1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
45
from bzrlib.bundle import serializer
46
from bzrlib.revisiontree import RevisionTree
47
from bzrlib.store.versioned import VersionedFileStore
48
from bzrlib.store.text import TextStore
49
from bzrlib.testament import Testament
50
from bzrlib.util import bencode
53
from bzrlib.decorators import needs_read_lock, needs_write_lock
54
from bzrlib.inter import InterObject
55
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
56
from bzrlib.symbol_versioning import (
59
from bzrlib.trace import mutter, mutter_callsite, note, warning
62
# Old formats display a warning, but only once
63
_deprecation_warning_done = False
66
class CommitBuilder(object):
67
"""Provides an interface to build up a commit.
69
This allows describing a tree to be committed without needing to
70
know the internals of the format of the repository.
73
# all clients should supply tree roots.
74
record_root_entry = True
75
# the default CommitBuilder does not manage trees whose root is versioned.
76
_versioned_root = False
78
def __init__(self, repository, parents, config, timestamp=None,
79
timezone=None, committer=None, revprops=None,
81
"""Initiate a CommitBuilder.
83
:param repository: Repository to commit to.
84
:param parents: Revision ids of the parents of the new revision.
85
:param config: Configuration to use.
86
:param timestamp: Optional timestamp recorded for commit.
87
:param timezone: Optional timezone for timestamp.
88
:param committer: Optional committer to set for commit.
89
:param revprops: Optional dictionary of revision properties.
90
:param revision_id: Optional revision id.
95
self._committer = self._config.username()
97
self._committer = committer
99
self.new_inventory = Inventory(None)
100
self._new_revision_id = revision_id
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self.__heads = graph.HeadsCache(repository.get_graph()).heads
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
raise AssertionError('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _heads(self, file_id, revision_ids):
192
"""Calculate the graph heads for revision_ids in the graph of file_id.
194
This can use either a per-file graph or a global revision graph as we
195
have an identity relationship between the two graphs.
197
return self.__heads(revision_ids)
199
def _check_root(self, ie, parent_invs, tree):
200
"""Helper for record_entry_contents.
202
:param ie: An entry being added.
203
:param parent_invs: The inventories of the parent revisions of the
205
:param tree: The tree that is being committed.
207
# In this revision format, root entries have no knit or weave When
208
# serializing out to disk and back in root.revision is always
210
ie.revision = self._new_revision_id
212
def _get_delta(self, ie, basis_inv, path):
213
"""Get a delta against the basis inventory for ie."""
214
if ie.file_id not in basis_inv:
216
return (None, path, ie.file_id, ie)
217
elif ie != basis_inv[ie.file_id]:
219
# TODO: avoid tis id2path call.
220
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
225
def record_entry_contents(self, ie, parent_invs, path, tree,
227
"""Record the content of ie from tree into the commit if needed.
229
Side effect: sets ie.revision when unchanged
231
:param ie: An inventory entry present in the commit.
232
:param parent_invs: The inventories of the parent revisions of the
234
:param path: The path the entry is at in the tree.
235
:param tree: The tree which contains this entry and should be used to
237
:param content_summary: Summary data from the tree about the paths
238
content - stat, length, exec, sha/link target. This is only
239
accessed when the entry has a revision of None - that is when it is
240
a candidate to commit.
241
:return: A tuple (change_delta, version_recorded). change_delta is
242
an inventory_delta change for this entry against the basis tree of
243
the commit, or None if no change occured against the basis tree.
244
version_recorded is True if a new version of the entry has been
245
recorded. For instance, committing a merge where a file was only
246
changed on the other side will return (delta, False).
248
if self.new_inventory.root is None:
249
if ie.parent_id is not None:
250
raise errors.RootMissing()
251
self._check_root(ie, parent_invs, tree)
252
if ie.revision is None:
253
kind = content_summary[0]
255
# ie is carried over from a prior commit
257
# XXX: repository specific check for nested tree support goes here - if
258
# the repo doesn't want nested trees we skip it ?
259
if (kind == 'tree-reference' and
260
not self.repository._format.supports_tree_reference):
261
# mismatch between commit builder logic and repository:
262
# this needs the entry creation pushed down into the builder.
263
raise NotImplementedError('Missing repository subtree support.')
264
self.new_inventory.add(ie)
266
# TODO: slow, take it out of the inner loop.
268
basis_inv = parent_invs[0]
270
basis_inv = Inventory(root_id=None)
272
# ie.revision is always None if the InventoryEntry is considered
273
# for committing. We may record the previous parents revision if the
274
# content is actually unchanged against a sole head.
275
if ie.revision is not None:
276
if not self._versioned_root and path == '':
277
# repositories that do not version the root set the root's
278
# revision to the new commit even when no change occurs, and
279
# this masks when a change may have occurred against the basis,
280
# so calculate if one happened.
281
if ie.file_id in basis_inv:
282
delta = (basis_inv.id2path(ie.file_id), path,
286
delta = (None, path, ie.file_id, ie)
289
# we don't need to commit this, because the caller already
290
# determined that an existing revision of this file is
292
return None, (ie.revision == self._new_revision_id)
293
# XXX: Friction: parent_candidates should return a list not a dict
294
# so that we don't have to walk the inventories again.
295
parent_candiate_entries = ie.parent_candidates(parent_invs)
296
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
298
for inv in parent_invs:
299
if ie.file_id in inv:
300
old_rev = inv[ie.file_id].revision
301
if old_rev in head_set:
302
heads.append(inv[ie.file_id].revision)
303
head_set.remove(inv[ie.file_id].revision)
306
# now we check to see if we need to write a new record to the
308
# We write a new entry unless there is one head to the ancestors, and
309
# the kind-derived content is unchanged.
311
# Cheapest check first: no ancestors, or more the one head in the
312
# ancestors, we write a new node.
316
# There is a single head, look it up for comparison
317
parent_entry = parent_candiate_entries[heads[0]]
318
# if the non-content specific data has changed, we'll be writing a
320
if (parent_entry.parent_id != ie.parent_id or
321
parent_entry.name != ie.name):
323
# now we need to do content specific checks:
325
# if the kind changed the content obviously has
326
if kind != parent_entry.kind:
329
if content_summary[2] is None:
330
raise ValueError("Files must not have executable = None")
332
if (# if the file length changed we have to store:
333
parent_entry.text_size != content_summary[1] or
334
# if the exec bit has changed we have to store:
335
parent_entry.executable != content_summary[2]):
337
elif parent_entry.text_sha1 == content_summary[3]:
338
# all meta and content is unchanged (using a hash cache
339
# hit to check the sha)
340
ie.revision = parent_entry.revision
341
ie.text_size = parent_entry.text_size
342
ie.text_sha1 = parent_entry.text_sha1
343
ie.executable = parent_entry.executable
344
return self._get_delta(ie, basis_inv, path), False
346
# Either there is only a hash change(no hash cache entry,
347
# or same size content change), or there is no change on
349
# Provide the parent's hash to the store layer, so that the
350
# content is unchanged we will not store a new node.
351
nostore_sha = parent_entry.text_sha1
353
# We want to record a new node regardless of the presence or
354
# absence of a content change in the file.
356
ie.executable = content_summary[2]
357
lines = tree.get_file(ie.file_id, path).readlines()
359
ie.text_sha1, ie.text_size = self._add_text_to_weave(
360
ie.file_id, lines, heads, nostore_sha)
361
except errors.ExistingContent:
362
# Turns out that the file content was unchanged, and we were
363
# only going to store a new node if it was changed. Carry over
365
ie.revision = parent_entry.revision
366
ie.text_size = parent_entry.text_size
367
ie.text_sha1 = parent_entry.text_sha1
368
ie.executable = parent_entry.executable
369
return self._get_delta(ie, basis_inv, path), False
370
elif kind == 'directory':
372
# all data is meta here, nothing specific to directory, so
374
ie.revision = parent_entry.revision
375
return self._get_delta(ie, basis_inv, path), False
377
self._add_text_to_weave(ie.file_id, lines, heads, None)
378
elif kind == 'symlink':
379
current_link_target = content_summary[3]
381
# symlink target is not generic metadata, check if it has
383
if current_link_target != parent_entry.symlink_target:
386
# unchanged, carry over.
387
ie.revision = parent_entry.revision
388
ie.symlink_target = parent_entry.symlink_target
389
return self._get_delta(ie, basis_inv, path), False
390
ie.symlink_target = current_link_target
392
self._add_text_to_weave(ie.file_id, lines, heads, None)
393
elif kind == 'tree-reference':
395
if content_summary[3] != parent_entry.reference_revision:
398
# unchanged, carry over.
399
ie.reference_revision = parent_entry.reference_revision
400
ie.revision = parent_entry.revision
401
return self._get_delta(ie, basis_inv, path), False
402
ie.reference_revision = content_summary[3]
404
self._add_text_to_weave(ie.file_id, lines, heads, None)
406
raise NotImplementedError('unknown kind')
407
ie.revision = self._new_revision_id
408
return self._get_delta(ie, basis_inv, path), True
410
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
411
# Note: as we read the content directly from the tree, we know its not
412
# been turned into unicode or badly split - but a broken tree
413
# implementation could give us bad output from readlines() so this is
414
# not a guarantee of safety. What would be better is always checking
415
# the content during test suite execution. RBC 20070912
416
parent_keys = tuple((file_id, parent) for parent in parents)
417
return self.repository.texts.add_lines(
418
(file_id, self._new_revision_id), parent_keys, new_lines,
419
nostore_sha=nostore_sha, random_id=self.random_revid,
420
check_content=False)[0:2]
423
class RootCommitBuilder(CommitBuilder):
424
"""This commitbuilder actually records the root id"""
426
# the root entry gets versioned properly by this builder.
427
_versioned_root = True
429
def _check_root(self, ie, parent_invs, tree):
430
"""Helper for record_entry_contents.
432
:param ie: An entry being added.
433
:param parent_invs: The inventories of the parent revisions of the
435
:param tree: The tree that is being committed.
439
######################################################################
442
class Repository(object):
443
"""Repository holding history for one or more branches.
445
The repository holds and retrieves historical information including
446
revisions and file history. It's normally accessed only by the Branch,
447
which views a particular line of development through that history.
449
The Repository builds on top of some byte storage facilies (the revisions,
450
signatures, inventories and texts attributes) and a Transport, which
451
respectively provide byte storage and a means to access the (possibly
454
The byte storage facilities are addressed via tuples, which we refer to
455
as 'keys' throughout the code base. Revision_keys, inventory_keys and
456
signature_keys are all 1-tuples: (revision_id,). text_keys are two-tuples:
457
(file_id, revision_id). We use this interface because it allows low
458
friction with the underlying code that implements disk indices, network
459
encoding and other parts of bzrlib.
461
:ivar revisions: A bzrlib.versionedfile.VersionedFiles instance containing
462
the serialised revisions for the repository. This can be used to obtain
463
revision graph information or to access raw serialised revisions.
464
The result of trying to insert data into the repository via this store
465
is undefined: it should be considered read-only except for implementors
467
:ivar signatures: A bzrlib.versionedfile.VersionedFiles instance containing
468
the serialised signatures for the repository. This can be used to
469
obtain access to raw serialised signatures. The result of trying to
470
insert data into the repository via this store is undefined: it should
471
be considered read-only except for implementors of repositories.
472
:ivar inventories: A bzrlib.versionedfile.VersionedFiles instance containing
473
the serialised inventories for the repository. This can be used to
474
obtain unserialised inventories. The result of trying to insert data
475
into the repository via this store is undefined: it should be
476
considered read-only except for implementors of repositories.
