1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
38
revision as _mod_revision,
43
from bzrlib.revisiontree import RevisionTree
44
from bzrlib.store.versioned import VersionedFileStore
45
from bzrlib.store.text import TextStore
46
from bzrlib.testament import Testament
50
from bzrlib.decorators import needs_read_lock, needs_write_lock
51
from bzrlib.inter import InterObject
52
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
53
from bzrlib.symbol_versioning import (
57
from bzrlib.trace import mutter, note, warning
60
# Old formats display a warning, but only once
61
_deprecation_warning_done = False
64
######################################################################
67
class Repository(object):
68
"""Repository holding history for one or more branches.
70
The repository holds and retrieves historical information including
71
revisions and file history. It's normally accessed only by the Branch,
72
which views a particular line of development through that history.
74
The Repository builds on top of Stores and a Transport, which respectively
75
describe the disk data format and the way of accessing the (possibly
79
_file_ids_altered_regex = lazy_regex.lazy_compile(
80
r'file_id="(?P<file_id>[^"]+)"'
81
r'.*revision="(?P<revision_id>[^"]+)"'
85
def add_inventory(self, revision_id, inv, parents):
86
"""Add the inventory inv to the repository as revision_id.
88
:param parents: The revision ids of the parents that revision_id
89
is known to have and are in the repository already.
91
returns the sha1 of the serialized inventory.
93
revision_id = osutils.safe_revision_id(revision_id)
94
_mod_revision.check_not_reserved_id(revision_id)
95
assert inv.revision_id is None or inv.revision_id == revision_id, \
96
"Mismatch between inventory revision" \
97
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
98
assert inv.root is not None
99
inv_text = self.serialise_inventory(inv)
100
inv_sha1 = osutils.sha_string(inv_text)
101
inv_vf = self.control_weaves.get_weave('inventory',
102
self.get_transaction())
103
self._inventory_add_lines(inv_vf, revision_id, parents,
104
osutils.split_lines(inv_text))
107
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines):
109
for parent in parents:
111
final_parents.append(parent)
113
inv_vf.add_lines(revision_id, final_parents, lines)
116
def add_revision(self, revision_id, rev, inv=None, config=None):
117
"""Add rev to the revision store as revision_id.
119
:param revision_id: the revision id to use.
120
:param rev: The revision object.
121
:param inv: The inventory for the revision. if None, it will be looked
122
up in the inventory storer
123
:param config: If None no digital signature will be created.
124
If supplied its signature_needed method will be used
125
to determine if a signature should be made.
127
revision_id = osutils.safe_revision_id(revision_id)
128
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
130
_mod_revision.check_not_reserved_id(revision_id)
131
if config is not None and config.signature_needed():
133
inv = self.get_inventory(revision_id)
134
plaintext = Testament(rev, inv).as_short_text()
135
self.store_revision_signature(
136
gpg.GPGStrategy(config), plaintext, revision_id)
137
if not revision_id in self.get_inventory_weave():
139
raise errors.WeaveRevisionNotPresent(revision_id,
140
self.get_inventory_weave())
142
# yes, this is not suitable for adding with ghosts.
143
self.add_inventory(revision_id, inv, rev.parent_ids)
144
self._revision_store.add_revision(rev, self.get_transaction())
147
def _all_possible_ids(self):
148
"""Return all the possible revisions that we could find."""
149
return self.get_inventory_weave().versions()
151
def all_revision_ids(self):
152
"""Returns a list of all the revision ids in the repository.
154
This is deprecated because code should generally work on the graph
155
reachable from a particular revision, and ignore any other revisions
156
that might be present. There is no direct replacement method.
158
return self._all_revision_ids()
161
def _all_revision_ids(self):
162
"""Returns a list of all the revision ids in the repository.
164
These are in as much topological order as the underlying store can
165
present: for weaves ghosts may lead to a lack of correctness until
166
the reweave updates the parents list.
168
if self._revision_store.text_store.listable():
169
return self._revision_store.all_revision_ids(self.get_transaction())
170
result = self._all_possible_ids()
171
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
172
# ids. (It should, since _revision_store's API should change to
173
# return utf8 revision_ids)
174
return self._eliminate_revisions_not_present(result)
176
def break_lock(self):
177
"""Break a lock if one is present from another instance.
179
Uses the ui factory to ask for confirmation if the lock may be from
182
self.control_files.break_lock()
185
def _eliminate_revisions_not_present(self, revision_ids):
186
"""Check every revision id in revision_ids to see if we have it.
188
Returns a set of the present revisions.
191
for id in revision_ids:
192
if self.has_revision(id):
197
def create(a_bzrdir):
198
"""Construct the current default format repository in a_bzrdir."""
199
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
201
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
202
"""instantiate a Repository.
204
:param _format: The format of the repository on disk.
205
:param a_bzrdir: The BzrDir of the repository.
207
In the future we will have a single api for all stores for
208
getting file texts, inventories and revisions, then
209
this construct will accept instances of those things.
211
super(Repository, self).__init__()
212
self._format = _format
213
# the following are part of the public API for Repository:
214
self.bzrdir = a_bzrdir
215
self.control_files = control_files
216
self._revision_store = _revision_store
217
self.text_store = text_store
218
# backwards compatibility
219
self.weave_store = text_store
220
# not right yet - should be more semantically clear ?
222
self.control_store = control_store
223
self.control_weaves = control_store
224
# TODO: make sure to construct the right store classes, etc, depending
225
# on whether escaping is required.
226
self._warn_if_deprecated()
229
return '%s(%r)' % (self.__class__.__name__,
230
self.bzrdir.transport.base)
233
return self.control_files.is_locked()
235
def lock_write(self, token=None):
236
"""Lock this repository for writing.
238
:param token: if this is already locked, then lock_write will fail
239
unless the token matches the existing lock.
240
:returns: a token if this instance supports tokens, otherwise None.
241
:raises TokenLockingNotSupported: when a token is given but this
242
instance doesn't support using token locks.
243
:raises MismatchedToken: if the specified token doesn't match the token
244
of the existing lock.
246
A token should be passed in if you know that you have locked the object
247
some other way, and need to synchronise this object's state with that
250
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
252
return self.control_files.lock_write(token=token)
255
self.control_files.lock_read()
257
def get_physical_lock_status(self):
258
return self.control_files.get_physical_lock_status()
260
def leave_lock_in_place(self):
261
"""Tell this repository not to release the physical lock when this
264
If lock_write doesn't return a token, then this method is not supported.
266
self.control_files.leave_in_place()
268
def dont_leave_lock_in_place(self):
269
"""Tell this repository to release the physical lock when this
270
object is unlocked, even if it didn't originally acquire it.
272
If lock_write doesn't return a token, then this method is not supported.
274
self.control_files.dont_leave_in_place()
277
def gather_stats(self, revid=None, committers=None):
278
"""Gather statistics from a revision id.
280
:param revid: The revision id to gather statistics from, if None, then
281
no revision specific statistics are gathered.
282
:param committers: Optional parameter controlling whether to grab
283
a count of committers from the revision specific statistics.
284
:return: A dictionary of statistics. Currently this contains:
285
committers: The number of committers if requested.
286
firstrev: A tuple with timestamp, timezone for the penultimate left
287
most ancestor of revid, if revid is not the NULL_REVISION.
288
latestrev: A tuple with timestamp, timezone for revid, if revid is
289
not the NULL_REVISION.
290
revisions: The total revision count in the repository.
291
size: An estimate disk size of the repository in bytes.
294
if revid and committers:
295
result['committers'] = 0
296
if revid and revid != _mod_revision.NULL_REVISION:
298
all_committers = set()
299
revisions = self.get_ancestry(revid)
300
# pop the leading None
302
first_revision = None
304
# ignore the revisions in the middle - just grab first and last
305
revisions = revisions[0], revisions[-1]
306
for revision in self.get_revisions(revisions):
307
if not first_revision:
308
first_revision = revision
310
all_committers.add(revision.committer)
311
last_revision = revision
313
result['committers'] = len(all_committers)
314
result['firstrev'] = (first_revision.timestamp,
315
first_revision.timezone)
316
result['latestrev'] = (last_revision.timestamp,
317
last_revision.timezone)
319
# now gather global repository information
320
if self.bzrdir.root_transport.listable():
321
c, t = self._revision_store.total_size(self.get_transaction())
322
result['revisions'] = c
327
def missing_revision_ids(self, other, revision_id=None):
328
"""Return the revision ids that other has that this does not.
