (first_revision.timestamp,
750
first_revision.timezone)
751
result['latestrev'] = (last_revision.timestamp,
752
last_revision.timezone)
754
# now gather global repository information
755
if self.bzrdir.root_transport.listable():
756
c, t = self._revision_store.total_size(self.get_transaction())
757
result['revisions'] = c
761
def get_data_stream(self, revision_ids):
762
raise NotImplementedError(self.get_data_stream)
764
def insert_data_stream(self, stream):
765
for item_key, bytes in stream:
766
if item_key[0] == 'file':
767
(file_id,) = item_key[1:]
768
knit = self.weave_store.get_weave_or_empty(
769
file_id, self.get_transaction())
770
elif item_key == ('inventory',):
771
knit = self.get_inventory_weave()
772
elif item_key == ('revisions',):
773
knit = self._revision_store.get_revision_file(
774
self.get_transaction())
775
elif item_key == ('signatures',):
776
knit = self._revision_store.get_signature_file(
777
self.get_transaction())
779
raise RepositoryDataStreamError(
780
"Unrecognised data stream key '%s'" % (item_key,))
781
decoded_list = bencode.bdecode(bytes)
782
format = decoded_list.pop(0)
785
for version, options, parents, some_bytes in decoded_list:
786
data_list.append((version, options, len(some_bytes), parents))
787
knit_bytes += some_bytes
788
knit.insert_data_stream(
789
(format, data_list, StringIO(knit_bytes).read))
792
def missing_revision_ids(self, other, revision_id=None):
793
"""Return the revision ids that other has that this does not.
795
These are returned in topological order.
797
revision_id: only return revision ids included by revision_id.
799
revision_id = osutils.safe_revision_id(revision_id)
800
return InterRepository.get(other, self).missing_revision_ids(revision_id)
804
"""Open the repository rooted at base.
806
For instance, if the repository is at URL/.bzr/repository,
807
Repository.open(URL) -> a Repository instance.
809
control = bzrdir.BzrDir.open(base)
810
return control.open_repository()
812
def copy_content_into(self, destination, revision_id=None):
813
"""Make a complete copy of the content in self into destination.
815
This is a destructive operation! Do not use it on existing
818
revision_id = osutils.safe_revision_id(revision_id)
819
return InterRepository.get(self, destination).copy_content(revision_id)
821
def commit_write_group(self):
822
"""Commit the contents accrued within the current write group.
824
:seealso: start_write_group.
826
if self._write_group is not self.get_transaction():
827
# has an unlock or relock occured ?
828
raise errors.BzrError('mismatched lock context %r and '
830
(self.get_transaction(), self._write_group))
831
self._commit_write_group()
832
self._write_group = None
834
def _commit_write_group(self):
835
"""Template method for per-repository write group cleanup.
837
This is called before the write group is considered to be
838
finished and should ensure that all data handed to the repository
839
for writing during the write group is safely committed (to the
840
extent possible considering file system caching etc).
843
def fetch(self, source, revision_id=None, pb=None):
844
"""Fetch the content required to construct revision_id from source.
846
If revision_id is None all content is copied.
848
revision_id = osutils.safe_revision_id(revision_id)
849
# fast path same-url fetch operations
850
if self.has_same_location(source):
851
# check that last_revision is in 'from' and then return a
853
if (revision_id is not None and
854
not _mod_revision.is_null(revision_id)):
855
self.get_revision(revision_id)
857
inter = InterRepository.get(source, self)
859
return inter.fetch(revision_id=revision_id, pb=pb)
860
except NotImplementedError:
861
raise errors.IncompatibleRepositories(source, self)
863
def create_bundle(self, target, base, fileobj, format=None):
864
return serializer.write_bundle(self, target, base, fileobj, format)
866
def get_commit_builder(self, branch, parents, config, timestamp=None,
867
timezone=None, committer=None, revprops=None,
869
"""Obtain a CommitBuilder for this repository.
871
:param branch: Branch to commit to.
872
:param parents: Revision ids of the parents of the new revision.
873
:param config: Configuration to use.
874
:param timestamp: Optional timestamp recorded for commit.
875
:param timezone: Optional timezone for timestamp.
876
:param committer: Optional committer to set for commit.
877
:param revprops: Optional dictionary of revision properties.
878
:param revision_id: Optional revision id.
880
revision_id = osutils.safe_revision_id(revision_id)
881
result = self._commit_builder_class(self, parents, config,
882
timestamp, timezone, committer, revprops, revision_id)
883
self.start_write_group()
887
if (self.control_files._lock_count == 1 and
888
self.control_files._lock_mode == 'w'):
889
if self._write_group is not None:
890
raise errors.BzrError(
891
'Must end write groups before releasing write locks.')
892
self.control_files.unlock()
895
def clone(self, a_bzrdir, revision_id=None):
896
"""Clone this repository into a_bzrdir using the current format.
898
Currently no check is made that the format of this repository and
899
the bzrdir format are compatible. FIXME RBC 20060201.
901
:return: The newly created destination repository.
903
# TODO: deprecate after 0.16; cloning this with all its settings is
904
# probably not very useful -- mbp 20070423
905
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
906
self.copy_content_into(dest_repo, revision_id)
909
def start_write_group(self):
910
"""Start a write group in the repository.
912
Write groups are used by repositories which do not have a 1:1 mapping
913
between file ids and backend store to manage the insertion of data from
914
both fetch and commit operations.
916
A write lock is required around the start_write_group/commit_write_group
917
for the support of lock-requiring repository formats.
919
One can only insert data into a repository inside a write group.
923
if not self.is_locked() or self.control_files._lock_mode != 'w':
924
raise errors.NotWriteLocked(self)
925
if self._write_group:
926
raise errors.BzrError('already in a write group')
927
self._start_write_group()
928
# so we can detect unlock/relock - the write group is now entered.
929
self._write_group = self.get_transaction()
931
def _start_write_group(self):
932
"""Template method for per-repository write group startup.
934
This is called before the write group is considered to be
939
def sprout(self, to_bzrdir, revision_id=None):
940
"""Create a descendent repository for new development.
942
Unlike clone, this does not copy the settings of the repository.
944
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
945
dest_repo.fetch(self, revision_id=revision_id)
948
def _create_sprouting_repo(self, a_bzrdir, shared):
949
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
950
# use target default format.
951
dest_repo = a_bzrdir.create_repository()
953
# Most control formats need the repository to be specifically
954
# created, but on some old all-in-one formats it's not needed
956
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
957
except errors.UninitializableFormat:
958
dest_repo = a_bzrdir.open_repository()
962
def has_revision(self, revision_id):
963
"""True if this repository has a copy of the revision."""
