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
23
from warnings import warn
25
from bzrlib import bzrdir, check, delta, gpg, errors, xml5, ui, transactions, osutils
26
from bzrlib.decorators import needs_read_lock, needs_write_lock
27
from bzrlib.errors import InvalidRevisionId
28
from bzrlib.graph import Graph
29
from bzrlib.inter import InterObject
30
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
31
from bzrlib.knit import KnitVersionedFile, KnitPlainFactory
32
from bzrlib.lockable_files import LockableFiles, TransportLock
33
from bzrlib.lockdir import LockDir
34
from bzrlib.osutils import (safe_unicode, rand_bytes, compact_date,
36
from bzrlib.revision import NULL_REVISION, Revision
37
from bzrlib.revisiontree import RevisionTree
38
from bzrlib.store.versioned import VersionedFileStore, WeaveStore
39
from bzrlib.store.text import TextStore
40
from bzrlib.symbol_versioning import (deprecated_method,
43
from bzrlib.testament import Testament
44
from bzrlib.trace import mutter, note
45
from bzrlib.tsort import topo_sort
46
from bzrlib.weave import WeaveFile
49
class Repository(object):
50
"""Repository holding history for one or more branches.
52
The repository holds and retrieves historical information including
53
revisions and file history. It's normally accessed only by the Branch,
54
which views a particular line of development through that history.
56
The Repository builds on top of Stores and a Transport, which respectively
57
describe the disk data format and the way of accessing the (possibly
62
def add_inventory(self, revid, inv, parents):
63
"""Add the inventory inv to the repository as revid.
65
:param parents: The revision ids of the parents that revid
66
is known to have and are in the repository already.
68
returns the sha1 of the serialized inventory.
70
assert inv.revision_id is None or inv.revision_id == revid, \
71
"Mismatch between inventory revision" \
72
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
73
inv_text = xml5.serializer_v5.write_inventory_to_string(inv)
74
inv_sha1 = osutils.sha_string(inv_text)
75
inv_vf = self.control_weaves.get_weave('inventory',
76
self.get_transaction())
77
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
80
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
82
for parent in parents:
84
final_parents.append(parent)
86
inv_vf.add_lines(revid, final_parents, lines)
89
def add_revision(self, rev_id, rev, inv=None, config=None):
90
"""Add rev to the revision store as rev_id.
92
:param rev_id: the revision id to use.
93
:param rev: The revision object.
94
:param inv: The inventory for the revision. if None, it will be looked
95
up in the inventory storer
96
:param config: If None no digital signature will be created.
97
If supplied its signature_needed method will be used
98
to determine if a signature should be made.
100
if config is not None and config.signature_needed():
102
inv = self.get_inventory(rev_id)
103
plaintext = Testament(rev, inv).as_short_text()
104
self.store_revision_signature(
105
gpg.GPGStrategy(config), plaintext, rev_id)
106
if not rev_id in self.get_inventory_weave():
108
raise errors.WeaveRevisionNotPresent(rev_id,
109
self.get_inventory_weave())
111
# yes, this is not suitable for adding with ghosts.
112
self.add_inventory(rev_id, inv, rev.parent_ids)
113
self._revision_store.add_revision(rev, self.get_transaction())
116
def _all_possible_ids(self):
117
"""Return all the possible revisions that we could find."""
118
return self.get_inventory_weave().versions()
120
def all_revision_ids(self):
121
"""Returns a list of all the revision ids in the repository.
123
This is deprecated because code should generally work on the graph
124
reachable from a particular revision, and ignore any other revisions
125
that might be present. There is no direct replacement method.
127
return self._all_revision_ids()
130
def _all_revision_ids(self):
131
"""Returns a list of all the revision ids in the repository.
133
These are in as much topological order as the underlying store can
134
present: for weaves ghosts may lead to a lack of correctness until
135
the reweave updates the parents list.
137
if self._revision_store.text_store.listable():
138
return self._revision_store.all_revision_ids(self.get_transaction())
139
result = self._all_possible_ids()
140
return self._eliminate_revisions_not_present(result)
142
def break_lock(self):
143
"""Break a lock if one is present from another instance.
145
Uses the ui factory to ask for confirmation if the lock may be from
148
self.control_files.break_lock()
151
def _eliminate_revisions_not_present(self, revision_ids):
152
"""Check every revision id in revision_ids to see if we have it.
154
Returns a set of the present revisions.
157
for id in revision_ids:
158
if self.has_revision(id):
163
def create(a_bzrdir):
164
"""Construct the current default format repository in a_bzrdir."""
165
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
167
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
168
"""instantiate a Repository.
170
:param _format: The format of the repository on disk.
171
:param a_bzrdir: The BzrDir of the repository.
173
In the future we will have a single api for all stores for
174
getting file texts, inventories and revisions, then
175
this construct will accept instances of those things.
177
super(Repository, self).__init__()
178
self._format = _format
179
# the following are part of the public API for Repository:
180
self.bzrdir = a_bzrdir
181
self.control_files = control_files
182
self._revision_store = _revision_store
183
self.text_store = text_store
184
# backwards compatibility
185
self.weave_store = text_store
186
# not right yet - should be more semantically clear ?
188
self.control_store = control_store
189
self.control_weaves = control_store
190
# TODO: make sure to construct the right store classes, etc, depending
191
# on whether escaping is required.
194
return '%s(%r)' % (self.__class__.__name__,
195
self.bzrdir.transport.base)
198
return self.control_files.is_locked()
200
def lock_write(self):
201
self.control_files.lock_write()
204
self.control_files.lock_read()
206
def get_physical_lock_status(self):
207
return self.control_files.get_physical_lock_status()
210
def missing_revision_ids(self, other, revision_id=None):
211
"""Return the revision ids that other has that this does not.
213
These are returned in topological order.
215
revision_id: only return revision ids included by revision_id.
217
return InterRepository.get(other, self).missing_revision_ids(revision_id)
221
"""Open the repository rooted at base.
223
For instance, if the repository is at URL/.bzr/repository,
224
Repository.open(URL) -> a Repository instance.
226
control = bzrdir.BzrDir.open(base)
227
return control.open_repository()
229
def copy_content_into(self, destination, revision_id=None, basis=None):
230
"""Make a complete copy of the content in self into destination.
232
This is a destructive operation! Do not use it on existing
235
return InterRepository.get(self, destination).copy_content(revision_id, basis)
237
def fetch(self, source, revision_id=None, pb=None):
238
"""Fetch the content required to construct revision_id from source.
240
If revision_id is None all content is copied.
242
return InterRepository.get(source, self).fetch(revision_id=revision_id,
245
def get_commit_builder(self, branch, parents, config, timestamp=None,
246
timezone=None, committer=None, revprops=None,
248
"""Obtain a CommitBuilder for this repository.
250
:param branch: Branch to commit to.
251
:param parents: Revision ids of the parents of the new revision.
252
:param config: Configuration to use.
253
:param timestamp: Optional timestamp recorded for commit.
254
:param timezone: Optional timezone for timestamp.
255
:param committer: Optional committer to set for commit.
256
:param revprops: Optional dictionary of revision properties.
257
:param revision_id: Optional revision id.
259
return CommitBuilder(self, parents, config, timestamp, timezone,
260
committer, revprops, revision_id)
263
self.control_files.unlock()
266
def clone(self, a_bzrdir, revision_id=None, basis=None):
267
"""Clone this repository into a_bzrdir using the current format.
269
Currently no check is made that the format of this repository and
270
the bzrdir format are compatible. FIXME RBC 20060201.
272
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
273
# use target default format.
274
result = a_bzrdir.create_repository()
275
# FIXME RBC 20060209 split out the repository type to avoid this check ?
276
elif isinstance(a_bzrdir._format,
277
(bzrdir.BzrDirFormat4,
278
bzrdir.BzrDirFormat5,
279
bzrdir.BzrDirFormat6)):
280
result = a_bzrdir.open_repository()
282
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
283
self.copy_content_into(result, revision_id, basis)
287
def has_revision(self, revision_id):
288
"""True if this repository has a copy of the revision."""
