1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
44
from bzrlib.bundle import serializer
45
from bzrlib.revisiontree import RevisionTree
46
from bzrlib.store.versioned import VersionedFileStore
47
from bzrlib.store.text import TextStore
48
from bzrlib.testament import Testament
51
from bzrlib.decorators import needs_read_lock, needs_write_lock
52
from bzrlib.inter import InterObject
53
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
54
from bzrlib.symbol_versioning import (
57
from bzrlib.trace import mutter, mutter_callsite, note, warning
60
# Old formats display a warning, but only once
61
_deprecation_warning_done = False
64
class CommitBuilder(object):
65
"""Provides an interface to build up a commit.
67
This allows describing a tree to be committed without needing to
68
know the internals of the format of the repository.
71
# all clients should supply tree roots.
72
record_root_entry = True
73
# the default CommitBuilder does not manage trees whose root is versioned.
74
_versioned_root = False
76
def __init__(self, repository, parents, config, timestamp=None,
77
timezone=None, committer=None, revprops=None,
79
"""Initiate a CommitBuilder.
81
:param repository: Repository to commit to.
82
:param parents: Revision ids of the parents of the new revision.
83
:param config: Configuration to use.
84
:param timestamp: Optional timestamp recorded for commit.
85
:param timezone: Optional timezone for timestamp.
86
:param committer: Optional committer to set for commit.
87
:param revprops: Optional dictionary of revision properties.
88
:param revision_id: Optional revision id.
93
self._committer = self._config.username()
95
assert isinstance(committer, basestring), type(committer)
96
self._committer = committer
98
self.new_inventory = Inventory(None)
99
self._new_revision_id = osutils.safe_revision_id(revision_id)
100
self.parents = parents
101
self.repository = repository
104
if revprops is not None:
105
self._revprops.update(revprops)
107
if timestamp is None:
108
timestamp = time.time()
109
# Restrict resolution to 1ms
110
self._timestamp = round(timestamp, 3)
113
self._timezone = osutils.local_time_offset()
115
self._timezone = int(timezone)
117
self._generate_revision_if_needed()
118
self._repo_graph = repository.get_graph()
120
def commit(self, message):
121
"""Make the actual commit.
123
:return: The revision id of the recorded revision.
125
rev = _mod_revision.Revision(
126
timestamp=self._timestamp,
127
timezone=self._timezone,
128
committer=self._committer,
130
inventory_sha1=self.inv_sha1,
131
revision_id=self._new_revision_id,
132
properties=self._revprops)
133
rev.parent_ids = self.parents
134
self.repository.add_revision(self._new_revision_id, rev,
135
self.new_inventory, self._config)
136
self.repository.commit_write_group()
137
return self._new_revision_id
140
"""Abort the commit that is being built.
142
self.repository.abort_write_group()
144
def revision_tree(self):
145
"""Return the tree that was just committed.
147
After calling commit() this can be called to get a RevisionTree
148
representing the newly committed tree. This is preferred to
149
calling Repository.revision_tree() because that may require
150
deserializing the inventory, while we already have a copy in
153
return RevisionTree(self.repository, self.new_inventory,
154
self._new_revision_id)
156
def finish_inventory(self):
157
"""Tell the builder that the inventory is finished."""
158
if self.new_inventory.root is None:
159
symbol_versioning.warn('Root entry should be supplied to'
160
' record_entry_contents, as of bzr 0.10.',
161
DeprecationWarning, stacklevel=2)
162
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
163
self.new_inventory.revision_id = self._new_revision_id
164
self.inv_sha1 = self.repository.add_inventory(
165
self._new_revision_id,
170
def _gen_revision_id(self):
171
"""Return new revision-id."""
172
return generate_ids.gen_revision_id(self._config.username(),
175
def _generate_revision_if_needed(self):
176
"""Create a revision id if None was supplied.
178
If the repository can not support user-specified revision ids
179
they should override this function and raise CannotSetRevisionId
180
if _new_revision_id is not None.
182
:raises: CannotSetRevisionId
184
if self._new_revision_id is None:
185
self._new_revision_id = self._gen_revision_id()
186
self.random_revid = True
188
self.random_revid = False
190
def _check_root(self, ie, parent_invs, tree):
191
"""Helper for record_entry_contents.
193
:param ie: An entry being added.
194
:param parent_invs: The inventories of the parent revisions of the
196
:param tree: The tree that is being committed.
198
# In this revision format, root entries have no knit or weave When
199
# serializing out to disk and back in root.revision is always
201
ie.revision = self._new_revision_id
203
def _get_delta(self, ie, basis_inv, path):
204
"""Get a delta against the basis inventory for ie."""
205
if ie.file_id not in basis_inv:
207
return (None, path, ie.file_id, ie)
208
elif ie != basis_inv[ie.file_id]:
210
# TODO: avoid tis id2path call.
211
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
216
def record_entry_contents(self, ie, parent_invs, path, tree,
218
"""Record the content of ie from tree into the commit if needed.
220
Side effect: sets ie.revision when unchanged
222
:param ie: An inventory entry present in the commit.
223
:param parent_invs: The inventories of the parent revisions of the
225
:param path: The path the entry is at in the tree.
226
:param tree: The tree which contains this entry and should be used to
228
:param content_summary: Summary data from the tree about the paths
229
content - stat, length, exec, sha/link target. This is only
230
accessed when the entry has a revision of None - that is when it is
231
a candidate to commit.
232
:return: A tuple (change_delta, version_recorded). change_delta is
233
an inventory_delta change for this entry against the basis tree of
234
the commit, or None if no change occured against the basis tree.
235
version_recorded is True if a new version of the entry has been
236
recorded. For instance, committing a merge where a file was only
237
changed on the other side will return (delta, False).
239
if self.new_inventory.root is None:
240
if ie.parent_id is not None:
241
raise errors.RootMissing()
242
self._check_root(ie, parent_invs, tree)
243
if ie.revision is None:
244
kind = content_summary[0]
246
# ie is carried over from a prior commit
248
# XXX: repository specific check for nested tree support goes here - if
249
# the repo doesn't want nested trees we skip it ?
250
if (kind == 'tree-reference' and
251
not self.repository._format.supports_tree_reference):
252
# mismatch between commit builder logic and repository:
253
# this needs the entry creation pushed down into the builder.
254
raise NotImplementedError('Missing repository subtree support.')
255
# transitional assert only, will remove before release.
256
assert ie.kind == kind
257
self.new_inventory.add(ie)
259
# TODO: slow, take it out of the inner loop.
261
basis_inv = parent_invs[0]
263
basis_inv = Inventory(root_id=None)
265
# ie.revision is always None if the InventoryEntry is considered
266
# for committing. We may record the previous parents revision if the
267
# content is actually unchanged against a sole head.
268
if ie.revision is not None:
269
if self._versioned_root or path != '':
270
# not considered for commit
273
# repositories that do not version the root set the root's
274
# revision to the new commit even when no change occurs, and
275
# this masks when a change may have occurred against the basis,
276
# so calculate if one happened.
277
if ie.file_id not in basis_inv:
279
delta = (None, path, ie.file_id, ie)
281
basis_id = basis_inv[ie.file_id]
282
if basis_id.name != '':
284
delta = (basis_inv.id2path(ie.file_id), path,
289
# not considered for commit, OR, for non-rich-root
290
return delta, ie.revision == self._new_revision_id and (path != '' or
291
self._versioned_root)
293
# XXX: Friction: parent_candidates should return a list not a dict
294
# so that we don't have to walk the inventories again.
295
parent_candiate_entries = ie.parent_candidates(parent_invs)
296
head_set = self._repo_graph.heads(parent_candiate_entries.keys())
298
for inv in parent_invs:
299
if ie.file_id in inv:
300
old_rev = inv[ie.file_id].revision
301
if old_rev in head_set:
302
heads.append(inv[ie.file_id].revision)
303
head_set.remove(inv[ie.file_id].revision)
306
# now we check to see if we need to write a new record to the
308
# We write a new entry unless there is one head to the ancestors, and
309
# the kind-derived content is unchanged.
311
# Cheapest check first: no ancestors, or more the one head in the
312
# ancestors, we write a new node.
316
# There is a single head, look it up for comparison
317
parent_entry = parent_candiate_entries[heads[0]]
318
# if the non-content specific data has changed, we'll be writing a
320
if (parent_entry.parent_id != ie.parent_id or
321
parent_entry.name != ie.name):
323
# now we need to do content specific checks:
325
# if the kind changed the content obviously has
326
if kind != parent_entry.kind:
330
if (# if the file length changed we have to store:
331
parent_entry.text_size != content_summary[1] or
332
# if the exec bit has changed we have to store:
333
parent_entry.executable != content_summary[2]):
335
elif parent_entry.text_sha1 == content_summary[3]:
