13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
17
19
from bzrlib.lazy_import import lazy_import
18
20
lazy_import(globals(), """
22
25
from bzrlib import (
32
38
revision as _mod_revision,
33
testament as _mod_testament,
37
from bzrlib.bundle import serializer
46
from bzrlib.decorators import needs_read_lock, needs_write_lock, only_raises
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
47
51
from bzrlib.inter import InterObject
48
from bzrlib.lock import _RelockDebugMixin, LogicalLockResult
49
from bzrlib.trace import (
50
log_exception_quietly, note, mutter, mutter_callsite, warning)
52
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
53
from bzrlib.symbol_versioning import (
57
from bzrlib.trace import mutter, note, warning
53
60
# Old formats display a warning, but only once
54
61
_deprecation_warning_done = False
57
class IsInWriteGroupError(errors.InternalBzrError):
59
_fmt = "May not refresh_data of repo %(repo)s while in a write group."
61
def __init__(self, repo):
62
errors.InternalBzrError.__init__(self, repo=repo)
65
class CommitBuilder(object):
66
"""Provides an interface to build up a commit.
68
This allows describing a tree to be committed without needing to
69
know the internals of the format of the repository.
72
# all clients should supply tree roots.
73
record_root_entry = True
74
# whether this commit builder supports the record_entry_contents interface
75
supports_record_entry_contents = False
77
def __init__(self, repository, parents, config, timestamp=None,
78
timezone=None, committer=None, revprops=None,
79
revision_id=None, lossy=False):
80
"""Initiate a CommitBuilder.
82
:param repository: Repository to commit to.
83
:param parents: Revision ids of the parents of the new revision.
84
:param timestamp: Optional timestamp recorded for commit.
85
:param timezone: Optional timezone for timestamp.
86
:param committer: Optional committer to set for commit.
87
:param revprops: Optional dictionary of revision properties.
88
:param revision_id: Optional revision id.
89
:param lossy: Whether to discard data that can not be natively
90
represented, when pushing to a foreign VCS
96
self._committer = self._config.username()
97
elif not isinstance(committer, unicode):
98
self._committer = committer.decode() # throw if non-ascii
100
self._committer = committer
102
self._new_revision_id = revision_id
103
self.parents = parents
104
self.repository = repository
107
if revprops is not None:
108
self._validate_revprops(revprops)
109
self._revprops.update(revprops)
111
if timestamp is None:
112
timestamp = time.time()
113
# Restrict resolution to 1ms
114
self._timestamp = round(timestamp, 3)
117
self._timezone = osutils.local_time_offset()
119
self._timezone = int(timezone)
121
self._generate_revision_if_needed()
123
def any_changes(self):
124
"""Return True if any entries were changed.
126
This includes merge-only changes. It is the core for the --unchanged
129
:return: True if any changes have occured.
131
raise NotImplementedError(self.any_changes)
133
def _validate_unicode_text(self, text, context):
134
"""Verify things like commit messages don't have bogus characters."""
136
raise ValueError('Invalid value for %s: %r' % (context, text))
138
def _validate_revprops(self, revprops):
139
for key, value in revprops.iteritems():
140
# We know that the XML serializers do not round trip '\r'
141
# correctly, so refuse to accept them
142
if not isinstance(value, basestring):
143
raise ValueError('revision property (%s) is not a valid'
144
' (unicode) string: %r' % (key, value))
145
self._validate_unicode_text(value,
146
'revision property (%s)' % (key,))
148
def commit(self, message):
149
"""Make the actual commit.
151
:return: The revision id of the recorded revision.
153
raise NotImplementedError(self.commit)
156
"""Abort the commit that is being built.
158
raise NotImplementedError(self.abort)
160
def revision_tree(self):
161
"""Return the tree that was just committed.
163
After calling commit() this can be called to get a
164
RevisionTree representing the newly committed tree. This is
165
preferred to calling Repository.revision_tree() because that may
166
require deserializing the inventory, while we already have a copy in
169
raise NotImplementedError(self.revision_tree)
171
def finish_inventory(self):
172
"""Tell the builder that the inventory is finished.
174
:return: The inventory id in the repository, which can be used with
175
repository.get_inventory.
177
raise NotImplementedError(self.finish_inventory)
179
def _gen_revision_id(self):
180
"""Return new revision-id."""
181
return generate_ids.gen_revision_id(self._committer, self._timestamp)
183
def _generate_revision_if_needed(self):
184
"""Create a revision id if None was supplied.
186
If the repository can not support user-specified revision ids
187
they should override this function and raise CannotSetRevisionId
188
if _new_revision_id is not None.