289
return self._revision_store.has_revision_id(revision_id,
290
self.get_transaction())
293
def get_revision_reconcile(self, revision_id):
294
"""'reconcile' helper routine that allows access to a revision always.
296
This variant of get_revision does not cross check the weave graph
297
against the revision one as get_revision does: but it should only
298
be used by reconcile, or reconcile-alike commands that are correcting
299
or testing the revision graph.
301
if not revision_id or not isinstance(revision_id, basestring):
302
raise InvalidRevisionId(revision_id=revision_id, branch=self)
303
return self._revision_store.get_revisions([revision_id],
304
self.get_transaction())[0]
306
def get_revisions(self, revision_ids):
307
return self._revision_store.get_revisions(revision_ids,
308
self.get_transaction())
311
def get_revision_xml(self, revision_id):
312
rev = self.get_revision(revision_id)
314
# the current serializer..
315
self._revision_store._serializer.write_revision(rev, rev_tmp)
317
return rev_tmp.getvalue()
320
def get_revision(self, revision_id):
321
"""Return the Revision object for a named revision"""
322
r = self.get_revision_reconcile(revision_id)
323
# weave corruption can lead to absent revision markers that should be
325
# the following test is reasonably cheap (it needs a single weave read)
326
# and the weave is cached in read transactions. In write transactions
327
# it is not cached but typically we only read a small number of
328
# revisions. For knits when they are introduced we will probably want
329
# to ensure that caching write transactions are in use.
330
inv = self.get_inventory_weave()
331
self._check_revision_parents(r, inv)
335
def get_deltas_for_revisions(self, revisions):
336
"""Produce a generator of revision deltas.
338
Note that the input is a sequence of REVISIONS, not revision_ids.
339
Trees will be held in memory until the generator exits.
340
Each delta is relative to the revision's lefthand predecessor.
342
required_trees = set()
343
for revision in revisions:
344
required_trees.add(revision.revision_id)
345
required_trees.update(revision.parent_ids[:1])
346
trees = dict((t.get_revision_id(), t) for
347
t in self.revision_trees(required_trees))
348
for revision in revisions:
349
if not revision.parent_ids:
350
old_tree = self.revision_tree(None)
352
old_tree = trees[revision.parent_ids[0]]
353
yield trees[revision.revision_id].changes_from(old_tree)
356
def get_revision_delta(self, revision_id):
357
"""Return the delta for one revision.
359
The delta is relative to the left-hand predecessor of the
362
r = self.get_revision(revision_id)
363
return list(self.get_deltas_for_revisions([r]))[0]
365
def _check_revision_parents(self, revision, inventory):
366
"""Private to Repository and Fetch.
368
This checks the parentage of revision in an inventory weave for
369
consistency and is only applicable to inventory-weave-for-ancestry
370
using repository formats & fetchers.
372
weave_parents = inventory.get_parents(revision.revision_id)
373
weave_names = inventory.versions()
374
for parent_id in revision.parent_ids:
375
if parent_id in weave_names:
376
# this parent must not be a ghost.
377
if not parent_id in weave_parents:
379
raise errors.CorruptRepository(self)
382
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
383
signature = gpg_strategy.sign(plaintext)
384
self._revision_store.add_revision_signature_text(revision_id,
386
self.get_transaction())
388
def fileids_altered_by_revision_ids(self, revision_ids):
389
"""Find the file ids and versions affected by revisions.
391
:param revisions: an iterable containing revision ids.
392
:return: a dictionary mapping altered file-ids to an iterable of
393
revision_ids. Each altered file-ids has the exact revision_ids that
394
altered it listed explicitly.
396
assert isinstance(self._format, (RepositoryFormat5,
399
RepositoryFormatKnit1)), \
400
("fileids_altered_by_revision_ids only supported for branches "
401
"which store inventory as unnested xml, not on %r" % self)
402
selected_revision_ids = set(revision_ids)
403
w = self.get_inventory_weave()
406
# this code needs to read every new line in every inventory for the
407
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
408
# not present in one of those inventories is unnecessary but not
409
# harmful because we are filtering by the revision id marker in the
410
# inventory lines : we only select file ids altered in one of those
411
# revisions. We don't need to see all lines in the inventory because
412
# only those added in an inventory in rev X can contain a revision=X
414
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
415
start = line.find('file_id="')+9
416
if start < 9: continue
417
end = line.find('"', start)
419
file_id = _unescape_xml(line[start:end])
421
start = line.find('revision="')+10
422
if start < 10: continue
423
end = line.find('"', start)
425
revision_id = _unescape_xml(line[start:end])
426
if revision_id in selected_revision_ids:
427
result.setdefault(file_id, set()).add(revision_id)
431
def get_inventory_weave(self):
432
return self.control_weaves.get_weave('inventory',
433
self.get_transaction())
436
def get_inventory(self, revision_id):
437
"""Get Inventory object by hash."""
438
return self.deserialise_inventory(
439
revision_id, self.get_inventory_xml(revision_id))
441
def deserialise_inventory(self, revision_id, xml):
442
"""Transform the xml into an inventory object.
444
:param revision_id: The expected revision id of the inventory.
445
:param xml: A serialised inventory.
447
result = xml5.serializer_v5.read_inventory_from_string(xml)
448
result.root.revision = revision_id
452
def get_inventory_xml(self, revision_id):
453
"""Get inventory XML as a file object."""
455
assert isinstance(revision_id, basestring), type(revision_id)
456
iw = self.get_inventory_weave()
457
return iw.get_text(revision_id)
459
raise errors.HistoryMissing(self, 'inventory', revision_id)
462
def get_inventory_sha1(self, revision_id):
463
"""Return the sha1 hash of the inventory entry
465
return self.get_revision(revision_id).inventory_sha1
468
def get_revision_graph(self, revision_id=None):
469
"""Return a dictionary containing the revision graph.
471
:param revision_id: The revision_id to get a graph from. If None, then
472
the entire revision graph is returned. This is a deprecated mode of
473
operation and will be removed in the future.
474
:return: a dictionary of revision_id->revision_parents_list.
476
# special case NULL_REVISION
477
if revision_id == NULL_REVISION:
479
weave = self.get_inventory_weave()
480
all_revisions = self._eliminate_revisions_not_present(weave.versions())
481
entire_graph = dict([(node, weave.get_parents(node)) for
482
node in all_revisions])
483
if revision_id is None:
485
elif revision_id not in entire_graph:
486
raise errors.NoSuchRevision(self, revision_id)
488
# add what can be reached from revision_id
490
pending = set([revision_id])
491
while len(pending) > 0:
493
result[node] = entire_graph[node]
494
for revision_id in result[node]:
495
if revision_id not in result:
496
pending.add(revision_id)
500
def get_revision_graph_with_ghosts(self, revision_ids=None):
501
"""Return a graph of the revisions with ghosts marked as applicable.
503
:param revision_ids: an iterable of revisions to graph or None for all.
504
:return: a Graph object with the graph reachable from revision_ids.
508
pending = set(self.all_revision_ids())
511
pending = set(revision_ids)
512
# special case NULL_REVISION
513
if NULL_REVISION in pending:
514
pending.remove(NULL_REVISION)
515
required = set(pending)
518
revision_id = pending.pop()
520
rev = self.get_revision(revision_id)
521
except errors.NoSuchRevision:
522
if revision_id in required:
525
result.add_ghost(revision_id)
527
for parent_id in rev.parent_ids:
528
# is this queued or done ?
529
if (parent_id not in pending and
530
parent_id not in done):
532
pending.add(parent_id)
533
result.add_node(revision_id, rev.parent_ids)
534
done.add(revision_id)
538
def get_revision_inventory(self, revision_id):
539
"""Return inventory of a past revision."""
540
# TODO: Unify this with get_inventory()
541
# bzr 0.0.6 and later imposes the constraint that the inventory_id
542
# must be the same as its revision, so this is trivial.
543
if revision_id is None:
544
# This does not make sense: if there is no revision,
545
# then it is the current tree inventory surely ?!
546
# and thus get_root_id() is something that looks at the last
547
# commit on the branch, and the get_root_id is an inventory check.
548
raise NotImplementedError
549
# return Inventory(self.get_root_id())
551
return self.get_inventory(revision_id)
555
"""Return True if this repository is flagged as a shared repository."""
