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(), """
21
from binascii import hexlify
22
from copy import deepcopy
41
revision as _mod_revision,
50
from bzrlib.osutils import (
55
from bzrlib.revisiontree import RevisionTree
56
from bzrlib.store.versioned import VersionedFileStore
57
from bzrlib.store.text import TextStore
58
from bzrlib.testament import Testament
61
from bzrlib.decorators import needs_read_lock, needs_write_lock
62
from bzrlib.inter import InterObject
63
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
64
from bzrlib.symbol_versioning import (
68
from bzrlib.trace import mutter, note, warning
71
# Old formats display a warning, but only once
72
_deprecation_warning_done = False
75
class Repository(object):
76
"""Repository holding history for one or more branches.
78
The repository holds and retrieves historical information including
79
revisions and file history. It's normally accessed only by the Branch,
80
which views a particular line of development through that history.
82
The Repository builds on top of Stores and a Transport, which respectively
83
describe the disk data format and the way of accessing the (possibly
87
_file_ids_altered_regex = lazy_regex.lazy_compile(
88
r'file_id="(?P<file_id>[^"]+)"'
89
r'.*revision="(?P<revision_id>[^"]+)"'
93
def add_inventory(self, revid, inv, parents):
94
"""Add the inventory inv to the repository as revid.
96
:param parents: The revision ids of the parents that revid
97
is known to have and are in the repository already.
99
returns the sha1 of the serialized inventory.
101
_mod_revision.check_not_reserved_id(revid)
102
assert inv.revision_id is None or inv.revision_id == revid, \
103
"Mismatch between inventory revision" \
104
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
105
assert inv.root is not None
106
inv_text = self.serialise_inventory(inv)
107
inv_sha1 = osutils.sha_string(inv_text)
108
inv_vf = self.control_weaves.get_weave('inventory',
109
self.get_transaction())
110
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
113
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
115
for parent in parents:
117
final_parents.append(parent)
119
inv_vf.add_lines(revid, final_parents, lines)
122
def add_revision(self, rev_id, rev, inv=None, config=None):
123
"""Add rev to the revision store as rev_id.
125
:param rev_id: the revision id to use.
126
:param rev: The revision object.
127
:param inv: The inventory for the revision. if None, it will be looked
128
up in the inventory storer
129
:param config: If None no digital signature will be created.
130
If supplied its signature_needed method will be used
131
to determine if a signature should be made.
133
_mod_revision.check_not_reserved_id(rev_id)
134
if config is not None and config.signature_needed():
136
inv = self.get_inventory(rev_id)
137
plaintext = Testament(rev, inv).as_short_text()
138
self.store_revision_signature(
139
gpg.GPGStrategy(config), plaintext, rev_id)
140
if not rev_id in self.get_inventory_weave():
142
raise errors.WeaveRevisionNotPresent(rev_id,
143
self.get_inventory_weave())
145
# yes, this is not suitable for adding with ghosts.
146
self.add_inventory(rev_id, inv, rev.parent_ids)
147
self._revision_store.add_revision(rev, self.get_transaction())
150
def _all_possible_ids(self):
151
"""Return all the possible revisions that we could find."""
152
return self.get_inventory_weave().versions()
154
def all_revision_ids(self):
155
"""Returns a list of all the revision ids in the repository.
157
This is deprecated because code should generally work on the graph
158
reachable from a particular revision, and ignore any other revisions
159
that might be present. There is no direct replacement method.
161
return self._all_revision_ids()
164
def _all_revision_ids(self):
165
"""Returns a list of all the revision ids in the repository.
167
These are in as much topological order as the underlying store can
168
present: for weaves ghosts may lead to a lack of correctness until
169
the reweave updates the parents list.
171
if self._revision_store.text_store.listable():
172
return self._revision_store.all_revision_ids(self.get_transaction())
173
result = self._all_possible_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()
227
self._serializer = xml5.serializer_v5
230
return '%s(%r)' % (self.__class__.__name__,
231
self.bzrdir.transport.base)
234
return self.control_files.is_locked()
236
def lock_write(self):
237
self.control_files.lock_write()
240
self.control_files.lock_read()
242
def get_physical_lock_status(self):
243
return self.control_files.get_physical_lock_status()
246
def missing_revision_ids(self, other, revision_id=None):
247
"""Return the revision ids that other has that this does not.
249
These are returned in topological order.
251
revision_id: only return revision ids included by revision_id.
253
return InterRepository.get(other, self).missing_revision_ids(revision_id)
257
"""Open the repository rooted at base.
259
For instance, if the repository is at URL/.bzr/repository,
260
Repository.open(URL) -> a Repository instance.
262
control = bzrdir.BzrDir.open(base)
263
return control.open_repository()
265
def copy_content_into(self, destination, revision_id=None, basis=None):
266
"""Make a complete copy of the content in self into destination.
268
This is a destructive operation! Do not use it on existing
271
return InterRepository.get(self, destination).copy_content(revision_id, basis)
273
def fetch(self, source, revision_id=None, pb=None):
274
"""Fetch the content required to construct revision_id from source.
276
If revision_id is None all content is copied.
278
return InterRepository.get(source, self).fetch(revision_id=revision_id,
281
def get_commit_builder(self, branch, parents, config, timestamp=None,
282
timezone=None, committer=None, revprops=None,
284
"""Obtain a CommitBuilder for this repository.
286
:param branch: Branch to commit to.
287
:param parents: Revision ids of the parents of the new revision.
288
:param config: Configuration to use.
289
:param timestamp: Optional timestamp recorded for commit.
290
:param timezone: Optional timezone for timestamp.
291
:param committer: Optional committer to set for commit.
292
:param revprops: Optional dictionary of revision properties.
293
:param revision_id: Optional revision id.
295
return _CommitBuilder(self, parents, config, timestamp, timezone,
296
committer, revprops, revision_id)
299
self.control_files.unlock()
302
def clone(self, a_bzrdir, revision_id=None, basis=None):
303
"""Clone this repository into a_bzrdir using the current format.
305
Currently no check is made that the format of this repository and
306
the bzrdir format are compatible. FIXME RBC 20060201.
308
:return: The newly created destination repository.
