1
# Copyright (C) 2005 Canonical Ltd
1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
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
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
17
from cStringIO import StringIO
19
from bzrlib.lockable_files import LockableFiles
20
from bzrlib.tree import EmptyTree
21
from bzrlib.revision import NULL_REVISION
22
from bzrlib.store import copy_all
23
from bzrlib.store.weave import WeaveStore
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
38
revision as _mod_revision,
43
from bzrlib.bundle import serializer
44
from bzrlib.revisiontree import RevisionTree
45
from bzrlib.store.versioned import VersionedFileStore
24
46
from bzrlib.store.text import TextStore
26
from bzrlib.tree import RevisionTree
27
from bzrlib.errors import InvalidRevisionId
28
47
from bzrlib.testament import Testament
31
def needs_read_lock(unbound):
32
"""Decorate unbound to take out and release a read lock."""
33
def decorated(self, *args, **kwargs):
34
self.control_files.lock_read()
36
return unbound(self, *args, **kwargs)
38
self.control_files.unlock()
42
def needs_write_lock(unbound):
43
"""Decorate unbound to take out and release a write lock."""
44
def decorated(self, *args, **kwargs):
45
self.control_files.lock_write()
47
return unbound(self, *args, **kwargs)
49
self.control_files.unlock()
51
from bzrlib.decorators import needs_read_lock, needs_write_lock
52
from bzrlib.inter import InterObject
53
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
54
from bzrlib.symbol_versioning import (
58
from bzrlib.trace import mutter, note, warning
61
# Old formats display a warning, but only once
62
_deprecation_warning_done = False
65
######################################################################
53
68
class Repository(object):
55
def __init__(self, transport, branch_format):
57
self.control_files = LockableFiles(transport.clone(bzrlib.BZRDIR), 'README')
59
dir_mode = self.control_files._dir_mode
60
file_mode = self.control_files._file_mode
62
def get_weave(name, prefixed=False):
64
name = bzrlib.BZRDIR + '/' + unicode(name)
67
relpath = self.control_files._escape(name)
68
weave_transport = transport.clone(relpath)
69
ws = WeaveStore(weave_transport, prefixed=prefixed,
72
if self.control_files._transport.should_cache():
73
ws.enable_cache = True
76
def get_store(name, compressed=True, prefixed=False):
77
# FIXME: This approach of assuming stores are all entirely compressed
78
# or entirely uncompressed is tidy, but breaks upgrade from
79
# some existing branches where there's a mixture; we probably
80
# still want the option to look for both.
82
name = bzrlib.BZRDIR + '/' + unicode(name)
85
relpath = self.control_files._escape(name)
86
store = TextStore(transport.clone(relpath),
87
prefixed=prefixed, compressed=compressed,
90
#if self._transport.should_cache():
91
# cache_path = os.path.join(self.cache_root, name)
92
# os.mkdir(cache_path)
93
# store = bzrlib.store.CachedStore(store, cache_path)
96
if branch_format == 4:
97
self.inventory_store = get_store('inventory-store')
98
self.text_store = get_store('text-store')
99
self.revision_store = get_store('revision-store')
100
elif branch_format == 5:
101
self.control_weaves = get_weave('')
102
self.weave_store = get_weave('weaves')
103
self.revision_store = get_store('revision-store', compressed=False)
104
elif branch_format == 6:
105
self.control_weaves = get_weave('')
106
self.weave_store = get_weave('weaves', prefixed=True)
107
self.revision_store = get_store('revision-store', compressed=False,
109
self.revision_store.register_suffix('sig')
111
def lock_write(self):
112
self.control_files.lock_write()
69
"""Repository holding history for one or more branches.
71
The repository holds and retrieves historical information including
72
revisions and file history. It's normally accessed only by the Branch,
73
which views a particular line of development through that history.
75
The Repository builds on top of Stores and a Transport, which respectively
76
describe the disk data format and the way of accessing the (possibly
80
_file_ids_altered_regex = lazy_regex.lazy_compile(
81
r'file_id="(?P<file_id>[^"]+)"'
82
r'.*revision="(?P<revision_id>[^"]+)"'
86
def add_inventory(self, revision_id, inv, parents):
87
"""Add the inventory inv to the repository as revision_id.
89
:param parents: The revision ids of the parents that revision_id
90
is known to have and are in the repository already.
92
returns the sha1 of the serialized inventory.
94
revision_id = osutils.safe_revision_id(revision_id)
95
_mod_revision.check_not_reserved_id(revision_id)
96
assert inv.revision_id is None or inv.revision_id == revision_id, \
97
"Mismatch between inventory revision" \
98
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
99
assert inv.root is not None
100
inv_text = self.serialise_inventory(inv)
101
inv_sha1 = osutils.sha_string(inv_text)
102
inv_vf = self.control_weaves.get_weave('inventory',
103
self.get_transaction())
104
self._inventory_add_lines(inv_vf, revision_id, parents,
105
osutils.split_lines(inv_text))
108
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines):
110
for parent in parents:
112
final_parents.append(parent)
114
inv_vf.add_lines(revision_id, final_parents, lines)
117
def add_revision(self, revision_id, rev, inv=None, config=None):
118
"""Add rev to the revision store as revision_id.
120
:param revision_id: the revision id to use.
121
:param rev: The revision object.
122
:param inv: The inventory for the revision. if None, it will be looked
123
up in the inventory storer
124
:param config: If None no digital signature will be created.
125
If supplied its signature_needed method will be used
126
to determine if a signature should be made.
128
revision_id = osutils.safe_revision_id(revision_id)
129
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
131
_mod_revision.check_not_reserved_id(revision_id)
132
if config is not None and config.signature_needed():
134
inv = self.get_inventory(revision_id)
135
plaintext = Testament(rev, inv).as_short_text()
136
self.store_revision_signature(
137
gpg.GPGStrategy(config), plaintext, revision_id)
138
if not revision_id in self.get_inventory_weave():
140
raise errors.WeaveRevisionNotPresent(revision_id,
141
self.get_inventory_weave())
143
# yes, this is not suitable for adding with ghosts.
144
self.add_inventory(revision_id, inv, rev.parent_ids)
145
self._revision_store.add_revision(rev, self.get_transaction())
147
def _add_revision_text(self, revision_id, text):
148
revision = self._revision_store._serializer.read_revision_from_string(
150
self._revision_store._add_revision(revision, StringIO(text),
151
self.get_transaction())
154
def _all_possible_ids(self):
155
"""Return all the possible revisions that we could find."""
156
return self.get_inventory_weave().versions()
158
def all_revision_ids(self):
159
"""Returns a list of all the revision ids in the repository.
161
This is deprecated because code should generally work on the graph
162
reachable from a particular revision, and ignore any other revisions
163
that might be present. There is no direct replacement method.
165
return self._all_revision_ids()
168
def _all_revision_ids(self):
169
"""Returns a list of all the revision ids in the repository.
171
These are in as much topological order as the underlying store can
172
present: for weaves ghosts may lead to a lack of correctness until
173
the reweave updates the parents list.
175
if self._revision_store.text_store.listable():
176
return self._revision_store.all_revision_ids(self.get_transaction())
177
result = self._all_possible_ids()
178
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
179
# ids. (It should, since _revision_store's API should change to
180
# return utf8 revision_ids)
181
return self._eliminate_revisions_not_present(result)
183
def break_lock(self):
184
"""Break a lock if one is present from another instance.
186
Uses the ui factory to ask for confirmation if the lock may be from
189
self.control_files.break_lock()
192
def _eliminate_revisions_not_present(self, revision_ids):
193
"""Check every revision id in revision_ids to see if we have it.
195
Returns a set of the present revisions.
198
for id in revision_ids:
199
if self.has_revision(id):
204
def create(a_bzrdir):
205
"""Construct the current default format repository in a_bzrdir."""
206
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
208
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
209
"""instantiate a Repository.
211
:param _format: The format of the repository on disk.
212
:param a_bzrdir: The BzrDir of the repository.
214
In the future we will have a single api for all stores for
215
getting file texts, inventories and revisions, then
216
this construct will accept instances of those things.
218
super(Repository, self).__init__()
219
self._format = _format
220
# the following are part of the public API for Repository:
221
self.bzrdir = a_bzrdir
222
self.control_files = control_files
223
self._revision_store = _revision_store
224
# backwards compatibility
225
self.weave_store = text_store
226
# not right yet - should be more semantically clear ?
228
self.control_store = control_store
229
self.control_weaves = control_store
230
# TODO: make sure to construct the right store classes, etc, depending
231
# on whether escaping is required.
232
self._warn_if_deprecated()
235
return '%s(%r)' % (self.__class__.__name__,
236
self.bzrdir.transport.base)
238
def __eq__(self, other):
239
if self.__class__ is not other.__class__:
241
return (self.control_files._transport.base ==
242
other.control_files._transport.base)
244
def __ne__(self, other):
245
return not self == other
248
return self.control_files.is_locked()
250
def lock_write(self, token=None):
251
"""Lock this repository for writing.