556
raise NotImplementedError(self.is_shared)
559
def reconcile(self, other=None, thorough=False):
560
"""Reconcile this repository."""
561
from bzrlib.reconcile import RepoReconciler
562
reconciler = RepoReconciler(self, thorough=thorough)
563
reconciler.reconcile()
567
def revision_tree(self, revision_id):
568
"""Return Tree for a revision on this branch.
570
`revision_id` may be None for the empty tree revision.
572
# TODO: refactor this to use an existing revision object
573
# so we don't need to read it in twice.
574
if revision_id is None or revision_id == NULL_REVISION:
575
return RevisionTree(self, Inventory(), NULL_REVISION)
577
inv = self.get_revision_inventory(revision_id)
578
return RevisionTree(self, inv, revision_id)
581
def revision_trees(self, revision_ids):
582
"""Return Tree for a revision on this branch.
584
`revision_id` may not be None or 'null:'"""
585
assert None not in revision_ids
586
assert NULL_REVISION not in revision_ids
587
texts = self.get_inventory_weave().get_texts(revision_ids)
588
for text, revision_id in zip(texts, revision_ids):
589
inv = self.deserialise_inventory(revision_id, text)
590
yield RevisionTree(self, inv, revision_id)
593
def get_ancestry(self, revision_id):
594
"""Return a list of revision-ids integrated by a revision.
596
The first element of the list is always None, indicating the origin
597
revision. This might change when we have history horizons, or
598
perhaps we should have a new API.
600
This is topologically sorted.
602
if revision_id is None:
604
if not self.has_revision(revision_id):
605
raise errors.NoSuchRevision(self, revision_id)
606
w = self.get_inventory_weave()
607
candidates = w.get_ancestry(revision_id)
608
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
611
def print_file(self, file, revision_id):
612
"""Print `file` to stdout.
614
FIXME RBC 20060125 as John Meinel points out this is a bad api
615
- it writes to stdout, it assumes that that is valid etc. Fix
616
by creating a new more flexible convenience function.
618
tree = self.revision_tree(revision_id)
619
# use inventory as it was in that revision
620
file_id = tree.inventory.path2id(file)
622
# TODO: jam 20060427 Write a test for this code path
623
# it had a bug in it, and was raising the wrong
625
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
626
tree.print_file(file_id)
628
def get_transaction(self):
629
return self.control_files.get_transaction()
631
def revision_parents(self, revid):
632
return self.get_inventory_weave().parent_names(revid)
635
def set_make_working_trees(self, new_value):
636
"""Set the policy flag for making working trees when creating branches.
638
This only applies to branches that use this repository.
640
The default is 'True'.
641
:param new_value: True to restore the default, False to disable making
644
raise NotImplementedError(self.set_make_working_trees)
646
def make_working_trees(self):
647
"""Returns the policy for making working trees on new branches."""
648
raise NotImplementedError(self.make_working_trees)
651
def sign_revision(self, revision_id, gpg_strategy):
652
plaintext = Testament.from_revision(self, revision_id).as_short_text()
653
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
656
def has_signature_for_revision_id(self, revision_id):
657
"""Query for a revision signature for revision_id in the repository."""
658
return self._revision_store.has_signature(revision_id,
659
self.get_transaction())
662
def get_signature_text(self, revision_id):
663
"""Return the text for a signature."""
664
return self._revision_store.get_signature_text(revision_id,
665
self.get_transaction())
668
def check(self, revision_ids):
669
"""Check consistency of all history of given revision_ids.
671
Different repository implementations should override _check().
673
:param revision_ids: A non-empty list of revision_ids whose ancestry
674
will be checked. Typically the last revision_id of a branch.
677
raise ValueError("revision_ids must be non-empty in %s.check"
679
return self._check(revision_ids)
681
def _check(self, revision_ids):
682
result = check.Check(self)
687
class AllInOneRepository(Repository):
688
"""Legacy support - the repository behaviour for all-in-one branches."""
690
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
691
# we reuse one control files instance.
692
dir_mode = a_bzrdir._control_files._dir_mode
693
file_mode = a_bzrdir._control_files._file_mode
695
def get_store(name, compressed=True, prefixed=False):
696
# FIXME: This approach of assuming stores are all entirely compressed
697
# or entirely uncompressed is tidy, but breaks upgrade from
698
# some existing branches where there's a mixture; we probably
699
# still want the option to look for both.
700
relpath = a_bzrdir._control_files._escape(name)
701
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
702
prefixed=prefixed, compressed=compressed,
705
#if self._transport.should_cache():
706
# cache_path = os.path.join(self.cache_root, name)
707
# os.mkdir(cache_path)
708
# store = bzrlib.store.CachedStore(store, cache_path)
711
# not broken out yet because the controlweaves|inventory_store
712
# and text_store | weave_store bits are still different.
713
if isinstance(_format, RepositoryFormat4):
714
# cannot remove these - there is still no consistent api
715
# which allows access to this old info.
716
self.inventory_store = get_store('inventory-store')
717
text_store = get_store('text-store')
718
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
722
"""AllInOne repositories cannot be shared."""
726
def set_make_working_trees(self, new_value):
727
"""Set the policy flag for making working trees when creating branches.
729
This only applies to branches that use this repository.
731
The default is 'True'.
732
:param new_value: True to restore the default, False to disable making
735
raise NotImplementedError(self.set_make_working_trees)
737
def make_working_trees(self):
738
"""Returns the policy for making working trees on new branches."""
742
def install_revision(repository, rev, revision_tree):
743
"""Install all revision data into a repository."""
746
for p_id in rev.parent_ids:
747
if repository.has_revision(p_id):
748
present_parents.append(p_id)
749
parent_trees[p_id] = repository.revision_tree(p_id)
751
parent_trees[p_id] = repository.revision_tree(None)
753
inv = revision_tree.inventory
755
# backwards compatability hack: skip the root id.
756
entries = inv.iter_entries()
758
# Add the texts that are not already present
759
for path, ie in entries:
760
w = repository.weave_store.get_weave_or_empty(ie.file_id,
761
repository.get_transaction())
762
if ie.revision not in w:
764
# FIXME: TODO: The following loop *may* be overlapping/duplicate
765
# with InventoryEntry.find_previous_heads(). if it is, then there
766
# is a latent bug here where the parents may have ancestors of each
768
for revision, tree in parent_trees.iteritems():
769
if ie.file_id not in tree:
771
parent_id = tree.inventory[ie.file_id].revision
772
if parent_id in text_parents:
774
text_parents.append(parent_id)
776
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
777
repository.get_transaction())
778
lines = revision_tree.get_file(ie.file_id).readlines()
779
vfile.add_lines(rev.revision_id, text_parents, lines)
781
# install the inventory
782
repository.add_inventory(rev.revision_id, inv, present_parents)
783
except errors.RevisionAlreadyPresent:
785
repository.add_revision(rev.revision_id, rev, inv)
788
class MetaDirRepository(Repository):
789
"""Repositories in the new meta-dir layout."""
791
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
792
super(MetaDirRepository, self).__init__(_format,
799
dir_mode = self.control_files._dir_mode
800
file_mode = self.control_files._file_mode
804
"""Return True if this repository is flagged as a shared repository."""
805
return self.control_files._transport.has('shared-storage')
808
def set_make_working_trees(self, new_value):
809
"""Set the policy flag for making working trees when creating branches.
811
This only applies to branches that use this repository.
813
The default is 'True'.
814
:param new_value: True to restore the default, False to disable making
819
self.control_files._transport.delete('no-working-trees')
820
except errors.NoSuchFile:
823
self.control_files.put_utf8('no-working-trees', '')
825
def make_working_trees(self):
826
"""Returns the policy for making working trees on new branches."""
827
return not self.control_files._transport.has('no-working-trees')
830
class KnitRepository(MetaDirRepository):
831
"""Knit format repository."""