310
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
311
# use target default format.
312
dest_repo = a_bzrdir.create_repository()
314
# Most control formats need the repository to be specifically
315
# created, but on some old all-in-one formats it's not needed
317
dest_repo = self._format.initialize(a_bzrdir, shared=self.is_shared())
318
except errors.UninitializableFormat:
319
dest_repo = a_bzrdir.open_repository()
320
self.copy_content_into(dest_repo, revision_id, basis)
324
def has_revision(self, revision_id):
325
"""True if this repository has a copy of the revision."""
326
return self._revision_store.has_revision_id(revision_id,
327
self.get_transaction())
330
def get_revision_reconcile(self, revision_id):
331
"""'reconcile' helper routine that allows access to a revision always.
333
This variant of get_revision does not cross check the weave graph
334
against the revision one as get_revision does: but it should only
335
be used by reconcile, or reconcile-alike commands that are correcting
336
or testing the revision graph.
338
if not revision_id or not isinstance(revision_id, basestring):
339
raise errors.InvalidRevisionId(revision_id=revision_id,
341
return self._revision_store.get_revisions([revision_id],
342
self.get_transaction())[0]
344
def get_revisions(self, revision_ids):
345
return self._revision_store.get_revisions(revision_ids,
346
self.get_transaction())
349
def get_revision_xml(self, revision_id):
350
rev = self.get_revision(revision_id)
352
# the current serializer..
353
self._revision_store._serializer.write_revision(rev, rev_tmp)
355
return rev_tmp.getvalue()
358
def get_revision(self, revision_id):
359
"""Return the Revision object for a named revision"""
360
r = self.get_revision_reconcile(revision_id)
361
# weave corruption can lead to absent revision markers that should be
363
# the following test is reasonably cheap (it needs a single weave read)
364
# and the weave is cached in read transactions. In write transactions
365
# it is not cached but typically we only read a small number of
366
# revisions. For knits when they are introduced we will probably want
367
# to ensure that caching write transactions are in use.
368
inv = self.get_inventory_weave()
369
self._check_revision_parents(r, inv)
373
def get_deltas_for_revisions(self, revisions):
374
"""Produce a generator of revision deltas.
376
Note that the input is a sequence of REVISIONS, not revision_ids.
377
Trees will be held in memory until the generator exits.
378
Each delta is relative to the revision's lefthand predecessor.
380
required_trees = set()
381
for revision in revisions:
382
required_trees.add(revision.revision_id)
383
required_trees.update(revision.parent_ids[:1])
384
trees = dict((t.get_revision_id(), t) for
385
t in self.revision_trees(required_trees))
386
for revision in revisions:
387
if not revision.parent_ids:
388
old_tree = self.revision_tree(None)
390
old_tree = trees[revision.parent_ids[0]]
391
yield trees[revision.revision_id].changes_from(old_tree)
394
def get_revision_delta(self, revision_id):
395
"""Return the delta for one revision.
397
The delta is relative to the left-hand predecessor of the
400
r = self.get_revision(revision_id)
401
return list(self.get_deltas_for_revisions([r]))[0]
403
def _check_revision_parents(self, revision, inventory):
404
"""Private to Repository and Fetch.
406
This checks the parentage of revision in an inventory weave for
407
consistency and is only applicable to inventory-weave-for-ancestry
408
using repository formats & fetchers.
410
weave_parents = inventory.get_parents(revision.revision_id)
411
weave_names = inventory.versions()
412
for parent_id in revision.parent_ids:
413
if parent_id in weave_names:
414
# this parent must not be a ghost.
415
if not parent_id in weave_parents:
417
raise errors.CorruptRepository(self)
420
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
421
signature = gpg_strategy.sign(plaintext)
422
self._revision_store.add_revision_signature_text(revision_id,
424
self.get_transaction())
426
def fileids_altered_by_revision_ids(self, revision_ids):
427
"""Find the file ids and versions affected by revisions.
429
:param revisions: an iterable containing revision ids.
430
:return: a dictionary mapping altered file-ids to an iterable of
431
revision_ids. Each altered file-ids has the exact revision_ids that
432
altered it listed explicitly.
434
assert self._serializer.support_altered_by_hack, \
435
("fileids_altered_by_revision_ids only supported for branches "
436
"which store inventory as unnested xml, not on %r" % self)
437
selected_revision_ids = set(revision_ids)
438
w = self.get_inventory_weave()
441
# this code needs to read every new line in every inventory for the
442
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
443
# not present in one of those inventories is unnecessary but not
444
# harmful because we are filtering by the revision id marker in the
445
# inventory lines : we only select file ids altered in one of those
446
# revisions. We don't need to see all lines in the inventory because
447
# only those added in an inventory in rev X can contain a revision=X
449
unescape_revid_cache = {}
450
unescape_fileid_cache = {}
452
# jam 20061218 In a big fetch, this handles hundreds of thousands
453
# of lines, so it has had a lot of inlining and optimizing done.
454
# Sorry that it is a little bit messy.
455
# Move several functions to be local variables, since this is a long
457
search = self._file_ids_altered_regex.search
458
unescape = _unescape_xml
459
setdefault = result.setdefault
460
pb = ui.ui_factory.nested_progress_bar()
462
for line in w.iter_lines_added_or_present_in_versions(
463
selected_revision_ids, pb=pb):
467
# One call to match.group() returning multiple items is quite a
468
# bit faster than 2 calls to match.group() each returning 1
469
file_id, revision_id = match.group('file_id', 'revision_id')
471
# Inlining the cache lookups helps a lot when you make 170,000
472
# lines and 350k ids, versus 8.4 unique ids.
473
# Using a cache helps in 2 ways:
474
# 1) Avoids unnecessary decoding calls
475
# 2) Re-uses cached strings, which helps in future set and
477
# (2) is enough that removing encoding entirely along with
478
# the cache (so we are using plain strings) results in no
479
# performance improvement.
481
revision_id = unescape_revid_cache[revision_id]
483
unescaped = unescape(revision_id)
484
unescape_revid_cache[revision_id] = unescaped
485
revision_id = unescaped
487
if revision_id in selected_revision_ids:
489
file_id = unescape_fileid_cache[file_id]
491
unescaped = unescape(file_id)
492
unescape_fileid_cache[file_id] = unescaped
494
setdefault(file_id, set()).add(revision_id)
500
def get_inventory_weave(self):
501
return self.control_weaves.get_weave('inventory',
502
self.get_transaction())
505
def get_inventory(self, revision_id):
506
"""Get Inventory object by hash."""
507
return self.deserialise_inventory(
508
revision_id, self.get_inventory_xml(revision_id))
510
def deserialise_inventory(self, revision_id, xml):
511
"""Transform the xml into an inventory object.
