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(), """
40
revision as _mod_revision,
46
from bzrlib.bundle import serializer
47
from bzrlib.revisiontree import RevisionTree
48
from bzrlib.store.versioned import VersionedFileStore
49
from bzrlib.store.text import TextStore
50
from bzrlib.testament import Testament
51
from bzrlib.util import bencode
54
from bzrlib.decorators import needs_read_lock, needs_write_lock
55
from bzrlib.inter import InterObject
56
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
57
from bzrlib.symbol_versioning import (
60
from bzrlib.trace import mutter, mutter_callsite, note, warning
63
# Old formats display a warning, but only once
64
_deprecation_warning_done = False
67
class CommitBuilder(object):
68
"""Provides an interface to build up a commit.
70
This allows describing a tree to be committed without needing to
71
know the internals of the format of the repository.
74
# all clients should supply tree roots.
75
record_root_entry = True
76
# the default CommitBuilder does not manage trees whose root is versioned.
77
_versioned_root = False
79
def __init__(self, repository, parents, config, timestamp=None,
80
timezone=None, committer=None, revprops=None,
82
"""Initiate a CommitBuilder.
84
:param repository: Repository to commit to.
85
:param parents: Revision ids of the parents of the new revision.
86
:param config: Configuration to use.
87
:param timestamp: Optional timestamp recorded for commit.
88
:param timezone: Optional timezone for timestamp.
89
:param committer: Optional committer to set for commit.
90
:param revprops: Optional dictionary of revision properties.
91
:param revision_id: Optional revision id.
96
self._committer = self._config.username()
98
assert isinstance(committer, basestring), type(committer)
99
self._committer = committer
101
self.new_inventory = Inventory(None)
102
self._new_revision_id = revision_id
103
self.parents = parents
104
self.repository = repository
107
if revprops is not None:
108
self._revprops.update(revprops)
110
if timestamp is None:
111
timestamp = time.time()
112
# Restrict resolution to 1ms
113
self._timestamp = round(timestamp, 3)
116
self._timezone = osutils.local_time_offset()
118
self._timezone = int(timezone)
120
self._generate_revision_if_needed()
121
self.__heads = graph.HeadsCache(repository.get_graph()).heads
123
def commit(self, message):
124
"""Make the actual commit.
126
:return: The revision id of the recorded revision.
128
rev = _mod_revision.Revision(
129
timestamp=self._timestamp,
130
timezone=self._timezone,
131
committer=self._committer,
133
inventory_sha1=self.inv_sha1,
134
revision_id=self._new_revision_id,
135
properties=self._revprops)
136
rev.parent_ids = self.parents
137
self.repository.add_revision(self._new_revision_id, rev,
138
self.new_inventory, self._config)
139
self.repository.commit_write_group()
140
return self._new_revision_id
143
"""Abort the commit that is being built.
145
self.repository.abort_write_group()
147
def revision_tree(self):
148
"""Return the tree that was just committed.
150
After calling commit() this can be called to get a RevisionTree
151
representing the newly committed tree. This is preferred to
152
calling Repository.revision_tree() because that may require
153
deserializing the inventory, while we already have a copy in
156
return RevisionTree(self.repository, self.new_inventory,
157
self._new_revision_id)
159
def finish_inventory(self):
160
"""Tell the builder that the inventory is finished."""
161
if self.new_inventory.root is None:
162
raise AssertionError('Root entry should be supplied to'
163
' record_entry_contents, as of bzr 0.10.',
164
DeprecationWarning, stacklevel=2)
165
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
166
self.new_inventory.revision_id = self._new_revision_id
167
self.inv_sha1 = self.repository.add_inventory(
168
self._new_revision_id,
173
def _gen_revision_id(self):
174
"""Return new revision-id."""
175
return generate_ids.gen_revision_id(self._config.username(),
178
def _generate_revision_if_needed(self):
179
"""Create a revision id if None was supplied.
181
If the repository can not support user-specified revision ids
182
they should override this function and raise CannotSetRevisionId
183
if _new_revision_id is not None.
185
:raises: CannotSetRevisionId
187
if self._new_revision_id is None:
188
self._new_revision_id = self._gen_revision_id()
189
self.random_revid = True
191
self.random_revid = False
193
def _heads(self, file_id, revision_ids):
194
"""Calculate the graph heads for revision_ids in the graph of file_id.
196
This can use either a per-file graph or a global revision graph as we
197
have an identity relationship between the two graphs.
199
return self.__heads(revision_ids)
201
def _check_root(self, ie, parent_invs, tree):
202
"""Helper for record_entry_contents.
204
:param ie: An entry being added.
205
:param parent_invs: The inventories of the parent revisions of the
207
:param tree: The tree that is being committed.
209
# In this revision format, root entries have no knit or weave When
210
# serializing out to disk and back in root.revision is always
212
ie.revision = self._new_revision_id
214
def _get_delta(self, ie, basis_inv, path):
215
"""Get a delta against the basis inventory for ie."""
216
if ie.file_id not in basis_inv:
218
return (None, path, ie.file_id, ie)
219
elif ie != basis_inv[ie.file_id]:
221
# TODO: avoid tis id2path call.
222
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
227
def record_entry_contents(self, ie, parent_invs, path, tree,
229
"""Record the content of ie from tree into the commit if needed.
231
Side effect: sets ie.revision when unchanged
233
:param ie: An inventory entry present in the commit.
234
:param parent_invs: The inventories of the parent revisions of the
236
:param path: The path the entry is at in the tree.
237
:param tree: The tree which contains this entry and should be used to
239
:param content_summary: Summary data from the tree about the paths
240
content - stat, length, exec, sha/link target. This is only
241
accessed when the entry has a revision of None - that is when it is
242
a candidate to commit.
243
:return: A tuple (change_delta, version_recorded). change_delta is
244
an inventory_delta change for this entry against the basis tree of
245
the commit, or None if no change occured against the basis tree.
246
version_recorded is True if a new version of the entry has been
247
recorded. For instance, committing a merge where a file was only
248
changed on the other side will return (delta, False).
250
if self.new_inventory.root is None:
251
if ie.parent_id is not None:
252
raise errors.RootMissing()
253
self._check_root(ie, parent_invs, tree)
254
if ie.revision is None:
255
kind = content_summary[0]
257
# ie is carried over from a prior commit
259
# XXX: repository specific check for nested tree support goes here - if
260
# the repo doesn't want nested trees we skip it ?
261
if (kind == 'tree-reference' and
262
not self.repository._format.supports_tree_reference):
263
# mismatch between commit builder logic and repository:
264
# this needs the entry creation pushed down into the builder.
265
raise NotImplementedError('Missing repository subtree support.')
266
self.new_inventory.add(ie)
268
# TODO: slow, take it out of the inner loop.
270
basis_inv = parent_invs[0]
272
basis_inv = Inventory(root_id=None)
274
# ie.revision is always None if the InventoryEntry is considered
275
# for committing. We may record the previous parents revision if the
276
# content is actually unchanged against a sole head.
277
if ie.revision is not None:
278
if not self._versioned_root and path == '':
279
# repositories that do not version the root set the root's
280
# revision to the new commit even when no change occurs, and
281
# this masks when a change may have occurred against the basis,
282
# so calculate if one happened.
283
if ie.file_id in basis_inv:
284
delta = (basis_inv.id2path(ie.file_id), path,
288
delta = (None, path, ie.file_id, ie)
291
# we don't need to commit this, because the caller already
292
# determined that an existing revision of this file is
294
return None, (ie.revision == self._new_revision_id)
295
# XXX: Friction: parent_candidates should return a list not a dict
296
# so that we don't have to walk the inventories again.
297
parent_candiate_entries = ie.parent_candidates(parent_invs)
298
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
300
for inv in parent_invs:
301
if ie.file_id in inv:
302
old_rev = inv[ie.file_id].revision
303
if old_rev in head_set:
304
heads.append(inv[ie.file_id].revision)
305
head_set.remove(inv[ie.file_id].revision)
308
# now we check to see if we need to write a new record to the
310
# We write a new entry unless there is one head to the ancestors, and
311
# the kind-derived content is unchanged.
313
# Cheapest check first: no ancestors, or more the one head in the
314
# ancestors, we write a new node.
318
# There is a single head, look it up for comparison
319
parent_entry = parent_candiate_entries[heads[0]]
320
# if the non-content specific data has changed, we'll be writing a
322
if (parent_entry.parent_id != ie.parent_id or
323
parent_entry.name != ie.name):
325
# now we need to do content specific checks:
327
# if the kind changed the content obviously has
328
if kind != parent_entry.kind:
331
assert content_summary[2] is not None, \
332
"Files must not have executable = None"
334
if (# if the file length changed we have to store:
335
parent_entry.text_size != content_summary[1] or
336
# if the exec bit has changed we have to store:
337
parent_entry.executable != content_summary[2]):
339
elif parent_entry.text_sha1 == content_summary[3]:
340
# all meta and content is unchanged (using a hash cache
341
# hit to check the sha)
342
ie.revision = parent_entry.revision
343
ie.text_size = parent_entry.text_size
344
ie.text_sha1 = parent_entry.text_sha1
345
ie.executable = parent_entry.executable
346
return self._get_delta(ie, basis_inv, path), False
348
# Either there is only a hash change(no hash cache entry,
349
# or same size content change), or there is no change on
351
# Provide the parent's hash to the store layer, so that the
352
# content is unchanged we will not store a new node.
353
nostore_sha = parent_entry.text_sha1
355
# We want to record a new node regardless of the presence or
356
# absence of a content change in the file.
358
ie.executable = content_summary[2]
359
lines = tree.get_file(ie.file_id, path).readlines()
361
ie.text_sha1, ie.text_size = self._add_text_to_weave(
362
ie.file_id, lines, heads, nostore_sha)
363
except errors.ExistingContent:
364
# Turns out that the file content was unchanged, and we were
365
# only going to store a new node if it was changed. Carry over
367
ie.revision = parent_entry.revision
368
ie.text_size = parent_entry.text_size
369
ie.text_sha1 = parent_entry.text_sha1
370
ie.executable = parent_entry.executable
371
return self._get_delta(ie, basis_inv, path), False
372
elif kind == 'directory':
374
# all data is meta here, nothing specific to directory, so
376
ie.revision = parent_entry.revision
377
return self._get_delta(ie, basis_inv, path), False
379
self._add_text_to_weave(ie.file_id, lines, heads, None)
380
elif kind == 'symlink':
381
current_link_target = content_summary[3]
383
# symlink target is not generic metadata, check if it has
385
if current_link_target != parent_entry.symlink_target:
388
# unchanged, carry over.
389
ie.revision = parent_entry.revision
390
ie.symlink_target = parent_entry.symlink_target
391
return self._get_delta(ie, basis_inv, path), False
392
ie.symlink_target = current_link_target
394
self._add_text_to_weave(ie.file_id, lines, heads, None)
395
elif kind == 'tree-reference':
397
if content_summary[3] != parent_entry.reference_revision:
400
# unchanged, carry over.
401
ie.reference_revision = parent_entry.reference_revision
402
ie.revision = parent_entry.revision
403
return self._get_delta(ie, basis_inv, path), False
404
ie.reference_revision = content_summary[3]
406
self._add_text_to_weave(ie.file_id, lines, heads, None)
408
raise NotImplementedError('unknown kind')
409
ie.revision = self._new_revision_id
410
return self._get_delta(ie, basis_inv, path), True
412
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
413
versionedfile = self.repository.weave_store.get_weave_or_empty(
414
file_id, self.repository.get_transaction())
415
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
416
# than add_lines, and allows committing when a parent is ghosted for
418
# Note: as we read the content directly from the tree, we know its not
419
# been turned into unicode or badly split - but a broken tree
420
# implementation could give us bad output from readlines() so this is
421
# not a guarantee of safety. What would be better is always checking
422
# the content during test suite execution. RBC 20070912
423
return versionedfile.add_lines_with_ghosts(
424
self._new_revision_id, parents, new_lines,
425
nostore_sha=nostore_sha, random_id=self.random_revid,
426
check_content=False)[0:2]
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
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 validator(which is a sha1 digest, though what is sha'd is
502
repository format specific) of the serialized inventory.