833
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
834
inv_vf.add_lines_with_ghosts(revid, parents, lines)
837
def _all_revision_ids(self):
838
"""See Repository.all_revision_ids()."""
839
# Knits get the revision graph from the index of the revision knit, so
840
# it's always possible even if they're on an unlistable transport.
841
return self._revision_store.all_revision_ids(self.get_transaction())
843
def fileid_involved_between_revs(self, from_revid, to_revid):
844
"""Find file_id(s) which are involved in the changes between revisions.
846
This determines the set of revisions which are involved, and then
847
finds all file ids affected by those revisions.
849
vf = self._get_revision_vf()
850
from_set = set(vf.get_ancestry(from_revid))
851
to_set = set(vf.get_ancestry(to_revid))
852
changed = to_set.difference(from_set)
853
return self._fileid_involved_by_set(changed)
855
def fileid_involved(self, last_revid=None):
856
"""Find all file_ids modified in the ancestry of last_revid.
858
:param last_revid: If None, last_revision() will be used.
861
changed = set(self.all_revision_ids())
863
changed = set(self.get_ancestry(last_revid))
866
return self._fileid_involved_by_set(changed)
869
def get_ancestry(self, revision_id):
870
"""Return a list of revision-ids integrated by a revision.
872
This is topologically sorted.
874
if revision_id is None:
876
vf = self._get_revision_vf()
878
return [None] + vf.get_ancestry(revision_id)
879
except errors.RevisionNotPresent:
880
raise errors.NoSuchRevision(self, revision_id)
883
def get_revision(self, revision_id):
884
"""Return the Revision object for a named revision"""
885
return self.get_revision_reconcile(revision_id)
888
def get_revision_graph(self, revision_id=None):
889
"""Return a dictionary containing the revision graph.
891
:param revision_id: The revision_id to get a graph from. If None, then
892
the entire revision graph is returned. This is a deprecated mode of
893
operation and will be removed in the future.
894
:return: a dictionary of revision_id->revision_parents_list.
896
# special case NULL_REVISION
897
if revision_id == NULL_REVISION:
899
weave = self._get_revision_vf()
900
entire_graph = weave.get_graph()
901
if revision_id is None:
902
return weave.get_graph()
903
elif revision_id not in weave:
904
raise errors.NoSuchRevision(self, revision_id)
906
# add what can be reached from revision_id
908
pending = set([revision_id])
909
while len(pending) > 0:
911
result[node] = weave.get_parents(node)
912
for revision_id in result[node]:
913
if revision_id not in result:
914
pending.add(revision_id)
918
def get_revision_graph_with_ghosts(self, revision_ids=None):
919
"""Return a graph of the revisions with ghosts marked as applicable.
921
:param revision_ids: an iterable of revisions to graph or None for all.
922
:return: a Graph object with the graph reachable from revision_ids.
925
vf = self._get_revision_vf()
926
versions = set(vf.versions())
928
pending = set(self.all_revision_ids())
931
pending = set(revision_ids)
932
# special case NULL_REVISION
933
if NULL_REVISION in pending:
934
pending.remove(NULL_REVISION)
935
required = set(pending)
938
revision_id = pending.pop()
939
if not revision_id in versions:
940
if revision_id in required:
941
raise errors.NoSuchRevision(self, revision_id)
943
result.add_ghost(revision_id)
944
# mark it as done so we don't try for it again.
945
done.add(revision_id)
947
parent_ids = vf.get_parents_with_ghosts(revision_id)
948
for parent_id in parent_ids:
949
# is this queued or done ?
950
if (parent_id not in pending and
951
parent_id not in done):
953
pending.add(parent_id)
954
result.add_node(revision_id, parent_ids)
955
done.add(revision_id)
958
def _get_revision_vf(self):
959
""":return: a versioned file containing the revisions."""
960
vf = self._revision_store.get_revision_file(self.get_transaction())
964
def reconcile(self, other=None, thorough=False):
965
"""Reconcile this repository."""
966
from bzrlib.reconcile import KnitReconciler
967
reconciler = KnitReconciler(self, thorough=thorough)
968
reconciler.reconcile()
971
def revision_parents(self, revision_id):
972
return self._get_revision_vf().get_parents(revision_id)
975
class RepositoryFormat(object):
976
"""A repository format.
978
Formats provide three things:
979
* An initialization routine to construct repository data on disk.
980
* a format string which is used when the BzrDir supports versioned
982
* an open routine which returns a Repository instance.
984
Formats are placed in an dict by their format string for reference
985
during opening. These should be subclasses of RepositoryFormat
988
Once a format is deprecated, just deprecate the initialize and open
989
methods on the format class. Do not deprecate the object, as the
990
object will be created every system load.
992
Common instance attributes:
993
_matchingbzrdir - the bzrdir format that the repository format was
994
originally written to work with. This can be used if manually
995
constructing a bzrdir and repository, or more commonly for test suite
999
_default_format = None
1000
"""The default format used for new repositories."""
1003
"""The known formats."""
1006
def find_format(klass, a_bzrdir):
1007
"""Return the format for the repository object in a_bzrdir."""
1009
transport = a_bzrdir.get_repository_transport(None)
1010
format_string = transport.get("format").read()
1011
return klass._formats[format_string]
1012
except errors.NoSuchFile:
1013
raise errors.NoRepositoryPresent(a_bzrdir)
1015
raise errors.UnknownFormatError(format=format_string)
1017
def _get_control_store(self, repo_transport, control_files):
1018
"""Return the control store for this repository."""
1019
raise NotImplementedError(self._get_control_store)
1022
def get_default_format(klass):
1023
"""Return the current default format."""
1024
return klass._default_format
1026
def get_format_string(self):
1027
"""Return the ASCII format string that identifies this format.
1029
Note that in pre format ?? repositories the format string is
1030
not permitted nor written to disk.
1032
raise NotImplementedError(self.get_format_string)
1034
def get_format_description(self):
1035
"""Return the short description for this format."""
1036
raise NotImplementedError(self.get_format_description)
1038
def _get_revision_store(self, repo_transport, control_files):
1039
"""Return the revision store object for this a_bzrdir."""
1040
raise NotImplementedError(self._get_revision_store)
1042
def _get_text_rev_store(self,
1049
"""Common logic for getting a revision store for a repository.
1051
see self._get_revision_store for the subclass-overridable method to
1052
get the store for a repository.
1054
from bzrlib.store.revision.text import TextRevisionStore
1055
dir_mode = control_files._dir_mode
1056
file_mode = control_files._file_mode
1057
text_store =TextStore(transport.clone(name),
1059
compressed=compressed,
1061
file_mode=file_mode)
1062
_revision_store = TextRevisionStore(text_store, serializer)
1063
return _revision_store
1065
def _get_versioned_file_store(self,
1070
versionedfile_class=WeaveFile,
1072
weave_transport = control_files._transport.clone(name)
1073
dir_mode = control_files._dir_mode
1074
file_mode = control_files._file_mode
1075
return VersionedFileStore(weave_transport, prefixed=prefixed,
1077
file_mode=file_mode,
1078
versionedfile_class=versionedfile_class,
1081
def initialize(self, a_bzrdir, shared=False):
1082
"""Initialize a repository of this format in a_bzrdir.
1084
:param a_bzrdir: The bzrdir to put the new repository in it.
1085
:param shared: The repository should be initialized as a sharable one.
1087
This may raise UninitializableFormat if shared repository are not
1088
compatible the a_bzrdir.
1091
def is_supported(self):
1092
"""Is this format supported?
1094
Supported formats must be initializable and openable.
1095
Unsupported formats may not support initialization or committing or
1096
some other features depending on the reason for not being supported.
1100
def open(self, a_bzrdir, _found=False):
1101
"""Return an instance of this format for the bzrdir a_bzrdir.
1103
_found is a private parameter, do not use it.
1105
raise NotImplementedError(self.open)
1108
def register_format(klass, format):
1109
klass._formats[format.get_format_string()] = format
1112
def set_default_format(klass, format):
1113
klass._default_format = format
1116
def unregister_format(klass, format):
1117
assert klass._formats[format.get_format_string()] is format
1118
del klass._formats[format.get_format_string()]
1121
class PreSplitOutRepositoryFormat(RepositoryFormat):
1122
"""Base class for the pre split out repository formats."""