477
:ivar texts: A bzrlib.versionedfile.VersionedFiles instance containing the
478
texts of files and directories for the repository. This can be used to
479
obtain file texts or file graphs. Note that Repository.iter_file_bytes
480
is usually a better interface for accessing file texts.
481
The result of trying to insert data into the repository via this store
482
is undefined: it should be considered read-only except for implementors
484
:ivar _transport: Transport for file access to repository, typically
485
pointing to .bzr/repository.
488
# What class to use for a CommitBuilder. Often its simpler to change this
489
# in a Repository class subclass rather than to override
490
# get_commit_builder.
491
_commit_builder_class = CommitBuilder
492
# The search regex used by xml based repositories to determine what things
493
# where changed in a single commit.
494
_file_ids_altered_regex = lazy_regex.lazy_compile(
495
r'file_id="(?P<file_id>[^"]+)"'
496
r'.* revision="(?P<revision_id>[^"]+)"'
499
def abort_write_group(self):
500
"""Commit the contents accrued within the current write group.
502
:seealso: start_write_group.
504
if self._write_group is not self.get_transaction():
505
# has an unlock or relock occured ?
506
raise errors.BzrError('mismatched lock context and write group.')
507
self._abort_write_group()
508
self._write_group = None
510
def _abort_write_group(self):
511
"""Template method for per-repository write group cleanup.
513
This is called during abort before the write group is considered to be
514
finished and should cleanup any internal state accrued during the write
515
group. There is no requirement that data handed to the repository be
516
*not* made available - this is not a rollback - but neither should any
517
attempt be made to ensure that data added is fully commited. Abort is
518
invoked when an error has occured so futher disk or network operations
519
may not be possible or may error and if possible should not be
523
def add_inventory(self, revision_id, inv, parents):
524
"""Add the inventory inv to the repository as revision_id.
526
:param parents: The revision ids of the parents that revision_id
527
is known to have and are in the repository already.
529
:returns: The validator(which is a sha1 digest, though what is sha'd is
530
repository format specific) of the serialized inventory.
532
if not self.is_in_write_group():
533
raise AssertionError("%r not in write group" % (self,))
534
_mod_revision.check_not_reserved_id(revision_id)
535
if not (inv.revision_id is None or inv.revision_id == revision_id):
536
raise AssertionError(
537
"Mismatch between inventory revision"
538
" id and insertion revid (%r, %r)"
539
% (inv.revision_id, revision_id))
541
raise AssertionError()
542
inv_lines = self._serialise_inventory_to_lines(inv)
543
return self._inventory_add_lines(revision_id, parents,
544
inv_lines, check_content=False)
546
def _inventory_add_lines(self, revision_id, parents, lines,
548
"""Store lines in inv_vf and return the sha1 of the inventory."""
549
parents = [(parent,) for parent in parents]
550
return self.inventories.add_lines((revision_id,), parents, lines,
551
check_content=check_content)[0]
553
def add_revision(self, revision_id, rev, inv=None, config=None):
554
"""Add rev to the revision store as revision_id.
556
:param revision_id: the revision id to use.
557
:param rev: The revision object.
558
:param inv: The inventory for the revision. if None, it will be looked
559
up in the inventory storer
560
:param config: If None no digital signature will be created.
561
If supplied its signature_needed method will be used
562
to determine if a signature should be made.
564
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
566
_mod_revision.check_not_reserved_id(revision_id)
567
if config is not None and config.signature_needed():
569
inv = self.get_inventory(revision_id)
570
plaintext = Testament(rev, inv).as_short_text()
571
self.store_revision_signature(
572
gpg.GPGStrategy(config), plaintext, revision_id)
573
# check inventory present
574
if not self.inventories.get_parent_map([(revision_id,)]):
576
raise errors.WeaveRevisionNotPresent(revision_id,
579
# yes, this is not suitable for adding with ghosts.
580
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
583
rev.inventory_sha1 = self.inventories.get_sha1s([(revision_id,)])[0]
584
self._add_revision(rev)
586
def _add_revision(self, revision):
587
text = self._serializer.write_revision_to_string(revision)
588
key = (revision.revision_id,)
589
parents = tuple((parent,) for parent in revision.parent_ids)
590
self.revisions.add_lines(key, parents, osutils.split_lines(text))
592
def all_revision_ids(self):
593
"""Returns a list of all the revision ids in the repository.
595
This is deprecated because code should generally work on the graph
596
reachable from a particular revision, and ignore any other revisions
597
that might be present. There is no direct replacement method.
599
if 'evil' in debug.debug_flags:
600
mutter_callsite(2, "all_revision_ids is linear with history.")
601
return self._all_revision_ids()
603
def _all_revision_ids(self):
604
"""Returns a list of all the revision ids in the repository.
606
These are in as much topological order as the underlying store can
609
raise NotImplementedError(self._all_revision_ids)
611
def break_lock(self):
612
"""Break a lock if one is present from another instance.
614
Uses the ui factory to ask for confirmation if the lock may be from
617
self.control_files.break_lock()
620
def _eliminate_revisions_not_present(self, revision_ids):
621
"""Check every revision id in revision_ids to see if we have it.
623
Returns a set of the present revisions.
626
graph = self.get_graph()
627
parent_map = graph.get_parent_map(revision_ids)
628
# The old API returned a list, should this actually be a set?
629
return parent_map.keys()
632
def create(a_bzrdir):
633
"""Construct the current default format repository in a_bzrdir."""
634
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
636
def __init__(self, _format, a_bzrdir, control_files):
637
"""instantiate a Repository.
639
:param _format: The format of the repository on disk.
640
:param a_bzrdir: The BzrDir of the repository.
641
:param revisions: The revisions store for the repository.
643
In the future we will have a single api for all stores for
644
getting file texts, inventories and revisions, then
645
this construct will accept instances of those things.
647
super(Repository, self).__init__()
648
self._format = _format
649
# the following are part of the public API for Repository:
650
self.bzrdir = a_bzrdir
651
self.control_files = control_files
652
self._transport = control_files._transport
653
self.base = self._transport.base
655
self._reconcile_does_inventory_gc = True
656
self._reconcile_fixes_text_parents = False
657
self._reconcile_backsup_inventory = True
658
# not right yet - should be more semantically clear ?
660
# TODO: make sure to construct the right store classes, etc, depending
661
# on whether escaping is required.
662
self._warn_if_deprecated()
663
self._write_group = None
666
return '%s(%r)' % (self.__class__.__name__,
669
def has_same_location(self, other):
670
"""Returns a boolean indicating if this repository is at the same
671
location as another repository.
673
This might return False even when two repository objects are accessing
674
the same physical repository via different URLs.
676
if self.__class__ is not other.__class__:
678
return (self._transport.base == other._transport.base)
680
def is_in_write_group(self):
681
"""Return True if there is an open write group.
683
:seealso: start_write_group.
685
return self._write_group is not None
688
return self.control_files.is_locked()
690
def is_write_locked(self):
691
"""Return True if this object is write locked."""
692
return self.is_locked() and self.control_files._lock_mode == 'w'
694
def lock_write(self, token=None):
695
"""Lock this repository for writing.
697
This causes caching within the repository obejct to start accumlating
698
data during reads, and allows a 'write_group' to be obtained. Write
699
groups must be used for actual data insertion.
701
:param token: if this is already locked, then lock_write will fail
702
unless the token matches the existing lock.
703
:returns: a token if this instance supports tokens, otherwise None.
704
:raises TokenLockingNotSupported: when a token is given but this
705
instance doesn't support using token locks.
706
:raises MismatchedToken: if the specified token doesn't match the token
707
of the existing lock.
708
:seealso: start_write_group.
710
A token should be passed in if you know that you have locked the object
711
some other way, and need to synchronise this object's state with that
714
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
716
result = self.control_files.lock_write(token=token)
721
self.control_files.lock_read()
724
def get_physical_lock_status(self):
725
return self.control_files.get_physical_lock_status()
727
def leave_lock_in_place(self):
728
"""Tell this repository not to release the physical lock when this
731
If lock_write doesn't return a token, then this method is not supported.
733
self.control_files.leave_in_place()
735
def dont_leave_lock_in_place(self):
736
"""Tell this repository to release the physical lock when this
737
object is unlocked, even if it didn't originally acquire it.
739
If lock_write doesn't return a token, then this method is not supported.
741
self.control_files.dont_leave_in_place()
744
def gather_stats(self, revid=None, committers=None):
745
"""Gather statistics from a revision id.
747
:param revid: The revision id to gather statistics from, if None, then
748
no revision specific statistics are gathered.
749
:param committers: Optional parameter controlling whether to grab
750
a count of committers from the revision specific statistics.
751
:return: A dictionary of statistics. Currently this contains:
752
committers: The number of committers if requested.
753
firstrev: A tuple with timestamp, timezone for the penultimate left
754
most ancestor of revid, if revid is not the NULL_REVISION.
755
latestrev: A tuple with timestamp, timezone for revid, if revid is
756
not the NULL_REVISION.
757
revisions: The total revision count in the repository.
758
size: An estimate disk size of the repository in bytes.
761
if revid and committers:
762
result['committers'] = 0
763
if revid and revid != _mod_revision.NULL_REVISION:
765
all_committers = set()
766
revisions = self.get_ancestry(revid)
767
# pop the leading None
769
first_revision = None
771
# ignore the revisions in the middle - just grab first and last
772
revisions = revisions[0], revisions[-1]
773
for revision in self.get_revisions(revisions):
774
if not first_revision:
775
first_revision = revision
777
all_committers.add(revision.committer)
778
last_revision = revision
780
result['committers'] = len(all_committers)
781
result['firstrev'] = (first_revision.timestamp,
782
first_revision.timezone)
783
result['latestrev'] = (last_revision.timestamp,
784
last_revision.timezone)
786
# now gather global repository information
787
# XXX: This is available for many repos regardless of listability.
788
if self.bzrdir.root_transport.listable():
789
# XXX: do we want to __define len__() ?
790
result['revisions'] = len(self.revisions.keys())
794
def find_branches(self, using=False):
795
"""Find branches underneath this repository.
797
This will include branches inside other branches.
799
:param using: If True, list only branches using this repository.
801
if using and not self.is_shared():
803
return [self.bzrdir.open_branch()]
804
except errors.NotBranchError:
806
class Evaluator(object):
809
self.first_call = True
811
def __call__(self, bzrdir):
812
# On the first call, the parameter is always the bzrdir
813
# containing the current repo.
814
if not self.first_call:
816
repository = bzrdir.open_repository()
817
except errors.NoRepositoryPresent:
820
return False, (None, repository)
821
self.first_call = False
823
value = (bzrdir.open_branch(), None)
824
except errors.NotBranchError:
829
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
830
self.bzrdir.root_transport, evaluate=Evaluator()):
831
if branch is not None:
832
branches.append(branch)
833
if not using and repository is not None:
834
branches.extend(repository.find_branches())
838
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
839
"""Return the revision ids that other has that this does not.
841
These are returned in topological order.
843
revision_id: only return revision ids included by revision_id.
845
return InterRepository.get(other, self).search_missing_revision_ids(
846
revision_id, find_ghosts)
848
@deprecated_method(symbol_versioning.one_two)
850
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
851
"""Return the revision ids that other has that this does not.
853
These are returned in topological order.
855
revision_id: only return revision ids included by revision_id.
857
keys = self.search_missing_revision_ids(
858
other, revision_id, find_ghosts).get_keys()
861
parents = other.get_graph().get_parent_map(keys)
864
return tsort.topo_sort(parents)
868
"""Open the repository rooted at base.