964
if 'evil' in debug.debug_flags:
965
mutter_callsite(3, "has_revision is a LBYL symptom.")
966
revision_id = osutils.safe_revision_id(revision_id)
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
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
995
for rev_id in revision_ids:
996
if not rev_id or not isinstance(rev_id, basestring):
997
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
998
revs = self._revision_store.get_revisions(revision_ids,
999
self.get_transaction())
1001
assert not isinstance(rev.revision_id, unicode)
1002
for parent_id in rev.parent_ids:
1003
assert not isinstance(parent_id, unicode)
1007
def get_revision_xml(self, revision_id):
1008
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1009
# would have already do it.
1010
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1011
revision_id = osutils.safe_revision_id(revision_id)
1012
rev = self.get_revision(revision_id)
1013
rev_tmp = StringIO()
1014
# the current serializer..
1015
self._revision_store._serializer.write_revision(rev, rev_tmp)
1017
return rev_tmp.getvalue()
1020
def get_deltas_for_revisions(self, revisions):
1021
"""Produce a generator of revision deltas.
1023
Note that the input is a sequence of REVISIONS, not revision_ids.
1024
Trees will be held in memory until the generator exits.
1025
Each delta is relative to the revision's lefthand predecessor.
1027
required_trees = set()
1028
for revision in revisions:
1029
required_trees.add(revision.revision_id)
1030
required_trees.update(revision.parent_ids[:1])
1031
trees = dict((t.get_revision_id(), t) for
1032
t in self.revision_trees(required_trees))
1033
for revision in revisions:
1034
if not revision.parent_ids:
1035
old_tree = self.revision_tree(None)
1037
old_tree = trees[revision.parent_ids[0]]
1038
yield trees[revision.revision_id].changes_from(old_tree)
1041
def get_revision_delta(self, revision_id):
1042
"""Return the delta for one revision.
1044
The delta is relative to the left-hand predecessor of the
1047
r = self.get_revision(revision_id)
1048
return list(self.get_deltas_for_revisions([r]))[0]
1051
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1052
revision_id = osutils.safe_revision_id(revision_id)
1053
signature = gpg_strategy.sign(plaintext)
1054
self._revision_store.add_revision_signature_text(revision_id,
1056
self.get_transaction())
1058
def fileids_altered_by_revision_ids(self, revision_ids):
1059
"""Find the file ids and versions affected by revisions.
1061
:param revisions: an iterable containing revision ids.
1062
:return: a dictionary mapping altered file-ids to an iterable of
1063
revision_ids. Each altered file-ids has the exact revision_ids that
1064
altered it listed explicitly.
1066
assert self._serializer.support_altered_by_hack, \
1067
("fileids_altered_by_revision_ids only supported for branches "
1068
"which store inventory as unnested xml, not on %r" % self)
1069
selected_revision_ids = set(osutils.safe_revision_id(r)
1070
for r in revision_ids)
1071
w = self.get_inventory_weave()
1074
# this code needs to read every new line in every inventory for the
1075
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1076
# not present in one of those inventories is unnecessary but not
1077
# harmful because we are filtering by the revision id marker in the
1078
# inventory lines : we only select file ids altered in one of those
1079
# revisions. We don't need to see all lines in the inventory because
1080
# only those added in an inventory in rev X can contain a revision=X
1082
unescape_revid_cache = {}
1083
unescape_fileid_cache = {}
1085
# jam 20061218 In a big fetch, this handles hundreds of thousands
1086
# of lines, so it has had a lot of inlining and optimizing done.
1087
# Sorry that it is a little bit messy.
1088
# Move several functions to be local variables, since this is a long
1090
search = self._file_ids_altered_regex.search
1091
unescape = _unescape_xml
1092
setdefault = result.setdefault
1093
pb = ui.ui_factory.nested_progress_bar()
1095
for line in w.iter_lines_added_or_present_in_versions(
1096
selected_revision_ids, pb=pb):
1097
match = search(line)
1100
# One call to match.group() returning multiple items is quite a
1101
# bit faster than 2 calls to match.group() each returning 1
1102
file_id, revision_id = match.group('file_id', 'revision_id')
1104
# Inlining the cache lookups helps a lot when you make 170,000
1105
# lines and 350k ids, versus 8.4 unique ids.
1106
# Using a cache helps in 2 ways:
1107
# 1) Avoids unnecessary decoding calls
1108
# 2) Re-uses cached strings, which helps in future set and
1110
# (2) is enough that removing encoding entirely along with
1111
# the cache (so we are using plain strings) results in no
1112
# performance improvement.
1114
revision_id = unescape_revid_cache[revision_id]
1116
unescaped = unescape(revision_id)
1117
unescape_revid_cache[revision_id] = unescaped
1118
revision_id = unescaped
1120
if revision_id in selected_revision_ids:
1122
file_id = unescape_fileid_cache[file_id]
1124
unescaped = unescape(file_id)
1125
unescape_fileid_cache[file_id] = unescaped
1127
setdefault(file_id, set()).add(revision_id)
1132
def iter_files_bytes(self, desired_files):
1133
"""Iterate through file versions.
1135
Files will not necessarily be returned in the order they occur in
1136
desired_files. No specific order is guaranteed.
1138
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1139
value supplied by the caller as part of desired_files. It should
1140
uniquely identify the file version in the caller's context. (Examples:
1141
an index number or a TreeTransform trans_id.)
1143
bytes_iterator is an iterable of bytestrings for the file. The
1144
kind of iterable and length of the bytestrings are unspecified, but for
1145
this implementation, it is a list of lines produced by
1146
VersionedFile.get_lines().
1148
:param desired_files: a list of (file_id, revision_id, identifier)
1151
transaction = self.get_transaction()
1152
for file_id, revision_id, callable_data in desired_files:
1154
weave = self.weave_store.get_weave(file_id, transaction)
1155
except errors.NoSuchFile:
1156
raise errors.NoSuchIdInRepository(self, file_id)
1157
yield callable_data, weave.get_lines(revision_id)
1159
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1160
"""Get an iterable listing the keys of all the data introduced by a set
1163
The keys will be ordered so that the corresponding items can be safely
1164
fetched and inserted in that order.
1166
:returns: An iterable producing tuples of (knit-kind, file-id,
1167
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1168
'revisions'. file-id is None unless knit-kind is 'file'.