330
These are returned in topological order.
332
revision_id: only return revision ids included by revision_id.
334
revision_id = osutils.safe_revision_id(revision_id)
335
return InterRepository.get(other, self).missing_revision_ids(revision_id)
339
"""Open the repository rooted at base.
341
For instance, if the repository is at URL/.bzr/repository,
342
Repository.open(URL) -> a Repository instance.
344
control = bzrdir.BzrDir.open(base)
345
return control.open_repository()
347
def copy_content_into(self, destination, revision_id=None):
348
"""Make a complete copy of the content in self into destination.
350
This is a destructive operation! Do not use it on existing
353
revision_id = osutils.safe_revision_id(revision_id)
354
return InterRepository.get(self, destination).copy_content(revision_id)
356
def fetch(self, source, revision_id=None, pb=None):
357
"""Fetch the content required to construct revision_id from source.
359
If revision_id is None all content is copied.
361
revision_id = osutils.safe_revision_id(revision_id)
362
inter = InterRepository.get(source, self)
364
return inter.fetch(revision_id=revision_id, pb=pb)
365
except NotImplementedError:
366
raise errors.IncompatibleRepositories(source, self)
368
def get_commit_builder(self, branch, parents, config, timestamp=None,
369
timezone=None, committer=None, revprops=None,
371
"""Obtain a CommitBuilder for this repository.
373
:param branch: Branch to commit to.
374
:param parents: Revision ids of the parents of the new revision.
375
:param config: Configuration to use.
376
:param timestamp: Optional timestamp recorded for commit.
377
:param timezone: Optional timezone for timestamp.
378
:param committer: Optional committer to set for commit.
379
:param revprops: Optional dictionary of revision properties.
380
:param revision_id: Optional revision id.
382
revision_id = osutils.safe_revision_id(revision_id)
383
return _CommitBuilder(self, parents, config, timestamp, timezone,
384
committer, revprops, revision_id)
387
self.control_files.unlock()
390
def clone(self, a_bzrdir, revision_id=None):
391
"""Clone this repository into a_bzrdir using the current format.
393
Currently no check is made that the format of this repository and
394
the bzrdir format are compatible. FIXME RBC 20060201.
396
:return: The newly created destination repository.
398
# TODO: deprecate after 0.16; cloning this with all its settings is
399
# probably not very useful -- mbp 20070423
400
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
401
self.copy_content_into(dest_repo, revision_id)
405
def sprout(self, to_bzrdir, revision_id=None):
406
"""Create a descendent repository for new development.
408
Unlike clone, this does not copy the settings of the repository.
410
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
411
dest_repo.fetch(self, revision_id=revision_id)
414
def _create_sprouting_repo(self, a_bzrdir, shared):
415
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
416
# use target default format.
417
dest_repo = a_bzrdir.create_repository()
419
# Most control formats need the repository to be specifically
420
# created, but on some old all-in-one formats it's not needed
422
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
423
except errors.UninitializableFormat:
424
dest_repo = a_bzrdir.open_repository()
428
def has_revision(self, revision_id):
429
"""True if this repository has a copy of the revision."""
430
revision_id = osutils.safe_revision_id(revision_id)
431
return self._revision_store.has_revision_id(revision_id,
432
self.get_transaction())
435
def get_revision_reconcile(self, revision_id):
436
"""'reconcile' helper routine that allows access to a revision always.
438
This variant of get_revision does not cross check the weave graph
439
against the revision one as get_revision does: but it should only
440
be used by reconcile, or reconcile-alike commands that are correcting
441
or testing the revision graph.
443
if not revision_id or not isinstance(revision_id, basestring):
444
raise errors.InvalidRevisionId(revision_id=revision_id,
446
return self.get_revisions([revision_id])[0]
449
def get_revisions(self, revision_ids):
450
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
451
revs = self._revision_store.get_revisions(revision_ids,
452
self.get_transaction())
454
assert not isinstance(rev.revision_id, unicode)
455
for parent_id in rev.parent_ids:
456
assert not isinstance(parent_id, unicode)
460
def get_revision_xml(self, revision_id):
461
# TODO: jam 20070210 This shouldn't be necessary since get_revision
462
# would have already do it.
463
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
464
revision_id = osutils.safe_revision_id(revision_id)
465
rev = self.get_revision(revision_id)
467
# the current serializer..
468
self._revision_store._serializer.write_revision(rev, rev_tmp)
470
return rev_tmp.getvalue()
473
def get_revision(self, revision_id):
474
"""Return the Revision object for a named revision"""
475
# TODO: jam 20070210 get_revision_reconcile should do this for us
476
revision_id = osutils.safe_revision_id(revision_id)
477
r = self.get_revision_reconcile(revision_id)
478
# weave corruption can lead to absent revision markers that should be
480
# the following test is reasonably cheap (it needs a single weave read)
481
# and the weave is cached in read transactions. In write transactions
482
# it is not cached but typically we only read a small number of
483
# revisions. For knits when they are introduced we will probably want
484
# to ensure that caching write transactions are in use.
485
inv = self.get_inventory_weave()
486
self._check_revision_parents(r, inv)
490
def get_deltas_for_revisions(self, revisions):
491
"""Produce a generator of revision deltas.
493
Note that the input is a sequence of REVISIONS, not revision_ids.
494
Trees will be held in memory until the generator exits.
495
Each delta is relative to the revision's lefthand predecessor.
497
required_trees = set()
498
for revision in revisions:
499
required_trees.add(revision.revision_id)
500
required_trees.update(revision.parent_ids[:1])
501
trees = dict((t.get_revision_id(), t) for
502
t in self.revision_trees(required_trees))
503
for revision in revisions:
504
if not revision.parent_ids:
505
old_tree = self.revision_tree(None)
507
old_tree = trees[revision.parent_ids[0]]
508
yield trees[revision.revision_id].changes_from(old_tree)
511
def get_revision_delta(self, revision_id):
512
"""Return the delta for one revision.
514
The delta is relative to the left-hand predecessor of the
517
r = self.get_revision(revision_id)
518
return list(self.get_deltas_for_revisions([r]))[0]
520
def _check_revision_parents(self, revision, inventory):
521
"""Private to Repository and Fetch.
523
This checks the parentage of revision in an inventory weave for
524
consistency and is only applicable to inventory-weave-for-ancestry
525
using repository formats & fetchers.
527
weave_parents = inventory.get_parents(revision.revision_id)
528
weave_names = inventory.versions()
529
for parent_id in revision.parent_ids:
530
if parent_id in weave_names:
531
# this parent must not be a ghost.
532
if not parent_id in weave_parents:
534
raise errors.CorruptRepository(self)
537
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
538
revision_id = osutils.safe_revision_id(revision_id)
539
signature = gpg_strategy.sign(plaintext)
540
self._revision_store.add_revision_signature_text(revision_id,
542
self.get_transaction())
544
def fileids_altered_by_revision_ids(self, revision_ids):
545
"""Find the file ids and versions affected by revisions.
547
:param revisions: an iterable containing revision ids.
548
:return: a dictionary mapping altered file-ids to an iterable of
549
revision_ids. Each altered file-ids has the exact revision_ids that
550
altered it listed explicitly.