513
:param revision_id: The expected revision id of the inventory.
514
:param xml: A serialised inventory.
516
result = self._serializer.read_inventory_from_string(xml)
517
result.root.revision = revision_id
520
def serialise_inventory(self, inv):
521
return self._serializer.write_inventory_to_string(inv)
524
def get_inventory_xml(self, revision_id):
525
"""Get inventory XML as a file object."""
527
assert isinstance(revision_id, basestring), type(revision_id)
528
iw = self.get_inventory_weave()
529
return iw.get_text(revision_id)
531
raise errors.HistoryMissing(self, 'inventory', revision_id)
534
def get_inventory_sha1(self, revision_id):
535
"""Return the sha1 hash of the inventory entry
537
return self.get_revision(revision_id).inventory_sha1
540
def get_revision_graph(self, revision_id=None):
541
"""Return a dictionary containing the revision graph.
543
:param revision_id: The revision_id to get a graph from. If None, then
544
the entire revision graph is returned. This is a deprecated mode of
545
operation and will be removed in the future.
546
:return: a dictionary of revision_id->revision_parents_list.
548
# special case NULL_REVISION
549
if revision_id == _mod_revision.NULL_REVISION:
551
a_weave = self.get_inventory_weave()
552
all_revisions = self._eliminate_revisions_not_present(
554
entire_graph = dict([(node, a_weave.get_parents(node)) for
555
node in all_revisions])
556
if revision_id is None:
558
elif revision_id not in entire_graph:
559
raise errors.NoSuchRevision(self, revision_id)
561
# add what can be reached from revision_id
563
pending = set([revision_id])
564
while len(pending) > 0:
566
result[node] = entire_graph[node]
567
for revision_id in result[node]:
568
if revision_id not in result:
569
pending.add(revision_id)
573
def get_revision_graph_with_ghosts(self, revision_ids=None):
574
"""Return a graph of the revisions with ghosts marked as applicable.
576
:param revision_ids: an iterable of revisions to graph or None for all.
577
:return: a Graph object with the graph reachable from revision_ids.
579
result = graph.Graph()
581
pending = set(self.all_revision_ids())
584
pending = set(revision_ids)
585
# special case NULL_REVISION
586
if _mod_revision.NULL_REVISION in pending:
587
pending.remove(_mod_revision.NULL_REVISION)
588
required = set(pending)
591
revision_id = pending.pop()
593
rev = self.get_revision(revision_id)
594
except errors.NoSuchRevision:
595
if revision_id in required:
598
result.add_ghost(revision_id)
600
for parent_id in rev.parent_ids:
601
# is this queued or done ?
602
if (parent_id not in pending and
603
parent_id not in done):
605
pending.add(parent_id)
606
result.add_node(revision_id, rev.parent_ids)
607
done.add(revision_id)
611
def get_revision_inventory(self, revision_id):
612
"""Return inventory of a past revision."""
613
# TODO: Unify this with get_inventory()
614
# bzr 0.0.6 and later imposes the constraint that the inventory_id
615
# must be the same as its revision, so this is trivial.
616
if revision_id is None:
617
# This does not make sense: if there is no revision,
618
# then it is the current tree inventory surely ?!
619
# and thus get_root_id() is something that looks at the last
620
# commit on the branch, and the get_root_id is an inventory check.
621
raise NotImplementedError
622
# return Inventory(self.get_root_id())
624
return self.get_inventory(revision_id)
628
"""Return True if this repository is flagged as a shared repository."""
629
raise NotImplementedError(self.is_shared)
632
def reconcile(self, other=None, thorough=False):
633
"""Reconcile this repository."""
634
from bzrlib.reconcile import RepoReconciler
635
reconciler = RepoReconciler(self, thorough=thorough)
636
reconciler.reconcile()
640
def revision_tree(self, revision_id):
641
"""Return Tree for a revision on this branch.
643
`revision_id` may be None for the empty tree revision.
645
# TODO: refactor this to use an existing revision object
646
# so we don't need to read it in twice.
647
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
648
return RevisionTree(self, Inventory(root_id=None),
649
_mod_revision.NULL_REVISION)
651
inv = self.get_revision_inventory(revision_id)
652
return RevisionTree(self, inv, revision_id)
655
def revision_trees(self, revision_ids):
656
"""Return Tree for a revision on this branch.
658
`revision_id` may not be None or 'null:'"""
659
assert None not in revision_ids
660
assert _mod_revision.NULL_REVISION not in revision_ids
661
texts = self.get_inventory_weave().get_texts(revision_ids)
662
for text, revision_id in zip(texts, revision_ids):
663
inv = self.deserialise_inventory(revision_id, text)
664
yield RevisionTree(self, inv, revision_id)
667
def get_ancestry(self, revision_id):
668
"""Return a list of revision-ids integrated by a revision.
670
The first element of the list is always None, indicating the origin
671
revision. This might change when we have history horizons, or
672
perhaps we should have a new API.
674
This is topologically sorted.
676
if revision_id is None:
678
if not self.has_revision(revision_id):
679
raise errors.NoSuchRevision(self, revision_id)
680
w = self.get_inventory_weave()
681
candidates = w.get_ancestry(revision_id)
682
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
685
def print_file(self, file, revision_id):
686
"""Print `file` to stdout.
688
FIXME RBC 20060125 as John Meinel points out this is a bad api
689
- it writes to stdout, it assumes that that is valid etc. Fix
690
by creating a new more flexible convenience function.
692
tree = self.revision_tree(revision_id)
693
# use inventory as it was in that revision
694
file_id = tree.inventory.path2id(file)
696
# TODO: jam 20060427 Write a test for this code path
697
# it had a bug in it, and was raising the wrong
699
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
700
tree.print_file(file_id)
702
def get_transaction(self):
703
return self.control_files.get_transaction()
705
def revision_parents(self, revid):
706
return self.get_inventory_weave().parent_names(revid)
709
def set_make_working_trees(self, new_value):
710
"""Set the policy flag for making working trees when creating branches.
712
This only applies to branches that use this repository.
714
The default is 'True'.
715
:param new_value: True to restore the default, False to disable making
718
raise NotImplementedError(self.set_make_working_trees)
720
def make_working_trees(self):
721
"""Returns the policy for making working trees on new branches."""