253
:param token: if this is already locked, then lock_write will fail
254
unless the token matches the existing lock.
255
:returns: a token if this instance supports tokens, otherwise None.
256
:raises TokenLockingNotSupported: when a token is given but this
257
instance doesn't support using token locks.
258
:raises MismatchedToken: if the specified token doesn't match the token
259
of the existing lock.
261
A token should be passed in if you know that you have locked the object
262
some other way, and need to synchronise this object's state with that
265
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
267
return self.control_files.lock_write(token=token)
114
269
def lock_read(self):
115
270
self.control_files.lock_read()
272
def get_physical_lock_status(self):
273
return self.control_files.get_physical_lock_status()
275
def leave_lock_in_place(self):
276
"""Tell this repository not to release the physical lock when this
279
If lock_write doesn't return a token, then this method is not supported.
281
self.control_files.leave_in_place()
283
def dont_leave_lock_in_place(self):
284
"""Tell this repository to release the physical lock when this
285
object is unlocked, even if it didn't originally acquire it.
287
If lock_write doesn't return a token, then this method is not supported.
289
self.control_files.dont_leave_in_place()
292
def gather_stats(self, revid=None, committers=None):
293
"""Gather statistics from a revision id.
295
:param revid: The revision id to gather statistics from, if None, then
296
no revision specific statistics are gathered.
297
:param committers: Optional parameter controlling whether to grab
298
a count of committers from the revision specific statistics.
299
:return: A dictionary of statistics. Currently this contains:
300
committers: The number of committers if requested.
301
firstrev: A tuple with timestamp, timezone for the penultimate left
302
most ancestor of revid, if revid is not the NULL_REVISION.
303
latestrev: A tuple with timestamp, timezone for revid, if revid is
304
not the NULL_REVISION.
305
revisions: The total revision count in the repository.
306
size: An estimate disk size of the repository in bytes.
309
if revid and committers:
310
result['committers'] = 0
311
if revid and revid != _mod_revision.NULL_REVISION:
313
all_committers = set()
314
revisions = self.get_ancestry(revid)
315
# pop the leading None
317
first_revision = None
319
# ignore the revisions in the middle - just grab first and last
320
revisions = revisions[0], revisions[-1]
321
for revision in self.get_revisions(revisions):
322
if not first_revision:
323
first_revision = revision
325
all_committers.add(revision.committer)
326
last_revision = revision
328
result['committers'] = len(all_committers)
329
result['firstrev'] = (first_revision.timestamp,
330
first_revision.timezone)
331
result['latestrev'] = (last_revision.timestamp,
332
last_revision.timezone)
334
# now gather global repository information
335
if self.bzrdir.root_transport.listable():
336
c, t = self._revision_store.total_size(self.get_transaction())
337
result['revisions'] = c
342
def missing_revision_ids(self, other, revision_id=None):
343
"""Return the revision ids that other has that this does not.
345
These are returned in topological order.
347
revision_id: only return revision ids included by revision_id.
349
revision_id = osutils.safe_revision_id(revision_id)
350
return InterRepository.get(other, self).missing_revision_ids(revision_id)
354
"""Open the repository rooted at base.
356
For instance, if the repository is at URL/.bzr/repository,
357
Repository.open(URL) -> a Repository instance.
359
control = bzrdir.BzrDir.open(base)
360
return control.open_repository()
362
def copy_content_into(self, destination, revision_id=None):
363
"""Make a complete copy of the content in self into destination.
365
This is a destructive operation! Do not use it on existing
368
revision_id = osutils.safe_revision_id(revision_id)
369
return InterRepository.get(self, destination).copy_content(revision_id)
371
def fetch(self, source, revision_id=None, pb=None):
372
"""Fetch the content required to construct revision_id from source.
374
If revision_id is None all content is copied.
376
revision_id = osutils.safe_revision_id(revision_id)
377
inter = InterRepository.get(source, self)
379
return inter.fetch(revision_id=revision_id, pb=pb)
380
except NotImplementedError:
381
raise errors.IncompatibleRepositories(source, self)
383
def create_bundle(self, target, base, fileobj, format=None):
384
return serializer.write_bundle(self, target, base, fileobj, format)
386
def get_commit_builder(self, branch, parents, config, timestamp=None,
387
timezone=None, committer=None, revprops=None,
389
"""Obtain a CommitBuilder for this repository.
391
:param branch: Branch to commit to.
392
:param parents: Revision ids of the parents of the new revision.
393
:param config: Configuration to use.
394
:param timestamp: Optional timestamp recorded for commit.
395
:param timezone: Optional timezone for timestamp.
396
:param committer: Optional committer to set for commit.
397
:param revprops: Optional dictionary of revision properties.
398
:param revision_id: Optional revision id.
400
revision_id = osutils.safe_revision_id(revision_id)
401
return _CommitBuilder(self, parents, config, timestamp, timezone,
402
committer, revprops, revision_id)
117
404
def unlock(self):
118
405
self.control_files.unlock()
120
def copy(self, destination):
121
destination.control_weaves.copy_multi(self.control_weaves,
123
copy_all(self.weave_store, destination.weave_store)
124
copy_all(self.revision_store, destination.revision_store)
408
def clone(self, a_bzrdir, revision_id=None):
409
"""Clone this repository into a_bzrdir using the current format.
411
Currently no check is made that the format of this repository and
412
the bzrdir format are compatible. FIXME RBC 20060201.
414
:return: The newly created destination repository.
416
# TODO: deprecate after 0.16; cloning this with all its settings is
417
# probably not very useful -- mbp 20070423
418
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
419
self.copy_content_into(dest_repo, revision_id)
423
def sprout(self, to_bzrdir, revision_id=None):
424
"""Create a descendent repository for new development.
426
Unlike clone, this does not copy the settings of the repository.
428
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
429
dest_repo.fetch(self, revision_id=revision_id)
432
def _create_sprouting_repo(self, a_bzrdir, shared):
433
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
434
# use target default format.
435
dest_repo = a_bzrdir.create_repository()
437
# Most control formats need the repository to be specifically
438
# created, but on some old all-in-one formats it's not needed
440
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
441
except errors.UninitializableFormat:
442
dest_repo = a_bzrdir.open_repository()
126
446
def has_revision(self, revision_id):
127
"""True if this branch has a copy of the revision.
129
This does not necessarily imply the revision is merge
130
or on the mainline."""
131
return (revision_id is None
132
or self.revision_store.has_id(revision_id))
447
"""True if this repository has a copy of the revision."""
448
revision_id = osutils.safe_revision_id(revision_id)
449
return self._revision_store.has_revision_id(revision_id,
450
self.get_transaction())
135
def get_revision_xml_file(self, revision_id):
136
"""Return XML file object for revision object."""
453
def get_revision_reconcile(self, revision_id):
454
"""'reconcile' helper routine that allows access to a revision always.
456
This variant of get_revision does not cross check the weave graph
457
against the revision one as get_revision does: but it should only
458
be used by reconcile, or reconcile-alike commands that are correcting
459
or testing the revision graph.
137
461
if not revision_id or not isinstance(revision_id, basestring):
138
raise InvalidRevisionId(revision_id=revision_id, branch=self)
140
return self.revision_store.get(revision_id)
141
except (IndexError, KeyError):
142
raise bzrlib.errors.NoSuchRevision(self, revision_id)
462
raise errors.InvalidRevisionId(revision_id=revision_id,
464
return self.get_revisions([revision_id])[0]
467
def get_revisions(self, revision_ids):
468
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
469
revs = self._revision_store.get_revisions(revision_ids,
470
self.get_transaction())
472
assert not isinstance(rev.revision_id, unicode)
473
for parent_id in rev.parent_ids:
474
assert not isinstance(parent_id, unicode)
144
478
def get_revision_xml(self, revision_id):
145
return self.get_revision_xml_file(revision_id).read()
479
# TODO: jam 20070210 This shouldn't be necessary since get_revision
480
# would have already do it.
481
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
482
revision_id = osutils.safe_revision_id(revision_id)
483
rev = self.get_revision(revision_id)
485
# the current serializer..
486
self._revision_store._serializer.write_revision(rev, rev_tmp)
488
return rev_tmp.getvalue()
147
491
def get_revision(self, revision_id):
148
492
"""Return the Revision object for a named revision"""
149
xml_file = self.get_revision_xml_file(revision_id)
152
r = bzrlib.xml5.serializer_v5.read_revision(xml_file)
153
except SyntaxError, e:
154
raise bzrlib.errors.BzrError('failed to unpack revision_xml',
158
assert r.revision_id == revision_id
493
# TODO: jam 20070210 get_revision_reconcile should do this for us
494
revision_id = osutils.safe_revision_id(revision_id)
495
r = self.get_revision_reconcile(revision_id)
496
# weave corruption can lead to absent revision markers that should be
498
# the following test is reasonably cheap (it needs a single weave read)
499
# and the weave is cached in read transactions. In write transactions
500
# it is not cached but typically we only read a small number of
501
# revisions. For knits when they are introduced we will probably want
502
# to ensure that caching write transactions are in use.
503
inv = self.get_inventory_weave()
504
self._check_revision_parents(r, inv)
161
def get_revision_sha1(self, revision_id):
162
"""Hash the stored value of a revision, and return it."""