1124
def initialize(self, a_bzrdir, shared=False, _internal=False):
1125
"""Create a weave repository.
1127
TODO: when creating split out bzr branch formats, move this to a common
1128
base for Format5, Format6. or something like that.
1130
from bzrlib.weavefile import write_weave_v5
1131
from bzrlib.weave import Weave
1134
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1137
# always initialized when the bzrdir is.
1138
return self.open(a_bzrdir, _found=True)
1140
# Create an empty weave
1142
write_weave_v5(Weave(), sio)
1143
empty_weave = sio.getvalue()
1145
mutter('creating repository in %s.', a_bzrdir.transport.base)
1146
dirs = ['revision-store', 'weaves']
1147
files = [('inventory.weave', StringIO(empty_weave)),
1150
# FIXME: RBC 20060125 don't peek under the covers
1151
# NB: no need to escape relative paths that are url safe.
1152
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1154
control_files.create_lock()
1155
control_files.lock_write()
1156
control_files._transport.mkdir_multi(dirs,
1157
mode=control_files._dir_mode)
1159
for file, content in files:
1160
control_files.put(file, content)
1162
control_files.unlock()
1163
return self.open(a_bzrdir, _found=True)
1165
def _get_control_store(self, repo_transport, control_files):
1166
"""Return the control store for this repository."""
1167
return self._get_versioned_file_store('',
1172
def _get_text_store(self, transport, control_files):
1173
"""Get a store for file texts for this format."""
1174
raise NotImplementedError(self._get_text_store)
1176
def open(self, a_bzrdir, _found=False):
1177
"""See RepositoryFormat.open()."""
1179
# we are being called directly and must probe.
1180
raise NotImplementedError
1182
repo_transport = a_bzrdir.get_repository_transport(None)
1183
control_files = a_bzrdir._control_files
1184
text_store = self._get_text_store(repo_transport, control_files)
1185
control_store = self._get_control_store(repo_transport, control_files)
1186
_revision_store = self._get_revision_store(repo_transport, control_files)
1187
return AllInOneRepository(_format=self,
1189
_revision_store=_revision_store,
1190
control_store=control_store,
1191
text_store=text_store)
1194
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1195
"""Bzr repository format 4.
1197
This repository format has:
1199
- TextStores for texts, inventories,revisions.
1201
This format is deprecated: it indexes texts using a text id which is
1202
removed in format 5; initialization and write support for this format
1207
super(RepositoryFormat4, self).__init__()
1208
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1210
def get_format_description(self):
1211
"""See RepositoryFormat.get_format_description()."""
1212
return "Repository format 4"
1214
def initialize(self, url, shared=False, _internal=False):
1215
"""Format 4 branches cannot be created."""
1216
raise errors.UninitializableFormat(self)
1218
def is_supported(self):
1219
"""Format 4 is not supported.
1221
It is not supported because the model changed from 4 to 5 and the
1222
conversion logic is expensive - so doing it on the fly was not
1227
def _get_control_store(self, repo_transport, control_files):
1228
"""Format 4 repositories have no formal control store at this point.
1230
This will cause any control-file-needing apis to fail - this is desired.
1234
def _get_revision_store(self, repo_transport, control_files):
1235
"""See RepositoryFormat._get_revision_store()."""
1236
from bzrlib.xml4 import serializer_v4
1237
return self._get_text_rev_store(repo_transport,
1240
serializer=serializer_v4)
1242
def _get_text_store(self, transport, control_files):
1243
"""See RepositoryFormat._get_text_store()."""
1246
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1247
"""Bzr control format 5.
1249
This repository format has:
1250
- weaves for file texts and inventory
1252
- TextStores for revisions and signatures.
1256
super(RepositoryFormat5, self).__init__()
1257
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1259
def get_format_description(self):
1260
"""See RepositoryFormat.get_format_description()."""
1261
return "Weave repository format 5"
1263
def _get_revision_store(self, repo_transport, control_files):
1264
"""See RepositoryFormat._get_revision_store()."""
1265
"""Return the revision store object for this a_bzrdir."""
1266
return self._get_text_rev_store(repo_transport,
1271
def _get_text_store(self, transport, control_files):
1272
"""See RepositoryFormat._get_text_store()."""
1273
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1276
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1277
"""Bzr control format 6.
1279
This repository format has:
1280
- weaves for file texts and inventory
1281
- hash subdirectory based stores.
1282
- TextStores for revisions and signatures.
1286
super(RepositoryFormat6, self).__init__()
1287
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1289
def get_format_description(self):
1290
"""See RepositoryFormat.get_format_description()."""
1291
return "Weave repository format 6"
1293
def _get_revision_store(self, repo_transport, control_files):
1294
"""See RepositoryFormat._get_revision_store()."""
1295
return self._get_text_rev_store(repo_transport,
1301
def _get_text_store(self, transport, control_files):
1302
"""See RepositoryFormat._get_text_store()."""
1303
return self._get_versioned_file_store('weaves', transport, control_files)
1306
class MetaDirRepositoryFormat(RepositoryFormat):
1307
"""Common base class for the new repositories using the metadir layout."""
1310
super(MetaDirRepositoryFormat, self).__init__()
1311
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1313
def _create_control_files(self, a_bzrdir):
1314
"""Create the required files and the initial control_files object."""
1315
# FIXME: RBC 20060125 don't peek under the covers
1316
# NB: no need to escape relative paths that are url safe.
1317
repository_transport = a_bzrdir.get_repository_transport(self)
1318
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1319
control_files.create_lock()
1320
return control_files
1322
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1323
"""Upload the initial blank content."""
1324
control_files = self._create_control_files(a_bzrdir)
1325
control_files.lock_write()
1327
control_files._transport.mkdir_multi(dirs,
1328
mode=control_files._dir_mode)
1329
for file, content in files:
1330
control_files.put(file, content)
1331
for file, content in utf8_files:
1332
control_files.put_utf8(file, content)
1334
control_files.put_utf8('shared-storage', '')
1336
control_files.unlock()
1339
class RepositoryFormat7(MetaDirRepositoryFormat):
1340
"""Bzr repository 7.
1342
This repository format has:
1343
- weaves for file texts and inventory
1344
- hash subdirectory based stores.
1345
- TextStores for revisions and signatures.
1346
- a format marker of its own
1347
- an optional 'shared-storage' flag
1348
- an optional 'no-working-trees' flag
1351
def _get_control_store(self, repo_transport, control_files):
1352
"""Return the control store for this repository."""
1353
return self._get_versioned_file_store('',
1358
def get_format_string(self):
1359
"""See RepositoryFormat.get_format_string()."""
1360
return "Bazaar-NG Repository format 7"
1362
def get_format_description(self):
1363
"""See RepositoryFormat.get_format_description()."""
1364
return "Weave repository format 7"
1366
def _get_revision_store(self, repo_transport, control_files):
1367
"""See RepositoryFormat._get_revision_store()."""
1368
return self._get_text_rev_store(repo_transport,
1375
def _get_text_store(self, transport, control_files):
1376
"""See RepositoryFormat._get_text_store()."""
1377
return self._get_versioned_file_store('weaves',
1381
def initialize(self, a_bzrdir, shared=False):
1382
"""Create a weave repository.
1384
:param shared: If true the repository will be initialized as a shared
1387
from bzrlib.weavefile import write_weave_v5
1388
from bzrlib.weave import Weave
1390
# Create an empty weave
1392
write_weave_v5(Weave(), sio)
1393
empty_weave = sio.getvalue()
1395
mutter('creating repository in %s.', a_bzrdir.transport.base)
1396
dirs = ['revision-store', 'weaves']
1397
files = [('inventory.weave', StringIO(empty_weave)),
1399
utf8_files = [('format', self.get_format_string())]
1401
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1402
return self.open(a_bzrdir=a_bzrdir, _found=True)
1404
def open(self, a_bzrdir, _found=False, _override_transport=None):
1405
"""See RepositoryFormat.open().