722
raise NotImplementedError(self.make_working_trees)
725
def sign_revision(self, revision_id, gpg_strategy):
726
plaintext = Testament.from_revision(self, revision_id).as_short_text()
727
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
730
def has_signature_for_revision_id(self, revision_id):
731
"""Query for a revision signature for revision_id in the repository."""
732
return self._revision_store.has_signature(revision_id,
733
self.get_transaction())
736
def get_signature_text(self, revision_id):
737
"""Return the text for a signature."""
738
return self._revision_store.get_signature_text(revision_id,
739
self.get_transaction())
742
def check(self, revision_ids):
743
"""Check consistency of all history of given revision_ids.
745
Different repository implementations should override _check().
747
:param revision_ids: A non-empty list of revision_ids whose ancestry
748
will be checked. Typically the last revision_id of a branch.
751
raise ValueError("revision_ids must be non-empty in %s.check"
753
return self._check(revision_ids)
755
def _check(self, revision_ids):
756
result = check.Check(self)
760
def _warn_if_deprecated(self):
761
global _deprecation_warning_done
762
if _deprecation_warning_done:
764
_deprecation_warning_done = True
765
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
766
% (self._format, self.bzrdir.transport.base))
768
def supports_rich_root(self):
769
return self._format.rich_root_data
771
def _check_ascii_revisionid(self, revision_id, method):
772
"""Private helper for ascii-only repositories."""
773
# weave repositories refuse to store revisionids that are non-ascii.
774
if revision_id is not None:
775
# weaves require ascii revision ids.
776
if isinstance(revision_id, unicode):
778
revision_id.encode('ascii')
779
except UnicodeEncodeError:
780
raise errors.NonAsciiRevisionId(method, self)
783
def install_revision(repository, rev, revision_tree):
784
"""Install all revision data into a repository."""
787
for p_id in rev.parent_ids:
788
if repository.has_revision(p_id):
789
present_parents.append(p_id)
790
parent_trees[p_id] = repository.revision_tree(p_id)
792
parent_trees[p_id] = repository.revision_tree(None)
794
inv = revision_tree.inventory
795
entries = inv.iter_entries()
796
# backwards compatability hack: skip the root id.
797
if not repository.supports_rich_root():
798
path, root = entries.next()
799
if root.revision != rev.revision_id:
800
raise errors.IncompatibleRevision(repr(repository))
801
# Add the texts that are not already present
802
for path, ie in entries:
803
w = repository.weave_store.get_weave_or_empty(ie.file_id,
804
repository.get_transaction())
805
if ie.revision not in w:
807
# FIXME: TODO: The following loop *may* be overlapping/duplicate
808
# with InventoryEntry.find_previous_heads(). if it is, then there
809
# is a latent bug here where the parents may have ancestors of each
811
for revision, tree in parent_trees.iteritems():
812
if ie.file_id not in tree:
814
parent_id = tree.inventory[ie.file_id].revision
815
if parent_id in text_parents:
817
text_parents.append(parent_id)
819
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
820
repository.get_transaction())
821
lines = revision_tree.get_file(ie.file_id).readlines()
822
vfile.add_lines(rev.revision_id, text_parents, lines)
824
# install the inventory
825
repository.add_inventory(rev.revision_id, inv, present_parents)
826
except errors.RevisionAlreadyPresent:
828
repository.add_revision(rev.revision_id, rev, inv)
831
class MetaDirRepository(Repository):
832
"""Repositories in the new meta-dir layout."""
834
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
835
super(MetaDirRepository, self).__init__(_format,
841
dir_mode = self.control_files._dir_mode
842
file_mode = self.control_files._file_mode
846
"""Return True if this repository is flagged as a shared repository."""
847
return self.control_files._transport.has('shared-storage')
850
def set_make_working_trees(self, new_value):
851
"""Set the policy flag for making working trees when creating branches.
853
This only applies to branches that use this repository.
855
The default is 'True'.
856
:param new_value: True to restore the default, False to disable making
861
self.control_files._transport.delete('no-working-trees')
862
except errors.NoSuchFile:
865
self.control_files.put_utf8('no-working-trees', '')
867
def make_working_trees(self):
868
"""Returns the policy for making working trees on new branches."""
869
return not self.control_files._transport.has('no-working-trees')
872
class RepositoryFormatRegistry(registry.Registry):
873
"""Registry of RepositoryFormats.
877
format_registry = RepositoryFormatRegistry()
878
"""Registry of formats, indexed by their identifying format string."""
881
class RepositoryFormat(object):
882
"""A repository format.
884
Formats provide three things:
885
* An initialization routine to construct repository data on disk.
886
* a format string which is used when the BzrDir supports versioned
888
* an open routine which returns a Repository instance.
890
Formats are placed in an dict by their format string for reference
891
during opening. These should be subclasses of RepositoryFormat
894
Once a format is deprecated, just deprecate the initialize and open
895
methods on the format class. Do not deprecate the object, as the
896
object will be created every system load.
898
Common instance attributes:
899
_matchingbzrdir - the bzrdir format that the repository format was
900
originally written to work with. This can be used if manually
901
constructing a bzrdir and repository, or more commonly for test suite
906
return "<%s>" % self.__class__.__name__
909
def find_format(klass, a_bzrdir):
910
"""Return the format for the repository object in a_bzrdir.
912
This is used by bzr native formats that have a "format" file in
913
the repository. Other methods may be used by different types of
917
transport = a_bzrdir.get_repository_transport(None)
918
format_string = transport.get("format").read()
919
return format_registry.get(format_string)
920
except errors.NoSuchFile:
921
raise errors.NoRepositoryPresent(a_bzrdir)
923
raise errors.UnknownFormatError(format=format_string)
926
@deprecated_method(symbol_versioning.zero_fourteen)
927
def set_default_format(klass, format):
928
klass._set_default_format(format)
931
def _set_default_format(klass, format):
932
"""Set the default format for new Repository creation.
934
The format must already be registered.
936
format_registry.default_key = format.get_format_string()
939
def register_format(klass, format):
940
format_registry.register(format.get_format_string(), format)
943
def unregister_format(klass, format):
944
format_registry.remove(format.get_format_string())
947
def get_default_format(klass):
948
"""Return the current default format."""
949
return format_registry.get(format_registry.default_key)
951
def _get_control_store(self, repo_transport, control_files):
952
"""Return the control store for this repository."""
953
raise NotImplementedError(self._get_control_store)
955
def get_format_string(self):
956
"""Return the ASCII format string that identifies this format.