870
For instance, if the repository is at URL/.bzr/repository,
871
Repository.open(URL) -> a Repository instance.
873
control = bzrdir.BzrDir.open(base)
874
return control.open_repository()
876
def copy_content_into(self, destination, revision_id=None):
877
"""Make a complete copy of the content in self into destination.
879
This is a destructive operation! Do not use it on existing
882
return InterRepository.get(self, destination).copy_content(revision_id)
884
def commit_write_group(self):
885
"""Commit the contents accrued within the current write group.
887
:seealso: start_write_group.
889
if self._write_group is not self.get_transaction():
890
# has an unlock or relock occured ?
891
raise errors.BzrError('mismatched lock context %r and '
893
(self.get_transaction(), self._write_group))
894
self._commit_write_group()
895
self._write_group = None
897
def _commit_write_group(self):
898
"""Template method for per-repository write group cleanup.
900
This is called before the write group is considered to be
901
finished and should ensure that all data handed to the repository
902
for writing during the write group is safely committed (to the
903
extent possible considering file system caching etc).
906
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
907
"""Fetch the content required to construct revision_id from source.
909
If revision_id is None all content is copied.
910
:param find_ghosts: Find and copy revisions in the source that are
911
ghosts in the target (and not reachable directly by walking out to
912
the first-present revision in target from revision_id).
914
# fast path same-url fetch operations
915
if self.has_same_location(source):
916
# check that last_revision is in 'from' and then return a
918
if (revision_id is not None and
919
not _mod_revision.is_null(revision_id)):
920
self.get_revision(revision_id)
922
inter = InterRepository.get(source, self)
924
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
925
except NotImplementedError:
926
raise errors.IncompatibleRepositories(source, self)
928
def create_bundle(self, target, base, fileobj, format=None):
929
return serializer.write_bundle(self, target, base, fileobj, format)
931
def get_commit_builder(self, branch, parents, config, timestamp=None,
932
timezone=None, committer=None, revprops=None,
934
"""Obtain a CommitBuilder for this repository.
936
:param branch: Branch to commit to.
937
:param parents: Revision ids of the parents of the new revision.
938
:param config: Configuration to use.
939
:param timestamp: Optional timestamp recorded for commit.
940
:param timezone: Optional timezone for timestamp.
941
:param committer: Optional committer to set for commit.
942
:param revprops: Optional dictionary of revision properties.
943
:param revision_id: Optional revision id.
945
result = self._commit_builder_class(self, parents, config,
946
timestamp, timezone, committer, revprops, revision_id)
947
self.start_write_group()
951
if (self.control_files._lock_count == 1 and
952
self.control_files._lock_mode == 'w'):
953
if self._write_group is not None:
954
self.abort_write_group()
955
self.control_files.unlock()
956
raise errors.BzrError(
957
'Must end write groups before releasing write locks.')
958
self.control_files.unlock()
961
def clone(self, a_bzrdir, revision_id=None):
962
"""Clone this repository into a_bzrdir using the current format.
964
Currently no check is made that the format of this repository and
965
the bzrdir format are compatible. FIXME RBC 20060201.
967
:return: The newly created destination repository.
969
# TODO: deprecate after 0.16; cloning this with all its settings is
970
# probably not very useful -- mbp 20070423
971
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
972
self.copy_content_into(dest_repo, revision_id)
975
def start_write_group(self):
976
"""Start a write group in the repository.
978
Write groups are used by repositories which do not have a 1:1 mapping
979
between file ids and backend store to manage the insertion of data from
980
both fetch and commit operations.
982
A write lock is required around the start_write_group/commit_write_group
983
for the support of lock-requiring repository formats.
985
One can only insert data into a repository inside a write group.
989
if not self.is_write_locked():
990
raise errors.NotWriteLocked(self)
991
if self._write_group:
992
raise errors.BzrError('already in a write group')
993
self._start_write_group()
994
# so we can detect unlock/relock - the write group is now entered.
995
self._write_group = self.get_transaction()
997
def _start_write_group(self):
998
"""Template method for per-repository write group startup.
1000
This is called before the write group is considered to be
1005
def sprout(self, to_bzrdir, revision_id=None):
1006
"""Create a descendent repository for new development.
1008
Unlike clone, this does not copy the settings of the repository.
1010
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1011
dest_repo.fetch(self, revision_id=revision_id)
1014
def _create_sprouting_repo(self, a_bzrdir, shared):
1015
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1016
# use target default format.
1017
dest_repo = a_bzrdir.create_repository()
1019
# Most control formats need the repository to be specifically
1020
# created, but on some old all-in-one formats it's not needed
1022
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1023
except errors.UninitializableFormat:
1024
dest_repo = a_bzrdir.open_repository()
1028
def has_revision(self, revision_id):
1029
"""True if this repository has a copy of the revision."""
1030
return revision_id in self.has_revisions((revision_id,))
1033
def has_revisions(self, revision_ids):
1034
"""Probe to find out the presence of multiple revisions.
1036
:param revision_ids: An iterable of revision_ids.
1037
:return: A set of the revision_ids that were present.
1039
parent_map = self.revisions.get_parent_map(
1040
[(rev_id,) for rev_id in revision_ids])
1042
if _mod_revision.NULL_REVISION in revision_ids:
1043
result.add(_mod_revision.NULL_REVISION)
1044
result.update([key[0] for key in parent_map])
1048
def get_revision(self, revision_id):
1049
"""Return the Revision object for a named revision."""
1050
return self.get_revisions([revision_id])[0]
1053
def get_revision_reconcile(self, revision_id):
1054
"""'reconcile' helper routine that allows access to a revision always.
1056
This variant of get_revision does not cross check the weave graph
1057
against the revision one as get_revision does: but it should only
1058
be used by reconcile, or reconcile-alike commands that are correcting
1059
or testing the revision graph.
1061
return self._get_revisions([revision_id])[0]
1064
def get_revisions(self, revision_ids):
1065
"""Get many revisions at once."""
1066
return self._get_revisions(revision_ids)
1069
def _get_revisions(self, revision_ids):
1070
"""Core work logic to get many revisions without sanity checks."""
1071
for rev_id in revision_ids:
1072
if not rev_id or not isinstance(rev_id, basestring):
1073
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1074
keys = [(key,) for key in revision_ids]
1075
stream = self.revisions.get_record_stream(keys, 'unordered', True)
1077
for record in stream:
1078
if record.storage_kind == 'absent':
1079
raise errors.NoSuchRevision(self, record.key[0])
1080
text = record.get_bytes_as('fulltext')
1081
rev = self._serializer.read_revision_from_string(text)
1082
revs[record.key[0]] = rev
1083
return [revs[revid] for revid in revision_ids]
1086
def get_revision_xml(self, revision_id):
1087
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1088
# would have already do it.
1089
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1090
rev = self.get_revision(revision_id)
1091
rev_tmp = StringIO()
1092
# the current serializer..
1093
self._serializer.write_revision(rev, rev_tmp)
1095
return rev_tmp.getvalue()
1097
def get_deltas_for_revisions(self, revisions):
1098
"""Produce a generator of revision deltas.
1100
Note that the input is a sequence of REVISIONS, not revision_ids.
1101
Trees will be held in memory until the generator exits.
1102
Each delta is relative to the revision's lefthand predecessor.
1104
required_trees = set()
1105
for revision in revisions:
1106
required_trees.add(revision.revision_id)
1107
required_trees.update(revision.parent_ids[:1])
1108
trees = dict((t.get_revision_id(), t) for
1109
t in self.revision_trees(required_trees))
1110
for revision in revisions:
1111
if not revision.parent_ids:
1112
old_tree = self.revision_tree(None)
1114
old_tree = trees[revision.parent_ids[0]]
1115
yield trees[revision.revision_id].changes_from(old_tree)
1118
def get_revision_delta(self, revision_id):
1119
"""Return the delta for one revision.
1121
The delta is relative to the left-hand predecessor of the
1124
r = self.get_revision(revision_id)
1125
return list(self.get_deltas_for_revisions([r]))[0]
1128
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1129
signature = gpg_strategy.sign(plaintext)
1130
self.add_signature_text(revision_id, signature)
1133
def add_signature_text(self, revision_id, signature):
1134
self.signatures.add_lines((revision_id,), (),
1135
osutils.split_lines(signature))
1137
def find_text_key_references(self):
1138
"""Find the text key references within the repository.
1140
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1141
revision_ids. Each altered file-ids has the exact revision_ids that
1142
altered it listed explicitly.
1143
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1144
to whether they were referred to by the inventory of the
1145
revision_id that they contain. The inventory texts from all present
1146
revision ids are assessed to generate this report.
1148
revision_keys = self.revisions.keys()
1149
w = self.inventories
1150
pb = ui.ui_factory.nested_progress_bar()
1152
return self._find_text_key_references_from_xml_inventory_lines(
1153
w.iter_lines_added_or_present_in_keys(revision_keys, pb=pb))
1157
def _find_text_key_references_from_xml_inventory_lines(self,
1159
"""Core routine for extracting references to texts from inventories.
1161
This performs the translation of xml lines to revision ids.
1163
:param line_iterator: An iterator of lines, origin_version_id
1164
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1165
to whether they were referred to by the inventory of the
1166
revision_id that they contain. Note that if that revision_id was
1167
not part of the line_iterator's output then False will be given -
1168
even though it may actually refer to that key.
1170
if not self._serializer.support_altered_by_hack:
1171
raise AssertionError(
1172
"_find_text_key_references_from_xml_inventory_lines only "
1173
"supported for branches which store inventory as unnested xml"
1174
", not on %r" % self)
1177
# this code needs to read every new line in every inventory for the
1178
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1179
# not present in one of those inventories is unnecessary but not
1180
# harmful because we are filtering by the revision id marker in the
1181
# inventory lines : we only select file ids altered in one of those
1182
# revisions. We don't need to see all lines in the inventory because
1183
# only those added in an inventory in rev X can contain a revision=X
1185
unescape_revid_cache = {}
1186
unescape_fileid_cache = {}
1188
# jam 20061218 In a big fetch, this handles hundreds of thousands
1189
# of lines, so it has had a lot of inlining and optimizing done.
1190
# Sorry that it is a little bit messy.
1191
# Move several functions to be local variables, since this is a long
1193
search = self._file_ids_altered_regex.search
1194
unescape = _unescape_xml
1195
setdefault = result.setdefault
1196
for line, line_key in line_iterator:
1197
match = search(line)
1200
# One call to match.group() returning multiple items is quite a
1201
# bit faster than 2 calls to match.group() each returning 1
1202
file_id, revision_id = match.group('file_id', 'revision_id')
1204
# Inlining the cache lookups helps a lot when you make 170,000
1205
# lines and 350k ids, versus 8.4 unique ids.
1206
# Using a cache helps in 2 ways:
1207
# 1) Avoids unnecessary decoding calls
1208
# 2) Re-uses cached strings, which helps in future set and
1210
# (2) is enough that removing encoding entirely along with
1211
# the cache (so we are using plain strings) results in no
1212
# performance improvement.
1214
revision_id = unescape_revid_cache[revision_id]
1216
unescaped = unescape(revision_id)
1217
unescape_revid_cache[revision_id] = unescaped
1218
revision_id = unescaped
1220
# Note that unconditionally unescaping means that we deserialise
1221
# every fileid, which for general 'pull' is not great, but we don't
1222
# really want to have some many fulltexts that this matters anyway.
1225
file_id = unescape_fileid_cache[file_id]
1227
unescaped = unescape(file_id)
1228
unescape_fileid_cache[file_id] = unescaped
1231
key = (file_id, revision_id)
1232
setdefault(key, False)
1233
if revision_id == line_key[-1]:
1237
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1239
"""Helper routine for fileids_altered_by_revision_ids.