190
:raises: CannotSetRevisionId
192
if self._new_revision_id is None:
193
self._new_revision_id = self._gen_revision_id()
194
self.random_revid = True
196
self.random_revid = False
198
def will_record_deletes(self):
199
"""Tell the commit builder that deletes are being notified.
201
This enables the accumulation of an inventory delta; for the resulting
202
commit to be valid, deletes against the basis MUST be recorded via
203
builder.record_delete().
205
raise NotImplementedError(self.will_record_deletes)
207
def record_iter_changes(self, tree, basis_revision_id, iter_changes):
208
"""Record a new tree via iter_changes.
210
:param tree: The tree to obtain text contents from for changed objects.
211
:param basis_revision_id: The revision id of the tree the iter_changes
212
has been generated against. Currently assumed to be the same
213
as self.parents[0] - if it is not, errors may occur.
214
:param iter_changes: An iter_changes iterator with the changes to apply
215
to basis_revision_id. The iterator must not include any items with
216
a current kind of None - missing items must be either filtered out
217
or errored-on beefore record_iter_changes sees the item.
218
:return: A generator of (file_id, relpath, fs_hash) tuples for use with
221
raise NotImplementedError(self.record_iter_changes)
224
class RepositoryWriteLockResult(LogicalLockResult):
225
"""The result of write locking a repository.
227
:ivar repository_token: The token obtained from the underlying lock, or
229
:ivar unlock: A callable which will unlock the lock.
232
def __init__(self, unlock, repository_token):
233
LogicalLockResult.__init__(self, unlock)
234
self.repository_token = repository_token
237
return "RepositoryWriteLockResult(%s, %s)" % (self.repository_token,
241
64
######################################################################
245
class Repository(_RelockDebugMixin, controldir.ControlComponent):
67
class Repository(object):
246
68
"""Repository holding history for one or more branches.
248
70
The repository holds and retrieves historical information including
249
71
revisions and file history. It's normally accessed only by the Branch,
250
72
which views a particular line of development through that history.
252
See VersionedFileRepository in bzrlib.vf_repository for the
253
base class for most Bazaar repositories.
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
256
def abort_write_group(self, suppress_errors=False):
257
"""Commit the contents accrued within the current write group.
259
:param suppress_errors: if true, abort_write_group will catch and log
260
unexpected errors that happen during the abort, rather than
261
allowing them to propagate. Defaults to False.
263
:seealso: start_write_group.
265
if self._write_group is not self.get_transaction():
266
# has an unlock or relock occured ?
269
'(suppressed) mismatched lock context and write group. %r, %r',
270
self._write_group, self.get_transaction())
272
raise errors.BzrError(
273
'mismatched lock context and write group. %r, %r' %
274
(self._write_group, self.get_transaction()))
276
self._abort_write_group()
277
except Exception, exc:
278
self._write_group = None
279
if not suppress_errors:
281
mutter('abort_write_group failed')
282
log_exception_quietly()
283
note('bzr: ERROR (ignored): %s', exc)
284
self._write_group = None
286
def _abort_write_group(self):
287
"""Template method for per-repository write group cleanup.
289
This is called during abort before the write group is considered to be
290
finished and should cleanup any internal state accrued during the write
291
group. There is no requirement that data handed to the repository be
292
*not* made available - this is not a rollback - but neither should any
293
attempt be made to ensure that data added is fully commited. Abort is
294
invoked when an error has occured so futher disk or network operations
295
may not be possible or may error and if possible should not be
299
def add_fallback_repository(self, repository):
300
"""Add a repository to use for looking up data not held locally.
302
:param repository: A repository.
304
raise NotImplementedError(self.add_fallback_repository)
306
def _check_fallback_repository(self, repository):
307
"""Check that this repository can fallback to repository safely.
309
Raise an error if not.
311
:param repository: A repository to fallback to.
313
return InterRepository._assert_same_model(self, repository)
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()
315
151
def all_revision_ids(self):
316
"""Returns a list of all the revision ids in the repository.
152
"""Returns a list of all the revision ids in the repository.
318
This is conceptually deprecated because code should generally work on
319
the graph reachable from a particular revision, and ignore any other
320
revisions that might be present. There is no direct replacement
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.
323
if 'evil' in debug.debug_flags:
324
mutter_callsite(2, "all_revision_ids is linear with history.")
325
158
return self._all_revision_ids()
327
161
def _all_revision_ids(self):
328
"""Returns a list of all the revision ids in the repository.
162
"""Returns a list of all the revision ids in the repository.
330
These are in as much topological order as the underlying store can
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.
333
raise NotImplementedError(self._all_revision_ids)
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)
335
176
def break_lock(self):
336
177
"""Break a lock if one is present from another instance.
