« back to all changes in this revision
Viewing changes to bzrlib/repository.py
- browse files at revision 963
- compare with another revision
- download diff
- download tarball
- view history from revision 963
- Committer: Martin Pool
- Date: 2005-07-22 22:37:53 UTC
- Revision ID: mbp@sourcefrog.net-20050722223753-7dced4e32d3ce21d
- add the start of a test for inventory file-id matching
- files added:
- bzrlib/mdiff.py
- notes
- patches
- plugins/changeset
- files removed:
- BRANCH.TODO
- bzrlib/benchmarks
- bzrlib/bundle
- bzrlib/bundle/old
- bzrlib/bundle/serializer
- bzrlib/doc
- bzrlib/doc/api
- bzrlib/export
- bzrlib/plugins/launchpad
- bzrlib/store
- bzrlib/store/revision
- bzrlib/store/versioned
- bzrlib/tests/blackbox
- bzrlib/tests/branch_implementations
- bzrlib/tests/bzrdir_implementations
- bzrlib/tests/interrepository_implementations
- bzrlib/tests/interversionedfile_implementations
- bzrlib/tests/repository_implementations
- bzrlib/tests/revisionstore_implementations
- bzrlib/tests/test_patches_data
- bzrlib/tests/tree_implementations
- bzrlib/tests/workingtree_implementations
- bzrlib/transport
- bzrlib/transport/http
- bzrlib/ui
- bzrlib/util
- bzrlib/util/configobj
- bzrlib/util/configobj/docs
- contrib/emacs
- tools/doc_generate
- files renamed:
- tools/doc_generate/autodoc_man.py => bzr-man.py
- bzrlib/bundle/__init__.py => bzrlib/changeset.py
- contrib/newinventory.py => bzrlib/newinventory.py
- bzrlib/tests/ => bzrlib/selftest/
- bzrlib/tests/blackbox/test_too_much.py => bzrlib/selftest/blackbox.py
- bzrlib/tests/test_plugins.py => bzrlib/selftest/plugins.py
- bzrlib/tests/branch_implementations/test_branch.py => bzrlib/selftest/testbranch.py
- bzrlib/tests/test_hashcache.py => bzrlib/selftest/testhashcache.py
- bzrlib/tests/test_inv.py => bzrlib/selftest/testinv.py
- bzrlib/tests/test_merge3.py => bzrlib/selftest/testmerge3.py
- bzrlib/tests/test_revisionnamespaces.py => bzrlib/selftest/testrevisionnamespaces.py
- bzrlib/tests/blackbox/test_status.py => bzrlib/selftest/teststatus.py
- bzrlib/tests/blackbox/test_versioning.py => bzrlib/selftest/versioning.py
- bzrlib/tests/test_whitebox.py => bzrlib/selftest/whitebox.py
- bzrlib/store/__init__.py => bzrlib/store.py
- bzrlib/xml_serializer.py => bzrlib/xml.py
- bzrlib/util/effbot/ => effbot/
- bzrlib/util/elementtree/ => elementtree/
- bzrlib/plugins/ => plugins/
- bzrlib/tests/test_weave.py => tools/testweave.py
- bzrlib/util/urlgrabber/ => urlgrabber/
- files modified:
- .bzrignore
- bzr *
- setup.py *
added
removed
bzrlib/repository.py
1
# Copyright (C) 2005, 2006 Canonical Ltd
2
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
7
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
12
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
17
from binascii import hexlify
18
from copy import deepcopy
19
from cStringIO import StringIO
20
import re
21
import time
22
from unittest import TestSuite
23
24
from bzrlib import bzrdir, check, delta, gpg, errors, xml5, ui, transactions, osutils
25
from bzrlib.decorators import needs_read_lock, needs_write_lock
26
from bzrlib.errors import InvalidRevisionId
27
from bzrlib.graph import Graph
28
from bzrlib.inter import InterObject
29
from bzrlib.inventory import Inventory
30
from bzrlib.knit import KnitVersionedFile, KnitPlainFactory
31
from bzrlib.lockable_files import LockableFiles, TransportLock
32
from bzrlib.lockdir import LockDir
33
from bzrlib.osutils import (safe_unicode, rand_bytes, compact_date,
34
local_time_offset)
35
from bzrlib.revision import NULL_REVISION, Revision
36
from bzrlib.store.versioned import VersionedFileStore, WeaveStore
37
from bzrlib.store.text import TextStore
38
from bzrlib.symbol_versioning import (deprecated_method,
39
zero_nine,
40
)
41
from bzrlib.testament import Testament
42
from bzrlib.trace import mutter, note
43
from bzrlib.tree import RevisionTree
44
from bzrlib.tsort import topo_sort
45
from bzrlib.weave import WeaveFile
46
47
48
class Repository(object):
49
"""Repository holding history for one or more branches.
50
51
The repository holds and retrieves historical information including
52
revisions and file history. It's normally accessed only by the Branch,
53
which views a particular line of development through that history.
54
55
The Repository builds on top of Stores and a Transport, which respectively
56
describe the disk data format and the way of accessing the (possibly
57
remote) disk.
58
"""
59
60
@needs_write_lock
61
def add_inventory(self, revid, inv, parents):
62
"""Add the inventory inv to the repository as revid.
63
64
:param parents: The revision ids of the parents that revid
65
is known to have and are in the repository already.
66
67
returns the sha1 of the serialized inventory.
68
"""
69
assert inv.revision_id is None or inv.revision_id == revid, \
70
"Mismatch between inventory revision" \
71
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
72
inv_text = xml5.serializer_v5.write_inventory_to_string(inv)
73
inv_sha1 = osutils.sha_string(inv_text)
74
inv_vf = self.control_weaves.get_weave('inventory',
75
self.get_transaction())
76
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
77
return inv_sha1
78
79
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
80
final_parents = []
81
for parent in parents:
82
if parent in inv_vf:
83
final_parents.append(parent)
84
85
inv_vf.add_lines(revid, final_parents, lines)
86
87
@needs_write_lock
88
def add_revision(self, rev_id, rev, inv=None, config=None):
89
"""Add rev to the revision store as rev_id.
90
91
:param rev_id: the revision id to use.
92
:param rev: The revision object.
93
:param inv: The inventory for the revision. if None, it will be looked
94
up in the inventory storer
95
:param config: If None no digital signature will be created.
96
If supplied its signature_needed method will be used
97
to determine if a signature should be made.
98
"""
99
if config is not None and config.signature_needed():
100
if inv is None:
101
inv = self.get_inventory(rev_id)
102
plaintext = Testament(rev, inv).as_short_text()
103
self.store_revision_signature(
104
gpg.GPGStrategy(config), plaintext, rev_id)
105
if not rev_id in self.get_inventory_weave():
106
if inv is None:
107
raise errors.WeaveRevisionNotPresent(rev_id,
108
self.get_inventory_weave())
109
else:
110
# yes, this is not suitable for adding with ghosts.
111
self.add_inventory(rev_id, inv, rev.parent_ids)
112
self._revision_store.add_revision(rev, self.get_transaction())
113
114
@needs_read_lock
115
def _all_possible_ids(self):
116
"""Return all the possible revisions that we could find."""
117
return self.get_inventory_weave().versions()
118
119
def all_revision_ids(self):
120
"""Returns a list of all the revision ids in the repository.
121
122
This is deprecated because code should generally work on the graph
123
reachable from a particular revision, and ignore any other revisions
124
that might be present. There is no direct replacement method.
125
"""
126
return self._all_revision_ids()
127
128
@needs_read_lock
129
def _all_revision_ids(self):
130
"""Returns a list of all the revision ids in the repository.
131
132
These are in as much topological order as the underlying store can
133
present: for weaves ghosts may lead to a lack of correctness until
134
the reweave updates the parents list.
135
"""
136
if self._revision_store.text_store.listable():
137
return self._revision_store.all_revision_ids(self.get_transaction())
138
result = self._all_possible_ids()
139
return self._eliminate_revisions_not_present(result)
140
141
def break_lock(self):
142
"""Break a lock if one is present from another instance.
143
144
Uses the ui factory to ask for confirmation if the lock may be from
145
an active process.
146
"""
147
self.control_files.break_lock()
148
149
@needs_read_lock
150
def _eliminate_revisions_not_present(self, revision_ids):
151
"""Check every revision id in revision_ids to see if we have it.
152
153
Returns a set of the present revisions.
154
"""
155
result = []
156
for id in revision_ids:
157
if self.has_revision(id):
158
result.append(id)
159
return result
160
161
@staticmethod
162
def create(a_bzrdir):
163
"""Construct the current default format repository in a_bzrdir."""
164
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
165
166
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
167
"""instantiate a Repository.
168
169
:param _format: The format of the repository on disk.
170
:param a_bzrdir: The BzrDir of the repository.
171
172
In the future we will have a single api for all stores for
173
getting file texts, inventories and revisions, then
174
this construct will accept instances of those things.
175
"""
176
super(Repository, self).__init__()
177
self._format = _format
178
# the following are part of the public API for Repository:
179
self.bzrdir = a_bzrdir
180
self.control_files = control_files
181
self._revision_store = _revision_store
182
self.text_store = text_store
183
# backwards compatibility
184
self.weave_store = text_store
185
# not right yet - should be more semantically clear ?
186
#
187
self.control_store = control_store
188
self.control_weaves = control_store
189
# TODO: make sure to construct the right store classes, etc, depending
190
# on whether escaping is required.
191
192
def __repr__(self):
193
return '%s(%r)' % (self.__class__.__name__,
194
self.bzrdir.transport.base)
195
196
def is_locked(self):
197
return self.control_files.is_locked()
198
199
def lock_write(self):
200
self.control_files.lock_write()
201
202
def lock_read(self):
203
self.control_files.lock_read()
204
205
def get_physical_lock_status(self):
206
return self.control_files.get_physical_lock_status()
207
208
@needs_read_lock
209
def missing_revision_ids(self, other, revision_id=None):
210
"""Return the revision ids that other has that this does not.