1407
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1408
repository at a slightly different url
1409
than normal. I.e. during 'upgrade'.
1412
format = RepositoryFormat.find_format(a_bzrdir)
1413
assert format.__class__ == self.__class__
1414
if _override_transport is not None:
1415
repo_transport = _override_transport
1417
repo_transport = a_bzrdir.get_repository_transport(None)
1418
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1419
text_store = self._get_text_store(repo_transport, control_files)
1420
control_store = self._get_control_store(repo_transport, control_files)
1421
_revision_store = self._get_revision_store(repo_transport, control_files)
1422
return MetaDirRepository(_format=self,
1424
control_files=control_files,
1425
_revision_store=_revision_store,
1426
control_store=control_store,
1427
text_store=text_store)
1430
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1431
"""Bzr repository knit format 1.
1433
This repository format has:
1434
- knits for file texts and inventory
1435
- hash subdirectory based stores.
1436
- knits for revisions and signatures
1437
- TextStores for revisions and signatures.
1438
- a format marker of its own
1439
- an optional 'shared-storage' flag
1440
- an optional 'no-working-trees' flag
1443
This format was introduced in bzr 0.8.
1446
def _get_control_store(self, repo_transport, control_files):
1447
"""Return the control store for this repository."""
1448
return VersionedFileStore(
1451
file_mode=control_files._file_mode,
1452
versionedfile_class=KnitVersionedFile,
1453
versionedfile_kwargs={'factory':KnitPlainFactory()},
1456
def get_format_string(self):
1457
"""See RepositoryFormat.get_format_string()."""
1458
return "Bazaar-NG Knit Repository Format 1"
1460
def get_format_description(self):
1461
"""See RepositoryFormat.get_format_description()."""
1462
return "Knit repository format 1"
1464
def _get_revision_store(self, repo_transport, control_files):
1465
"""See RepositoryFormat._get_revision_store()."""
1466
from bzrlib.store.revision.knit import KnitRevisionStore
1467
versioned_file_store = VersionedFileStore(
1469
file_mode=control_files._file_mode,
1472
versionedfile_class=KnitVersionedFile,
1473
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1476
return KnitRevisionStore(versioned_file_store)
1478
def _get_text_store(self, transport, control_files):
1479
"""See RepositoryFormat._get_text_store()."""
1480
return self._get_versioned_file_store('knits',
1483
versionedfile_class=KnitVersionedFile,
1486
def initialize(self, a_bzrdir, shared=False):
1487
"""Create a knit format 1 repository.
1489
:param a_bzrdir: bzrdir to contain the new repository; must already
1491
:param shared: If true the repository will be initialized as a shared
1494
mutter('creating repository in %s.', a_bzrdir.transport.base)
1495
dirs = ['revision-store', 'knits']
1497
utf8_files = [('format', self.get_format_string())]
1499
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1500
repo_transport = a_bzrdir.get_repository_transport(None)
1501
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1502
control_store = self._get_control_store(repo_transport, control_files)
1503
transaction = transactions.WriteTransaction()
1504
# trigger a write of the inventory store.
1505
control_store.get_weave_or_empty('inventory', transaction)
1506
_revision_store = self._get_revision_store(repo_transport, control_files)
1507
_revision_store.has_revision_id('A', transaction)
1508
_revision_store.get_signature_file(transaction)
1509
return self.open(a_bzrdir=a_bzrdir, _found=True)
1511
def open(self, a_bzrdir, _found=False, _override_transport=None):
1512
"""See RepositoryFormat.open().
1514
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1515
repository at a slightly different url
1516
than normal. I.e. during 'upgrade'.
1519
format = RepositoryFormat.find_format(a_bzrdir)
1520
assert format.__class__ == self.__class__
1521
if _override_transport is not None:
1522
repo_transport = _override_transport
1524
repo_transport = a_bzrdir.get_repository_transport(None)
1525
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1526
text_store = self._get_text_store(repo_transport, control_files)
1527
control_store = self._get_control_store(repo_transport, control_files)
1528
_revision_store = self._get_revision_store(repo_transport, control_files)
1529
return KnitRepository(_format=self,
1531
control_files=control_files,
1532
_revision_store=_revision_store,
1533
control_store=control_store,
1534
text_store=text_store)
1537
# formats which have no format string are not discoverable
1538
# and not independently creatable, so are not registered.
1539
RepositoryFormat.register_format(RepositoryFormat7())
1540
_default_format = RepositoryFormatKnit1()
1541
RepositoryFormat.register_format(_default_format)
1542
RepositoryFormat.set_default_format(_default_format)
1543
_legacy_formats = [RepositoryFormat4(),
1544
RepositoryFormat5(),
1545
RepositoryFormat6()]
1548
class InterRepository(InterObject):
1549
"""This class represents operations taking place between two repositories.
1551
Its instances have methods like copy_content and fetch, and contain
1552
references to the source and target repositories these operations can be
1555
Often we will provide convenience methods on 'repository' which carry out
1556
operations with another repository - they will always forward to
1557
InterRepository.get(other).method_name(parameters).
1561
"""The available optimised InterRepository types."""
1564
def copy_content(self, revision_id=None, basis=None):
1565
"""Make a complete copy of the content in self into destination.
1567
This is a destructive operation! Do not use it on existing
1570
:param revision_id: Only copy the content needed to construct
1571
revision_id and its parents.
1572
:param basis: Copy the needed data preferentially from basis.
1575
self.target.set_make_working_trees(self.source.make_working_trees())
1576
except NotImplementedError:
1578
# grab the basis available data
1579
if basis is not None:
1580
self.target.fetch(basis, revision_id=revision_id)
1581
# but don't bother fetching if we have the needed data now.
1582
if (revision_id not in (None, NULL_REVISION) and
1583
self.target.has_revision(revision_id)):
1585
self.target.fetch(self.source, revision_id=revision_id)
1588
def fetch(self, revision_id=None, pb=None):
1589
"""Fetch the content required to construct revision_id.
1591
The content is copied from source to target.
1593
:param revision_id: if None all content is copied, if NULL_REVISION no
1595
:param pb: optional progress bar to use for progress reports. If not
1596
provided a default one will be created.
1598
Returns the copied revision count and the failed revisions in a tuple:
1601
from bzrlib.fetch import GenericRepoFetcher
1602
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1603
self.source, self.source._format, self.target, self.target._format)
1604
f = GenericRepoFetcher(to_repository=self.target,
1605
from_repository=self.source,
1606
last_revision=revision_id,
1608
return f.count_copied, f.failed_revisions
1611
def missing_revision_ids(self, revision_id=None):
1612
"""Return the revision ids that source has that target does not.
1614
These are returned in topological order.
1616
:param revision_id: only return revision ids included by this
1619
# generic, possibly worst case, slow code path.
1620
target_ids = set(self.target.all_revision_ids())
1621
if revision_id is not None:
1622
source_ids = self.source.get_ancestry(revision_id)
1623
assert source_ids[0] == None
1626
source_ids = self.source.all_revision_ids()
1627
result_set = set(source_ids).difference(target_ids)
1628
# this may look like a no-op: its not. It preserves the ordering
1629
# other_ids had while only returning the members from other_ids
1630
# that we've decided we need.
1631
return [rev_id for rev_id in source_ids if rev_id in result_set]
1634
class InterWeaveRepo(InterRepository):
1635
"""Optimised code paths between Weave based repositories."""
1637
_matching_repo_format = RepositoryFormat7()
1638
"""Repository format for testing with."""
1641
def is_compatible(source, target):
1642
"""Be compatible with known Weave formats.
1644
We don't test for the stores being of specific types because that
1645
could lead to confusing results, and there is no need to be
1649
return (isinstance(source._format, (RepositoryFormat5,
1651
RepositoryFormat7)) and
1652
isinstance(target._format, (RepositoryFormat5,
1654
RepositoryFormat7)))
1655
except AttributeError:
1659
def copy_content(self, revision_id=None, basis=None):
1660
"""See InterRepository.copy_content()."""