1170
# XXX: it's a bit weird to control the inventory weave caching in this
1171
# generator. Ideally the caching would be done in fetch.py I think. Or
1172
# maybe this generator should explicitly have the contract that it
1173
# should not be iterated until the previously yielded item has been
1176
inv_w = self.get_inventory_weave()
1177
inv_w.enable_cache()
1179
# file ids that changed
1180
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1182
num_file_ids = len(file_ids)
1183
for file_id, altered_versions in file_ids.iteritems():
1184
if _files_pb is not None:
1185
_files_pb.update("fetch texts", count, num_file_ids)
1187
yield ("file", file_id, altered_versions)
1188
# We're done with the files_pb. Note that it finished by the caller,
1189
# just as it was created by the caller.
1193
yield ("inventory", None, revision_ids)
1197
revisions_with_signatures = set()
1198
for rev_id in revision_ids:
1200
self.get_signature_text(rev_id)
1201
except errors.NoSuchRevision:
1205
revisions_with_signatures.add(rev_id)
1207
yield ("signatures", None, revisions_with_signatures)
1210
yield ("revisions", None, revision_ids)
1213
def get_inventory_weave(self):
1214
return self.control_weaves.get_weave('inventory',
1215
self.get_transaction())
1218
def get_inventory(self, revision_id):
1219
"""Get Inventory object by hash."""
1220
# TODO: jam 20070210 Technically we don't need to sanitize, since all
1221
# called functions must sanitize.
1222
revision_id = osutils.safe_revision_id(revision_id)
1223
return self.deserialise_inventory(
1224
revision_id, self.get_inventory_xml(revision_id))
1226
def deserialise_inventory(self, revision_id, xml):
1227
"""Transform the xml into an inventory object.
1229
:param revision_id: The expected revision id of the inventory.
1230
:param xml: A serialised inventory.
1232
revision_id = osutils.safe_revision_id(revision_id)
1233
result = self._serializer.read_inventory_from_string(xml)
1234
result.root.revision = revision_id
1237
def serialise_inventory(self, inv):
1238
return self._serializer.write_inventory_to_string(inv)
1240
def _serialise_inventory_to_lines(self, inv):
1241
return self._serializer.write_inventory_to_lines(inv)
1243
def get_serializer_format(self):
1244
return self._serializer.format_num
1247
def get_inventory_xml(self, revision_id):
1248
"""Get inventory XML as a file object."""
1249
revision_id = osutils.safe_revision_id(revision_id)
1251
assert isinstance(revision_id, str), type(revision_id)
1252
iw = self.get_inventory_weave()
1253
return iw.get_text(revision_id)
1255
raise errors.HistoryMissing(self, 'inventory', revision_id)
1258
def get_inventory_sha1(self, revision_id):
1259
"""Return the sha1 hash of the inventory entry
1261
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
1262
revision_id = osutils.safe_revision_id(revision_id)
1263
return self.get_revision(revision_id).inventory_sha1
1266
def get_revision_graph(self, revision_id=None):
1267
"""Return a dictionary containing the revision graph.
1269
NB: This method should not be used as it accesses the entire graph all
1270
at once, which is much more data than most operations should require.
1272
:param revision_id: The revision_id to get a graph from. If None, then
1273
the entire revision graph is returned. This is a deprecated mode of
1274
operation and will be removed in the future.
1275
:return: a dictionary of revision_id->revision_parents_list.
1277
raise NotImplementedError(self.get_revision_graph)
1280
def get_revision_graph_with_ghosts(self, revision_ids=None):
1281
"""Return a graph of the revisions with ghosts marked as applicable.
1283
:param revision_ids: an iterable of revisions to graph or None for all.
1284
:return: a Graph object with the graph reachable from revision_ids.
1286
if 'evil' in debug.debug_flags:
1288
"get_revision_graph_with_ghosts scales with size of history.")
1289
result = deprecated_graph.Graph()
1290
if not revision_ids:
1291
pending = set(self.all_revision_ids())
1294
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
1295
# special case NULL_REVISION
1296
if _mod_revision.NULL_REVISION in pending:
1297
pending.remove(_mod_revision.NULL_REVISION)
1298
required = set(pending)
1301
revision_id = pending.pop()
1303
rev = self.get_revision(revision_id)
1304
except errors.NoSuchRevision:
1305
if revision_id in required:
1308
result.add_ghost(revision_id)
1310
for parent_id in rev.parent_ids:
1311
# is this queued or done ?
1312
if (parent_id not in pending and
1313
parent_id not in done):
1315
pending.add(parent_id)
1316
result.add_node(revision_id, rev.parent_ids)
1317
done.add(revision_id)
1320
def _get_history_vf(self):
1321
"""Get a versionedfile whose history graph reflects all revisions.
1323
For weave repositories, this is the inventory weave.
1325
return self.get_inventory_weave()
1327
def iter_reverse_revision_history(self, revision_id):
1328
"""Iterate backwards through revision ids in the lefthand history
1330
:param revision_id: The revision id to start with. All its lefthand
1331
ancestors will be traversed.
1333
revision_id = osutils.safe_revision_id(revision_id)
1334
if revision_id in (None, _mod_revision.NULL_REVISION):
1336
next_id = revision_id
1337
versionedfile = self._get_history_vf()
1340
parents = versionedfile.get_parents(next_id)
1341
if len(parents) == 0:
1344
next_id = parents[0]
1347
def get_revision_inventory(self, revision_id):
1348
"""Return inventory of a past revision."""
1349
# TODO: Unify this with get_inventory()
1350
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1351
# must be the same as its revision, so this is trivial.
1352
if revision_id is None:
1353
# This does not make sense: if there is no revision,
1354
# then it is the current tree inventory surely ?!
1355
# and thus get_root_id() is something that looks at the last
1356
# commit on the branch, and the get_root_id is an inventory check.
1357
raise NotImplementedError
1358
# return Inventory(self.get_root_id())
1360
return self.get_inventory(revision_id)
1363
def is_shared(self):
1364
"""Return True if this repository is flagged as a shared repository."""
1365
raise NotImplementedError(self.is_shared)
1368
def reconcile(self, other=None, thorough=False):
1369
"""Reconcile this repository."""
1370
from bzrlib.reconcile import RepoReconciler
1371
reconciler = RepoReconciler(self, thorough=thorough)
1372
reconciler.reconcile()
1375
def _refresh_data(self):
1376
"""Helper called from lock_* to ensure coherency with disk.
1378
The default implementation does nothing; it is however possible
1379
for repositories to maintain loaded indices across multiple locks
1380
by checking inside their implementation of this method to see
1381
whether their indices are still valid. This depends of course on
1382
the disk format being validatable in this manner.
1386
def revision_tree(self, revision_id):
1387
"""Return Tree for a revision on this branch.
1389
`revision_id` may be None for the empty tree revision.
1391
# TODO: refactor this to use an existing revision object
1392
# so we don't need to read it in twice.
1393
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1394
return RevisionTree(self, Inventory(root_id=None),
1395
_mod_revision.NULL_REVISION)
1397
revision_id = osutils.safe_revision_id(revision_id)
1398
inv = self.get_revision_inventory(revision_id)
1399
return RevisionTree(self, inv, revision_id)
1402
def revision_trees(self, revision_ids):
1403
"""Return Tree for a revision on this branch.