1241
This performs the translation of xml lines to revision ids.
1243
:param line_iterator: An iterator of lines, origin_version_id
1244
:param revision_ids: The revision ids to filter for. This should be a
1245
set or other type which supports efficient __contains__ lookups, as
1246
the revision id from each parsed line will be looked up in the
1247
revision_ids filter.
1248
:return: a dictionary mapping altered file-ids to an iterable of
1249
revision_ids. Each altered file-ids has the exact revision_ids that
1250
altered it listed explicitly.
1253
setdefault = result.setdefault
1255
self._find_text_key_references_from_xml_inventory_lines(
1256
line_iterator).iterkeys():
1257
# once data is all ensured-consistent; then this is
1258
# if revision_id == version_id
1259
if key[-1:] in revision_ids:
1260
setdefault(key[0], set()).add(key[-1])
1263
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1264
"""Find the file ids and versions affected by revisions.
1266
:param revisions: an iterable containing revision ids.
1267
:param _inv_weave: The inventory weave from this repository or None.
1268
If None, the inventory weave will be opened automatically.
1269
:return: a dictionary mapping altered file-ids to an iterable of
1270
revision_ids. Each altered file-ids has the exact revision_ids that
1271
altered it listed explicitly.
1273
selected_keys = set((revid,) for revid in revision_ids)
1274
w = _inv_weave or self.inventories
1275
pb = ui.ui_factory.nested_progress_bar()
1277
return self._find_file_ids_from_xml_inventory_lines(
1278
w.iter_lines_added_or_present_in_keys(
1279
selected_keys, pb=pb),
1284
def iter_files_bytes(self, desired_files):
1285
"""Iterate through file versions.
1287
Files will not necessarily be returned in the order they occur in
1288
desired_files. No specific order is guaranteed.
1290
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1291
value supplied by the caller as part of desired_files. It should
1292
uniquely identify the file version in the caller's context. (Examples:
1293
an index number or a TreeTransform trans_id.)
1295
bytes_iterator is an iterable of bytestrings for the file. The
1296
kind of iterable and length of the bytestrings are unspecified, but for
1297
this implementation, it is a list of bytes produced by
1298
VersionedFile.get_record_stream().
1300
:param desired_files: a list of (file_id, revision_id, identifier)
1303
transaction = self.get_transaction()
1305
for file_id, revision_id, callable_data in desired_files:
1306
text_keys[(file_id, revision_id)] = callable_data
1307
for record in self.texts.get_record_stream(text_keys, 'unordered', True):
1308
if record.storage_kind == 'absent':
1309
raise errors.RevisionNotPresent(record.key, self)
1310
yield text_keys[record.key], record.get_bytes_as('fulltext')
1312
def _generate_text_key_index(self, text_key_references=None,
1314
"""Generate a new text key index for the repository.
1316
This is an expensive function that will take considerable time to run.
1318
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1319
list of parents, also text keys. When a given key has no parents,
1320
the parents list will be [NULL_REVISION].
1322
# All revisions, to find inventory parents.
1323
if ancestors is None:
1324
graph = self.get_graph()
1325
ancestors = graph.get_parent_map(self.all_revision_ids())
1326
if text_key_references is None:
1327
text_key_references = self.find_text_key_references()
1328
pb = ui.ui_factory.nested_progress_bar()
1330
return self._do_generate_text_key_index(ancestors,
1331
text_key_references, pb)
1335
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1336
"""Helper for _generate_text_key_index to avoid deep nesting."""
1337
revision_order = tsort.topo_sort(ancestors)
1338
invalid_keys = set()
1340
for revision_id in revision_order:
1341
revision_keys[revision_id] = set()
1342
text_count = len(text_key_references)
1343
# a cache of the text keys to allow reuse; costs a dict of all the
1344
# keys, but saves a 2-tuple for every child of a given key.
1346
for text_key, valid in text_key_references.iteritems():
1348
invalid_keys.add(text_key)
1350
revision_keys[text_key[1]].add(text_key)
1351
text_key_cache[text_key] = text_key
1352
del text_key_references
1354
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1355
NULL_REVISION = _mod_revision.NULL_REVISION
1356
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1357
# too small for large or very branchy trees. However, for 55K path
1358
# trees, it would be easy to use too much memory trivially. Ideally we
1359
# could gauge this by looking at available real memory etc, but this is
1360
# always a tricky proposition.
1361
inventory_cache = lru_cache.LRUCache(10)
1362
batch_size = 10 # should be ~150MB on a 55K path tree
1363
batch_count = len(revision_order) / batch_size + 1
1365
pb.update("Calculating text parents.", processed_texts, text_count)
1366
for offset in xrange(batch_count):
1367
to_query = revision_order[offset * batch_size:(offset + 1) *
1371
for rev_tree in self.revision_trees(to_query):
1372
revision_id = rev_tree.get_revision_id()
1373
parent_ids = ancestors[revision_id]
1374
for text_key in revision_keys[revision_id]:
1375
pb.update("Calculating text parents.", processed_texts)
1376
processed_texts += 1
1377
candidate_parents = []
1378
for parent_id in parent_ids:
1379
parent_text_key = (text_key[0], parent_id)
1381
check_parent = parent_text_key not in \
1382
revision_keys[parent_id]
1384
# the parent parent_id is a ghost:
1385
check_parent = False
1386
# truncate the derived graph against this ghost.
1387
parent_text_key = None
1389
# look at the parent commit details inventories to
1390
# determine possible candidates in the per file graph.
1393
inv = inventory_cache[parent_id]
1395
inv = self.revision_tree(parent_id).inventory
1396
inventory_cache[parent_id] = inv
1397
parent_entry = inv._byid.get(text_key[0], None)
1398
if parent_entry is not None:
1400
text_key[0], parent_entry.revision)
1402
parent_text_key = None
1403
if parent_text_key is not None:
1404
candidate_parents.append(
1405
text_key_cache[parent_text_key])
1406
parent_heads = text_graph.heads(candidate_parents)
1407
new_parents = list(parent_heads)
1408
new_parents.sort(key=lambda x:candidate_parents.index(x))
1409
if new_parents == []:
1410
new_parents = [NULL_REVISION]
1411
text_index[text_key] = new_parents
1413
for text_key in invalid_keys:
1414
text_index[text_key] = [NULL_REVISION]
1417
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1418
"""Get an iterable listing the keys of all the data introduced by a set
1421
The keys will be ordered so that the corresponding items can be safely
1422
fetched and inserted in that order.
1424
:returns: An iterable producing tuples of (knit-kind, file-id,
1425
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1426
'revisions'. file-id is None unless knit-kind is 'file'.
1428
# XXX: it's a bit weird to control the inventory weave caching in this
1429
# generator. Ideally the caching would be done in fetch.py I think. Or
1430
# maybe this generator should explicitly have the contract that it
1431
# should not be iterated until the previously yielded item has been
1433
inv_w = self.inventories
1435
# file ids that changed
1436
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1438
num_file_ids = len(file_ids)
1439
for file_id, altered_versions in file_ids.iteritems():
1440
if _files_pb is not None:
1441
_files_pb.update("fetch texts", count, num_file_ids)
1443
yield ("file", file_id, altered_versions)
1444
# We're done with the files_pb. Note that it finished by the caller,
1445
# just as it was created by the caller.
1449
yield ("inventory", None, revision_ids)
1452
revisions_with_signatures = set()
1453
for rev_id in revision_ids:
1455
self.get_signature_text(rev_id)
1456
except errors.NoSuchRevision:
1460
revisions_with_signatures.add(rev_id)
1461
yield ("signatures", None, revisions_with_signatures)
1464
yield ("revisions", None, revision_ids)
1467
def get_inventory(self, revision_id):
1468
"""Get Inventory object by revision id."""
1469
return self.iter_inventories([revision_id]).next()
1471
def iter_inventories(self, revision_ids):
1472
"""Get many inventories by revision_ids.
1474
This will buffer some or all of the texts used in constructing the
1475
inventories in memory, but will only parse a single inventory at a
1478
:return: An iterator of inventories.
1480
if ((None in revision_ids)
1481
or (_mod_revision.NULL_REVISION in revision_ids)):
1482
raise ValueError('cannot get null revision inventory')
1483
return self._iter_inventories(revision_ids)
1485
def _iter_inventories(self, revision_ids):
1486
"""single-document based inventory iteration."""
1487
for text, revision_id in self._iter_inventory_xmls(revision_ids):
1488
yield self.deserialise_inventory(revision_id, text)
1490
def _iter_inventory_xmls(self, revision_ids):
1491
keys = [(revision_id,) for revision_id in revision_ids]
1492
stream = self.inventories.get_record_stream(keys, 'unordered', True)
1494
for record in stream:
1495
if record.storage_kind != 'absent':
1496
texts[record.key] = record.get_bytes_as('fulltext')
1498
raise errors.NoSuchRevision(self, record.key)
1500
yield texts[key], key[-1]
1502
def deserialise_inventory(self, revision_id, xml):
1503
"""Transform the xml into an inventory object.
1505
:param revision_id: The expected revision id of the inventory.
1506
:param xml: A serialised inventory.
1508
result = self._serializer.read_inventory_from_string(xml, revision_id)
1509
if result.revision_id != revision_id:
1510
raise AssertionError('revision id mismatch %s != %s' % (
1511
result.revision_id, revision_id))
1514
def serialise_inventory(self, inv):
1515
return self._serializer.write_inventory_to_string(inv)
1517
def _serialise_inventory_to_lines(self, inv):
1518
return self._serializer.write_inventory_to_lines(inv)
1520
def get_serializer_format(self):
1521
return self._serializer.format_num
1524
def get_inventory_xml(self, revision_id):
1525
"""Get inventory XML as a file object."""
1526
texts = self._iter_inventory_xmls([revision_id])
1528
text, revision_id = texts.next()
1529
except StopIteration:
1530
raise errors.HistoryMissing(self, 'inventory', revision_id)
1534
def get_inventory_sha1(self, revision_id):
1535
"""Return the sha1 hash of the inventory entry
1537
return self.get_revision(revision_id).inventory_sha1
1539
def iter_reverse_revision_history(self, revision_id):
1540
"""Iterate backwards through revision ids in the lefthand history
1542
:param revision_id: The revision id to start with. All its lefthand
1543
ancestors will be traversed.
1545
graph = self.get_graph()
1546
next_id = revision_id
1548
if next_id in (None, _mod_revision.NULL_REVISION):
1551
# Note: The following line may raise KeyError in the event of
1552
# truncated history. We decided not to have a try:except:raise
1553
# RevisionNotPresent here until we see a use for it, because of the
1554
# cost in an inner loop that is by its very nature O(history).
1555
# Robert Collins 20080326
1556
parents = graph.get_parent_map([next_id])[next_id]
1557
if len(parents) == 0:
1560
next_id = parents[0]
1563
def get_revision_inventory(self, revision_id):
1564
"""Return inventory of a past revision."""
1565
# TODO: Unify this with get_inventory()
1566
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1567
# must be the same as its revision, so this is trivial.
1568
if revision_id is None:
1569
# This does not make sense: if there is no revision,
1570
# then it is the current tree inventory surely ?!
1571
# and thus get_root_id() is something that looks at the last
1572
# commit on the branch, and the get_root_id is an inventory check.
1573
raise NotImplementedError
1574
# return Inventory(self.get_root_id())
1576
return self.get_inventory(revision_id)
1579
def is_shared(self):
1580
"""Return True if this repository is flagged as a shared repository."""