552
assert self._serializer.support_altered_by_hack, \
553
("fileids_altered_by_revision_ids only supported for branches "
554
"which store inventory as unnested xml, not on %r" % self)
555
selected_revision_ids = set(osutils.safe_revision_id(r)
556
for r in revision_ids)
557
w = self.get_inventory_weave()
560
# this code needs to read every new line in every inventory for the
561
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
562
# not present in one of those inventories is unnecessary but not
563
# harmful because we are filtering by the revision id marker in the
564
# inventory lines : we only select file ids altered in one of those
565
# revisions. We don't need to see all lines in the inventory because
566
# only those added in an inventory in rev X can contain a revision=X
568
unescape_revid_cache = {}
569
unescape_fileid_cache = {}
571
# jam 20061218 In a big fetch, this handles hundreds of thousands
572
# of lines, so it has had a lot of inlining and optimizing done.
573
# Sorry that it is a little bit messy.
574
# Move several functions to be local variables, since this is a long
576
search = self._file_ids_altered_regex.search
577
unescape = _unescape_xml
578
setdefault = result.setdefault
579
pb = ui.ui_factory.nested_progress_bar()
581
for line in w.iter_lines_added_or_present_in_versions(
582
selected_revision_ids, pb=pb):
586
# One call to match.group() returning multiple items is quite a
587
# bit faster than 2 calls to match.group() each returning 1
588
file_id, revision_id = match.group('file_id', 'revision_id')
590
# Inlining the cache lookups helps a lot when you make 170,000
591
# lines and 350k ids, versus 8.4 unique ids.
592
# Using a cache helps in 2 ways:
593
# 1) Avoids unnecessary decoding calls
594
# 2) Re-uses cached strings, which helps in future set and
596
# (2) is enough that removing encoding entirely along with
597
# the cache (so we are using plain strings) results in no
598
# performance improvement.
600
revision_id = unescape_revid_cache[revision_id]
602
unescaped = unescape(revision_id)
603
unescape_revid_cache[revision_id] = unescaped
604
revision_id = unescaped
606
if revision_id in selected_revision_ids:
608
file_id = unescape_fileid_cache[file_id]
610
unescaped = unescape(file_id)
611
unescape_fileid_cache[file_id] = unescaped
613
setdefault(file_id, set()).add(revision_id)
619
def get_inventory_weave(self):
620
return self.control_weaves.get_weave('inventory',
621
self.get_transaction())
624
def get_inventory(self, revision_id):
625
"""Get Inventory object by hash."""
626
# TODO: jam 20070210 Technically we don't need to sanitize, since all
627
# called functions must sanitize.
628
revision_id = osutils.safe_revision_id(revision_id)
629
return self.deserialise_inventory(
630
revision_id, self.get_inventory_xml(revision_id))
632
def deserialise_inventory(self, revision_id, xml):
633
"""Transform the xml into an inventory object.
635
:param revision_id: The expected revision id of the inventory.
636
:param xml: A serialised inventory.
638
revision_id = osutils.safe_revision_id(revision_id)
639
result = self._serializer.read_inventory_from_string(xml)
640
result.root.revision = revision_id
643
def serialise_inventory(self, inv):
644
return self._serializer.write_inventory_to_string(inv)
647
def get_inventory_xml(self, revision_id):
648
"""Get inventory XML as a file object."""
649
revision_id = osutils.safe_revision_id(revision_id)
651
assert isinstance(revision_id, str), type(revision_id)
652
iw = self.get_inventory_weave()
653
return iw.get_text(revision_id)
655
raise errors.HistoryMissing(self, 'inventory', revision_id)
658
def get_inventory_sha1(self, revision_id):
659
"""Return the sha1 hash of the inventory entry
661
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
662
revision_id = osutils.safe_revision_id(revision_id)
663
return self.get_revision(revision_id).inventory_sha1
666
def get_revision_graph(self, revision_id=None):
667
"""Return a dictionary containing the revision graph.
669
:param revision_id: The revision_id to get a graph from. If None, then
670
the entire revision graph is returned. This is a deprecated mode of
671
operation and will be removed in the future.
672
:return: a dictionary of revision_id->revision_parents_list.
674
# special case NULL_REVISION
675
if revision_id == _mod_revision.NULL_REVISION:
677
revision_id = osutils.safe_revision_id(revision_id)
678
a_weave = self.get_inventory_weave()
679
all_revisions = self._eliminate_revisions_not_present(
681
entire_graph = dict([(node, a_weave.get_parents(node)) for
682
node in all_revisions])
683
if revision_id is None:
685
elif revision_id not in entire_graph:
686
raise errors.NoSuchRevision(self, revision_id)
688
# add what can be reached from revision_id
690
pending = set([revision_id])
691
while len(pending) > 0:
693
result[node] = entire_graph[node]
694
for revision_id in result[node]:
695
if revision_id not in result:
696
pending.add(revision_id)
700
def get_revision_graph_with_ghosts(self, revision_ids=None):
701
"""Return a graph of the revisions with ghosts marked as applicable.
703
:param revision_ids: an iterable of revisions to graph or None for all.
704
:return: a Graph object with the graph reachable from revision_ids.
706
result = graph.Graph()
708
pending = set(self.all_revision_ids())
711
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
712
# special case NULL_REVISION
713
if _mod_revision.NULL_REVISION in pending:
714
pending.remove(_mod_revision.NULL_REVISION)
715
required = set(pending)
718
revision_id = pending.pop()
720
rev = self.get_revision(revision_id)
721
except errors.NoSuchRevision:
722
if revision_id in required:
725
result.add_ghost(revision_id)
727
for parent_id in rev.parent_ids:
728
# is this queued or done ?
729
if (parent_id not in pending and
730
parent_id not in done):
732
pending.add(parent_id)
733
result.add_node(revision_id, rev.parent_ids)
734
done.add(revision_id)
737
def _get_history_vf(self):
738
"""Get a versionedfile whose history graph reflects all revisions.
740
For weave repositories, this is the inventory weave.
742
return self.get_inventory_weave()
744
def iter_reverse_revision_history(self, revision_id):
745
"""Iterate backwards through revision ids in the lefthand history
747
:param revision_id: The revision id to start with. All its lefthand
748
ancestors will be traversed.
750
revision_id = osutils.safe_revision_id(revision_id)
751
if revision_id in (None, _mod_revision.NULL_REVISION):
753
next_id = revision_id
754
versionedfile = self._get_history_vf()
757
parents = versionedfile.get_parents(next_id)
758
if len(parents) == 0:
764
def get_revision_inventory(self, revision_id):
765
"""Return inventory of a past revision."""
766
# TODO: Unify this with get_inventory()
767
# bzr 0.0.6 and later imposes the constraint that the inventory_id
768
# must be the same as its revision, so this is trivial.
769
if revision_id is None:
770
# This does not make sense: if there is no revision,
771
# then it is the current tree inventory surely ?!
772
# and thus get_root_id() is something that looks at the last
773
# commit on the branch, and the get_root_id is an inventory check.
774
raise NotImplementedError
775
# return Inventory(self.get_root_id())
777
return self.get_inventory(revision_id)
781
"""Return True if this repository is flagged as a shared repository."""
782
raise NotImplementedError(self.is_shared)
785
def reconcile(self, other=None, thorough=False):
786
"""Reconcile this repository."""
787
from bzrlib.reconcile import RepoReconciler
788
reconciler = RepoReconciler(self, thorough=thorough)
789
reconciler.reconcile()
793
def revision_tree(self, revision_id):
794
"""Return Tree for a revision on this branch.
796
`revision_id` may be None for the empty tree revision.
798
# TODO: refactor this to use an existing revision object
799
# so we don't need to read it in twice.
800
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
801
return RevisionTree(self, Inventory(root_id=None),
802
_mod_revision.NULL_REVISION)
804
revision_id = osutils.safe_revision_id(revision_id)
805
inv = self.get_revision_inventory(revision_id)
806
return RevisionTree(self, inv, revision_id)
809
def revision_trees(self, revision_ids):
810
"""Return Tree for a revision on this branch.
812
`revision_id` may not be None or 'null:'"""
813
assert None not in revision_ids
814
assert _mod_revision.NULL_REVISION not in revision_ids
815
texts = self.get_inventory_weave().get_texts(revision_ids)
816
for text, revision_id in zip(texts, revision_ids):
817
inv = self.deserialise_inventory(revision_id, text)
818
yield RevisionTree(self, inv, revision_id)
821
def get_ancestry(self, revision_id):
822
"""Return a list of revision-ids integrated by a revision.
824
The first element of the list is always None, indicating the origin
825
revision. This might change when we have history horizons, or
826
perhaps we should have a new API.
828
This is topologically sorted.
830
if revision_id is None:
832
revision_id = osutils.safe_revision_id(revision_id)
833
if not self.has_revision(revision_id):
834
raise errors.NoSuchRevision(self, revision_id)
835
w = self.get_inventory_weave()
836
candidates = w.get_ancestry(revision_id)
837
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
840
def print_file(self, file, revision_id):
841
"""Print `file` to stdout.