341
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):
344
197
def create(a_bzrdir):
345
198
"""Construct the current default format repository in a_bzrdir."""
346
199
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
348
def __init__(self, _format, a_bzrdir, control_files):
201
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
349
202
"""instantiate a Repository.
351
204
:param _format: The format of the repository on disk.
352
205
:param a_bzrdir: The BzrDir of the repository.
353
:param control_files: Control files to use for locking, etc.
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.
355
# In the future we will have a single api for all stores for
356
# getting file texts, inventories and revisions, then
357
# this construct will accept instances of those things.
358
211
super(Repository, self).__init__()
359
212
self._format = _format
360
213
# the following are part of the public API for Repository:
361
214
self.bzrdir = a_bzrdir
362
215
self.control_files = control_files
363
self._transport = control_files._transport
364
self.base = self._transport.base
366
self._write_group = None
367
# Additional places to query for data.
368
self._fallback_repositories = []
371
def user_transport(self):
372
return self.bzrdir.user_transport
375
def control_transport(self):
376
return self._transport
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()
378
228
def __repr__(self):
379
if self._fallback_repositories:
380
return '%s(%r, fallback_repositories=%r)' % (
381
self.__class__.__name__,
383
self._fallback_repositories)
385
return '%s(%r)' % (self.__class__.__name__,
388
def _has_same_fallbacks(self, other_repo):
389
"""Returns true if the repositories have the same fallbacks."""
390
my_fb = self._fallback_repositories
391
other_fb = other_repo._fallback_repositories
392
if len(my_fb) != len(other_fb):
394
for f, g in zip(my_fb, other_fb):
395
if not f.has_same_location(g):
399
def has_same_location(self, other):
400
"""Returns a boolean indicating if this repository is at the same
401
location as another repository.
403
This might return False even when two repository objects are accessing
404
the same physical repository via different URLs.
406
if self.__class__ is not other.__class__:
408
return (self._transport.base == other._transport.base)
410
def is_in_write_group(self):
411
"""Return True if there is an open write group.
413
:seealso: start_write_group.
415
return self._write_group is not None
229
return '%s(%r)' % (self.__class__.__name__,
230
self.bzrdir.transport.base)
417
232
def is_locked(self):
418
233
return self.control_files.is_locked()
420
def is_write_locked(self):
421
"""Return True if this object is write locked."""
422
return self.is_locked() and self.control_files._lock_mode == 'w'
424
235
def lock_write(self, token=None):
425
236
"""Lock this repository for writing.
427
This causes caching within the repository obejct to start accumlating
428
data during reads, and allows a 'write_group' to be obtained. Write
429
groups must be used for actual data insertion.
431
A token should be passed in if you know that you have locked the object
432
some other way, and need to synchronise this object's state with that
435
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
437
238
:param token: if this is already locked, then lock_write will fail
438
239
unless the token matches the existing lock.
439
240
:returns: a token if this instance supports tokens, otherwise None.
609
347
def copy_content_into(self, destination, revision_id=None):
610
348
"""Make a complete copy of the content in self into destination.
612
This is a destructive operation! Do not use it on existing
350
This is a destructive operation! Do not use it on existing
353
revision_id = osutils.safe_revision_id(revision_id)
615
354
return InterRepository.get(self, destination).copy_content(revision_id)
617
def commit_write_group(self):
618
"""Commit the contents accrued within the current write group.
620
:seealso: start_write_group.
622
:return: it may return an opaque hint that can be passed to 'pack'.
624
if self._write_group is not self.get_transaction():
625
# has an unlock or relock occured ?
626
raise errors.BzrError('mismatched lock context %r and '
628
(self.get_transaction(), self._write_group))
629
result = self._commit_write_group()
630
self._write_group = None
633
def _commit_write_group(self):
634
"""Template method for per-repository write group cleanup.
636
This is called before the write group is considered to be
637
finished and should ensure that all data handed to the repository
638
for writing during the write group is safely committed (to the
639
extent possible considering file system caching etc).
642
def suspend_write_group(self):
643
raise errors.UnsuspendableWriteGroup(self)
645
def refresh_data(self):
646
"""Re-read any data needed to synchronise with disk.
648
This method is intended to be called after another repository instance
649
(such as one used by a smart server) has inserted data into the
650
repository. On all repositories this will work outside of write groups.
651
Some repository formats (pack and newer for bzrlib native formats)
652
support refresh_data inside write groups. If called inside a write
653
group on a repository that does not support refreshing in a write group
654
IsInWriteGroupError will be raised.
