1
# Copyright (C) 2005, 2006 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 binascii import hexlify
18
from copy import deepcopy
19
from cStringIO import StringIO
22
from unittest import TestSuite
24
import bzrlib.bzrdir as bzrdir
25
from bzrlib.decorators import needs_read_lock, needs_write_lock
26
import bzrlib.errors as errors
27
from bzrlib.errors import InvalidRevisionId
28
import bzrlib.gpg as gpg
29
from bzrlib.graph import Graph
30
from bzrlib.inter import InterObject
31
from bzrlib.inventory import Inventory
32
from bzrlib.knit import KnitVersionedFile, KnitPlainFactory
33
from bzrlib.lockable_files import LockableFiles, TransportLock
34
from bzrlib.lockdir import LockDir
35
from bzrlib.osutils import (safe_unicode, rand_bytes, compact_date,
37
from bzrlib.revision import NULL_REVISION, Revision
38
from bzrlib.store.versioned import VersionedFileStore, WeaveStore
39
from bzrlib.store.text import TextStore
40
from bzrlib.symbol_versioning import *
41
from bzrlib.trace import mutter, note
42
from bzrlib.tree import RevisionTree
43
from bzrlib.tsort import topo_sort
44
from bzrlib.testament import Testament
45
from bzrlib.tree import EmptyTree
47
from bzrlib.weave import WeaveFile
51
class Repository(object):
52
"""Repository holding history for one or more branches.
54
The repository holds and retrieves historical information including
55
revisions and file history. It's normally accessed only by the Branch,
56
which views a particular line of development through that history.
58
The Repository builds on top of Stores and a Transport, which respectively
59
describe the disk data format and the way of accessing the (possibly
64
def add_inventory(self, revid, inv, parents):
65
"""Add the inventory inv to the repository as revid.
67
:param parents: The revision ids of the parents that revid
68
is known to have and are in the repository already.
70
returns the sha1 of the serialized inventory.
72
assert inv.revision_id is None or inv.revision_id == revid, \
73
"Mismatch between inventory revision" \
74
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
75
inv_text = bzrlib.xml5.serializer_v5.write_inventory_to_string(inv)
76
inv_sha1 = bzrlib.osutils.sha_string(inv_text)
77
inv_vf = self.control_weaves.get_weave('inventory',
78
self.get_transaction())
79
self._inventory_add_lines(inv_vf, revid, parents, bzrlib.osutils.split_lines(inv_text))
82
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
84
for parent in parents:
86
final_parents.append(parent)
88
inv_vf.add_lines(revid, final_parents, lines)
91
def add_revision(self, rev_id, rev, inv=None, config=None):
92
"""Add rev to the revision store as rev_id.
94
:param rev_id: the revision id to use.
95
:param rev: The revision object.
96
:param inv: The inventory for the revision. if None, it will be looked
97
up in the inventory storer
98
:param config: If None no digital signature will be created.
99
If supplied its signature_needed method will be used
100
to determine if a signature should be made.
102
if config is not None and config.signature_needed():
104
inv = self.get_inventory(rev_id)
105
plaintext = Testament(rev, inv).as_short_text()
106
self.store_revision_signature(
107
gpg.GPGStrategy(config), plaintext, rev_id)
108
if not rev_id in self.get_inventory_weave():
110
raise errors.WeaveRevisionNotPresent(rev_id,
111
self.get_inventory_weave())
113
# yes, this is not suitable for adding with ghosts.
114
self.add_inventory(rev_id, inv, rev.parent_ids)
115
self._revision_store.add_revision(rev, self.get_transaction())
118
def _all_possible_ids(self):
119
"""Return all the possible revisions that we could find."""
120
return self.get_inventory_weave().versions()
122
@deprecated_method(zero_nine)
123
def all_revision_ids(self):
124
"""Returns a list of all the revision ids in the repository.
126
This is deprecated because code should generally work on the graph
127
reachable from a particular revision, and ignore any other revisions
128
that might be present. There is no direct replacement method.
130
return self._all_revision_ids()
133
def _all_revision_ids(self):
134
"""Returns a list of all the revision ids in the repository.
136
These are in as much topological order as the underlying store can
137
present: for weaves ghosts may lead to a lack of correctness until
138
the reweave updates the parents list.
140
if self._revision_store.text_store.listable():
141
return self._revision_store.all_revision_ids(self.get_transaction())
142
result = self._all_possible_ids()
143
return self._eliminate_revisions_not_present(result)
145
def break_lock(self):
146
"""Break a lock if one is present from another instance.
148
Uses the ui factory to ask for confirmation if the lock may be from
151
self.control_files.break_lock()
154
def _eliminate_revisions_not_present(self, revision_ids):
155
"""Check every revision id in revision_ids to see if we have it.
157
Returns a set of the present revisions.
160
for id in revision_ids:
161
if self.has_revision(id):
166
def create(a_bzrdir):
167
"""Construct the current default format repository in a_bzrdir."""
168
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
170
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
171
"""instantiate a Repository.
173
:param _format: The format of the repository on disk.
174
:param a_bzrdir: The BzrDir of the repository.
176
In the future we will have a single api for all stores for
177
getting file texts, inventories and revisions, then
178
this construct will accept instances of those things.
180
super(Repository, self).__init__()
181
self._format = _format
182
# the following are part of the public API for Repository:
183
self.bzrdir = a_bzrdir
184
self.control_files = control_files
185
self._revision_store = _revision_store
186
self.text_store = text_store
187
# backwards compatibility
188
self.weave_store = text_store
189
# not right yet - should be more semantically clear ?
191
self.control_store = control_store
192
self.control_weaves = control_store
193
# TODO: make sure to construct the right store classes, etc, depending
194
# on whether escaping is required.
197
return '%s(%r)' % (self.__class__.__name__,
198
self.bzrdir.transport.base)
201
return self.control_files.is_locked()
203
def lock_write(self):
204
self.control_files.lock_write()
207
self.control_files.lock_read()
209
def get_physical_lock_status(self):
210
return self.control_files.get_physical_lock_status()
213
def missing_revision_ids(self, other, revision_id=None):
214
"""Return the revision ids that other has that this does not.
216
These are returned in topological order.
218
revision_id: only return revision ids included by revision_id.
220
return InterRepository.get(other, self).missing_revision_ids(revision_id)
224
"""Open the repository rooted at base.
226
For instance, if the repository is at URL/.bzr/repository,
227
Repository.open(URL) -> a Repository instance.
229
control = bzrlib.bzrdir.BzrDir.open(base)
230
return control.open_repository()
232
def copy_content_into(self, destination, revision_id=None, basis=None):
233
"""Make a complete copy of the content in self into destination.
235
This is a destructive operation! Do not use it on existing
238
return InterRepository.get(self, destination).copy_content(revision_id, basis)
240
def fetch(self, source, revision_id=None, pb=None):
241
"""Fetch the content required to construct revision_id from source.
243
If revision_id is None all content is copied.
245
return InterRepository.get(source, self).fetch(revision_id=revision_id,
248
def get_commit_builder(self, branch, parents, config, timestamp=None,
249
timezone=None, committer=None, revprops=None,
251
"""Obtain a CommitBuilder for this repository.
253
:param branch: Branch to commit to.
254
:param parents: Revision ids of the parents of the new revision.
255
:param config: Configuration to use.
256
:param timestamp: Optional timestamp recorded for commit.
257
:param timezone: Optional timezone for timestamp.
258
:param committer: Optional committer to set for commit.
259
:param revprops: Optional dictionary of revision properties.
260
:param revision_id: Optional revision id.
262
return CommitBuilder(self, parents, config, timestamp, timezone,
263
committer, revprops, revision_id)
266
self.control_files.unlock()
269
def clone(self, a_bzrdir, revision_id=None, basis=None):
270
"""Clone this repository into a_bzrdir using the current format.
272
Currently no check is made that the format of this repository and
273
the bzrdir format are compatible. FIXME RBC 20060201.