211
212
These are returned in topological order.
213
214
revision_id: only return revision ids included by revision_id.
215
"""
216
return InterRepository.get(other, self).missing_revision_ids(revision_id)
217
218
@staticmethod
219
def open(base):
220
"""Open the repository rooted at base.
221
222
For instance, if the repository is at URL/.bzr/repository,
223
Repository.open(URL) -> a Repository instance.
224
"""
225
control = bzrdir.BzrDir.open(base)
226
return control.open_repository()
227
228
def copy_content_into(self, destination, revision_id=None, basis=None):
229
"""Make a complete copy of the content in self into destination.
230
231
This is a destructive operation! Do not use it on existing
232
repositories.
233
"""
234
return InterRepository.get(self, destination).copy_content(revision_id, basis)
235
236
def fetch(self, source, revision_id=None, pb=None):
237
"""Fetch the content required to construct revision_id from source.
238
239
If revision_id is None all content is copied.
240
"""
241
return InterRepository.get(source, self).fetch(revision_id=revision_id,
242
pb=pb)
243
244
def get_commit_builder(self, branch, parents, config, timestamp=None,
245
timezone=None, committer=None, revprops=None,
246
revision_id=None):
247
"""Obtain a CommitBuilder for this repository.
248
249
:param branch: Branch to commit to.
250
:param parents: Revision ids of the parents of the new revision.
251
:param config: Configuration to use.
252
:param timestamp: Optional timestamp recorded for commit.
253
:param timezone: Optional timezone for timestamp.
254
:param committer: Optional committer to set for commit.
255
:param revprops: Optional dictionary of revision properties.
256
:param revision_id: Optional revision id.
257
"""
258
return CommitBuilder(self, parents, config, timestamp, timezone,
259
committer, revprops, revision_id)
260
261
def unlock(self):
262
self.control_files.unlock()
263
264
@needs_read_lock
265
def clone(self, a_bzrdir, revision_id=None, basis=None):
266
"""Clone this repository into a_bzrdir using the current format.
267
268
Currently no check is made that the format of this repository and
269
the bzrdir format are compatible. FIXME RBC 20060201.
270
"""
271
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
272
# use target default format.
273
result = a_bzrdir.create_repository()
274
# FIXME RBC 20060209 split out the repository type to avoid this check ?
275
elif isinstance(a_bzrdir._format,
276
(bzrdir.BzrDirFormat4,
277
bzrdir.BzrDirFormat5,
278
bzrdir.BzrDirFormat6)):
279
result = a_bzrdir.open_repository()
280
else:
281
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
282
self.copy_content_into(result, revision_id, basis)
283
return result
284
285
@needs_read_lock
286
def has_revision(self, revision_id):
287
"""True if this repository has a copy of the revision."""
288
return self._revision_store.has_revision_id(revision_id,
289
self.get_transaction())
290
291
@needs_read_lock
292
def get_revision_reconcile(self, revision_id):
293
"""'reconcile' helper routine that allows access to a revision always.
294
295
This variant of get_revision does not cross check the weave graph
296
against the revision one as get_revision does: but it should only
297
be used by reconcile, or reconcile-alike commands that are correcting
298
or testing the revision graph.
299
"""
300
if not revision_id or not isinstance(revision_id, basestring):
301
raise InvalidRevisionId(revision_id=revision_id, branch=self)
302
return self._revision_store.get_revisions([revision_id],
303
self.get_transaction())[0]
304
@needs_read_lock
305
def get_revisions(self, revision_ids):
306
return self._revision_store.get_revisions(revision_ids,
307
self.get_transaction())
308
309
@needs_read_lock
310
def get_revision_xml(self, revision_id):
311
rev = self.get_revision(revision_id)
312
rev_tmp = StringIO()
313
# the current serializer..
314
self._revision_store._serializer.write_revision(rev, rev_tmp)
315
rev_tmp.seek(0)
316
return rev_tmp.getvalue()
317
318
@needs_read_lock
319
def get_revision(self, revision_id):
320
"""Return the Revision object for a named revision"""
321
r = self.get_revision_reconcile(revision_id)
322
# weave corruption can lead to absent revision markers that should be
323
# present.
324
# the following test is reasonably cheap (it needs a single weave read)
325
# and the weave is cached in read transactions. In write transactions
326
# it is not cached but typically we only read a small number of
327
# revisions. For knits when they are introduced we will probably want
328
# to ensure that caching write transactions are in use.
329
inv = self.get_inventory_weave()
330
self._check_revision_parents(r, inv)
331
return r
332
333
@needs_read_lock
334
def get_deltas_for_revisions(self, revisions):
335
"""Produce a generator of revision deltas.
336
337
Note that the input is a sequence of REVISIONS, not revision_ids.
338
Trees will be held in memory until the generator exits.
339
Each delta is relative to the revision's lefthand predecessor.
340
"""
341
required_trees = set()
342
for revision in revisions:
343
required_trees.add(revision.revision_id)
344
required_trees.update(revision.parent_ids[:1])
345
trees = dict((t.get_revision_id(), t) for
346
t in self.revision_trees(required_trees))
347
for revision in revisions:
348
if not revision.parent_ids:
349
old_tree = self.revision_tree(None)
350
else:
351
old_tree = trees[revision.parent_ids[0]]
352
yield delta.compare_trees(old_tree, trees[revision.revision_id])
353
354
@needs_read_lock
355
def get_revision_delta(self, revision_id):
356
"""Return the delta for one revision.
357
358
The delta is relative to the left-hand predecessor of the
359
revision.
360
"""
361
r = self.get_revision(revision_id)
362
return list(self.get_deltas_for_revisions([r]))[0]
363
364
def _check_revision_parents(self, revision, inventory):
365
"""Private to Repository and Fetch.
366
367
This checks the parentage of revision in an inventory weave for
368
consistency and is only applicable to inventory-weave-for-ancestry
369
using repository formats & fetchers.
370
"""
371
weave_parents = inventory.get_parents(revision.revision_id)
372
weave_names = inventory.versions()
373
for parent_id in revision.parent_ids:
374
if parent_id in weave_names:
375
# this parent must not be a ghost.
376
if not parent_id in weave_parents:
377
# but it is a ghost
378
raise errors.CorruptRepository(self)
379
380
@needs_write_lock
381
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
382
signature = gpg_strategy.sign(plaintext)
383
self._revision_store.add_revision_signature_text(revision_id,
384
signature,
385
self.get_transaction())
386
387
def fileids_altered_by_revision_ids(self, revision_ids):
388
"""Find the file ids and versions affected by revisions.
389
390
:param revisions: an iterable containing revision ids.
391
:return: a dictionary mapping altered file-ids to an iterable of
392
revision_ids. Each altered file-ids has the exact revision_ids that
393
altered it listed explicitly.
394
"""
395
assert isinstance(self._format, (RepositoryFormat5,
396
RepositoryFormat6,
397
RepositoryFormat7,
398
RepositoryFormatKnit1)), \
399
("fileids_altered_by_revision_ids only supported for branches "
400
"which store inventory as unnested xml, not on %r" % self)
401
selected_revision_ids = set(revision_ids)
402
w = self.get_inventory_weave()
403
result = {}
404
405
# this code needs to read every new line in every inventory for the
406
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
407
# not present in one of those inventories is unnecessary but not
408
# harmful because we are filtering by the revision id marker in the
409
# inventory lines : we only select file ids altered in one of those
410
# revisions. We don't need to see all lines in the inventory because
411
# only those added in an inventory in rev X can contain a revision=X
412
# line.
413
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
414
start = line.find('file_id="')+9
415
if start < 9: continue
416
end = line.find('"', start)
417
assert end>= 0
418
file_id = _unescape_xml(line[start:end])
419
420
start = line.find('revision="')+10
421
if start < 10: continue
422
end = line.find('"', start)
423
assert end>= 0
424
revision_id = _unescape_xml(line[start:end])
425
if revision_id in selected_revision_ids:
426
result.setdefault(file_id, set()).add(revision_id)
427
return result
428
429
@needs_read_lock
430
def get_inventory_weave(self):
431
return self.control_weaves.get_weave('inventory',
432
self.get_transaction())
433
434
@needs_read_lock
435
def get_inventory(self, revision_id):
436
"""Get Inventory object by hash."""
437
return self.deserialise_inventory(
438
revision_id, self.get_inventory_xml(revision_id))
439
440
def deserialise_inventory(self, revision_id, xml):
441
"""Transform the xml into an inventory object.
442
443
:param revision_id: The expected revision id of the inventory.
444
:param xml: A serialised inventory.
445
"""
446
return xml5.serializer_v5.read_inventory_from_string(xml)
447
448
@needs_read_lock
449
def get_inventory_xml(self, revision_id):
450
"""Get inventory XML as a file object."""
451
try:
452
assert isinstance(revision_id, basestring), type(revision_id)
453
iw = self.get_inventory_weave()
454
return iw.get_text(revision_id)
455
except IndexError:
456
raise errors.HistoryMissing(self, 'inventory', revision_id)
457
458
@needs_read_lock
459
def get_inventory_sha1(self, revision_id):
460
"""Return the sha1 hash of the inventory entry
461
"""
462
return self.get_revision(revision_id).inventory_sha1
463
464
@needs_read_lock
465
def get_revision_graph(self, revision_id=None):
466
"""Return a dictionary containing the revision graph.
467
468
:param revision_id: The revision_id to get a graph from. If None, then
469
the entire revision graph is returned. This is a deprecated mode of
470
operation and will be removed in the future.
471
:return: a dictionary of revision_id->revision_parents_list.