163
# In the future, revision entries will be signed. At that
164
# point, it is probably best *not* to include the signature
165
# in the revision hash. Because that lets you re-sign
166
# the revision, (add signatures/remove signatures) and still
167
# have all hash pointers stay consistent.
168
# But for now, just hash the contents.
169
return bzrlib.osutils.sha_file(self.get_revision_xml_file(revision_id))
508
def get_deltas_for_revisions(self, revisions):
509
"""Produce a generator of revision deltas.
511
Note that the input is a sequence of REVISIONS, not revision_ids.
512
Trees will be held in memory until the generator exits.
513
Each delta is relative to the revision's lefthand predecessor.
515
required_trees = set()
516
for revision in revisions:
517
required_trees.add(revision.revision_id)
518
required_trees.update(revision.parent_ids[:1])
519
trees = dict((t.get_revision_id(), t) for
520
t in self.revision_trees(required_trees))
521
for revision in revisions:
522
if not revision.parent_ids:
523
old_tree = self.revision_tree(None)
525
old_tree = trees[revision.parent_ids[0]]
526
yield trees[revision.revision_id].changes_from(old_tree)
529
def get_revision_delta(self, revision_id):
530
"""Return the delta for one revision.
532
The delta is relative to the left-hand predecessor of the
535
r = self.get_revision(revision_id)
536
return list(self.get_deltas_for_revisions([r]))[0]
538
def _check_revision_parents(self, revision, inventory):
539
"""Private to Repository and Fetch.
541
This checks the parentage of revision in an inventory weave for
542
consistency and is only applicable to inventory-weave-for-ancestry
543
using repository formats & fetchers.
545
weave_parents = inventory.get_parents(revision.revision_id)
546
weave_names = inventory.versions()
547
for parent_id in revision.parent_ids:
548
if parent_id in weave_names:
549
# this parent must not be a ghost.
550
if not parent_id in weave_parents:
552
raise errors.CorruptRepository(self)
171
554
@needs_write_lock
172
555
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
173
self.revision_store.add(StringIO(gpg_strategy.sign(plaintext)),
556
revision_id = osutils.safe_revision_id(revision_id)
557
signature = gpg_strategy.sign(plaintext)
558
self._revision_store.add_revision_signature_text(revision_id,
560
self.get_transaction())
562
def fileids_altered_by_revision_ids(self, revision_ids):
563
"""Find the file ids and versions affected by revisions.
565
:param revisions: an iterable containing revision ids.
566
:return: a dictionary mapping altered file-ids to an iterable of
567
revision_ids. Each altered file-ids has the exact revision_ids that
568
altered it listed explicitly.
570
assert self._serializer.support_altered_by_hack, \
571
("fileids_altered_by_revision_ids only supported for branches "
572
"which store inventory as unnested xml, not on %r" % self)
573
selected_revision_ids = set(osutils.safe_revision_id(r)
574
for r in revision_ids)
575
w = self.get_inventory_weave()
578
# this code needs to read every new line in every inventory for the
579
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
580
# not present in one of those inventories is unnecessary but not
581
# harmful because we are filtering by the revision id marker in the
582
# inventory lines : we only select file ids altered in one of those
583
# revisions. We don't need to see all lines in the inventory because
584
# only those added in an inventory in rev X can contain a revision=X
586
unescape_revid_cache = {}
587
unescape_fileid_cache = {}
589
# jam 20061218 In a big fetch, this handles hundreds of thousands
590
# of lines, so it has had a lot of inlining and optimizing done.
591
# Sorry that it is a little bit messy.
592
# Move several functions to be local variables, since this is a long
594
search = self._file_ids_altered_regex.search
595
unescape = _unescape_xml
596
setdefault = result.setdefault
597
pb = ui.ui_factory.nested_progress_bar()
599
for line in w.iter_lines_added_or_present_in_versions(
600
selected_revision_ids, pb=pb):
604
# One call to match.group() returning multiple items is quite a
605
# bit faster than 2 calls to match.group() each returning 1
606
file_id, revision_id = match.group('file_id', 'revision_id')
608
# Inlining the cache lookups helps a lot when you make 170,000
609
# lines and 350k ids, versus 8.4 unique ids.
610
# Using a cache helps in 2 ways:
611
# 1) Avoids unnecessary decoding calls
612
# 2) Re-uses cached strings, which helps in future set and
614
# (2) is enough that removing encoding entirely along with
615
# the cache (so we are using plain strings) results in no
616
# performance improvement.
618
revision_id = unescape_revid_cache[revision_id]
620
unescaped = unescape(revision_id)
621
unescape_revid_cache[revision_id] = unescaped
622
revision_id = unescaped
624
if revision_id in selected_revision_ids:
626
file_id = unescape_fileid_cache[file_id]
628
unescaped = unescape(file_id)
629
unescape_fileid_cache[file_id] = unescaped
631
setdefault(file_id, set()).add(revision_id)
176
637
def get_inventory_weave(self):
177
638
return self.control_weaves.get_weave('inventory',
178
639
self.get_transaction())
180
642
def get_inventory(self, revision_id):
181
643
"""Get Inventory object by hash."""
182
xml = self.get_inventory_xml(revision_id)
183
return bzrlib.xml5.serializer_v5.read_inventory_from_string(xml)
644
# TODO: jam 20070210 Technically we don't need to sanitize, since all
645
# called functions must sanitize.
646
revision_id = osutils.safe_revision_id(revision_id)
647
return self.deserialise_inventory(
648
revision_id, self.get_inventory_xml(revision_id))
650
def deserialise_inventory(self, revision_id, xml):
651
"""Transform the xml into an inventory object.
653
:param revision_id: The expected revision id of the inventory.
654
:param xml: A serialised inventory.
656
revision_id = osutils.safe_revision_id(revision_id)
657
result = self._serializer.read_inventory_from_string(xml)
658
result.root.revision = revision_id
661
def serialise_inventory(self, inv):
662
return self._serializer.write_inventory_to_string(inv)
664
def get_serializer_format(self):
665
return self._serializer.format_num
185
668
def get_inventory_xml(self, revision_id):
186
669
"""Get inventory XML as a file object."""
670
revision_id = osutils.safe_revision_id(revision_id)
188
assert isinstance(revision_id, basestring), type(revision_id)
672
assert isinstance(revision_id, str), type(revision_id)
189
673
iw = self.get_inventory_weave()
190
return iw.get_text(iw.lookup(revision_id))
674
return iw.get_text(revision_id)
191
675
except IndexError:
192
raise bzrlib.errors.HistoryMissing(self, 'inventory', revision_id)
676
raise errors.HistoryMissing(self, 'inventory', revision_id)
194
679
def get_inventory_sha1(self, revision_id):
195
680
"""Return the sha1 hash of the inventory entry
682
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
683
revision_id = osutils.safe_revision_id(revision_id)
197
684
return self.get_revision(revision_id).inventory_sha1
687
def get_revision_graph(self, revision_id=None):
688
"""Return a dictionary containing the revision graph.
690
:param revision_id: The revision_id to get a graph from. If None, then
691
the entire revision graph is returned. This is a deprecated mode of
692
operation and will be removed in the future.
693
:return: a dictionary of revision_id->revision_parents_list.
695
# special case NULL_REVISION
696
if revision_id == _mod_revision.NULL_REVISION:
698
revision_id = osutils.safe_revision_id(revision_id)
699
a_weave = self.get_inventory_weave()
700
all_revisions = self._eliminate_revisions_not_present(
702
entire_graph = dict([(node, tuple(a_weave.get_parents(node))) for
703
node in all_revisions])
704
if revision_id is None:
706
elif revision_id not in entire_graph:
707
raise errors.NoSuchRevision(self, revision_id)
709
# add what can be reached from revision_id
711
pending = set([revision_id])
712
while len(pending) > 0:
714
result[node] = entire_graph[node]
715
for revision_id in result[node]:
716
if revision_id not in result:
717
pending.add(revision_id)
721
def get_revision_graph_with_ghosts(self, revision_ids=None):
722
"""Return a graph of the revisions with ghosts marked as applicable.