275
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
276
# use target default format.
277
result = a_bzrdir.create_repository()
278
# FIXME RBC 20060209 split out the repository type to avoid this check ?
279
elif isinstance(a_bzrdir._format,
280
(bzrlib.bzrdir.BzrDirFormat4,
281
bzrlib.bzrdir.BzrDirFormat5,
282
bzrlib.bzrdir.BzrDirFormat6)):
283
result = a_bzrdir.open_repository()
285
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
286
self.copy_content_into(result, revision_id, basis)
290
def has_revision(self, revision_id):
291
"""True if this repository has a copy of the revision."""
292
return self._revision_store.has_revision_id(revision_id,
293
self.get_transaction())
296
def get_revision_reconcile(self, revision_id):
297
"""'reconcile' helper routine that allows access to a revision always.
299
This variant of get_revision does not cross check the weave graph
300
against the revision one as get_revision does: but it should only
301
be used by reconcile, or reconcile-alike commands that are correcting
302
or testing the revision graph.
304
if not revision_id or not isinstance(revision_id, basestring):
305
raise InvalidRevisionId(revision_id=revision_id, branch=self)
306
return self._revision_store.get_revision(revision_id,
307
self.get_transaction())
310
def get_revision_xml(self, revision_id):
311
rev = self.get_revision(revision_id)
313
# the current serializer..
314
self._revision_store._serializer.write_revision(rev, rev_tmp)
316
return rev_tmp.getvalue()
319
def get_revision(self, revision_id):
320
"""Return the Revision object for a named revision"""
321
r = self.get_revision_reconcile(revision_id)
322
# weave corruption can lead to absent revision markers that should be
324
# the following test is reasonably cheap (it needs a single weave read)
325
# and the weave is cached in read transactions. In write transactions
326
# it is not cached but typically we only read a small number of
327
# revisions. For knits when they are introduced we will probably want
328
# to ensure that caching write transactions are in use.
329
inv = self.get_inventory_weave()
330
self._check_revision_parents(r, inv)
333
def _check_revision_parents(self, revision, inventory):
334
"""Private to Repository and Fetch.
336
This checks the parentage of revision in an inventory weave for
337
consistency and is only applicable to inventory-weave-for-ancestry
338
using repository formats & fetchers.
340
weave_parents = inventory.get_parents(revision.revision_id)
341
weave_names = inventory.versions()
342
for parent_id in revision.parent_ids:
343
if parent_id in weave_names:
344
# this parent must not be a ghost.
345
if not parent_id in weave_parents:
347
raise errors.CorruptRepository(self)
350
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
351
signature = gpg_strategy.sign(plaintext)
352
self._revision_store.add_revision_signature_text(revision_id,
354
self.get_transaction())
356
def fileids_altered_by_revision_ids(self, revision_ids):
357
"""Find the file ids and versions affected by revisions.
359
:param revisions: an iterable containing revision ids.
360
:return: a dictionary mapping altered file-ids to an iterable of
361
revision_ids. Each altered file-ids has the exact revision_ids that
362
altered it listed explicitly.
364
assert isinstance(self._format, (RepositoryFormat5,
367
RepositoryFormatKnit1)), \
368
("fileids_altered_by_revision_ids only supported for branches "
369
"which store inventory as unnested xml, not on %r" % self)
370
selected_revision_ids = set(revision_ids)
371
w = self.get_inventory_weave()
374
# this code needs to read every new line in every inventory for the
375
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
376
# not present in one of those inventories is unnecessary but not
377
# harmful because we are filtering by the revision id marker in the
378
# inventory lines : we only select file ids altered in one of those
379
# revisions. We don't need to see all lines in the inventory because
380
# only those added in an inventory in rev X can contain a revision=X
382
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
383
start = line.find('file_id="')+9
384
if start < 9: continue
385
end = line.find('"', start)
387
file_id = _unescape_xml(line[start:end])
389
start = line.find('revision="')+10
390
if start < 10: continue
391
end = line.find('"', start)
393
revision_id = _unescape_xml(line[start:end])
394
if revision_id in selected_revision_ids:
395
result.setdefault(file_id, set()).add(revision_id)
399
def get_inventory_weave(self):
400
return self.control_weaves.get_weave('inventory',
401
self.get_transaction())
404
def get_inventory(self, revision_id):
405
"""Get Inventory object by hash."""
406
return self.deserialise_inventory(
407
revision_id, self.get_inventory_xml(revision_id))
409
def deserialise_inventory(self, revision_id, xml):
410
"""Transform the xml into an inventory object.
412
:param revision_id: The expected revision id of the inventory.
413
:param xml: A serialised inventory.
415
return bzrlib.xml5.serializer_v5.read_inventory_from_string(xml)
418
def get_inventory_xml(self, revision_id):
419
"""Get inventory XML as a file object."""
421
assert isinstance(revision_id, basestring), type(revision_id)
422
iw = self.get_inventory_weave()
423
return iw.get_text(revision_id)
425
raise bzrlib.errors.HistoryMissing(self, 'inventory', revision_id)
428
def get_inventory_sha1(self, revision_id):
429
"""Return the sha1 hash of the inventory entry
431
return self.get_revision(revision_id).inventory_sha1
434
def get_revision_graph(self, revision_id=None):
435
"""Return a dictionary containing the revision graph.
437
:return: a dictionary of revision_id->revision_parents_list.
439
weave = self.get_inventory_weave()
440
all_revisions = self._eliminate_revisions_not_present(weave.versions())
441
entire_graph = dict([(node, weave.get_parents(node)) for
442
node in all_revisions])
443
if revision_id is None:
445
elif revision_id not in entire_graph:
446
raise errors.NoSuchRevision(self, revision_id)
448
# add what can be reached from revision_id
450
pending = set([revision_id])
451
while len(pending) > 0:
453
result[node] = entire_graph[node]
454
for revision_id in result[node]:
455
if revision_id not in result:
456
pending.add(revision_id)
460
def get_revision_graph_with_ghosts(self, revision_ids=None):
461
"""Return a graph of the revisions with ghosts marked as applicable.
463
:param revision_ids: an iterable of revisions to graph or None for all.
464
:return: a Graph object with the graph reachable from revision_ids.
468
pending = set(self.all_revision_ids())
471
pending = set(revision_ids)
472
required = set(revision_ids)
475
revision_id = pending.pop()
477
rev = self.get_revision(revision_id)
478
except errors.NoSuchRevision:
479
if revision_id in required:
482
result.add_ghost(revision_id)
484
for parent_id in rev.parent_ids:
485
# is this queued or done ?
486
if (parent_id not in pending and
487
parent_id not in done):
489
pending.add(parent_id)
490
result.add_node(revision_id, rev.parent_ids)
491
done.add(revision_id)
495
def get_revision_inventory(self, revision_id):
496
"""Return inventory of a past revision."""
497
# TODO: Unify this with get_inventory()
498
# bzr 0.0.6 and later imposes the constraint that the inventory_id
499
# must be the same as its revision, so this is trivial.
500
if revision_id is None:
501
# This does not make sense: if there is no revision,
502
# then it is the current tree inventory surely ?!
503
# and thus get_root_id() is something that looks at the last
504
# commit on the branch, and the get_root_id is an inventory check.
505
raise NotImplementedError
506
# return Inventory(self.get_root_id())
508
return self.get_inventory(revision_id)
512
"""Return True if this repository is flagged as a shared repository."""
513
raise NotImplementedError(self.is_shared)
516
def reconcile(self, other=None, thorough=False):
517
"""Reconcile this repository."""
518
from bzrlib.reconcile import RepoReconciler
519
reconciler = RepoReconciler(self, thorough=thorough)
520
reconciler.reconcile()
524
def revision_tree(self, revision_id):
525
"""Return Tree for a revision on this branch.
527
`revision_id` may be None for the null revision, in which case
528
an `EmptyTree` is returned."""