1405
`revision_id` may not be None or 'null:'"""
1406
assert None not in revision_ids
1407
assert _mod_revision.NULL_REVISION not in revision_ids
1408
texts = self.get_inventory_weave().get_texts(revision_ids)
1409
for text, revision_id in zip(texts, revision_ids):
1410
inv = self.deserialise_inventory(revision_id, text)
1411
yield RevisionTree(self, inv, revision_id)
1414
def get_ancestry(self, revision_id, topo_sorted=True):
1415
"""Return a list of revision-ids integrated by a revision.
1417
The first element of the list is always None, indicating the origin
1418
revision. This might change when we have history horizons, or
1419
perhaps we should have a new API.
1421
This is topologically sorted.
1423
if _mod_revision.is_null(revision_id):
1425
revision_id = osutils.safe_revision_id(revision_id)
1426
if not self.has_revision(revision_id):
1427
raise errors.NoSuchRevision(self, revision_id)
1428
w = self.get_inventory_weave()
1429
candidates = w.get_ancestry(revision_id, topo_sorted)
1430
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1433
"""Compress the data within the repository.
1435
This operation only makes sense for some repository types. For other
1436
types it should be a no-op that just returns.
1438
This stub method does not require a lock, but subclasses should use
1439
@needs_write_lock as this is a long running call its reasonable to
1440
implicitly lock for the user.
1444
def print_file(self, file, revision_id):
1445
"""Print `file` to stdout.
1447
FIXME RBC 20060125 as John Meinel points out this is a bad api
1448
- it writes to stdout, it assumes that that is valid etc. Fix
1449
by creating a new more flexible convenience function.
1451
revision_id = osutils.safe_revision_id(revision_id)
1452
tree = self.revision_tree(revision_id)
1453
# use inventory as it was in that revision
1454
file_id = tree.inventory.path2id(file)
1456
# TODO: jam 20060427 Write a test for this code path
1457
# it had a bug in it, and was raising the wrong
1459
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1460
tree.print_file(file_id)
1462
def get_transaction(self):
1463
return self.control_files.get_transaction()
1465
def revision_parents(self, revision_id):
1466
revision_id = osutils.safe_revision_id(revision_id)
1467
return self.get_inventory_weave().parent_names(revision_id)
1469
def get_parents(self, revision_ids):
1470
"""See StackedParentsProvider.get_parents"""
1472
for revision_id in revision_ids:
1473
if revision_id == _mod_revision.NULL_REVISION:
1477
parents = self.get_revision(revision_id).parent_ids
1478
except errors.NoSuchRevision:
1481
if len(parents) == 0:
1482
parents = [_mod_revision.NULL_REVISION]
1483
parents_list.append(parents)
1486
def _make_parents_provider(self):
1489
def get_graph(self, other_repository=None):
1490
"""Return the graph walker for this repository format"""
1491
parents_provider = self._make_parents_provider()
1492
if (other_repository is not None and
1493
other_repository.bzrdir.transport.base !=
1494
self.bzrdir.transport.base):
1495
parents_provider = graph._StackedParentsProvider(
1496
[parents_provider, other_repository._make_parents_provider()])
1497
return graph.Graph(parents_provider)
1500
def set_make_working_trees(self, new_value):
1501
"""Set the policy flag for making working trees when creating branches.
1503
This only applies to branches that use this repository.
1505
The default is 'True'.
1506
:param new_value: True to restore the default, False to disable making
1509
raise NotImplementedError(self.set_make_working_trees)
1511
def make_working_trees(self):
1512
"""Returns the policy for making working trees on new branches."""
1513
raise NotImplementedError(self.make_working_trees)
1516
def sign_revision(self, revision_id, gpg_strategy):
1517
revision_id = osutils.safe_revision_id(revision_id)
1518
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1519
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1522
def has_signature_for_revision_id(self, revision_id):
1523
"""Query for a revision signature for revision_id in the repository."""
1524
revision_id = osutils.safe_revision_id(revision_id)
1525
return self._revision_store.has_signature(revision_id,
1526
self.get_transaction())
1529
def get_signature_text(self, revision_id):
1530
"""Return the text for a signature."""
1531
revision_id = osutils.safe_revision_id(revision_id)
1532
return self._revision_store.get_signature_text(revision_id,
1533
self.get_transaction())
1536
def check(self, revision_ids):
1537
"""Check consistency of all history of given revision_ids.
1539
Different repository implementations should override _check().
1541
:param revision_ids: A non-empty list of revision_ids whose ancestry
1542
will be checked. Typically the last revision_id of a branch.
1544
if not revision_ids:
1545
raise ValueError("revision_ids must be non-empty in %s.check"
1547
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
1548
return self._check(revision_ids)
1550
def _check(self, revision_ids):
1551
result = check.Check(self)
1555
def _warn_if_deprecated(self):
1556
global _deprecation_warning_done
1557
if _deprecation_warning_done:
1559
_deprecation_warning_done = True
1560
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1561
% (self._format, self.bzrdir.transport.base))
1563
def supports_rich_root(self):
1564
return self._format.rich_root_data
1566
def _check_ascii_revisionid(self, revision_id, method):
1567
"""Private helper for ascii-only repositories."""
1568
# weave repositories refuse to store revisionids that are non-ascii.
1569
if revision_id is not None:
1570
# weaves require ascii revision ids.
1571
if isinstance(revision_id, unicode):
1573
revision_id.encode('ascii')
1574
except UnicodeEncodeError:
1575
raise errors.NonAsciiRevisionId(method, self)
1578
revision_id.decode('ascii')
1579
except UnicodeDecodeError:
1580
raise errors.NonAsciiRevisionId(method, self)
1584
# remove these delegates a while after bzr 0.15
1585
def __make_delegated(name, from_module):
1586
def _deprecated_repository_forwarder():
1587
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1588
% (name, from_module),
1591
m = __import__(from_module, globals(), locals(), [name])
1593
return getattr(m, name)
1594
except AttributeError:
1595
raise AttributeError('module %s has no name %s'
1597
globals()[name] = _deprecated_repository_forwarder
1600
'AllInOneRepository',
1601
'WeaveMetaDirRepository',
1602
'PreSplitOutRepositoryFormat',
1603
'RepositoryFormat4',
1604
'RepositoryFormat5',
1605
'RepositoryFormat6',
1606
'RepositoryFormat7',
1608
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1612
'RepositoryFormatKnit',
1613
'RepositoryFormatKnit1',
1615
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1618
def install_revision(repository, rev, revision_tree):
1619
"""Install all revision data into a repository."""