336
# all meta and content is unchanged (using a hash cache
337
# hit to check the sha)
338
ie.revision = parent_entry.revision
339
ie.text_size = parent_entry.text_size
340
ie.text_sha1 = parent_entry.text_sha1
341
ie.executable = parent_entry.executable
342
return self._get_delta(ie, basis_inv, path), False
344
# Either there is only a hash change(no hash cache entry,
345
# or same size content change), or there is no change on
347
# Provide the parent's hash to the store layer, so that the
348
# content is unchanged we will not store a new node.
349
nostore_sha = parent_entry.text_sha1
351
# We want to record a new node regardless of the presence or
352
# absence of a content change in the file.
354
ie.executable = content_summary[2]
355
lines = tree.get_file(ie.file_id, path).readlines()
357
ie.text_sha1, ie.text_size = self._add_text_to_weave(
358
ie.file_id, lines, heads, nostore_sha)
359
except errors.ExistingContent:
360
# Turns out that the file content was unchanged, and we were
361
# only going to store a new node if it was changed. Carry over
363
ie.revision = parent_entry.revision
364
ie.text_size = parent_entry.text_size
365
ie.text_sha1 = parent_entry.text_sha1
366
ie.executable = parent_entry.executable
367
return self._get_delta(ie, basis_inv, path), False
368
elif kind == 'directory':
370
# all data is meta here, nothing specific to directory, so
372
ie.revision = parent_entry.revision
373
return self._get_delta(ie, basis_inv, path), False
375
self._add_text_to_weave(ie.file_id, lines, heads, None)
376
elif kind == 'symlink':
377
current_link_target = content_summary[3]
379
# symlink target is not generic metadata, check if it has
381
if current_link_target != parent_entry.symlink_target:
384
# unchanged, carry over.
385
ie.revision = parent_entry.revision
386
ie.symlink_target = parent_entry.symlink_target
387
return self._get_delta(ie, basis_inv, path), False
388
ie.symlink_target = current_link_target
390
self._add_text_to_weave(ie.file_id, lines, heads, None)
391
elif kind == 'tree-reference':
393
if content_summary[3] != parent_entry.reference_revision:
396
# unchanged, carry over.
397
ie.reference_revision = parent_entry.reference_revision
398
ie.revision = parent_entry.revision
399
return self._get_delta(ie, basis_inv, path), False
400
ie.reference_revision = content_summary[3]
402
self._add_text_to_weave(ie.file_id, lines, heads, None)
404
raise NotImplementedError('unknown kind')
405
ie.revision = self._new_revision_id
406
return self._get_delta(ie, basis_inv, path), True
408
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
409
versionedfile = self.repository.weave_store.get_weave_or_empty(
410
file_id, self.repository.get_transaction())
411
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
412
# than add_lines, and allows committing when a parent is ghosted for
414
# Note: as we read the content directly from the tree, we know its not
415
# been turned into unicode or badly split - but a broken tree
416
# implementation could give us bad output from readlines() so this is
417
# not a guarantee of safety. What would be better is always checking
418
# the content during test suite execution. RBC 20070912
420
return versionedfile.add_lines_with_ghosts(
421
self._new_revision_id, parents, new_lines,
422
nostore_sha=nostore_sha, random_id=self.random_revid,
423
check_content=False)[0:2]
425
versionedfile.clear_cache()
428
class RootCommitBuilder(CommitBuilder):
429
"""This commitbuilder actually records the root id"""
431
# the root entry gets versioned properly by this builder.
432
_versioned_root = True
434
def _check_root(self, ie, parent_invs, tree):
435
"""Helper for record_entry_contents.
437
:param ie: An entry being added.
438
:param parent_invs: The inventories of the parent revisions of the
440
:param tree: The tree that is being committed.
444
######################################################################
447
class Repository(object):
448
"""Repository holding history for one or more branches.
450
The repository holds and retrieves historical information including
451
revisions and file history. It's normally accessed only by the Branch,
452
which views a particular line of development through that history.
454
The Repository builds on top of Stores and a Transport, which respectively
455
describe the disk data format and the way of accessing the (possibly
459
# What class to use for a CommitBuilder. Often its simpler to change this
460
# in a Repository class subclass rather than to override
461
# get_commit_builder.
462
_commit_builder_class = CommitBuilder
463
# The search regex used by xml based repositories to determine what things
464
# where changed in a single commit.
465
_file_ids_altered_regex = lazy_regex.lazy_compile(
466
r'file_id="(?P<file_id>[^"]+)"'
467
r'.* revision="(?P<revision_id>[^"]+)"'
470
def abort_write_group(self):
471
"""Commit the contents accrued within the current write group.
473
:seealso: start_write_group.
475
if self._write_group is not self.get_transaction():
476
# has an unlock or relock occured ?
477
raise errors.BzrError('mismatched lock context and write group.')
478
self._abort_write_group()
479
self._write_group = None
481
def _abort_write_group(self):
482
"""Template method for per-repository write group cleanup.
484
This is called during abort before the write group is considered to be
485
finished and should cleanup any internal state accrued during the write
486
group. There is no requirement that data handed to the repository be
487
*not* made available - this is not a rollback - but neither should any
488
attempt be made to ensure that data added is fully commited. Abort is
489
invoked when an error has occured so futher disk or network operations
490
may not be possible or may error and if possible should not be
495
def add_inventory(self, revision_id, inv, parents):
496
"""Add the inventory inv to the repository as revision_id.
498
:param parents: The revision ids of the parents that revision_id
499
is known to have and are in the repository already.
501
returns the sha1 of the serialized inventory.
503
assert self.is_in_write_group()
504
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
762
def missing_revision_ids(self, other, revision_id=None):
763
"""Return the revision ids that other has that this does not.
765
These are returned in topological order.
767
revision_id: only return revision ids included by revision_id.
769
revision_id = osutils.safe_revision_id(revision_id)
770
return InterRepository.get(other, self).missing_revision_ids(revision_id)
774
"""Open the repository rooted at base.
776
For instance, if the repository is at URL/.bzr/repository,
777
Repository.open(URL) -> a Repository instance.
779
control = bzrdir.BzrDir.open(base)
780
return control.open_repository()
782
def copy_content_into(self, destination, revision_id=None):
783
"""Make a complete copy of the content in self into destination.
785
This is a destructive operation! Do not use it on existing
788
revision_id = osutils.safe_revision_id(revision_id)
789
return InterRepository.get(self, destination).copy_content(revision_id)
791
def commit_write_group(self):
792
"""Commit the contents accrued within the current write group.
794
:seealso: start_write_group.
796
if self._write_group is not self.get_transaction():
797
# has an unlock or relock occured ?
798
raise errors.BzrError('mismatched lock context %r and '
800
(self.get_transaction(), self._write_group))
801
self._commit_write_group()
802
self._write_group = None
804
def _commit_write_group(self):
805
"""Template method for per-repository write group cleanup.
807
This is called before the write group is considered to be
808
finished and should ensure that all data handed to the repository
809
for writing during the write group is safely committed (to the
810
extent possible considering file system caching etc).
813
def fetch(self, source, revision_id=None, pb=None):
814
"""Fetch the content required to construct revision_id from source.
816
If revision_id is None all content is copied.
818
revision_id = osutils.safe_revision_id(revision_id)
819
# fast path same-url fetch operations
820
if self.has_same_location(source):
821
# check that last_revision is in 'from' and then return a
823
if (revision_id is not None and
824
not _mod_revision.is_null(revision_id)):
825
self.get_revision(revision_id)
827
inter = InterRepository.get(source, self)
829
return inter.fetch(revision_id=revision_id, pb=pb)
830
except NotImplementedError:
831
raise errors.IncompatibleRepositories(source, self)
833
def create_bundle(self, target, base, fileobj, format=None):
834
return serializer.write_bundle(self, target, base, fileobj, format)
836
def get_commit_builder(self, branch, parents, config, timestamp=None,
837
timezone=None, committer=None, revprops=None,
839
"""Obtain a CommitBuilder for this repository.
841
:param branch: Branch to commit to.
842
:param parents: Revision ids of the parents of the new revision.
843
:param config: Configuration to use.
844
:param timestamp: Optional timestamp recorded for commit.
845
:param timezone: Optional timezone for timestamp.
846
:param committer: Optional committer to set for commit.
847
:param revprops: Optional dictionary of revision properties.
848
:param revision_id: Optional revision id.
850
revision_id = osutils.safe_revision_id(revision_id)
851
result = self._commit_builder_class(self, parents, config,
852
timestamp, timezone, committer, revprops, revision_id)
853
self.start_write_group()
857
if (self.control_files._lock_count == 1 and
858
self.control_files._lock_mode == 'w'):
859
if self._write_group is not None:
860
raise errors.BzrError(
861
'Must end write groups before releasing write locks.')
862
self.control_files.unlock()
865
def clone(self, a_bzrdir, revision_id=None):
866
"""Clone this repository into a_bzrdir using the current format.