529
# TODO: refactor this to use an existing revision object
530
# so we don't need to read it in twice.
531
if revision_id is None or revision_id == NULL_REVISION:
534
inv = self.get_revision_inventory(revision_id)
535
return RevisionTree(self, inv, revision_id)
538
def get_ancestry(self, revision_id):
539
"""Return a list of revision-ids integrated by a revision.
541
The first element of the list is always None, indicating the origin
542
revision. This might change when we have history horizons, or
543
perhaps we should have a new API.
545
This is topologically sorted.
547
if revision_id is None:
549
if not self.has_revision(revision_id):
550
raise errors.NoSuchRevision(self, revision_id)
551
w = self.get_inventory_weave()
552
candidates = w.get_ancestry(revision_id)
553
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
556
def print_file(self, file, revision_id):
557
"""Print `file` to stdout.
559
FIXME RBC 20060125 as John Meinel points out this is a bad api
560
- it writes to stdout, it assumes that that is valid etc. Fix
561
by creating a new more flexible convenience function.
563
tree = self.revision_tree(revision_id)
564
# use inventory as it was in that revision
565
file_id = tree.inventory.path2id(file)
567
# TODO: jam 20060427 Write a test for this code path
568
# it had a bug in it, and was raising the wrong
570
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
571
tree.print_file(file_id)
573
def get_transaction(self):
574
return self.control_files.get_transaction()
576
def revision_parents(self, revid):
577
return self.get_inventory_weave().parent_names(revid)
580
def set_make_working_trees(self, new_value):
581
"""Set the policy flag for making working trees when creating branches.
583
This only applies to branches that use this repository.
585
The default is 'True'.
586
:param new_value: True to restore the default, False to disable making
589
raise NotImplementedError(self.set_make_working_trees)
591
def make_working_trees(self):
592
"""Returns the policy for making working trees on new branches."""
593
raise NotImplementedError(self.make_working_trees)
596
def sign_revision(self, revision_id, gpg_strategy):
597
plaintext = Testament.from_revision(self, revision_id).as_short_text()
598
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
601
def has_signature_for_revision_id(self, revision_id):
602
"""Query for a revision signature for revision_id in the repository."""
603
return self._revision_store.has_signature(revision_id,
604
self.get_transaction())
607
def get_signature_text(self, revision_id):
608
"""Return the text for a signature."""
609
return self._revision_store.get_signature_text(revision_id,
610
self.get_transaction())
613
def check(self, revision_ids):
614
"""Check consistency of all history of given revision_ids.
616
Different repository implementations should override _check().
618
:param revision_ids: A non-empty list of revision_ids whose ancestry
619
will be checked. Typically the last revision_id of a branch.
622
raise ValueError("revision_ids must be non-empty in %s.check"
624
return self._check(revision_ids)
626
def _check(self, revision_ids):
627
result = bzrlib.check.Check(self)
632
class AllInOneRepository(Repository):
633
"""Legacy support - the repository behaviour for all-in-one branches."""
635
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
636
# we reuse one control files instance.
637
dir_mode = a_bzrdir._control_files._dir_mode
638
file_mode = a_bzrdir._control_files._file_mode
640
def get_store(name, compressed=True, prefixed=False):
641
# FIXME: This approach of assuming stores are all entirely compressed
642
# or entirely uncompressed is tidy, but breaks upgrade from
643
# some existing branches where there's a mixture; we probably
644
# still want the option to look for both.
645
relpath = a_bzrdir._control_files._escape(name)
646
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
647
prefixed=prefixed, compressed=compressed,
650
#if self._transport.should_cache():
651
# cache_path = os.path.join(self.cache_root, name)
652
# os.mkdir(cache_path)
653
# store = bzrlib.store.CachedStore(store, cache_path)
656
# not broken out yet because the controlweaves|inventory_store
657
# and text_store | weave_store bits are still different.
658
if isinstance(_format, RepositoryFormat4):
659
# cannot remove these - there is still no consistent api
660
# which allows access to this old info.
661
self.inventory_store = get_store('inventory-store')
662
text_store = get_store('text-store')
663
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
667
"""AllInOne repositories cannot be shared."""
671
def set_make_working_trees(self, new_value):
672
"""Set the policy flag for making working trees when creating branches.
674
This only applies to branches that use this repository.
676
The default is 'True'.
677
:param new_value: True to restore the default, False to disable making
680
raise NotImplementedError(self.set_make_working_trees)
682
def make_working_trees(self):
683
"""Returns the policy for making working trees on new branches."""
687
def install_revision(repository, rev, revision_tree):
688
"""Install all revision data into a repository."""
691
for p_id in rev.parent_ids:
692
if repository.has_revision(p_id):
693
present_parents.append(p_id)
694
parent_trees[p_id] = repository.revision_tree(p_id)
696
parent_trees[p_id] = EmptyTree()
698
inv = revision_tree.inventory
700
# Add the texts that are not already present
701
for path, ie in inv.iter_entries():
702
w = repository.weave_store.get_weave_or_empty(ie.file_id,
703
repository.get_transaction())
704
if ie.revision not in w:
706
# FIXME: TODO: The following loop *may* be overlapping/duplicate
707
# with InventoryEntry.find_previous_heads(). if it is, then there
708
# is a latent bug here where the parents may have ancestors of each
710
for revision, tree in parent_trees.iteritems():
711
if ie.file_id not in tree:
713
parent_id = tree.inventory[ie.file_id].revision
714
if parent_id in text_parents:
716
text_parents.append(parent_id)
718
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
719
repository.get_transaction())
720
lines = revision_tree.get_file(ie.file_id).readlines()
721
vfile.add_lines(rev.revision_id, text_parents, lines)
723
# install the inventory
724
repository.add_inventory(rev.revision_id, inv, present_parents)
725
except errors.RevisionAlreadyPresent:
727
repository.add_revision(rev.revision_id, rev, inv)
730
class MetaDirRepository(Repository):
731
"""Repositories in the new meta-dir layout."""
733
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
734
super(MetaDirRepository, self).__init__(_format,
741
dir_mode = self.control_files._dir_mode
742
file_mode = self.control_files._file_mode
746
"""Return True if this repository is flagged as a shared repository."""
747
return self.control_files._transport.has('shared-storage')
750
def set_make_working_trees(self, new_value):
751
"""Set the policy flag for making working trees when creating branches.
753
This only applies to branches that use this repository.
755
The default is 'True'.
756
:param new_value: True to restore the default, False to disable making
761
self.control_files._transport.delete('no-working-trees')
762
except errors.NoSuchFile:
765
self.control_files.put_utf8('no-working-trees', '')
767
def make_working_trees(self):
768
"""Returns the policy for making working trees on new branches."""
769
return not self.control_files._transport.has('no-working-trees')
772
class KnitRepository(MetaDirRepository):
773
"""Knit format repository."""
775
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
776
inv_vf.add_lines_with_ghosts(revid, parents, lines)
779
def _all_revision_ids(self):
780
"""See Repository.all_revision_ids()."""
781
# Knits get the revision graph from the index of the revision knit, so
782
# it's always possible even if they're on an unlistable transport.
783
return self._revision_store.all_revision_ids(self.get_transaction())
785
def fileid_involved_between_revs(self, from_revid, to_revid):
786
"""Find file_id(s) which are involved in the changes between revisions.
788
This determines the set of revisions which are involved, and then
789
finds all file ids affected by those revisions.
791
vf = self._get_revision_vf()
792
from_set = set(vf.get_ancestry(from_revid))
793
to_set = set(vf.get_ancestry(to_revid))
794
changed = to_set.difference(from_set)
795
return self._fileid_involved_by_set(changed)
797
def fileid_involved(self, last_revid=None):
798
"""Find all file_ids modified in the ancestry of last_revid.
800
:param last_revid: If None, last_revision() will be used.