958
Note that in pre format ?? repositories the format string is
959
not permitted nor written to disk.
961
raise NotImplementedError(self.get_format_string)
963
def get_format_description(self):
964
"""Return the short description for this format."""
965
raise NotImplementedError(self.get_format_description)
967
def _get_revision_store(self, repo_transport, control_files):
968
"""Return the revision store object for this a_bzrdir."""
969
raise NotImplementedError(self._get_revision_store)
971
def _get_text_rev_store(self,
978
"""Common logic for getting a revision store for a repository.
980
see self._get_revision_store for the subclass-overridable method to
981
get the store for a repository.
983
from bzrlib.store.revision.text import TextRevisionStore
984
dir_mode = control_files._dir_mode
985
file_mode = control_files._file_mode
986
text_store =TextStore(transport.clone(name),
988
compressed=compressed,
991
_revision_store = TextRevisionStore(text_store, serializer)
992
return _revision_store
994
# TODO: this shouldn't be in the base class, it's specific to things that
995
# use weaves or knits -- mbp 20070207
996
def _get_versioned_file_store(self,
1001
versionedfile_class=weave.WeaveFile,
1002
versionedfile_kwargs={},
1004
weave_transport = control_files._transport.clone(name)
1005
dir_mode = control_files._dir_mode
1006
file_mode = control_files._file_mode
1007
return VersionedFileStore(weave_transport, prefixed=prefixed,
1009
file_mode=file_mode,
1010
versionedfile_class=versionedfile_class,
1011
versionedfile_kwargs=versionedfile_kwargs,
1014
def initialize(self, a_bzrdir, shared=False):
1015
"""Initialize a repository of this format in a_bzrdir.
1017
:param a_bzrdir: The bzrdir to put the new repository in it.
1018
:param shared: The repository should be initialized as a sharable one.
1020
This may raise UninitializableFormat if shared repository are not
1021
compatible the a_bzrdir.
1024
def is_supported(self):
1025
"""Is this format supported?
1027
Supported formats must be initializable and openable.
1028
Unsupported formats may not support initialization or committing or
1029
some other features depending on the reason for not being supported.
1033
def check_conversion_target(self, target_format):
1034
raise NotImplementedError(self.check_conversion_target)
1036
def open(self, a_bzrdir, _found=False):
1037
"""Return an instance of this format for the bzrdir a_bzrdir.
1039
_found is a private parameter, do not use it.
1041
raise NotImplementedError(self.open)
1044
class MetaDirRepositoryFormat(RepositoryFormat):
1045
"""Common base class for the new repositories using the metadir layout."""
1047
rich_root_data = False
1050
super(MetaDirRepositoryFormat, self).__init__()
1051
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1053
def _create_control_files(self, a_bzrdir):
1054
"""Create the required files and the initial control_files object."""
1055
# FIXME: RBC 20060125 don't peek under the covers
1056
# NB: no need to escape relative paths that are url safe.
1057
repository_transport = a_bzrdir.get_repository_transport(self)
1058
control_files = lockable_files.LockableFiles(repository_transport,
1059
'lock', lockdir.LockDir)
1060
control_files.create_lock()
1061
return control_files
1063
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1064
"""Upload the initial blank content."""
1065
control_files = self._create_control_files(a_bzrdir)
1066
control_files.lock_write()
1068
control_files._transport.mkdir_multi(dirs,
1069
mode=control_files._dir_mode)
1070
for file, content in files:
1071
control_files.put(file, content)
1072
for file, content in utf8_files:
1073
control_files.put_utf8(file, content)
1075
control_files.put_utf8('shared-storage', '')
1077
control_files.unlock()
1080
# formats which have no format string are not discoverable
1081
# and not independently creatable, so are not registered. They're
1082
# all in bzrlib.repofmt.weaverepo now.
1083
format_registry.register_lazy(
1084
'Bazaar-NG Repository format 7',
1085
'bzrlib.repofmt.weaverepo',
1086
'RepositoryFormat7_instance'
1088
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1089
# default control directory format
1091
format_registry.register_lazy(
1092
'Bazaar-NG Knit Repository Format 1',
1093
'bzrlib.repofmt.knitrepo',
1094
'RepositoryFormatKnit1_instance',
1096
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1098
format_registry.register_lazy(
1099
'Bazaar Knit Repository Format 2\n',
1100
'bzrlib.repofmt.knitrepo',
1101
'RepositoryFormatKnit2_instance',
1105
class InterRepository(InterObject):
1106
"""This class represents operations taking place between two repositories.
1108
Its instances have methods like copy_content and fetch, and contain
1109
references to the source and target repositories these operations can be
1112
Often we will provide convenience methods on 'repository' which carry out
1113
operations with another repository - they will always forward to
1114
InterRepository.get(other).method_name(parameters).
1118
"""The available optimised InterRepository types."""
1120
def copy_content(self, revision_id=None, basis=None):
1121
raise NotImplementedError(self.copy_content)
1123
def fetch(self, revision_id=None, pb=None):
1124
"""Fetch the content required to construct revision_id.
1126
The content is copied from self.source to self.target.
1128
:param revision_id: if None all content is copied, if NULL_REVISION no
1130
:param pb: optional progress bar to use for progress reports. If not
1131
provided a default one will be created.
1133
Returns the copied revision count and the failed revisions in a tuple:
1136
raise NotImplementedError(self.fetch)
1139
def missing_revision_ids(self, revision_id=None):
1140
"""Return the revision ids that source has that target does not.
1142
These are returned in topological order.
1144
:param revision_id: only return revision ids included by this
1147
# generic, possibly worst case, slow code path.
1148
target_ids = set(self.target.all_revision_ids())
1149
if revision_id is not None:
1150
source_ids = self.source.get_ancestry(revision_id)
1151
assert source_ids[0] is None
1154
source_ids = self.source.all_revision_ids()
1155
result_set = set(source_ids).difference(target_ids)
1156
# this may look like a no-op: its not. It preserves the ordering
1157
# other_ids had while only returning the members from other_ids
1158
# that we've decided we need.
1159
return [rev_id for rev_id in source_ids if rev_id in result_set]
1162
class InterSameDataRepository(InterRepository):
1163
"""Code for converting between repositories that represent the same data.
1165
Data format and model must match for this to work.
1169
def _get_matching_repo_format(self):
1170
"""Repository format for testing with."""