1581
raise NotImplementedError(self.is_shared)
1584
def reconcile(self, other=None, thorough=False):
1585
"""Reconcile this repository."""
1586
from bzrlib.reconcile import RepoReconciler
1587
reconciler = RepoReconciler(self, thorough=thorough)
1588
reconciler.reconcile()
1591
def _refresh_data(self):
1592
"""Helper called from lock_* to ensure coherency with disk.
1594
The default implementation does nothing; it is however possible
1595
for repositories to maintain loaded indices across multiple locks
1596
by checking inside their implementation of this method to see
1597
whether their indices are still valid. This depends of course on
1598
the disk format being validatable in this manner.
1602
def revision_tree(self, revision_id):
1603
"""Return Tree for a revision on this branch.
1605
`revision_id` may be None for the empty tree revision.
1607
# TODO: refactor this to use an existing revision object
1608
# so we don't need to read it in twice.
1609
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1610
return RevisionTree(self, Inventory(root_id=None),
1611
_mod_revision.NULL_REVISION)
1613
inv = self.get_revision_inventory(revision_id)
1614
return RevisionTree(self, inv, revision_id)
1616
def revision_trees(self, revision_ids):
1617
"""Return Tree for a revision on this branch.
1619
`revision_id` may not be None or 'null:'"""
1620
inventories = self.iter_inventories(revision_ids)
1621
for inv in inventories:
1622
yield RevisionTree(self, inv, inv.revision_id)
1625
def get_ancestry(self, revision_id, topo_sorted=True):
1626
"""Return a list of revision-ids integrated by a revision.
1628
The first element of the list is always None, indicating the origin
1629
revision. This might change when we have history horizons, or
1630
perhaps we should have a new API.
1632
This is topologically sorted.
1634
if _mod_revision.is_null(revision_id):
1636
if not self.has_revision(revision_id):
1637
raise errors.NoSuchRevision(self, revision_id)
1638
graph = self.get_graph()
1640
search = graph._make_breadth_first_searcher([revision_id])
1643
found, ghosts = search.next_with_ghosts()
1644
except StopIteration:
1647
if _mod_revision.NULL_REVISION in keys:
1648
keys.remove(_mod_revision.NULL_REVISION)
1650
parent_map = graph.get_parent_map(keys)
1651
keys = tsort.topo_sort(parent_map)
1652
return [None] + list(keys)
1655
"""Compress the data within the repository.
1657
This operation only makes sense for some repository types. For other
1658
types it should be a no-op that just returns.
1660
This stub method does not require a lock, but subclasses should use
1661
@needs_write_lock as this is a long running call its reasonable to
1662
implicitly lock for the user.
1666
def print_file(self, file, revision_id):
1667
"""Print `file` to stdout.
1669
FIXME RBC 20060125 as John Meinel points out this is a bad api
1670
- it writes to stdout, it assumes that that is valid etc. Fix
1671
by creating a new more flexible convenience function.
1673
tree = self.revision_tree(revision_id)
1674
# use inventory as it was in that revision
1675
file_id = tree.inventory.path2id(file)
1677
# TODO: jam 20060427 Write a test for this code path
1678
# it had a bug in it, and was raising the wrong
1680
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1681
tree.print_file(file_id)
1683
def get_transaction(self):
1684
return self.control_files.get_transaction()
1686
@deprecated_method(symbol_versioning.one_one)
1687
def get_parents(self, revision_ids):
1688
"""See StackedParentsProvider.get_parents"""
1689
parent_map = self.get_parent_map(revision_ids)
1690
return [parent_map.get(r, None) for r in revision_ids]
1692
def get_parent_map(self, keys):
1693
"""See graph._StackedParentsProvider.get_parent_map"""
1695
for revision_id in keys:
1696
if revision_id is None:
1697
raise ValueError('get_parent_map(None) is not valid')
1698
if revision_id == _mod_revision.NULL_REVISION:
1699
parent_map[revision_id] = ()
1702
parent_id_list = self.get_revision(revision_id).parent_ids
1703
except errors.NoSuchRevision:
1706
if len(parent_id_list) == 0:
1707
parent_ids = (_mod_revision.NULL_REVISION,)
1709
parent_ids = tuple(parent_id_list)
1710
parent_map[revision_id] = parent_ids
1713
def _make_parents_provider(self):
1716
def get_graph(self, other_repository=None):
1717
"""Return the graph walker for this repository format"""
1718
parents_provider = self._make_parents_provider()
1719
if (other_repository is not None and
1720
not self.has_same_location(other_repository)):
1721
parents_provider = graph._StackedParentsProvider(
1722
[parents_provider, other_repository._make_parents_provider()])
1723
return graph.Graph(parents_provider)
1725
def _get_versioned_file_checker(self):
1726
"""Return an object suitable for checking versioned files."""
1727
return _VersionedFileChecker(self)
1729
def revision_ids_to_search_result(self, result_set):
1730
"""Convert a set of revision ids to a graph SearchResult."""
1731
result_parents = set()
1732
for parents in self.get_graph().get_parent_map(
1733
result_set).itervalues():
1734
result_parents.update(parents)
1735
included_keys = result_set.intersection(result_parents)
1736
start_keys = result_set.difference(included_keys)
1737
exclude_keys = result_parents.difference(result_set)
1738
result = graph.SearchResult(start_keys, exclude_keys,
1739
len(result_set), result_set)
1743
def set_make_working_trees(self, new_value):
1744
"""Set the policy flag for making working trees when creating branches.
1746
This only applies to branches that use this repository.
1748
The default is 'True'.
1749
:param new_value: True to restore the default, False to disable making
1752
raise NotImplementedError(self.set_make_working_trees)
1754
def make_working_trees(self):
1755
"""Returns the policy for making working trees on new branches."""
1756
raise NotImplementedError(self.make_working_trees)
1759
def sign_revision(self, revision_id, gpg_strategy):
1760
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1761
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1764
def has_signature_for_revision_id(self, revision_id):
1765
"""Query for a revision signature for revision_id in the repository."""
1766
if not self.has_revision(revision_id):
1767
raise errors.NoSuchRevision(self, revision_id)
1768
sig_present = (1 == len(
1769
self.signatures.get_parent_map([(revision_id,)])))
1773
def get_signature_text(self, revision_id):
1774
"""Return the text for a signature."""
1775
stream = self.signatures.get_record_stream([(revision_id,)],
1777
record = stream.next()
1778
if record.storage_kind == 'absent':
1779
raise errors.NoSuchRevision(self, revision_id)
1780
return record.get_bytes_as('fulltext')
1783
def check(self, revision_ids=None):
1784
"""Check consistency of all history of given revision_ids.
1786
Different repository implementations should override _check().
1788
:param revision_ids: A non-empty list of revision_ids whose ancestry
1789
will be checked. Typically the last revision_id of a branch.
1791
return self._check(revision_ids)
1793
def _check(self, revision_ids):
1794
result = check.Check(self)
1798
def _warn_if_deprecated(self):
1799
global _deprecation_warning_done
1800
if _deprecation_warning_done:
1802
_deprecation_warning_done = True
1803
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1804
% (self._format, self.bzrdir.transport.base))
1806
def supports_rich_root(self):
1807
return self._format.rich_root_data
1809
def _check_ascii_revisionid(self, revision_id, method):
1810
"""Private helper for ascii-only repositories."""
1811
# weave repositories refuse to store revisionids that are non-ascii.
1812
if revision_id is not None:
1813
# weaves require ascii revision ids.
1814
if isinstance(revision_id, unicode):
1816
revision_id.encode('ascii')
1817
except UnicodeEncodeError:
1818
raise errors.NonAsciiRevisionId(method, self)
1821
revision_id.decode('ascii')
1822
except UnicodeDecodeError:
1823
raise errors.NonAsciiRevisionId(method, self)
1825
def revision_graph_can_have_wrong_parents(self):
1826
"""Is it possible for this repository to have a revision graph with
1829
If True, then this repository must also implement
1830
_find_inconsistent_revision_parents so that check and reconcile can
1831
check for inconsistencies before proceeding with other checks that may
1832
depend on the revision index being consistent.
1834
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1837
# remove these delegates a while after bzr 0.15
1838
def __make_delegated(name, from_module):
1839
def _deprecated_repository_forwarder():
1840
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1841
% (name, from_module),
1844
m = __import__(from_module, globals(), locals(), [name])
1846
return getattr(m, name)
1847
except AttributeError:
1848
raise AttributeError('module %s has no name %s'
1850
globals()[name] = _deprecated_repository_forwarder
1853
'AllInOneRepository',
1854
'WeaveMetaDirRepository',
1855
'PreSplitOutRepositoryFormat',
1856
'RepositoryFormat4',
1857
'RepositoryFormat5',
1858
'RepositoryFormat6',
1859
'RepositoryFormat7',
1861
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1865
'RepositoryFormatKnit',
1866
'RepositoryFormatKnit1',
1868
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1871
def install_revision(repository, rev, revision_tree):
1872
"""Install all revision data into a repository."""
1873
install_revisions(repository, [(rev, revision_tree, None)])
1876
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1877
"""Install all revision data into a repository.
1879
Accepts an iterable of revision, tree, signature tuples. The signature
1882
repository.start_write_group()
1884
for n, (revision, revision_tree, signature) in enumerate(iterable):
1885
_install_revision(repository, revision, revision_tree, signature)
1887
pb.update('Transferring revisions', n + 1, num_revisions)
1889
repository.abort_write_group()
1892
repository.commit_write_group()
1895
def _install_revision(repository, rev, revision_tree, signature):
1896
"""Install all revision data into a repository."""
1897
present_parents = []
1899
for p_id in rev.parent_ids:
1900
if repository.has_revision(p_id):
1901
present_parents.append(p_id)
1902
parent_trees[p_id] = repository.revision_tree(p_id)
1904
parent_trees[p_id] = repository.revision_tree(None)
1906
inv = revision_tree.inventory
1907
entries = inv.iter_entries()
1908
# backwards compatibility hack: skip the root id.
1909
if not repository.supports_rich_root():
1910
path, root = entries.next()
1911
if root.revision != rev.revision_id:
1912
raise errors.IncompatibleRevision(repr(repository))
1914
for path, ie in entries:
1915
text_keys[(ie.file_id, ie.revision)] = ie
1916
text_parent_map = repository.texts.get_parent_map(text_keys)
1917
missing_texts = set(text_keys) - set(text_parent_map)
1918
# Add the texts that are not already present
1919
for text_key in missing_texts:
1920
ie = text_keys[text_key]
1922
# FIXME: TODO: The following loop overlaps/duplicates that done by
1923
# commit to determine parents. There is a latent/real bug here where
1924
# the parents inserted are not those commit would do - in particular
1925
# they are not filtered by heads(). RBC, AB
1926
for revision, tree in parent_trees.iteritems():
1927
if ie.file_id not in tree:
1929
parent_id = tree.inventory[ie.file_id].revision
1930
if parent_id in text_parents:
1932
text_parents.append((ie.file_id, parent_id))
1933
lines = revision_tree.get_file(ie.file_id).readlines()
1934
repository.texts.add_lines(text_key, text_parents, lines)
1936
# install the inventory
1937
repository.add_inventory(rev.revision_id, inv, present_parents)
1938
except errors.RevisionAlreadyPresent:
1940
if signature is not None:
1941
repository.add_signature_text(rev.revision_id, signature)
1942
repository.add_revision(rev.revision_id, rev, inv)
1945
class MetaDirRepository(Repository):
1946
"""Repositories in the new meta-dir layout.