1620
present_parents = []
1622
for p_id in rev.parent_ids:
1623
if repository.has_revision(p_id):
1624
present_parents.append(p_id)
1625
parent_trees[p_id] = repository.revision_tree(p_id)
1627
parent_trees[p_id] = repository.revision_tree(None)
1629
inv = revision_tree.inventory
1630
entries = inv.iter_entries()
1631
# backwards compatibility hack: skip the root id.
1632
if not repository.supports_rich_root():
1633
path, root = entries.next()
1634
if root.revision != rev.revision_id:
1635
raise errors.IncompatibleRevision(repr(repository))
1636
# Add the texts that are not already present
1637
for path, ie in entries:
1638
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1639
repository.get_transaction())
1640
if ie.revision not in w:
1642
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1643
# with InventoryEntry.find_previous_heads(). if it is, then there
1644
# is a latent bug here where the parents may have ancestors of each
1646
for revision, tree in parent_trees.iteritems():
1647
if ie.file_id not in tree:
1649
parent_id = tree.inventory[ie.file_id].revision
1650
if parent_id in text_parents:
1652
text_parents.append(parent_id)
1654
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1655
repository.get_transaction())
1656
lines = revision_tree.get_file(ie.file_id).readlines()
1657
vfile.add_lines(rev.revision_id, text_parents, lines)
1659
# install the inventory
1660
repository.add_inventory(rev.revision_id, inv, present_parents)
1661
except errors.RevisionAlreadyPresent:
1663
repository.add_revision(rev.revision_id, rev, inv)
1666
class MetaDirRepository(Repository):
1667
"""Repositories in the new meta-dir layout."""
1669
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1670
super(MetaDirRepository, self).__init__(_format,
1676
dir_mode = self.control_files._dir_mode
1677
file_mode = self.control_files._file_mode
1680
def is_shared(self):
1681
"""Return True if this repository is flagged as a shared repository."""
1682
return self.control_files._transport.has('shared-storage')
1685
def set_make_working_trees(self, new_value):
1686
"""Set the policy flag for making working trees when creating branches.
1688
This only applies to branches that use this repository.
1690
The default is 'True'.
1691
:param new_value: True to restore the default, False to disable making
1696
self.control_files._transport.delete('no-working-trees')
1697
except errors.NoSuchFile:
1700
self.control_files.put_utf8('no-working-trees', '')
1702
def make_working_trees(self):
1703
"""Returns the policy for making working trees on new branches."""
1704
return not self.control_files._transport.has('no-working-trees')
1707
class RepositoryFormatRegistry(registry.Registry):
1708
"""Registry of RepositoryFormats.
1711
def get(self, format_string):
1712
r = registry.Registry.get(self, format_string)
1718
format_registry = RepositoryFormatRegistry()
1719
"""Registry of formats, indexed by their identifying format string.
1721
This can contain either format instances themselves, or classes/factories that
1722
can be called to obtain one.
1726
#####################################################################
1727
# Repository Formats
1729
class RepositoryFormat(object):
1730
"""A repository format.
1732
Formats provide three things:
1733
* An initialization routine to construct repository data on disk.
1734
* a format string which is used when the BzrDir supports versioned
1736
* an open routine which returns a Repository instance.
1738
Formats are placed in an dict by their format string for reference
1739
during opening. These should be subclasses of RepositoryFormat
1742
Once a format is deprecated, just deprecate the initialize and open
1743
methods on the format class. Do not deprecate the object, as the
1744
object will be created every system load.
1746
Common instance attributes:
1747
_matchingbzrdir - the bzrdir format that the repository format was
1748
originally written to work with. This can be used if manually
1749
constructing a bzrdir and repository, or more commonly for test suite
1754
return "<%s>" % self.__class__.__name__
1756
def __eq__(self, other):
1757
# format objects are generally stateless
1758
return isinstance(other, self.__class__)
1760
def __ne__(self, other):
1761
return not self == other
1764
def find_format(klass, a_bzrdir):
1765
"""Return the format for the repository object in a_bzrdir.
1767
This is used by bzr native formats that have a "format" file in
1768
the repository. Other methods may be used by different types of
1772
transport = a_bzrdir.get_repository_transport(None)
1773
format_string = transport.get("format").read()
1774
return format_registry.get(format_string)
1775
except errors.NoSuchFile:
1776
raise errors.NoRepositoryPresent(a_bzrdir)
1778
raise errors.UnknownFormatError(format=format_string)
1781
def register_format(klass, format):
1782
format_registry.register(format.get_format_string(), format)
1785
def unregister_format(klass, format):
1786
format_registry.remove(format.get_format_string())
1789
def get_default_format(klass):
1790
"""Return the current default format."""
1791
from bzrlib import bzrdir
1792
return bzrdir.format_registry.make_bzrdir('default').repository_format
1794
def _get_control_store(self, repo_transport, control_files):
1795
"""Return the control store for this repository."""
1796
raise NotImplementedError(self._get_control_store)
1798
def get_format_string(self):
1799
"""Return the ASCII format string that identifies this format.
1801
Note that in pre format ?? repositories the format string is
1802
not permitted nor written to disk.
1804
raise NotImplementedError(self.get_format_string)
1806
def get_format_description(self):
1807
"""Return the short description for this format."""
1808
raise NotImplementedError(self.get_format_description)
1810
def _get_revision_store(self, repo_transport, control_files):
1811
"""Return the revision store object for this a_bzrdir."""
1812
raise NotImplementedError(self._get_revision_store)
1814
def _get_text_rev_store(self,
1821
"""Common logic for getting a revision store for a repository.
1823
see self._get_revision_store for the subclass-overridable method to
1824
get the store for a repository.
1826
from bzrlib.store.revision.text import TextRevisionStore
1827
dir_mode = control_files._dir_mode
1828
file_mode = control_files._file_mode
1829
text_store = TextStore(transport.clone(name),
1831
compressed=compressed,
1833
file_mode=file_mode)
1834
_revision_store = TextRevisionStore(text_store, serializer)
1835
return _revision_store
1837
# TODO: this shouldn't be in the base class, it's specific to things that
1838
# use weaves or knits -- mbp 20070207
1839
def _get_versioned_file_store(self,
1844
versionedfile_class=None,
1845
versionedfile_kwargs={},
1847
if versionedfile_class is None:
1848
versionedfile_class = self._versionedfile_class
1849
weave_transport = control_files._transport.clone(name)
1850
dir_mode = control_files._dir_mode
1851
file_mode = control_files._file_mode
1852
return VersionedFileStore(weave_transport, prefixed=prefixed,
1854
file_mode=file_mode,
1855
versionedfile_class=versionedfile_class,
1856
versionedfile_kwargs=versionedfile_kwargs,
1859
def initialize(self, a_bzrdir, shared=False):
1860
"""Initialize a repository of this format in a_bzrdir.