803
changed = set(self.all_revision_ids())
805
changed = set(self.get_ancestry(last_revid))
808
return self._fileid_involved_by_set(changed)
811
def get_ancestry(self, revision_id):
812
"""Return a list of revision-ids integrated by a revision.
814
This is topologically sorted.
816
if revision_id is None:
818
vf = self._get_revision_vf()
820
return [None] + vf.get_ancestry(revision_id)
821
except errors.RevisionNotPresent:
822
raise errors.NoSuchRevision(self, revision_id)
825
def get_revision(self, revision_id):
826
"""Return the Revision object for a named revision"""
827
return self.get_revision_reconcile(revision_id)
830
def get_revision_graph(self, revision_id=None):
831
"""Return a dictionary containing the revision graph.
833
:return: a dictionary of revision_id->revision_parents_list.
835
weave = self._get_revision_vf()
836
entire_graph = weave.get_graph()
837
if revision_id is None:
838
return weave.get_graph()
839
elif revision_id not in weave:
840
raise errors.NoSuchRevision(self, revision_id)
842
# add what can be reached from revision_id
844
pending = set([revision_id])
845
while len(pending) > 0:
847
result[node] = weave.get_parents(node)
848
for revision_id in result[node]:
849
if revision_id not in result:
850
pending.add(revision_id)
854
def get_revision_graph_with_ghosts(self, revision_ids=None):
855
"""Return a graph of the revisions with ghosts marked as applicable.
857
:param revision_ids: an iterable of revisions to graph or None for all.
858
:return: a Graph object with the graph reachable from revision_ids.
861
vf = self._get_revision_vf()
862
versions = set(vf.versions())
864
pending = set(self.all_revision_ids())
867
pending = set(revision_ids)
868
required = set(revision_ids)
871
revision_id = pending.pop()
872
if not revision_id in versions:
873
if revision_id in required:
874
raise errors.NoSuchRevision(self, revision_id)
876
result.add_ghost(revision_id)
877
# mark it as done so we don't try for it again.
878
done.add(revision_id)
880
parent_ids = vf.get_parents_with_ghosts(revision_id)
881
for parent_id in parent_ids:
882
# is this queued or done ?
883
if (parent_id not in pending and
884
parent_id not in done):
886
pending.add(parent_id)
887
result.add_node(revision_id, parent_ids)
888
done.add(revision_id)
891
def _get_revision_vf(self):
892
""":return: a versioned file containing the revisions."""
893
vf = self._revision_store.get_revision_file(self.get_transaction())
897
def reconcile(self, other=None, thorough=False):
898
"""Reconcile this repository."""
899
from bzrlib.reconcile import KnitReconciler
900
reconciler = KnitReconciler(self, thorough=thorough)
901
reconciler.reconcile()
904
def revision_parents(self, revid):
905
return self._get_revision_vf().get_parents(rev_id)
907
class RepositoryFormat(object):
908
"""A repository format.
910
Formats provide three things:
911
* An initialization routine to construct repository data on disk.
912
* a format string which is used when the BzrDir supports versioned
914
* an open routine which returns a Repository instance.
916
Formats are placed in an dict by their format string for reference
917
during opening. These should be subclasses of RepositoryFormat
920
Once a format is deprecated, just deprecate the initialize and open
921
methods on the format class. Do not deprecate the object, as the
922
object will be created every system load.
924
Common instance attributes:
925
_matchingbzrdir - the bzrdir format that the repository format was
926
originally written to work with. This can be used if manually
927
constructing a bzrdir and repository, or more commonly for test suite
931
_default_format = None
932
"""The default format used for new repositories."""
935
"""The known formats."""
938
def find_format(klass, a_bzrdir):
939
"""Return the format for the repository object in a_bzrdir."""
941
transport = a_bzrdir.get_repository_transport(None)
942
format_string = transport.get("format").read()
943
return klass._formats[format_string]
944
except errors.NoSuchFile:
945
raise errors.NoRepositoryPresent(a_bzrdir)
947
raise errors.UnknownFormatError(format_string)
949
def _get_control_store(self, repo_transport, control_files):
950
"""Return the control store for this repository."""
951
raise NotImplementedError(self._get_control_store)
954
def get_default_format(klass):
955
"""Return the current default format."""
956
return klass._default_format
958
def get_format_string(self):
959
"""Return the ASCII format string that identifies this format.
961
Note that in pre format ?? repositories the format string is
962
not permitted nor written to disk.
964
raise NotImplementedError(self.get_format_string)
966
def get_format_description(self):
967
"""Return the short description for this format."""
968
raise NotImplementedError(self.get_format_description)
970
def _get_revision_store(self, repo_transport, control_files):
971
"""Return the revision store object for this a_bzrdir."""
972
raise NotImplementedError(self._get_revision_store)
974
def _get_text_rev_store(self,
981
"""Common logic for getting a revision store for a repository.
983
see self._get_revision_store for the subclass-overridable method to
984
get the store for a repository.
986
from bzrlib.store.revision.text import TextRevisionStore
987
dir_mode = control_files._dir_mode
988
file_mode = control_files._file_mode
989
text_store =TextStore(transport.clone(name),
991
compressed=compressed,
994
_revision_store = TextRevisionStore(text_store, serializer)
995
return _revision_store
997
def _get_versioned_file_store(self,
1002
versionedfile_class=WeaveFile,
1004
weave_transport = control_files._transport.clone(name)
1005
dir_mode = control_files._dir_mode
1006
file_mode = control_files._file_mode
1007
return VersionedFileStore(weave_transport, prefixed=prefixed,
1009
file_mode=file_mode,
1010
versionedfile_class=versionedfile_class,
1013
def initialize(self, a_bzrdir, shared=False):
1014
"""Initialize a repository of this format in a_bzrdir.
1016
:param a_bzrdir: The bzrdir to put the new repository in it.
1017
:param shared: The repository should be initialized as a sharable one.
1019
This may raise UninitializableFormat if shared repository are not
1020
compatible the a_bzrdir.
1023
def is_supported(self):
1024
"""Is this format supported?
1026
Supported formats must be initializable and openable.
1027
Unsupported formats may not support initialization or committing or
1028
some other features depending on the reason for not being supported.
1032
def open(self, a_bzrdir, _found=False):
1033
"""Return an instance of this format for the bzrdir a_bzrdir.
1035
_found is a private parameter, do not use it.
1037
raise NotImplementedError(self.open)
1040
def register_format(klass, format):
1041
klass._formats[format.get_format_string()] = format
1044
def set_default_format(klass, format):
1045
klass._default_format = format
1048
def unregister_format(klass, format):
1049
assert klass._formats[format.get_format_string()] is format
1050
del klass._formats[format.get_format_string()]
1053
class PreSplitOutRepositoryFormat(RepositoryFormat):
1054
"""Base class for the pre split out repository formats."""
1056
def initialize(self, a_bzrdir, shared=False, _internal=False):
1057
"""Create a weave repository.
1059
TODO: when creating split out bzr branch formats, move this to a common
1060
base for Format5, Format6. or something like that.
1062
from bzrlib.weavefile import write_weave_v5
1063
from bzrlib.weave import Weave
1066
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1069
# always initialized when the bzrdir is.
1070
return self.open(a_bzrdir, _found=True)
1072
# Create an empty weave
1074
bzrlib.weavefile.write_weave_v5(Weave(), sio)
1075
empty_weave = sio.getvalue()
1077
mutter('creating repository in %s.', a_bzrdir.transport.base)
1078
dirs = ['revision-store', 'weaves']
1079
files = [('inventory.weave', StringIO(empty_weave)),
1082
# FIXME: RBC 20060125 don't peek under the covers
1083
# NB: no need to escape relative paths that are url safe.
1084
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1086
control_files.create_lock()
1087
control_files.lock_write()
1088
control_files._transport.mkdir_multi(dirs,
1089
mode=control_files._dir_mode)
1091
for file, content in files:
1092
control_files.put(file, content)
1094
control_files.unlock()
1095
return self.open(a_bzrdir, _found=True)
1097
def _get_control_store(self, repo_transport, control_files):
1098
"""Return the control store for this repository."""