1948
:ivar _transport: Transport for access to repository control files,
1949
typically pointing to .bzr/repository.
1952
def __init__(self, _format, a_bzrdir, control_files):
1953
super(MetaDirRepository, self).__init__(_format, a_bzrdir, control_files)
1954
self._transport = control_files._transport
1957
def is_shared(self):
1958
"""Return True if this repository is flagged as a shared repository."""
1959
return self._transport.has('shared-storage')
1962
def set_make_working_trees(self, new_value):
1963
"""Set the policy flag for making working trees when creating branches.
1965
This only applies to branches that use this repository.
1967
The default is 'True'.
1968
:param new_value: True to restore the default, False to disable making
1973
self._transport.delete('no-working-trees')
1974
except errors.NoSuchFile:
1977
self._transport.put_bytes('no-working-trees', '',
1978
mode=self.bzrdir._get_file_mode())
1980
def make_working_trees(self):
1981
"""Returns the policy for making working trees on new branches."""
1982
return not self._transport.has('no-working-trees')
1985
class MetaDirVersionedFileRepository(MetaDirRepository):
1986
"""Repositories in a meta-dir, that work via versioned file objects."""
1988
def __init__(self, _format, a_bzrdir, control_files):
1989
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
1993
class RepositoryFormatRegistry(registry.Registry):
1994
"""Registry of RepositoryFormats."""
1996
def get(self, format_string):
1997
r = registry.Registry.get(self, format_string)
2003
format_registry = RepositoryFormatRegistry()
2004
"""Registry of formats, indexed by their identifying format string.
2006
This can contain either format instances themselves, or classes/factories that
2007
can be called to obtain one.
2011
#####################################################################
2012
# Repository Formats
2014
class RepositoryFormat(object):
2015
"""A repository format.
2017
Formats provide three things:
2018
* An initialization routine to construct repository data on disk.
2019
* a format string which is used when the BzrDir supports versioned
2021
* an open routine which returns a Repository instance.
2023
There is one and only one Format subclass for each on-disk format. But
2024
there can be one Repository subclass that is used for several different
2025
formats. The _format attribute on a Repository instance can be used to
2026
determine the disk format.
2028
Formats are placed in an dict by their format string for reference
2029
during opening. These should be subclasses of RepositoryFormat
2032
Once a format is deprecated, just deprecate the initialize and open
2033
methods on the format class. Do not deprecate the object, as the
2034
object will be created every system load.
2036
Common instance attributes:
2037
_matchingbzrdir - the bzrdir format that the repository format was
2038
originally written to work with. This can be used if manually
2039
constructing a bzrdir and repository, or more commonly for test suite
2043
# Set to True or False in derived classes. True indicates that the format
2044
# supports ghosts gracefully.
2045
supports_ghosts = None
2046
# Can this repository be given external locations to lookup additional
2047
# data. Set to True or False in derived classes.
2048
supports_external_lookups = None
2051
return "<%s>" % self.__class__.__name__
2053
def __eq__(self, other):
2054
# format objects are generally stateless
2055
return isinstance(other, self.__class__)
2057
def __ne__(self, other):
2058
return not self == other
2061
def find_format(klass, a_bzrdir):
2062
"""Return the format for the repository object in a_bzrdir.
2064
This is used by bzr native formats that have a "format" file in
2065
the repository. Other methods may be used by different types of
2069
transport = a_bzrdir.get_repository_transport(None)
2070
format_string = transport.get("format").read()
2071
return format_registry.get(format_string)
2072
except errors.NoSuchFile:
2073
raise errors.NoRepositoryPresent(a_bzrdir)
2075
raise errors.UnknownFormatError(format=format_string,
2079
def register_format(klass, format):
2080
format_registry.register(format.get_format_string(), format)
2083
def unregister_format(klass, format):
2084
format_registry.remove(format.get_format_string())
2087
def get_default_format(klass):
2088
"""Return the current default format."""
2089
from bzrlib import bzrdir
2090
return bzrdir.format_registry.make_bzrdir('default').repository_format
2092
def get_format_string(self):
2093
"""Return the ASCII format string that identifies this format.
2095
Note that in pre format ?? repositories the format string is
2096
not permitted nor written to disk.
2098
raise NotImplementedError(self.get_format_string)
2100
def get_format_description(self):
2101
"""Return the short description for this format."""
2102
raise NotImplementedError(self.get_format_description)
2104
# TODO: this shouldn't be in the base class, it's specific to things that
2105
# use weaves or knits -- mbp 20070207
2106
def _get_versioned_file_store(self,
2111
versionedfile_class=None,
2112
versionedfile_kwargs={},
2114
if versionedfile_class is None:
2115
versionedfile_class = self._versionedfile_class
2116
weave_transport = control_files._transport.clone(name)
2117
dir_mode = control_files._dir_mode
2118
file_mode = control_files._file_mode
2119
return VersionedFileStore(weave_transport, prefixed=prefixed,
2121
file_mode=file_mode,
2122
versionedfile_class=versionedfile_class,
2123
versionedfile_kwargs=versionedfile_kwargs,
2126
def initialize(self, a_bzrdir, shared=False):
2127
"""Initialize a repository of this format in a_bzrdir.
2129
:param a_bzrdir: The bzrdir to put the new repository in it.
2130
:param shared: The repository should be initialized as a sharable one.
2131
:returns: The new repository object.
2133
This may raise UninitializableFormat if shared repository are not
2134
compatible the a_bzrdir.
2136
raise NotImplementedError(self.initialize)
2138
def is_supported(self):
2139
"""Is this format supported?
2141
Supported formats must be initializable and openable.
2142
Unsupported formats may not support initialization or committing or
2143
some other features depending on the reason for not being supported.
2147
def check_conversion_target(self, target_format):
2148
raise NotImplementedError(self.check_conversion_target)
2150
def open(self, a_bzrdir, _found=False):
2151
"""Return an instance of this format for the bzrdir a_bzrdir.
2153
_found is a private parameter, do not use it.
2155
raise NotImplementedError(self.open)
2158
class MetaDirRepositoryFormat(RepositoryFormat):
2159
"""Common base class for the new repositories using the metadir layout."""
2161
rich_root_data = False
2162
supports_tree_reference = False
2163
supports_external_lookups = False
2164
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2167
super(MetaDirRepositoryFormat, self).__init__()
2169
def _create_control_files(self, a_bzrdir):
2170
"""Create the required files and the initial control_files object."""
2171
# FIXME: RBC 20060125 don't peek under the covers
2172
# NB: no need to escape relative paths that are url safe.
2173
repository_transport = a_bzrdir.get_repository_transport(self)
2174
control_files = lockable_files.LockableFiles(repository_transport,
2175
'lock', lockdir.LockDir)
2176
control_files.create_lock()
2177
return control_files
2179
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2180
"""Upload the initial blank content."""
2181
control_files = self._create_control_files(a_bzrdir)
2182
control_files.lock_write()
2183
transport = control_files._transport
2185
utf8_files += [('shared-storage', '')]
2187
transport.mkdir_multi(dirs, mode=a_bzrdir._get_dir_mode())
2188
for (filename, content_stream) in files:
2189
transport.put_file(filename, content_stream,
2190
mode=a_bzrdir._get_file_mode())
2191
for (filename, content_bytes) in utf8_files:
2192
transport.put_bytes_non_atomic(filename, content_bytes,
2193
mode=a_bzrdir._get_file_mode())
2195
control_files.unlock()
2198
# formats which have no format string are not discoverable
2199
# and not independently creatable, so are not registered. They're
2200
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2201
# needed, it's constructed directly by the BzrDir. Non-native formats where
2202
# the repository is not separately opened are similar.
2204
format_registry.register_lazy(
2205
'Bazaar-NG Repository format 7',
2206
'bzrlib.repofmt.weaverepo',
2210
format_registry.register_lazy(
2211
'Bazaar-NG Knit Repository Format 1',
2212
'bzrlib.repofmt.knitrepo',
2213
'RepositoryFormatKnit1',
2216
format_registry.register_lazy(
2217
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2218
'bzrlib.repofmt.knitrepo',
2219
'RepositoryFormatKnit3',
2222
format_registry.register_lazy(
2223
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2224
'bzrlib.repofmt.knitrepo',
2225
'RepositoryFormatKnit4',
2228
# Pack-based formats. There is one format for pre-subtrees, and one for
2229
# post-subtrees to allow ease of testing.
2230
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2231
format_registry.register_lazy(
2232
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2233
'bzrlib.repofmt.pack_repo',
2234
'RepositoryFormatKnitPack1',
2236
format_registry.register_lazy(
2237
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2238
'bzrlib.repofmt.pack_repo',
2239
'RepositoryFormatKnitPack3',
2241
format_registry.register_lazy(
2242
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2243
'bzrlib.repofmt.pack_repo',
2244
'RepositoryFormatKnitPack4',
2246
# Development formats.
2248
# development 0 - stub to introduce development versioning scheme.
2249
format_registry.register_lazy(
2250
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2251
'bzrlib.repofmt.pack_repo',
2252
'RepositoryFormatPackDevelopment0',
2254
format_registry.register_lazy(
2255
("Bazaar development format 0 with subtree support "
2256
"(needs bzr.dev from before 1.3)\n"),
2257
'bzrlib.repofmt.pack_repo',
2258
'RepositoryFormatPackDevelopment0Subtree',
2260
# 1.3->1.4 go below here
2263
class InterRepository(InterObject):
2264
"""This class represents operations taking place between two repositories.
2266
Its instances have methods like copy_content and fetch, and contain
2267
references to the source and target repositories these operations can be
2270
Often we will provide convenience methods on 'repository' which carry out
2271
operations with another repository - they will always forward to
2272
InterRepository.get(other).method_name(parameters).
2276
"""The available optimised InterRepository types."""
2278
def copy_content(self, revision_id=None):
2279
raise NotImplementedError(self.copy_content)
2281
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2282
"""Fetch the content required to construct revision_id.
2284
The content is copied from self.source to self.target.
2286
:param revision_id: if None all content is copied, if NULL_REVISION no
2288
:param pb: optional progress bar to use for progress reports. If not
2289
provided a default one will be created.
2291
Returns the copied revision count and the failed revisions in a tuple:
2294
raise NotImplementedError(self.fetch)
2296
def _walk_to_common_revisions(self, revision_ids):
2297
"""Walk out from revision_ids in source to revisions target has.
2299
:param revision_ids: The start point for the search.
2300
:return: A set of revision ids.
2302
target_graph = self.target.get_graph()
2303
revision_ids = frozenset(revision_ids)
2304
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2305
return graph.SearchResult(revision_ids, set(), 0, set())
2306
missing_revs = set()
2307
source_graph = self.source.get_graph()
2308
# ensure we don't pay silly lookup costs.
2309
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2310
null_set = frozenset([_mod_revision.NULL_REVISION])
2313
next_revs, ghosts = searcher.next_with_ghosts()
2314
except StopIteration:
2316
if revision_ids.intersection(ghosts):
2317
absent_ids = set(revision_ids.intersection(ghosts))
2318
# If all absent_ids are present in target, no error is needed.
2319
absent_ids.difference_update(
2320
set(target_graph.get_parent_map(absent_ids)))
2322
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2323
# we don't care about other ghosts as we can't fetch them and
2324
# haven't been asked to.
2325
next_revs = set(next_revs)
2326
# we always have NULL_REVISION present.
2327
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2328
missing_revs.update(next_revs - have_revs)
2329
searcher.stop_searching_any(have_revs)
2330
return searcher.get_result()
2332
@deprecated_method(symbol_versioning.one_two)
2334
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2335
"""Return the revision ids that source has that target does not.