1862
:param a_bzrdir: The bzrdir to put the new repository in it.
1863
:param shared: The repository should be initialized as a sharable one.
1864
:returns: The new repository object.
1866
This may raise UninitializableFormat if shared repository are not
1867
compatible the a_bzrdir.
1869
raise NotImplementedError(self.initialize)
1871
def is_supported(self):
1872
"""Is this format supported?
1874
Supported formats must be initializable and openable.
1875
Unsupported formats may not support initialization or committing or
1876
some other features depending on the reason for not being supported.
1880
def check_conversion_target(self, target_format):
1881
raise NotImplementedError(self.check_conversion_target)
1883
def open(self, a_bzrdir, _found=False):
1884
"""Return an instance of this format for the bzrdir a_bzrdir.
1886
_found is a private parameter, do not use it.
1888
raise NotImplementedError(self.open)
1891
class MetaDirRepositoryFormat(RepositoryFormat):
1892
"""Common base class for the new repositories using the metadir layout."""
1894
rich_root_data = False
1895
supports_tree_reference = False
1896
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1899
super(MetaDirRepositoryFormat, self).__init__()
1901
def _create_control_files(self, a_bzrdir):
1902
"""Create the required files and the initial control_files object."""
1903
# FIXME: RBC 20060125 don't peek under the covers
1904
# NB: no need to escape relative paths that are url safe.
1905
repository_transport = a_bzrdir.get_repository_transport(self)
1906
control_files = lockable_files.LockableFiles(repository_transport,
1907
'lock', lockdir.LockDir)
1908
control_files.create_lock()
1909
return control_files
1911
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1912
"""Upload the initial blank content."""
1913
control_files = self._create_control_files(a_bzrdir)
1914
control_files.lock_write()
1916
control_files._transport.mkdir_multi(dirs,
1917
mode=control_files._dir_mode)
1918
for file, content in files:
1919
control_files.put(file, content)
1920
for file, content in utf8_files:
1921
control_files.put_utf8(file, content)
1923
control_files.put_utf8('shared-storage', '')
1925
control_files.unlock()
1928
# formats which have no format string are not discoverable
1929
# and not independently creatable, so are not registered. They're
1930
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1931
# needed, it's constructed directly by the BzrDir. Non-native formats where
1932
# the repository is not separately opened are similar.
1934
format_registry.register_lazy(
1935
'Bazaar-NG Repository format 7',
1936
'bzrlib.repofmt.weaverepo',
1940
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1941
# default control directory format
1942
format_registry.register_lazy(
1943
'Bazaar-NG Knit Repository Format 1',
1944
'bzrlib.repofmt.knitrepo',
1945
'RepositoryFormatKnit1',
1947
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1949
format_registry.register_lazy(
1950
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1951
'bzrlib.repofmt.knitrepo',
1952
'RepositoryFormatKnit3',
1956
class InterRepository(InterObject):
1957
"""This class represents operations taking place between two repositories.
1959
Its instances have methods like copy_content and fetch, and contain
1960
references to the source and target repositories these operations can be
1963
Often we will provide convenience methods on 'repository' which carry out
1964
operations with another repository - they will always forward to
1965
InterRepository.get(other).method_name(parameters).
1969
"""The available optimised InterRepository types."""
1971
def copy_content(self, revision_id=None):
1972
raise NotImplementedError(self.copy_content)
1974
def fetch(self, revision_id=None, pb=None):
1975
"""Fetch the content required to construct revision_id.
1977
The content is copied from self.source to self.target.
1979
:param revision_id: if None all content is copied, if NULL_REVISION no
1981
:param pb: optional progress bar to use for progress reports. If not
1982
provided a default one will be created.
1984
Returns the copied revision count and the failed revisions in a tuple:
1987
raise NotImplementedError(self.fetch)
1990
def missing_revision_ids(self, revision_id=None):
1991
"""Return the revision ids that source has that target does not.
1993
These are returned in topological order.
1995
:param revision_id: only return revision ids included by this
1998
# generic, possibly worst case, slow code path.
1999
target_ids = set(self.target.all_revision_ids())
2000
if revision_id is not None:
2001
# TODO: jam 20070210 InterRepository is internal enough that it
2002
# should assume revision_ids are already utf-8
2003
revision_id = osutils.safe_revision_id(revision_id)
2004
source_ids = self.source.get_ancestry(revision_id)
2005
assert source_ids[0] is None
2008
source_ids = self.source.all_revision_ids()
2009
result_set = set(source_ids).difference(target_ids)
2010
# this may look like a no-op: its not. It preserves the ordering
2011
# other_ids had while only returning the members from other_ids
2012
# that we've decided we need.
2013
return [rev_id for rev_id in source_ids if rev_id in result_set]
2016
def _same_model(source, target):
2017
"""True if source and target have the same data representation."""
2018
if source.supports_rich_root() != target.supports_rich_root():
2020
if source._serializer != target._serializer:
2025
class InterSameDataRepository(InterRepository):
2026
"""Code for converting between repositories that represent the same data.
2028
Data format and model must match for this to work.
2032
def _get_repo_format_to_test(self):
2033
"""Repository format for testing with.
2035
InterSameData can pull from subtree to subtree and from non-subtree to
2036
non-subtree, so we test this with the richest repository format.
2038
from bzrlib.repofmt import knitrepo
2039
return knitrepo.RepositoryFormatKnit3()
2042
def is_compatible(source, target):
2043
return InterRepository._same_model(source, target)
2046
def copy_content(self, revision_id=None):
2047
"""Make a complete copy of the content in self into destination.
2049
This copies both the repository's revision data, and configuration information
2050
such as the make_working_trees setting.
2052
This is a destructive operation! Do not use it on existing
2055
:param revision_id: Only copy the content needed to construct
2056
revision_id and its parents.
2059
self.target.set_make_working_trees(self.source.make_working_trees())
2060
except NotImplementedError:
2062
# TODO: jam 20070210 This is fairly internal, so we should probably
2063
# just assert that revision_id is not unicode.
2064
revision_id = osutils.safe_revision_id(revision_id)
2065
# but don't bother fetching if we have the needed data now.
2066
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2067
self.target.has_revision(revision_id)):
2069
self.target.fetch(self.source, revision_id=revision_id)
2072
def fetch(self, revision_id=None, pb=None):
2073
"""See InterRepository.fetch()."""