1171
return RepositoryFormat.get_default_format()
1174
def is_compatible(source, target):
1175
if not isinstance(source, Repository):
1177
if not isinstance(target, Repository):
1179
if source._format.rich_root_data == target._format.rich_root_data:
1185
def copy_content(self, revision_id=None, basis=None):
1186
"""Make a complete copy of the content in self into destination.
1188
This is a destructive operation! Do not use it on existing
1191
:param revision_id: Only copy the content needed to construct
1192
revision_id and its parents.
1193
:param basis: Copy the needed data preferentially from basis.
1196
self.target.set_make_working_trees(self.source.make_working_trees())
1197
except NotImplementedError:
1199
# grab the basis available data
1200
if basis is not None:
1201
self.target.fetch(basis, revision_id=revision_id)
1202
# but don't bother fetching if we have the needed data now.
1203
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1204
self.target.has_revision(revision_id)):
1206
self.target.fetch(self.source, revision_id=revision_id)
1209
def fetch(self, revision_id=None, pb=None):
1210
"""See InterRepository.fetch()."""
1211
from bzrlib.fetch import GenericRepoFetcher
1212
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1213
self.source, self.source._format, self.target,
1214
self.target._format)
1215
f = GenericRepoFetcher(to_repository=self.target,
1216
from_repository=self.source,
1217
last_revision=revision_id,
1219
return f.count_copied, f.failed_revisions
1222
class InterKnitRepo(InterSameDataRepository):
1223
"""Optimised code paths between Knit based repositories."""
1226
def _get_matching_repo_format(self):
1227
from bzrlib.repofmt import knitrepo
1228
return knitrepo.RepositoryFormatKnit1()
1231
def is_compatible(source, target):
1232
"""Be compatible with known Knit formats.
1234
We don't test for the stores being of specific types because that
1235
could lead to confusing results, and there is no need to be
1238
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1240
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1241
isinstance(target._format, (RepositoryFormatKnit1)))
1242
except AttributeError:
1246
def fetch(self, revision_id=None, pb=None):
1247
"""See InterRepository.fetch()."""
1248
from bzrlib.fetch import KnitRepoFetcher
1249
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1250
self.source, self.source._format, self.target, self.target._format)
1251
f = KnitRepoFetcher(to_repository=self.target,
1252
from_repository=self.source,
1253
last_revision=revision_id,
1255
return f.count_copied, f.failed_revisions
1258
def missing_revision_ids(self, revision_id=None):
1259
"""See InterRepository.missing_revision_ids()."""
1260
if revision_id is not None:
1261
source_ids = self.source.get_ancestry(revision_id)
1262
assert source_ids[0] is None
1265
source_ids = self.source._all_possible_ids()
1266
source_ids_set = set(source_ids)
1267
# source_ids is the worst possible case we may need to pull.
1268
# now we want to filter source_ids against what we actually
1269
# have in target, but don't try to check for existence where we know
1270
# we do not have a revision as that would be pointless.
1271
target_ids = set(self.target._all_possible_ids())
1272
possibly_present_revisions = target_ids.intersection(source_ids_set)
1273
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1274
required_revisions = source_ids_set.difference(actually_present_revisions)
1275
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1276
if revision_id is not None:
1277
# we used get_ancestry to determine source_ids then we are assured all
1278
# revisions referenced are present as they are installed in topological order.
1279
# and the tip revision was validated by get_ancestry.
1280
return required_topo_revisions
1282
# if we just grabbed the possibly available ids, then
1283
# we only have an estimate of whats available and need to validate
1284
# that against the revision records.
1285
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1288
class InterModel1and2(InterRepository):
1291
def _get_matching_repo_format(self):
1295
def is_compatible(source, target):
1296
if not isinstance(source, Repository):
1298
if not isinstance(target, Repository):
1300
if not source._format.rich_root_data and target._format.rich_root_data:
1306
def fetch(self, revision_id=None, pb=None):
1307
"""See InterRepository.fetch()."""
1308
from bzrlib.fetch import Model1toKnit2Fetcher
1309
f = Model1toKnit2Fetcher(to_repository=self.target,
1310
from_repository=self.source,
1311
last_revision=revision_id,
1313
return f.count_copied, f.failed_revisions
1316
def copy_content(self, revision_id=None, basis=None):
1317
"""Make a complete copy of the content in self into destination.
1319
This is a destructive operation! Do not use it on existing
1322
:param revision_id: Only copy the content needed to construct
1323
revision_id and its parents.
1324
:param basis: Copy the needed data preferentially from basis.
1327
self.target.set_make_working_trees(self.source.make_working_trees())
1328
except NotImplementedError:
1330
# grab the basis available data
1331
if basis is not None:
1332
self.target.fetch(basis, revision_id=revision_id)
1333
# but don't bother fetching if we have the needed data now.
1334
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1335
self.target.has_revision(revision_id)):
1337
self.target.fetch(self.source, revision_id=revision_id)
1340
class InterKnit1and2(InterKnitRepo):
1343
def _get_matching_repo_format(self):
1347
def is_compatible(source, target):
1348
"""Be compatible with Knit1 source and Knit2 target"""
1349
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit2
1351
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1352
RepositoryFormatKnit2
1353
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1354
isinstance(target._format, (RepositoryFormatKnit2)))
1355
except AttributeError:
1359
def fetch(self, revision_id=None, pb=None):
1360
"""See InterRepository.fetch()."""
1361
from bzrlib.fetch import Knit1to2Fetcher
1362
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1363
self.source, self.source._format, self.target,
1364
self.target._format)
1365
f = Knit1to2Fetcher(to_repository=self.target,
1366
from_repository=self.source,
1367
last_revision=revision_id,
1369
return f.count_copied, f.failed_revisions
1372
InterRepository.register_optimiser(InterSameDataRepository)
1373
InterRepository.register_optimiser(InterKnitRepo)
1374
InterRepository.register_optimiser(InterModel1and2)
1375
InterRepository.register_optimiser(InterKnit1and2)
1378
class RepositoryTestProviderAdapter(object):
1379
"""A tool to generate a suite testing multiple repository formats at once.
1381
This is done by copying the test once for each transport and injecting
1382
the transport_server, transport_readonly_server, and bzrdir_format and
1383
repository_format classes into each copy. Each copy is also given a new id()
1384
to make it easy to identify.