2337
These are returned in topological order.
2339
:param revision_id: only return revision ids included by this
2341
:param find_ghosts: If True find missing revisions in deep history
2342
rather than just finding the surface difference.
2344
return list(self.search_missing_revision_ids(
2345
revision_id, find_ghosts).get_keys())
2348
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2349
"""Return the revision ids that source has that target does not.
2351
:param revision_id: only return revision ids included by this
2353
:param find_ghosts: If True find missing revisions in deep history
2354
rather than just finding the surface difference.
2355
:return: A bzrlib.graph.SearchResult.
2357
# stop searching at found target revisions.
2358
if not find_ghosts and revision_id is not None:
2359
return self._walk_to_common_revisions([revision_id])
2360
# generic, possibly worst case, slow code path.
2361
target_ids = set(self.target.all_revision_ids())
2362
if revision_id is not None:
2363
source_ids = self.source.get_ancestry(revision_id)
2364
if source_ids[0] is not None:
2365
raise AssertionError()
2368
source_ids = self.source.all_revision_ids()
2369
result_set = set(source_ids).difference(target_ids)
2370
return self.source.revision_ids_to_search_result(result_set)
2373
def _same_model(source, target):
2374
"""True if source and target have the same data representation."""
2375
if source.supports_rich_root() != target.supports_rich_root():
2377
if source._serializer != target._serializer:
2382
class InterSameDataRepository(InterRepository):
2383
"""Code for converting between repositories that represent the same data.
2385
Data format and model must match for this to work.
2389
def _get_repo_format_to_test(self):
2390
"""Repository format for testing with.
2392
InterSameData can pull from subtree to subtree and from non-subtree to
2393
non-subtree, so we test this with the richest repository format.
2395
from bzrlib.repofmt import knitrepo
2396
return knitrepo.RepositoryFormatKnit3()
2399
def is_compatible(source, target):
2400
return InterRepository._same_model(source, target)
2403
def copy_content(self, revision_id=None):
2404
"""Make a complete copy of the content in self into destination.
2406
This copies both the repository's revision data, and configuration information
2407
such as the make_working_trees setting.
2409
This is a destructive operation! Do not use it on existing
2412
:param revision_id: Only copy the content needed to construct
2413
revision_id and its parents.
2416
self.target.set_make_working_trees(self.source.make_working_trees())
2417
except NotImplementedError:
2419
# but don't bother fetching if we have the needed data now.
2420
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2421
self.target.has_revision(revision_id)):
2423
self.target.fetch(self.source, revision_id=revision_id)
2426
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2427
"""See InterRepository.fetch()."""
2428
from bzrlib.fetch import GenericRepoFetcher
2429
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2430
self.source, self.source._format, self.target,
2431
self.target._format)
2432
f = GenericRepoFetcher(to_repository=self.target,
2433
from_repository=self.source,
2434
last_revision=revision_id,
2435
pb=pb, find_ghosts=find_ghosts)
2436
return f.count_copied, f.failed_revisions
2439
class InterWeaveRepo(InterSameDataRepository):
2440
"""Optimised code paths between Weave based repositories.
2442
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2443
implemented lazy inter-object optimisation.
2447
def _get_repo_format_to_test(self):
2448
from bzrlib.repofmt import weaverepo
2449
return weaverepo.RepositoryFormat7()
2452
def is_compatible(source, target):
2453
"""Be compatible with known Weave formats.
2455
We don't test for the stores being of specific types because that
2456
could lead to confusing results, and there is no need to be
2459
from bzrlib.repofmt.weaverepo import (
2465
return (isinstance(source._format, (RepositoryFormat5,
2467
RepositoryFormat7)) and
2468
isinstance(target._format, (RepositoryFormat5,
2470
RepositoryFormat7)))
2471
except AttributeError:
2475
def copy_content(self, revision_id=None):
2476
"""See InterRepository.copy_content()."""
2477
# weave specific optimised path:
2479
self.target.set_make_working_trees(self.source.make_working_trees())
2480
except (errors.RepositoryUpgradeRequired, NotImplemented):
2482
# FIXME do not peek!
2483
if self.source._transport.listable():
2484
pb = ui.ui_factory.nested_progress_bar()
2486
self.target.texts.insert_record_stream(
2487
self.source.texts.get_record_stream(
2488
self.source.texts.keys(), 'topological', False))
2489
pb.update('copying inventory', 0, 1)
2490
self.target.inventories.insert_record_stream(
2491
self.source.inventories.get_record_stream(
2492
self.source.inventories.keys(), 'topological', False))
2493
self.target.signatures.insert_record_stream(
2494
self.source.signatures.get_record_stream(
2495
self.source.signatures.keys(),
2497
self.target.revisions.insert_record_stream(
2498
self.source.revisions.get_record_stream(
2499
self.source.revisions.keys(),
2500
'topological', True))
2504
self.target.fetch(self.source, revision_id=revision_id)
2507
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2508
"""See InterRepository.fetch()."""
2509
from bzrlib.fetch import GenericRepoFetcher
2510
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2511
self.source, self.source._format, self.target, self.target._format)
2512
f = GenericRepoFetcher(to_repository=self.target,
2513
from_repository=self.source,
2514
last_revision=revision_id,
2515
pb=pb, find_ghosts=find_ghosts)
2516
return f.count_copied, f.failed_revisions
2519
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2520
"""See InterRepository.missing_revision_ids()."""
2521
# we want all revisions to satisfy revision_id in source.
2522
# but we don't want to stat every file here and there.
2523
# we want then, all revisions other needs to satisfy revision_id
2524
# checked, but not those that we have locally.
2525
# so the first thing is to get a subset of the revisions to
2526
# satisfy revision_id in source, and then eliminate those that
2527
# we do already have.
2528
# this is slow on high latency connection to self, but as as this
2529
# disk format scales terribly for push anyway due to rewriting
2530
# inventory.weave, this is considered acceptable.
2532
if revision_id is not None:
2533
source_ids = self.source.get_ancestry(revision_id)
2534
if source_ids[0] is not None:
2535
raise AssertionError()
2538
source_ids = self.source._all_possible_ids()
2539
source_ids_set = set(source_ids)
2540
# source_ids is the worst possible case we may need to pull.
2541
# now we want to filter source_ids against what we actually
2542
# have in target, but don't try to check for existence where we know
2543
# we do not have a revision as that would be pointless.
2544
target_ids = set(self.target._all_possible_ids())
2545
possibly_present_revisions = target_ids.intersection(source_ids_set)
2546
actually_present_revisions = set(
2547
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2548
required_revisions = source_ids_set.difference(actually_present_revisions)
2549
if revision_id is not None:
2550
# we used get_ancestry to determine source_ids then we are assured all
2551
# revisions referenced are present as they are installed in topological order.
2552
# and the tip revision was validated by get_ancestry.
2553
result_set = required_revisions
2555
# if we just grabbed the possibly available ids, then
2556
# we only have an estimate of whats available and need to validate
2557
# that against the revision records.
2559
self.source._eliminate_revisions_not_present(required_revisions))
2560
return self.source.revision_ids_to_search_result(result_set)
2563
class InterKnitRepo(InterSameDataRepository):
2564
"""Optimised code paths between Knit based repositories."""
2567
def _get_repo_format_to_test(self):
2568
from bzrlib.repofmt import knitrepo
2569
return knitrepo.RepositoryFormatKnit1()
2572
def is_compatible(source, target):
2573
"""Be compatible with known Knit formats.
2575
We don't test for the stores being of specific types because that
2576
could lead to confusing results, and there is no need to be
2579
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2581
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2582
isinstance(target._format, RepositoryFormatKnit))
2583
except AttributeError:
2585
return are_knits and InterRepository._same_model(source, target)
2588
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2589
"""See InterRepository.fetch()."""
2590
from bzrlib.fetch import KnitRepoFetcher
2591
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2592
self.source, self.source._format, self.target, self.target._format)
2593
f = KnitRepoFetcher(to_repository=self.target,
2594
from_repository=self.source,
2595
last_revision=revision_id,
2596
pb=pb, find_ghosts=find_ghosts)
2597
return f.count_copied, f.failed_revisions
2600
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2601
"""See InterRepository.missing_revision_ids()."""
2602
if revision_id is not None:
2603
source_ids = self.source.get_ancestry(revision_id)
2604
if source_ids[0] is not None:
2605
raise AssertionError()
2608
source_ids = self.source.all_revision_ids()
2609
source_ids_set = set(source_ids)
2610
# source_ids is the worst possible case we may need to pull.
2611
# now we want to filter source_ids against what we actually
2612
# have in target, but don't try to check for existence where we know
2613
# we do not have a revision as that would be pointless.
2614
target_ids = set(self.target.all_revision_ids())
2615
possibly_present_revisions = target_ids.intersection(source_ids_set)
2616
actually_present_revisions = set(
2617
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2618
required_revisions = source_ids_set.difference(actually_present_revisions)
2619
if revision_id is not None:
2620
# we used get_ancestry to determine source_ids then we are assured all
2621
# revisions referenced are present as they are installed in topological order.
2622
# and the tip revision was validated by get_ancestry.
2623
result_set = required_revisions
2625
# if we just grabbed the possibly available ids, then
2626
# we only have an estimate of whats available and need to validate
2627
# that against the revision records.
2629
self.source._eliminate_revisions_not_present(required_revisions))
2630
return self.source.revision_ids_to_search_result(result_set)
2633
class InterPackRepo(InterSameDataRepository):
2634
"""Optimised code paths between Pack based repositories."""
2637
def _get_repo_format_to_test(self):
2638
from bzrlib.repofmt import pack_repo
2639
return pack_repo.RepositoryFormatKnitPack1()
2642
def is_compatible(source, target):
2643
"""Be compatible with known Pack formats.
2645
We don't test for the stores being of specific types because that
2646
could lead to confusing results, and there is no need to be
2649
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2651
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2652
isinstance(target._format, RepositoryFormatPack))
2653
except AttributeError:
2655
return are_packs and InterRepository._same_model(source, target)
2658
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2659
"""See InterRepository.fetch()."""
2660
from bzrlib.repofmt.pack_repo import Packer
2661
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2662
self.source, self.source._format, self.target, self.target._format)
2663
self.count_copied = 0
2664
if revision_id is None:
2666
# everything to do - use pack logic
2667
# to fetch from all packs to one without
2668
# inventory parsing etc, IFF nothing to be copied is in the target.
2670
revision_ids = self.source.all_revision_ids()
2671
revision_keys = [(revid,) for revid in revision_ids]
2672
index = self.target._pack_collection.revision_index.combined_index
2673
present_revision_ids = set(item[1][0] for item in
2674
index.iter_entries(revision_keys))
2675
revision_ids = set(revision_ids) - present_revision_ids
2676
# implementing the TODO will involve:
2677
# - detecting when all of a pack is selected
2678
# - avoiding as much as possible pre-selection, so the
2679
# more-core routines such as create_pack_from_packs can filter in
2680
# a just-in-time fashion. (though having a HEADS list on a
2681
# repository might make this a lot easier, because we could
2682
# sensibly detect 'new revisions' without doing a full index scan.
2683
elif _mod_revision.is_null(revision_id):
2688
revision_ids = self.search_missing_revision_ids(revision_id,
2689
find_ghosts=find_ghosts).get_keys()
2690
except errors.NoSuchRevision:
2691
raise errors.InstallFailed([revision_id])
2692
if len(revision_ids) == 0:
2694
packs = self.source._pack_collection.all_packs()
2695
pack = Packer(self.target._pack_collection, packs, '.fetch',
2696
revision_ids).pack()
2697
if pack is not None:
2698
self.target._pack_collection._save_pack_names()
2699
# Trigger an autopack. This may duplicate effort as we've just done
2700
# a pack creation, but for now it is simpler to think about as
2701
# 'upload data, then repack if needed'.
2702
self.target._pack_collection.autopack()
2703
return (pack.get_revision_count(), [])
2708
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2709
"""See InterRepository.missing_revision_ids().