868
Currently no check is made that the format of this repository and
869
the bzrdir format are compatible. FIXME RBC 20060201.
871
:return: The newly created destination repository.
873
# TODO: deprecate after 0.16; cloning this with all its settings is
874
# probably not very useful -- mbp 20070423
875
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
876
self.copy_content_into(dest_repo, revision_id)
879
def start_write_group(self):
880
"""Start a write group in the repository.
882
Write groups are used by repositories which do not have a 1:1 mapping
883
between file ids and backend store to manage the insertion of data from
884
both fetch and commit operations.
886
A write lock is required around the start_write_group/commit_write_group
887
for the support of lock-requiring repository formats.
889
One can only insert data into a repository inside a write group.
893
if not self.is_locked() or self.control_files._lock_mode != 'w':
894
raise errors.NotWriteLocked(self)
895
if self._write_group:
896
raise errors.BzrError('already in a write group')
897
self._start_write_group()
898
# so we can detect unlock/relock - the write group is now entered.
899
self._write_group = self.get_transaction()
901
def _start_write_group(self):
902
"""Template method for per-repository write group startup.
904
This is called before the write group is considered to be
909
def sprout(self, to_bzrdir, revision_id=None):
910
"""Create a descendent repository for new development.
912
Unlike clone, this does not copy the settings of the repository.
914
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
915
dest_repo.fetch(self, revision_id=revision_id)
918
def _create_sprouting_repo(self, a_bzrdir, shared):
919
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
920
# use target default format.
921
dest_repo = a_bzrdir.create_repository()
923
# Most control formats need the repository to be specifically
924
# created, but on some old all-in-one formats it's not needed
926
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
927
except errors.UninitializableFormat:
928
dest_repo = a_bzrdir.open_repository()
932
def has_revision(self, revision_id):
933
"""True if this repository has a copy of the revision."""
934
if 'evil' in debug.debug_flags:
935
mutter_callsite(3, "has_revision is a LBYL symptom.")
936
revision_id = osutils.safe_revision_id(revision_id)
937
return self._revision_store.has_revision_id(revision_id,
938
self.get_transaction())
941
def get_revision(self, revision_id):
942
"""Return the Revision object for a named revision."""
943
return self.get_revisions([revision_id])[0]
946
def get_revision_reconcile(self, revision_id):
947
"""'reconcile' helper routine that allows access to a revision always.
949
This variant of get_revision does not cross check the weave graph
950
against the revision one as get_revision does: but it should only
951
be used by reconcile, or reconcile-alike commands that are correcting
952
or testing the revision graph.
954
return self._get_revisions([revision_id])[0]
957
def get_revisions(self, revision_ids):
958
"""Get many revisions at once."""
959
return self._get_revisions(revision_ids)
962
def _get_revisions(self, revision_ids):
963
"""Core work logic to get many revisions without sanity checks."""
964
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
965
for rev_id in revision_ids:
966
if not rev_id or not isinstance(rev_id, basestring):
967
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
968
revs = self._revision_store.get_revisions(revision_ids,
969
self.get_transaction())
971
assert not isinstance(rev.revision_id, unicode)
972
for parent_id in rev.parent_ids:
973
assert not isinstance(parent_id, unicode)
977
def get_revision_xml(self, revision_id):
978
# TODO: jam 20070210 This shouldn't be necessary since get_revision
979
# would have already do it.
980
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
981
revision_id = osutils.safe_revision_id(revision_id)
982
rev = self.get_revision(revision_id)
984
# the current serializer..
985
self._revision_store._serializer.write_revision(rev, rev_tmp)
987
return rev_tmp.getvalue()
990
def get_deltas_for_revisions(self, revisions):
991
"""Produce a generator of revision deltas.
993
Note that the input is a sequence of REVISIONS, not revision_ids.
994
Trees will be held in memory until the generator exits.
995
Each delta is relative to the revision's lefthand predecessor.
997
required_trees = set()
998
for revision in revisions:
999
required_trees.add(revision.revision_id)
1000
required_trees.update(revision.parent_ids[:1])
1001
trees = dict((t.get_revision_id(), t) for
1002
t in self.revision_trees(required_trees))
1003
for revision in revisions:
1004
if not revision.parent_ids:
1005
old_tree = self.revision_tree(None)
1007
old_tree = trees[revision.parent_ids[0]]
1008
yield trees[revision.revision_id].changes_from(old_tree)
1011
def get_revision_delta(self, revision_id):
1012
"""Return the delta for one revision.
1014
The delta is relative to the left-hand predecessor of the
1017
r = self.get_revision(revision_id)
1018
return list(self.get_deltas_for_revisions([r]))[0]
1021
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1022
revision_id = osutils.safe_revision_id(revision_id)
1023
signature = gpg_strategy.sign(plaintext)
1024
self._revision_store.add_revision_signature_text(revision_id,
1026
self.get_transaction())
1028
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1030
"""Helper routine for fileids_altered_by_revision_ids.
1032
This performs the translation of xml lines to revision ids.
1034
:param line_iterator: An iterator of lines
1035
:param revision_ids: The revision ids to filter for.
1036
:return: a dictionary mapping altered file-ids to an iterable of
1037
revision_ids. Each altered file-ids has the exact revision_ids that
1038
altered it listed explicitly.
1042
# this code needs to read every new line in every inventory for the
1043
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1044
# not present in one of those inventories is unnecessary but not
1045
# harmful because we are filtering by the revision id marker in the
1046
# inventory lines : we only select file ids altered in one of those
1047
# revisions. We don't need to see all lines in the inventory because
1048
# only those added in an inventory in rev X can contain a revision=X
1050
unescape_revid_cache = {}
1051
unescape_fileid_cache = {}
1053
# jam 20061218 In a big fetch, this handles hundreds of thousands
1054
# of lines, so it has had a lot of inlining and optimizing done.
1055
# Sorry that it is a little bit messy.
1056
# Move several functions to be local variables, since this is a long
1058
search = self._file_ids_altered_regex.search
1059
unescape = _unescape_xml
1060
setdefault = result.setdefault
1061
for line in line_iterator:
1062
match = search(line)
1065
# One call to match.group() returning multiple items is quite a
1066
# bit faster than 2 calls to match.group() each returning 1
1067
file_id, revision_id = match.group('file_id', 'revision_id')
1069
# Inlining the cache lookups helps a lot when you make 170,000
1070
# lines and 350k ids, versus 8.4 unique ids.
1071
# Using a cache helps in 2 ways:
1072
# 1) Avoids unnecessary decoding calls
1073
# 2) Re-uses cached strings, which helps in future set and
1075
# (2) is enough that removing encoding entirely along with
1076
# the cache (so we are using plain strings) results in no
1077
# performance improvement.
1079
revision_id = unescape_revid_cache[revision_id]
1081
unescaped = unescape(revision_id)
1082
unescape_revid_cache[revision_id] = unescaped
1083
revision_id = unescaped
1085
if revision_id in revision_ids:
1087
file_id = unescape_fileid_cache[file_id]
1089
unescaped = unescape(file_id)
1090
unescape_fileid_cache[file_id] = unescaped
1092
setdefault(file_id, set()).add(revision_id)
1095
def fileids_altered_by_revision_ids(self, revision_ids):
1096
"""Find the file ids and versions affected by revisions.
1098
:param revisions: an iterable containing revision ids.
1099
:return: a dictionary mapping altered file-ids to an iterable of
1100
revision_ids. Each altered file-ids has the exact revision_ids that
1101
altered it listed explicitly.
1103
assert self._serializer.support_altered_by_hack, \
1104
("fileids_altered_by_revision_ids only supported for branches "
1105
"which store inventory as unnested xml, not on %r" % self)
1106
selected_revision_ids = set(osutils.safe_revision_id(r)
1107
for r in revision_ids)
1108
w = self.get_inventory_weave()
1109
pb = ui.ui_factory.nested_progress_bar()
1111
return self._find_file_ids_from_xml_inventory_lines(
1112
w.iter_lines_added_or_present_in_versions(
1113
selected_revision_ids, pb=pb),
1114
selected_revision_ids)
1118
def iter_files_bytes(self, desired_files):
1119
"""Iterate through file versions.
1121
Files will not necessarily be returned in the order they occur in
1122
desired_files. No specific order is guaranteed.
1124
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1125
value supplied by the caller as part of desired_files. It should
1126
uniquely identify the file version in the caller's context. (Examples:
1127
an index number or a TreeTransform trans_id.)
1129
bytes_iterator is an iterable of bytestrings for the file. The
1130
kind of iterable and length of the bytestrings are unspecified, but for
1131
this implementation, it is a list of lines produced by
1132
VersionedFile.get_lines().