843
FIXME RBC 20060125 as John Meinel points out this is a bad api
844
- it writes to stdout, it assumes that that is valid etc. Fix
845
by creating a new more flexible convenience function.
847
revision_id = osutils.safe_revision_id(revision_id)
848
tree = self.revision_tree(revision_id)
849
# use inventory as it was in that revision
850
file_id = tree.inventory.path2id(file)
852
# TODO: jam 20060427 Write a test for this code path
853
# it had a bug in it, and was raising the wrong
855
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
856
tree.print_file(file_id)
858
def get_transaction(self):
859
return self.control_files.get_transaction()
861
def revision_parents(self, revision_id):
862
revision_id = osutils.safe_revision_id(revision_id)
863
return self.get_inventory_weave().parent_names(revision_id)
866
def set_make_working_trees(self, new_value):
867
"""Set the policy flag for making working trees when creating branches.
869
This only applies to branches that use this repository.
871
The default is 'True'.
872
:param new_value: True to restore the default, False to disable making
875
raise NotImplementedError(self.set_make_working_trees)
877
def make_working_trees(self):
878
"""Returns the policy for making working trees on new branches."""
879
raise NotImplementedError(self.make_working_trees)
882
def sign_revision(self, revision_id, gpg_strategy):
883
revision_id = osutils.safe_revision_id(revision_id)
884
plaintext = Testament.from_revision(self, revision_id).as_short_text()
885
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
888
def has_signature_for_revision_id(self, revision_id):
889
"""Query for a revision signature for revision_id in the repository."""
890
revision_id = osutils.safe_revision_id(revision_id)
891
return self._revision_store.has_signature(revision_id,
892
self.get_transaction())
895
def get_signature_text(self, revision_id):
896
"""Return the text for a signature."""
897
revision_id = osutils.safe_revision_id(revision_id)
898
return self._revision_store.get_signature_text(revision_id,
899
self.get_transaction())
902
def check(self, revision_ids):
903
"""Check consistency of all history of given revision_ids.
905
Different repository implementations should override _check().
907
:param revision_ids: A non-empty list of revision_ids whose ancestry
908
will be checked. Typically the last revision_id of a branch.
911
raise ValueError("revision_ids must be non-empty in %s.check"
913
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
914
return self._check(revision_ids)
916
def _check(self, revision_ids):
917
result = check.Check(self)
921
def _warn_if_deprecated(self):
922
global _deprecation_warning_done
923
if _deprecation_warning_done:
925
_deprecation_warning_done = True
926
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
927
% (self._format, self.bzrdir.transport.base))
929
def supports_rich_root(self):
930
return self._format.rich_root_data
932
def _check_ascii_revisionid(self, revision_id, method):
933
"""Private helper for ascii-only repositories."""
934
# weave repositories refuse to store revisionids that are non-ascii.
935
if revision_id is not None:
936
# weaves require ascii revision ids.
937
if isinstance(revision_id, unicode):
939
revision_id.encode('ascii')
940
except UnicodeEncodeError:
941
raise errors.NonAsciiRevisionId(method, self)
944
revision_id.decode('ascii')
945
except UnicodeDecodeError:
946
raise errors.NonAsciiRevisionId(method, self)
950
# remove these delegates a while after bzr 0.15
951
def __make_delegated(name, from_module):
952
def _deprecated_repository_forwarder():
953
symbol_versioning.warn('%s moved to %s in bzr 0.15'
954
% (name, from_module),
957
m = __import__(from_module, globals(), locals(), [name])
959
return getattr(m, name)
960
except AttributeError:
961
raise AttributeError('module %s has no name %s'
963
globals()[name] = _deprecated_repository_forwarder
966
'AllInOneRepository',
967
'WeaveMetaDirRepository',
968
'PreSplitOutRepositoryFormat',
974
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
978
'RepositoryFormatKnit',
979
'RepositoryFormatKnit1',
981
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
984
def install_revision(repository, rev, revision_tree):
985
"""Install all revision data into a repository."""
988
for p_id in rev.parent_ids:
989
if repository.has_revision(p_id):
990
present_parents.append(p_id)
991
parent_trees[p_id] = repository.revision_tree(p_id)
993
parent_trees[p_id] = repository.revision_tree(None)
995
inv = revision_tree.inventory
996
entries = inv.iter_entries()
997
# backwards compatability hack: skip the root id.
998
if not repository.supports_rich_root():
999
path, root = entries.next()
1000
if root.revision != rev.revision_id:
1001
raise errors.IncompatibleRevision(repr(repository))
1002
# Add the texts that are not already present
1003
for path, ie in entries:
1004
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1005
repository.get_transaction())
1006
if ie.revision not in w:
1008
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1009
# with InventoryEntry.find_previous_heads(). if it is, then there
1010
# is a latent bug here where the parents may have ancestors of each
1012
for revision, tree in parent_trees.iteritems():
1013
if ie.file_id not in tree:
1015
parent_id = tree.inventory[ie.file_id].revision
1016
if parent_id in text_parents:
1018
text_parents.append(parent_id)
1020
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1021
repository.get_transaction())
1022
lines = revision_tree.get_file(ie.file_id).readlines()
1023
vfile.add_lines(rev.revision_id, text_parents, lines)
1025
# install the inventory
1026
repository.add_inventory(rev.revision_id, inv, present_parents)
1027
except errors.RevisionAlreadyPresent:
1029
repository.add_revision(rev.revision_id, rev, inv)
1032
class MetaDirRepository(Repository):
1033
"""Repositories in the new meta-dir layout."""
1035
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1036
super(MetaDirRepository, self).__init__(_format,
1042
dir_mode = self.control_files._dir_mode
1043
file_mode = self.control_files._file_mode
1046
def is_shared(self):
1047
"""Return True if this repository is flagged as a shared repository."""
1048
return self.control_files._transport.has('shared-storage')
1051
def set_make_working_trees(self, new_value):
1052
"""Set the policy flag for making working trees when creating branches.
1054
This only applies to branches that use this repository.
1056
The default is 'True'.
1057
:param new_value: True to restore the default, False to disable making
1062
self.control_files._transport.delete('no-working-trees')
1063
except errors.NoSuchFile:
1066
self.control_files.put_utf8('no-working-trees', '')
1068
def make_working_trees(self):
1069
"""Returns the policy for making working trees on new branches."""
1070
return not self.control_files._transport.has('no-working-trees')
1073
class RepositoryFormatRegistry(registry.Registry):
1074
"""Registry of RepositoryFormats.
1077
def get(self, format_string):
1078
r = registry.Registry.get(self, format_string)
1084
format_registry = RepositoryFormatRegistry()
1085
"""Registry of formats, indexed by their identifying format string.
1087
This can contain either format instances themselves, or classes/factories that
1088
can be called to obtain one.
1092
#####################################################################
1093
# Repository Formats
1095
class RepositoryFormat(object):
1096
"""A repository format.
1098
Formats provide three things:
1099
* An initialization routine to construct repository data on disk.
1100
* a format string which is used when the BzrDir supports versioned
1102
* an open routine which returns a Repository instance.
1104
Formats are placed in an dict by their format string for reference
1105
during opening. These should be subclasses of RepositoryFormat
1108
Once a format is deprecated, just deprecate the initialize and open
1109
methods on the format class. Do not deprecate the object, as the
1110
object will be created every system load.
1112
Common instance attributes:
1113
_matchingbzrdir - the bzrdir format that the repository format was
1114
originally written to work with. This can be used if manually
1115
constructing a bzrdir and repository, or more commonly for test suite
1120
return "<%s>" % self.__class__.__name__
1122
def __eq__(self, other):
1123
# format objects are generally stateless
1124
return isinstance(other, self.__class__)
1126
def __ne__(self, other):
1127
return not self == other
1130
def find_format(klass, a_bzrdir):
1131
"""Return the format for the repository object in a_bzrdir.
1133
This is used by bzr native formats that have a "format" file in
1134
the repository. Other methods may be used by different types of
1138
transport = a_bzrdir.get_repository_transport(None)
1139
format_string = transport.get("format").read()
1140
return format_registry.get(format_string)
1141
except errors.NoSuchFile:
1142
raise errors.NoRepositoryPresent(a_bzrdir)
1144
raise errors.UnknownFormatError(format=format_string)
1147
def register_format(klass, format):
1148
format_registry.register(format.get_format_string(), format)
1151
def unregister_format(klass, format):
1152
format_registry.remove(format.get_format_string())
1155
def get_default_format(klass):
1156
"""Return the current default format."""