724
:param revision_ids: an iterable of revisions to graph or None for all.
725
:return: a Graph object with the graph reachable from revision_ids.
727
result = deprecated_graph.Graph()
729
pending = set(self.all_revision_ids())
732
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
733
# special case NULL_REVISION
734
if _mod_revision.NULL_REVISION in pending:
735
pending.remove(_mod_revision.NULL_REVISION)
736
required = set(pending)
739
revision_id = pending.pop()
741
rev = self.get_revision(revision_id)
742
except errors.NoSuchRevision:
743
if revision_id in required:
746
result.add_ghost(revision_id)
748
for parent_id in rev.parent_ids:
749
# is this queued or done ?
750
if (parent_id not in pending and
751
parent_id not in done):
753
pending.add(parent_id)
754
result.add_node(revision_id, rev.parent_ids)
755
done.add(revision_id)
758
def _get_history_vf(self):
759
"""Get a versionedfile whose history graph reflects all revisions.
761
For weave repositories, this is the inventory weave.
763
return self.get_inventory_weave()
765
def iter_reverse_revision_history(self, revision_id):
766
"""Iterate backwards through revision ids in the lefthand history
768
:param revision_id: The revision id to start with. All its lefthand
769
ancestors will be traversed.
771
revision_id = osutils.safe_revision_id(revision_id)
772
if revision_id in (None, _mod_revision.NULL_REVISION):
774
next_id = revision_id
775
versionedfile = self._get_history_vf()
778
parents = versionedfile.get_parents(next_id)
779
if len(parents) == 0:
199
785
def get_revision_inventory(self, revision_id):
200
786
"""Return inventory of a past revision."""
201
787
# TODO: Unify this with get_inventory()
202
788
# bzr 0.0.6 and later imposes the constraint that the inventory_id
203
789
# must be the same as its revision, so this is trivial.
204
if revision_id == None:
790
if revision_id is None:
205
791
# This does not make sense: if there is no revision,
206
792
# then it is the current tree inventory surely ?!
207
793
# and thus get_root_id() is something that looks at the last
212
798
return self.get_inventory(revision_id)
802
"""Return True if this repository is flagged as a shared repository."""
803
raise NotImplementedError(self.is_shared)
806
def reconcile(self, other=None, thorough=False):
807
"""Reconcile this repository."""
808
from bzrlib.reconcile import RepoReconciler
809
reconciler = RepoReconciler(self, thorough=thorough)
810
reconciler.reconcile()
214
814
def revision_tree(self, revision_id):
215
815
"""Return Tree for a revision on this branch.
217
`revision_id` may be None for the null revision, in which case
218
an `EmptyTree` is returned."""
817
`revision_id` may be None for the empty tree revision.
219
819
# TODO: refactor this to use an existing revision object
220
820
# so we don't need to read it in twice.
221
if revision_id == None or revision_id == NULL_REVISION:
821
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
822
return RevisionTree(self, Inventory(root_id=None),
823
_mod_revision.NULL_REVISION)
825
revision_id = osutils.safe_revision_id(revision_id)
224
826
inv = self.get_revision_inventory(revision_id)
225
827
return RevisionTree(self, inv, revision_id)
227
def get_ancestry(self, revision_id):
830
def revision_trees(self, revision_ids):
831
"""Return Tree for a revision on this branch.
833
`revision_id` may not be None or 'null:'"""
834
assert None not in revision_ids
835
assert _mod_revision.NULL_REVISION not in revision_ids
836
texts = self.get_inventory_weave().get_texts(revision_ids)
837
for text, revision_id in zip(texts, revision_ids):
838
inv = self.deserialise_inventory(revision_id, text)
839
yield RevisionTree(self, inv, revision_id)
842
def get_ancestry(self, revision_id, topo_sorted=True):
228
843
"""Return a list of revision-ids integrated by a revision.
845
The first element of the list is always None, indicating the origin
846
revision. This might change when we have history horizons, or
847
perhaps we should have a new API.
230
849
This is topologically sorted.
232
if revision_id is None:
851
if _mod_revision.is_null(revision_id):
853
revision_id = osutils.safe_revision_id(revision_id)
854
if not self.has_revision(revision_id):
855
raise errors.NoSuchRevision(self, revision_id)
234
856
w = self.get_inventory_weave()
235
return [None] + map(w.idx_to_name,
236
w.inclusions([w.lookup(revision_id)]))
857
candidates = w.get_ancestry(revision_id, topo_sorted)
858
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
861
"""Compress the data within the repository.
863
This operation only makes sense for some repository types. For other
864
types it should be a no-op that just returns.
866
This stub method does not require a lock, but subclasses should use
867
@needs_write_lock as this is a long running call its reasonable to
868
implicitly lock for the user.
239
872
def print_file(self, file, revision_id):
240
"""Print `file` to stdout."""
873
"""Print `file` to stdout.
875
FIXME RBC 20060125 as John Meinel points out this is a bad api
876
- it writes to stdout, it assumes that that is valid etc. Fix
877
by creating a new more flexible convenience function.
879
revision_id = osutils.safe_revision_id(revision_id)
241
880
tree = self.revision_tree(revision_id)
242
881
# use inventory as it was in that revision
243
882
file_id = tree.inventory.path2id(file)
245
raise BzrError("%r is not present in revision %s" % (file, revno))
247
revno = self.revision_id_to_revno(revision_id)
248
except errors.NoSuchRevision:
249
# TODO: This should not be BzrError,
250
# but NoSuchFile doesn't fit either
251
raise BzrError('%r is not present in revision %s'
252
% (file, revision_id))
254
raise BzrError('%r is not present in revision %s'
884
# TODO: jam 20060427 Write a test for this code path
885
# it had a bug in it, and was raising the wrong
887
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
256
888
tree.print_file(file_id)
258
890
def get_transaction(self):
259
891
return self.control_files.get_transaction()
893
def revision_parents(self, revision_id):
894
revision_id = osutils.safe_revision_id(revision_id)
895
return self.get_inventory_weave().parent_names(revision_id)
897
def get_parents(self, revision_ids):
898
"""See StackedParentsProvider.get_parents"""
900
for revision_id in revision_ids:
901
if revision_id == _mod_revision.NULL_REVISION:
905
parents = self.get_revision(revision_id).parent_ids
906
except errors.NoSuchRevision:
909
if len(parents) == 0:
910
parents = [_mod_revision.NULL_REVISION]
911
parents_list.append(parents)
914
def _make_parents_provider(self):
917
def get_graph(self, other_repository=None):
918
"""Return the graph walker for this repository format"""
919
parents_provider = self._make_parents_provider()
920
if (other_repository is not None and
921
other_repository.bzrdir.transport.base !=
922
self.bzrdir.transport.base):
923
parents_provider = graph._StackedParentsProvider(
924
[parents_provider, other_repository._make_parents_provider()])
925
return graph.Graph(parents_provider)
928
def set_make_working_trees(self, new_value):
929
"""Set the policy flag for making working trees when creating branches.
931
This only applies to branches that use this repository.
933
The default is 'True'.
934
:param new_value: True to restore the default, False to disable making
937
raise NotImplementedError(self.set_make_working_trees)
939
def make_working_trees(self):
940
"""Returns the policy for making working trees on new branches."""
941
raise NotImplementedError(self.make_working_trees)
261
944
def sign_revision(self, revision_id, gpg_strategy):
945
revision_id = osutils.safe_revision_id(revision_id)
262
946
plaintext = Testament.from_revision(self, revision_id).as_short_text()
263
947
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
950
def has_signature_for_revision_id(self, revision_id):
951
"""Query for a revision signature for revision_id in the repository."""
952
revision_id = osutils.safe_revision_id(revision_id)
953
return self._revision_store.has_signature(revision_id,
954
self.get_transaction())
957
def get_signature_text(self, revision_id):
958
"""Return the text for a signature."""
959
revision_id = osutils.safe_revision_id(revision_id)
960
return self._revision_store.get_signature_text(revision_id,
961
self.get_transaction())
964
def check(self, revision_ids):
965
"""Check consistency of all history of given revision_ids.
967
Different repository implementations should override _check().
969
:param revision_ids: A non-empty list of revision_ids whose ancestry
970
will be checked. Typically the last revision_id of a branch.
973
raise ValueError("revision_ids must be non-empty in %s.check"
975
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
976
return self._check(revision_ids)
978
def _check(self, revision_ids):
979
result = check.Check(self)
983
def _warn_if_deprecated(self):
984
global _deprecation_warning_done
985
if _deprecation_warning_done:
987
_deprecation_warning_done = True
988
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
989
% (self._format, self.bzrdir.transport.base))
991
def supports_rich_root(self):
992
return self._format.rich_root_data
994
def _check_ascii_revisionid(self, revision_id, method):
995
"""Private helper for ascii-only repositories."""