1134
:param desired_files: a list of (file_id, revision_id, identifier)
1137
transaction = self.get_transaction()
1138
for file_id, revision_id, callable_data in desired_files:
1140
weave = self.weave_store.get_weave(file_id, transaction)
1141
except errors.NoSuchFile:
1142
raise errors.NoSuchIdInRepository(self, file_id)
1143
yield callable_data, weave.get_lines(revision_id)
1145
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1146
"""Get an iterable listing the keys of all the data introduced by a set
1149
The keys will be ordered so that the corresponding items can be safely
1150
fetched and inserted in that order.
1152
:returns: An iterable producing tuples of (knit-kind, file-id,
1153
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1154
'revisions'. file-id is None unless knit-kind is 'file'.
1156
# XXX: it's a bit weird to control the inventory weave caching in this
1157
# generator. Ideally the caching would be done in fetch.py I think. Or
1158
# maybe this generator should explicitly have the contract that it
1159
# should not be iterated until the previously yielded item has been
1161
inv_w = self.get_inventory_weave()
1162
inv_w.enable_cache()
1164
# file ids that changed
1165
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1167
num_file_ids = len(file_ids)
1168
for file_id, altered_versions in file_ids.iteritems():
1169
if _files_pb is not None:
1170
_files_pb.update("fetch texts", count, num_file_ids)
1172
yield ("file", file_id, altered_versions)
1173
# We're done with the files_pb. Note that it finished by the caller,
1174
# just as it was created by the caller.
1178
yield ("inventory", None, revision_ids)
1182
revisions_with_signatures = set()
1183
for rev_id in revision_ids:
1185
self.get_signature_text(rev_id)
1186
except errors.NoSuchRevision:
1190
revisions_with_signatures.add(rev_id)
1191
yield ("signatures", None, revisions_with_signatures)
1194
yield ("revisions", None, revision_ids)
1197
def get_inventory_weave(self):
1198
return self.control_weaves.get_weave('inventory',
1199
self.get_transaction())
1202
def get_inventory(self, revision_id):
1203
"""Get Inventory object by hash."""
1204
# TODO: jam 20070210 Technically we don't need to sanitize, since all
1205
# called functions must sanitize.
1206
revision_id = osutils.safe_revision_id(revision_id)
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
revision_id = osutils.safe_revision_id(revision_id)
1217
result = self._serializer.read_inventory_from_string(xml, revision_id)
1220
def serialise_inventory(self, inv):
1221
return self._serializer.write_inventory_to_string(inv)
1223
def _serialise_inventory_to_lines(self, inv):
1224
return self._serializer.write_inventory_to_lines(inv)
1226
def get_serializer_format(self):
1227
return self._serializer.format_num
1230
def get_inventory_xml(self, revision_id):
1231
"""Get inventory XML as a file object."""
1232
revision_id = osutils.safe_revision_id(revision_id)
1234
assert isinstance(revision_id, str), type(revision_id)
1235
iw = self.get_inventory_weave()
1236
return iw.get_text(revision_id)
1238
raise errors.HistoryMissing(self, 'inventory', revision_id)
1241
def get_inventory_sha1(self, revision_id):
1242
"""Return the sha1 hash of the inventory entry
1244
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
1245
revision_id = osutils.safe_revision_id(revision_id)
1246
return self.get_revision(revision_id).inventory_sha1
1249
def get_revision_graph(self, revision_id=None):
1250
"""Return a dictionary containing the revision graph.
1252
NB: This method should not be used as it accesses the entire graph all
1253
at once, which is much more data than most operations should require.
1255
:param revision_id: The revision_id to get a graph from. If None, then
1256
the entire revision graph is returned. This is a deprecated mode of
1257
operation and will be removed in the future.
1258
:return: a dictionary of revision_id->revision_parents_list.
1260
raise NotImplementedError(self.get_revision_graph)
1263
def get_revision_graph_with_ghosts(self, revision_ids=None):
1264
"""Return a graph of the revisions with ghosts marked as applicable.
1266
:param revision_ids: an iterable of revisions to graph or None for all.
1267
:return: a Graph object with the graph reachable from revision_ids.
1269
if 'evil' in debug.debug_flags:
1271
"get_revision_graph_with_ghosts scales with size of history.")
1272
result = deprecated_graph.Graph()
1273
if not revision_ids:
1274
pending = set(self.all_revision_ids())
1277
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
1278
# special case NULL_REVISION
1279
if _mod_revision.NULL_REVISION in pending:
1280
pending.remove(_mod_revision.NULL_REVISION)
1281
required = set(pending)
1284
revision_id = pending.pop()
1286
rev = self.get_revision(revision_id)
1287
except errors.NoSuchRevision:
1288
if revision_id in required:
1291
result.add_ghost(revision_id)
1293
for parent_id in rev.parent_ids:
1294
# is this queued or done ?
1295
if (parent_id not in pending and
1296
parent_id not in done):
1298
pending.add(parent_id)
1299
result.add_node(revision_id, rev.parent_ids)
1300
done.add(revision_id)
1303
def _get_history_vf(self):
1304
"""Get a versionedfile whose history graph reflects all revisions.
1306
For weave repositories, this is the inventory weave.
1308
return self.get_inventory_weave()
1310
def iter_reverse_revision_history(self, revision_id):
1311
"""Iterate backwards through revision ids in the lefthand history
1313
:param revision_id: The revision id to start with. All its lefthand
1314
ancestors will be traversed.
1316
revision_id = osutils.safe_revision_id(revision_id)
1317
if revision_id in (None, _mod_revision.NULL_REVISION):
1319
next_id = revision_id
1320
versionedfile = self._get_history_vf()
1323
parents = versionedfile.get_parents(next_id)
1324
if len(parents) == 0:
1327
next_id = parents[0]
1330
def get_revision_inventory(self, revision_id):
1331
"""Return inventory of a past revision."""
1332
# TODO: Unify this with get_inventory()
1333
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1334
# must be the same as its revision, so this is trivial.
1335
if revision_id is None:
1336
# This does not make sense: if there is no revision,
1337
# then it is the current tree inventory surely ?!
1338
# and thus get_root_id() is something that looks at the last
1339
# commit on the branch, and the get_root_id is an inventory check.
1340
raise NotImplementedError
1341
# return Inventory(self.get_root_id())
1343
return self.get_inventory(revision_id)
1346
def is_shared(self):
1347
"""Return True if this repository is flagged as a shared repository."""
1348
raise NotImplementedError(self.is_shared)
1351
def reconcile(self, other=None, thorough=False):
1352
"""Reconcile this repository."""
1353
from bzrlib.reconcile import RepoReconciler
1354
reconciler = RepoReconciler(self, thorough=thorough)
1355
reconciler.reconcile()
1358
def _refresh_data(self):
1359
"""Helper called from lock_* to ensure coherency with disk.
1361
The default implementation does nothing; it is however possible
1362
for repositories to maintain loaded indices across multiple locks
1363
by checking inside their implementation of this method to see
1364
whether their indices are still valid. This depends of course on
1365
the disk format being validatable in this manner.
1369
def revision_tree(self, revision_id):
1370
"""Return Tree for a revision on this branch.
1372
`revision_id` may be None for the empty tree revision.
1374
# TODO: refactor this to use an existing revision object
1375
# so we don't need to read it in twice.
1376
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1377
return RevisionTree(self, Inventory(root_id=None),
1378
_mod_revision.NULL_REVISION)
1380
revision_id = osutils.safe_revision_id(revision_id)
1381
inv = self.get_revision_inventory(revision_id)
1382
return RevisionTree(self, inv, revision_id)
1385
def revision_trees(self, revision_ids):
1386
"""Return Tree for a revision on this branch.
1388
`revision_id` may not be None or 'null:'"""
1389
assert None not in revision_ids
1390
assert _mod_revision.NULL_REVISION not in revision_ids
1391
texts = self.get_inventory_weave().get_texts(revision_ids)
1392
for text, revision_id in zip(texts, revision_ids):
1393
inv = self.deserialise_inventory(revision_id, text)
1394
yield RevisionTree(self, inv, revision_id)
1397
def get_ancestry(self, revision_id, topo_sorted=True):
1398
"""Return a list of revision-ids integrated by a revision.
1400
The first element of the list is always None, indicating the origin
1401
revision. This might change when we have history horizons, or
1402
perhaps we should have a new API.
1404
This is topologically sorted.
1406
if _mod_revision.is_null(revision_id):
1408
revision_id = osutils.safe_revision_id(revision_id)
1409
if not self.has_revision(revision_id):
1410
raise errors.NoSuchRevision(self, revision_id)
1411
w = self.get_inventory_weave()
1412
candidates = w.get_ancestry(revision_id, topo_sorted)
1413
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1416
"""Compress the data within the repository.
1418
This operation only makes sense for some repository types. For other
1419
types it should be a no-op that just returns.
1421
This stub method does not require a lock, but subclasses should use
1422
@needs_write_lock as this is a long running call its reasonable to
1423
implicitly lock for the user.
1427
def print_file(self, file, revision_id):
1428
"""Print `file` to stdout.