504
assert self.is_in_write_group()
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]
525
def add_revision(self, revision_id, rev, inv=None, config=None):
526
"""Add rev to the revision store as revision_id.
528
:param revision_id: the revision id to use.
529
:param rev: The revision object.
530
:param inv: The inventory for the revision. if None, it will be looked
531
up in the inventory storer
532
:param config: If None no digital signature will be created.
533
If supplied its signature_needed method will be used
534
to determine if a signature should be made.
536
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
538
_mod_revision.check_not_reserved_id(revision_id)
539
if config is not None and config.signature_needed():
541
inv = self.get_inventory(revision_id)
542
plaintext = Testament(rev, inv).as_short_text()
543
self.store_revision_signature(
544
gpg.GPGStrategy(config), plaintext, revision_id)
545
if not revision_id in self.get_inventory_weave():
547
raise errors.WeaveRevisionNotPresent(revision_id,
548
self.get_inventory_weave())
550
# yes, this is not suitable for adding with ghosts.
551
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
553
self._revision_store.add_revision(rev, self.get_transaction())
555
def _add_revision_text(self, revision_id, text):
556
revision = self._revision_store._serializer.read_revision_from_string(
558
self._revision_store._add_revision(revision, StringIO(text),
559
self.get_transaction())
561
def all_revision_ids(self):
562
"""Returns a list of all the revision ids in the repository.
564
This is deprecated because code should generally work on the graph
565
reachable from a particular revision, and ignore any other revisions
566
that might be present. There is no direct replacement method.
568
if 'evil' in debug.debug_flags:
569
mutter_callsite(2, "all_revision_ids is linear with history.")
570
return self._all_revision_ids()
572
def _all_revision_ids(self):
573
"""Returns a list of all the revision ids in the repository.
575
These are in as much topological order as the underlying store can
578
raise NotImplementedError(self._all_revision_ids)
580
def break_lock(self):
581
"""Break a lock if one is present from another instance.
583
Uses the ui factory to ask for confirmation if the lock may be from
586
self.control_files.break_lock()
589
def _eliminate_revisions_not_present(self, revision_ids):
590
"""Check every revision id in revision_ids to see if we have it.
592
Returns a set of the present revisions.
595
graph = self.get_graph()
596
parent_map = graph.get_parent_map(revision_ids)
597
# The old API returned a list, should this actually be a set?
598
return parent_map.keys()
601
def create(a_bzrdir):
602
"""Construct the current default format repository in a_bzrdir."""
603
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
605
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
606
"""instantiate a Repository.
608
:param _format: The format of the repository on disk.
609
:param a_bzrdir: The BzrDir of the repository.
611
In the future we will have a single api for all stores for
612
getting file texts, inventories and revisions, then
613
this construct will accept instances of those things.
615
super(Repository, self).__init__()
616
self._format = _format
617
# the following are part of the public API for Repository:
618
self.bzrdir = a_bzrdir
619
self.control_files = control_files
620
self._revision_store = _revision_store
621
# backwards compatibility
622
self.weave_store = text_store
624
self._reconcile_does_inventory_gc = True
625
self._reconcile_fixes_text_parents = False
626
self._reconcile_backsup_inventory = True
627
# not right yet - should be more semantically clear ?
629
self.control_store = control_store
630
self.control_weaves = control_store
631
# TODO: make sure to construct the right store classes, etc, depending
632
# on whether escaping is required.
633
self._warn_if_deprecated()
634
self._write_group = None
635
self.base = control_files._transport.base
638
return '%s(%r)' % (self.__class__.__name__,
641
def has_same_location(self, other):
642
"""Returns a boolean indicating if this repository is at the same
643
location as another repository.
645
This might return False even when two repository objects are accessing
646
the same physical repository via different URLs.
648
if self.__class__ is not other.__class__:
650
return (self.control_files._transport.base ==
651
other.control_files._transport.base)
653
def is_in_write_group(self):
654
"""Return True if there is an open write group.
656
:seealso: start_write_group.
658
return self._write_group is not None
661
return self.control_files.is_locked()
663
def is_write_locked(self):
664
"""Return True if this object is write locked."""
665
return self.is_locked() and self.control_files._lock_mode == 'w'
667
def lock_write(self, token=None):
668
"""Lock this repository for writing.
670
This causes caching within the repository obejct to start accumlating
671
data during reads, and allows a 'write_group' to be obtained. Write
672
groups must be used for actual data insertion.
674
:param token: if this is already locked, then lock_write will fail
675
unless the token matches the existing lock.
676
:returns: a token if this instance supports tokens, otherwise None.
677
:raises TokenLockingNotSupported: when a token is given but this
678
instance doesn't support using token locks.
679
:raises MismatchedToken: if the specified token doesn't match the token
680
of the existing lock.
681
:seealso: start_write_group.
683
A token should be passed in if you know that you have locked the object
684
some other way, and need to synchronise this object's state with that
687
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
689
result = self.control_files.lock_write(token=token)
694
self.control_files.lock_read()
697
def get_physical_lock_status(self):
698
return self.control_files.get_physical_lock_status()
700
def leave_lock_in_place(self):
701
"""Tell this repository not to release the physical lock when this
704
If lock_write doesn't return a token, then this method is not supported.
706
self.control_files.leave_in_place()
708
def dont_leave_lock_in_place(self):
709
"""Tell this repository to release the physical lock when this
710
object is unlocked, even if it didn't originally acquire it.
712
If lock_write doesn't return a token, then this method is not supported.
714
self.control_files.dont_leave_in_place()
717
def gather_stats(self, revid=None, committers=None):
718
"""Gather statistics from a revision id.
720
:param revid: The revision id to gather statistics from, if None, then
721
no revision specific statistics are gathered.
722
:param committers: Optional parameter controlling whether to grab
723
a count of committers from the revision specific statistics.
724
:return: A dictionary of statistics. Currently this contains:
725
committers: The number of committers if requested.
726
firstrev: A tuple with timestamp, timezone for the penultimate left
727
most ancestor of revid, if revid is not the NULL_REVISION.
728
latestrev: A tuple with timestamp, timezone for revid, if revid is
729
not the NULL_REVISION.
730
revisions: The total revision count in the repository.
731
size: An estimate disk size of the repository in bytes.
734
if revid and committers:
735
result['committers'] = 0
736
if revid and revid != _mod_revision.NULL_REVISION:
738
all_committers = set()
739
revisions = self.get_ancestry(revid)
740
# pop the leading None
742
first_revision = None
744
# ignore the revisions in the middle - just grab first and last
745
revisions = revisions[0], revisions[-1]
746
for revision in self.get_revisions(revisions):
747
if not first_revision:
748
first_revision = revision
750
all_committers.add(revision.committer)
751
last_revision = revision
753
result['committers'] = len(all_committers)
754
result['firstrev'] = (first_revision.timestamp,
755
first_revision.timezone)
756
result['latestrev'] = (last_revision.timestamp,
757
last_revision.timezone)
759
# now gather global repository information
760
if self.bzrdir.root_transport.listable():
761
c, t = self._revision_store.total_size(self.get_transaction())
762
result['revisions'] = c
766
def find_branches(self, using=False):
767
"""Find branches underneath this repository.
769
This will include branches inside other branches.
771
:param using: If True, list only branches using this repository.
773
if using and not self.is_shared():
775
return [self.bzrdir.open_branch()]
776
except errors.NotBranchError:
778
class Evaluator(object):
781
self.first_call = True
783
def __call__(self, bzrdir):
784
# On the first call, the parameter is always the bzrdir
785
# containing the current repo.
786
if not self.first_call:
788
repository = bzrdir.open_repository()
789
except errors.NoRepositoryPresent:
792
return False, (None, repository)
793
self.first_call = False
795
value = (bzrdir.open_branch(), None)
796
except errors.NotBranchError:
801
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
802
self.bzrdir.root_transport, evaluate=Evaluator()):
803
if branch is not None:
804
branches.append(branch)
805
if not using and repository is not None:
806
branches.extend(repository.find_branches())
809
def get_data_stream(self, revision_ids):
810
raise NotImplementedError(self.get_data_stream)
812
def get_data_stream_for_search(self, search_result):
813
"""Get a data stream that can be inserted to a repository.
815
:param search_result: A bzrlib.graph.SearchResult selecting the
817
:return: A data stream that can be inserted into a repository using
820
raise NotImplementedError(self.get_data_stream_for_search)
822
def insert_data_stream(self, stream):
823
"""XXX What does this really do?
825
Is it a substitute for fetch?
826
Should it manage its own write group ?
828
for item_key, bytes in stream:
829
if item_key[0] == 'file':
830
(file_id,) = item_key[1:]
831
knit = self.weave_store.get_weave_or_empty(
832
file_id, self.get_transaction())
833
elif item_key == ('inventory',):
834
knit = self.get_inventory_weave()
835
elif item_key == ('revisions',):
836
knit = self._revision_store.get_revision_file(
837
self.get_transaction())
838
elif item_key == ('signatures',):
839
knit = self._revision_store.get_signature_file(
840
self.get_transaction())
842
raise errors.RepositoryDataStreamError(
843
"Unrecognised data stream key '%s'" % (item_key,))
844
decoded_list = bencode.bdecode(bytes)
845
format = decoded_list.pop(0)
848
for version, options, parents, some_bytes in decoded_list:
849
data_list.append((version, options, len(some_bytes), parents))
850
knit_bytes += some_bytes
851
buffer = StringIO(knit_bytes)
852
def reader_func(count):
856
return buffer.read(count)
857
knit.insert_data_stream(
858
(format, data_list, reader_func))
861
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
862
"""Return the revision ids that other has that this does not.
864
These are returned in topological order.
866
revision_id: only return revision ids included by revision_id.
868
return InterRepository.get(other, self).search_missing_revision_ids(
869
revision_id, find_ghosts)
871
@deprecated_method(symbol_versioning.one_two)
873
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
874
"""Return the revision ids that other has that this does not.
876
These are returned in topological order.
878
revision_id: only return revision ids included by revision_id.
880
keys = self.search_missing_revision_ids(
881
other, revision_id, find_ghosts).get_keys()
884
parents = other.get_graph().get_parent_map(keys)
887
return tsort.topo_sort(parents)
891
"""Open the repository rooted at base.
893
For instance, if the repository is at URL/.bzr/repository,
894
Repository.open(URL) -> a Repository instance.
896
control = bzrdir.BzrDir.open(base)
897
return control.open_repository()
899
def copy_content_into(self, destination, revision_id=None):
900
"""Make a complete copy of the content in self into destination.
902
This is a destructive operation! Do not use it on existing
905
return InterRepository.get(self, destination).copy_content(revision_id)
907
def commit_write_group(self):
908
"""Commit the contents accrued within the current write group.
910
:seealso: start_write_group.
912
if self._write_group is not self.get_transaction():
913
# has an unlock or relock occured ?
914
raise errors.BzrError('mismatched lock context %r and '
916
(self.get_transaction(), self._write_group))
917
self._commit_write_group()
918
self._write_group = None
920
def _commit_write_group(self):
921
"""Template method for per-repository write group cleanup.