1387
def __init__(self, transport_server, transport_readonly_server, formats):
1388
self._transport_server = transport_server
1389
self._transport_readonly_server = transport_readonly_server
1390
self._formats = formats
1392
def adapt(self, test):
1393
result = unittest.TestSuite()
1394
for repository_format, bzrdir_format in self._formats:
1395
new_test = deepcopy(test)
1396
new_test.transport_server = self._transport_server
1397
new_test.transport_readonly_server = self._transport_readonly_server
1398
new_test.bzrdir_format = bzrdir_format
1399
new_test.repository_format = repository_format
1400
def make_new_test_id():
1401
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1402
return lambda: new_id
1403
new_test.id = make_new_test_id()
1404
result.addTest(new_test)
1408
class InterRepositoryTestProviderAdapter(object):
1409
"""A tool to generate a suite testing multiple inter repository formats.
1411
This is done by copying the test once for each interrepo provider and injecting
1412
the transport_server, transport_readonly_server, repository_format and
1413
repository_to_format classes into each copy.
1414
Each copy is also given a new id() to make it easy to identify.
1417
def __init__(self, transport_server, transport_readonly_server, formats):
1418
self._transport_server = transport_server
1419
self._transport_readonly_server = transport_readonly_server
1420
self._formats = formats
1422
def adapt(self, test):
1423
result = unittest.TestSuite()
1424
for interrepo_class, repository_format, repository_format_to in self._formats:
1425
new_test = deepcopy(test)
1426
new_test.transport_server = self._transport_server
1427
new_test.transport_readonly_server = self._transport_readonly_server
1428
new_test.interrepo_class = interrepo_class
1429
new_test.repository_format = repository_format
1430
new_test.repository_format_to = repository_format_to
1431
def make_new_test_id():
1432
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1433
return lambda: new_id
1434
new_test.id = make_new_test_id()
1435
result.addTest(new_test)
1439
def default_test_list():
1440
"""Generate the default list of interrepo permutations to test."""
1441
from bzrlib.repofmt import knitrepo, weaverepo
1443
# test the default InterRepository between format 6 and the current
1445
# XXX: robertc 20060220 reinstate this when there are two supported
1446
# formats which do not have an optimal code path between them.
1447
#result.append((InterRepository,
1448
# RepositoryFormat6(),
1449
# RepositoryFormatKnit1()))
1450
for optimiser_class in InterRepository._optimisers:
1451
format_to_test = optimiser_class._get_matching_repo_format()
1452
if format_to_test is not None:
1453
result.append((optimiser_class,
1454
format_to_test, format_to_test))
1455
# if there are specific combinations we want to use, we can add them
1457
result.append((InterModel1and2,
1458
weaverepo.RepositoryFormat5(),
1459
knitrepo.RepositoryFormatKnit2()))
1460
result.append((InterKnit1and2,
1461
knitrepo.RepositoryFormatKnit1(),
1462
knitrepo.RepositoryFormatKnit2()))
1466
class CopyConverter(object):
1467
"""A repository conversion tool which just performs a copy of the content.
1469
This is slow but quite reliable.
1472
def __init__(self, target_format):
1473
"""Create a CopyConverter.
1475
:param target_format: The format the resulting repository should be.
1477
self.target_format = target_format
1479
def convert(self, repo, pb):
1480
"""Perform the conversion of to_convert, giving feedback via pb.
1482
:param to_convert: The disk object to convert.
1483
:param pb: a progress bar to use for progress information.
1488
# this is only useful with metadir layouts - separated repo content.
1489
# trigger an assertion if not such
1490
repo._format.get_format_string()
1491
self.repo_dir = repo.bzrdir
1492
self.step('Moving repository to repository.backup')
1493
self.repo_dir.transport.move('repository', 'repository.backup')
1494
backup_transport = self.repo_dir.transport.clone('repository.backup')
1495
repo._format.check_conversion_target(self.target_format)
1496
self.source_repo = repo._format.open(self.repo_dir,
1498
_override_transport=backup_transport)
1499
self.step('Creating new repository')
1500
converted = self.target_format.initialize(self.repo_dir,
1501
self.source_repo.is_shared())
1502
converted.lock_write()
1504
self.step('Copying content into repository.')
1505
self.source_repo.copy_content_into(converted)
1508
self.step('Deleting old repository content.')
1509
self.repo_dir.transport.delete_tree('repository.backup')
1510
self.pb.note('repository converted')
1512
def step(self, message):
1513
"""Update the pb by a step."""
1515
self.pb.update(message, self.count, self.total)
1518
class CommitBuilder(object):
1519
"""Provides an interface to build up a commit.
1521
This allows describing a tree to be committed without needing to
1522
know the internals of the format of the repository.
1525
record_root_entry = False
1526
def __init__(self, repository, parents, config, timestamp=None,
1527
timezone=None, committer=None, revprops=None,
1529
"""Initiate a CommitBuilder.
1531
:param repository: Repository to commit to.
1532
:param parents: Revision ids of the parents of the new revision.
1533
:param config: Configuration to use.
1534
:param timestamp: Optional timestamp recorded for commit.
1535
:param timezone: Optional timezone for timestamp.
1536
:param committer: Optional committer to set for commit.
1537
:param revprops: Optional dictionary of revision properties.
1538
:param revision_id: Optional revision id.
1540
self._config = config
1542
if committer is None:
1543
self._committer = self._config.username()
1545
assert isinstance(committer, basestring), type(committer)
1546
self._committer = committer
1548
self.new_inventory = Inventory(None)
1549
self._new_revision_id = revision_id
1550
self.parents = parents
1551
self.repository = repository
1554
if revprops is not None:
1555
self._revprops.update(revprops)
1557
if timestamp is None:
1558
timestamp = time.time()
1559
# Restrict resolution to 1ms
1560
self._timestamp = round(timestamp, 3)
1562
if timezone is None:
1563
self._timezone = local_time_offset()
1565
self._timezone = int(timezone)
1567
self._generate_revision_if_needed()
1569
def commit(self, message):
1570
"""Make the actual commit.
1572
:return: The revision id of the recorded revision.
1574
rev = _mod_revision.Revision(
1575
timestamp=self._timestamp,
1576
timezone=self._timezone,
1577
committer=self._committer,
1579
inventory_sha1=self.inv_sha1,
1580
revision_id=self._new_revision_id,
1581
properties=self._revprops)
1582
rev.parent_ids = self.parents
1583
self.repository.add_revision(self._new_revision_id, rev,
1584
self.new_inventory, self._config)
1585
return self._new_revision_id
1587
def revision_tree(self):
1588
"""Return the tree that was just committed.