1430
FIXME RBC 20060125 as John Meinel points out this is a bad api
1431
- it writes to stdout, it assumes that that is valid etc. Fix
1432
by creating a new more flexible convenience function.
1434
revision_id = osutils.safe_revision_id(revision_id)
1435
tree = self.revision_tree(revision_id)
1436
# use inventory as it was in that revision
1437
file_id = tree.inventory.path2id(file)
1439
# TODO: jam 20060427 Write a test for this code path
1440
# it had a bug in it, and was raising the wrong
1442
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1443
tree.print_file(file_id)
1445
def get_transaction(self):
1446
return self.control_files.get_transaction()
1448
def revision_parents(self, revision_id):
1449
revision_id = osutils.safe_revision_id(revision_id)
1450
return self.get_inventory_weave().parent_names(revision_id)
1452
def get_parents(self, revision_ids):
1453
"""See StackedParentsProvider.get_parents"""
1455
for revision_id in revision_ids:
1456
if revision_id == _mod_revision.NULL_REVISION:
1460
parents = self.get_revision(revision_id).parent_ids
1461
except errors.NoSuchRevision:
1464
if len(parents) == 0:
1465
parents = [_mod_revision.NULL_REVISION]
1466
parents_list.append(parents)
1469
def _make_parents_provider(self):
1472
def get_graph(self, other_repository=None):
1473
"""Return the graph walker for this repository format"""
1474
parents_provider = self._make_parents_provider()
1475
if (other_repository is not None and
1476
other_repository.bzrdir.transport.base !=
1477
self.bzrdir.transport.base):
1478
parents_provider = graph._StackedParentsProvider(
1479
[parents_provider, other_repository._make_parents_provider()])
1480
return graph.Graph(parents_provider)
1483
def set_make_working_trees(self, new_value):
1484
"""Set the policy flag for making working trees when creating branches.
1486
This only applies to branches that use this repository.
1488
The default is 'True'.
1489
:param new_value: True to restore the default, False to disable making
1492
raise NotImplementedError(self.set_make_working_trees)
1494
def make_working_trees(self):
1495
"""Returns the policy for making working trees on new branches."""
1496
raise NotImplementedError(self.make_working_trees)
1499
def sign_revision(self, revision_id, gpg_strategy):
1500
revision_id = osutils.safe_revision_id(revision_id)
1501
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1502
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1505
def has_signature_for_revision_id(self, revision_id):
1506
"""Query for a revision signature for revision_id in the repository."""
1507
revision_id = osutils.safe_revision_id(revision_id)
1508
return self._revision_store.has_signature(revision_id,
1509
self.get_transaction())
1512
def get_signature_text(self, revision_id):
1513
"""Return the text for a signature."""
1514
revision_id = osutils.safe_revision_id(revision_id)
1515
return self._revision_store.get_signature_text(revision_id,
1516
self.get_transaction())
1519
def check(self, revision_ids):
1520
"""Check consistency of all history of given revision_ids.
1522
Different repository implementations should override _check().
1524
:param revision_ids: A non-empty list of revision_ids whose ancestry
1525
will be checked. Typically the last revision_id of a branch.
1527
if not revision_ids:
1528
raise ValueError("revision_ids must be non-empty in %s.check"
1530
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
1531
return self._check(revision_ids)
1533
def _check(self, revision_ids):
1534
result = check.Check(self)
1538
def _warn_if_deprecated(self):
1539
global _deprecation_warning_done
1540
if _deprecation_warning_done:
1542
_deprecation_warning_done = True
1543
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1544
% (self._format, self.bzrdir.transport.base))
1546
def supports_rich_root(self):
1547
return self._format.rich_root_data
1549
def _check_ascii_revisionid(self, revision_id, method):
1550
"""Private helper for ascii-only repositories."""
1551
# weave repositories refuse to store revisionids that are non-ascii.
1552
if revision_id is not None:
1553
# weaves require ascii revision ids.
1554
if isinstance(revision_id, unicode):
1556
revision_id.encode('ascii')
1557
except UnicodeEncodeError:
1558
raise errors.NonAsciiRevisionId(method, self)
1561
revision_id.decode('ascii')
1562
except UnicodeDecodeError:
1563
raise errors.NonAsciiRevisionId(method, self)
1567
# remove these delegates a while after bzr 0.15
1568
def __make_delegated(name, from_module):
1569
def _deprecated_repository_forwarder():
1570
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1571
% (name, from_module),
1574
m = __import__(from_module, globals(), locals(), [name])
1576
return getattr(m, name)
1577
except AttributeError:
1578
raise AttributeError('module %s has no name %s'
1580
globals()[name] = _deprecated_repository_forwarder
1583
'AllInOneRepository',
1584
'WeaveMetaDirRepository',
1585
'PreSplitOutRepositoryFormat',
1586
'RepositoryFormat4',
1587
'RepositoryFormat5',
1588
'RepositoryFormat6',
1589
'RepositoryFormat7',
1591
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1595
'RepositoryFormatKnit',
1596
'RepositoryFormatKnit1',
1598
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1601
def install_revision(repository, rev, revision_tree):
1602
"""Install all revision data into a repository."""
1603
repository.start_write_group()
1605
_install_revision(repository, rev, revision_tree)
1607
repository.abort_write_group()
1610
repository.commit_write_group()
1613
def _install_revision(repository, rev, revision_tree):
1614
"""Install all revision data into a repository."""
1615
present_parents = []
1617
for p_id in rev.parent_ids:
1618
if repository.has_revision(p_id):
1619
present_parents.append(p_id)
1620
parent_trees[p_id] = repository.revision_tree(p_id)
1622
parent_trees[p_id] = repository.revision_tree(None)
1624
inv = revision_tree.inventory
1625
entries = inv.iter_entries()
1626
# backwards compatibility hack: skip the root id.
1627
if not repository.supports_rich_root():
1628
path, root = entries.next()
1629
if root.revision != rev.revision_id:
1630
raise errors.IncompatibleRevision(repr(repository))
1631
# Add the texts that are not already present
1632
for path, ie in entries:
1633
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1634
repository.get_transaction())
1635
if ie.revision not in w:
1637
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1638
# with InventoryEntry.find_previous_heads(). if it is, then there
1639
# is a latent bug here where the parents may have ancestors of each
1641
for revision, tree in parent_trees.iteritems():
1642
if ie.file_id not in tree:
1644
parent_id = tree.inventory[ie.file_id].revision
1645
if parent_id in text_parents:
1647
text_parents.append(parent_id)
1649
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1650
repository.get_transaction())
1651
lines = revision_tree.get_file(ie.file_id).readlines()
1652
vfile.add_lines(rev.revision_id, text_parents, lines)
1654
# install the inventory
1655
repository.add_inventory(rev.revision_id, inv, present_parents)
1656
except errors.RevisionAlreadyPresent:
1658
repository.add_revision(rev.revision_id, rev, inv)
1661
class MetaDirRepository(Repository):
1662
"""Repositories in the new meta-dir layout."""
1664
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1665
super(MetaDirRepository, self).__init__(_format,
1671
dir_mode = self.control_files._dir_mode
1672
file_mode = self.control_files._file_mode
1675
def is_shared(self):
1676
"""Return True if this repository is flagged as a shared repository."""
1677
return self.control_files._transport.has('shared-storage')
1680
def set_make_working_trees(self, new_value):
1681
"""Set the policy flag for making working trees when creating branches.
1683
This only applies to branches that use this repository.
1685
The default is 'True'.
1686
:param new_value: True to restore the default, False to disable making
1691
self.control_files._transport.delete('no-working-trees')
1692
except errors.NoSuchFile:
1695
self.control_files.put_utf8('no-working-trees', '')
1697
def make_working_trees(self):
1698
"""Returns the policy for making working trees on new branches."""
1699
return not self.control_files._transport.has('no-working-trees')
1702
class RepositoryFormatRegistry(registry.Registry):
1703
"""Registry of RepositoryFormats."""
1705
def get(self, format_string):
1706
r = registry.Registry.get(self, format_string)
1712
format_registry = RepositoryFormatRegistry()
1713
"""Registry of formats, indexed by their identifying format string.
1715
This can contain either format instances themselves, or classes/factories that
1716
can be called to obtain one.
1720
#####################################################################
1721
# Repository Formats
1723
class RepositoryFormat(object):
1724
"""A repository format.
1726
Formats provide three things:
1727
* An initialization routine to construct repository data on disk.
1728
* a format string which is used when the BzrDir supports versioned
1730
* an open routine which returns a Repository instance.
1732
There is one and only one format for every disk format. The actual
1733
repository types do not indicate disk format at all - only repo._format can
1734
be used to determine the disk format of a Repository instance.
1736
Formats are placed in an dict by their format string for reference
1737
during opening. These should be subclasses of RepositoryFormat
1740
Once a format is deprecated, just deprecate the initialize and open
1741
methods on the format class. Do not deprecate the object, as the
1742
object will be created every system load.