2074
from bzrlib.fetch import GenericRepoFetcher
2075
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2076
self.source, self.source._format, self.target,
2077
self.target._format)
2078
# TODO: jam 20070210 This should be an assert, not a translate
2079
revision_id = osutils.safe_revision_id(revision_id)
2080
f = GenericRepoFetcher(to_repository=self.target,
2081
from_repository=self.source,
2082
last_revision=revision_id,
2084
return f.count_copied, f.failed_revisions
2087
class InterWeaveRepo(InterSameDataRepository):
2088
"""Optimised code paths between Weave based repositories.
2090
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2091
implemented lazy inter-object optimisation.
2095
def _get_repo_format_to_test(self):
2096
from bzrlib.repofmt import weaverepo
2097
return weaverepo.RepositoryFormat7()
2100
def is_compatible(source, target):
2101
"""Be compatible with known Weave formats.
2103
We don't test for the stores being of specific types because that
2104
could lead to confusing results, and there is no need to be
2107
from bzrlib.repofmt.weaverepo import (
2113
return (isinstance(source._format, (RepositoryFormat5,
2115
RepositoryFormat7)) and
2116
isinstance(target._format, (RepositoryFormat5,
2118
RepositoryFormat7)))
2119
except AttributeError:
2123
def copy_content(self, revision_id=None):
2124
"""See InterRepository.copy_content()."""
2125
# weave specific optimised path:
2126
# TODO: jam 20070210 Internal, should be an assert, not translate
2127
revision_id = osutils.safe_revision_id(revision_id)
2129
self.target.set_make_working_trees(self.source.make_working_trees())
2130
except NotImplementedError:
2132
# FIXME do not peek!
2133
if self.source.control_files._transport.listable():
2134
pb = ui.ui_factory.nested_progress_bar()
2136
self.target.weave_store.copy_all_ids(
2137
self.source.weave_store,
2139
from_transaction=self.source.get_transaction(),
2140
to_transaction=self.target.get_transaction())
2141
pb.update('copying inventory', 0, 1)
2142
self.target.control_weaves.copy_multi(
2143
self.source.control_weaves, ['inventory'],
2144
from_transaction=self.source.get_transaction(),
2145
to_transaction=self.target.get_transaction())
2146
self.target._revision_store.text_store.copy_all_ids(
2147
self.source._revision_store.text_store,
2152
self.target.fetch(self.source, revision_id=revision_id)
2155
def fetch(self, revision_id=None, pb=None):
2156
"""See InterRepository.fetch()."""
2157
from bzrlib.fetch import GenericRepoFetcher
2158
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2159
self.source, self.source._format, self.target, self.target._format)
2160
# TODO: jam 20070210 This should be an assert, not a translate
2161
revision_id = osutils.safe_revision_id(revision_id)
2162
f = GenericRepoFetcher(to_repository=self.target,
2163
from_repository=self.source,
2164
last_revision=revision_id,
2166
return f.count_copied, f.failed_revisions
2169
def missing_revision_ids(self, revision_id=None):
2170
"""See InterRepository.missing_revision_ids()."""
2171
# we want all revisions to satisfy revision_id in source.
2172
# but we don't want to stat every file here and there.
2173
# we want then, all revisions other needs to satisfy revision_id
2174
# checked, but not those that we have locally.
2175
# so the first thing is to get a subset of the revisions to
2176
# satisfy revision_id in source, and then eliminate those that
2177
# we do already have.
2178
# this is slow on high latency connection to self, but as as this
2179
# disk format scales terribly for push anyway due to rewriting
2180
# inventory.weave, this is considered acceptable.
2182
if revision_id is not None:
2183
source_ids = self.source.get_ancestry(revision_id)
2184
assert source_ids[0] is None
2187
source_ids = self.source._all_possible_ids()
2188
source_ids_set = set(source_ids)
2189
# source_ids is the worst possible case we may need to pull.
2190
# now we want to filter source_ids against what we actually
2191
# have in target, but don't try to check for existence where we know
2192
# we do not have a revision as that would be pointless.
2193
target_ids = set(self.target._all_possible_ids())
2194
possibly_present_revisions = target_ids.intersection(source_ids_set)
2195
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2196
required_revisions = source_ids_set.difference(actually_present_revisions)
2197
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2198
if revision_id is not None:
2199
# we used get_ancestry to determine source_ids then we are assured all
2200
# revisions referenced are present as they are installed in topological order.
2201
# and the tip revision was validated by get_ancestry.
2202
return required_topo_revisions
2204
# if we just grabbed the possibly available ids, then
2205
# we only have an estimate of whats available and need to validate
2206
# that against the revision records.
2207
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2210
class InterKnitRepo(InterSameDataRepository):
2211
"""Optimised code paths between Knit based repositories."""
2214
def _get_repo_format_to_test(self):
2215
from bzrlib.repofmt import knitrepo
2216
return knitrepo.RepositoryFormatKnit1()
2219
def is_compatible(source, target):
2220
"""Be compatible with known Knit formats.
2222
We don't test for the stores being of specific types because that
2223
could lead to confusing results, and there is no need to be
2226
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2228
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2229
isinstance(target._format, RepositoryFormatKnit))
2230
except AttributeError:
2232
return are_knits and InterRepository._same_model(source, target)
2235
def fetch(self, revision_id=None, pb=None):
2236
"""See InterRepository.fetch()."""
2237
from bzrlib.fetch import KnitRepoFetcher
2238
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2239
self.source, self.source._format, self.target, self.target._format)
2240
# TODO: jam 20070210 This should be an assert, not a translate
2241
revision_id = osutils.safe_revision_id(revision_id)
2242
f = KnitRepoFetcher(to_repository=self.target,
2243
from_repository=self.source,
2244
last_revision=revision_id,
2246
return f.count_copied, f.failed_revisions
2249
def missing_revision_ids(self, revision_id=None):
2250
"""See InterRepository.missing_revision_ids()."""
2251
if revision_id is not None:
2252
source_ids = self.source.get_ancestry(revision_id)
2253
assert source_ids[0] is None
2256
source_ids = self.source.all_revision_ids()
2257
source_ids_set = set(source_ids)
2258
# source_ids is the worst possible case we may need to pull.
2259
# now we want to filter source_ids against what we actually
2260
# have in target, but don't try to check for existence where we know
2261
# we do not have a revision as that would be pointless.
2262
target_ids = set(self.target.all_revision_ids())
2263
possibly_present_revisions = target_ids.intersection(source_ids_set)
2264
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2265
required_revisions = source_ids_set.difference(actually_present_revisions)
2266
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2267
if revision_id is not None:
2268
# we used get_ancestry to determine source_ids then we are assured all
2269
# revisions referenced are present as they are installed in topological order.
2270
# and the tip revision was validated by get_ancestry.
2271
return required_topo_revisions
2273
# if we just grabbed the possibly available ids, then
2274
# we only have an estimate of whats available and need to validate
2275
# that against the revision records.
2276
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2279
class InterModel1and2(InterRepository):
2282
def _get_repo_format_to_test(self):
2286
def is_compatible(source, target):
2287
if not source.supports_rich_root() and target.supports_rich_root():
2293
def fetch(self, revision_id=None, pb=None):
2294
"""See InterRepository.fetch()."""