923
This is called before the write group is considered to be
924
finished and should ensure that all data handed to the repository
925
for writing during the write group is safely committed (to the
926
extent possible considering file system caching etc).
929
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
930
"""Fetch the content required to construct revision_id from source.
932
If revision_id is None all content is copied.
933
:param find_ghosts: Find and copy revisions in the source that are
934
ghosts in the target (and not reachable directly by walking out to
935
the first-present revision in target from revision_id).
937
# fast path same-url fetch operations
938
if self.has_same_location(source):
939
# check that last_revision is in 'from' and then return a
941
if (revision_id is not None and
942
not _mod_revision.is_null(revision_id)):
943
self.get_revision(revision_id)
945
inter = InterRepository.get(source, self)
947
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
948
except NotImplementedError:
949
raise errors.IncompatibleRepositories(source, self)
951
def create_bundle(self, target, base, fileobj, format=None):
952
return serializer.write_bundle(self, target, base, fileobj, format)
954
def get_commit_builder(self, branch, parents, config, timestamp=None,
955
timezone=None, committer=None, revprops=None,
957
"""Obtain a CommitBuilder for this repository.
959
:param branch: Branch to commit to.
960
:param parents: Revision ids of the parents of the new revision.
961
:param config: Configuration to use.
962
:param timestamp: Optional timestamp recorded for commit.
963
:param timezone: Optional timezone for timestamp.
964
:param committer: Optional committer to set for commit.
965
:param revprops: Optional dictionary of revision properties.
966
:param revision_id: Optional revision id.
968
result = self._commit_builder_class(self, parents, config,
969
timestamp, timezone, committer, revprops, revision_id)
970
self.start_write_group()
974
if (self.control_files._lock_count == 1 and
975
self.control_files._lock_mode == 'w'):
976
if self._write_group is not None:
977
self.abort_write_group()
978
self.control_files.unlock()
979
raise errors.BzrError(
980
'Must end write groups before releasing write locks.')
981
self.control_files.unlock()
984
def clone(self, a_bzrdir, revision_id=None):
985
"""Clone this repository into a_bzrdir using the current format.
987
Currently no check is made that the format of this repository and
988
the bzrdir format are compatible. FIXME RBC 20060201.
990
:return: The newly created destination repository.
992
# TODO: deprecate after 0.16; cloning this with all its settings is
993
# probably not very useful -- mbp 20070423
994
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
995
self.copy_content_into(dest_repo, revision_id)
998
def start_write_group(self):
999
"""Start a write group in the repository.
1001
Write groups are used by repositories which do not have a 1:1 mapping
1002
between file ids and backend store to manage the insertion of data from
1003
both fetch and commit operations.
1005
A write lock is required around the start_write_group/commit_write_group
1006
for the support of lock-requiring repository formats.
1008
One can only insert data into a repository inside a write group.
1012
if not self.is_write_locked():
1013
raise errors.NotWriteLocked(self)
1014
if self._write_group:
1015
raise errors.BzrError('already in a write group')
1016
self._start_write_group()
1017
# so we can detect unlock/relock - the write group is now entered.
1018
self._write_group = self.get_transaction()
1020
def _start_write_group(self):
1021
"""Template method for per-repository write group startup.
1023
This is called before the write group is considered to be
1028
def sprout(self, to_bzrdir, revision_id=None):
1029
"""Create a descendent repository for new development.
1031
Unlike clone, this does not copy the settings of the repository.
1033
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1034
dest_repo.fetch(self, revision_id=revision_id)
1037
def _create_sprouting_repo(self, a_bzrdir, shared):
1038
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1039
# use target default format.
1040
dest_repo = a_bzrdir.create_repository()
1042
# Most control formats need the repository to be specifically
1043
# created, but on some old all-in-one formats it's not needed
1045
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1046
except errors.UninitializableFormat:
1047
dest_repo = a_bzrdir.open_repository()
1051
def has_revision(self, revision_id):
1052
"""True if this repository has a copy of the revision."""
1053
return revision_id in self.has_revisions((revision_id,))
1055
def has_revisions(self, revision_ids):
1056
"""Probe to find out the presence of multiple revisions.
1058
:param revision_ids: An iterable of revision_ids.
1059
:return: A set of the revision_ids that were present.
1061
raise NotImplementedError(self.has_revisions)
1063
return self._revision_store.has_revision_id(revision_id,
1064
self.get_transaction())
1067
def get_revision(self, revision_id):
1068
"""Return the Revision object for a named revision."""
1069
return self.get_revisions([revision_id])[0]
1072
def get_revision_reconcile(self, revision_id):
1073
"""'reconcile' helper routine that allows access to a revision always.
1075
This variant of get_revision does not cross check the weave graph
1076
against the revision one as get_revision does: but it should only
1077
be used by reconcile, or reconcile-alike commands that are correcting
1078
or testing the revision graph.
1080
return self._get_revisions([revision_id])[0]
1083
def get_revisions(self, revision_ids):
1084
"""Get many revisions at once."""
1085
return self._get_revisions(revision_ids)
1088
def _get_revisions(self, revision_ids):
1089
"""Core work logic to get many revisions without sanity checks."""
1090
for rev_id in revision_ids:
1091
if not rev_id or not isinstance(rev_id, basestring):
1092
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1093
revs = self._revision_store.get_revisions(revision_ids,
1094
self.get_transaction())
1096
assert not isinstance(rev.revision_id, unicode)
1097
for parent_id in rev.parent_ids:
1098
assert not isinstance(parent_id, unicode)
1102
def get_revision_xml(self, revision_id):
1103
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1104
# would have already do it.
1105
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1106
rev = self.get_revision(revision_id)
1107
rev_tmp = StringIO()
1108
# the current serializer..
1109
self._revision_store._serializer.write_revision(rev, rev_tmp)
1111
return rev_tmp.getvalue()
1113
def get_deltas_for_revisions(self, revisions):
1114
"""Produce a generator of revision deltas.
1116
Note that the input is a sequence of REVISIONS, not revision_ids.
1117
Trees will be held in memory until the generator exits.
1118
Each delta is relative to the revision's lefthand predecessor.
1120
required_trees = set()
1121
for revision in revisions:
1122
required_trees.add(revision.revision_id)
1123
required_trees.update(revision.parent_ids[:1])
1124
trees = dict((t.get_revision_id(), t) for
1125
t in self.revision_trees(required_trees))
1126
for revision in revisions:
1127
if not revision.parent_ids:
1128
old_tree = self.revision_tree(None)
1130
old_tree = trees[revision.parent_ids[0]]
1131
yield trees[revision.revision_id].changes_from(old_tree)
1134
def get_revision_delta(self, revision_id):
1135
"""Return the delta for one revision.
1137
The delta is relative to the left-hand predecessor of the
1140
r = self.get_revision(revision_id)
1141
return list(self.get_deltas_for_revisions([r]))[0]
1144
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1145
signature = gpg_strategy.sign(plaintext)
1146
self.add_signature_text(revision_id, signature)
1149
def add_signature_text(self, revision_id, signature):
1150
self._revision_store.add_revision_signature_text(revision_id,
1152
self.get_transaction())
1154
def find_text_key_references(self):
1155
"""Find the text key references within the repository.
1157
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1158
revision_ids. Each altered file-ids has the exact revision_ids that
1159
altered it listed explicitly.
1160
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1161
to whether they were referred to by the inventory of the
1162
revision_id that they contain. The inventory texts from all present
1163
revision ids are assessed to generate this report.
1165
revision_ids = self.all_revision_ids()
1166
w = self.get_inventory_weave()
1167
pb = ui.ui_factory.nested_progress_bar()
1169
return self._find_text_key_references_from_xml_inventory_lines(
1170
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1174
def _find_text_key_references_from_xml_inventory_lines(self,
1176
"""Core routine for extracting references to texts from inventories.
1178
This performs the translation of xml lines to revision ids.
1180
:param line_iterator: An iterator of lines, origin_version_id
1181
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1182
to whether they were referred to by the inventory of the
1183
revision_id that they contain. Note that if that revision_id was
1184
not part of the line_iterator's output then False will be given -
1185
even though it may actually refer to that key.
1187
if not self._serializer.support_altered_by_hack:
1188
raise AssertionError(
1189
"_find_text_key_references_from_xml_inventory_lines only "
1190
"supported for branches which store inventory as unnested xml"
1191
", not on %r" % self)
1194
# this code needs to read every new line in every inventory for the
1195
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1196
# not present in one of those inventories is unnecessary but not
1197
# harmful because we are filtering by the revision id marker in the
1198
# inventory lines : we only select file ids altered in one of those
1199
# revisions. We don't need to see all lines in the inventory because
1200
# only those added in an inventory in rev X can contain a revision=X
1202
unescape_revid_cache = {}
1203
unescape_fileid_cache = {}
1205
# jam 20061218 In a big fetch, this handles hundreds of thousands
1206
# of lines, so it has had a lot of inlining and optimizing done.
1207
# Sorry that it is a little bit messy.
1208
# Move several functions to be local variables, since this is a long
1210
search = self._file_ids_altered_regex.search
1211
unescape = _unescape_xml
1212
setdefault = result.setdefault
1213
for line, version_id in line_iterator:
1214
match = search(line)
1217
# One call to match.group() returning multiple items is quite a
1218
# bit faster than 2 calls to match.group() each returning 1
1219
file_id, revision_id = match.group('file_id', 'revision_id')
1221
# Inlining the cache lookups helps a lot when you make 170,000
1222
# lines and 350k ids, versus 8.4 unique ids.
1223
# Using a cache helps in 2 ways:
1224
# 1) Avoids unnecessary decoding calls
1225
# 2) Re-uses cached strings, which helps in future set and
1227
# (2) is enough that removing encoding entirely along with
1228
# the cache (so we are using plain strings) results in no
1229
# performance improvement.
1231
revision_id = unescape_revid_cache[revision_id]
1233
unescaped = unescape(revision_id)
1234
unescape_revid_cache[revision_id] = unescaped
1235
revision_id = unescaped
1237
# Note that unconditionally unescaping means that we deserialise
1238
# every fileid, which for general 'pull' is not great, but we don't
1239
# really want to have some many fulltexts that this matters anyway.
1242
file_id = unescape_fileid_cache[file_id]
1244
unescaped = unescape(file_id)
1245
unescape_fileid_cache[file_id] = unescaped
1248
key = (file_id, revision_id)
1249
setdefault(key, False)
1250
if revision_id == version_id:
1254
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1256
"""Helper routine for fileids_altered_by_revision_ids.
1258
This performs the translation of xml lines to revision ids.
1260
:param line_iterator: An iterator of lines, origin_version_id
1261
:param revision_ids: The revision ids to filter for. This should be a
1262
set or other type which supports efficient __contains__ lookups, as
1263
the revision id from each parsed line will be looked up in the
1264
revision_ids filter.
1265
:return: a dictionary mapping altered file-ids to an iterable of
1266
revision_ids. Each altered file-ids has the exact revision_ids that
1267
altered it listed explicitly.
1270
setdefault = result.setdefault
1271
for file_id, revision_id in \
1272
self._find_text_key_references_from_xml_inventory_lines(
1273
line_iterator).iterkeys():
1274
# once data is all ensured-consistent; then this is
1275
# if revision_id == version_id
1276
if revision_id in revision_ids:
1277
setdefault(file_id, set()).add(revision_id)
1280
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1281
"""Find the file ids and versions affected by revisions.
1283
:param revisions: an iterable containing revision ids.
1284
:param _inv_weave: The inventory weave from this repository or None.