1744
Common instance attributes:
1745
_matchingbzrdir - the bzrdir format that the repository format was
1746
originally written to work with. This can be used if manually
1747
constructing a bzrdir and repository, or more commonly for test suite
1752
return "<%s>" % self.__class__.__name__
1754
def __eq__(self, other):
1755
# format objects are generally stateless
1756
return isinstance(other, self.__class__)
1758
def __ne__(self, other):
1759
return not self == other
1762
def find_format(klass, a_bzrdir):
1763
"""Return the format for the repository object in a_bzrdir.
1765
This is used by bzr native formats that have a "format" file in
1766
the repository. Other methods may be used by different types of
1770
transport = a_bzrdir.get_repository_transport(None)
1771
format_string = transport.get("format").read()
1772
return format_registry.get(format_string)
1773
except errors.NoSuchFile:
1774
raise errors.NoRepositoryPresent(a_bzrdir)
1776
raise errors.UnknownFormatError(format=format_string)
1779
def register_format(klass, format):
1780
format_registry.register(format.get_format_string(), format)
1783
def unregister_format(klass, format):
1784
format_registry.remove(format.get_format_string())
1787
def get_default_format(klass):
1788
"""Return the current default format."""
1789
from bzrlib import bzrdir
1790
return bzrdir.format_registry.make_bzrdir('default').repository_format
1792
def _get_control_store(self, repo_transport, control_files):
1793
"""Return the control store for this repository."""
1794
raise NotImplementedError(self._get_control_store)
1796
def get_format_string(self):
1797
"""Return the ASCII format string that identifies this format.
1799
Note that in pre format ?? repositories the format string is
1800
not permitted nor written to disk.
1802
raise NotImplementedError(self.get_format_string)
1804
def get_format_description(self):
1805
"""Return the short description for this format."""
1806
raise NotImplementedError(self.get_format_description)
1808
def _get_revision_store(self, repo_transport, control_files):
1809
"""Return the revision store object for this a_bzrdir."""
1810
raise NotImplementedError(self._get_revision_store)
1812
def _get_text_rev_store(self,
1819
"""Common logic for getting a revision store for a repository.
1821
see self._get_revision_store for the subclass-overridable method to
1822
get the store for a repository.
1824
from bzrlib.store.revision.text import TextRevisionStore
1825
dir_mode = control_files._dir_mode
1826
file_mode = control_files._file_mode
1827
text_store = TextStore(transport.clone(name),
1829
compressed=compressed,
1831
file_mode=file_mode)
1832
_revision_store = TextRevisionStore(text_store, serializer)
1833
return _revision_store
1835
# TODO: this shouldn't be in the base class, it's specific to things that
1836
# use weaves or knits -- mbp 20070207
1837
def _get_versioned_file_store(self,
1842
versionedfile_class=None,
1843
versionedfile_kwargs={},
1845
if versionedfile_class is None:
1846
versionedfile_class = self._versionedfile_class
1847
weave_transport = control_files._transport.clone(name)
1848
dir_mode = control_files._dir_mode
1849
file_mode = control_files._file_mode
1850
return VersionedFileStore(weave_transport, prefixed=prefixed,
1852
file_mode=file_mode,
1853
versionedfile_class=versionedfile_class,
1854
versionedfile_kwargs=versionedfile_kwargs,
1857
def initialize(self, a_bzrdir, shared=False):
1858
"""Initialize a repository of this format in a_bzrdir.
1860
:param a_bzrdir: The bzrdir to put the new repository in it.
1861
:param shared: The repository should be initialized as a sharable one.
1862
:returns: The new repository object.
1864
This may raise UninitializableFormat if shared repository are not
1865
compatible the a_bzrdir.
1867
raise NotImplementedError(self.initialize)
1869
def is_supported(self):
1870
"""Is this format supported?
1872
Supported formats must be initializable and openable.
1873
Unsupported formats may not support initialization or committing or
1874
some other features depending on the reason for not being supported.
1878
def check_conversion_target(self, target_format):
1879
raise NotImplementedError(self.check_conversion_target)
1881
def open(self, a_bzrdir, _found=False):
1882
"""Return an instance of this format for the bzrdir a_bzrdir.
1884
_found is a private parameter, do not use it.
1886
raise NotImplementedError(self.open)
1889
class MetaDirRepositoryFormat(RepositoryFormat):
1890
"""Common base class for the new repositories using the metadir layout."""
1892
rich_root_data = False
1893
supports_tree_reference = False
1894
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1897
super(MetaDirRepositoryFormat, self).__init__()
1899
def _create_control_files(self, a_bzrdir):
1900
"""Create the required files and the initial control_files object."""
1901
# FIXME: RBC 20060125 don't peek under the covers
1902
# NB: no need to escape relative paths that are url safe.
1903
repository_transport = a_bzrdir.get_repository_transport(self)
1904
control_files = lockable_files.LockableFiles(repository_transport,
1905
'lock', lockdir.LockDir)
1906
control_files.create_lock()
1907
return control_files
1909
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1910
"""Upload the initial blank content."""
1911
control_files = self._create_control_files(a_bzrdir)
1912
control_files.lock_write()
1914
control_files._transport.mkdir_multi(dirs,
1915
mode=control_files._dir_mode)
1916
for file, content in files:
1917
control_files.put(file, content)
1918
for file, content in utf8_files:
1919
control_files.put_utf8(file, content)
1921
control_files.put_utf8('shared-storage', '')
1923
control_files.unlock()
1926
# formats which have no format string are not discoverable
1927
# and not independently creatable, so are not registered. They're
1928
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1929
# needed, it's constructed directly by the BzrDir. Non-native formats where
1930
# the repository is not separately opened are similar.
1932
format_registry.register_lazy(
1933
'Bazaar-NG Repository format 7',
1934
'bzrlib.repofmt.weaverepo',
1938
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1939
# default control directory format
1940
format_registry.register_lazy(
1941
'Bazaar-NG Knit Repository Format 1',
1942
'bzrlib.repofmt.knitrepo',
1943
'RepositoryFormatKnit1',
1945
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1947
format_registry.register_lazy(
1948
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1949
'bzrlib.repofmt.knitrepo',
1950
'RepositoryFormatKnit3',
1953
# Experimental formats. These make no guarantee about data stability.
1954
# There is one format for pre-subtrees, and one for post-subtrees to
1955
# allow ease of testing.
1956
format_registry.register_lazy(
1957
'Bazaar Experimental no-subtrees\n',
1958
'bzrlib.repofmt.pack_repo',
1959
'RepositoryFormatGraphKnit1',
1961
format_registry.register_lazy(
1962
'Bazaar Experimental subtrees\n',
1963
'bzrlib.repofmt.pack_repo',
1964
'RepositoryFormatGraphKnit3',
1968
class InterRepository(InterObject):
1969
"""This class represents operations taking place between two repositories.
1971
Its instances have methods like copy_content and fetch, and contain
1972
references to the source and target repositories these operations can be
1975
Often we will provide convenience methods on 'repository' which carry out
1976
operations with another repository - they will always forward to
1977
InterRepository.get(other).method_name(parameters).
1981
"""The available optimised InterRepository types."""
1983
def copy_content(self, revision_id=None):
1984
raise NotImplementedError(self.copy_content)
1986
def fetch(self, revision_id=None, pb=None):
1987
"""Fetch the content required to construct revision_id.
1989
The content is copied from self.source to self.target.
1991
:param revision_id: if None all content is copied, if NULL_REVISION no
1993
:param pb: optional progress bar to use for progress reports. If not
1994
provided a default one will be created.
1996
Returns the copied revision count and the failed revisions in a tuple:
1999
raise NotImplementedError(self.fetch)
2002
def missing_revision_ids(self, revision_id=None):
2003
"""Return the revision ids that source has that target does not.
2005
These are returned in topological order.
2007
:param revision_id: only return revision ids included by this
2010
# generic, possibly worst case, slow code path.
2011
target_ids = set(self.target.all_revision_ids())
2012
if revision_id is not None:
2013
# TODO: jam 20070210 InterRepository is internal enough that it
2014
# should assume revision_ids are already utf-8
2015
revision_id = osutils.safe_revision_id(revision_id)
2016
source_ids = self.source.get_ancestry(revision_id)
2017
assert source_ids[0] is None
2020
source_ids = self.source.all_revision_ids()
2021
result_set = set(source_ids).difference(target_ids)
2022
# this may look like a no-op: its not. It preserves the ordering
2023
# other_ids had while only returning the members from other_ids
2024
# that we've decided we need.
2025
return [rev_id for rev_id in source_ids if rev_id in result_set]
2028
def _same_model(source, target):
2029
"""True if source and target have the same data representation."""
2030
if source.supports_rich_root() != target.supports_rich_root():
2032
if source._serializer != target._serializer:
2037
class InterSameDataRepository(InterRepository):
2038
"""Code for converting between repositories that represent the same data.