1157
from bzrlib import bzrdir
1158
return bzrdir.format_registry.make_bzrdir('default').repository_format
1160
def _get_control_store(self, repo_transport, control_files):
1161
"""Return the control store for this repository."""
1162
raise NotImplementedError(self._get_control_store)
1164
def get_format_string(self):
1165
"""Return the ASCII format string that identifies this format.
1167
Note that in pre format ?? repositories the format string is
1168
not permitted nor written to disk.
1170
raise NotImplementedError(self.get_format_string)
1172
def get_format_description(self):
1173
"""Return the short description for this format."""
1174
raise NotImplementedError(self.get_format_description)
1176
def _get_revision_store(self, repo_transport, control_files):
1177
"""Return the revision store object for this a_bzrdir."""
1178
raise NotImplementedError(self._get_revision_store)
1180
def _get_text_rev_store(self,
1187
"""Common logic for getting a revision store for a repository.
1189
see self._get_revision_store for the subclass-overridable method to
1190
get the store for a repository.
1192
from bzrlib.store.revision.text import TextRevisionStore
1193
dir_mode = control_files._dir_mode
1194
file_mode = control_files._file_mode
1195
text_store = TextStore(transport.clone(name),
1197
compressed=compressed,
1199
file_mode=file_mode)
1200
_revision_store = TextRevisionStore(text_store, serializer)
1201
return _revision_store
1203
# TODO: this shouldn't be in the base class, it's specific to things that
1204
# use weaves or knits -- mbp 20070207
1205
def _get_versioned_file_store(self,
1210
versionedfile_class=None,
1211
versionedfile_kwargs={},
1213
if versionedfile_class is None:
1214
versionedfile_class = self._versionedfile_class
1215
weave_transport = control_files._transport.clone(name)
1216
dir_mode = control_files._dir_mode
1217
file_mode = control_files._file_mode
1218
return VersionedFileStore(weave_transport, prefixed=prefixed,
1220
file_mode=file_mode,
1221
versionedfile_class=versionedfile_class,
1222
versionedfile_kwargs=versionedfile_kwargs,
1225
def initialize(self, a_bzrdir, shared=False):
1226
"""Initialize a repository of this format in a_bzrdir.
1228
:param a_bzrdir: The bzrdir to put the new repository in it.
1229
:param shared: The repository should be initialized as a sharable one.
1230
:returns: The new repository object.
1232
This may raise UninitializableFormat if shared repository are not
1233
compatible the a_bzrdir.
1235
raise NotImplementedError(self.initialize)
1237
def is_supported(self):
1238
"""Is this format supported?
1240
Supported formats must be initializable and openable.
1241
Unsupported formats may not support initialization or committing or
1242
some other features depending on the reason for not being supported.
1246
def check_conversion_target(self, target_format):
1247
raise NotImplementedError(self.check_conversion_target)
1249
def open(self, a_bzrdir, _found=False):
1250
"""Return an instance of this format for the bzrdir a_bzrdir.
1252
_found is a private parameter, do not use it.
1254
raise NotImplementedError(self.open)
1257
class MetaDirRepositoryFormat(RepositoryFormat):
1258
"""Common base class for the new repositories using the metadir layout."""
1260
rich_root_data = False
1261
supports_tree_reference = False
1262
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1265
super(MetaDirRepositoryFormat, self).__init__()
1267
def _create_control_files(self, a_bzrdir):
1268
"""Create the required files and the initial control_files object."""
1269
# FIXME: RBC 20060125 don't peek under the covers
1270
# NB: no need to escape relative paths that are url safe.
1271
repository_transport = a_bzrdir.get_repository_transport(self)
1272
control_files = lockable_files.LockableFiles(repository_transport,
1273
'lock', lockdir.LockDir)
1274
control_files.create_lock()
1275
return control_files
1277
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1278
"""Upload the initial blank content."""
1279
control_files = self._create_control_files(a_bzrdir)
1280
control_files.lock_write()
1282
control_files._transport.mkdir_multi(dirs,
1283
mode=control_files._dir_mode)
1284
for file, content in files:
1285
control_files.put(file, content)
1286
for file, content in utf8_files:
1287
control_files.put_utf8(file, content)
1289
control_files.put_utf8('shared-storage', '')
1291
control_files.unlock()
1294
# formats which have no format string are not discoverable
1295
# and not independently creatable, so are not registered. They're
1296
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1297
# needed, it's constructed directly by the BzrDir. Non-native formats where
1298
# the repository is not separately opened are similar.
1300
format_registry.register_lazy(
1301
'Bazaar-NG Repository format 7',
1302
'bzrlib.repofmt.weaverepo',
1305
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1306
# default control directory format
1308
format_registry.register_lazy(
1309
'Bazaar-NG Knit Repository Format 1',
1310
'bzrlib.repofmt.knitrepo',
1311
'RepositoryFormatKnit1',
1313
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1315
format_registry.register_lazy(
1316
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1317
'bzrlib.repofmt.knitrepo',
1318
'RepositoryFormatKnit3',
1322
class InterRepository(InterObject):
1323
"""This class represents operations taking place between two repositories.
1325
Its instances have methods like copy_content and fetch, and contain
1326
references to the source and target repositories these operations can be
1329
Often we will provide convenience methods on 'repository' which carry out
1330
operations with another repository - they will always forward to
1331
InterRepository.get(other).method_name(parameters).
1335
"""The available optimised InterRepository types."""
1337
def copy_content(self, revision_id=None):
1338
raise NotImplementedError(self.copy_content)
1340
def fetch(self, revision_id=None, pb=None):
1341
"""Fetch the content required to construct revision_id.
1343
The content is copied from self.source to self.target.
1345
:param revision_id: if None all content is copied, if NULL_REVISION no
1347
:param pb: optional progress bar to use for progress reports. If not
1348
provided a default one will be created.
1350
Returns the copied revision count and the failed revisions in a tuple:
1353
raise NotImplementedError(self.fetch)
1356
def missing_revision_ids(self, revision_id=None):
1357
"""Return the revision ids that source has that target does not.
1359
These are returned in topological order.
1361
:param revision_id: only return revision ids included by this
1364
# generic, possibly worst case, slow code path.
1365
target_ids = set(self.target.all_revision_ids())
1366
if revision_id is not None:
1367
# TODO: jam 20070210 InterRepository is internal enough that it
1368
# should assume revision_ids are already utf-8
1369
revision_id = osutils.safe_revision_id(revision_id)
1370
source_ids = self.source.get_ancestry(revision_id)
1371
assert source_ids[0] is None
1374
source_ids = self.source.all_revision_ids()
1375
result_set = set(source_ids).difference(target_ids)
1376
# this may look like a no-op: its not. It preserves the ordering
1377
# other_ids had while only returning the members from other_ids
1378
# that we've decided we need.
1379
return [rev_id for rev_id in source_ids if rev_id in result_set]
1382
class InterSameDataRepository(InterRepository):
1383
"""Code for converting between repositories that represent the same data.
1385
Data format and model must match for this to work.
1389
def _get_repo_format_to_test(self):
1390
"""Repository format for testing with."""
1391
return RepositoryFormat.get_default_format()
1394
def is_compatible(source, target):
1395
if source.supports_rich_root() != target.supports_rich_root():
1397
if source._serializer != target._serializer:
1402
def copy_content(self, revision_id=None):
1403
"""Make a complete copy of the content in self into destination.
1405
This copies both the repository's revision data, and configuration information
1406
such as the make_working_trees setting.
1408
This is a destructive operation! Do not use it on existing
1411
:param revision_id: Only copy the content needed to construct
1412
revision_id and its parents.
1415
self.target.set_make_working_trees(self.source.make_working_trees())
1416
except NotImplementedError:
1418
# TODO: jam 20070210 This is fairly internal, so we should probably
1419
# just assert that revision_id is not unicode.
1420
revision_id = osutils.safe_revision_id(revision_id)
1421
# but don't bother fetching if we have the needed data now.
1422
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1423
self.target.has_revision(revision_id)):
1425
self.target.fetch(self.source, revision_id=revision_id)
1428
def fetch(self, revision_id=None, pb=None):
1429
"""See InterRepository.fetch()."""