1285
If None, the inventory weave will be opened automatically.
1286
:return: a dictionary mapping altered file-ids to an iterable of
1287
revision_ids. Each altered file-ids has the exact revision_ids that
1288
altered it listed explicitly.
1290
selected_revision_ids = set(revision_ids)
1291
w = _inv_weave or self.get_inventory_weave()
1292
pb = ui.ui_factory.nested_progress_bar()
1294
return self._find_file_ids_from_xml_inventory_lines(
1295
w.iter_lines_added_or_present_in_versions(
1296
selected_revision_ids, pb=pb),
1297
selected_revision_ids)
1301
def iter_files_bytes(self, desired_files):
1302
"""Iterate through file versions.
1304
Files will not necessarily be returned in the order they occur in
1305
desired_files. No specific order is guaranteed.
1307
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1308
value supplied by the caller as part of desired_files. It should
1309
uniquely identify the file version in the caller's context. (Examples:
1310
an index number or a TreeTransform trans_id.)
1312
bytes_iterator is an iterable of bytestrings for the file. The
1313
kind of iterable and length of the bytestrings are unspecified, but for
1314
this implementation, it is a list of lines produced by
1315
VersionedFile.get_lines().
1317
:param desired_files: a list of (file_id, revision_id, identifier)
1320
transaction = self.get_transaction()
1321
for file_id, revision_id, callable_data in desired_files:
1323
weave = self.weave_store.get_weave(file_id, transaction)
1324
except errors.NoSuchFile:
1325
raise errors.NoSuchIdInRepository(self, file_id)
1326
yield callable_data, weave.get_lines(revision_id)
1328
def _generate_text_key_index(self, text_key_references=None,
1330
"""Generate a new text key index for the repository.
1332
This is an expensive function that will take considerable time to run.
1334
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1335
list of parents, also text keys. When a given key has no parents,
1336
the parents list will be [NULL_REVISION].
1338
# All revisions, to find inventory parents.
1339
if ancestors is None:
1340
graph = self.get_graph()
1341
ancestors = graph.get_parent_map(self.all_revision_ids())
1342
if text_key_references is None:
1343
text_key_references = self.find_text_key_references()
1344
pb = ui.ui_factory.nested_progress_bar()
1346
return self._do_generate_text_key_index(ancestors,
1347
text_key_references, pb)
1351
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1352
"""Helper for _generate_text_key_index to avoid deep nesting."""
1353
revision_order = tsort.topo_sort(ancestors)
1354
invalid_keys = set()
1356
for revision_id in revision_order:
1357
revision_keys[revision_id] = set()
1358
text_count = len(text_key_references)
1359
# a cache of the text keys to allow reuse; costs a dict of all the
1360
# keys, but saves a 2-tuple for every child of a given key.
1362
for text_key, valid in text_key_references.iteritems():
1364
invalid_keys.add(text_key)
1366
revision_keys[text_key[1]].add(text_key)
1367
text_key_cache[text_key] = text_key
1368
del text_key_references
1370
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1371
NULL_REVISION = _mod_revision.NULL_REVISION
1372
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1373
# too small for large or very branchy trees. However, for 55K path
1374
# trees, it would be easy to use too much memory trivially. Ideally we
1375
# could gauge this by looking at available real memory etc, but this is
1376
# always a tricky proposition.
1377
inventory_cache = lru_cache.LRUCache(10)
1378
batch_size = 10 # should be ~150MB on a 55K path tree
1379
batch_count = len(revision_order) / batch_size + 1
1381
pb.update("Calculating text parents.", processed_texts, text_count)
1382
for offset in xrange(batch_count):
1383
to_query = revision_order[offset * batch_size:(offset + 1) *
1387
for rev_tree in self.revision_trees(to_query):
1388
revision_id = rev_tree.get_revision_id()
1389
parent_ids = ancestors[revision_id]
1390
for text_key in revision_keys[revision_id]:
1391
pb.update("Calculating text parents.", processed_texts)
1392
processed_texts += 1
1393
candidate_parents = []
1394
for parent_id in parent_ids:
1395
parent_text_key = (text_key[0], parent_id)
1397
check_parent = parent_text_key not in \
1398
revision_keys[parent_id]
1400
# the parent parent_id is a ghost:
1401
check_parent = False
1402
# truncate the derived graph against this ghost.
1403
parent_text_key = None
1405
# look at the parent commit details inventories to
1406
# determine possible candidates in the per file graph.
1409
inv = inventory_cache[parent_id]
1411
inv = self.revision_tree(parent_id).inventory
1412
inventory_cache[parent_id] = inv
1413
parent_entry = inv._byid.get(text_key[0], None)
1414
if parent_entry is not None:
1416
text_key[0], parent_entry.revision)
1418
parent_text_key = None
1419
if parent_text_key is not None:
1420
candidate_parents.append(
1421
text_key_cache[parent_text_key])
1422
parent_heads = text_graph.heads(candidate_parents)
1423
new_parents = list(parent_heads)
1424
new_parents.sort(key=lambda x:candidate_parents.index(x))
1425
if new_parents == []:
1426
new_parents = [NULL_REVISION]
1427
text_index[text_key] = new_parents
1429
for text_key in invalid_keys:
1430
text_index[text_key] = [NULL_REVISION]
1433
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1434
"""Get an iterable listing the keys of all the data introduced by a set
1437
The keys will be ordered so that the corresponding items can be safely
1438
fetched and inserted in that order.
1440
:returns: An iterable producing tuples of (knit-kind, file-id,
1441
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1442
'revisions'. file-id is None unless knit-kind is 'file'.
1444
# XXX: it's a bit weird to control the inventory weave caching in this
1445
# generator. Ideally the caching would be done in fetch.py I think. Or
1446
# maybe this generator should explicitly have the contract that it
1447
# should not be iterated until the previously yielded item has been
1449
inv_w = self.get_inventory_weave()
1451
# file ids that changed
1452
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1454
num_file_ids = len(file_ids)
1455
for file_id, altered_versions in file_ids.iteritems():
1456
if _files_pb is not None:
1457
_files_pb.update("fetch texts", count, num_file_ids)
1459
yield ("file", file_id, altered_versions)
1460
# We're done with the files_pb. Note that it finished by the caller,
1461
# just as it was created by the caller.
1465
yield ("inventory", None, revision_ids)
1468
revisions_with_signatures = set()
1469
for rev_id in revision_ids:
1471
self.get_signature_text(rev_id)
1472
except errors.NoSuchRevision:
1476
revisions_with_signatures.add(rev_id)
1477
yield ("signatures", None, revisions_with_signatures)
1480
yield ("revisions", None, revision_ids)
1483
def get_inventory_weave(self):
1484
return self.control_weaves.get_weave('inventory',
1485
self.get_transaction())
1488
def get_inventory(self, revision_id):
1489
"""Get Inventory object by revision id."""
1490
return self.iter_inventories([revision_id]).next()
1492
def iter_inventories(self, revision_ids):
1493
"""Get many inventories by revision_ids.
1495
This will buffer some or all of the texts used in constructing the
1496
inventories in memory, but will only parse a single inventory at a
1499
:return: An iterator of inventories.
1501
assert None not in revision_ids
1502
assert _mod_revision.NULL_REVISION not in revision_ids
1503
return self._iter_inventories(revision_ids)
1505
def _iter_inventories(self, revision_ids):
1506
"""single-document based inventory iteration."""
1507
texts = self.get_inventory_weave().get_texts(revision_ids)
1508
for text, revision_id in zip(texts, revision_ids):
1509
yield self.deserialise_inventory(revision_id, text)
1511
def deserialise_inventory(self, revision_id, xml):
1512
"""Transform the xml into an inventory object.
1514
:param revision_id: The expected revision id of the inventory.
1515
:param xml: A serialised inventory.
1517
result = self._serializer.read_inventory_from_string(xml, revision_id)
1518
if result.revision_id != revision_id:
1519
raise AssertionError('revision id mismatch %s != %s' % (
1520
result.revision_id, revision_id))
1523
def serialise_inventory(self, inv):
1524
return self._serializer.write_inventory_to_string(inv)
1526
def _serialise_inventory_to_lines(self, inv):
1527
return self._serializer.write_inventory_to_lines(inv)
1529
def get_serializer_format(self):
1530
return self._serializer.format_num
1533
def get_inventory_xml(self, revision_id):
1534
"""Get inventory XML as a file object."""
1536
assert isinstance(revision_id, str), type(revision_id)
1537
iw = self.get_inventory_weave()
1538
return iw.get_text(revision_id)
1540
raise errors.HistoryMissing(self, 'inventory', revision_id)
1543
def get_inventory_sha1(self, revision_id):
1544
"""Return the sha1 hash of the inventory entry
1546
return self.get_revision(revision_id).inventory_sha1
1549
@deprecated_method(symbol_versioning.one_four)
1550
def get_revision_graph(self, revision_id=None):
1551
"""Return a dictionary containing the revision graph.
1553
NB: This method should not be used as it accesses the entire graph all
1554
at once, which is much more data than most operations should require.
1556
:param revision_id: The revision_id to get a graph from. If None, then
1557
the entire revision graph is returned. This is a deprecated mode of
1558
operation and will be removed in the future.
1559
:return: a dictionary of revision_id->revision_parents_list.
1561
raise NotImplementedError(self.get_revision_graph)
1564
@deprecated_method(symbol_versioning.one_three)
1565
def get_revision_graph_with_ghosts(self, revision_ids=None):
1566
"""Return a graph of the revisions with ghosts marked as applicable.
1568
:param revision_ids: an iterable of revisions to graph or None for all.
1569
:return: a Graph object with the graph reachable from revision_ids.
1571
if 'evil' in debug.debug_flags:
1573
"get_revision_graph_with_ghosts scales with size of history.")
1574
result = deprecated_graph.Graph()
1575
if not revision_ids:
1576
pending = set(self.all_revision_ids())
1579
pending = set(revision_ids)
1580
# special case NULL_REVISION
1581
if _mod_revision.NULL_REVISION in pending:
1582
pending.remove(_mod_revision.NULL_REVISION)
1583
required = set(pending)
1586
revision_id = pending.pop()
1588
rev = self.get_revision(revision_id)
1589
except errors.NoSuchRevision:
1590
if revision_id in required:
1593
result.add_ghost(revision_id)
1595
for parent_id in rev.parent_ids:
1596
# is this queued or done ?
1597
if (parent_id not in pending and
1598
parent_id not in done):
1600
pending.add(parent_id)
1601
result.add_node(revision_id, rev.parent_ids)
1602
done.add(revision_id)
1605
def iter_reverse_revision_history(self, revision_id):
1606
"""Iterate backwards through revision ids in the lefthand history
1608
:param revision_id: The revision id to start with. All its lefthand
1609
ancestors will be traversed.
1611
graph = self.get_graph()
1612
next_id = revision_id
1614
if next_id in (None, _mod_revision.NULL_REVISION):
1617
# Note: The following line may raise KeyError in the event of
1618
# truncated history. We decided not to have a try:except:raise
1619
# RevisionNotPresent here until we see a use for it, because of the
1620
# cost in an inner loop that is by its very nature O(history).
1621
# Robert Collins 20080326
1622
parents = graph.get_parent_map([next_id])[next_id]
1623
if len(parents) == 0:
1626
next_id = parents[0]
1629
def get_revision_inventory(self, revision_id):
1630
"""Return inventory of a past revision."""