2040
Data format and model must match for this to work.
2044
def _get_repo_format_to_test(self):
2045
"""Repository format for testing with.
2047
InterSameData can pull from subtree to subtree and from non-subtree to
2048
non-subtree, so we test this with the richest repository format.
2050
from bzrlib.repofmt import knitrepo
2051
return knitrepo.RepositoryFormatKnit3()
2054
def is_compatible(source, target):
2055
return InterRepository._same_model(source, target)
2058
def copy_content(self, revision_id=None):
2059
"""Make a complete copy of the content in self into destination.
2061
This copies both the repository's revision data, and configuration information
2062
such as the make_working_trees setting.
2064
This is a destructive operation! Do not use it on existing
2067
:param revision_id: Only copy the content needed to construct
2068
revision_id and its parents.
2071
self.target.set_make_working_trees(self.source.make_working_trees())
2072
except NotImplementedError:
2074
# TODO: jam 20070210 This is fairly internal, so we should probably
2075
# just assert that revision_id is not unicode.
2076
revision_id = osutils.safe_revision_id(revision_id)
2077
# but don't bother fetching if we have the needed data now.
2078
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2079
self.target.has_revision(revision_id)):
2081
self.target.fetch(self.source, revision_id=revision_id)
2084
def fetch(self, revision_id=None, pb=None):
2085
"""See InterRepository.fetch()."""
2086
from bzrlib.fetch import GenericRepoFetcher
2087
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2088
self.source, self.source._format, self.target,
2089
self.target._format)
2090
# TODO: jam 20070210 This should be an assert, not a translate
2091
revision_id = osutils.safe_revision_id(revision_id)
2092
f = GenericRepoFetcher(to_repository=self.target,
2093
from_repository=self.source,
2094
last_revision=revision_id,
2096
return f.count_copied, f.failed_revisions
2099
class InterWeaveRepo(InterSameDataRepository):
2100
"""Optimised code paths between Weave based repositories.
2102
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2103
implemented lazy inter-object optimisation.
2107
def _get_repo_format_to_test(self):
2108
from bzrlib.repofmt import weaverepo
2109
return weaverepo.RepositoryFormat7()
2112
def is_compatible(source, target):
2113
"""Be compatible with known Weave formats.
2115
We don't test for the stores being of specific types because that
2116
could lead to confusing results, and there is no need to be
2119
from bzrlib.repofmt.weaverepo import (
2125
return (isinstance(source._format, (RepositoryFormat5,
2127
RepositoryFormat7)) and
2128
isinstance(target._format, (RepositoryFormat5,
2130
RepositoryFormat7)))
2131
except AttributeError:
2135
def copy_content(self, revision_id=None):
2136
"""See InterRepository.copy_content()."""
2137
# weave specific optimised path:
2138
# TODO: jam 20070210 Internal, should be an assert, not translate
2139
revision_id = osutils.safe_revision_id(revision_id)
2141
self.target.set_make_working_trees(self.source.make_working_trees())
2142
except NotImplementedError:
2144
# FIXME do not peek!
2145
if self.source.control_files._transport.listable():
2146
pb = ui.ui_factory.nested_progress_bar()
2148
self.target.weave_store.copy_all_ids(
2149
self.source.weave_store,
2151
from_transaction=self.source.get_transaction(),
2152
to_transaction=self.target.get_transaction())
2153
pb.update('copying inventory', 0, 1)
2154
self.target.control_weaves.copy_multi(
2155
self.source.control_weaves, ['inventory'],
2156
from_transaction=self.source.get_transaction(),
2157
to_transaction=self.target.get_transaction())
2158
self.target._revision_store.text_store.copy_all_ids(
2159
self.source._revision_store.text_store,
2164
self.target.fetch(self.source, revision_id=revision_id)
2167
def fetch(self, revision_id=None, pb=None):
2168
"""See InterRepository.fetch()."""
2169
from bzrlib.fetch import GenericRepoFetcher
2170
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2171
self.source, self.source._format, self.target, self.target._format)
2172
# TODO: jam 20070210 This should be an assert, not a translate
2173
revision_id = osutils.safe_revision_id(revision_id)
2174
f = GenericRepoFetcher(to_repository=self.target,
2175
from_repository=self.source,
2176
last_revision=revision_id,
2178
return f.count_copied, f.failed_revisions
2181
def missing_revision_ids(self, revision_id=None):
2182
"""See InterRepository.missing_revision_ids()."""
2183
# we want all revisions to satisfy revision_id in source.
2184
# but we don't want to stat every file here and there.
2185
# we want then, all revisions other needs to satisfy revision_id
2186
# checked, but not those that we have locally.
2187
# so the first thing is to get a subset of the revisions to
2188
# satisfy revision_id in source, and then eliminate those that
2189
# we do already have.
2190
# this is slow on high latency connection to self, but as as this
2191
# disk format scales terribly for push anyway due to rewriting
2192
# inventory.weave, this is considered acceptable.
2194
if revision_id is not None:
2195
source_ids = self.source.get_ancestry(revision_id)
2196
assert source_ids[0] is None
2199
source_ids = self.source._all_possible_ids()
2200
source_ids_set = set(source_ids)
2201
# source_ids is the worst possible case we may need to pull.
2202
# now we want to filter source_ids against what we actually
2203
# have in target, but don't try to check for existence where we know
2204
# we do not have a revision as that would be pointless.
2205
target_ids = set(self.target._all_possible_ids())
2206
possibly_present_revisions = target_ids.intersection(source_ids_set)
2207
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2208
required_revisions = source_ids_set.difference(actually_present_revisions)
2209
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2210
if revision_id is not None:
2211
# we used get_ancestry to determine source_ids then we are assured all
2212
# revisions referenced are present as they are installed in topological order.
2213
# and the tip revision was validated by get_ancestry.
2214
return required_topo_revisions
2216
# if we just grabbed the possibly available ids, then
2217
# we only have an estimate of whats available and need to validate
2218
# that against the revision records.
2219
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2222
class InterKnitRepo(InterSameDataRepository):
2223
"""Optimised code paths between Knit based repositories."""
2226
def _get_repo_format_to_test(self):
2227
from bzrlib.repofmt import knitrepo
2228
return knitrepo.RepositoryFormatKnit1()
2231
def is_compatible(source, target):
2232
"""Be compatible with known Knit formats.
2234
We don't test for the stores being of specific types because that
2235
could lead to confusing results, and there is no need to be
2238
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2240
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2241
isinstance(target._format, RepositoryFormatKnit))
2242
except AttributeError:
2244
return are_knits and InterRepository._same_model(source, target)
2247
def fetch(self, revision_id=None, pb=None):
2248
"""See InterRepository.fetch()."""
2249
from bzrlib.fetch import KnitRepoFetcher
2250
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2251
self.source, self.source._format, self.target, self.target._format)
2252
# TODO: jam 20070210 This should be an assert, not a translate
2253
revision_id = osutils.safe_revision_id(revision_id)
2254
f = KnitRepoFetcher(to_repository=self.target,
2255
from_repository=self.source,
2256
last_revision=revision_id,
2258
return f.count_copied, f.failed_revisions
2261
def missing_revision_ids(self, revision_id=None):
2262
"""See InterRepository.missing_revision_ids()."""
2263
if revision_id is not None:
2264
source_ids = self.source.get_ancestry(revision_id)
2265
assert source_ids[0] is None
2268
source_ids = self.source.all_revision_ids()
2269
source_ids_set = set(source_ids)
2270
# source_ids is the worst possible case we may need to pull.
2271
# now we want to filter source_ids against what we actually
2272
# have in target, but don't try to check for existence where we know
2273
# we do not have a revision as that would be pointless.
2274
target_ids = set(self.target.all_revision_ids())
2275
possibly_present_revisions = target_ids.intersection(source_ids_set)
2276
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2277
required_revisions = source_ids_set.difference(actually_present_revisions)
2278
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2279
if revision_id is not None:
2280
# we used get_ancestry to determine source_ids then we are assured all
2281
# revisions referenced are present as they are installed in topological order.
2282
# and the tip revision was validated by get_ancestry.
2283
return required_topo_revisions
2285
# if we just grabbed the possibly available ids, then
2286
# we only have an estimate of whats available and need to validate
2287
# that against the revision records.
2288
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2291
class InterPackRepo(InterSameDataRepository):
2292
"""Optimised code paths between Pack based repositories."""
2295
def _get_repo_format_to_test(self):
2296
from bzrlib.repofmt import pack_repo
2297
return pack_repo.RepositoryFormatGraphKnit1()
2300
def is_compatible(source, target):
2301
"""Be compatible with known Pack formats.
2303
We don't test for the stores being of specific types because that
2304
could lead to confusing results, and there is no need to be
2307
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2309
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2310
isinstance(target._format, RepositoryFormatPack))
2311
except AttributeError:
2313
return are_packs and InterRepository._same_model(source, target)
2316
def fetch(self, revision_id=None, pb=None):
2317
"""See InterRepository.fetch()."""