1430
from bzrlib.fetch import GenericRepoFetcher
1431
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1432
self.source, self.source._format, self.target,
1433
self.target._format)
1434
# TODO: jam 20070210 This should be an assert, not a translate
1435
revision_id = osutils.safe_revision_id(revision_id)
1436
f = GenericRepoFetcher(to_repository=self.target,
1437
from_repository=self.source,
1438
last_revision=revision_id,
1440
return f.count_copied, f.failed_revisions
1443
class InterWeaveRepo(InterSameDataRepository):
1444
"""Optimised code paths between Weave based repositories."""
1447
def _get_repo_format_to_test(self):
1448
from bzrlib.repofmt import weaverepo
1449
return weaverepo.RepositoryFormat7()
1452
def is_compatible(source, target):
1453
"""Be compatible with known Weave formats.
1455
We don't test for the stores being of specific types because that
1456
could lead to confusing results, and there is no need to be
1459
from bzrlib.repofmt.weaverepo import (
1465
return (isinstance(source._format, (RepositoryFormat5,
1467
RepositoryFormat7)) and
1468
isinstance(target._format, (RepositoryFormat5,
1470
RepositoryFormat7)))
1471
except AttributeError:
1475
def copy_content(self, revision_id=None):
1476
"""See InterRepository.copy_content()."""
1477
# weave specific optimised path:
1478
# TODO: jam 20070210 Internal, should be an assert, not translate
1479
revision_id = osutils.safe_revision_id(revision_id)
1481
self.target.set_make_working_trees(self.source.make_working_trees())
1482
except NotImplementedError:
1484
# FIXME do not peek!
1485
if self.source.control_files._transport.listable():
1486
pb = ui.ui_factory.nested_progress_bar()
1488
self.target.weave_store.copy_all_ids(
1489
self.source.weave_store,
1491
from_transaction=self.source.get_transaction(),
1492
to_transaction=self.target.get_transaction())
1493
pb.update('copying inventory', 0, 1)
1494
self.target.control_weaves.copy_multi(
1495
self.source.control_weaves, ['inventory'],
1496
from_transaction=self.source.get_transaction(),
1497
to_transaction=self.target.get_transaction())
1498
self.target._revision_store.text_store.copy_all_ids(
1499
self.source._revision_store.text_store,
1504
self.target.fetch(self.source, revision_id=revision_id)
1507
def fetch(self, revision_id=None, pb=None):
1508
"""See InterRepository.fetch()."""
1509
from bzrlib.fetch import GenericRepoFetcher
1510
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1511
self.source, self.source._format, self.target, self.target._format)
1512
# TODO: jam 20070210 This should be an assert, not a translate
1513
revision_id = osutils.safe_revision_id(revision_id)
1514
f = GenericRepoFetcher(to_repository=self.target,
1515
from_repository=self.source,
1516
last_revision=revision_id,
1518
return f.count_copied, f.failed_revisions
1521
def missing_revision_ids(self, revision_id=None):
1522
"""See InterRepository.missing_revision_ids()."""
1523
# we want all revisions to satisfy revision_id in source.
1524
# but we don't want to stat every file here and there.
1525
# we want then, all revisions other needs to satisfy revision_id
1526
# checked, but not those that we have locally.
1527
# so the first thing is to get a subset of the revisions to
1528
# satisfy revision_id in source, and then eliminate those that
1529
# we do already have.
1530
# this is slow on high latency connection to self, but as as this
1531
# disk format scales terribly for push anyway due to rewriting
1532
# inventory.weave, this is considered acceptable.
1534
if revision_id is not None:
1535
source_ids = self.source.get_ancestry(revision_id)
1536
assert source_ids[0] is None
1539
source_ids = self.source._all_possible_ids()
1540
source_ids_set = set(source_ids)
1541
# source_ids is the worst possible case we may need to pull.
1542
# now we want to filter source_ids against what we actually
1543
# have in target, but don't try to check for existence where we know
1544
# we do not have a revision as that would be pointless.
1545
target_ids = set(self.target._all_possible_ids())
1546
possibly_present_revisions = target_ids.intersection(source_ids_set)
1547
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1548
required_revisions = source_ids_set.difference(actually_present_revisions)
1549
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1550
if revision_id is not None:
1551
# we used get_ancestry to determine source_ids then we are assured all
1552
# revisions referenced are present as they are installed in topological order.
1553
# and the tip revision was validated by get_ancestry.
1554
return required_topo_revisions
1556
# if we just grabbed the possibly available ids, then
1557
# we only have an estimate of whats available and need to validate
1558
# that against the revision records.
1559
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1562
class InterKnitRepo(InterSameDataRepository):
1563
"""Optimised code paths between Knit based repositories."""
1566
def _get_repo_format_to_test(self):
1567
from bzrlib.repofmt import knitrepo
1568
return knitrepo.RepositoryFormatKnit1()
1571
def is_compatible(source, target):
1572
"""Be compatible with known Knit formats.
1574
We don't test for the stores being of specific types because that
1575
could lead to confusing results, and there is no need to be
1578
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1580
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1581
isinstance(target._format, (RepositoryFormatKnit1)))
1582
except AttributeError:
1586
def fetch(self, revision_id=None, pb=None):
1587
"""See InterRepository.fetch()."""
1588
from bzrlib.fetch import KnitRepoFetcher
1589
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1590
self.source, self.source._format, self.target, self.target._format)
1591
# TODO: jam 20070210 This should be an assert, not a translate
1592
revision_id = osutils.safe_revision_id(revision_id)
1593
f = KnitRepoFetcher(to_repository=self.target,
1594
from_repository=self.source,
1595
last_revision=revision_id,
1597
return f.count_copied, f.failed_revisions
1600
def missing_revision_ids(self, revision_id=None):
1601
"""See InterRepository.missing_revision_ids()."""
1602
if revision_id is not None:
1603
source_ids = self.source.get_ancestry(revision_id)
1604
assert source_ids[0] is None
1607
source_ids = self.source._all_possible_ids()
1608
source_ids_set = set(source_ids)
1609
# source_ids is the worst possible case we may need to pull.
1610
# now we want to filter source_ids against what we actually
1611
# have in target, but don't try to check for existence where we know
1612
# we do not have a revision as that would be pointless.
1613
target_ids = set(self.target._all_possible_ids())
1614
possibly_present_revisions = target_ids.intersection(source_ids_set)
1615
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1616
required_revisions = source_ids_set.difference(actually_present_revisions)
1617
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1618
if revision_id is not None:
1619
# we used get_ancestry to determine source_ids then we are assured all
1620
# revisions referenced are present as they are installed in topological order.
1621
# and the tip revision was validated by get_ancestry.
1622
return required_topo_revisions
1624
# if we just grabbed the possibly available ids, then
1625
# we only have an estimate of whats available and need to validate
1626
# that against the revision records.
1627
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1630
class InterModel1and2(InterRepository):
1633
def _get_repo_format_to_test(self):
1637
def is_compatible(source, target):
1638
if not source.supports_rich_root() and target.supports_rich_root():
1644
def fetch(self, revision_id=None, pb=None):
1645
"""See InterRepository.fetch()."""
1646
from bzrlib.fetch import Model1toKnit2Fetcher
1647
# TODO: jam 20070210 This should be an assert, not a translate
1648
revision_id = osutils.safe_revision_id(revision_id)
1649
f = Model1toKnit2Fetcher(to_repository=self.target,
1650
from_repository=self.source,
1651
last_revision=revision_id,
1653
return f.count_copied, f.failed_revisions
1656
def copy_content(self, revision_id=None):
1657
"""Make a complete copy of the content in self into destination.
1659
This is a destructive operation! Do not use it on existing
1662
:param revision_id: Only copy the content needed to construct
1663
revision_id and its parents.