1099
return self._get_versioned_file_store('',
1104
def _get_text_store(self, transport, control_files):
1105
"""Get a store for file texts for this format."""
1106
raise NotImplementedError(self._get_text_store)
1108
def open(self, a_bzrdir, _found=False):
1109
"""See RepositoryFormat.open()."""
1111
# we are being called directly and must probe.
1112
raise NotImplementedError
1114
repo_transport = a_bzrdir.get_repository_transport(None)
1115
control_files = a_bzrdir._control_files
1116
text_store = self._get_text_store(repo_transport, control_files)
1117
control_store = self._get_control_store(repo_transport, control_files)
1118
_revision_store = self._get_revision_store(repo_transport, control_files)
1119
return AllInOneRepository(_format=self,
1121
_revision_store=_revision_store,
1122
control_store=control_store,
1123
text_store=text_store)
1126
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1127
"""Bzr repository format 4.
1129
This repository format has:
1131
- TextStores for texts, inventories,revisions.
1133
This format is deprecated: it indexes texts using a text id which is
1134
removed in format 5; initialization and write support for this format
1139
super(RepositoryFormat4, self).__init__()
1140
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat4()
1142
def get_format_description(self):
1143
"""See RepositoryFormat.get_format_description()."""
1144
return "Repository format 4"
1146
def initialize(self, url, shared=False, _internal=False):
1147
"""Format 4 branches cannot be created."""
1148
raise errors.UninitializableFormat(self)
1150
def is_supported(self):
1151
"""Format 4 is not supported.
1153
It is not supported because the model changed from 4 to 5 and the
1154
conversion logic is expensive - so doing it on the fly was not
1159
def _get_control_store(self, repo_transport, control_files):
1160
"""Format 4 repositories have no formal control store at this point.
1162
This will cause any control-file-needing apis to fail - this is desired.
1166
def _get_revision_store(self, repo_transport, control_files):
1167
"""See RepositoryFormat._get_revision_store()."""
1168
from bzrlib.xml4 import serializer_v4
1169
return self._get_text_rev_store(repo_transport,
1172
serializer=serializer_v4)
1174
def _get_text_store(self, transport, control_files):
1175
"""See RepositoryFormat._get_text_store()."""
1178
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1179
"""Bzr control format 5.
1181
This repository format has:
1182
- weaves for file texts and inventory
1184
- TextStores for revisions and signatures.
1188
super(RepositoryFormat5, self).__init__()
1189
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat5()
1191
def get_format_description(self):
1192
"""See RepositoryFormat.get_format_description()."""
1193
return "Weave repository format 5"
1195
def _get_revision_store(self, repo_transport, control_files):
1196
"""See RepositoryFormat._get_revision_store()."""
1197
"""Return the revision store object for this a_bzrdir."""
1198
return self._get_text_rev_store(repo_transport,
1203
def _get_text_store(self, transport, control_files):
1204
"""See RepositoryFormat._get_text_store()."""
1205
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1208
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1209
"""Bzr control format 6.
1211
This repository format has:
1212
- weaves for file texts and inventory
1213
- hash subdirectory based stores.
1214
- TextStores for revisions and signatures.
1218
super(RepositoryFormat6, self).__init__()
1219
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat6()
1221
def get_format_description(self):
1222
"""See RepositoryFormat.get_format_description()."""
1223
return "Weave repository format 6"
1225
def _get_revision_store(self, repo_transport, control_files):
1226
"""See RepositoryFormat._get_revision_store()."""
1227
return self._get_text_rev_store(repo_transport,
1233
def _get_text_store(self, transport, control_files):
1234
"""See RepositoryFormat._get_text_store()."""
1235
return self._get_versioned_file_store('weaves', transport, control_files)
1238
class MetaDirRepositoryFormat(RepositoryFormat):
1239
"""Common base class for the new repositories using the metadir layout."""
1242
super(MetaDirRepositoryFormat, self).__init__()
1243
self._matchingbzrdir = bzrlib.bzrdir.BzrDirMetaFormat1()
1245
def _create_control_files(self, a_bzrdir):
1246
"""Create the required files and the initial control_files object."""
1247
# FIXME: RBC 20060125 don't peek under the covers
1248
# NB: no need to escape relative paths that are url safe.
1249
repository_transport = a_bzrdir.get_repository_transport(self)
1250
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1251
control_files.create_lock()
1252
return control_files
1254
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1255
"""Upload the initial blank content."""
1256
control_files = self._create_control_files(a_bzrdir)
1257
control_files.lock_write()
1259
control_files._transport.mkdir_multi(dirs,
1260
mode=control_files._dir_mode)
1261
for file, content in files:
1262
control_files.put(file, content)
1263
for file, content in utf8_files:
1264
control_files.put_utf8(file, content)
1266
control_files.put_utf8('shared-storage', '')
1268
control_files.unlock()
1271
class RepositoryFormat7(MetaDirRepositoryFormat):
1272
"""Bzr repository 7.
1274
This repository format has:
1275
- weaves for file texts and inventory
1276
- hash subdirectory based stores.
1277
- TextStores for revisions and signatures.
1278
- a format marker of its own
1279
- an optional 'shared-storage' flag
1280
- an optional 'no-working-trees' flag
1283
def _get_control_store(self, repo_transport, control_files):
1284
"""Return the control store for this repository."""
1285
return self._get_versioned_file_store('',
1290
def get_format_string(self):
1291
"""See RepositoryFormat.get_format_string()."""
1292
return "Bazaar-NG Repository format 7"
1294
def get_format_description(self):
1295
"""See RepositoryFormat.get_format_description()."""
1296
return "Weave repository format 7"
1298
def _get_revision_store(self, repo_transport, control_files):
1299
"""See RepositoryFormat._get_revision_store()."""
1300
return self._get_text_rev_store(repo_transport,
1307
def _get_text_store(self, transport, control_files):
1308
"""See RepositoryFormat._get_text_store()."""
1309
return self._get_versioned_file_store('weaves',
1313
def initialize(self, a_bzrdir, shared=False):
1314
"""Create a weave repository.
1316
:param shared: If true the repository will be initialized as a shared
1319
from bzrlib.weavefile import write_weave_v5
1320
from bzrlib.weave import Weave
1322
# Create an empty weave
1324
bzrlib.weavefile.write_weave_v5(Weave(), sio)
1325
empty_weave = sio.getvalue()
1327
mutter('creating repository in %s.', a_bzrdir.transport.base)
1328
dirs = ['revision-store', 'weaves']
1329
files = [('inventory.weave', StringIO(empty_weave)),
1331
utf8_files = [('format', self.get_format_string())]
1333
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1334
return self.open(a_bzrdir=a_bzrdir, _found=True)
1336
def open(self, a_bzrdir, _found=False, _override_transport=None):
1337
"""See RepositoryFormat.open().
1339
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1340
repository at a slightly different url
1341
than normal. I.e. during 'upgrade'.
1344
format = RepositoryFormat.find_format(a_bzrdir)
1345
assert format.__class__ == self.__class__
1346
if _override_transport is not None:
1347
repo_transport = _override_transport
1349
repo_transport = a_bzrdir.get_repository_transport(None)
1350
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1351
text_store = self._get_text_store(repo_transport, control_files)
1352
control_store = self._get_control_store(repo_transport, control_files)
1353
_revision_store = self._get_revision_store(repo_transport, control_files)
1354
return MetaDirRepository(_format=self,
1356
control_files=control_files,
1357
_revision_store=_revision_store,
1358
control_store=control_store,
1359
text_store=text_store)
1362
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1363
"""Bzr repository knit format 1.
1365
This repository format has:
1366
- knits for file texts and inventory
1367
- hash subdirectory based stores.
1368
- knits for revisions and signatures
1369
- TextStores for revisions and signatures.