996
# weave repositories refuse to store revisionids that are non-ascii.
997
if revision_id is not None:
998
# weaves require ascii revision ids.
999
if isinstance(revision_id, unicode):
1001
revision_id.encode('ascii')
1002
except UnicodeEncodeError:
1003
raise errors.NonAsciiRevisionId(method, self)
1006
revision_id.decode('ascii')
1007
except UnicodeDecodeError:
1008
raise errors.NonAsciiRevisionId(method, self)
1012
# remove these delegates a while after bzr 0.15
1013
def __make_delegated(name, from_module):
1014
def _deprecated_repository_forwarder():
1015
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1016
% (name, from_module),
1019
m = __import__(from_module, globals(), locals(), [name])
1021
return getattr(m, name)
1022
except AttributeError:
1023
raise AttributeError('module %s has no name %s'
1025
globals()[name] = _deprecated_repository_forwarder
1028
'AllInOneRepository',
1029
'WeaveMetaDirRepository',
1030
'PreSplitOutRepositoryFormat',
1031
'RepositoryFormat4',
1032
'RepositoryFormat5',
1033
'RepositoryFormat6',
1034
'RepositoryFormat7',
1036
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1040
'RepositoryFormatKnit',
1041
'RepositoryFormatKnit1',
1043
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1046
def install_revision(repository, rev, revision_tree):
1047
"""Install all revision data into a repository."""
1048
present_parents = []
1050
for p_id in rev.parent_ids:
1051
if repository.has_revision(p_id):
1052
present_parents.append(p_id)
1053
parent_trees[p_id] = repository.revision_tree(p_id)
1055
parent_trees[p_id] = repository.revision_tree(None)
1057
inv = revision_tree.inventory
1058
entries = inv.iter_entries()
1059
# backwards compatability hack: skip the root id.
1060
if not repository.supports_rich_root():
1061
path, root = entries.next()
1062
if root.revision != rev.revision_id:
1063
raise errors.IncompatibleRevision(repr(repository))
1064
# Add the texts that are not already present
1065
for path, ie in entries:
1066
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1067
repository.get_transaction())
1068
if ie.revision not in w:
1070
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1071
# with InventoryEntry.find_previous_heads(). if it is, then there
1072
# is a latent bug here where the parents may have ancestors of each
1074
for revision, tree in parent_trees.iteritems():
1075
if ie.file_id not in tree:
1077
parent_id = tree.inventory[ie.file_id].revision
1078
if parent_id in text_parents:
1080
text_parents.append(parent_id)
1082
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1083
repository.get_transaction())
1084
lines = revision_tree.get_file(ie.file_id).readlines()
1085
vfile.add_lines(rev.revision_id, text_parents, lines)
1087
# install the inventory
1088
repository.add_inventory(rev.revision_id, inv, present_parents)
1089
except errors.RevisionAlreadyPresent:
1091
repository.add_revision(rev.revision_id, rev, inv)
1094
class MetaDirRepository(Repository):
1095
"""Repositories in the new meta-dir layout."""
1097
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1098
super(MetaDirRepository, self).__init__(_format,
1104
dir_mode = self.control_files._dir_mode
1105
file_mode = self.control_files._file_mode
1108
def is_shared(self):
1109
"""Return True if this repository is flagged as a shared repository."""
1110
return self.control_files._transport.has('shared-storage')
1113
def set_make_working_trees(self, new_value):
1114
"""Set the policy flag for making working trees when creating branches.
1116
This only applies to branches that use this repository.
1118
The default is 'True'.
1119
:param new_value: True to restore the default, False to disable making
1124
self.control_files._transport.delete('no-working-trees')
1125
except errors.NoSuchFile:
1128
self.control_files.put_utf8('no-working-trees', '')
1130
def make_working_trees(self):
1131
"""Returns the policy for making working trees on new branches."""
1132
return not self.control_files._transport.has('no-working-trees')
1135
class RepositoryFormatRegistry(registry.Registry):
1136
"""Registry of RepositoryFormats.
1139
def get(self, format_string):
1140
r = registry.Registry.get(self, format_string)
1146
format_registry = RepositoryFormatRegistry()
1147
"""Registry of formats, indexed by their identifying format string.
1149
This can contain either format instances themselves, or classes/factories that
1150
can be called to obtain one.
1154
#####################################################################
1155
# Repository Formats
1157
class RepositoryFormat(object):
1158
"""A repository format.
1160
Formats provide three things:
1161
* An initialization routine to construct repository data on disk.
1162
* a format string which is used when the BzrDir supports versioned
1164
* an open routine which returns a Repository instance.
1166
Formats are placed in an dict by their format string for reference
1167
during opening. These should be subclasses of RepositoryFormat
1170
Once a format is deprecated, just deprecate the initialize and open
1171
methods on the format class. Do not deprecate the object, as the
1172
object will be created every system load.
1174
Common instance attributes:
1175
_matchingbzrdir - the bzrdir format that the repository format was
1176
originally written to work with. This can be used if manually
1177
constructing a bzrdir and repository, or more commonly for test suite
1182
return "<%s>" % self.__class__.__name__
1184
def __eq__(self, other):
1185
# format objects are generally stateless
1186
return isinstance(other, self.__class__)
1188
def __ne__(self, other):
1189
return not self == other
1192
def find_format(klass, a_bzrdir):
1193
"""Return the format for the repository object in a_bzrdir.
1195
This is used by bzr native formats that have a "format" file in
1196
the repository. Other methods may be used by different types of
1200
transport = a_bzrdir.get_repository_transport(None)
1201
format_string = transport.get("format").read()
1202
return format_registry.get(format_string)
1203
except errors.NoSuchFile:
1204
raise errors.NoRepositoryPresent(a_bzrdir)
1206
raise errors.UnknownFormatError(format=format_string)
1209
def register_format(klass, format):
1210
format_registry.register(format.get_format_string(), format)
1213
def unregister_format(klass, format):
1214
format_registry.remove(format.get_format_string())
1217
def get_default_format(klass):
1218
"""Return the current default format."""
1219
from bzrlib import bzrdir
1220
return bzrdir.format_registry.make_bzrdir('default').repository_format
1222
def _get_control_store(self, repo_transport, control_files):
1223
"""Return the control store for this repository."""
1224
raise NotImplementedError(self._get_control_store)
1226
def get_format_string(self):
1227
"""Return the ASCII format string that identifies this format.
1229
Note that in pre format ?? repositories the format string is
1230
not permitted nor written to disk.
1232
raise NotImplementedError(self.get_format_string)
1234
def get_format_description(self):
1235
"""Return the short description for this format."""
1236
raise NotImplementedError(self.get_format_description)
1238
def _get_revision_store(self, repo_transport, control_files):
1239
"""Return the revision store object for this a_bzrdir."""
1240
raise NotImplementedError(self._get_revision_store)
1242
def _get_text_rev_store(self,
1249
"""Common logic for getting a revision store for a repository.
1251
see self._get_revision_store for the subclass-overridable method to
1252
get the store for a repository.
1254
from bzrlib.store.revision.text import TextRevisionStore
1255
dir_mode = control_files._dir_mode
1256
file_mode = control_files._file_mode
1257
text_store = TextStore(transport.clone(name),
1259
compressed=compressed,
1261
file_mode=file_mode)
1262
_revision_store = TextRevisionStore(text_store, serializer)
1263
return _revision_store
1265
# TODO: this shouldn't be in the base class, it's specific to things that
1266
# use weaves or knits -- mbp 20070207
1267
def _get_versioned_file_store(self,
1272
versionedfile_class=None,
1273
versionedfile_kwargs={},
1275
if versionedfile_class is None:
1276
versionedfile_class = self._versionedfile_class
1277
weave_transport = control_files._transport.clone(name)
1278
dir_mode = control_files._dir_mode
1279
file_mode = control_files._file_mode
1280
return VersionedFileStore(weave_transport, prefixed=prefixed,
1282
file_mode=file_mode,
1283
versionedfile_class=versionedfile_class,
1284
versionedfile_kwargs=versionedfile_kwargs,
1287
def initialize(self, a_bzrdir, shared=False):
1288
"""Initialize a repository of this format in a_bzrdir.
1290
:param a_bzrdir: The bzrdir to put the new repository in it.
1291
:param shared: The repository should be initialized as a sharable one.
1292
:returns: The new repository object.
1294
This may raise UninitializableFormat if shared repository are not
1295
compatible the a_bzrdir.
1297
raise NotImplementedError(self.initialize)
1299
def is_supported(self):
1300
"""Is this format supported?
1302
Supported formats must be initializable and openable.
1303
Unsupported formats may not support initialization or committing or
1304
some other features depending on the reason for not being supported.