1666
self.target.set_make_working_trees(self.source.make_working_trees())
1667
except NotImplementedError:
1669
# TODO: jam 20070210 Internal, assert, don't translate
1670
revision_id = osutils.safe_revision_id(revision_id)
1671
# but don't bother fetching if we have the needed data now.
1672
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1673
self.target.has_revision(revision_id)):
1675
self.target.fetch(self.source, revision_id=revision_id)
1678
class InterKnit1and2(InterKnitRepo):
1681
def _get_repo_format_to_test(self):
1685
def is_compatible(source, target):
1686
"""Be compatible with Knit1 source and Knit3 target"""
1687
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
1689
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1690
RepositoryFormatKnit3
1691
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1692
isinstance(target._format, (RepositoryFormatKnit3)))
1693
except AttributeError:
1697
def fetch(self, revision_id=None, pb=None):
1698
"""See InterRepository.fetch()."""
1699
from bzrlib.fetch import Knit1to2Fetcher
1700
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1701
self.source, self.source._format, self.target,
1702
self.target._format)
1703
# TODO: jam 20070210 This should be an assert, not a translate
1704
revision_id = osutils.safe_revision_id(revision_id)
1705
f = Knit1to2Fetcher(to_repository=self.target,
1706
from_repository=self.source,
1707
last_revision=revision_id,
1709
return f.count_copied, f.failed_revisions
1712
class InterRemoteRepository(InterRepository):
1713
"""Code for converting between RemoteRepository objects.
1715
This just gets an non-remote repository from the RemoteRepository, and calls
1716
InterRepository.get again.
1719
def __init__(self, source, target):
1720
if isinstance(source, remote.RemoteRepository):
1721
source._ensure_real()
1722
real_source = source._real_repository
1724
real_source = source
1725
if isinstance(target, remote.RemoteRepository):
1726
target._ensure_real()
1727
real_target = target._real_repository
1729
real_target = target
1730
self.real_inter = InterRepository.get(real_source, real_target)
1733
def is_compatible(source, target):
1734
if isinstance(source, remote.RemoteRepository):
1736
if isinstance(target, remote.RemoteRepository):
1740
def copy_content(self, revision_id=None):
1741
self.real_inter.copy_content(revision_id=revision_id)
1743
def fetch(self, revision_id=None, pb=None):
1744
self.real_inter.fetch(revision_id=revision_id, pb=pb)
1747
def _get_repo_format_to_test(self):
1751
InterRepository.register_optimiser(InterSameDataRepository)
1752
InterRepository.register_optimiser(InterWeaveRepo)
1753
InterRepository.register_optimiser(InterKnitRepo)
1754
InterRepository.register_optimiser(InterModel1and2)
1755
InterRepository.register_optimiser(InterKnit1and2)
1756
InterRepository.register_optimiser(InterRemoteRepository)
1759
class RepositoryTestProviderAdapter(object):
1760
"""A tool to generate a suite testing multiple repository formats at once.
1762
This is done by copying the test once for each transport and injecting
1763
the transport_server, transport_readonly_server, and bzrdir_format and
1764
repository_format classes into each copy. Each copy is also given a new id()
1765
to make it easy to identify.
1768
def __init__(self, transport_server, transport_readonly_server, formats,
1769
vfs_transport_factory=None):
1770
self._transport_server = transport_server
1771
self._transport_readonly_server = transport_readonly_server
1772
self._vfs_transport_factory = vfs_transport_factory
1773
self._formats = formats
1775
def adapt(self, test):
1776
result = unittest.TestSuite()
1777
for repository_format, bzrdir_format in self._formats:
1778
from copy import deepcopy
1779
new_test = deepcopy(test)
1780
new_test.transport_server = self._transport_server
1781
new_test.transport_readonly_server = self._transport_readonly_server
1782
# Only override the test's vfs_transport_factory if one was
1783
# specified, otherwise just leave the default in place.
1784
if self._vfs_transport_factory:
1785
new_test.vfs_transport_factory = self._vfs_transport_factory
1786
new_test.bzrdir_format = bzrdir_format
1787
new_test.repository_format = repository_format
1788
def make_new_test_id():
1789
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1790
return lambda: new_id
1791
new_test.id = make_new_test_id()
1792
result.addTest(new_test)
1796
class InterRepositoryTestProviderAdapter(object):
1797
"""A tool to generate a suite testing multiple inter repository formats.
1799
This is done by copying the test once for each interrepo provider and injecting
1800
the transport_server, transport_readonly_server, repository_format and
1801
repository_to_format classes into each copy.
1802
Each copy is also given a new id() to make it easy to identify.
1805
def __init__(self, transport_server, transport_readonly_server, formats):
1806
self._transport_server = transport_server
1807
self._transport_readonly_server = transport_readonly_server
1808
self._formats = formats
1810
def adapt(self, test):
1811
result = unittest.TestSuite()
1812
for interrepo_class, repository_format, repository_format_to in self._formats:
1813
from copy import deepcopy
1814
new_test = deepcopy(test)
1815
new_test.transport_server = self._transport_server
1816
new_test.transport_readonly_server = self._transport_readonly_server
1817
new_test.interrepo_class = interrepo_class
1818
new_test.repository_format = repository_format
1819
new_test.repository_format_to = repository_format_to
1820
def make_new_test_id():
1821
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1822
return lambda: new_id
1823
new_test.id = make_new_test_id()
1824
result.addTest(new_test)
1828
def default_test_list():
1829
"""Generate the default list of interrepo permutations to test."""
1830
from bzrlib.repofmt import knitrepo, weaverepo
1832
# test the default InterRepository between format 6 and the current
1834
# XXX: robertc 20060220 reinstate this when there are two supported
1835
# formats which do not have an optimal code path between them.
1836
#result.append((InterRepository,
1837
# RepositoryFormat6(),
1838
# RepositoryFormatKnit1()))
1839
for optimiser_class in InterRepository._optimisers:
1840
format_to_test = optimiser_class._get_repo_format_to_test()
1841
if format_to_test is not None:
1842
result.append((optimiser_class,
1843
format_to_test, format_to_test))
1844
# if there are specific combinations we want to use, we can add them
1846
result.append((InterModel1and2,
1847
weaverepo.RepositoryFormat5(),
1848
knitrepo.RepositoryFormatKnit3()))
1849
result.append((InterKnit1and2,
1850
knitrepo.RepositoryFormatKnit1(),
1851
knitrepo.RepositoryFormatKnit3()))
1855
class CopyConverter(object):
1856
"""A repository conversion tool which just performs a copy of the content.
1858
This is slow but quite reliable.
1861
def __init__(self, target_format):
1862
"""Create a CopyConverter.
1864
:param target_format: The format the resulting repository should be.
1866
self.target_format = target_format
1868
def convert(self, repo, pb):
1869
"""Perform the conversion of to_convert, giving feedback via pb.
1871
:param to_convert: The disk object to convert.
1872
:param pb: a progress bar to use for progress information.
1877
# this is only useful with metadir layouts - separated repo content.
1878
# trigger an assertion if not such
1879
repo._format.get_format_string()
1880
self.repo_dir = repo.bzrdir
1881
self.step('Moving repository to repository.backup')
1882
self.repo_dir.transport.move('repository', 'repository.backup')
1883
backup_transport = self.repo_dir.transport.clone('repository.backup')
1884
repo._format.check_conversion_target(self.target_format)
1885
self.source_repo = repo._format.open(self.repo_dir,
1887
_override_transport=backup_transport)
1888
self.step('Creating new repository')
1889
converted = self.target_format.initialize(self.repo_dir,
1890
self.source_repo.is_shared())
1891
converted.lock_write()
1893
self.step('Copying content into repository.')
1894
self.source_repo.copy_content_into(converted)
1897
self.step('Deleting old repository content.')
1898
self.repo_dir.transport.delete_tree('repository.backup')
1899
self.pb.note('repository converted')
1901
def step(self, message):
1902
"""Update the pb by a step."""
1904
self.pb.update(message, self.count, self.total)
1907
class CommitBuilder(object):
1908
"""Provides an interface to build up a commit.
1910
This allows describing a tree to be committed without needing to
1911
know the internals of the format of the repository.
1914
record_root_entry = False
1915
def __init__(self, repository, parents, config, timestamp=None,
1916
timezone=None, committer=None, revprops=None,
1918
"""Initiate a CommitBuilder.
1920
:param repository: Repository to commit to.
1921
:param parents: Revision ids of the parents of the new revision.
1922
:param config: Configuration to use.
1923
:param timestamp: Optional timestamp recorded for commit.