1370
- a format marker of its own
1371
- an optional 'shared-storage' flag
1372
- an optional 'no-working-trees' flag
1375
This format was introduced in bzr 0.8.
1378
def _get_control_store(self, repo_transport, control_files):
1379
"""Return the control store for this repository."""
1380
return VersionedFileStore(
1383
file_mode=control_files._file_mode,
1384
versionedfile_class=KnitVersionedFile,
1385
versionedfile_kwargs={'factory':KnitPlainFactory()},
1388
def get_format_string(self):
1389
"""See RepositoryFormat.get_format_string()."""
1390
return "Bazaar-NG Knit Repository Format 1"
1392
def get_format_description(self):
1393
"""See RepositoryFormat.get_format_description()."""
1394
return "Knit repository format 1"
1396
def _get_revision_store(self, repo_transport, control_files):
1397
"""See RepositoryFormat._get_revision_store()."""
1398
from bzrlib.store.revision.knit import KnitRevisionStore
1399
versioned_file_store = VersionedFileStore(
1401
file_mode=control_files._file_mode,
1404
versionedfile_class=KnitVersionedFile,
1405
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1408
return KnitRevisionStore(versioned_file_store)
1410
def _get_text_store(self, transport, control_files):
1411
"""See RepositoryFormat._get_text_store()."""
1412
return self._get_versioned_file_store('knits',
1415
versionedfile_class=KnitVersionedFile,
1418
def initialize(self, a_bzrdir, shared=False):
1419
"""Create a knit format 1 repository.
1421
:param a_bzrdir: bzrdir to contain the new repository; must already
1423
:param shared: If true the repository will be initialized as a shared
1426
mutter('creating repository in %s.', a_bzrdir.transport.base)
1427
dirs = ['revision-store', 'knits']
1429
utf8_files = [('format', self.get_format_string())]
1431
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1432
repo_transport = a_bzrdir.get_repository_transport(None)
1433
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1434
control_store = self._get_control_store(repo_transport, control_files)
1435
transaction = bzrlib.transactions.WriteTransaction()
1436
# trigger a write of the inventory store.
1437
control_store.get_weave_or_empty('inventory', transaction)
1438
_revision_store = self._get_revision_store(repo_transport, control_files)
1439
_revision_store.has_revision_id('A', transaction)
1440
_revision_store.get_signature_file(transaction)
1441
return self.open(a_bzrdir=a_bzrdir, _found=True)
1443
def open(self, a_bzrdir, _found=False, _override_transport=None):
1444
"""See RepositoryFormat.open().
1446
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1447
repository at a slightly different url
1448
than normal. I.e. during 'upgrade'.
1451
format = RepositoryFormat.find_format(a_bzrdir)
1452
assert format.__class__ == self.__class__
1453
if _override_transport is not None:
1454
repo_transport = _override_transport
1456
repo_transport = a_bzrdir.get_repository_transport(None)
1457
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1458
text_store = self._get_text_store(repo_transport, control_files)
1459
control_store = self._get_control_store(repo_transport, control_files)
1460
_revision_store = self._get_revision_store(repo_transport, control_files)
1461
return KnitRepository(_format=self,
1463
control_files=control_files,
1464
_revision_store=_revision_store,
1465
control_store=control_store,
1466
text_store=text_store)
1469
# formats which have no format string are not discoverable
1470
# and not independently creatable, so are not registered.
1471
RepositoryFormat.register_format(RepositoryFormat7())
1472
_default_format = RepositoryFormatKnit1()
1473
RepositoryFormat.register_format(_default_format)
1474
RepositoryFormat.set_default_format(_default_format)
1475
_legacy_formats = [RepositoryFormat4(),
1476
RepositoryFormat5(),
1477
RepositoryFormat6()]
1480
class InterRepository(InterObject):
1481
"""This class represents operations taking place between two repositories.
1483
Its instances have methods like copy_content and fetch, and contain
1484
references to the source and target repositories these operations can be
1487
Often we will provide convenience methods on 'repository' which carry out
1488
operations with another repository - they will always forward to
1489
InterRepository.get(other).method_name(parameters).
1493
"""The available optimised InterRepository types."""
1496
def copy_content(self, revision_id=None, basis=None):
1497
"""Make a complete copy of the content in self into destination.
1499
This is a destructive operation! Do not use it on existing
1502
:param revision_id: Only copy the content needed to construct
1503
revision_id and its parents.
1504
:param basis: Copy the needed data preferentially from basis.
1507
self.target.set_make_working_trees(self.source.make_working_trees())
1508
except NotImplementedError:
1510
# grab the basis available data
1511
if basis is not None:
1512
self.target.fetch(basis, revision_id=revision_id)
1513
# but don't bother fetching if we have the needed data now.
1514
if (revision_id not in (None, NULL_REVISION) and
1515
self.target.has_revision(revision_id)):
1517
self.target.fetch(self.source, revision_id=revision_id)
1519
def _double_lock(self, lock_source, lock_target):
1520
"""Take out too locks, rolling back the first if the second throws."""
1525
# we want to ensure that we don't leave source locked by mistake.
1526
# and any error on target should not confuse source.
1527
self.source.unlock()
1531
def fetch(self, revision_id=None, pb=None):
1532
"""Fetch the content required to construct revision_id.
1534
The content is copied from source to target.
1536
:param revision_id: if None all content is copied, if NULL_REVISION no
1538
:param pb: optional progress bar to use for progress reports. If not
1539
provided a default one will be created.
1541
Returns the copied revision count and the failed revisions in a tuple:
1544
from bzrlib.fetch import GenericRepoFetcher
1545
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1546
self.source, self.source._format, self.target, self.target._format)
1547
f = GenericRepoFetcher(to_repository=self.target,
1548
from_repository=self.source,
1549
last_revision=revision_id,
1551
return f.count_copied, f.failed_revisions
1553
def lock_read(self):
1554
"""Take out a logical read lock.
1556
This will lock the source branch and the target branch. The source gets
1557
a read lock and the target a read lock.
1559
self._double_lock(self.source.lock_read, self.target.lock_read)
1561
def lock_write(self):
1562
"""Take out a logical write lock.
1564
This will lock the source branch and the target branch. The source gets
1565
a read lock and the target a write lock.
1567
self._double_lock(self.source.lock_read, self.target.lock_write)
1570
def missing_revision_ids(self, revision_id=None):
1571
"""Return the revision ids that source has that target does not.
1573
These are returned in topological order.
1575
:param revision_id: only return revision ids included by this
1578
# generic, possibly worst case, slow code path.
1579
target_ids = set(self.target.all_revision_ids())
1580
if revision_id is not None:
1581
source_ids = self.source.get_ancestry(revision_id)
1582
assert source_ids[0] == None
1585
source_ids = self.source.all_revision_ids()
1586
result_set = set(source_ids).difference(target_ids)
1587
# this may look like a no-op: its not. It preserves the ordering
1588
# other_ids had while only returning the members from other_ids
1589
# that we've decided we need.
1590
return [rev_id for rev_id in source_ids if rev_id in result_set]
1593
"""Release the locks on source and target."""
1595
self.target.unlock()
1597
self.source.unlock()
1600
class InterWeaveRepo(InterRepository):
1601
"""Optimised code paths between Weave based repositories."""
1603
_matching_repo_format = RepositoryFormat7()
1604
"""Repository format for testing with."""
1607
def is_compatible(source, target):
1608
"""Be compatible with known Weave formats.
1610
We don't test for the stores being of specific types because that
1611
could lead to confusing results, and there is no need to be
1615
return (isinstance(source._format, (RepositoryFormat5,
1617
RepositoryFormat7)) and
1618
isinstance(target._format, (RepositoryFormat5,
1620
RepositoryFormat7)))
1621
except AttributeError:
1625
def copy_content(self, revision_id=None, basis=None):
1626
"""See InterRepository.copy_content()."""