1590
After calling commit() this can be called to get a RevisionTree
1591
representing the newly committed tree. This is preferred to
1592
calling Repository.revision_tree() because that may require
1593
deserializing the inventory, while we already have a copy in
1596
return RevisionTree(self.repository, self.new_inventory,
1597
self._new_revision_id)
1599
def finish_inventory(self):
1600
"""Tell the builder that the inventory is finished."""
1601
if self.new_inventory.root is None:
1602
symbol_versioning.warn('Root entry should be supplied to'
1603
' record_entry_contents, as of bzr 0.10.',
1604
DeprecationWarning, stacklevel=2)
1605
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1606
self.new_inventory.revision_id = self._new_revision_id
1607
self.inv_sha1 = self.repository.add_inventory(
1608
self._new_revision_id,
1613
def _gen_revision_id(self):
1614
"""Return new revision-id."""
1615
return generate_ids.gen_revision_id(self._config.username(),
1618
def _generate_revision_if_needed(self):
1619
"""Create a revision id if None was supplied.
1621
If the repository can not support user-specified revision ids
1622
they should override this function and raise CannotSetRevisionId
1623
if _new_revision_id is not None.
1625
:raises: CannotSetRevisionId
1627
if self._new_revision_id is None:
1628
self._new_revision_id = self._gen_revision_id()
1630
def record_entry_contents(self, ie, parent_invs, path, tree):
1631
"""Record the content of ie from tree into the commit if needed.
1633
Side effect: sets ie.revision when unchanged
1635
:param ie: An inventory entry present in the commit.
1636
:param parent_invs: The inventories of the parent revisions of the
1638
:param path: The path the entry is at in the tree.
1639
:param tree: The tree which contains this entry and should be used to
1642
if self.new_inventory.root is None and ie.parent_id is not None:
1643
symbol_versioning.warn('Root entry should be supplied to'
1644
' record_entry_contents, as of bzr 0.10.',
1645
DeprecationWarning, stacklevel=2)
1646
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
1648
self.new_inventory.add(ie)
1650
# ie.revision is always None if the InventoryEntry is considered
1651
# for committing. ie.snapshot will record the correct revision
1652
# which may be the sole parent if it is untouched.
1653
if ie.revision is not None:
1656
# In this revision format, root entries have no knit or weave
1657
if ie is self.new_inventory.root:
1658
# When serializing out to disk and back in
1659
# root.revision is always _new_revision_id
1660
ie.revision = self._new_revision_id
1662
previous_entries = ie.find_previous_heads(
1664
self.repository.weave_store,
1665
self.repository.get_transaction())
1666
# we are creating a new revision for ie in the history store
1668
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1670
def modified_directory(self, file_id, file_parents):
1671
"""Record the presence of a symbolic link.
1673
:param file_id: The file_id of the link to record.
1674
:param file_parents: The per-file parent revision ids.
1676
self._add_text_to_weave(file_id, [], file_parents.keys())
1678
def modified_file_text(self, file_id, file_parents,
1679
get_content_byte_lines, text_sha1=None,
1681
"""Record the text of file file_id
1683
:param file_id: The file_id of the file to record the text of.
1684
:param file_parents: The per-file parent revision ids.
1685
:param get_content_byte_lines: A callable which will return the byte
1687
:param text_sha1: Optional SHA1 of the file contents.
1688
:param text_size: Optional size of the file contents.
1690
# mutter('storing text of file {%s} in revision {%s} into %r',
1691
# file_id, self._new_revision_id, self.repository.weave_store)
1692
# special case to avoid diffing on renames or
1694
if (len(file_parents) == 1
1695
and text_sha1 == file_parents.values()[0].text_sha1
1696
and text_size == file_parents.values()[0].text_size):
1697
previous_ie = file_parents.values()[0]
1698
versionedfile = self.repository.weave_store.get_weave(file_id,
1699
self.repository.get_transaction())
1700
versionedfile.clone_text(self._new_revision_id,
1701
previous_ie.revision, file_parents.keys())
1702
return text_sha1, text_size
1704
new_lines = get_content_byte_lines()
1705
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
1706
# should return the SHA1 and size
1707
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
1708
return osutils.sha_strings(new_lines), \
1709
sum(map(len, new_lines))
1711
def modified_link(self, file_id, file_parents, link_target):
1712
"""Record the presence of a symbolic link.
1714
:param file_id: The file_id of the link to record.
1715
:param file_parents: The per-file parent revision ids.
1716
:param link_target: Target location of this link.
1718
self._add_text_to_weave(file_id, [], file_parents.keys())
1720
def _add_text_to_weave(self, file_id, new_lines, parents):
1721
versionedfile = self.repository.weave_store.get_weave_or_empty(
1722
file_id, self.repository.get_transaction())
1723
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
1724
versionedfile.clear_cache()
1727
class _CommitBuilder(CommitBuilder):
1728
"""Temporary class so old CommitBuilders are detected properly
1730
Note: CommitBuilder works whether or not root entry is recorded.
1733
record_root_entry = True
1736
class RootCommitBuilder(CommitBuilder):
1737
"""This commitbuilder actually records the root id"""
1739
record_root_entry = True
1741
def record_entry_contents(self, ie, parent_invs, path, tree):
1742
"""Record the content of ie from tree into the commit if needed.
1744
Side effect: sets ie.revision when unchanged
1746
:param ie: An inventory entry present in the commit.
1747
:param parent_invs: The inventories of the parent revisions of the
1749
:param path: The path the entry is at in the tree.
1750
:param tree: The tree which contains this entry and should be used to
1753
assert self.new_inventory.root is not None or ie.parent_id is None
1754
self.new_inventory.add(ie)
1756
# ie.revision is always None if the InventoryEntry is considered
1757
# for committing. ie.snapshot will record the correct revision
1758
# which may be the sole parent if it is untouched.
1759
if ie.revision is not None:
1762
previous_entries = ie.find_previous_heads(
1764
self.repository.weave_store,
1765
self.repository.get_transaction())
1766
# we are creating a new revision for ie in the history store
1768
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1780
def _unescaper(match, _map=_unescape_map):
1781
return _map[match.group(1)]
1787
def _unescape_xml(data):
1788
"""Unescape predefined XML entities in a string of data."""
1790
if _unescape_re is None:
1791
_unescape_re = re.compile('\&([^;]*);')
1792
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