1661
# weave specific optimised path:
1662
if basis is not None:
1663
# copy the basis in, then fetch remaining data.
1664
basis.copy_content_into(self.target, revision_id)
1665
# the basis copy_content_into could miss-set this.
1667
self.target.set_make_working_trees(self.source.make_working_trees())
1668
except NotImplementedError:
1670
self.target.fetch(self.source, revision_id=revision_id)
1673
self.target.set_make_working_trees(self.source.make_working_trees())
1674
except NotImplementedError:
1676
# FIXME do not peek!
1677
if self.source.control_files._transport.listable():
1678
pb = ui.ui_factory.nested_progress_bar()
1680
self.target.weave_store.copy_all_ids(
1681
self.source.weave_store,
1683
from_transaction=self.source.get_transaction(),
1684
to_transaction=self.target.get_transaction())
1685
pb.update('copying inventory', 0, 1)
1686
self.target.control_weaves.copy_multi(
1687
self.source.control_weaves, ['inventory'],
1688
from_transaction=self.source.get_transaction(),
1689
to_transaction=self.target.get_transaction())
1690
self.target._revision_store.text_store.copy_all_ids(
1691
self.source._revision_store.text_store,
1696
self.target.fetch(self.source, revision_id=revision_id)
1699
def fetch(self, revision_id=None, pb=None):
1700
"""See InterRepository.fetch()."""
1701
from bzrlib.fetch import GenericRepoFetcher
1702
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1703
self.source, self.source._format, self.target, self.target._format)
1704
f = GenericRepoFetcher(to_repository=self.target,
1705
from_repository=self.source,
1706
last_revision=revision_id,
1708
return f.count_copied, f.failed_revisions
1711
def missing_revision_ids(self, revision_id=None):
1712
"""See InterRepository.missing_revision_ids()."""
1713
# we want all revisions to satisfy revision_id in source.
1714
# but we don't want to stat every file here and there.
1715
# we want then, all revisions other needs to satisfy revision_id
1716
# checked, but not those that we have locally.
1717
# so the first thing is to get a subset of the revisions to
1718
# satisfy revision_id in source, and then eliminate those that
1719
# we do already have.
1720
# this is slow on high latency connection to self, but as as this
1721
# disk format scales terribly for push anyway due to rewriting
1722
# inventory.weave, this is considered acceptable.
1724
if revision_id is not None:
1725
source_ids = self.source.get_ancestry(revision_id)
1726
assert source_ids[0] == None
1729
source_ids = self.source._all_possible_ids()
1730
source_ids_set = set(source_ids)
1731
# source_ids is the worst possible case we may need to pull.
1732
# now we want to filter source_ids against what we actually
1733
# have in target, but don't try to check for existence where we know
1734
# we do not have a revision as that would be pointless.
1735
target_ids = set(self.target._all_possible_ids())
1736
possibly_present_revisions = target_ids.intersection(source_ids_set)
1737
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1738
required_revisions = source_ids_set.difference(actually_present_revisions)
1739
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1740
if revision_id is not None:
1741
# we used get_ancestry to determine source_ids then we are assured all
1742
# revisions referenced are present as they are installed in topological order.
1743
# and the tip revision was validated by get_ancestry.
1744
return required_topo_revisions
1746
# if we just grabbed the possibly available ids, then
1747
# we only have an estimate of whats available and need to validate
1748
# that against the revision records.
1749
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1752
class InterKnitRepo(InterRepository):
1753
"""Optimised code paths between Knit based repositories."""
1755
_matching_repo_format = RepositoryFormatKnit1()
1756
"""Repository format for testing with."""
1759
def is_compatible(source, target):
1760
"""Be compatible with known Knit formats.
1762
We don't test for the stores being of specific types because that
1763
could lead to confusing results, and there is no need to be
1767
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1768
isinstance(target._format, (RepositoryFormatKnit1)))
1769
except AttributeError:
1773
def fetch(self, revision_id=None, pb=None):
1774
"""See InterRepository.fetch()."""
1775
from bzrlib.fetch import KnitRepoFetcher
1776
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1777
self.source, self.source._format, self.target, self.target._format)
1778
f = KnitRepoFetcher(to_repository=self.target,
1779
from_repository=self.source,
1780
last_revision=revision_id,
1782
return f.count_copied, f.failed_revisions
1785
def missing_revision_ids(self, revision_id=None):
1786
"""See InterRepository.missing_revision_ids()."""
1787
if revision_id is not None:
1788
source_ids = self.source.get_ancestry(revision_id)
1789
assert source_ids[0] == None
1792
source_ids = self.source._all_possible_ids()
1793
source_ids_set = set(source_ids)
1794
# source_ids is the worst possible case we may need to pull.
1795
# now we want to filter source_ids against what we actually
1796
# have in target, but don't try to check for existence where we know
1797
# we do not have a revision as that would be pointless.
1798
target_ids = set(self.target._all_possible_ids())
1799
possibly_present_revisions = target_ids.intersection(source_ids_set)
1800
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1801
required_revisions = source_ids_set.difference(actually_present_revisions)
1802
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1803
if revision_id is not None:
1804
# we used get_ancestry to determine source_ids then we are assured all
1805
# revisions referenced are present as they are installed in topological order.
1806
# and the tip revision was validated by get_ancestry.
1807
return required_topo_revisions
1809
# if we just grabbed the possibly available ids, then
1810
# we only have an estimate of whats available and need to validate
1811
# that against the revision records.
1812
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1814
InterRepository.register_optimiser(InterWeaveRepo)
1815
InterRepository.register_optimiser(InterKnitRepo)
1818
class RepositoryTestProviderAdapter(object):
1819
"""A tool to generate a suite testing multiple repository formats at once.
1821
This is done by copying the test once for each transport and injecting
1822
the transport_server, transport_readonly_server, and bzrdir_format and
1823
repository_format classes into each copy. Each copy is also given a new id()
1824
to make it easy to identify.
1827
def __init__(self, transport_server, transport_readonly_server, formats):
1828
self._transport_server = transport_server
1829
self._transport_readonly_server = transport_readonly_server
1830
self._formats = formats
1832
def adapt(self, test):
1833
result = TestSuite()
1834
for repository_format, bzrdir_format in self._formats:
1835
new_test = deepcopy(test)
1836
new_test.transport_server = self._transport_server
1837
new_test.transport_readonly_server = self._transport_readonly_server
1838
new_test.bzrdir_format = bzrdir_format
1839
new_test.repository_format = repository_format
1840
def make_new_test_id():
1841
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1842
return lambda: new_id
1843
new_test.id = make_new_test_id()
1844
result.addTest(new_test)
1848
class InterRepositoryTestProviderAdapter(object):
1849
"""A tool to generate a suite testing multiple inter repository formats.
1851
This is done by copying the test once for each interrepo provider and injecting
1852
the transport_server, transport_readonly_server, repository_format and
1853
repository_to_format classes into each copy.
1854
Each copy is also given a new id() to make it easy to identify.
1857
def __init__(self, transport_server, transport_readonly_server, formats):
1858
self._transport_server = transport_server
1859
self._transport_readonly_server = transport_readonly_server
1860
self._formats = formats
1862
def adapt(self, test):
1863
result = TestSuite()
1864
for interrepo_class, repository_format, repository_format_to in self._formats:
1865
new_test = deepcopy(test)
1866
new_test.transport_server = self._transport_server
1867
new_test.transport_readonly_server = self._transport_readonly_server
1868
new_test.interrepo_class = interrepo_class
1869
new_test.repository_format = repository_format
1870
new_test.repository_format_to = repository_format_to
1871
def make_new_test_id():
1872
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1873
return lambda: new_id
1874
new_test.id = make_new_test_id()
1875
result.addTest(new_test)
1879
def default_test_list():
1880
"""Generate the default list of interrepo permutations to test."""
1882
# test the default InterRepository between format 6 and the current
1884
# XXX: robertc 20060220 reinstate this when there are two supported
1885
# formats which do not have an optimal code path between them.
1886
result.append((InterRepository,
1887
RepositoryFormat6(),
1888
RepositoryFormatKnit1()))
1889
for optimiser in InterRepository._optimisers:
1890
result.append((optimiser,
1891
optimiser._matching_repo_format,
1892
optimiser._matching_repo_format
1894
# if there are specific combinations we want to use, we can add them
1899
class CopyConverter(object):
1900
"""A repository conversion tool which just performs a copy of the content.