1627
# weave specific optimised path:
1628
if basis is not None:
1629
# copy the basis in, then fetch remaining data.
1630
basis.copy_content_into(self.target, revision_id)
1631
# the basis copy_content_into could miss-set this.
1633
self.target.set_make_working_trees(self.source.make_working_trees())
1634
except NotImplementedError:
1636
self.target.fetch(self.source, revision_id=revision_id)
1639
self.target.set_make_working_trees(self.source.make_working_trees())
1640
except NotImplementedError:
1642
# FIXME do not peek!
1643
if self.source.control_files._transport.listable():
1644
pb = bzrlib.ui.ui_factory.nested_progress_bar()
1646
self.target.weave_store.copy_all_ids(
1647
self.source.weave_store,
1649
from_transaction=self.source.get_transaction(),
1650
to_transaction=self.target.get_transaction())
1651
pb.update('copying inventory', 0, 1)
1652
self.target.control_weaves.copy_multi(
1653
self.source.control_weaves, ['inventory'],
1654
from_transaction=self.source.get_transaction(),
1655
to_transaction=self.target.get_transaction())
1656
self.target._revision_store.text_store.copy_all_ids(
1657
self.source._revision_store.text_store,
1662
self.target.fetch(self.source, revision_id=revision_id)
1665
def fetch(self, revision_id=None, pb=None):
1666
"""See InterRepository.fetch()."""
1667
from bzrlib.fetch import GenericRepoFetcher
1668
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1669
self.source, self.source._format, self.target, self.target._format)
1670
f = GenericRepoFetcher(to_repository=self.target,
1671
from_repository=self.source,
1672
last_revision=revision_id,
1674
return f.count_copied, f.failed_revisions
1677
def missing_revision_ids(self, revision_id=None):
1678
"""See InterRepository.missing_revision_ids()."""
1679
# we want all revisions to satisfy revision_id in source.
1680
# but we don't want to stat every file here and there.
1681
# we want then, all revisions other needs to satisfy revision_id
1682
# checked, but not those that we have locally.
1683
# so the first thing is to get a subset of the revisions to
1684
# satisfy revision_id in source, and then eliminate those that
1685
# we do already have.
1686
# this is slow on high latency connection to self, but as as this
1687
# disk format scales terribly for push anyway due to rewriting
1688
# inventory.weave, this is considered acceptable.
1690
if revision_id is not None:
1691
source_ids = self.source.get_ancestry(revision_id)
1692
assert source_ids[0] == None
1695
source_ids = self.source._all_possible_ids()
1696
source_ids_set = set(source_ids)
1697
# source_ids is the worst possible case we may need to pull.
1698
# now we want to filter source_ids against what we actually
1699
# have in target, but don't try to check for existence where we know
1700
# we do not have a revision as that would be pointless.
1701
target_ids = set(self.target._all_possible_ids())
1702
possibly_present_revisions = target_ids.intersection(source_ids_set)
1703
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1704
required_revisions = source_ids_set.difference(actually_present_revisions)
1705
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1706
if revision_id is not None:
1707
# we used get_ancestry to determine source_ids then we are assured all
1708
# revisions referenced are present as they are installed in topological order.
1709
# and the tip revision was validated by get_ancestry.
1710
return required_topo_revisions
1712
# if we just grabbed the possibly available ids, then
1713
# we only have an estimate of whats available and need to validate
1714
# that against the revision records.
1715
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1718
class InterKnitRepo(InterRepository):
1719
"""Optimised code paths between Knit based repositories."""
1721
_matching_repo_format = RepositoryFormatKnit1()
1722
"""Repository format for testing with."""
1725
def is_compatible(source, target):
1726
"""Be compatible with known Knit formats.
1728
We don't test for the stores being of specific types because that
1729
could lead to confusing results, and there is no need to be
1733
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1734
isinstance(target._format, (RepositoryFormatKnit1)))
1735
except AttributeError:
1739
def fetch(self, revision_id=None, pb=None):
1740
"""See InterRepository.fetch()."""
1741
from bzrlib.fetch import KnitRepoFetcher
1742
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1743
self.source, self.source._format, self.target, self.target._format)
1744
f = KnitRepoFetcher(to_repository=self.target,
1745
from_repository=self.source,
1746
last_revision=revision_id,
1748
return f.count_copied, f.failed_revisions
1751
def missing_revision_ids(self, revision_id=None):
1752
"""See InterRepository.missing_revision_ids()."""
1753
if revision_id is not None:
1754
source_ids = self.source.get_ancestry(revision_id)
1755
assert source_ids[0] == None
1758
source_ids = self.source._all_possible_ids()
1759
source_ids_set = set(source_ids)
1760
# source_ids is the worst possible case we may need to pull.
1761
# now we want to filter source_ids against what we actually
1762
# have in target, but don't try to check for existence where we know
1763
# we do not have a revision as that would be pointless.
1764
target_ids = set(self.target._all_possible_ids())
1765
possibly_present_revisions = target_ids.intersection(source_ids_set)
1766
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1767
required_revisions = source_ids_set.difference(actually_present_revisions)
1768
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1769
if revision_id is not None:
1770
# we used get_ancestry to determine source_ids then we are assured all
1771
# revisions referenced are present as they are installed in topological order.
1772
# and the tip revision was validated by get_ancestry.
1773
return required_topo_revisions
1775
# if we just grabbed the possibly available ids, then
1776
# we only have an estimate of whats available and need to validate
1777
# that against the revision records.
1778
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1780
InterRepository.register_optimiser(InterWeaveRepo)
1781
InterRepository.register_optimiser(InterKnitRepo)
1784
class RepositoryTestProviderAdapter(object):
1785
"""A tool to generate a suite testing multiple repository formats at once.
1787
This is done by copying the test once for each transport and injecting
1788
the transport_server, transport_readonly_server, and bzrdir_format and
1789
repository_format classes into each copy. Each copy is also given a new id()
1790
to make it easy to identify.
1793
def __init__(self, transport_server, transport_readonly_server, formats):
1794
self._transport_server = transport_server
1795
self._transport_readonly_server = transport_readonly_server
1796
self._formats = formats
1798
def adapt(self, test):
1799
result = TestSuite()
1800
for repository_format, bzrdir_format in self._formats:
1801
new_test = deepcopy(test)
1802
new_test.transport_server = self._transport_server
1803
new_test.transport_readonly_server = self._transport_readonly_server
1804
new_test.bzrdir_format = bzrdir_format
1805
new_test.repository_format = repository_format
1806
def make_new_test_id():
1807
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1808
return lambda: new_id
1809
new_test.id = make_new_test_id()
1810
result.addTest(new_test)
1814
class InterRepositoryTestProviderAdapter(object):
1815
"""A tool to generate a suite testing multiple inter repository formats.
1817
This is done by copying the test once for each interrepo provider and injecting
1818
the transport_server, transport_readonly_server, repository_format and
1819
repository_to_format classes into each copy.
1820
Each copy is also given a new id() to make it easy to identify.
1823
def __init__(self, transport_server, transport_readonly_server, formats):
1824
self._transport_server = transport_server
1825
self._transport_readonly_server = transport_readonly_server
1826
self._formats = formats
1828
def adapt(self, test):
1829
result = TestSuite()
1830
for interrepo_class, repository_format, repository_format_to in self._formats:
1831
new_test = deepcopy(test)
1832
new_test.transport_server = self._transport_server
1833
new_test.transport_readonly_server = self._transport_readonly_server
1834
new_test.interrepo_class = interrepo_class
1835
new_test.repository_format = repository_format
1836
new_test.repository_format_to = repository_format_to
1837
def make_new_test_id():
1838
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1839
return lambda: new_id
1840
new_test.id = make_new_test_id()
1841
result.addTest(new_test)
1845
def default_test_list():
1846
"""Generate the default list of interrepo permutations to test."""