2318
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2319
self.source, self.source._format, self.target, self.target._format)
2320
# TODO: jam 20070210 This should be an assert, not a translate
2321
revision_id = osutils.safe_revision_id(revision_id)
2322
self.count_copied = 0
2323
if revision_id is None:
2325
# everything to do - use pack logic
2326
# to fetch from all packs to one without
2327
# inventory parsing etc, IFF nothing to be copied is in the target.
2329
revision_ids = self.source.all_revision_ids()
2330
# implementing the TODO will involve:
2331
# - detecting when all of a pack is selected
2332
# - avoiding as much as possible pre-selection, so the
2333
# more-core routines such as create_pack_from_packs can filter in
2334
# a just-in-time fashion. (though having a HEADS list on a
2335
# repository might make this a lot easier, because we could
2336
# sensibly detect 'new revisions' without doing a full index scan.
2337
elif _mod_revision.is_null(revision_id):
2342
revision_ids = self.missing_revision_ids(revision_id)
2343
except errors.NoSuchRevision:
2344
raise errors.InstallFailed([revision_id])
2345
packs = self.source._packs.all_pack_details()
2346
revision_index_map = self.source._packs._revision_index_map(packs)
2347
inventory_index_map = self.source._packs._inv_index_map(packs)
2348
text_index_map = self.source._packs._text_index_map(packs)
2349
signature_index_map = self.source._packs._signature_index_map(packs)
2350
pack = self.target._packs.create_pack_from_packs(
2352
inventory_index_map,
2354
signature_index_map,
2358
if pack is not None:
2359
self.target._packs._save_pack_names()
2360
self.target._packs.add_pack_to_memory(pack)
2361
# Trigger an autopack. This may duplicate effort as we've just done
2362
# a pack creation, but for now it is simpler to think about as
2363
# 'upload data, then repack if needed'.
2364
self.target._packs.autopack()
2365
return pack.get_revision_count()
2370
def missing_revision_ids(self, revision_id=None):
2371
"""See InterRepository.missing_revision_ids()."""
2372
if revision_id is not None:
2373
source_ids = self.source.get_ancestry(revision_id)
2374
assert source_ids[0] is None
2377
source_ids = self.source.all_revision_ids()
2378
source_ids_set = set(source_ids)
2379
# source_ids is the worst possible case we may need to pull.
2380
# now we want to filter source_ids against what we actually
2381
# have in target, but don't try to check for existence where we know
2382
# we do not have a revision as that would be pointless.
2383
target_ids = set(self.target.all_revision_ids())
2384
actually_present_revisions = target_ids.intersection(source_ids_set)
2385
required_revisions = source_ids_set.difference(actually_present_revisions)
2386
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2387
return required_topo_revisions
2390
class InterModel1and2(InterRepository):
2393
def _get_repo_format_to_test(self):
2397
def is_compatible(source, target):
2398
if not source.supports_rich_root() and target.supports_rich_root():
2404
def fetch(self, revision_id=None, pb=None):
2405
"""See InterRepository.fetch()."""
2406
from bzrlib.fetch import Model1toKnit2Fetcher
2407
# TODO: jam 20070210 This should be an assert, not a translate
2408
revision_id = osutils.safe_revision_id(revision_id)
2409
f = Model1toKnit2Fetcher(to_repository=self.target,
2410
from_repository=self.source,
2411
last_revision=revision_id,
2413
return f.count_copied, f.failed_revisions
2416
def copy_content(self, revision_id=None):
2417
"""Make a complete copy of the content in self into destination.
2419
This is a destructive operation! Do not use it on existing
2422
:param revision_id: Only copy the content needed to construct
2423
revision_id and its parents.
2426
self.target.set_make_working_trees(self.source.make_working_trees())
2427
except NotImplementedError:
2429
# TODO: jam 20070210 Internal, assert, don't translate
2430
revision_id = osutils.safe_revision_id(revision_id)
2431
# but don't bother fetching if we have the needed data now.
2432
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2433
self.target.has_revision(revision_id)):
2435
self.target.fetch(self.source, revision_id=revision_id)
2438
class InterKnit1and2(InterKnitRepo):
2441
def _get_repo_format_to_test(self):
2445
def is_compatible(source, target):
2446
"""Be compatible with Knit1 source and Knit3 target"""
2447
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2449
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2450
RepositoryFormatKnit3)
2451
from bzrlib.repofmt.pack_repo import (RepositoryFormatGraphKnit1,
2452
RepositoryFormatGraphKnit3)
2453
return (isinstance(source._format, RepositoryFormatKnit1) and
2454
isinstance(target._format, RepositoryFormatKnit3) or
2455
isinstance(source._format, (RepositoryFormatGraphKnit1)) and
2456
isinstance(target._format, (RepositoryFormatGraphKnit3))
2458
except AttributeError:
2462
def fetch(self, revision_id=None, pb=None):
2463
"""See InterRepository.fetch()."""
2464
from bzrlib.fetch import Knit1to2Fetcher
2465
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2466
self.source, self.source._format, self.target,
2467
self.target._format)
2468
# TODO: jam 20070210 This should be an assert, not a translate
2469
revision_id = osutils.safe_revision_id(revision_id)
2470
f = Knit1to2Fetcher(to_repository=self.target,
2471
from_repository=self.source,
2472
last_revision=revision_id,
2474
return f.count_copied, f.failed_revisions
2477
class InterRemoteRepository(InterRepository):
2478
"""Code for converting between RemoteRepository objects.
2480
This just gets an non-remote repository from the RemoteRepository, and calls
2481
InterRepository.get again.
2484
def __init__(self, source, target):
2485
if isinstance(source, remote.RemoteRepository):
2486
source._ensure_real()
2487
real_source = source._real_repository
2489
real_source = source
2490
if isinstance(target, remote.RemoteRepository):
2491
target._ensure_real()
2492
real_target = target._real_repository
2494
real_target = target
2495
self.real_inter = InterRepository.get(real_source, real_target)
2498
def is_compatible(source, target):
2499
if isinstance(source, remote.RemoteRepository):
2501
if isinstance(target, remote.RemoteRepository):
2505
def copy_content(self, revision_id=None):
2506
self.real_inter.copy_content(revision_id=revision_id)
2508
def fetch(self, revision_id=None, pb=None):
2509
self.real_inter.fetch(revision_id=revision_id, pb=pb)
2512
def _get_repo_format_to_test(self):
2516
InterRepository.register_optimiser(InterSameDataRepository)
2517
InterRepository.register_optimiser(InterWeaveRepo)
2518
InterRepository.register_optimiser(InterKnitRepo)
2519
InterRepository.register_optimiser(InterModel1and2)
2520
InterRepository.register_optimiser(InterKnit1and2)
2521
InterRepository.register_optimiser(InterPackRepo)
2522
InterRepository.register_optimiser(InterRemoteRepository)
2525
class CopyConverter(object):
2526
"""A repository conversion tool which just performs a copy of the content.
2528
This is slow but quite reliable.
2531
def __init__(self, target_format):
2532
"""Create a CopyConverter.
2534
:param target_format: The format the resulting repository should be.
2536
self.target_format = target_format
2538
def convert(self, repo, pb):
2539
"""Perform the conversion of to_convert, giving feedback via pb.
2541
:param to_convert: The disk object to convert.
2542
:param pb: a progress bar to use for progress information.
2547
# this is only useful with metadir layouts - separated repo content.
2548
# trigger an assertion if not such
2549
repo._format.get_format_string()
2550
self.repo_dir = repo.bzrdir
2551
self.step('Moving repository to repository.backup')
2552
self.repo_dir.transport.move('repository', 'repository.backup')
2553
backup_transport = self.repo_dir.transport.clone('repository.backup')
2554
repo._format.check_conversion_target(self.target_format)
2555
self.source_repo = repo._format.open(self.repo_dir,
2557
_override_transport=backup_transport)
2558
self.step('Creating new repository')
2559
converted = self.target_format.initialize(self.repo_dir,
2560
self.source_repo.is_shared())
2561
converted.lock_write()
2563
self.step('Copying content into repository.')
2564
self.source_repo.copy_content_into(converted)
2567
self.step('Deleting old repository content.')
2568
self.repo_dir.transport.delete_tree('repository.backup')
2569
self.pb.note('repository converted')
2571
def step(self, message):
2572
"""Update the pb by a step."""
2574
self.pb.update(message, self.count, self.total)
2586
def _unescaper(match, _map=_unescape_map):
2587
code = match.group(1)
2591
if not code.startswith('#'):
2593
return unichr(int(code[1:])).encode('utf8')
2599
def _unescape_xml(data):
2600
"""Unescape predefined XML entities in a string of data."""
2602
if _unescape_re is None:
2603
_unescape_re = re.compile('\&([^;]*);')
2604
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