472
"""
473
# special case NULL_REVISION
474
if revision_id == NULL_REVISION:
475
return {}
476
weave = self.get_inventory_weave()
477
all_revisions = self._eliminate_revisions_not_present(weave.versions())
478
entire_graph = dict([(node, weave.get_parents(node)) for
479
node in all_revisions])
480
if revision_id is None:
481
return entire_graph
482
elif revision_id not in entire_graph:
483
raise errors.NoSuchRevision(self, revision_id)
484
else:
485
# add what can be reached from revision_id
486
result = {}
487
pending = set([revision_id])
488
while len(pending) > 0:
489
node = pending.pop()
490
result[node] = entire_graph[node]
491
for revision_id in result[node]:
492
if revision_id not in result:
493
pending.add(revision_id)
494
return result
495
496
@needs_read_lock
497
def get_revision_graph_with_ghosts(self, revision_ids=None):
498
"""Return a graph of the revisions with ghosts marked as applicable.
499
500
:param revision_ids: an iterable of revisions to graph or None for all.
501
:return: a Graph object with the graph reachable from revision_ids.
502
"""
503
result = Graph()
504
if not revision_ids:
505
pending = set(self.all_revision_ids())
506
required = set([])
507
else:
508
pending = set(revision_ids)
509
# special case NULL_REVISION
510
if NULL_REVISION in pending:
511
pending.remove(NULL_REVISION)
512
required = set(pending)
513
done = set([])
514
while len(pending):
515
revision_id = pending.pop()
516
try:
517
rev = self.get_revision(revision_id)
518
except errors.NoSuchRevision:
519
if revision_id in required:
520
raise
521
# a ghost
522
result.add_ghost(revision_id)
523
continue
524
for parent_id in rev.parent_ids:
525
# is this queued or done ?
526
if (parent_id not in pending and
527
parent_id not in done):
528
# no, queue it.
529
pending.add(parent_id)
530
result.add_node(revision_id, rev.parent_ids)
531
done.add(revision_id)
532
return result
533
534
@needs_read_lock
535
def get_revision_inventory(self, revision_id):
536
"""Return inventory of a past revision."""
537
# TODO: Unify this with get_inventory()
538
# bzr 0.0.6 and later imposes the constraint that the inventory_id
539
# must be the same as its revision, so this is trivial.
540
if revision_id is None:
541
# This does not make sense: if there is no revision,
542
# then it is the current tree inventory surely ?!
543
# and thus get_root_id() is something that looks at the last
544
# commit on the branch, and the get_root_id is an inventory check.
545
raise NotImplementedError
546
# return Inventory(self.get_root_id())
547
else:
548
return self.get_inventory(revision_id)
549
550
@needs_read_lock
551
def is_shared(self):
552
"""Return True if this repository is flagged as a shared repository."""
553
raise NotImplementedError(self.is_shared)
554
555
@needs_write_lock
556
def reconcile(self, other=None, thorough=False):
557
"""Reconcile this repository."""
558
from bzrlib.reconcile import RepoReconciler
559
reconciler = RepoReconciler(self, thorough=thorough)
560
reconciler.reconcile()
561
return reconciler
562
563
@needs_read_lock
564
def revision_tree(self, revision_id):
565
"""Return Tree for a revision on this branch.
566
567
`revision_id` may be None for the empty tree revision.
568
"""
569
# TODO: refactor this to use an existing revision object
570
# so we don't need to read it in twice.
571
if revision_id is None or revision_id == NULL_REVISION:
572
return RevisionTree(self, Inventory(), NULL_REVISION)
573
else:
574
inv = self.get_revision_inventory(revision_id)
575
return RevisionTree(self, inv, revision_id)
576
577
@needs_read_lock
578
def revision_trees(self, revision_ids):
579
"""Return Tree for a revision on this branch.
580
581
`revision_id` may not be None or 'null:'"""
582
assert None not in revision_ids
583
assert NULL_REVISION not in revision_ids
584
texts = self.get_inventory_weave().get_texts(revision_ids)
585
for text, revision_id in zip(texts, revision_ids):
586
inv = self.deserialise_inventory(revision_id, text)
587
yield RevisionTree(self, inv, revision_id)
588
589
@needs_read_lock
590
def get_ancestry(self, revision_id):
591
"""Return a list of revision-ids integrated by a revision.
592
593
The first element of the list is always None, indicating the origin
594
revision. This might change when we have history horizons, or
595
perhaps we should have a new API.
596
597
This is topologically sorted.
598
"""
599
if revision_id is None:
600
return [None]
601
if not self.has_revision(revision_id):
602
raise errors.NoSuchRevision(self, revision_id)
603
w = self.get_inventory_weave()
604
candidates = w.get_ancestry(revision_id)
605
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
606
607
@needs_read_lock
608
def print_file(self, file, revision_id):
609
"""Print `file` to stdout.
610
611
FIXME RBC 20060125 as John Meinel points out this is a bad api
612
- it writes to stdout, it assumes that that is valid etc. Fix
613
by creating a new more flexible convenience function.
614
"""
615
tree = self.revision_tree(revision_id)
616
# use inventory as it was in that revision
617
file_id = tree.inventory.path2id(file)
618
if not file_id:
619
# TODO: jam 20060427 Write a test for this code path
620
# it had a bug in it, and was raising the wrong
621
# exception.
622
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
623
tree.print_file(file_id)
624
625
def get_transaction(self):
626
return self.control_files.get_transaction()
627
628
def revision_parents(self, revid):
629
return self.get_inventory_weave().parent_names(revid)
630
631
@needs_write_lock
632
def set_make_working_trees(self, new_value):
633
"""Set the policy flag for making working trees when creating branches.
634
635
This only applies to branches that use this repository.
636
637
The default is 'True'.
638
:param new_value: True to restore the default, False to disable making
639
working trees.
640
"""
641
raise NotImplementedError(self.set_make_working_trees)
642
643
def make_working_trees(self):
644
"""Returns the policy for making working trees on new branches."""
645
raise NotImplementedError(self.make_working_trees)
646
647
@needs_write_lock
648
def sign_revision(self, revision_id, gpg_strategy):
649
plaintext = Testament.from_revision(self, revision_id).as_short_text()
650
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
651
652
@needs_read_lock
653
def has_signature_for_revision_id(self, revision_id):
654
"""Query for a revision signature for revision_id in the repository."""
655
return self._revision_store.has_signature(revision_id,
656
self.get_transaction())
657
658
@needs_read_lock
659
def get_signature_text(self, revision_id):
660
"""Return the text for a signature."""
661
return self._revision_store.get_signature_text(revision_id,
662
self.get_transaction())
663
664
@needs_read_lock
665
def check(self, revision_ids):
666
"""Check consistency of all history of given revision_ids.
667
668
Different repository implementations should override _check().
669
670
:param revision_ids: A non-empty list of revision_ids whose ancestry
671
will be checked. Typically the last revision_id of a branch.
672
"""
673
if not revision_ids:
674
raise ValueError("revision_ids must be non-empty in %s.check"
675
% (self,))
676
return self._check(revision_ids)
677
678
def _check(self, revision_ids):
679
result = check.Check(self)
680
result.check()
681
return result
682
683
684
class AllInOneRepository(Repository):
685
"""Legacy support - the repository behaviour for all-in-one branches."""
686
687
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
688
# we reuse one control files instance.
689
dir_mode = a_bzrdir._control_files._dir_mode
690
file_mode = a_bzrdir._control_files._file_mode
691
692
def get_store(name, compressed=True, prefixed=False):
693
# FIXME: This approach of assuming stores are all entirely compressed
694
# or entirely uncompressed is tidy, but breaks upgrade from
695
# some existing branches where there's a mixture; we probably
696
# still want the option to look for both.
697
relpath = a_bzrdir._control_files._escape(name)
698
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
699
prefixed=prefixed, compressed=compressed,
700
dir_mode=dir_mode,
701
file_mode=file_mode)
702
#if self._transport.should_cache():
703
# cache_path = os.path.join(self.cache_root, name)
704
# os.mkdir(cache_path)
705
# store = bzrlib.store.CachedStore(store, cache_path)
706
return store
707
708
# not broken out yet because the controlweaves|inventory_store
709
# and text_store | weave_store bits are still different.
710
if isinstance(_format, RepositoryFormat4):
711
# cannot remove these - there is still no consistent api
712
# which allows access to this old info.
713
self.inventory_store = get_store('inventory-store')
714
text_store = get_store('text-store')
715
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
716
717
@needs_read_lock
718
def is_shared(self):
719
"""AllInOne repositories cannot be shared."""
720
return False
721
722
@needs_write_lock
723
def set_make_working_trees(self, new_value):
724
"""Set the policy flag for making working trees when creating branches.
725
726
This only applies to branches that use this repository.
727
728
The default is 'True'.
729
:param new_value: True to restore the default, False to disable making
730
working trees.
731
"""
732
raise NotImplementedError(self.set_make_working_trees)
733
734
def make_working_trees(self):
735
"""Returns the policy for making working trees on new branches."""
736
return True
737
738
739
def install_revision(repository, rev, revision_tree):
740
"""Install all revision data into a repository."""
741
present_parents = []
742
parent_trees = {}
743
for p_id in rev.parent_ids:
744
if repository.has_revision(p_id):
745
present_parents.append(p_id)
746
parent_trees[p_id] = repository.revision_tree(p_id)
747
else:
748
parent_trees[p_id] = repository.revision_tree(None)
749
750
inv = revision_tree.inventory
751
752
# backwards compatability hack: skip the root id.
753
entries = inv.iter_entries()
754
entries.next()
755
# Add the texts that are not already present
756
for path, ie in entries:
757
w = repository.weave_store.get_weave_or_empty(ie.file_id,
758
repository.get_transaction())
759
if ie.revision not in w:
760
text_parents = []
761
# FIXME: TODO: The following loop *may* be overlapping/duplicate
762
# with InventoryEntry.find_previous_heads(). if it is, then there
763
# is a latent bug here where the parents may have ancestors of each
764
# other. RBC, AB
765
for revision, tree in parent_trees.iteritems():
766
if ie.file_id not in tree:
767
continue
768
parent_id = tree.inventory[ie.file_id].revision
769
if parent_id in text_parents:
770
continue
771
text_parents.append(parent_id)
772
773
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
774
repository.get_transaction())
775
lines = revision_tree.get_file(ie.file_id).readlines()
776
vfile.add_lines(rev.revision_id, text_parents, lines)
777
try:
778
# install the inventory
779
repository.add_inventory(rev.revision_id, inv, present_parents)
780
except errors.RevisionAlreadyPresent:
781
pass
782
repository.add_revision(rev.revision_id, rev, inv)
783
784
785
class MetaDirRepository(Repository):
786
"""Repositories in the new meta-dir layout."""