1902
This is slow but quite reliable.
1905
def __init__(self, target_format):
1906
"""Create a CopyConverter.
1908
:param target_format: The format the resulting repository should be.
1910
self.target_format = target_format
1912
def convert(self, repo, pb):
1913
"""Perform the conversion of to_convert, giving feedback via pb.
1915
:param to_convert: The disk object to convert.
1916
:param pb: a progress bar to use for progress information.
1921
# this is only useful with metadir layouts - separated repo content.
1922
# trigger an assertion if not such
1923
repo._format.get_format_string()
1924
self.repo_dir = repo.bzrdir
1925
self.step('Moving repository to repository.backup')
1926
self.repo_dir.transport.move('repository', 'repository.backup')
1927
backup_transport = self.repo_dir.transport.clone('repository.backup')
1928
self.source_repo = repo._format.open(self.repo_dir,
1930
_override_transport=backup_transport)
1931
self.step('Creating new repository')
1932
converted = self.target_format.initialize(self.repo_dir,
1933
self.source_repo.is_shared())
1934
converted.lock_write()
1936
self.step('Copying content into repository.')
1937
self.source_repo.copy_content_into(converted)
1940
self.step('Deleting old repository content.')
1941
self.repo_dir.transport.delete_tree('repository.backup')
1942
self.pb.note('repository converted')
1944
def step(self, message):
1945
"""Update the pb by a step."""
1947
self.pb.update(message, self.count, self.total)
1950
class CommitBuilder(object):
1951
"""Provides an interface to build up a commit.
1953
This allows describing a tree to be committed without needing to
1954
know the internals of the format of the repository.
1956
def __init__(self, repository, parents, config, timestamp=None,
1957
timezone=None, committer=None, revprops=None,
1959
"""Initiate a CommitBuilder.
1961
:param repository: Repository to commit to.
1962
:param parents: Revision ids of the parents of the new revision.
1963
:param config: Configuration to use.
1964
:param timestamp: Optional timestamp recorded for commit.
1965
:param timezone: Optional timezone for timestamp.
1966
:param committer: Optional committer to set for commit.
1967
:param revprops: Optional dictionary of revision properties.
1968
:param revision_id: Optional revision id.
1970
self._config = config
1972
if committer is None:
1973
self._committer = self._config.username()
1975
assert isinstance(committer, basestring), type(committer)
1976
self._committer = committer
1978
self.new_inventory = Inventory(None)
1979
self._new_revision_id = revision_id
1980
self.parents = parents
1981
self.repository = repository
1984
if revprops is not None:
1985
self._revprops.update(revprops)
1987
if timestamp is None:
1988
timestamp = time.time()
1989
# Restrict resolution to 1ms
1990
self._timestamp = round(timestamp, 3)
1992
if timezone is None:
1993
self._timezone = local_time_offset()
1995
self._timezone = int(timezone)
1997
self._generate_revision_if_needed()
1999
def commit(self, message):
2000
"""Make the actual commit.
2002
:return: The revision id of the recorded revision.
2004
rev = Revision(timestamp=self._timestamp,
2005
timezone=self._timezone,
2006
committer=self._committer,
2008
inventory_sha1=self.inv_sha1,
2009
revision_id=self._new_revision_id,
2010
properties=self._revprops)
2011
rev.parent_ids = self.parents
2012
self.repository.add_revision(self._new_revision_id, rev,
2013
self.new_inventory, self._config)
2014
return self._new_revision_id
2016
def finish_inventory(self):
2017
"""Tell the builder that the inventory is finished."""
2018
if self.new_inventory.root is None:
2019
warn('Root entry should be supplied to record_entry_contents, as'
2021
DeprecationWarning, stacklevel=2)
2022
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2023
self.new_inventory.revision_id = self._new_revision_id
2024
self.inv_sha1 = self.repository.add_inventory(
2025
self._new_revision_id,
2030
def _gen_revision_id(self):
2031
"""Return new revision-id."""
2032
s = '%s-%s-' % (self._config.user_email(),
2033
compact_date(self._timestamp))
2034
s += hexlify(rand_bytes(8))
2037
def _generate_revision_if_needed(self):
2038
"""Create a revision id if None was supplied.
2040
If the repository can not support user-specified revision ids
2041
they should override this function and raise UnsupportedOperation
2042
if _new_revision_id is not None.
2044
:raises: UnsupportedOperation
2046
if self._new_revision_id is None:
2047
self._new_revision_id = self._gen_revision_id()
2049
def record_entry_contents(self, ie, parent_invs, path, tree):
2050
"""Record the content of ie from tree into the commit if needed.
2052
Side effect: sets ie.revision when unchanged
2054
:param ie: An inventory entry present in the commit.
2055
:param parent_invs: The inventories of the parent revisions of the
2057
:param path: The path the entry is at in the tree.
2058
:param tree: The tree which contains this entry and should be used to
2061
self.new_inventory.add(ie)
2063
# ie.revision is always None if the InventoryEntry is considered
2064
# for committing. ie.snapshot will record the correct revision
2065
# which may be the sole parent if it is untouched.
2066
if ie.revision is not None:
2069
# In this revision format, root entries have no knit or weave
2070
if ie is self.new_inventory.root:
2071
if len(parent_invs):
2072
ie.revision = parent_invs[0].root.revision
2076
previous_entries = ie.find_previous_heads(
2078
self.repository.weave_store,
2079
self.repository.get_transaction())
2080
# we are creating a new revision for ie in the history store
2082
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2084
def modified_directory(self, file_id, file_parents):
2085
"""Record the presence of a symbolic link.
2087
:param file_id: The file_id of the link to record.
2088
:param file_parents: The per-file parent revision ids.
2090
self._add_text_to_weave(file_id, [], file_parents.keys())
2092
def modified_file_text(self, file_id, file_parents,
2093
get_content_byte_lines, text_sha1=None,
2095
"""Record the text of file file_id
2097
:param file_id: The file_id of the file to record the text of.
2098
:param file_parents: The per-file parent revision ids.
2099
:param get_content_byte_lines: A callable which will return the byte
2101
:param text_sha1: Optional SHA1 of the file contents.
2102
:param text_size: Optional size of the file contents.
2104
# mutter('storing text of file {%s} in revision {%s} into %r',
2105
# file_id, self._new_revision_id, self.repository.weave_store)
2106
# special case to avoid diffing on renames or
2108
if (len(file_parents) == 1
2109
and text_sha1 == file_parents.values()[0].text_sha1
2110
and text_size == file_parents.values()[0].text_size):
2111
previous_ie = file_parents.values()[0]
2112
versionedfile = self.repository.weave_store.get_weave(file_id,
2113
self.repository.get_transaction())
2114
versionedfile.clone_text(self._new_revision_id,
2115
previous_ie.revision, file_parents.keys())
2116
return text_sha1, text_size
2118
new_lines = get_content_byte_lines()
2119
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2120
# should return the SHA1 and size
2121
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2122
return osutils.sha_strings(new_lines), \
2123
sum(map(len, new_lines))
2125
def modified_link(self, file_id, file_parents, link_target):
2126
"""Record the presence of a symbolic link.
2128
:param file_id: The file_id of the link to record.
2129
:param file_parents: The per-file parent revision ids.
2130
:param link_target: Target location of this link.
2132
self._add_text_to_weave(file_id, [], file_parents.keys())
2134
def _add_text_to_weave(self, file_id, new_lines, parents):
2135
versionedfile = self.repository.weave_store.get_weave_or_empty(
2136
file_id, self.repository.get_transaction())
2137
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2138
versionedfile.clear_cache()
2150
def _unescaper(match, _map=_unescape_map):
2151
return _map[match.group(1)]
2157
def _unescape_xml(data):
2158
"""Unescape predefined XML entities in a string of data."""
2160
if _unescape_re is None:
2161
_unescape_re = re.compile('\&([^;]*);')
2162
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