1308
def check_conversion_target(self, target_format):
1309
raise NotImplementedError(self.check_conversion_target)
1311
def open(self, a_bzrdir, _found=False):
1312
"""Return an instance of this format for the bzrdir a_bzrdir.
1314
_found is a private parameter, do not use it.
1316
raise NotImplementedError(self.open)
1319
class MetaDirRepositoryFormat(RepositoryFormat):
1320
"""Common base class for the new repositories using the metadir layout."""
1322
rich_root_data = False
1323
supports_tree_reference = False
1324
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1327
super(MetaDirRepositoryFormat, self).__init__()
1329
def _create_control_files(self, a_bzrdir):
1330
"""Create the required files and the initial control_files object."""
1331
# FIXME: RBC 20060125 don't peek under the covers
1332
# NB: no need to escape relative paths that are url safe.
1333
repository_transport = a_bzrdir.get_repository_transport(self)
1334
control_files = lockable_files.LockableFiles(repository_transport,
1335
'lock', lockdir.LockDir)
1336
control_files.create_lock()
1337
return control_files
1339
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1340
"""Upload the initial blank content."""
1341
control_files = self._create_control_files(a_bzrdir)
1342
control_files.lock_write()
1344
control_files._transport.mkdir_multi(dirs,
1345
mode=control_files._dir_mode)
1346
for file, content in files:
1347
control_files.put(file, content)
1348
for file, content in utf8_files:
1349
control_files.put_utf8(file, content)
1351
control_files.put_utf8('shared-storage', '')
1353
control_files.unlock()
1356
# formats which have no format string are not discoverable
1357
# and not independently creatable, so are not registered. They're
1358
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1359
# needed, it's constructed directly by the BzrDir. Non-native formats where
1360
# the repository is not separately opened are similar.
1362
format_registry.register_lazy(
1363
'Bazaar-NG Repository format 7',
1364
'bzrlib.repofmt.weaverepo',
1367
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1368
# default control directory format
1370
format_registry.register_lazy(
1371
'Bazaar-NG Knit Repository Format 1',
1372
'bzrlib.repofmt.knitrepo',
1373
'RepositoryFormatKnit1',
1375
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1377
format_registry.register_lazy(
1378
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1379
'bzrlib.repofmt.knitrepo',
1380
'RepositoryFormatKnit3',
1384
class InterRepository(InterObject):
1385
"""This class represents operations taking place between two repositories.
1387
Its instances have methods like copy_content and fetch, and contain
1388
references to the source and target repositories these operations can be
1391
Often we will provide convenience methods on 'repository' which carry out
1392
operations with another repository - they will always forward to
1393
InterRepository.get(other).method_name(parameters).
1397
"""The available optimised InterRepository types."""
1399
def copy_content(self, revision_id=None):
1400
raise NotImplementedError(self.copy_content)
1402
def fetch(self, revision_id=None, pb=None):
1403
"""Fetch the content required to construct revision_id.
1405
The content is copied from self.source to self.target.
1407
:param revision_id: if None all content is copied, if NULL_REVISION no
1409
:param pb: optional progress bar to use for progress reports. If not
1410
provided a default one will be created.
1412
Returns the copied revision count and the failed revisions in a tuple:
1415
raise NotImplementedError(self.fetch)
1418
def missing_revision_ids(self, revision_id=None):
1419
"""Return the revision ids that source has that target does not.
1421
These are returned in topological order.
1423
:param revision_id: only return revision ids included by this
1426
# generic, possibly worst case, slow code path.
1427
target_ids = set(self.target.all_revision_ids())
1428
if revision_id is not None:
1429
# TODO: jam 20070210 InterRepository is internal enough that it
1430
# should assume revision_ids are already utf-8
1431
revision_id = osutils.safe_revision_id(revision_id)
1432
source_ids = self.source.get_ancestry(revision_id)
1433
assert source_ids[0] is None
1436
source_ids = self.source.all_revision_ids()
1437
result_set = set(source_ids).difference(target_ids)
1438
# this may look like a no-op: its not. It preserves the ordering
1439
# other_ids had while only returning the members from other_ids
1440
# that we've decided we need.
1441
return [rev_id for rev_id in source_ids if rev_id in result_set]
1444
class InterSameDataRepository(InterRepository):
1445
"""Code for converting between repositories that represent the same data.
1447
Data format and model must match for this to work.
1451
def _get_repo_format_to_test(self):
1452
"""Repository format for testing with."""
1453
return RepositoryFormat.get_default_format()
1456
def is_compatible(source, target):
1457
if source.supports_rich_root() != target.supports_rich_root():
1459
if source._serializer != target._serializer:
1464
def copy_content(self, revision_id=None):
1465
"""Make a complete copy of the content in self into destination.
1467
This copies both the repository's revision data, and configuration information
1468
such as the make_working_trees setting.
1470
This is a destructive operation! Do not use it on existing
1473
:param revision_id: Only copy the content needed to construct
1474
revision_id and its parents.
1477
self.target.set_make_working_trees(self.source.make_working_trees())
1478
except NotImplementedError:
1480
# TODO: jam 20070210 This is fairly internal, so we should probably
1481
# just assert that revision_id is not unicode.
1482
revision_id = osutils.safe_revision_id(revision_id)
1483
# but don't bother fetching if we have the needed data now.
1484
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1485
self.target.has_revision(revision_id)):
1487
self.target.fetch(self.source, revision_id=revision_id)
1490
def fetch(self, revision_id=None, pb=None):
1491
"""See InterRepository.fetch()."""
1492
from bzrlib.fetch import GenericRepoFetcher
1493
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1494
self.source, self.source._format, self.target,
1495
self.target._format)
1496
# TODO: jam 20070210 This should be an assert, not a translate
1497
revision_id = osutils.safe_revision_id(revision_id)
1498
f = GenericRepoFetcher(to_repository=self.target,
1499
from_repository=self.source,
1500
last_revision=revision_id,
1502
return f.count_copied, f.failed_revisions
1505
class InterWeaveRepo(InterSameDataRepository):
1506
"""Optimised code paths between Weave based repositories."""
1509
def _get_repo_format_to_test(self):
1510
from bzrlib.repofmt import weaverepo
1511
return weaverepo.RepositoryFormat7()
1514
def is_compatible(source, target):
1515
"""Be compatible with known Weave formats.
1517
We don't test for the stores being of specific types because that
1518
could lead to confusing results, and there is no need to be
1521
from bzrlib.repofmt.weaverepo import (
1527
return (isinstance(source._format, (RepositoryFormat5,
1529
RepositoryFormat7)) and
1530
isinstance(target._format, (RepositoryFormat5,
1532
RepositoryFormat7)))
1533
except AttributeError:
1537
def copy_content(self, revision_id=None):
1538
"""See InterRepository.copy_content()."""
1539
# weave specific optimised path:
1540
# TODO: jam 20070210 Internal, should be an assert, not translate
1541
revision_id = osutils.safe_revision_id(revision_id)
1543
self.target.set_make_working_trees(self.source.make_working_trees())
1544
except NotImplementedError:
1546
# FIXME do not peek!
1547
if self.source.control_files._transport.listable():
1548
pb = ui.ui_factory.nested_progress_bar()
1550
self.target.weave_store.copy_all_ids(
1551
self.source.weave_store,
1553
from_transaction=self.source.get_transaction(),
1554
to_transaction=self.target.get_transaction())
1555
pb.update('copying inventory', 0, 1)
1556
self.target.control_weaves.copy_multi(
1557
self.source.control_weaves, ['inventory'],
1558
from_transaction=self.source.get_transaction(),
1559
to_transaction=self.target.get_transaction())
1560
self.target._revision_store.text_store.copy_all_ids(
1561
self.source._revision_store.text_store,
1566
self.target.fetch(self.source, revision_id=revision_id)
1569
def fetch(self, revision_id=None, pb=None):
1570
"""See InterRepository.fetch()."""
1571
from bzrlib.fetch import GenericRepoFetcher
1572
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1573
self.source, self.source._format, self.target, self.target._format)
1574
# TODO: jam 20070210 This should be an assert, not a translate
1575
revision_id = osutils.safe_revision_id(revision_id)
1576
f = GenericRepoFetcher(to_repository=self.target,
1577
from_repository=self.source,
1578
last_revision=revision_id,
1580
return f.count_copied, f.failed_revisions
1583
def missing_revision_ids(self, revision_id=None):
1584
"""See InterRepository.missing_revision_ids()."""