1924
:param timezone: Optional timezone for timestamp.
1925
:param committer: Optional committer to set for commit.
1926
:param revprops: Optional dictionary of revision properties.
1927
:param revision_id: Optional revision id.
1929
self._config = config
1931
if committer is None:
1932
self._committer = self._config.username()
1934
assert isinstance(committer, basestring), type(committer)
1935
self._committer = committer
1937
self.new_inventory = Inventory(None)
1938
self._new_revision_id = osutils.safe_revision_id(revision_id)
1939
self.parents = parents
1940
self.repository = repository
1943
if revprops is not None:
1944
self._revprops.update(revprops)
1946
if timestamp is None:
1947
timestamp = time.time()
1948
# Restrict resolution to 1ms
1949
self._timestamp = round(timestamp, 3)
1951
if timezone is None:
1952
self._timezone = osutils.local_time_offset()
1954
self._timezone = int(timezone)
1956
self._generate_revision_if_needed()
1958
def commit(self, message):
1959
"""Make the actual commit.
1961
:return: The revision id of the recorded revision.
1963
rev = _mod_revision.Revision(
1964
timestamp=self._timestamp,
1965
timezone=self._timezone,
1966
committer=self._committer,
1968
inventory_sha1=self.inv_sha1,
1969
revision_id=self._new_revision_id,
1970
properties=self._revprops)
1971
rev.parent_ids = self.parents
1972
self.repository.add_revision(self._new_revision_id, rev,
1973
self.new_inventory, self._config)
1974
return self._new_revision_id
1976
def revision_tree(self):
1977
"""Return the tree that was just committed.
1979
After calling commit() this can be called to get a RevisionTree
1980
representing the newly committed tree. This is preferred to
1981
calling Repository.revision_tree() because that may require
1982
deserializing the inventory, while we already have a copy in
1985
return RevisionTree(self.repository, self.new_inventory,
1986
self._new_revision_id)
1988
def finish_inventory(self):
1989
"""Tell the builder that the inventory is finished."""
1990
if self.new_inventory.root is None:
1991
symbol_versioning.warn('Root entry should be supplied to'
1992
' record_entry_contents, as of bzr 0.10.',
1993
DeprecationWarning, stacklevel=2)
1994
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1995
self.new_inventory.revision_id = self._new_revision_id
1996
self.inv_sha1 = self.repository.add_inventory(
1997
self._new_revision_id,
2002
def _gen_revision_id(self):
2003
"""Return new revision-id."""
2004
return generate_ids.gen_revision_id(self._config.username(),
2007
def _generate_revision_if_needed(self):
2008
"""Create a revision id if None was supplied.
2010
If the repository can not support user-specified revision ids
2011
they should override this function and raise CannotSetRevisionId
2012
if _new_revision_id is not None.
2014
:raises: CannotSetRevisionId
2016
if self._new_revision_id is None:
2017
self._new_revision_id = self._gen_revision_id()
2019
def record_entry_contents(self, ie, parent_invs, path, tree):
2020
"""Record the content of ie from tree into the commit if needed.
2022
Side effect: sets ie.revision when unchanged
2024
:param ie: An inventory entry present in the commit.
2025
:param parent_invs: The inventories of the parent revisions of the
2027
:param path: The path the entry is at in the tree.
2028
:param tree: The tree which contains this entry and should be used to
2031
if self.new_inventory.root is None and ie.parent_id is not None:
2032
symbol_versioning.warn('Root entry should be supplied to'
2033
' record_entry_contents, as of bzr 0.10.',
2034
DeprecationWarning, stacklevel=2)
2035
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2037
self.new_inventory.add(ie)
2039
# ie.revision is always None if the InventoryEntry is considered
2040
# for committing. ie.snapshot will record the correct revision
2041
# which may be the sole parent if it is untouched.
2042
if ie.revision is not None:
2045
# In this revision format, root entries have no knit or weave
2046
if ie is self.new_inventory.root:
2047
# When serializing out to disk and back in
2048
# root.revision is always _new_revision_id
2049
ie.revision = self._new_revision_id
2051
previous_entries = ie.find_previous_heads(
2053
self.repository.weave_store,
2054
self.repository.get_transaction())
2055
# we are creating a new revision for ie in the history store
2057
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2059
def modified_directory(self, file_id, file_parents):
2060
"""Record the presence of a symbolic link.
2062
:param file_id: The file_id of the link to record.
2063
:param file_parents: The per-file parent revision ids.
2065
self._add_text_to_weave(file_id, [], file_parents.keys())
2067
def modified_reference(self, file_id, file_parents):
2068
"""Record the modification of a reference.
2070
:param file_id: The file_id of the link to record.
2071
:param file_parents: The per-file parent revision ids.
2073
self._add_text_to_weave(file_id, [], file_parents.keys())
2075
def modified_file_text(self, file_id, file_parents,
2076
get_content_byte_lines, text_sha1=None,
2078
"""Record the text of file file_id
2080
:param file_id: The file_id of the file to record the text of.
2081
:param file_parents: The per-file parent revision ids.
2082
:param get_content_byte_lines: A callable which will return the byte
2084
:param text_sha1: Optional SHA1 of the file contents.
2085
:param text_size: Optional size of the file contents.
2087
# mutter('storing text of file {%s} in revision {%s} into %r',
2088
# file_id, self._new_revision_id, self.repository.weave_store)
2089
# special case to avoid diffing on renames or
2091
if (len(file_parents) == 1
2092
and text_sha1 == file_parents.values()[0].text_sha1
2093
and text_size == file_parents.values()[0].text_size):
2094
previous_ie = file_parents.values()[0]
2095
versionedfile = self.repository.weave_store.get_weave(file_id,
2096
self.repository.get_transaction())
2097
versionedfile.clone_text(self._new_revision_id,
2098
previous_ie.revision, file_parents.keys())
2099
return text_sha1, text_size
2101
new_lines = get_content_byte_lines()
2102
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2103
# should return the SHA1 and size
2104
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2105
return osutils.sha_strings(new_lines), \
2106
sum(map(len, new_lines))
2108
def modified_link(self, file_id, file_parents, link_target):
2109
"""Record the presence of a symbolic link.
2111
:param file_id: The file_id of the link to record.
2112
:param file_parents: The per-file parent revision ids.
2113
:param link_target: Target location of this link.
2115
self._add_text_to_weave(file_id, [], file_parents.keys())
2117
def _add_text_to_weave(self, file_id, new_lines, parents):
2118
versionedfile = self.repository.weave_store.get_weave_or_empty(
2119
file_id, self.repository.get_transaction())
2120
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2121
versionedfile.clear_cache()
2124
class _CommitBuilder(CommitBuilder):
2125
"""Temporary class so old CommitBuilders are detected properly
2127
Note: CommitBuilder works whether or not root entry is recorded.
2130
record_root_entry = True
2133
class RootCommitBuilder(CommitBuilder):
2134
"""This commitbuilder actually records the root id"""
2136
record_root_entry = True
2138
def record_entry_contents(self, ie, parent_invs, path, tree):
2139
"""Record the content of ie from tree into the commit if needed.
2141
Side effect: sets ie.revision when unchanged
2143
:param ie: An inventory entry present in the commit.
2144
:param parent_invs: The inventories of the parent revisions of the
2146
:param path: The path the entry is at in the tree.
2147
:param tree: The tree which contains this entry and should be used to
2150
assert self.new_inventory.root is not None or ie.parent_id is None
2151
self.new_inventory.add(ie)
2153
# ie.revision is always None if the InventoryEntry is considered
2154
# for committing. ie.snapshot will record the correct revision
2155
# which may be the sole parent if it is untouched.
2156
if ie.revision is not None:
2159
previous_entries = ie.find_previous_heads(
2161
self.repository.weave_store,
2162
self.repository.get_transaction())
2163
# we are creating a new revision for ie in the history store
2165
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2177
def _unescaper(match, _map=_unescape_map):
2178
code = match.group(1)
2182
if not code.startswith('#'):
2184
return unichr(int(code[1:])).encode('utf8')
2190
def _unescape_xml(data):
2191
"""Unescape predefined XML entities in a string of data."""
2193
if _unescape_re is None:
2194
_unescape_re = re.compile('\&([^;]*);')
2195
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