1631
# TODO: Unify this with get_inventory()
1632
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1633
# must be the same as its revision, so this is trivial.
1634
if revision_id is None:
1635
# This does not make sense: if there is no revision,
1636
# then it is the current tree inventory surely ?!
1637
# and thus get_root_id() is something that looks at the last
1638
# commit on the branch, and the get_root_id is an inventory check.
1639
raise NotImplementedError
1640
# return Inventory(self.get_root_id())
1642
return self.get_inventory(revision_id)
1645
def is_shared(self):
1646
"""Return True if this repository is flagged as a shared repository."""
1647
raise NotImplementedError(self.is_shared)
1650
def reconcile(self, other=None, thorough=False):
1651
"""Reconcile this repository."""
1652
from bzrlib.reconcile import RepoReconciler
1653
reconciler = RepoReconciler(self, thorough=thorough)
1654
reconciler.reconcile()
1657
def _refresh_data(self):
1658
"""Helper called from lock_* to ensure coherency with disk.
1660
The default implementation does nothing; it is however possible
1661
for repositories to maintain loaded indices across multiple locks
1662
by checking inside their implementation of this method to see
1663
whether their indices are still valid. This depends of course on
1664
the disk format being validatable in this manner.
1668
def revision_tree(self, revision_id):
1669
"""Return Tree for a revision on this branch.
1671
`revision_id` may be None for the empty tree revision.
1673
# TODO: refactor this to use an existing revision object
1674
# so we don't need to read it in twice.
1675
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1676
return RevisionTree(self, Inventory(root_id=None),
1677
_mod_revision.NULL_REVISION)
1679
inv = self.get_revision_inventory(revision_id)
1680
return RevisionTree(self, inv, revision_id)
1682
def revision_trees(self, revision_ids):
1683
"""Return Tree for a revision on this branch.
1685
`revision_id` may not be None or 'null:'"""
1686
inventories = self.iter_inventories(revision_ids)
1687
for inv in inventories:
1688
yield RevisionTree(self, inv, inv.revision_id)
1691
def get_ancestry(self, revision_id, topo_sorted=True):
1692
"""Return a list of revision-ids integrated by a revision.
1694
The first element of the list is always None, indicating the origin
1695
revision. This might change when we have history horizons, or
1696
perhaps we should have a new API.
1698
This is topologically sorted.
1700
if _mod_revision.is_null(revision_id):
1702
if not self.has_revision(revision_id):
1703
raise errors.NoSuchRevision(self, revision_id)
1704
w = self.get_inventory_weave()
1705
candidates = w.get_ancestry(revision_id, topo_sorted)
1706
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1709
"""Compress the data within the repository.
1711
This operation only makes sense for some repository types. For other
1712
types it should be a no-op that just returns.
1714
This stub method does not require a lock, but subclasses should use
1715
@needs_write_lock as this is a long running call its reasonable to
1716
implicitly lock for the user.
1720
def print_file(self, file, revision_id):
1721
"""Print `file` to stdout.
1723
FIXME RBC 20060125 as John Meinel points out this is a bad api
1724
- it writes to stdout, it assumes that that is valid etc. Fix
1725
by creating a new more flexible convenience function.
1727
tree = self.revision_tree(revision_id)
1728
# use inventory as it was in that revision
1729
file_id = tree.inventory.path2id(file)
1731
# TODO: jam 20060427 Write a test for this code path
1732
# it had a bug in it, and was raising the wrong
1734
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1735
tree.print_file(file_id)
1737
def get_transaction(self):
1738
return self.control_files.get_transaction()
1740
def revision_parents(self, revision_id):
1741
return self.get_inventory_weave().parent_names(revision_id)
1743
@deprecated_method(symbol_versioning.one_one)
1744
def get_parents(self, revision_ids):
1745
"""See StackedParentsProvider.get_parents"""
1746
parent_map = self.get_parent_map(revision_ids)
1747
return [parent_map.get(r, None) for r in revision_ids]
1749
def get_parent_map(self, keys):
1750
"""See graph._StackedParentsProvider.get_parent_map"""
1752
for revision_id in keys:
1753
if revision_id == _mod_revision.NULL_REVISION:
1754
parent_map[revision_id] = ()
1757
parent_id_list = self.get_revision(revision_id).parent_ids
1758
except errors.NoSuchRevision:
1761
if len(parent_id_list) == 0:
1762
parent_ids = (_mod_revision.NULL_REVISION,)
1764
parent_ids = tuple(parent_id_list)
1765
parent_map[revision_id] = parent_ids
1768
def _make_parents_provider(self):
1771
def get_graph(self, other_repository=None):
1772
"""Return the graph walker for this repository format"""
1773
parents_provider = self._make_parents_provider()
1774
if (other_repository is not None and
1775
not self.has_same_location(other_repository)):
1776
parents_provider = graph._StackedParentsProvider(
1777
[parents_provider, other_repository._make_parents_provider()])
1778
return graph.Graph(parents_provider)
1780
def _get_versioned_file_checker(self):
1781
"""Return an object suitable for checking versioned files."""
1782
return _VersionedFileChecker(self)
1784
def revision_ids_to_search_result(self, result_set):
1785
"""Convert a set of revision ids to a graph SearchResult."""
1786
result_parents = set()
1787
for parents in self.get_graph().get_parent_map(
1788
result_set).itervalues():
1789
result_parents.update(parents)
1790
included_keys = result_set.intersection(result_parents)
1791
start_keys = result_set.difference(included_keys)
1792
exclude_keys = result_parents.difference(result_set)
1793
result = graph.SearchResult(start_keys, exclude_keys,
1794
len(result_set), result_set)
1798
def set_make_working_trees(self, new_value):
1799
"""Set the policy flag for making working trees when creating branches.
1801
This only applies to branches that use this repository.
1803
The default is 'True'.
1804
:param new_value: True to restore the default, False to disable making
1807
raise NotImplementedError(self.set_make_working_trees)
1809
def make_working_trees(self):
1810
"""Returns the policy for making working trees on new branches."""
1811
raise NotImplementedError(self.make_working_trees)
1814
def sign_revision(self, revision_id, gpg_strategy):
1815
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1816
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1819
def has_signature_for_revision_id(self, revision_id):
1820
"""Query for a revision signature for revision_id in the repository."""
1821
return self._revision_store.has_signature(revision_id,
1822
self.get_transaction())
1825
def get_signature_text(self, revision_id):
1826
"""Return the text for a signature."""
1827
return self._revision_store.get_signature_text(revision_id,
1828
self.get_transaction())
1831
def check(self, revision_ids=None):
1832
"""Check consistency of all history of given revision_ids.
1834
Different repository implementations should override _check().
1836
:param revision_ids: A non-empty list of revision_ids whose ancestry
1837
will be checked. Typically the last revision_id of a branch.
1839
return self._check(revision_ids)
1841
def _check(self, revision_ids):
1842
result = check.Check(self)
1846
def _warn_if_deprecated(self):
1847
global _deprecation_warning_done
1848
if _deprecation_warning_done:
1850
_deprecation_warning_done = True
1851
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1852
% (self._format, self.bzrdir.transport.base))
1854
def supports_rich_root(self):
1855
return self._format.rich_root_data
1857
def _check_ascii_revisionid(self, revision_id, method):
1858
"""Private helper for ascii-only repositories."""
1859
# weave repositories refuse to store revisionids that are non-ascii.
1860
if revision_id is not None:
1861
# weaves require ascii revision ids.
1862
if isinstance(revision_id, unicode):
1864
revision_id.encode('ascii')
1865
except UnicodeEncodeError:
1866
raise errors.NonAsciiRevisionId(method, self)
1869
revision_id.decode('ascii')
1870
except UnicodeDecodeError:
1871
raise errors.NonAsciiRevisionId(method, self)
1873
def revision_graph_can_have_wrong_parents(self):
1874
"""Is it possible for this repository to have a revision graph with
1877
If True, then this repository must also implement
1878
_find_inconsistent_revision_parents so that check and reconcile can
1879
check for inconsistencies before proceeding with other checks that may
1880
depend on the revision index being consistent.
1882
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1885
# remove these delegates a while after bzr 0.15
1886
def __make_delegated(name, from_module):
1887
def _deprecated_repository_forwarder():
1888
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1889
% (name, from_module),
1892
m = __import__(from_module, globals(), locals(), [name])
1894
return getattr(m, name)
1895
except AttributeError:
1896
raise AttributeError('module %s has no name %s'
1898
globals()[name] = _deprecated_repository_forwarder
1901
'AllInOneRepository',
1902
'WeaveMetaDirRepository',
1903
'PreSplitOutRepositoryFormat',
1904
'RepositoryFormat4',
1905
'RepositoryFormat5',
1906
'RepositoryFormat6',
1907
'RepositoryFormat7',
1909
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1913
'RepositoryFormatKnit',
1914
'RepositoryFormatKnit1',
1916
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1919
def install_revision(repository, rev, revision_tree):
1920
"""Install all revision data into a repository."""
1921
install_revisions(repository, [(rev, revision_tree, None)])
1924
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1925
"""Install all revision data into a repository.
1927
Accepts an iterable of revision, tree, signature tuples. The signature
1930
repository.start_write_group()
1932
for n, (revision, revision_tree, signature) in enumerate(iterable):
1933
_install_revision(repository, revision, revision_tree, signature)
1935
pb.update('Transferring revisions', n + 1, num_revisions)
1937
repository.abort_write_group()
1940
repository.commit_write_group()
1943
def _install_revision(repository, rev, revision_tree, signature):
1944
"""Install all revision data into a repository."""
1945
present_parents = []
1947
for p_id in rev.parent_ids:
1948
if repository.has_revision(p_id):
1949
present_parents.append(p_id)
1950
parent_trees[p_id] = repository.revision_tree(p_id)
1952
parent_trees[p_id] = repository.revision_tree(None)
1954
inv = revision_tree.inventory
1955
entries = inv.iter_entries()
1956
# backwards compatibility hack: skip the root id.
1957
if not repository.supports_rich_root():
1958
path, root = entries.next()
1959
if root.revision != rev.revision_id:
1960
raise errors.IncompatibleRevision(repr(repository))
1961
# Add the texts that are not already present
1962
for path, ie in entries:
1963
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1964
repository.get_transaction())
1965
if ie.revision not in w:
1967
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1968
# with InventoryEntry.find_previous_heads(). if it is, then there
1969
# is a latent bug here where the parents may have ancestors of each
1971
for revision, tree in parent_trees.iteritems():
1972
if ie.file_id not in tree:
1974
parent_id = tree.inventory[ie.file_id].revision
1975
if parent_id in text_parents:
1977
text_parents.append(parent_id)
1979
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1980
repository.get_transaction())
1981
lines = revision_tree.get_file(ie.file_id).readlines()
1982
vfile.add_lines(rev.revision_id, text_parents, lines)
1984
# install the inventory
1985
repository.add_inventory(rev.revision_id, inv, present_parents)
1986
except errors.RevisionAlreadyPresent:
1988
if signature is not None:
1989
repository.add_signature_text(rev.revision_id, signature)
1990
repository.add_revision(rev.revision_id, rev, inv)
1993
class MetaDirRepository(Repository):
1994
"""Repositories in the new meta-dir layout."""
1996
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1997
super(MetaDirRepository, self).__init__(_format,
2003
dir_mode = self.control_files._dir_mode
2004
file_mode = self.control_files._file_mode
2007
def is_shared(self):
2008
"""Return True if this repository is flagged as a shared repository."""
2009
return self.control_files._transport.has('shared-storage')
2012
def set_make_working_trees(self, new_value):
2013
"""Set the policy flag for making working trees when creating branches.
2015
This only applies to branches that use this repository.
2017
The default is 'True'.