1585
# we want all revisions to satisfy revision_id in source.
1586
# but we don't want to stat every file here and there.
1587
# we want then, all revisions other needs to satisfy revision_id
1588
# checked, but not those that we have locally.
1589
# so the first thing is to get a subset of the revisions to
1590
# satisfy revision_id in source, and then eliminate those that
1591
# we do already have.
1592
# this is slow on high latency connection to self, but as as this
1593
# disk format scales terribly for push anyway due to rewriting
1594
# inventory.weave, this is considered acceptable.
1596
if revision_id is not None:
1597
source_ids = self.source.get_ancestry(revision_id)
1598
assert source_ids[0] is None
1601
source_ids = self.source._all_possible_ids()
1602
source_ids_set = set(source_ids)
1603
# source_ids is the worst possible case we may need to pull.
1604
# now we want to filter source_ids against what we actually
1605
# have in target, but don't try to check for existence where we know
1606
# we do not have a revision as that would be pointless.
1607
target_ids = set(self.target._all_possible_ids())
1608
possibly_present_revisions = target_ids.intersection(source_ids_set)
1609
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1610
required_revisions = source_ids_set.difference(actually_present_revisions)
1611
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1612
if revision_id is not None:
1613
# we used get_ancestry to determine source_ids then we are assured all
1614
# revisions referenced are present as they are installed in topological order.
1615
# and the tip revision was validated by get_ancestry.
1616
return required_topo_revisions
1618
# if we just grabbed the possibly available ids, then
1619
# we only have an estimate of whats available and need to validate
1620
# that against the revision records.
1621
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1624
class InterKnitRepo(InterSameDataRepository):
1625
"""Optimised code paths between Knit based repositories."""
1628
def _get_repo_format_to_test(self):
1629
from bzrlib.repofmt import knitrepo
1630
return knitrepo.RepositoryFormatKnit1()
1633
def is_compatible(source, target):
1634
"""Be compatible with known Knit formats.
1636
We don't test for the stores being of specific types because that
1637
could lead to confusing results, and there is no need to be
1640
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1642
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1643
isinstance(target._format, (RepositoryFormatKnit1)))
1644
except AttributeError:
1648
def fetch(self, revision_id=None, pb=None):
1649
"""See InterRepository.fetch()."""
1650
from bzrlib.fetch import KnitRepoFetcher
1651
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1652
self.source, self.source._format, self.target, self.target._format)
1653
# TODO: jam 20070210 This should be an assert, not a translate
1654
revision_id = osutils.safe_revision_id(revision_id)
1655
f = KnitRepoFetcher(to_repository=self.target,
1656
from_repository=self.source,
1657
last_revision=revision_id,
1659
return f.count_copied, f.failed_revisions
1662
def missing_revision_ids(self, revision_id=None):
1663
"""See InterRepository.missing_revision_ids()."""
1664
if revision_id is not None:
1665
source_ids = self.source.get_ancestry(revision_id)
1666
assert source_ids[0] is None
1669
source_ids = self.source._all_possible_ids()
1670
source_ids_set = set(source_ids)
1671
# source_ids is the worst possible case we may need to pull.
1672
# now we want to filter source_ids against what we actually
1673
# have in target, but don't try to check for existence where we know
1674
# we do not have a revision as that would be pointless.
1675
target_ids = set(self.target._all_possible_ids())
1676
possibly_present_revisions = target_ids.intersection(source_ids_set)
1677
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1678
required_revisions = source_ids_set.difference(actually_present_revisions)
1679
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1680
if revision_id is not None:
1681
# we used get_ancestry to determine source_ids then we are assured all
1682
# revisions referenced are present as they are installed in topological order.
1683
# and the tip revision was validated by get_ancestry.
1684
return required_topo_revisions
1686
# if we just grabbed the possibly available ids, then
1687
# we only have an estimate of whats available and need to validate
1688
# that against the revision records.
1689
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1692
class InterModel1and2(InterRepository):
1695
def _get_repo_format_to_test(self):
1699
def is_compatible(source, target):
1700
if not source.supports_rich_root() and target.supports_rich_root():
1706
def fetch(self, revision_id=None, pb=None):
1707
"""See InterRepository.fetch()."""
1708
from bzrlib.fetch import Model1toKnit2Fetcher
1709
# TODO: jam 20070210 This should be an assert, not a translate
1710
revision_id = osutils.safe_revision_id(revision_id)
1711
f = Model1toKnit2Fetcher(to_repository=self.target,
1712
from_repository=self.source,
1713
last_revision=revision_id,
1715
return f.count_copied, f.failed_revisions
1718
def copy_content(self, revision_id=None):
1719
"""Make a complete copy of the content in self into destination.
1721
This is a destructive operation! Do not use it on existing
1724
:param revision_id: Only copy the content needed to construct
1725
revision_id and its parents.
1728
self.target.set_make_working_trees(self.source.make_working_trees())
1729
except NotImplementedError:
1731
# TODO: jam 20070210 Internal, assert, don't translate
1732
revision_id = osutils.safe_revision_id(revision_id)
1733
# but don't bother fetching if we have the needed data now.
1734
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1735
self.target.has_revision(revision_id)):
1737
self.target.fetch(self.source, revision_id=revision_id)
1740
class InterKnit1and2(InterKnitRepo):
1743
def _get_repo_format_to_test(self):
1747
def is_compatible(source, target):
1748
"""Be compatible with Knit1 source and Knit3 target"""
1749
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
1751
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1752
RepositoryFormatKnit3
1753
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1754
isinstance(target._format, (RepositoryFormatKnit3)))
1755
except AttributeError:
1759
def fetch(self, revision_id=None, pb=None):
1760
"""See InterRepository.fetch()."""
1761
from bzrlib.fetch import Knit1to2Fetcher
1762
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1763
self.source, self.source._format, self.target,
1764
self.target._format)
1765
# TODO: jam 20070210 This should be an assert, not a translate
1766
revision_id = osutils.safe_revision_id(revision_id)
1767
f = Knit1to2Fetcher(to_repository=self.target,
1768
from_repository=self.source,
1769
last_revision=revision_id,
1771
return f.count_copied, f.failed_revisions
1774
class InterRemoteRepository(InterRepository):
1775
"""Code for converting between RemoteRepository objects.
1777
This just gets an non-remote repository from the RemoteRepository, and calls
1778
InterRepository.get again.
1781
def __init__(self, source, target):
1782
if isinstance(source, remote.RemoteRepository):
1783
source._ensure_real()
1784
real_source = source._real_repository
1786
real_source = source
1787
if isinstance(target, remote.RemoteRepository):
1788
target._ensure_real()
1789
real_target = target._real_repository
1791
real_target = target
1792
self.real_inter = InterRepository.get(real_source, real_target)
1795
def is_compatible(source, target):
1796
if isinstance(source, remote.RemoteRepository):
1798
if isinstance(target, remote.RemoteRepository):
1802
def copy_content(self, revision_id=None):
1803
self.real_inter.copy_content(revision_id=revision_id)
1805
def fetch(self, revision_id=None, pb=None):
1806
self.real_inter.fetch(revision_id=revision_id, pb=pb)
1809
def _get_repo_format_to_test(self):
1813
InterRepository.register_optimiser(InterSameDataRepository)
1814
InterRepository.register_optimiser(InterWeaveRepo)
1815
InterRepository.register_optimiser(InterKnitRepo)
1816
InterRepository.register_optimiser(InterModel1and2)
1817
InterRepository.register_optimiser(InterKnit1and2)
1818
InterRepository.register_optimiser(InterRemoteRepository)
1821
class CopyConverter(object):
1822
"""A repository conversion tool which just performs a copy of the content.
1824
This is slow but quite reliable.
1827
def __init__(self, target_format):
1828
"""Create a CopyConverter.
1830
:param target_format: The format the resulting repository should be.
1832
self.target_format = target_format
1834
def convert(self, repo, pb):
1835
"""Perform the conversion of to_convert, giving feedback via pb.
1837
:param to_convert: The disk object to convert.
1838
:param pb: a progress bar to use for progress information.
1843
# this is only useful with metadir layouts - separated repo content.
1844
# trigger an assertion if not such
1845
repo._format.get_format_string()
1846
self.repo_dir = repo.bzrdir
1847
self.step('Moving repository to repository.backup')
1848
self.repo_dir.transport.move('repository', 'repository.backup')
1849
backup_transport = self.repo_dir.transport.clone('repository.backup')
1850
repo._format.check_conversion_target(self.target_format)
1851
self.source_repo = repo._format.open(self.repo_dir,
1853
_override_transport=backup_transport)
1854
self.step('Creating new repository')
1855
converted = self.target_format.initialize(self.repo_dir,
1856
self.source_repo.is_shared())
1857
converted.lock_write()
1859
self.step('Copying content into repository.')
1860
self.source_repo.copy_content_into(converted)
1863
self.step('Deleting old repository content.')
1864
self.repo_dir.transport.delete_tree('repository.backup')
1865
self.pb.note('repository converted')
1867
def step(self, message):
1868
"""Update the pb by a step."""
1870
self.pb.update(message, self.count, self.total)
1873
class CommitBuilder(object):
1874
"""Provides an interface to build up a commit.