787
788
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
789
super(MetaDirRepository, self).__init__(_format,
790
a_bzrdir,
791
control_files,
792
_revision_store,
793
control_store,
794
text_store)
795
796
dir_mode = self.control_files._dir_mode
797
file_mode = self.control_files._file_mode
798
799
@needs_read_lock
800
def is_shared(self):
801
"""Return True if this repository is flagged as a shared repository."""
802
return self.control_files._transport.has('shared-storage')
803
804
@needs_write_lock
805
def set_make_working_trees(self, new_value):
806
"""Set the policy flag for making working trees when creating branches.
807
808
This only applies to branches that use this repository.
809
810
The default is 'True'.
811
:param new_value: True to restore the default, False to disable making
812
working trees.
813
"""
814
if new_value:
815
try:
816
self.control_files._transport.delete('no-working-trees')
817
except errors.NoSuchFile:
818
pass
819
else:
820
self.control_files.put_utf8('no-working-trees', '')
821
822
def make_working_trees(self):
823
"""Returns the policy for making working trees on new branches."""
824
return not self.control_files._transport.has('no-working-trees')
825
826
827
class KnitRepository(MetaDirRepository):
828
"""Knit format repository."""
829
830
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
831
inv_vf.add_lines_with_ghosts(revid, parents, lines)
832
833
@needs_read_lock
834
def _all_revision_ids(self):
835
"""See Repository.all_revision_ids()."""
836
# Knits get the revision graph from the index of the revision knit, so
837
# it's always possible even if they're on an unlistable transport.
838
return self._revision_store.all_revision_ids(self.get_transaction())
839
840
def fileid_involved_between_revs(self, from_revid, to_revid):
841
"""Find file_id(s) which are involved in the changes between revisions.
842
843
This determines the set of revisions which are involved, and then
844
finds all file ids affected by those revisions.
845
"""
846
vf = self._get_revision_vf()
847
from_set = set(vf.get_ancestry(from_revid))
848
to_set = set(vf.get_ancestry(to_revid))
849
changed = to_set.difference(from_set)
850
return self._fileid_involved_by_set(changed)
851
852
def fileid_involved(self, last_revid=None):
853
"""Find all file_ids modified in the ancestry of last_revid.
854
855
:param last_revid: If None, last_revision() will be used.
856
"""
857
if not last_revid:
858
changed = set(self.all_revision_ids())
859
else:
860
changed = set(self.get_ancestry(last_revid))
861
if None in changed:
862
changed.remove(None)
863
return self._fileid_involved_by_set(changed)
864
865
@needs_read_lock
866
def get_ancestry(self, revision_id):
867
"""Return a list of revision-ids integrated by a revision.
868
869
This is topologically sorted.
870
"""
871
if revision_id is None:
872
return [None]
873
vf = self._get_revision_vf()
874
try:
875
return [None] + vf.get_ancestry(revision_id)
876
except errors.RevisionNotPresent:
877
raise errors.NoSuchRevision(self, revision_id)
878
879
@needs_read_lock
880
def get_revision(self, revision_id):
881
"""Return the Revision object for a named revision"""
882
return self.get_revision_reconcile(revision_id)
883
884
@needs_read_lock
885
def get_revision_graph(self, revision_id=None):
886
"""Return a dictionary containing the revision graph.
887
888
:param revision_id: The revision_id to get a graph from. If None, then
889
the entire revision graph is returned. This is a deprecated mode of
890
operation and will be removed in the future.
891
:return: a dictionary of revision_id->revision_parents_list.
892
"""
893
# special case NULL_REVISION
894
if revision_id == NULL_REVISION:
895
return {}
896
weave = self._get_revision_vf()
897
entire_graph = weave.get_graph()
898
if revision_id is None:
899
return weave.get_graph()
900
elif revision_id not in weave:
901
raise errors.NoSuchRevision(self, revision_id)
902
else:
903
# add what can be reached from revision_id
904
result = {}
905
pending = set([revision_id])
906
while len(pending) > 0:
907
node = pending.pop()
908
result[node] = weave.get_parents(node)
909
for revision_id in result[node]:
910
if revision_id not in result:
911
pending.add(revision_id)
912
return result
913
914
@needs_read_lock
915
def get_revision_graph_with_ghosts(self, revision_ids=None):
916
"""Return a graph of the revisions with ghosts marked as applicable.
917
918
:param revision_ids: an iterable of revisions to graph or None for all.
919
:return: a Graph object with the graph reachable from revision_ids.
920
"""
921
result = Graph()
922
vf = self._get_revision_vf()
923
versions = set(vf.versions())
924
if not revision_ids:
925
pending = set(self.all_revision_ids())
926
required = set([])
927
else:
928
pending = set(revision_ids)
929
# special case NULL_REVISION
930
if NULL_REVISION in pending:
931
pending.remove(NULL_REVISION)
932
required = set(pending)
933
done = set([])
934
while len(pending):
935
revision_id = pending.pop()
936
if not revision_id in versions:
937
if revision_id in required:
938
raise errors.NoSuchRevision(self, revision_id)
939
# a ghost
940
result.add_ghost(revision_id)
941
# mark it as done so we don't try for it again.
942
done.add(revision_id)
943
continue
944
parent_ids = vf.get_parents_with_ghosts(revision_id)
945
for parent_id in parent_ids:
946
# is this queued or done ?
947
if (parent_id not in pending and
948
parent_id not in done):
949
# no, queue it.
950
pending.add(parent_id)
951
result.add_node(revision_id, parent_ids)
952
done.add(revision_id)
953
return result
954
955
def _get_revision_vf(self):
956
""":return: a versioned file containing the revisions."""
957
vf = self._revision_store.get_revision_file(self.get_transaction())
958
return vf
959
960
@needs_write_lock
961
def reconcile(self, other=None, thorough=False):
962
"""Reconcile this repository."""
963
from bzrlib.reconcile import KnitReconciler
964
reconciler = KnitReconciler(self, thorough=thorough)
965
reconciler.reconcile()
966
return reconciler
967
968
def revision_parents(self, revision_id):
969
return self._get_revision_vf().get_parents(revision_id)
970
971
972
class RepositoryFormat(object):
973
"""A repository format.
974
975
Formats provide three things:
976
* An initialization routine to construct repository data on disk.
977
* a format string which is used when the BzrDir supports versioned
978
children.
979
* an open routine which returns a Repository instance.
980
981
Formats are placed in an dict by their format string for reference
982
during opening. These should be subclasses of RepositoryFormat
983
for consistency.
984
985
Once a format is deprecated, just deprecate the initialize and open
986
methods on the format class. Do not deprecate the object, as the
987
object will be created every system load.
988
989
Common instance attributes:
990
_matchingbzrdir - the bzrdir format that the repository format was
991
originally written to work with. This can be used if manually
992
constructing a bzrdir and repository, or more commonly for test suite
993
parameterisation.
994
"""
995
996
_default_format = None
997
"""The default format used for new repositories."""
998
999
_formats = {}
1000
"""The known formats."""
1001
1002
@classmethod
1003
def find_format(klass, a_bzrdir):
1004
"""Return the format for the repository object in a_bzrdir."""
1005
try:
1006
transport = a_bzrdir.get_repository_transport(None)
1007
format_string = transport.get("format").read()
1008
return klass._formats[format_string]
1009
except errors.NoSuchFile:
1010
raise errors.NoRepositoryPresent(a_bzrdir)
1011
except KeyError:
1012
raise errors.UnknownFormatError(format=format_string)
1013
1014
def _get_control_store(self, repo_transport, control_files):
1015
"""Return the control store for this repository."""
1016
raise NotImplementedError(self._get_control_store)
1017
1018
@classmethod
1019
def get_default_format(klass):
1020
"""Return the current default format."""
1021
return klass._default_format
1022
1023
def get_format_string(self):
1024
"""Return the ASCII format string that identifies this format.
1025
1026
Note that in pre format ?? repositories the format string is
1027
not permitted nor written to disk.
1028
"""
1029
raise NotImplementedError(self.get_format_string)
1030
1031
def get_format_description(self):
1032
"""Return the short description for this format."""
1033
raise NotImplementedError(self.get_format_description)
1034
1035
def _get_revision_store(self, repo_transport, control_files):
1036
"""Return the revision store object for this a_bzrdir."""
1037
raise NotImplementedError(self._get_revision_store)
1038
1039
def _get_text_rev_store(self,
1040
transport,
1041
control_files,
1042
name,
1043
compressed=True,
1044
prefixed=False,
1045
serializer=None):
1046
"""Common logic for getting a revision store for a repository.
1047
1048
see self._get_revision_store for the subclass-overridable method to
1049
get the store for a repository.
1050
"""
1051
from bzrlib.store.revision.text import TextRevisionStore
1052
dir_mode = control_files._dir_mode
1053
file_mode = control_files._file_mode
1054
text_store =TextStore(transport.clone(name),
1055
prefixed=prefixed,
1056
compressed=compressed,
1057
dir_mode=dir_mode,
1058
file_mode=file_mode)
1059
_revision_store = TextRevisionStore(text_store, serializer)
1060
return _revision_store
1061
1062
def _get_versioned_file_store(self,
1063
name,
1064
transport,
1065
control_files,
1066
prefixed=True,
1067
versionedfile_class=WeaveFile,
1068
escaped=False):
1069
weave_transport = control_files._transport.clone(name)
1070
dir_mode = control_files._dir_mode
1071
file_mode = control_files._file_mode
1072
return VersionedFileStore(weave_transport, prefixed=prefixed,
1073
dir_mode=dir_mode,
1074
file_mode=file_mode,
1075
versionedfile_class=versionedfile_class,
1076
escaped=escaped)
1077
1078
def initialize(self, a_bzrdir, shared=False):
1079
"""Initialize a repository of this format in a_bzrdir.