2295
from bzrlib.fetch import Model1toKnit2Fetcher
2296
# TODO: jam 20070210 This should be an assert, not a translate
2297
revision_id = osutils.safe_revision_id(revision_id)
2298
f = Model1toKnit2Fetcher(to_repository=self.target,
2299
from_repository=self.source,
2300
last_revision=revision_id,
2302
return f.count_copied, f.failed_revisions
2305
def copy_content(self, revision_id=None):
2306
"""Make a complete copy of the content in self into destination.
2308
This is a destructive operation! Do not use it on existing
2311
:param revision_id: Only copy the content needed to construct
2312
revision_id and its parents.
2315
self.target.set_make_working_trees(self.source.make_working_trees())
2316
except NotImplementedError:
2318
# TODO: jam 20070210 Internal, assert, don't translate
2319
revision_id = osutils.safe_revision_id(revision_id)
2320
# but don't bother fetching if we have the needed data now.
2321
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2322
self.target.has_revision(revision_id)):
2324
self.target.fetch(self.source, revision_id=revision_id)
2327
class InterKnit1and2(InterKnitRepo):
2330
def _get_repo_format_to_test(self):
2334
def is_compatible(source, target):
2335
"""Be compatible with Knit1 source and Knit3 target"""
2336
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2338
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
2339
RepositoryFormatKnit3
2340
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2341
isinstance(target._format, (RepositoryFormatKnit3)))
2342
except AttributeError:
2346
def fetch(self, revision_id=None, pb=None):
2347
"""See InterRepository.fetch()."""
2348
from bzrlib.fetch import Knit1to2Fetcher
2349
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2350
self.source, self.source._format, self.target,
2351
self.target._format)
2352
# TODO: jam 20070210 This should be an assert, not a translate
2353
revision_id = osutils.safe_revision_id(revision_id)
2354
f = Knit1to2Fetcher(to_repository=self.target,
2355
from_repository=self.source,
2356
last_revision=revision_id,
2358
return f.count_copied, f.failed_revisions
2361
class InterRemoteToOther(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 not isinstance(source, remote.RemoteRepository):
2371
source._ensure_real()
2372
real_source = source._real_repository
2373
# Is source's model compatible with target's model, and are they the
2374
# same format? Currently we can only optimise fetching from an
2375
# identical model & format repo.
2376
assert not isinstance(real_source, remote.RemoteRepository), (
2377
"We don't support remote repos backed by remote repos yet.")
2378
return real_source._format == target._format
2381
def fetch(self, revision_id=None, pb=None):
2382
"""See InterRepository.fetch()."""
2383
from bzrlib.fetch import RemoteToOtherFetcher
2384
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2385
self.source, self.target, self.target._format)
2386
# TODO: jam 20070210 This should be an assert, not a translate
2387
revision_id = osutils.safe_revision_id(revision_id)
2388
f = RemoteToOtherFetcher(to_repository=self.target,
2389
from_repository=self.source,
2390
last_revision=revision_id,
2392
return f.count_copied, f.failed_revisions
2395
def _get_repo_format_to_test(self):
2399
class InterOtherToRemote(InterRepository):
2401
def __init__(self, source, target):
2402
InterRepository.__init__(self, source, target)
2403
self._real_inter = None
2406
def is_compatible(source, target):
2407
if isinstance(target, remote.RemoteRepository):
2411
def _ensure_real_inter(self):
2412
if self._real_inter is None:
2413
self.target._ensure_real()
2414
real_target = self.target._real_repository
2415
self._real_inter = InterRepository.get(self.source, real_target)
2417
def copy_content(self, revision_id=None):
2418
self._ensure_real_inter()
2419
self._real_inter.copy_content(revision_id=revision_id)
2421
def fetch(self, revision_id=None, pb=None):
2422
self._ensure_real_inter()
2423
self._real_inter.fetch(revision_id=revision_id, pb=pb)
2426
def _get_repo_format_to_test(self):
2430
InterRepository.register_optimiser(InterSameDataRepository)
2431
InterRepository.register_optimiser(InterWeaveRepo)
2432
InterRepository.register_optimiser(InterKnitRepo)
2433
InterRepository.register_optimiser(InterModel1and2)
2434
InterRepository.register_optimiser(InterKnit1and2)
2435
InterRepository.register_optimiser(InterRemoteToOther)
2436
InterRepository.register_optimiser(InterOtherToRemote)
2439
class CopyConverter(object):
2440
"""A repository conversion tool which just performs a copy of the content.
2442
This is slow but quite reliable.
2445
def __init__(self, target_format):
2446
"""Create a CopyConverter.
2448
:param target_format: The format the resulting repository should be.
2450
self.target_format = target_format
2452
def convert(self, repo, pb):
2453
"""Perform the conversion of to_convert, giving feedback via pb.
2455
:param to_convert: The disk object to convert.
2456
:param pb: a progress bar to use for progress information.
2461
# this is only useful with metadir layouts - separated repo content.
2462
# trigger an assertion if not such
2463
repo._format.get_format_string()
2464
self.repo_dir = repo.bzrdir
2465
self.step('Moving repository to repository.backup')
2466
self.repo_dir.transport.move('repository', 'repository.backup')
2467
backup_transport = self.repo_dir.transport.clone('repository.backup')
2468
repo._format.check_conversion_target(self.target_format)
2469
self.source_repo = repo._format.open(self.repo_dir,
2471
_override_transport=backup_transport)
2472
self.step('Creating new repository')
2473
converted = self.target_format.initialize(self.repo_dir,
2474
self.source_repo.is_shared())
2475
converted.lock_write()
2477
self.step('Copying content into repository.')
2478
self.source_repo.copy_content_into(converted)
2481
self.step('Deleting old repository content.')
2482
self.repo_dir.transport.delete_tree('repository.backup')
2483
self.pb.note('repository converted')
2485
def step(self, message):
2486
"""Update the pb by a step."""
2488
self.pb.update(message, self.count, self.total)
2500
def _unescaper(match, _map=_unescape_map):
2501
code = match.group(1)
2505
if not code.startswith('#'):
2507
return unichr(int(code[1:])).encode('utf8')
2513
def _unescape_xml(data):
2514
"""Unescape predefined XML entities in a string of data."""
2516
if _unescape_re is None:
2517
_unescape_re = re.compile('\&([^;]*);')
2518
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