1876
This allows describing a tree to be committed without needing to
1877
know the internals of the format of the repository.
1880
record_root_entry = False
1881
def __init__(self, repository, parents, config, timestamp=None,
1882
timezone=None, committer=None, revprops=None,
1884
"""Initiate a CommitBuilder.
1886
:param repository: Repository to commit to.
1887
:param parents: Revision ids of the parents of the new revision.
1888
:param config: Configuration to use.
1889
:param timestamp: Optional timestamp recorded for commit.
1890
:param timezone: Optional timezone for timestamp.
1891
:param committer: Optional committer to set for commit.
1892
:param revprops: Optional dictionary of revision properties.
1893
:param revision_id: Optional revision id.
1895
self._config = config
1897
if committer is None:
1898
self._committer = self._config.username()
1900
assert isinstance(committer, basestring), type(committer)
1901
self._committer = committer
1903
self.new_inventory = Inventory(None)
1904
self._new_revision_id = osutils.safe_revision_id(revision_id)
1905
self.parents = parents
1906
self.repository = repository
1909
if revprops is not None:
1910
self._revprops.update(revprops)
1912
if timestamp is None:
1913
timestamp = time.time()
1914
# Restrict resolution to 1ms
1915
self._timestamp = round(timestamp, 3)
1917
if timezone is None:
1918
self._timezone = osutils.local_time_offset()
1920
self._timezone = int(timezone)
1922
self._generate_revision_if_needed()
1924
def commit(self, message):
1925
"""Make the actual commit.
1927
:return: The revision id of the recorded revision.
1929
rev = _mod_revision.Revision(
1930
timestamp=self._timestamp,
1931
timezone=self._timezone,
1932
committer=self._committer,
1934
inventory_sha1=self.inv_sha1,
1935
revision_id=self._new_revision_id,
1936
properties=self._revprops)
1937
rev.parent_ids = self.parents
1938
self.repository.add_revision(self._new_revision_id, rev,
1939
self.new_inventory, self._config)
1940
return self._new_revision_id
1942
def revision_tree(self):
1943
"""Return the tree that was just committed.
1945
After calling commit() this can be called to get a RevisionTree
1946
representing the newly committed tree. This is preferred to
1947
calling Repository.revision_tree() because that may require
1948
deserializing the inventory, while we already have a copy in
1951
return RevisionTree(self.repository, self.new_inventory,
1952
self._new_revision_id)
1954
def finish_inventory(self):
1955
"""Tell the builder that the inventory is finished."""
1956
if self.new_inventory.root is None:
1957
symbol_versioning.warn('Root entry should be supplied to'
1958
' record_entry_contents, as of bzr 0.10.',
1959
DeprecationWarning, stacklevel=2)
1960
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1961
self.new_inventory.revision_id = self._new_revision_id
1962
self.inv_sha1 = self.repository.add_inventory(
1963
self._new_revision_id,
1968
def _gen_revision_id(self):
1969
"""Return new revision-id."""
1970
return generate_ids.gen_revision_id(self._config.username(),
1973
def _generate_revision_if_needed(self):
1974
"""Create a revision id if None was supplied.
1976
If the repository can not support user-specified revision ids
1977
they should override this function and raise CannotSetRevisionId
1978
if _new_revision_id is not None.
1980
:raises: CannotSetRevisionId
1982
if self._new_revision_id is None:
1983
self._new_revision_id = self._gen_revision_id()
1985
def record_entry_contents(self, ie, parent_invs, path, tree):
1986
"""Record the content of ie from tree into the commit if needed.
1988
Side effect: sets ie.revision when unchanged
1990
:param ie: An inventory entry present in the commit.
1991
:param parent_invs: The inventories of the parent revisions of the
1993
:param path: The path the entry is at in the tree.
1994
:param tree: The tree which contains this entry and should be used to
1997
if self.new_inventory.root is None and ie.parent_id is not None:
1998
symbol_versioning.warn('Root entry should be supplied to'
1999
' record_entry_contents, as of bzr 0.10.',
2000
DeprecationWarning, stacklevel=2)
2001
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2003
self.new_inventory.add(ie)
2005
# ie.revision is always None if the InventoryEntry is considered
2006
# for committing. ie.snapshot will record the correct revision
2007
# which may be the sole parent if it is untouched.
2008
if ie.revision is not None:
2011
# In this revision format, root entries have no knit or weave
2012
if ie is self.new_inventory.root:
2013
# When serializing out to disk and back in
2014
# root.revision is always _new_revision_id
2015
ie.revision = self._new_revision_id
2017
previous_entries = ie.find_previous_heads(
2019
self.repository.weave_store,
2020
self.repository.get_transaction())
2021
# we are creating a new revision for ie in the history store
2023
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2025
def modified_directory(self, file_id, file_parents):
2026
"""Record the presence of a symbolic link.
2028
:param file_id: The file_id of the link to record.
2029
:param file_parents: The per-file parent revision ids.
2031
self._add_text_to_weave(file_id, [], file_parents.keys())
2033
def modified_reference(self, file_id, file_parents):
2034
"""Record the modification of a reference.
2036
:param file_id: The file_id of the link to record.
2037
:param file_parents: The per-file parent revision ids.
2039
self._add_text_to_weave(file_id, [], file_parents.keys())
2041
def modified_file_text(self, file_id, file_parents,
2042
get_content_byte_lines, text_sha1=None,
2044
"""Record the text of file file_id
2046
:param file_id: The file_id of the file to record the text of.
2047
:param file_parents: The per-file parent revision ids.
2048
:param get_content_byte_lines: A callable which will return the byte
2050
:param text_sha1: Optional SHA1 of the file contents.
2051
:param text_size: Optional size of the file contents.
2053
# mutter('storing text of file {%s} in revision {%s} into %r',
2054
# file_id, self._new_revision_id, self.repository.weave_store)
2055
# special case to avoid diffing on renames or
2057
if (len(file_parents) == 1
2058
and text_sha1 == file_parents.values()[0].text_sha1
2059
and text_size == file_parents.values()[0].text_size):
2060
previous_ie = file_parents.values()[0]
2061
versionedfile = self.repository.weave_store.get_weave(file_id,
2062
self.repository.get_transaction())
2063
versionedfile.clone_text(self._new_revision_id,
2064
previous_ie.revision, file_parents.keys())
2065
return text_sha1, text_size
2067
new_lines = get_content_byte_lines()
2068
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2069
# should return the SHA1 and size
2070
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2071
return osutils.sha_strings(new_lines), \
2072
sum(map(len, new_lines))
2074
def modified_link(self, file_id, file_parents, link_target):
2075
"""Record the presence of a symbolic link.
2077
:param file_id: The file_id of the link to record.
2078
:param file_parents: The per-file parent revision ids.
2079
:param link_target: Target location of this link.
2081
self._add_text_to_weave(file_id, [], file_parents.keys())
2083
def _add_text_to_weave(self, file_id, new_lines, parents):
2084
versionedfile = self.repository.weave_store.get_weave_or_empty(
2085
file_id, self.repository.get_transaction())
2086
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2087
versionedfile.clear_cache()
2090
class _CommitBuilder(CommitBuilder):
2091
"""Temporary class so old CommitBuilders are detected properly
2093
Note: CommitBuilder works whether or not root entry is recorded.
2096
record_root_entry = True
2099
class RootCommitBuilder(CommitBuilder):
2100
"""This commitbuilder actually records the root id"""
2102
record_root_entry = True
2104
def record_entry_contents(self, ie, parent_invs, path, tree):
2105
"""Record the content of ie from tree into the commit if needed.
2107
Side effect: sets ie.revision when unchanged
2109
:param ie: An inventory entry present in the commit.
2110
:param parent_invs: The inventories of the parent revisions of the
2112
:param path: The path the entry is at in the tree.
2113
:param tree: The tree which contains this entry and should be used to
2116
assert self.new_inventory.root is not None or ie.parent_id is None
2117
self.new_inventory.add(ie)
2119
# ie.revision is always None if the InventoryEntry is considered
2120
# for committing. ie.snapshot will record the correct revision
2121
# which may be the sole parent if it is untouched.
2122
if ie.revision is not None:
2125
previous_entries = ie.find_previous_heads(
2127
self.repository.weave_store,
2128
self.repository.get_transaction())
2129
# we are creating a new revision for ie in the history store
2131
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2143
def _unescaper(match, _map=_unescape_map):
2144
code = match.group(1)
2148
if not code.startswith('#'):
2150
return unichr(int(code[1:])).encode('utf8')
2156
def _unescape_xml(data):
2157
"""Unescape predefined XML entities in a string of data."""
2159
if _unescape_re is None:
2160
_unescape_re = re.compile('\&([^;]*);')
2161
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