1080
1081
:param a_bzrdir: The bzrdir to put the new repository in it.
1082
:param shared: The repository should be initialized as a sharable one.
1083
1084
This may raise UninitializableFormat if shared repository are not
1085
compatible the a_bzrdir.
1086
"""
1087
1088
def is_supported(self):
1089
"""Is this format supported?
1090
1091
Supported formats must be initializable and openable.
1092
Unsupported formats may not support initialization or committing or
1093
some other features depending on the reason for not being supported.
1094
"""
1095
return True
1096
1097
def open(self, a_bzrdir, _found=False):
1098
"""Return an instance of this format for the bzrdir a_bzrdir.
1099
1100
_found is a private parameter, do not use it.
1101
"""
1102
raise NotImplementedError(self.open)
1103
1104
@classmethod
1105
def register_format(klass, format):
1106
klass._formats[format.get_format_string()] = format
1107
1108
@classmethod
1109
def set_default_format(klass, format):
1110
klass._default_format = format
1111
1112
@classmethod
1113
def unregister_format(klass, format):
1114
assert klass._formats[format.get_format_string()] is format
1115
del klass._formats[format.get_format_string()]
1116
1117
1118
class PreSplitOutRepositoryFormat(RepositoryFormat):
1119
"""Base class for the pre split out repository formats."""
1120
1121
def initialize(self, a_bzrdir, shared=False, _internal=False):
1122
"""Create a weave repository.
1123
1124
TODO: when creating split out bzr branch formats, move this to a common
1125
base for Format5, Format6. or something like that.
1126
"""
1127
from bzrlib.weavefile import write_weave_v5
1128
from bzrlib.weave import Weave
1129
1130
if shared:
1131
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1132
1133
if not _internal:
1134
# always initialized when the bzrdir is.
1135
return self.open(a_bzrdir, _found=True)
1136
1137
# Create an empty weave
1138
sio = StringIO()
1139
write_weave_v5(Weave(), sio)
1140
empty_weave = sio.getvalue()
1141
1142
mutter('creating repository in %s.', a_bzrdir.transport.base)
1143
dirs = ['revision-store', 'weaves']
1144
files = [('inventory.weave', StringIO(empty_weave)),
1145
]
1146
1147
# FIXME: RBC 20060125 don't peek under the covers
1148
# NB: no need to escape relative paths that are url safe.
1149
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1150
TransportLock)
1151
control_files.create_lock()
1152
control_files.lock_write()
1153
control_files._transport.mkdir_multi(dirs,
1154
mode=control_files._dir_mode)
1155
try:
1156
for file, content in files:
1157
control_files.put(file, content)
1158
finally:
1159
control_files.unlock()
1160
return self.open(a_bzrdir, _found=True)
1161
1162
def _get_control_store(self, repo_transport, control_files):
1163
"""Return the control store for this repository."""
1164
return self._get_versioned_file_store('',
1165
repo_transport,
1166
control_files,
1167
prefixed=False)
1168
1169
def _get_text_store(self, transport, control_files):
1170
"""Get a store for file texts for this format."""
1171
raise NotImplementedError(self._get_text_store)
1172
1173
def open(self, a_bzrdir, _found=False):
1174
"""See RepositoryFormat.open()."""
1175
if not _found:
1176
# we are being called directly and must probe.
1177
raise NotImplementedError
1178
1179
repo_transport = a_bzrdir.get_repository_transport(None)
1180
control_files = a_bzrdir._control_files
1181
text_store = self._get_text_store(repo_transport, control_files)
1182
control_store = self._get_control_store(repo_transport, control_files)
1183
_revision_store = self._get_revision_store(repo_transport, control_files)
1184
return AllInOneRepository(_format=self,
1185
a_bzrdir=a_bzrdir,
1186
_revision_store=_revision_store,
1187
control_store=control_store,
1188
text_store=text_store)
1189
1190
1191
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1192
"""Bzr repository format 4.
1193
1194
This repository format has:
1195
- flat stores
1196
- TextStores for texts, inventories,revisions.
1197
1198
This format is deprecated: it indexes texts using a text id which is
1199
removed in format 5; initialization and write support for this format
1200
has been removed.
1201
"""
1202
1203
def __init__(self):
1204
super(RepositoryFormat4, self).__init__()
1205
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1206
1207
def get_format_description(self):
1208
"""See RepositoryFormat.get_format_description()."""
1209
return "Repository format 4"
1210
1211
def initialize(self, url, shared=False, _internal=False):
1212
"""Format 4 branches cannot be created."""
1213
raise errors.UninitializableFormat(self)
1214
1215
def is_supported(self):
1216
"""Format 4 is not supported.
1217
1218
It is not supported because the model changed from 4 to 5 and the
1219
conversion logic is expensive - so doing it on the fly was not
1220
feasible.
1221
"""
1222
return False
1223
1224
def _get_control_store(self, repo_transport, control_files):
1225
"""Format 4 repositories have no formal control store at this point.
1226
1227
This will cause any control-file-needing apis to fail - this is desired.
1228
"""
1229
return None
1230
1231
def _get_revision_store(self, repo_transport, control_files):
1232
"""See RepositoryFormat._get_revision_store()."""
1233
from bzrlib.xml4 import serializer_v4
1234
return self._get_text_rev_store(repo_transport,
1235
control_files,
1236
'revision-store',
1237
serializer=serializer_v4)
1238
1239
def _get_text_store(self, transport, control_files):
1240
"""See RepositoryFormat._get_text_store()."""
1241
1242
1243
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1244
"""Bzr control format 5.
1245
1246
This repository format has:
1247
- weaves for file texts and inventory
1248
- flat stores
1249
- TextStores for revisions and signatures.
1250
"""
1251
1252
def __init__(self):
1253
super(RepositoryFormat5, self).__init__()
1254
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1255
1256
def get_format_description(self):
1257
"""See RepositoryFormat.get_format_description()."""
1258
return "Weave repository format 5"
1259
1260
def _get_revision_store(self, repo_transport, control_files):
1261
"""See RepositoryFormat._get_revision_store()."""
1262
"""Return the revision store object for this a_bzrdir."""
1263
return self._get_text_rev_store(repo_transport,
1264
control_files,
1265
'revision-store',
1266
compressed=False)
1267
1268
def _get_text_store(self, transport, control_files):
1269
"""See RepositoryFormat._get_text_store()."""
1270
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1271
1272
1273
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1274
"""Bzr control format 6.
1275
1276
This repository format has:
1277
- weaves for file texts and inventory
1278
- hash subdirectory based stores.
1279
- TextStores for revisions and signatures.
1280
"""
1281
1282
def __init__(self):
1283
super(RepositoryFormat6, self).__init__()
1284
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1285
1286
def get_format_description(self):
1287
"""See RepositoryFormat.get_format_description()."""
1288
return "Weave repository format 6"
1289
1290
def _get_revision_store(self, repo_transport, control_files):
1291
"""See RepositoryFormat._get_revision_store()."""
1292
return self._get_text_rev_store(repo_transport,
1293
control_files,
1294
'revision-store',
1295
compressed=False,
1296
prefixed=True)
1297
1298
def _get_text_store(self, transport, control_files):
1299
"""See RepositoryFormat._get_text_store()."""
1300
return self._get_versioned_file_store('weaves', transport, control_files)
1301
1302
1303
class MetaDirRepositoryFormat(RepositoryFormat):
1304
"""Common base class for the new repositories using the metadir layout."""
1305
1306
def __init__(self):
1307
super(MetaDirRepositoryFormat, self).__init__()
1308
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1309
1310
def _create_control_files(self, a_bzrdir):
1311
"""Create the required files and the initial control_files object."""
1312
# FIXME: RBC 20060125 don't peek under the covers
1313
# NB: no need to escape relative paths that are url safe.
1314
repository_transport = a_bzrdir.get_repository_transport(self)
1315
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1316
control_files.create_lock()
1317
return control_files
1318
1319
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1320
"""Upload the initial blank content."""
1321
control_files = self._create_control_files(a_bzrdir)
1322
control_files.lock_write()
1323
try:
1324
control_files._transport.mkdir_multi(dirs,
1325
mode=control_files._dir_mode)
1326
for file, content in files:
1327
control_files.put(file, content)
1328
for file, content in utf8_files:
1329
control_files.put_utf8(file, content)
1330
if shared == True:
1331
control_files.put_utf8('shared-storage', '')
1332
finally:
1333
control_files.unlock()
1334
1335
1336
class RepositoryFormat7(MetaDirRepositoryFormat):
1337
"""Bzr repository 7.
1338
1339
This repository format has:
1340
- weaves for file texts and inventory
1341
- hash subdirectory based stores.
1342
- TextStores for revisions and signatures.
1343
- a format marker of its own
1344
- an optional 'shared-storage' flag
1345
- an optional 'no-working-trees' flag
1346
"""
1347
1348
def _get_control_store(self, repo_transport, control_files):
1349
"""Return the control store for this repository."""
1350
return self._get_versioned_file_store('',
1351
repo_transport,
1352
control_files,
1353
prefixed=False)
1354
1355
def get_format_string(self):
1356
"""See RepositoryFormat.get_format_string()."""
1357
return "Bazaar-NG Repository format 7"
1358
1359
def get_format_description(self):
1360
"""See RepositoryFormat.get_format_description()."""
1361
return "Weave repository format 7"
1362
1363
def _get_revision_store(self, repo_transport, control_files):
1364
"""See RepositoryFormat._get_revision_store()."""