658
def resume_write_group(self, tokens):
659
if not self.is_write_locked():
660
raise errors.NotWriteLocked(self)
661
if self._write_group:
662
raise errors.BzrError('already in a write group')
663
self._resume_write_group(tokens)
664
# so we can detect unlock/relock - the write group is now entered.
665
self._write_group = self.get_transaction()
667
def _resume_write_group(self, tokens):
668
raise errors.UnsuspendableWriteGroup(self)
670
def fetch(self, source, revision_id=None, find_ghosts=False,
356
def fetch(self, source, revision_id=None, pb=None):
672
357
"""Fetch the content required to construct revision_id from source.
674
If revision_id is None and fetch_spec is None, then all content is
677
fetch() may not be used when the repository is in a write group -
678
either finish the current write group before using fetch, or use
679
fetch before starting the write group.
681
:param find_ghosts: Find and copy revisions in the source that are
682
ghosts in the target (and not reachable directly by walking out to
683
the first-present revision in target from revision_id).
684
:param revision_id: If specified, all the content needed for this
685
revision ID will be copied to the target. Fetch will determine for
686
itself which content needs to be copied.
687
:param fetch_spec: If specified, a SearchResult or
688
PendingAncestryResult that describes which revisions to copy. This
689
allows copying multiple heads at once. Mutually exclusive with
359
If revision_id is None all content is copied.
692
if fetch_spec is not None and revision_id is not None:
693
raise AssertionError(
694
"fetch_spec and revision_id are mutually exclusive.")
695
if self.is_in_write_group():
696
raise errors.InternalBzrError(
697
"May not fetch while in a write group.")
698
# fast path same-url fetch operations
699
# TODO: lift out to somewhere common with RemoteRepository
700
# <https://bugs.launchpad.net/bzr/+bug/401646>
701
if (self.has_same_location(source)
702
and fetch_spec is None
703
and self._has_same_fallbacks(source)):
704
# check that last_revision is in 'from' and then return a
706
if (revision_id is not None and
707
not _mod_revision.is_null(revision_id)):
708
self.get_revision(revision_id)
361
revision_id = osutils.safe_revision_id(revision_id)
710
362
inter = InterRepository.get(source, self)
711
return inter.fetch(revision_id=revision_id,
712
find_ghosts=find_ghosts, fetch_spec=fetch_spec)
714
def create_bundle(self, target, base, fileobj, format=None):
715
return serializer.write_bundle(self, target, base, fileobj, format)
717
def get_commit_builder(self, branch, parents, config, timestamp=None,
718
timezone=None, committer=None, revprops=None,
719
revision_id=None, lossy=False):
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,
720
371
"""Obtain a CommitBuilder for this repository.
722
373
:param branch: Branch to commit to.
723
374
:param parents: Revision ids of the parents of the new revision.
724
375
:param config: Configuration to use.
817
428
def has_revision(self, revision_id):
818
429
"""True if this repository has a copy of the revision."""
819
return revision_id in self.has_revisions((revision_id,))
822
def has_revisions(self, revision_ids):
823
"""Probe to find out the presence of multiple revisions.
825
:param revision_ids: An iterable of revision_ids.
826
:return: A set of the revision_ids that were present.
828
raise NotImplementedError(self.has_revisions)
831
def get_revision(self, revision_id):
832
"""Return the Revision object for a named revision."""
833
return self.get_revisions([revision_id])[0]
430
revision_id = osutils.safe_revision_id(revision_id)
431
return self._revision_store.has_revision_id(revision_id,
432
self.get_transaction())
835
435
def get_revision_reconcile(self, revision_id):
836
436
"""'reconcile' helper routine that allows access to a revision always.
838
438
This variant of get_revision does not cross check the weave graph
839
439
against the revision one as get_revision does: but it should only
840
440
be used by reconcile, or reconcile-alike commands that are correcting
841
441
or testing the revision graph.
843
raise NotImplementedError(self.get_revision_reconcile)
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]
845
449
def get_revisions(self, revision_ids):
846
"""Get many revisions at once.
848
Repositories that need to check data on every revision read should
849
subclass this method.
851
raise NotImplementedError(self.get_revisions)
853
def get_deltas_for_revisions(self, revisions, specific_fileids=None):
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):
854
491
"""Produce a generator of revision deltas.
856
493
Note that the input is a sequence of REVISIONS, not revision_ids.
857
494
Trees will be held in memory until the generator exits.
858
495
Each delta is relative to the revision's lefthand predecessor.
860
:param specific_fileids: if not None, the result is filtered
861
so that only those file-ids, their parents and their
862
children are included.