2018
:param new_value: True to restore the default, False to disable making
2023
self.control_files._transport.delete('no-working-trees')
2024
except errors.NoSuchFile:
2027
self.control_files.put_utf8('no-working-trees', '')
2029
def make_working_trees(self):
2030
"""Returns the policy for making working trees on new branches."""
2031
return not self.control_files._transport.has('no-working-trees')
2034
class MetaDirVersionedFileRepository(MetaDirRepository):
2035
"""Repositories in a meta-dir, that work via versioned file objects."""
2037
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
2038
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
2039
control_files, _revision_store, control_store, text_store)
2040
_revision_store.get_scope = self.get_transaction
2041
control_store.get_scope = self.get_transaction
2042
text_store.get_scope = self.get_transaction
2045
class RepositoryFormatRegistry(registry.Registry):
2046
"""Registry of RepositoryFormats."""
2048
def get(self, format_string):
2049
r = registry.Registry.get(self, format_string)
2055
format_registry = RepositoryFormatRegistry()
2056
"""Registry of formats, indexed by their identifying format string.
2058
This can contain either format instances themselves, or classes/factories that
2059
can be called to obtain one.
2063
#####################################################################
2064
# Repository Formats
2066
class RepositoryFormat(object):
2067
"""A repository format.
2069
Formats provide three things:
2070
* An initialization routine to construct repository data on disk.
2071
* a format string which is used when the BzrDir supports versioned
2073
* an open routine which returns a Repository instance.
2075
There is one and only one Format subclass for each on-disk format. But
2076
there can be one Repository subclass that is used for several different
2077
formats. The _format attribute on a Repository instance can be used to
2078
determine the disk format.
2080
Formats are placed in an dict by their format string for reference
2081
during opening. These should be subclasses of RepositoryFormat
2084
Once a format is deprecated, just deprecate the initialize and open
2085
methods on the format class. Do not deprecate the object, as the
2086
object will be created every system load.
2088
Common instance attributes:
2089
_matchingbzrdir - the bzrdir format that the repository format was
2090
originally written to work with. This can be used if manually
2091
constructing a bzrdir and repository, or more commonly for test suite
2095
# Set to True or False in derived classes. True indicates that the format
2096
# supports ghosts gracefully.
2097
supports_ghosts = None
2098
# Can this repository be given external locations to lookup additional
2099
# data. Set to True or False in derived classes.
2100
supports_external_lookups = None
2103
return "<%s>" % self.__class__.__name__
2105
def __eq__(self, other):
2106
# format objects are generally stateless
2107
return isinstance(other, self.__class__)
2109
def __ne__(self, other):
2110
return not self == other
2113
def find_format(klass, a_bzrdir):
2114
"""Return the format for the repository object in a_bzrdir.
2116
This is used by bzr native formats that have a "format" file in
2117
the repository. Other methods may be used by different types of
2121
transport = a_bzrdir.get_repository_transport(None)
2122
format_string = transport.get("format").read()
2123
return format_registry.get(format_string)
2124
except errors.NoSuchFile:
2125
raise errors.NoRepositoryPresent(a_bzrdir)
2127
raise errors.UnknownFormatError(format=format_string,
2131
def register_format(klass, format):
2132
format_registry.register(format.get_format_string(), format)
2135
def unregister_format(klass, format):
2136
format_registry.remove(format.get_format_string())
2139
def get_default_format(klass):
2140
"""Return the current default format."""
2141
from bzrlib import bzrdir
2142
return bzrdir.format_registry.make_bzrdir('default').repository_format
2144
def _get_control_store(self, repo_transport, control_files):
2145
"""Return the control store for this repository."""
2146
raise NotImplementedError(self._get_control_store)
2148
def get_format_string(self):
2149
"""Return the ASCII format string that identifies this format.
2151
Note that in pre format ?? repositories the format string is
2152
not permitted nor written to disk.
2154
raise NotImplementedError(self.get_format_string)
2156
def get_format_description(self):
2157
"""Return the short description for this format."""
2158
raise NotImplementedError(self.get_format_description)
2160
def _get_revision_store(self, repo_transport, control_files):
2161
"""Return the revision store object for this a_bzrdir."""
2162
raise NotImplementedError(self._get_revision_store)
2164
def _get_text_rev_store(self,
2171
"""Common logic for getting a revision store for a repository.
2173
see self._get_revision_store for the subclass-overridable method to
2174
get the store for a repository.
2176
from bzrlib.store.revision.text import TextRevisionStore
2177
dir_mode = control_files._dir_mode
2178
file_mode = control_files._file_mode
2179
text_store = TextStore(transport.clone(name),
2181
compressed=compressed,
2183
file_mode=file_mode)
2184
_revision_store = TextRevisionStore(text_store, serializer)
2185
return _revision_store
2187
# TODO: this shouldn't be in the base class, it's specific to things that
2188
# use weaves or knits -- mbp 20070207
2189
def _get_versioned_file_store(self,
2194
versionedfile_class=None,
2195
versionedfile_kwargs={},
2197
if versionedfile_class is None:
2198
versionedfile_class = self._versionedfile_class
2199
weave_transport = control_files._transport.clone(name)
2200
dir_mode = control_files._dir_mode
2201
file_mode = control_files._file_mode
2202
return VersionedFileStore(weave_transport, prefixed=prefixed,
2204
file_mode=file_mode,
2205
versionedfile_class=versionedfile_class,
2206
versionedfile_kwargs=versionedfile_kwargs,
2209
def initialize(self, a_bzrdir, shared=False):
2210
"""Initialize a repository of this format in a_bzrdir.
2212
:param a_bzrdir: The bzrdir to put the new repository in it.
2213
:param shared: The repository should be initialized as a sharable one.
2214
:returns: The new repository object.
2216
This may raise UninitializableFormat if shared repository are not
2217
compatible the a_bzrdir.
2219
raise NotImplementedError(self.initialize)
2221
def is_supported(self):
2222
"""Is this format supported?
2224
Supported formats must be initializable and openable.
2225
Unsupported formats may not support initialization or committing or
2226
some other features depending on the reason for not being supported.
2230
def check_conversion_target(self, target_format):
2231
raise NotImplementedError(self.check_conversion_target)
2233
def open(self, a_bzrdir, _found=False):
2234
"""Return an instance of this format for the bzrdir a_bzrdir.
2236
_found is a private parameter, do not use it.
2238
raise NotImplementedError(self.open)
2241
class MetaDirRepositoryFormat(RepositoryFormat):
2242
"""Common base class for the new repositories using the metadir layout."""
2244
rich_root_data = False
2245
supports_tree_reference = False
2246
supports_external_lookups = False
2247
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2250
super(MetaDirRepositoryFormat, self).__init__()
2252
def _create_control_files(self, a_bzrdir):
2253
"""Create the required files and the initial control_files object."""
2254
# FIXME: RBC 20060125 don't peek under the covers
2255
# NB: no need to escape relative paths that are url safe.
2256
repository_transport = a_bzrdir.get_repository_transport(self)
2257
control_files = lockable_files.LockableFiles(repository_transport,
2258
'lock', lockdir.LockDir)
2259
control_files.create_lock()
2260
return control_files
2262
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2263
"""Upload the initial blank content."""
2264
control_files = self._create_control_files(a_bzrdir)
2265
control_files.lock_write()
2267
control_files._transport.mkdir_multi(dirs,
2268
mode=control_files._dir_mode)
2269
for file, content in files:
2270
control_files.put(file, content)
2271
for file, content in utf8_files:
2272
control_files.put_utf8(file, content)
2274
control_files.put_utf8('shared-storage', '')
2276
control_files.unlock()
2279
# formats which have no format string are not discoverable
2280
# and not independently creatable, so are not registered. They're
2281
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2282
# needed, it's constructed directly by the BzrDir. Non-native formats where
2283
# the repository is not separately opened are similar.
2285
format_registry.register_lazy(
2286
'Bazaar-NG Repository format 7',
2287
'bzrlib.repofmt.weaverepo',
2291
format_registry.register_lazy(
2292
'Bazaar-NG Knit Repository Format 1',
2293
'bzrlib.repofmt.knitrepo',
2294
'RepositoryFormatKnit1',
2297
format_registry.register_lazy(
2298
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2299
'bzrlib.repofmt.knitrepo',
2300
'RepositoryFormatKnit3',
2303
format_registry.register_lazy(
2304
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2305
'bzrlib.repofmt.knitrepo',
2306
'RepositoryFormatKnit4',
2309
# Pack-based formats. There is one format for pre-subtrees, and one for
2310
# post-subtrees to allow ease of testing.
2311
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2312
format_registry.register_lazy(
2313
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2314
'bzrlib.repofmt.pack_repo',
2315
'RepositoryFormatKnitPack1',
2317
format_registry.register_lazy(
2318
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2319
'bzrlib.repofmt.pack_repo',
2320
'RepositoryFormatKnitPack3',
2322
format_registry.register_lazy(
2323
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2324
'bzrlib.repofmt.pack_repo',
2325
'RepositoryFormatKnitPack4',
2327
# Development formats.
2329
# development 0 - stub to introduce development versioning scheme.
2330
format_registry.register_lazy(
2331
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2332
'bzrlib.repofmt.pack_repo',
2333
'RepositoryFormatPackDevelopment0',
2335
format_registry.register_lazy(
2336
("Bazaar development format 0 with subtree support "
2337
"(needs bzr.dev from before 1.3)\n"),
2338
'bzrlib.repofmt.pack_repo',
2339
'RepositoryFormatPackDevelopment0Subtree',
2341
# 1.3->1.4 go below here
2344
class InterRepository(InterObject):
2345
"""This class represents operations taking place between two repositories.
2347
Its instances have methods like copy_content and fetch, and contain
2348
references to the source and target repositories these operations can be
2351
Often we will provide convenience methods on 'repository' which carry out
2352
operations with another repository - they will always forward to
2353
InterRepository.get(other).method_name(parameters).
2357
"""The available optimised InterRepository types."""
2359
def copy_content(self, revision_id=None):
2360
raise NotImplementedError(self.copy_content)
2362
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2363
"""Fetch the content required to construct revision_id.
2365
The content is copied from self.source to self.target.
2367
:param revision_id: if None all content is copied, if NULL_REVISION no
2369
:param pb: optional progress bar to use for progress reports. If not
2370
provided a default one will be created.
2372
Returns the copied revision count and the failed revisions in a tuple:
2375
raise NotImplementedError(self.fetch)
2377
def _walk_to_common_revisions(self, revision_ids):
2378
"""Walk out from revision_ids in source to revisions target has.
2380
:param revision_ids: The start point for the search.
2381
:return: A set of revision ids.
2383
graph = self.source.get_graph()
2384
target_graph = self.target.get_graph()
2385
missing_revs = set()
2386
# ensure we don't pay silly lookup costs.
2387
revision_ids = frozenset(revision_ids)
2388
searcher = graph._make_breadth_first_searcher(revision_ids)
2389
null_set = frozenset([_mod_revision.NULL_REVISION])
2392
next_revs, ghosts = searcher.next_with_ghosts()
2393
except StopIteration:
2395
if revision_ids.intersection(ghosts):
2396
absent_ids = set(revision_ids.intersection(ghosts))
2397
# If all absent_ids are present in target, no error is needed.
2398
absent_ids.difference_update(
2399
set(target_graph.get_parent_map(absent_ids)))
2401
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2402
# we don't care about other ghosts as we can't fetch them and
2403
# haven't been asked to.
2404
next_revs = set(next_revs)
2405
# we always have NULL_REVISION present.
2406
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2407
missing_revs.update(next_revs - have_revs)
2408
searcher.stop_searching_any(have_revs)
2409
return searcher.get_result()
2411
@deprecated_method(symbol_versioning.one_two)
2413
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2414
"""Return the revision ids that source has that target does not.