1365
return self._get_text_rev_store(repo_transport,
1366
control_files,
1367
'revision-store',
1368
compressed=False,
1369
prefixed=True,
1370
)
1371
1372
def _get_text_store(self, transport, control_files):
1373
"""See RepositoryFormat._get_text_store()."""
1374
return self._get_versioned_file_store('weaves',
1375
transport,
1376
control_files)
1377
1378
def initialize(self, a_bzrdir, shared=False):
1379
"""Create a weave repository.
1380
1381
:param shared: If true the repository will be initialized as a shared
1382
repository.
1383
"""
1384
from bzrlib.weavefile import write_weave_v5
1385
from bzrlib.weave import Weave
1386
1387
# Create an empty weave
1388
sio = StringIO()
1389
write_weave_v5(Weave(), sio)
1390
empty_weave = sio.getvalue()
1391
1392
mutter('creating repository in %s.', a_bzrdir.transport.base)
1393
dirs = ['revision-store', 'weaves']
1394
files = [('inventory.weave', StringIO(empty_weave)),
1395
]
1396
utf8_files = [('format', self.get_format_string())]
1397
1398
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1399
return self.open(a_bzrdir=a_bzrdir, _found=True)
1400
1401
def open(self, a_bzrdir, _found=False, _override_transport=None):
1402
"""See RepositoryFormat.open().
1403
1404
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1405
repository at a slightly different url
1406
than normal. I.e. during 'upgrade'.
1407
"""
1408
if not _found:
1409
format = RepositoryFormat.find_format(a_bzrdir)
1410
assert format.__class__ == self.__class__
1411
if _override_transport is not None:
1412
repo_transport = _override_transport
1413
else:
1414
repo_transport = a_bzrdir.get_repository_transport(None)
1415
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1416
text_store = self._get_text_store(repo_transport, control_files)
1417
control_store = self._get_control_store(repo_transport, control_files)
1418
_revision_store = self._get_revision_store(repo_transport, control_files)
1419
return MetaDirRepository(_format=self,
1420
a_bzrdir=a_bzrdir,
1421
control_files=control_files,
1422
_revision_store=_revision_store,
1423
control_store=control_store,
1424
text_store=text_store)
1425
1426
1427
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1428
"""Bzr repository knit format 1.
1429
1430
This repository format has:
1431
- knits for file texts and inventory
1432
- hash subdirectory based stores.
1433
- knits for revisions and signatures
1434
- TextStores for revisions and signatures.
1435
- a format marker of its own
1436
- an optional 'shared-storage' flag
1437
- an optional 'no-working-trees' flag
1438
- a LockDir lock
1439
1440
This format was introduced in bzr 0.8.
1441
"""
1442
1443
def _get_control_store(self, repo_transport, control_files):
1444
"""Return the control store for this repository."""
1445
return VersionedFileStore(
1446
repo_transport,
1447
prefixed=False,
1448
file_mode=control_files._file_mode,
1449
versionedfile_class=KnitVersionedFile,
1450
versionedfile_kwargs={'factory':KnitPlainFactory()},
1451
)
1452
1453
def get_format_string(self):
1454
"""See RepositoryFormat.get_format_string()."""
1455
return "Bazaar-NG Knit Repository Format 1"
1456
1457
def get_format_description(self):
1458
"""See RepositoryFormat.get_format_description()."""
1459
return "Knit repository format 1"
1460
1461
def _get_revision_store(self, repo_transport, control_files):
1462
"""See RepositoryFormat._get_revision_store()."""
1463
from bzrlib.store.revision.knit import KnitRevisionStore
1464
versioned_file_store = VersionedFileStore(
1465
repo_transport,
1466
file_mode=control_files._file_mode,
1467
prefixed=False,
1468
precious=True,
1469
versionedfile_class=KnitVersionedFile,
1470
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1471
escaped=True,
1472
)
1473
return KnitRevisionStore(versioned_file_store)
1474
1475
def _get_text_store(self, transport, control_files):
1476
"""See RepositoryFormat._get_text_store()."""
1477
return self._get_versioned_file_store('knits',
1478
transport,
1479
control_files,
1480
versionedfile_class=KnitVersionedFile,
1481
escaped=True)
1482
1483
def initialize(self, a_bzrdir, shared=False):
1484
"""Create a knit format 1 repository.
1485
1486
:param a_bzrdir: bzrdir to contain the new repository; must already
1487
be initialized.
1488
:param shared: If true the repository will be initialized as a shared
1489
repository.
1490
"""
1491
mutter('creating repository in %s.', a_bzrdir.transport.base)
1492
dirs = ['revision-store', 'knits']
1493
files = []
1494
utf8_files = [('format', self.get_format_string())]
1495
1496
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1497
repo_transport = a_bzrdir.get_repository_transport(None)
1498
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1499
control_store = self._get_control_store(repo_transport, control_files)
1500
transaction = transactions.WriteTransaction()
1501
# trigger a write of the inventory store.
1502
control_store.get_weave_or_empty('inventory', transaction)
1503
_revision_store = self._get_revision_store(repo_transport, control_files)
1504
_revision_store.has_revision_id('A', transaction)
1505
_revision_store.get_signature_file(transaction)
1506
return self.open(a_bzrdir=a_bzrdir, _found=True)
1507
1508
def open(self, a_bzrdir, _found=False, _override_transport=None):
1509
"""See RepositoryFormat.open().
1510
1511
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1512
repository at a slightly different url
1513
than normal. I.e. during 'upgrade'.
1514
"""
1515
if not _found:
1516
format = RepositoryFormat.find_format(a_bzrdir)
1517
assert format.__class__ == self.__class__
1518
if _override_transport is not None:
1519
repo_transport = _override_transport
1520
else:
1521
repo_transport = a_bzrdir.get_repository_transport(None)
1522
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1523
text_store = self._get_text_store(repo_transport, control_files)
1524
control_store = self._get_control_store(repo_transport, control_files)
1525
_revision_store = self._get_revision_store(repo_transport, control_files)
1526
return KnitRepository(_format=self,
1527
a_bzrdir=a_bzrdir,
1528
control_files=control_files,
1529
_revision_store=_revision_store,
1530
control_store=control_store,
1531
text_store=text_store)
1532
1533
1534
# formats which have no format string are not discoverable
1535
# and not independently creatable, so are not registered.
1536
RepositoryFormat.register_format(RepositoryFormat7())
1537
_default_format = RepositoryFormatKnit1()
1538
RepositoryFormat.register_format(_default_format)
1539
RepositoryFormat.set_default_format(_default_format)
1540
_legacy_formats = [RepositoryFormat4(),
1541
RepositoryFormat5(),
1542
RepositoryFormat6()]
1543
1544
1545
class InterRepository(InterObject):
1546
"""This class represents operations taking place between two repositories.
1547
1548
Its instances have methods like copy_content and fetch, and contain
1549
references to the source and target repositories these operations can be
1550
carried out on.
1551
1552
Often we will provide convenience methods on 'repository' which carry out
1553
operations with another repository - they will always forward to
1554
InterRepository.get(other).method_name(parameters).
1555
"""
1556
1557
_optimisers = set()
1558
"""The available optimised InterRepository types."""
1559
1560
@needs_write_lock
1561
def copy_content(self, revision_id=None, basis=None):
1562
"""Make a complete copy of the content in self into destination.
1563
1564
This is a destructive operation! Do not use it on existing
1565
repositories.
1566
1567
:param revision_id: Only copy the content needed to construct
1568
revision_id and its parents.
1569
:param basis: Copy the needed data preferentially from basis.
1570
"""
1571
try:
1572
self.target.set_make_working_trees(self.source.make_working_trees())
1573
except NotImplementedError:
1574
pass
1575
# grab the basis available data
1576
if basis is not None:
1577
self.target.fetch(basis, revision_id=revision_id)
1578
# but don't bother fetching if we have the needed data now.
1579
if (revision_id not in (None, NULL_REVISION) and
1580
self.target.has_revision(revision_id)):
1581
return
1582
self.target.fetch(self.source, revision_id=revision_id)
1583
1584
def _double_lock(self, lock_source, lock_target):
1585
"""Take out too locks, rolling back the first if the second throws."""
1586
lock_source()
1587
try:
1588
lock_target()
1589
except Exception:
1590
# we want to ensure that we don't leave source locked by mistake.
1591
# and any error on target should not confuse source.
1592
self.source.unlock()
1593
raise
1594
1595
@needs_write_lock
1596
def fetch(self, revision_id=None, pb=None):
1597
"""Fetch the content required to construct revision_id.
1598
1599
The content is copied from source to target.
1600
1601
:param revision_id: if None all content is copied, if NULL_REVISION no
1602
content is copied.
1603
:param pb: optional progress bar to use for progress reports. If not
1604
provided a default one will be created.
1605
1606
Returns the copied revision count and the failed revisions in a tuple:
1607
(copied, failures).
1608
"""
1609
from bzrlib.fetch import GenericRepoFetcher
1610
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1611
self.source, self.source._format, self.target, self.target._format)
1612
f = GenericRepoFetcher(to_repository=self.target,
1613
from_repository=self.source,
1614
last_revision=revision_id,
1615
pb=pb)
1616
return f.count_copied, f.failed_revisions
1617
1618
def lock_read(self):
1619
"""Take out a logical read lock.
1620
1621
This will lock the source branch and the target branch. The source gets
1622
a read lock and the target a read lock.
1623
"""
1624
self._double_lock(self.source.lock_read, self.target.lock_read)
1625
1626
def lock_write(self):
1627
"""Take out a logical write lock.
1628
1629
This will lock the source branch and the target branch. The source gets
1630
a read lock and the target a write lock.
1631
"""
1632
self._double_lock(self.source.lock_read, self.target.lock_write)
1633
1634
@needs_read_lock
1635
def missing_revision_ids(self, revision_id=None):
1636
"""Return the revision ids that source has that target does not.
1637
1638
These are returned in topological order.