864
# Get the revision-ids of interest
865
497
required_trees = set()
866
498
for revision in revisions:
867
499
required_trees.add(revision.revision_id)
868
500
required_trees.update(revision.parent_ids[:1])
870
# Get the matching filtered trees. Note that it's more
871
# efficient to pass filtered trees to changes_from() rather
872
# than doing the filtering afterwards. changes_from() could
873
# arguably do the filtering itself but it's path-based, not
874
# file-id based, so filtering before or afterwards is
876
if specific_fileids is None:
877
trees = dict((t.get_revision_id(), t) for
878
t in self.revision_trees(required_trees))
880
trees = dict((t.get_revision_id(), t) for
881
t in self._filtered_revision_trees(required_trees,
884
# Calculate the deltas
501
trees = dict((t.get_revision_id(), t) for
502
t in self.revision_trees(required_trees))
885
503
for revision in revisions:
886
504
if not revision.parent_ids:
887
old_tree = self.revision_tree(_mod_revision.NULL_REVISION)
505
old_tree = self.revision_tree(None)
889
507
old_tree = trees[revision.parent_ids[0]]
890
508
yield trees[revision.revision_id].changes_from(old_tree)
893
def get_revision_delta(self, revision_id, specific_fileids=None):
511
def get_revision_delta(self, revision_id):
894
512
"""Return the delta for one revision.
896
514
The delta is relative to the left-hand predecessor of the
899
:param specific_fileids: if not None, the result is filtered
900
so that only those file-ids, their parents and their
901
children are included.
903
517
r = self.get_revision(revision_id)
904
return list(self.get_deltas_for_revisions([r],
905
specific_fileids=specific_fileids))[0]
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)
907
536
@needs_write_lock
908
537
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
538
revision_id = osutils.safe_revision_id(revision_id)
909
539
signature = gpg_strategy.sign(plaintext)
910
self.add_signature_text(revision_id, signature)
912
def add_signature_text(self, revision_id, signature):
913
"""Store a signature text for a revision.
915
:param revision_id: Revision id of the revision
916
:param signature: Signature text.
918
raise NotImplementedError(self.add_signature_text)
920
def _find_parent_ids_of_revisions(self, revision_ids):
921
"""Find all parent ids that are mentioned in the revision graph.
923
:return: set of revisions that are parents of revision_ids which are
924
not part of revision_ids themselves
926
parent_map = self.get_parent_map(revision_ids)
928
map(parent_ids.update, parent_map.itervalues())
929
parent_ids.difference_update(revision_ids)
930
parent_ids.discard(_mod_revision.NULL_REVISION)
540
self._revision_store.add_revision_signature_text(revision_id,
542
self.get_transaction())
933
544
def fileids_altered_by_revision_ids(self, revision_ids):
934
545
"""Find the file ids and versions affected by revisions.
936
547
:param revisions: an iterable containing revision ids.
937
548
:return: a dictionary mapping altered file-ids to an iterable of
938
revision_ids. Each altered file-ids has the exact revision_ids
939
that altered it listed explicitly.
941
raise NotImplementedError(self.fileids_altered_by_revision_ids)
943
def iter_files_bytes(self, desired_files):
944
"""Iterate through file versions.
946
Files will not necessarily be returned in the order they occur in
947
desired_files. No specific order is guaranteed.
949
Yields pairs of identifier, bytes_iterator. identifier is an opaque
950
value supplied by the caller as part of desired_files. It should
951
uniquely identify the file version in the caller's context. (Examples:
952
an index number or a TreeTransform trans_id.)
954
:param desired_files: a list of (file_id, revision_id, identifier)
957
raise NotImplementedError(self.iter_files_bytes)
959
def get_rev_id_for_revno(self, revno, known_pair):
960
"""Return the revision id of a revno, given a later (revno, revid)
961
pair in the same history.
963
:return: if found (True, revid). If the available history ran out
964
before reaching the revno, then this returns
965
(False, (closest_revno, closest_revid)).
967
known_revno, known_revid = known_pair
968
partial_history = [known_revid]
969
distance_from_known = known_revno - revno
970
if distance_from_known < 0:
972
'requested revno (%d) is later than given known revno (%d)'
973
% (revno, known_revno))
976
self, partial_history, stop_index=distance_from_known)
977
except errors.RevisionNotPresent, err:
978
if err.revision_id == known_revid:
979
# The start revision (known_revid) wasn't found.
981
# This is a stacked repository with no fallbacks, or a there's a
982
# left-hand ghost. Either way, even though the revision named in
983
# the error isn't in this repo, we know it's the next step in this
985
partial_history.append(err.revision_id)
986
if len(partial_history) <= distance_from_known:
987
# Didn't find enough history to get a revid for the revno.
988
earliest_revno = known_revno - len(partial_history) + 1
989
return (False, (earliest_revno, partial_history[-1]))
990
if len(partial_history) - 1 > distance_from_known:
991
raise AssertionError('_iter_for_revno returned too much history')
992
return (True, partial_history[-1])
994
@symbol_versioning.deprecated_method(symbol_versioning.deprecated_in((2, 4, 0)))
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()
995
744
def iter_reverse_revision_history(self, revision_id):
996
745
"""Iterate backwards through revision ids in the lefthand history
998
747
:param revision_id: The revision id to start with. All its lefthand
999
748
ancestors will be traversed.