1848
# test the default InterRepository between format 6 and the current
1850
# XXX: robertc 20060220 reinstate this when there are two supported
1851
# formats which do not have an optimal code path between them.
1852
result.append((InterRepository,
1853
RepositoryFormat6(),
1854
RepositoryFormatKnit1()))
1855
for optimiser in InterRepository._optimisers:
1856
result.append((optimiser,
1857
optimiser._matching_repo_format,
1858
optimiser._matching_repo_format
1860
# if there are specific combinations we want to use, we can add them
1865
class CopyConverter(object):
1866
"""A repository conversion tool which just performs a copy of the content.
1868
This is slow but quite reliable.
1871
def __init__(self, target_format):
1872
"""Create a CopyConverter.
1874
:param target_format: The format the resulting repository should be.
1876
self.target_format = target_format
1878
def convert(self, repo, pb):
1879
"""Perform the conversion of to_convert, giving feedback via pb.
1881
:param to_convert: The disk object to convert.
1882
:param pb: a progress bar to use for progress information.
1887
# this is only useful with metadir layouts - separated repo content.
1888
# trigger an assertion if not such
1889
repo._format.get_format_string()
1890
self.repo_dir = repo.bzrdir
1891
self.step('Moving repository to repository.backup')
1892
self.repo_dir.transport.move('repository', 'repository.backup')
1893
backup_transport = self.repo_dir.transport.clone('repository.backup')
1894
self.source_repo = repo._format.open(self.repo_dir,
1896
_override_transport=backup_transport)
1897
self.step('Creating new repository')
1898
converted = self.target_format.initialize(self.repo_dir,
1899
self.source_repo.is_shared())
1900
converted.lock_write()
1902
self.step('Copying content into repository.')
1903
self.source_repo.copy_content_into(converted)
1906
self.step('Deleting old repository content.')
1907
self.repo_dir.transport.delete_tree('repository.backup')
1908
self.pb.note('repository converted')
1910
def step(self, message):
1911
"""Update the pb by a step."""
1913
self.pb.update(message, self.count, self.total)
1916
class CommitBuilder(object):
1917
"""Provides an interface to build up a commit.
1919
This allows describing a tree to be committed without needing to
1920
know the internals of the format of the repository.
1922
def __init__(self, repository, parents, config, timestamp=None,
1923
timezone=None, committer=None, revprops=None,
1925
"""Initiate a CommitBuilder.
1927
:param repository: Repository to commit to.
1928
:param parents: Revision ids of the parents of the new revision.
1929
:param config: Configuration to use.
1930
:param timestamp: Optional timestamp recorded for commit.
1931
:param timezone: Optional timezone for timestamp.
1932
:param committer: Optional committer to set for commit.
1933
:param revprops: Optional dictionary of revision properties.
1934
:param revision_id: Optional revision id.
1936
self._config = config
1938
if committer is None:
1939
self._committer = self._config.username()
1941
assert isinstance(committer, basestring), type(committer)
1942
self._committer = committer
1944
self.new_inventory = Inventory()
1945
self._new_revision_id = revision_id
1946
self.parents = parents
1947
self.repository = repository
1950
if revprops is not None:
1951
self._revprops.update(revprops)
1953
if timestamp is None:
1954
self._timestamp = time.time()
1956
self._timestamp = long(timestamp)
1958
if timezone is None:
1959
self._timezone = local_time_offset()
1961
self._timezone = int(timezone)
1963
self._generate_revision_if_needed()
1965
def commit(self, message):
1966
"""Make the actual commit.
1968
:return: The revision id of the recorded revision.
1970
rev = Revision(timestamp=self._timestamp,
1971
timezone=self._timezone,
1972
committer=self._committer,
1974
inventory_sha1=self.inv_sha1,
1975
revision_id=self._new_revision_id,
1976
properties=self._revprops)
1977
rev.parent_ids = self.parents
1978
self.repository.add_revision(self._new_revision_id, rev,
1979
self.new_inventory, self._config)
1980
return self._new_revision_id
1982
def finish_inventory(self):
1983
"""Tell the builder that the inventory is finished."""
1984
self.new_inventory.revision_id = self._new_revision_id
1985
self.inv_sha1 = self.repository.add_inventory(
1986
self._new_revision_id,
1991
def _gen_revision_id(self):
1992
"""Return new revision-id."""
1993
s = '%s-%s-' % (self._config.user_email(),
1994
compact_date(self._timestamp))
1995
s += hexlify(rand_bytes(8))
1998
def _generate_revision_if_needed(self):
1999
"""Create a revision id if None was supplied.
2001
If the repository can not support user-specified revision ids
2002
they should override this function and raise UnsupportedOperation
2003
if _new_revision_id is not None.
2005
:raises: UnsupportedOperation
2007
if self._new_revision_id is None:
2008
self._new_revision_id = self._gen_revision_id()
2010
def record_entry_contents(self, ie, parent_invs, path, tree):
2011
"""Record the content of ie from tree into the commit if needed.
2013
:param ie: An inventory entry present in the commit.
2014
:param parent_invs: The inventories of the parent revisions of the
2016
:param path: The path the entry is at in the tree.
2017
:param tree: The tree which contains this entry and should be used to
2020
self.new_inventory.add(ie)
2022
# ie.revision is always None if the InventoryEntry is considered
2023
# for committing. ie.snapshot will record the correct revision
2024
# which may be the sole parent if it is untouched.
2025
if ie.revision is not None:
2027
previous_entries = ie.find_previous_heads(
2029
self.repository.weave_store,
2030
self.repository.get_transaction())
2031
# we are creating a new revision for ie in the history store
2033
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2035
def modified_directory(self, file_id, file_parents):
2036
"""Record the presence of a symbolic link.
2038
:param file_id: The file_id of the link to record.
2039
:param file_parents: The per-file parent revision ids.
2041
self._add_text_to_weave(file_id, [], file_parents.keys())
2043
def modified_file_text(self, file_id, file_parents,
2044
get_content_byte_lines, text_sha1=None,
2046
"""Record the text of file file_id
2048
:param file_id: The file_id of the file to record the text of.
2049
:param file_parents: The per-file parent revision ids.
2050
:param get_content_byte_lines: A callable which will return the byte
2052
:param text_sha1: Optional SHA1 of the file contents.
2053
:param text_size: Optional size of the file contents.
2055
mutter('storing text of file {%s} in revision {%s} into %r',
2056
file_id, self._new_revision_id, self.repository.weave_store)
2057
# special case to avoid diffing on renames or
2059
if (len(file_parents) == 1
2060
and text_sha1 == file_parents.values()[0].text_sha1
2061
and text_size == file_parents.values()[0].text_size):
2062
previous_ie = file_parents.values()[0]
2063
versionedfile = self.repository.weave_store.get_weave(file_id,
2064
self.repository.get_transaction())
2065
versionedfile.clone_text(self._new_revision_id,
2066
previous_ie.revision, file_parents.keys())
2067
return text_sha1, text_size
2069
new_lines = get_content_byte_lines()
2070
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2071
# should return the SHA1 and size
2072
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2073
return bzrlib.osutils.sha_strings(new_lines), \
2074
sum(map(len, new_lines))
2076
def modified_link(self, file_id, file_parents, link_target):
2077
"""Record the presence of a symbolic link.
2079
:param file_id: The file_id of the link to record.
2080
:param file_parents: The per-file parent revision ids.
2081
:param link_target: Target location of this link.
2083
self._add_text_to_weave(file_id, [], file_parents.keys())
2085
def _add_text_to_weave(self, file_id, new_lines, parents):
2086
versionedfile = self.repository.weave_store.get_weave_or_empty(
2087
file_id, self.repository.get_transaction())
2088
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2089
versionedfile.clear_cache()
2092
# Copied from xml.sax.saxutils
2093
def _unescape_xml(data):
2094
"""Unescape &, <, and > in a string of data.
2096
data = data.replace("<", "<")
2097
data = data.replace(">", ">")
2098
# must do ampersand last
2099
return data.replace("&", "&")