1639
1640
:param revision_id: only return revision ids included by this
1641
revision_id.
1642
"""
1643
# generic, possibly worst case, slow code path.
1644
target_ids = set(self.target.all_revision_ids())
1645
if revision_id is not None:
1646
source_ids = self.source.get_ancestry(revision_id)
1647
assert source_ids[0] == None
1648
source_ids.pop(0)
1649
else:
1650
source_ids = self.source.all_revision_ids()
1651
result_set = set(source_ids).difference(target_ids)
1652
# this may look like a no-op: its not. It preserves the ordering
1653
# other_ids had while only returning the members from other_ids
1654
# that we've decided we need.
1655
return [rev_id for rev_id in source_ids if rev_id in result_set]
1656
1657
def unlock(self):
1658
"""Release the locks on source and target."""
1659
try:
1660
self.target.unlock()
1661
finally:
1662
self.source.unlock()
1663
1664
1665
class InterWeaveRepo(InterRepository):
1666
"""Optimised code paths between Weave based repositories."""
1667
1668
_matching_repo_format = RepositoryFormat7()
1669
"""Repository format for testing with."""
1670
1671
@staticmethod
1672
def is_compatible(source, target):
1673
"""Be compatible with known Weave formats.
1674
1675
We don't test for the stores being of specific types because that
1676
could lead to confusing results, and there is no need to be
1677
overly general.
1678
"""
1679
try:
1680
return (isinstance(source._format, (RepositoryFormat5,
1681
RepositoryFormat6,
1682
RepositoryFormat7)) and
1683
isinstance(target._format, (RepositoryFormat5,
1684
RepositoryFormat6,
1685
RepositoryFormat7)))
1686
except AttributeError:
1687
return False
1688
1689
@needs_write_lock
1690
def copy_content(self, revision_id=None, basis=None):
1691
"""See InterRepository.copy_content()."""
1692
# weave specific optimised path:
1693
if basis is not None:
1694
# copy the basis in, then fetch remaining data.
1695
basis.copy_content_into(self.target, revision_id)
1696
# the basis copy_content_into could miss-set this.
1697
try:
1698
self.target.set_make_working_trees(self.source.make_working_trees())
1699
except NotImplementedError:
1700
pass
1701
self.target.fetch(self.source, revision_id=revision_id)
1702
else:
1703
try:
1704
self.target.set_make_working_trees(self.source.make_working_trees())
1705
except NotImplementedError:
1706
pass
1707
# FIXME do not peek!
1708
if self.source.control_files._transport.listable():
1709
pb = ui.ui_factory.nested_progress_bar()
1710
try:
1711
self.target.weave_store.copy_all_ids(
1712
self.source.weave_store,
1713
pb=pb,
1714
from_transaction=self.source.get_transaction(),
1715
to_transaction=self.target.get_transaction())
1716
pb.update('copying inventory', 0, 1)
1717
self.target.control_weaves.copy_multi(
1718
self.source.control_weaves, ['inventory'],
1719
from_transaction=self.source.get_transaction(),
1720
to_transaction=self.target.get_transaction())
1721
self.target._revision_store.text_store.copy_all_ids(
1722
self.source._revision_store.text_store,
1723
pb=pb)
1724
finally:
1725
pb.finished()
1726
else:
1727
self.target.fetch(self.source, revision_id=revision_id)
1728
1729
@needs_write_lock
1730
def fetch(self, revision_id=None, pb=None):
1731
"""See InterRepository.fetch()."""
1732
from bzrlib.fetch import GenericRepoFetcher
1733
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1734
self.source, self.source._format, self.target, self.target._format)
1735
f = GenericRepoFetcher(to_repository=self.target,
1736
from_repository=self.source,
1737
last_revision=revision_id,
1738
pb=pb)
1739
return f.count_copied, f.failed_revisions
1740
1741
@needs_read_lock
1742
def missing_revision_ids(self, revision_id=None):
1743
"""See InterRepository.missing_revision_ids()."""
1744
# we want all revisions to satisfy revision_id in source.
1745
# but we don't want to stat every file here and there.
1746
# we want then, all revisions other needs to satisfy revision_id
1747
# checked, but not those that we have locally.
1748
# so the first thing is to get a subset of the revisions to
1749
# satisfy revision_id in source, and then eliminate those that
1750
# we do already have.
1751
# this is slow on high latency connection to self, but as as this
1752
# disk format scales terribly for push anyway due to rewriting
1753
# inventory.weave, this is considered acceptable.
1754
# - RBC 20060209
1755
if revision_id is not None:
1756
source_ids = self.source.get_ancestry(revision_id)
1757
assert source_ids[0] == None
1758
source_ids.pop(0)
1759
else:
1760
source_ids = self.source._all_possible_ids()
1761
source_ids_set = set(source_ids)
1762
# source_ids is the worst possible case we may need to pull.
1763
# now we want to filter source_ids against what we actually
1764
# have in target, but don't try to check for existence where we know
1765
# we do not have a revision as that would be pointless.
1766
target_ids = set(self.target._all_possible_ids())
1767
possibly_present_revisions = target_ids.intersection(source_ids_set)
1768
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1769
required_revisions = source_ids_set.difference(actually_present_revisions)
1770
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1771
if revision_id is not None:
1772
# we used get_ancestry to determine source_ids then we are assured all
1773
# revisions referenced are present as they are installed in topological order.
1774
# and the tip revision was validated by get_ancestry.
1775
return required_topo_revisions
1776
else:
1777
# if we just grabbed the possibly available ids, then
1778
# we only have an estimate of whats available and need to validate
1779
# that against the revision records.
1780
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1781
1782
1783
class InterKnitRepo(InterRepository):
1784
"""Optimised code paths between Knit based repositories."""
1785
1786
_matching_repo_format = RepositoryFormatKnit1()
1787
"""Repository format for testing with."""
1788
1789
@staticmethod
1790
def is_compatible(source, target):
1791
"""Be compatible with known Knit formats.
1792
1793
We don't test for the stores being of specific types because that
1794
could lead to confusing results, and there is no need to be
1795
overly general.
1796
"""
1797
try:
1798
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1799
isinstance(target._format, (RepositoryFormatKnit1)))
1800
except AttributeError:
1801
return False
1802
1803
@needs_write_lock
1804
def fetch(self, revision_id=None, pb=None):
1805
"""See InterRepository.fetch()."""
1806
from bzrlib.fetch import KnitRepoFetcher
1807
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1808
self.source, self.source._format, self.target, self.target._format)
1809
f = KnitRepoFetcher(to_repository=self.target,
1810
from_repository=self.source,
1811
last_revision=revision_id,
1812
pb=pb)
1813
return f.count_copied, f.failed_revisions
1814
1815
@needs_read_lock
1816
def missing_revision_ids(self, revision_id=None):
1817
"""See InterRepository.missing_revision_ids()."""
1818
if revision_id is not None:
1819
source_ids = self.source.get_ancestry(revision_id)
1820
assert source_ids[0] == None
1821
source_ids.pop(0)
1822
else:
1823
source_ids = self.source._all_possible_ids()
1824
source_ids_set = set(source_ids)
1825
# source_ids is the worst possible case we may need to pull.
1826
# now we want to filter source_ids against what we actually
1827
# have in target, but don't try to check for existence where we know
1828
# we do not have a revision as that would be pointless.
1829
target_ids = set(self.target._all_possible_ids())
1830
possibly_present_revisions = target_ids.intersection(source_ids_set)
1831
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1832
required_revisions = source_ids_set.difference(actually_present_revisions)
1833
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1834
if revision_id is not None:
1835
# we used get_ancestry to determine source_ids then we are assured all
1836
# revisions referenced are present as they are installed in topological order.
1837
# and the tip revision was validated by get_ancestry.
1838
return required_topo_revisions
1839
else:
1840
# if we just grabbed the possibly available ids, then
1841
# we only have an estimate of whats available and need to validate
1842
# that against the revision records.
1843
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1844
1845
InterRepository.register_optimiser(InterWeaveRepo)
1846
InterRepository.register_optimiser(InterKnitRepo)
1847
1848
1849
class RepositoryTestProviderAdapter(object):
1850
"""A tool to generate a suite testing multiple repository formats at once.
1851
1852
This is done by copying the test once for each transport and injecting
1853
the transport_server, transport_readonly_server, and bzrdir_format and
1854
repository_format classes into each copy. Each copy is also given a new id()
1855
to make it easy to identify.
1856
"""
1857
1858
def __init__(self, transport_server, transport_readonly_server, formats):
1859
self._transport_server = transport_server
1860
self._transport_readonly_server = transport_readonly_server
1861
self._formats = formats
1862
1863
def adapt(self, test):
1864
result = TestSuite()
1865
for repository_format, bzrdir_format in self._formats:
1866
new_test = deepcopy(test)
1867
new_test.transport_server = self._transport_server
1868
new_test.transport_readonly_server = self._transport_readonly_server
1869
new_test.bzrdir_format = bzrdir_format
1870
new_test.repository_format = repository_format
1871
def make_new_test_id():
1872
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1873
return lambda: new_id
1874
new_test.id = make_new_test_id()
1875
result.addTest(new_test)
1876
return result
1877
1878
1879
class InterRepositoryTestProviderAdapter(object):
1880
"""A tool to generate a suite testing multiple inter repository formats.
1881
1882
This is done by copying the test once for each interrepo provider and injecting
1883
the transport_server, transport_readonly_server, repository_format and
1884
repository_to_format classes into each copy.
1885
Each copy is also given a new id() to make it easy to identify.