2416
These are returned in topological order.
2418
:param revision_id: only return revision ids included by this
2420
:param find_ghosts: If True find missing revisions in deep history
2421
rather than just finding the surface difference.
2423
return list(self.search_missing_revision_ids(
2424
revision_id, find_ghosts).get_keys())
2427
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2428
"""Return the revision ids that source has that target does not.
2430
:param revision_id: only return revision ids included by this
2432
:param find_ghosts: If True find missing revisions in deep history
2433
rather than just finding the surface difference.
2434
:return: A bzrlib.graph.SearchResult.
2436
# stop searching at found target revisions.
2437
if not find_ghosts and revision_id is not None:
2438
return self._walk_to_common_revisions([revision_id])
2439
# generic, possibly worst case, slow code path.
2440
target_ids = set(self.target.all_revision_ids())
2441
if revision_id is not None:
2442
source_ids = self.source.get_ancestry(revision_id)
2443
assert source_ids[0] is None
2446
source_ids = self.source.all_revision_ids()
2447
result_set = set(source_ids).difference(target_ids)
2448
return self.source.revision_ids_to_search_result(result_set)
2451
def _same_model(source, target):
2452
"""True if source and target have the same data representation."""
2453
if source.supports_rich_root() != target.supports_rich_root():
2455
if source._serializer != target._serializer:
2460
class InterSameDataRepository(InterRepository):
2461
"""Code for converting between repositories that represent the same data.
2463
Data format and model must match for this to work.
2467
def _get_repo_format_to_test(self):
2468
"""Repository format for testing with.
2470
InterSameData can pull from subtree to subtree and from non-subtree to
2471
non-subtree, so we test this with the richest repository format.
2473
from bzrlib.repofmt import knitrepo
2474
return knitrepo.RepositoryFormatKnit3()
2477
def is_compatible(source, target):
2478
return InterRepository._same_model(source, target)
2481
def copy_content(self, revision_id=None):
2482
"""Make a complete copy of the content in self into destination.
2484
This copies both the repository's revision data, and configuration information
2485
such as the make_working_trees setting.
2487
This is a destructive operation! Do not use it on existing
2490
:param revision_id: Only copy the content needed to construct
2491
revision_id and its parents.
2494
self.target.set_make_working_trees(self.source.make_working_trees())
2495
except NotImplementedError:
2497
# but don't bother fetching if we have the needed data now.
2498
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2499
self.target.has_revision(revision_id)):
2501
self.target.fetch(self.source, revision_id=revision_id)
2504
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2505
"""See InterRepository.fetch()."""
2506
from bzrlib.fetch import GenericRepoFetcher
2507
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2508
self.source, self.source._format, self.target,
2509
self.target._format)
2510
f = GenericRepoFetcher(to_repository=self.target,
2511
from_repository=self.source,
2512
last_revision=revision_id,
2513
pb=pb, find_ghosts=find_ghosts)
2514
return f.count_copied, f.failed_revisions
2517
class InterWeaveRepo(InterSameDataRepository):
2518
"""Optimised code paths between Weave based repositories.
2520
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2521
implemented lazy inter-object optimisation.
2525
def _get_repo_format_to_test(self):
2526
from bzrlib.repofmt import weaverepo
2527
return weaverepo.RepositoryFormat7()
2530
def is_compatible(source, target):
2531
"""Be compatible with known Weave formats.
2533
We don't test for the stores being of specific types because that
2534
could lead to confusing results, and there is no need to be
2537
from bzrlib.repofmt.weaverepo import (
2543
return (isinstance(source._format, (RepositoryFormat5,
2545
RepositoryFormat7)) and
2546
isinstance(target._format, (RepositoryFormat5,
2548
RepositoryFormat7)))
2549
except AttributeError:
2553
def copy_content(self, revision_id=None):
2554
"""See InterRepository.copy_content()."""
2555
# weave specific optimised path:
2557
self.target.set_make_working_trees(self.source.make_working_trees())
2558
except (errors.RepositoryUpgradeRequired, NotImplemented):
2560
# FIXME do not peek!
2561
if self.source.control_files._transport.listable():
2562
pb = ui.ui_factory.nested_progress_bar()
2564
self.target.weave_store.copy_all_ids(
2565
self.source.weave_store,
2567
from_transaction=self.source.get_transaction(),
2568
to_transaction=self.target.get_transaction())
2569
pb.update('copying inventory', 0, 1)
2570
self.target.control_weaves.copy_multi(
2571
self.source.control_weaves, ['inventory'],
2572
from_transaction=self.source.get_transaction(),
2573
to_transaction=self.target.get_transaction())
2574
self.target._revision_store.text_store.copy_all_ids(
2575
self.source._revision_store.text_store,
2580
self.target.fetch(self.source, revision_id=revision_id)
2583
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2584
"""See InterRepository.fetch()."""
2585
from bzrlib.fetch import GenericRepoFetcher
2586
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2587
self.source, self.source._format, self.target, self.target._format)
2588
f = GenericRepoFetcher(to_repository=self.target,
2589
from_repository=self.source,
2590
last_revision=revision_id,
2591
pb=pb, find_ghosts=find_ghosts)
2592
return f.count_copied, f.failed_revisions
2595
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2596
"""See InterRepository.missing_revision_ids()."""
2597
# we want all revisions to satisfy revision_id in source.
2598
# but we don't want to stat every file here and there.
2599
# we want then, all revisions other needs to satisfy revision_id
2600
# checked, but not those that we have locally.
2601
# so the first thing is to get a subset of the revisions to
2602
# satisfy revision_id in source, and then eliminate those that
2603
# we do already have.
2604
# this is slow on high latency connection to self, but as as this
2605
# disk format scales terribly for push anyway due to rewriting
2606
# inventory.weave, this is considered acceptable.
2608
if revision_id is not None:
2609
source_ids = self.source.get_ancestry(revision_id)
2610
assert source_ids[0] is None
2613
source_ids = self.source._all_possible_ids()
2614
source_ids_set = set(source_ids)
2615
# source_ids is the worst possible case we may need to pull.
2616
# now we want to filter source_ids against what we actually
2617
# have in target, but don't try to check for existence where we know
2618
# we do not have a revision as that would be pointless.
2619
target_ids = set(self.target._all_possible_ids())
2620
possibly_present_revisions = target_ids.intersection(source_ids_set)
2621
actually_present_revisions = set(
2622
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2623
required_revisions = source_ids_set.difference(actually_present_revisions)
2624
if revision_id is not None:
2625
# we used get_ancestry to determine source_ids then we are assured all
2626
# revisions referenced are present as they are installed in topological order.
2627
# and the tip revision was validated by get_ancestry.
2628
result_set = required_revisions
2630
# if we just grabbed the possibly available ids, then
2631
# we only have an estimate of whats available and need to validate
2632
# that against the revision records.
2634
self.source._eliminate_revisions_not_present(required_revisions))
2635
return self.source.revision_ids_to_search_result(result_set)
2638
class InterKnitRepo(InterSameDataRepository):
2639
"""Optimised code paths between Knit based repositories."""
2642
def _get_repo_format_to_test(self):
2643
from bzrlib.repofmt import knitrepo
2644
return knitrepo.RepositoryFormatKnit1()
2647
def is_compatible(source, target):
2648
"""Be compatible with known Knit formats.
2650
We don't test for the stores being of specific types because that
2651
could lead to confusing results, and there is no need to be
2654
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2656
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2657
isinstance(target._format, RepositoryFormatKnit))
2658
except AttributeError:
2660
return are_knits and InterRepository._same_model(source, target)
2663
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2664
"""See InterRepository.fetch()."""
2665
from bzrlib.fetch import KnitRepoFetcher
2666
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2667
self.source, self.source._format, self.target, self.target._format)
2668
f = KnitRepoFetcher(to_repository=self.target,
2669
from_repository=self.source,
2670
last_revision=revision_id,
2671
pb=pb, find_ghosts=find_ghosts)
2672
return f.count_copied, f.failed_revisions
2675
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2676
"""See InterRepository.missing_revision_ids()."""
2677
if revision_id is not None:
2678
source_ids = self.source.get_ancestry(revision_id)
2679
assert source_ids[0] is None
2682
source_ids = self.source.all_revision_ids()
2683
source_ids_set = set(source_ids)
2684
# source_ids is the worst possible case we may need to pull.
2685
# now we want to filter source_ids against what we actually
2686
# have in target, but don't try to check for existence where we know
2687
# we do not have a revision as that would be pointless.
2688
target_ids = set(self.target.all_revision_ids())
2689
possibly_present_revisions = target_ids.intersection(source_ids_set)
2690
actually_present_revisions = set(
2691
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2692
required_revisions = source_ids_set.difference(actually_present_revisions)
2693
if revision_id is not None:
2694
# we used get_ancestry to determine source_ids then we are assured all
2695
# revisions referenced are present as they are installed in topological order.
2696
# and the tip revision was validated by get_ancestry.
2697
result_set = required_revisions
2699
# if we just grabbed the possibly available ids, then
2700
# we only have an estimate of whats available and need to validate
2701
# that against the revision records.
2703
self.source._eliminate_revisions_not_present(required_revisions))
2704
return self.source.revision_ids_to_search_result(result_set)
2707
class InterPackRepo(InterSameDataRepository):
2708
"""Optimised code paths between Pack based repositories."""
2711
def _get_repo_format_to_test(self):
2712
from bzrlib.repofmt import pack_repo
2713
return pack_repo.RepositoryFormatKnitPack1()
2716
def is_compatible(source, target):
2717
"""Be compatible with known Pack formats.
2719
We don't test for the stores being of specific types because that
2720
could lead to confusing results, and there is no need to be
2723
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2725
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2726
isinstance(target._format, RepositoryFormatPack))
2727
except AttributeError:
2729
return are_packs and InterRepository._same_model(source, target)
2732
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2733
"""See InterRepository.fetch()."""
2734
from bzrlib.repofmt.pack_repo import Packer
2735
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2736
self.source, self.source._format, self.target, self.target._format)
2737
self.count_copied = 0
2738
if revision_id is None:
2740
# everything to do - use pack logic
2741
# to fetch from all packs to one without
2742
# inventory parsing etc, IFF nothing to be copied is in the target.
2744
revision_ids = self.source.all_revision_ids()
2745
# implementing the TODO will involve:
2746
# - detecting when all of a pack is selected
2747
# - avoiding as much as possible pre-selection, so the
2748
# more-core routines such as create_pack_from_packs can filter in
2749
# a just-in-time fashion. (though having a HEADS list on a
2750
# repository might make this a lot easier, because we could
2751
# sensibly detect 'new revisions' without doing a full index scan.
2752
elif _mod_revision.is_null(revision_id):
2757
revision_ids = self.search_missing_revision_ids(revision_id,
2758
find_ghosts=find_ghosts).get_keys()
2759
except errors.NoSuchRevision:
2760
raise errors.InstallFailed([revision_id])
2761
packs = self.source._pack_collection.all_packs()
2762
pack = Packer(self.target._pack_collection, packs, '.fetch',
2763
revision_ids).pack()
2764
if pack is not None:
2765
self.target._pack_collection._save_pack_names()
2766
# Trigger an autopack. This may duplicate effort as we've just done
2767
# a pack creation, but for now it is simpler to think about as
2768
# 'upload data, then repack if needed'.
2769
self.target._pack_collection.autopack()
2770
return (pack.get_revision_count(), [])
2775
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2776
"""See InterRepository.missing_revision_ids().