1001
graph = self.get_graph()
1002
stop_revisions = (None, _mod_revision.NULL_REVISION)
1003
return graph.iter_lefthand_ancestry(revision_id, stop_revisions)
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)
1005
780
def is_shared(self):
1006
781
"""Return True if this repository is flagged as a shared repository."""
1007
782
raise NotImplementedError(self.is_shared)
1014
789
reconciler.reconcile()
1015
790
return reconciler
1017
def _refresh_data(self):
1018
"""Helper called from lock_* to ensure coherency with disk.
1020
The default implementation does nothing; it is however possible
1021
for repositories to maintain loaded indices across multiple locks
1022
by checking inside their implementation of this method to see
1023
whether their indices are still valid. This depends of course on
1024
the disk format being validatable in this manner. This method is
1025
also called by the refresh_data() public interface to cause a refresh
1026
to occur while in a write lock so that data inserted by a smart server
1027
push operation is visible on the client's instance of the physical
1031
792
@needs_read_lock
1032
793
def revision_tree(self, revision_id):
1033
794
"""Return Tree for a revision on this branch.
1035
`revision_id` may be NULL_REVISION for the empty tree revision.
796
`revision_id` may be None for the empty tree revision.
1037
raise NotImplementedError(self.revision_tree)
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)
1039
809
def revision_trees(self, revision_ids):
1040
"""Return Trees for revisions in this repository.
810
"""Return Tree for a revision on this branch.
1042
:param revision_ids: a sequence of revision-ids;
1043
a revision-id may not be None or 'null:'
1045
raise NotImplementedError(self.revision_trees)
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)
1047
820
@needs_read_lock
1048
@symbol_versioning.deprecated_method(
1049
symbol_versioning.deprecated_in((2, 4, 0)))
1050
def get_ancestry(self, revision_id, topo_sorted=True):
821
def get_ancestry(self, revision_id):
1051
822
"""Return a list of revision-ids integrated by a revision.
1053
The first element of the list is always None, indicating the origin
1054
revision. This might change when we have history horizons, or
824
The first element of the list is always None, indicating the origin
825
revision. This might change when we have history horizons, or
1055
826
perhaps we should have a new API.
1057
828
This is topologically sorted.
1059
if 'evil' in debug.debug_flags:
1060
mutter_callsite(2, "get_ancestry is linear with history.")
1061
if _mod_revision.is_null(revision_id):
830
if revision_id is None:
832
revision_id = osutils.safe_revision_id(revision_id)
1063
833
if not self.has_revision(revision_id):
1064
834
raise errors.NoSuchRevision(self, revision_id)
1065
graph = self.get_graph()
1067
search = graph._make_breadth_first_searcher([revision_id])
1070
found, ghosts = search.next_with_ghosts()
1071
except StopIteration:
1074
if _mod_revision.NULL_REVISION in keys:
1075
keys.remove(_mod_revision.NULL_REVISION)
1077
parent_map = graph.get_parent_map(keys)
1078
keys = tsort.topo_sort(parent_map)
1079
return [None] + list(keys)
1081
def pack(self, hint=None, clean_obsolete_packs=False):
1082
"""Compress the data within the repository.
1084
This operation only makes sense for some repository types. For other
1085
types it should be a no-op that just returns.
1087
This stub method does not require a lock, but subclasses should use
1088
@needs_write_lock as this is a long running call it's reasonable to
1089
implicitly lock for the user.
1091
:param hint: If not supplied, the whole repository is packed.
1092
If supplied, the repository may use the hint parameter as a
1093
hint for the parts of the repository to pack. A hint can be
1094
obtained from the result of commit_write_group(). Out of
1095
date hints are simply ignored, because concurrent operations
1096
can obsolete them rapidly.
1098
:param clean_obsolete_packs: Clean obsolete packs immediately after
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)
1102
858
def get_transaction(self):
1103
859
return self.control_files.get_transaction()
1105
def get_parent_map(self, revision_ids):
1106
"""See graph.StackedParentsProvider.get_parent_map"""
1107
raise NotImplementedError(self.get_parent_map)
1109
def _get_parent_map_no_fallbacks(self, revision_ids):
1110
"""Same as Repository.get_parent_map except doesn't query fallbacks."""