2711
:param find_ghosts: Find ghosts throughout the ancestry of
2714
if not find_ghosts and revision_id is not None:
2715
return self._walk_to_common_revisions([revision_id])
2716
elif revision_id is not None:
2717
source_ids = self.source.get_ancestry(revision_id)
2718
if source_ids[0] is not None:
2719
raise AssertionError()
2722
source_ids = self.source.all_revision_ids()
2723
# source_ids is the worst possible case we may need to pull.
2724
# now we want to filter source_ids against what we actually
2725
# have in target, but don't try to check for existence where we know
2726
# we do not have a revision as that would be pointless.
2727
target_ids = set(self.target.all_revision_ids())
2728
result_set = set(source_ids).difference(target_ids)
2729
return self.source.revision_ids_to_search_result(result_set)
2732
class InterModel1and2(InterRepository):
2735
def _get_repo_format_to_test(self):
2739
def is_compatible(source, target):
2740
if not source.supports_rich_root() and target.supports_rich_root():
2746
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2747
"""See InterRepository.fetch()."""
2748
from bzrlib.fetch import Model1toKnit2Fetcher
2749
f = Model1toKnit2Fetcher(to_repository=self.target,
2750
from_repository=self.source,
2751
last_revision=revision_id,
2752
pb=pb, find_ghosts=find_ghosts)
2753
return f.count_copied, f.failed_revisions
2756
def copy_content(self, revision_id=None):
2757
"""Make a complete copy of the content in self into destination.
2759
This is a destructive operation! Do not use it on existing
2762
:param revision_id: Only copy the content needed to construct
2763
revision_id and its parents.
2766
self.target.set_make_working_trees(self.source.make_working_trees())
2767
except NotImplementedError:
2769
# but don't bother fetching if we have the needed data now.
2770
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2771
self.target.has_revision(revision_id)):
2773
self.target.fetch(self.source, revision_id=revision_id)
2776
class InterKnit1and2(InterKnitRepo):
2779
def _get_repo_format_to_test(self):
2783
def is_compatible(source, target):
2784
"""Be compatible with Knit1 source and Knit3 target"""
2785
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2787
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2788
RepositoryFormatKnit3)
2789
from bzrlib.repofmt.pack_repo import (
2790
RepositoryFormatKnitPack1,
2791
RepositoryFormatKnitPack3,
2792
RepositoryFormatPackDevelopment0,
2793
RepositoryFormatPackDevelopment0Subtree,
2796
RepositoryFormatKnit1,
2797
RepositoryFormatKnitPack1,
2798
RepositoryFormatPackDevelopment0,
2801
RepositoryFormatKnit3,
2802
RepositoryFormatKnitPack3,
2803
RepositoryFormatPackDevelopment0Subtree,
2805
return (isinstance(source._format, nosubtrees) and
2806
isinstance(target._format, subtrees))
2807
except AttributeError:
2811
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2812
"""See InterRepository.fetch()."""
2813
from bzrlib.fetch import Knit1to2Fetcher
2814
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2815
self.source, self.source._format, self.target,
2816
self.target._format)
2817
f = Knit1to2Fetcher(to_repository=self.target,
2818
from_repository=self.source,
2819
last_revision=revision_id,
2820
pb=pb, find_ghosts=find_ghosts)
2821
return f.count_copied, f.failed_revisions
2824
class InterDifferingSerializer(InterKnitRepo):
2827
def _get_repo_format_to_test(self):
2831
def is_compatible(source, target):
2832
"""Be compatible with Knit2 source and Knit3 target"""
2833
if source.supports_rich_root() != target.supports_rich_root():
2835
# Ideally, we'd support fetching if the source had no tree references
2836
# even if it supported them...
2837
if (getattr(source, '_format.supports_tree_reference', False) and
2838
not getattr(target, '_format.supports_tree_reference', False)):
2843
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2844
"""See InterRepository.fetch()."""
2845
revision_ids = self.target.search_missing_revision_ids(self.source,
2846
revision_id, find_ghosts=find_ghosts).get_keys()
2847
revision_ids = tsort.topo_sort(
2848
self.source.get_graph().get_parent_map(revision_ids))
2849
def revisions_iterator():
2850
for current_revision_id in revision_ids:
2851
revision = self.source.get_revision(current_revision_id)
2852
tree = self.source.revision_tree(current_revision_id)
2854
signature = self.source.get_signature_text(
2855
current_revision_id)
2856
except errors.NoSuchRevision:
2858
yield revision, tree, signature
2860
my_pb = ui.ui_factory.nested_progress_bar()
2865
install_revisions(self.target, revisions_iterator(),
2866
len(revision_ids), pb)
2868
if my_pb is not None:
2870
return len(revision_ids), 0
2873
class InterOtherToRemote(InterRepository):
2875
def __init__(self, source, target):
2876
InterRepository.__init__(self, source, target)
2877
self._real_inter = None
2880
def is_compatible(source, target):
2881
if isinstance(target, remote.RemoteRepository):
2885
def _ensure_real_inter(self):
2886
if self._real_inter is None:
2887
self.target._ensure_real()
2888
real_target = self.target._real_repository
2889
self._real_inter = InterRepository.get(self.source, real_target)
2891
def copy_content(self, revision_id=None):
2892
self._ensure_real_inter()
2893
self._real_inter.copy_content(revision_id=revision_id)
2895
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2896
self._ensure_real_inter()
2897
self._real_inter.fetch(revision_id=revision_id, pb=pb,
2898
find_ghosts=find_ghosts)
2901
def _get_repo_format_to_test(self):
2905
InterRepository.register_optimiser(InterDifferingSerializer)
2906
InterRepository.register_optimiser(InterSameDataRepository)
2907
InterRepository.register_optimiser(InterWeaveRepo)
2908
InterRepository.register_optimiser(InterKnitRepo)
2909
InterRepository.register_optimiser(InterModel1and2)
2910
InterRepository.register_optimiser(InterKnit1and2)
2911
InterRepository.register_optimiser(InterPackRepo)
2912
InterRepository.register_optimiser(InterOtherToRemote)
2915
class CopyConverter(object):
2916
"""A repository conversion tool which just performs a copy of the content.
2918
This is slow but quite reliable.
2921
def __init__(self, target_format):
2922
"""Create a CopyConverter.
2924
:param target_format: The format the resulting repository should be.
2926
self.target_format = target_format
2928
def convert(self, repo, pb):
2929
"""Perform the conversion of to_convert, giving feedback via pb.
2931
:param to_convert: The disk object to convert.
2932
:param pb: a progress bar to use for progress information.
2937
# this is only useful with metadir layouts - separated repo content.
2938
# trigger an assertion if not such
2939
repo._format.get_format_string()
2940
self.repo_dir = repo.bzrdir
2941
self.step('Moving repository to repository.backup')
2942
self.repo_dir.transport.move('repository', 'repository.backup')
2943
backup_transport = self.repo_dir.transport.clone('repository.backup')
2944
repo._format.check_conversion_target(self.target_format)
2945
self.source_repo = repo._format.open(self.repo_dir,
2947
_override_transport=backup_transport)
2948
self.step('Creating new repository')
2949
converted = self.target_format.initialize(self.repo_dir,
2950
self.source_repo.is_shared())
2951
converted.lock_write()
2953
self.step('Copying content into repository.')
2954
self.source_repo.copy_content_into(converted)
2957
self.step('Deleting old repository content.')
2958
self.repo_dir.transport.delete_tree('repository.backup')
2959
self.pb.note('repository converted')
2961
def step(self, message):
2962
"""Update the pb by a step."""
2964
self.pb.update(message, self.count, self.total)
2976
def _unescaper(match, _map=_unescape_map):
2977
code = match.group(1)
2981
if not code.startswith('#'):
2983
return unichr(int(code[1:])).encode('utf8')
2989
def _unescape_xml(data):
2990
"""Unescape predefined XML entities in a string of data."""
2992
if _unescape_re is None:
2993
_unescape_re = re.compile('\&([^;]*);')
2994
return _unescape_re.sub(_unescaper, data)
2997
class _VersionedFileChecker(object):
2999
def __init__(self, repository):
3000
self.repository = repository
3001
self.text_index = self.repository._generate_text_key_index()
3003
def calculate_file_version_parents(self, text_key):
3004
"""Calculate the correct parents for a file version according to
3007
parent_keys = self.text_index[text_key]
3008
if parent_keys == [_mod_revision.NULL_REVISION]:
3010
return tuple(parent_keys)
3012
def check_file_version_parents(self, texts, progress_bar=None):
3013
"""Check the parents stored in a versioned file are correct.
3015
It also detects file versions that are not referenced by their
3016
corresponding revision's inventory.
3018
:returns: A tuple of (wrong_parents, dangling_file_versions).
3019
wrong_parents is a dict mapping {revision_id: (stored_parents,
3020
correct_parents)} for each revision_id where the stored parents
3021
are not correct. dangling_file_versions is a set of (file_id,
3022
revision_id) tuples for versions that are present in this versioned
3023
file, but not used by the corresponding inventory.
3026
self.file_ids = set([file_id for file_id, _ in
3027
self.text_index.iterkeys()])
3028
# text keys is now grouped by file_id
3029
n_weaves = len(self.file_ids)
3030
files_in_revisions = {}
3031
revisions_of_files = {}
3032
n_versions = len(self.text_index)
3033
progress_bar.update('loading text store', 0, n_versions)
3034
parent_map = self.repository.texts.get_parent_map(self.text_index)
3035
# On unlistable transports this could well be empty/error...
3036
text_keys = self.repository.texts.keys()
3037
unused_keys = frozenset(text_keys) - set(self.text_index)
3038
for num, key in enumerate(self.text_index.iterkeys()):
3039
if progress_bar is not None:
3040
progress_bar.update('checking text graph', num, n_versions)
3041
correct_parents = self.calculate_file_version_parents(key)
3043
knit_parents = parent_map[key]
3044
except errors.RevisionNotPresent:
3047
if correct_parents != knit_parents:
3048
wrong_parents[key] = (knit_parents, correct_parents)
3049
return wrong_parents, unused_keys
3052
def _old_get_graph(repository, revision_id):
3053
"""DO NOT USE. That is all. I'm serious."""
3054
graph = repository.get_graph()
3055
revision_graph = dict(((key, value) for key, value in
3056
graph.iter_ancestry([revision_id]) if value is not None))
3057
return _strip_NULL_ghosts(revision_graph)
3060
def _strip_NULL_ghosts(revision_graph):
3061
"""Also don't use this. more compatibility code for unmigrated clients."""
3062
# Filter ghosts, and null:
3063
if _mod_revision.NULL_REVISION in revision_graph:
3064
del revision_graph[_mod_revision.NULL_REVISION]
3065
for key, parents in revision_graph.items():
3066
revision_graph[key] = tuple(parent for parent in parents if parent
3068
return revision_graph
added
removed
bzrlib/repository.py