1886
"""
1887
1888
def __init__(self, transport_server, transport_readonly_server, formats):
1889
self._transport_server = transport_server
1890
self._transport_readonly_server = transport_readonly_server
1891
self._formats = formats
1892
1893
def adapt(self, test):
1894
result = TestSuite()
1895
for interrepo_class, repository_format, repository_format_to in self._formats:
1896
new_test = deepcopy(test)
1897
new_test.transport_server = self._transport_server
1898
new_test.transport_readonly_server = self._transport_readonly_server
1899
new_test.interrepo_class = interrepo_class
1900
new_test.repository_format = repository_format
1901
new_test.repository_format_to = repository_format_to
1902
def make_new_test_id():
1903
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1904
return lambda: new_id
1905
new_test.id = make_new_test_id()
1906
result.addTest(new_test)
1907
return result
1908
1909
@staticmethod
1910
def default_test_list():
1911
"""Generate the default list of interrepo permutations to test."""
1912
result = []
1913
# test the default InterRepository between format 6 and the current
1914
# default format.
1915
# XXX: robertc 20060220 reinstate this when there are two supported
1916
# formats which do not have an optimal code path between them.
1917
result.append((InterRepository,
1918
RepositoryFormat6(),
1919
RepositoryFormatKnit1()))
1920
for optimiser in InterRepository._optimisers:
1921
result.append((optimiser,
1922
optimiser._matching_repo_format,
1923
optimiser._matching_repo_format
1924
))
1925
# if there are specific combinations we want to use, we can add them
1926
# here.
1927
return result
1928
1929
1930
class CopyConverter(object):
1931
"""A repository conversion tool which just performs a copy of the content.
1932
1933
This is slow but quite reliable.
1934
"""
1935
1936
def __init__(self, target_format):
1937
"""Create a CopyConverter.
1938
1939
:param target_format: The format the resulting repository should be.
1940
"""
1941
self.target_format = target_format
1942
1943
def convert(self, repo, pb):
1944
"""Perform the conversion of to_convert, giving feedback via pb.
1945
1946
:param to_convert: The disk object to convert.
1947
:param pb: a progress bar to use for progress information.
1948
"""
1949
self.pb = pb
1950
self.count = 0
1951
self.total = 4
1952
# this is only useful with metadir layouts - separated repo content.
1953
# trigger an assertion if not such
1954
repo._format.get_format_string()
1955
self.repo_dir = repo.bzrdir
1956
self.step('Moving repository to repository.backup')
1957
self.repo_dir.transport.move('repository', 'repository.backup')
1958
backup_transport = self.repo_dir.transport.clone('repository.backup')
1959
self.source_repo = repo._format.open(self.repo_dir,
1960
_found=True,
1961
_override_transport=backup_transport)
1962
self.step('Creating new repository')
1963
converted = self.target_format.initialize(self.repo_dir,
1964
self.source_repo.is_shared())
1965
converted.lock_write()
1966
try:
1967
self.step('Copying content into repository.')
1968
self.source_repo.copy_content_into(converted)
1969
finally:
1970
converted.unlock()
1971
self.step('Deleting old repository content.')
1972
self.repo_dir.transport.delete_tree('repository.backup')
1973
self.pb.note('repository converted')
1974
1975
def step(self, message):
1976
"""Update the pb by a step."""
1977
self.count +=1
1978
self.pb.update(message, self.count, self.total)
1979
1980
1981
class CommitBuilder(object):
1982
"""Provides an interface to build up a commit.
1983
1984
This allows describing a tree to be committed without needing to
1985
know the internals of the format of the repository.
1986
"""
1987
def __init__(self, repository, parents, config, timestamp=None,
1988
timezone=None, committer=None, revprops=None,
1989
revision_id=None):
1990
"""Initiate a CommitBuilder.
1991
1992
:param repository: Repository to commit to.
1993
:param parents: Revision ids of the parents of the new revision.
1994
:param config: Configuration to use.
1995
:param timestamp: Optional timestamp recorded for commit.
1996
:param timezone: Optional timezone for timestamp.
1997
:param committer: Optional committer to set for commit.
1998
:param revprops: Optional dictionary of revision properties.
1999
:param revision_id: Optional revision id.
2000
"""
2001
self._config = config
2002
2003
if committer is None:
2004
self._committer = self._config.username()
2005
else:
2006
assert isinstance(committer, basestring), type(committer)
2007
self._committer = committer
2008
2009
self.new_inventory = Inventory()
2010
self._new_revision_id = revision_id
2011
self.parents = parents
2012
self.repository = repository
2013
2014
self._revprops = {}
2015
if revprops is not None:
2016
self._revprops.update(revprops)
2017
2018
if timestamp is None:
2019
timestamp = time.time()
2020
# Restrict resolution to 1ms
2021
self._timestamp = round(timestamp, 3)
2022
2023
if timezone is None:
2024
self._timezone = local_time_offset()
2025
else:
2026
self._timezone = int(timezone)
2027
2028
self._generate_revision_if_needed()
2029
2030
def commit(self, message):
2031
"""Make the actual commit.
2032
2033
:return: The revision id of the recorded revision.
2034
"""
2035
rev = Revision(timestamp=self._timestamp,
2036
timezone=self._timezone,
2037
committer=self._committer,
2038
message=message,
2039
inventory_sha1=self.inv_sha1,
2040
revision_id=self._new_revision_id,
2041
properties=self._revprops)
2042
rev.parent_ids = self.parents
2043
self.repository.add_revision(self._new_revision_id, rev,
2044
self.new_inventory, self._config)
2045
return self._new_revision_id
2046
2047
def finish_inventory(self):
2048
"""Tell the builder that the inventory is finished."""
2049
self.new_inventory.revision_id = self._new_revision_id
2050
self.inv_sha1 = self.repository.add_inventory(
2051
self._new_revision_id,
2052
self.new_inventory,
2053
self.parents
2054
)
2055
2056
def _gen_revision_id(self):
2057
"""Return new revision-id."""
2058
s = '%s-%s-' % (self._config.user_email(),
2059
compact_date(self._timestamp))
2060
s += hexlify(rand_bytes(8))
2061
return s
2062
2063
def _generate_revision_if_needed(self):
2064
"""Create a revision id if None was supplied.
2065
2066
If the repository can not support user-specified revision ids
2067
they should override this function and raise UnsupportedOperation
2068
if _new_revision_id is not None.
2069
2070
:raises: UnsupportedOperation
2071
"""
2072
if self._new_revision_id is None:
2073
self._new_revision_id = self._gen_revision_id()
2074
2075
def record_entry_contents(self, ie, parent_invs, path, tree):
2076
"""Record the content of ie from tree into the commit if needed.
2077
2078
:param ie: An inventory entry present in the commit.
2079
:param parent_invs: The inventories of the parent revisions of the
2080
commit.
2081
:param path: The path the entry is at in the tree.
2082
:param tree: The tree which contains this entry and should be used to
2083
obtain content.
2084
"""
2085
self.new_inventory.add(ie)
2086
2087
# ie.revision is always None if the InventoryEntry is considered
2088
# for committing. ie.snapshot will record the correct revision
2089
# which may be the sole parent if it is untouched.
2090
if ie.revision is not None:
2091
return
2092
previous_entries = ie.find_previous_heads(
2093
parent_invs,
2094
self.repository.weave_store,
2095
self.repository.get_transaction())
2096
# we are creating a new revision for ie in the history store
2097
# and inventory.
2098
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2099
2100
def modified_directory(self, file_id, file_parents):
2101
"""Record the presence of a symbolic link.
2102
2103
:param file_id: The file_id of the link to record.
2104
:param file_parents: The per-file parent revision ids.
2105
"""
2106
self._add_text_to_weave(file_id, [], file_parents.keys())
2107
2108
def modified_file_text(self, file_id, file_parents,
2109
get_content_byte_lines, text_sha1=None,
2110
text_size=None):
2111
"""Record the text of file file_id
2112
2113
:param file_id: The file_id of the file to record the text of.
2114
:param file_parents: The per-file parent revision ids.
2115
:param get_content_byte_lines: A callable which will return the byte
2116
lines for the file.
2117
:param text_sha1: Optional SHA1 of the file contents.
2118
:param text_size: Optional size of the file contents.
2119
"""
2120
mutter('storing text of file {%s} in revision {%s} into %r',
2121
file_id, self._new_revision_id, self.repository.weave_store)
2122
# special case to avoid diffing on renames or
2123
# reparenting
2124
if (len(file_parents) == 1
2125
and text_sha1 == file_parents.values()[0].text_sha1
2126
and text_size == file_parents.values()[0].text_size):
2127
previous_ie = file_parents.values()[0]
2128
versionedfile = self.repository.weave_store.get_weave(file_id,
2129
self.repository.get_transaction())
2130
versionedfile.clone_text(self._new_revision_id,
2131
previous_ie.revision, file_parents.keys())
2132
return text_sha1, text_size
2133
else:
2134
new_lines = get_content_byte_lines()
2135
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2136
# should return the SHA1 and size
2137
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2138
return osutils.sha_strings(new_lines), \
2139
sum(map(len, new_lines))
2140
2141
def modified_link(self, file_id, file_parents, link_target):
2142
"""Record the presence of a symbolic link.
2143
2144
:param file_id: The file_id of the link to record.
2145
:param file_parents: The per-file parent revision ids.
2146
:param link_target: Target location of this link.
2147
"""
2148
self._add_text_to_weave(file_id, [], file_parents.keys())
2149
2150
def _add_text_to_weave(self, file_id, new_lines, parents):
2151
versionedfile = self.repository.weave_store.get_weave_or_empty(
2152
file_id, self.repository.get_transaction())
2153
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2154
versionedfile.clear_cache()
2155
2156
2157
_unescape_map = {
2158
'apos':"'",
2159
'quot':'"',
2160
'amp':'&',
2161
'lt':'<',
2162
'gt':'>'
2163
}
2164
2165
2166
def _unescaper(match, _map=_unescape_map):
2167
return _map[match.group(1)]
2168
2169
2170
_unescape_re = None
2171
2172
2173
def _unescape_xml(data):
2174
"""Unescape predefined XML entities in a string of data."""
2175
global _unescape_re
2176
if _unescape_re is None:
2177
_unescape_re = re.compile('\&([^;]*);')
2178
return _unescape_re.sub(_unescaper, data)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1