2778
:param find_ghosts: Find ghosts throughout the ancestry of
2781
if not find_ghosts and revision_id is not None:
2782
return self._walk_to_common_revisions([revision_id])
2783
elif revision_id is not None:
2784
source_ids = self.source.get_ancestry(revision_id)
2785
assert source_ids[0] is None
2788
source_ids = self.source.all_revision_ids()
2789
# source_ids is the worst possible case we may need to pull.
2790
# now we want to filter source_ids against what we actually
2791
# have in target, but don't try to check for existence where we know
2792
# we do not have a revision as that would be pointless.
2793
target_ids = set(self.target.all_revision_ids())
2794
result_set = set(source_ids).difference(target_ids)
2795
return self.source.revision_ids_to_search_result(result_set)
2798
class InterModel1and2(InterRepository):
2801
def _get_repo_format_to_test(self):
2805
def is_compatible(source, target):
2806
if not source.supports_rich_root() and target.supports_rich_root():
2812
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2813
"""See InterRepository.fetch()."""
2814
from bzrlib.fetch import Model1toKnit2Fetcher
2815
f = Model1toKnit2Fetcher(to_repository=self.target,
2816
from_repository=self.source,
2817
last_revision=revision_id,
2818
pb=pb, find_ghosts=find_ghosts)
2819
return f.count_copied, f.failed_revisions
2822
def copy_content(self, revision_id=None):
2823
"""Make a complete copy of the content in self into destination.
2825
This is a destructive operation! Do not use it on existing
2828
:param revision_id: Only copy the content needed to construct
2829
revision_id and its parents.
2832
self.target.set_make_working_trees(self.source.make_working_trees())
2833
except NotImplementedError:
2835
# but don't bother fetching if we have the needed data now.
2836
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2837
self.target.has_revision(revision_id)):
2839
self.target.fetch(self.source, revision_id=revision_id)
2842
class InterKnit1and2(InterKnitRepo):
2845
def _get_repo_format_to_test(self):
2849
def is_compatible(source, target):
2850
"""Be compatible with Knit1 source and Knit3 target"""
2851
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2853
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2854
RepositoryFormatKnit3)
2855
from bzrlib.repofmt.pack_repo import (
2856
RepositoryFormatKnitPack1,
2857
RepositoryFormatKnitPack3,
2858
RepositoryFormatPackDevelopment0,
2859
RepositoryFormatPackDevelopment0Subtree,
2862
RepositoryFormatKnit1,
2863
RepositoryFormatKnitPack1,
2864
RepositoryFormatPackDevelopment0,
2867
RepositoryFormatKnit3,
2868
RepositoryFormatKnitPack3,
2869
RepositoryFormatPackDevelopment0Subtree,
2871
return (isinstance(source._format, nosubtrees) and
2872
isinstance(target._format, subtrees))
2873
except AttributeError:
2877
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2878
"""See InterRepository.fetch()."""
2879
from bzrlib.fetch import Knit1to2Fetcher
2880
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2881
self.source, self.source._format, self.target,
2882
self.target._format)
2883
f = Knit1to2Fetcher(to_repository=self.target,
2884
from_repository=self.source,
2885
last_revision=revision_id,
2886
pb=pb, find_ghosts=find_ghosts)
2887
return f.count_copied, f.failed_revisions
2890
class InterDifferingSerializer(InterKnitRepo):
2893
def _get_repo_format_to_test(self):
2897
def is_compatible(source, target):
2898
"""Be compatible with Knit2 source and Knit3 target"""
2899
if source.supports_rich_root() != target.supports_rich_root():
2901
# Ideally, we'd support fetching if the source had no tree references
2902
# even if it supported them...
2903
if (getattr(source, '_format.supports_tree_reference', False) and
2904
not getattr(target, '_format.supports_tree_reference', False)):
2909
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2910
"""See InterRepository.fetch()."""
2911
revision_ids = self.target.search_missing_revision_ids(self.source,
2912
revision_id, find_ghosts=find_ghosts).get_keys()
2913
revision_ids = tsort.topo_sort(
2914
self.source.get_graph().get_parent_map(revision_ids))
2915
def revisions_iterator():
2916
for current_revision_id in revision_ids:
2917
revision = self.source.get_revision(current_revision_id)
2918
tree = self.source.revision_tree(current_revision_id)
2920
signature = self.source.get_signature_text(
2921
current_revision_id)
2922
except errors.NoSuchRevision:
2924
yield revision, tree, signature
2926
my_pb = ui.ui_factory.nested_progress_bar()
2931
install_revisions(self.target, revisions_iterator(),
2932
len(revision_ids), pb)
2934
if my_pb is not None:
2936
return len(revision_ids), 0
2939
class InterRemoteToOther(InterRepository):
2941
def __init__(self, source, target):
2942
InterRepository.__init__(self, source, target)
2943
self._real_inter = None
2946
def is_compatible(source, target):
2947
if not isinstance(source, remote.RemoteRepository):
2949
# Is source's model compatible with target's model?
2950
source._ensure_real()
2951
real_source = source._real_repository
2952
assert not isinstance(real_source, remote.RemoteRepository), (
2953
"We don't support remote repos backed by remote repos yet.")
2954
return InterRepository._same_model(real_source, target)
2957
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2958
"""See InterRepository.fetch()."""
2959
from bzrlib.fetch import RemoteToOtherFetcher
2960
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2961
self.source, self.target, self.target._format)
2962
# TODO: jam 20070210 This should be an assert, not a translate
2963
revision_id = osutils.safe_revision_id(revision_id)
2964
f = RemoteToOtherFetcher(to_repository=self.target,
2965
from_repository=self.source,
2966
last_revision=revision_id,
2967
pb=pb, find_ghosts=find_ghosts)
2968
return f.count_copied, f.failed_revisions
2971
def _get_repo_format_to_test(self):
2975
class InterOtherToRemote(InterRepository):
2977
def __init__(self, source, target):
2978
InterRepository.__init__(self, source, target)
2979
self._real_inter = None
2982
def is_compatible(source, target):
2983
if isinstance(target, remote.RemoteRepository):
2987
def _ensure_real_inter(self):
2988
if self._real_inter is None:
2989
self.target._ensure_real()
2990
real_target = self.target._real_repository
2991
self._real_inter = InterRepository.get(self.source, real_target)
2993
def copy_content(self, revision_id=None):
2994
self._ensure_real_inter()
2995
self._real_inter.copy_content(revision_id=revision_id)
2997
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2998
self._ensure_real_inter()
2999
self._real_inter.fetch(revision_id=revision_id, pb=pb,
3000
find_ghosts=find_ghosts)
3003
def _get_repo_format_to_test(self):
3007
InterRepository.register_optimiser(InterDifferingSerializer)
3008
InterRepository.register_optimiser(InterSameDataRepository)
3009
InterRepository.register_optimiser(InterWeaveRepo)
3010
InterRepository.register_optimiser(InterKnitRepo)
3011
InterRepository.register_optimiser(InterModel1and2)
3012
InterRepository.register_optimiser(InterKnit1and2)
3013
InterRepository.register_optimiser(InterPackRepo)
3014
InterRepository.register_optimiser(InterRemoteToOther)
3015
InterRepository.register_optimiser(InterOtherToRemote)
3018
class CopyConverter(object):
3019
"""A repository conversion tool which just performs a copy of the content.
3021
This is slow but quite reliable.
3024
def __init__(self, target_format):
3025
"""Create a CopyConverter.
3027
:param target_format: The format the resulting repository should be.
3029
self.target_format = target_format
3031
def convert(self, repo, pb):
3032
"""Perform the conversion of to_convert, giving feedback via pb.
3034
:param to_convert: The disk object to convert.
3035
:param pb: a progress bar to use for progress information.
3040
# this is only useful with metadir layouts - separated repo content.
3041
# trigger an assertion if not such
3042
repo._format.get_format_string()
3043
self.repo_dir = repo.bzrdir
3044
self.step('Moving repository to repository.backup')
3045
self.repo_dir.transport.move('repository', 'repository.backup')
3046
backup_transport = self.repo_dir.transport.clone('repository.backup')
3047
repo._format.check_conversion_target(self.target_format)
3048
self.source_repo = repo._format.open(self.repo_dir,
3050
_override_transport=backup_transport)
3051
self.step('Creating new repository')
3052
converted = self.target_format.initialize(self.repo_dir,
3053
self.source_repo.is_shared())
3054
converted.lock_write()
3056
self.step('Copying content into repository.')
3057
self.source_repo.copy_content_into(converted)
3060
self.step('Deleting old repository content.')
3061
self.repo_dir.transport.delete_tree('repository.backup')
3062
self.pb.note('repository converted')
3064
def step(self, message):
3065
"""Update the pb by a step."""
3067
self.pb.update(message, self.count, self.total)
3079
def _unescaper(match, _map=_unescape_map):
3080
code = match.group(1)
3084
if not code.startswith('#'):
3086
return unichr(int(code[1:])).encode('utf8')
3092
def _unescape_xml(data):
3093
"""Unescape predefined XML entities in a string of data."""
3095
if _unescape_re is None:
3096
_unescape_re = re.compile('\&([^;]*);')
3097
return _unescape_re.sub(_unescaper, data)
3100
class _VersionedFileChecker(object):
3102
def __init__(self, repository):
3103
self.repository = repository
3104
self.text_index = self.repository._generate_text_key_index()
3106
def calculate_file_version_parents(self, revision_id, file_id):
3107
"""Calculate the correct parents for a file version according to
3110
parent_keys = self.text_index[(file_id, revision_id)]
3111
if parent_keys == [_mod_revision.NULL_REVISION]:
3113
# strip the file_id, for the weave api
3114
return tuple([revision_id for file_id, revision_id in parent_keys])
3116
def check_file_version_parents(self, weave, file_id):
3117
"""Check the parents stored in a versioned file are correct.
3119
It also detects file versions that are not referenced by their
3120
corresponding revision's inventory.
3122
:returns: A tuple of (wrong_parents, dangling_file_versions).
3123
wrong_parents is a dict mapping {revision_id: (stored_parents,
3124
correct_parents)} for each revision_id where the stored parents
3125
are not correct. dangling_file_versions is a set of (file_id,
3126
revision_id) tuples for versions that are present in this versioned
3127
file, but not used by the corresponding inventory.
3130
unused_versions = set()
3131
versions = weave.versions()
3132
parent_map = weave.get_parent_map(versions)
3133
for num, revision_id in enumerate(versions):
3135
correct_parents = self.calculate_file_version_parents(
3136
revision_id, file_id)
3138
# The version is not part of the used keys.
3139
unused_versions.add(revision_id)
3142
knit_parents = tuple(parent_map[revision_id])
3143
except errors.RevisionNotPresent:
3145
if correct_parents != knit_parents:
3146
wrong_parents[revision_id] = (knit_parents, correct_parents)
3147
return wrong_parents, unused_versions
3150
def _old_get_graph(repository, revision_id):
3151
"""DO NOT USE. That is all. I'm serious."""
3152
graph = repository.get_graph()
3153
revision_graph = dict(((key, value) for key, value in
3154
graph.iter_ancestry([revision_id]) if value is not None))
3155
return _strip_NULL_ghosts(revision_graph)
3158
def _strip_NULL_ghosts(revision_graph):
3159
"""Also don't use this. more compatibility code for unmigrated clients."""
3160
# Filter ghosts, and null:
3161
if _mod_revision.NULL_REVISION in revision_graph:
3162
del revision_graph[_mod_revision.NULL_REVISION]
3163
for key, parents in revision_graph.items():
3164
revision_graph[key] = tuple(parent for parent in parents if parent
3166
return revision_graph
added
removed
bzrlib/repository.py