1111
# revisions index works in keys; this just works in revisions
1112
# therefore wrap and unwrap
1115
for revision_id in revision_ids:
1116
if revision_id == _mod_revision.NULL_REVISION:
1117
result[revision_id] = ()
1118
elif revision_id is None:
1119
raise ValueError('get_parent_map(None) is not valid')
1121
query_keys.append((revision_id ,))
1122
vf = self.revisions.without_fallbacks()
1123
for ((revision_id,), parent_keys) in \
1124
vf.get_parent_map(query_keys).iteritems():
1126
result[revision_id] = tuple([parent_revid
1127
for (parent_revid,) in parent_keys])
1129
result[revision_id] = (_mod_revision.NULL_REVISION,)
1132
def _make_parents_provider(self):
1133
if not self._format.supports_external_lookups:
1135
return graph.StackedParentsProvider(_LazyListJoin(
1136
[self._make_parents_provider_unstacked()],
1137
self._fallback_repositories))
1139
def _make_parents_provider_unstacked(self):
1140
return graph.CallableToParentsProviderAdapter(
1141
self._get_parent_map_no_fallbacks)
1144
def get_known_graph_ancestry(self, revision_ids):
1145
"""Return the known graph for a set of revision ids and their ancestors.
1147
raise NotImplementedError(self.get_known_graph_ancestry)
1149
def get_file_graph(self):
1150
"""Return the graph walker for files."""
1151
raise NotImplementedError(self.get_file_graph)
1153
def get_graph(self, other_repository=None):
1154
"""Return the graph walker for this repository format"""
1155
parents_provider = self._make_parents_provider()
1156
if (other_repository is not None and
1157
not self.has_same_location(other_repository)):
1158
parents_provider = graph.StackedParentsProvider(
1159
[parents_provider, other_repository._make_parents_provider()])
1160
return graph.Graph(parents_provider)
1162
def revision_ids_to_search_result(self, result_set):
1163
"""Convert a set of revision ids to a graph SearchResult."""
1164
result_parents = set()
1165
for parents in self.get_graph().get_parent_map(
1166
result_set).itervalues():
1167
result_parents.update(parents)
1168
included_keys = result_set.intersection(result_parents)
1169
start_keys = result_set.difference(included_keys)
1170
exclude_keys = result_parents.difference(result_set)
1171
result = graph.SearchResult(start_keys, exclude_keys,
1172
len(result_set), result_set)
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)
1175
865
@needs_write_lock
1176
866
def set_make_working_trees(self, new_value):
1695
1334
_optimisers = []
1696
1335
"""The available optimised InterRepository types."""
1699
1337
def copy_content(self, revision_id=None):
1700
"""Make a complete copy of the content in self into destination.
1702
This is a destructive operation! Do not use it on existing
1705
:param revision_id: Only copy the content needed to construct
1706
revision_id and its parents.
1709
self.target.set_make_working_trees(self.source.make_working_trees())
1710
except NotImplementedError:
1712
self.target.fetch(self.source, revision_id=revision_id)
1715
def fetch(self, revision_id=None, find_ghosts=False,
1338
raise NotImplementedError(self.copy_content)
1340
def fetch(self, revision_id=None, pb=None):
1717
1341
"""Fetch the content required to construct revision_id.
1719
1343
The content is copied from self.source to self.target.
1721
1345
:param revision_id: if None all content is copied, if NULL_REVISION no
1722
1346
content is copied.
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:
1725
1353
raise NotImplementedError(self.fetch)
1727
1355
@needs_read_lock
1728
def search_missing_revision_ids(self,
1729
revision_id=symbol_versioning.DEPRECATED_PARAMETER,
1730
find_ghosts=True, revision_ids=None, if_present_ids=None,
1356
def missing_revision_ids(self, revision_id=None):
1732
1357
"""Return the revision ids that source has that target does not.
1359
These are returned in topological order.
1734
1361
:param revision_id: only return revision ids included by this
1736
:param revision_ids: return revision ids included by these
1737
revision_ids. NoSuchRevision will be raised if any of these
1738
revisions are not present.
1739
:param if_present_ids: like revision_ids, but will not cause
1740
NoSuchRevision if any of these are absent, instead they will simply
1741
not be in the result. This is useful for e.g. finding revisions
1742
to fetch for tags, which may reference absent revisions.
1743
:param find_ghosts: If True find missing revisions in deep history
1744
rather than just finding the surface difference.
1745
:param limit: Maximum number of revisions to return, topologically
1747
:return: A bzrlib.graph.SearchResult.
1749
raise NotImplementedError(self.search_missing_revision_ids)
1752
def _same_model(source, target):
1753
"""True if source and target have the same data representation.
1755
Note: this is always called on the base class; overriding it in a
1756
subclass will have no effect.
1759
InterRepository._assert_same_model(source, target)
1761
except errors.IncompatibleRepositories, e:
1765
def _assert_same_model(source, target):
1766
"""Raise an exception if two repositories do not use the same model.
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):
1768
1395
if source.supports_rich_root() != target.supports_rich_root():
1769
raise errors.IncompatibleRepositories(source, target,
1770
"different rich-root support")
1771
1397
if source._serializer != target._serializer:
1772
raise errors.IncompatibleRepositories(source, target,
1773
"different serializers")
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()))
1776
1855
class CopyConverter(object):
1777
1856
"""A repository conversion tool which just performs a copy of the content.
1779
1858
This is slow but quite reliable.
1785
1864
:param target_format: The format the resulting repository should be.
1787
1866
self.target_format = target_format
1789
1868
def convert(self, repo, pb):
1790
1869
"""Perform the conversion of to_convert, giving feedback via pb.
1792
1871
:param to_convert: The disk object to convert.
1793
1872
:param pb: a progress bar to use for progress information.
1795
pb = ui.ui_factory.nested_progress_bar()
1798
1877
# this is only useful with metadir layouts - separated repo content.
1799
1878
# trigger an assertion if not such
1800
1879
repo._format.get_format_string()
1801
1880
self.repo_dir = repo.bzrdir
1802
pb.update('Moving repository to repository.backup')
1881
self.step('Moving repository to repository.backup')
1803
1882
self.repo_dir.transport.move('repository', 'repository.backup')
1804
1883
backup_transport = self.repo_dir.transport.clone('repository.backup')
1805
1884
repo._format.check_conversion_target(self.target_format)
1806
1885
self.source_repo = repo._format.open(self.repo_dir,
1808
1887
_override_transport=backup_transport)
1809
pb.update('Creating new repository')
1888
self.step('Creating new repository')
1810
1889
converted = self.target_format.initialize(self.repo_dir,
1811
1890
self.source_repo.is_shared())
1812
1891
converted.lock_write()
1814
pb.update('Copying content')
1893
self.step('Copying content into repository.')
1815
1894
self.source_repo.copy_content_into(converted)
1817
1896
converted.unlock()
1818
pb.update('Deleting old repository content')
1897
self.step('Deleting old repository content.')
1819
1898
self.repo_dir.transport.delete_tree('repository.backup')
1820
ui.ui_factory.note('repository converted')
1824
def _strip_NULL_ghosts(revision_graph):
1825
"""Also don't use this. more compatibility code for unmigrated clients."""
1826
# Filter ghosts, and null:
1827
if _mod_revision.NULL_REVISION in revision_graph:
1828
del revision_graph[_mod_revision.NULL_REVISION]
1829
for key, parents in revision_graph.items():
1830
revision_graph[key] = tuple(parent for parent in parents if parent
1832
return revision_graph
1835
def _iter_for_revno(repo, partial_history_cache, stop_index=None,
1836
stop_revision=None):
1837
"""Extend the partial history to include a given index
1839
If a stop_index is supplied, stop when that index has been reached.
1840
If a stop_revision is supplied, stop when that revision is
1841
encountered. Otherwise, stop when the beginning of history is
1844
:param stop_index: The index which should be present. When it is
1845
present, history extension will stop.
1846
:param stop_revision: The revision id which should be present. When
1847
it is encountered, history extension will stop.
1849
start_revision = partial_history_cache[-1]
1850
graph = repo.get_graph()
1851
iterator = graph.iter_lefthand_ancestry(start_revision,
1852
(_mod_revision.NULL_REVISION,))
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)
1854
# skip the last revision in the list
1857
if (stop_index is not None and
1858
len(partial_history_cache) > stop_index):
1860
if partial_history_cache[-1] == stop_revision:
1862
revision_id = iterator.next()
1863
partial_history_cache.append(revision_id)
1864
except StopIteration:
1869
class _LazyListJoin(object):
1870
"""An iterable yielding the contents of many lists as one list.
1872
Each iterator made from this will reflect the current contents of the lists
1873
at the time the iterator is made.
1875
This is used by Repository's _make_parents_provider implementation so that
1878
pp = repo._make_parents_provider() # uses a list of fallback repos
1879
pp.add_fallback_repository(other_repo) # appends to that list
1880
result = pp.get_parent_map(...)
1881
# The result will include revs from other_repo
1884
def __init__(self, *list_parts):
1885
self.list_parts = list_parts
1889
for list_part in self.list_parts:
1890
full_list.extend(list_part)
1891
return iter(full_list)
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
