1
# Copyright (C) 2006-2010 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
"""Reconcilers are able to fix some potential data errors in a branch."""
34
from bzrlib.trace import mutter
35
from bzrlib.tsort import topo_sort
36
from bzrlib.versionedfile import AdapterFactory, FulltextContentFactory
39
def reconcile(dir, canonicalize_chks=False):
40
"""Reconcile the data in dir.
42
Currently this is limited to a inventory 'reweave'.
44
This is a convenience method, for using a Reconciler object.
46
Directly using Reconciler is recommended for library users that
47
desire fine grained control or analysis of the found issues.
49
:param canonicalize_chks: Make sure CHKs are in canonical form.
51
reconciler = Reconciler(dir, canonicalize_chks=canonicalize_chks)
52
reconciler.reconcile()
55
class Reconciler(object):
56
"""Reconcilers are used to reconcile existing data."""
58
def __init__(self, dir, other=None, canonicalize_chks=False):
59
"""Create a Reconciler."""
61
self.canonicalize_chks = canonicalize_chks
64
"""Perform reconciliation.
66
After reconciliation the following attributes document found issues:
67
inconsistent_parents: The number of revisions in the repository whose
68
ancestry was being reported incorrectly.
69
garbage_inventories: The number of inventory objects without revisions
70
that were garbage collected.
71
fixed_branch_history: None if there was no branch, False if the branch
72
history was correct, True if the branch history
73
needed to be re-normalized.
75
self.pb = ui.ui_factory.nested_progress_bar()
82
"""Helper function for performing reconciliation."""
83
self._reconcile_branch()
84
self._reconcile_repository()
86
def _reconcile_branch(self):
88
self.branch = self.bzrdir.open_branch()
89
except errors.NotBranchError:
90
# Nothing to check here
91
self.fixed_branch_history = None
93
ui.ui_factory.note('Reconciling branch %s' % self.branch.base)
94
branch_reconciler = self.branch.reconcile(thorough=True)
95
self.fixed_branch_history = branch_reconciler.fixed_history
97
def _reconcile_repository(self):
98
self.repo = self.bzrdir.find_repository()
99
ui.ui_factory.note('Reconciling repository %s' %
101
self.pb.update("Reconciling repository", 0, 1)
102
if self.canonicalize_chks:
104
self.repo.reconcile_canonicalize_chks
105
except AttributeError:
106
raise errors.BzrError(
107
"%s cannot canonicalize CHKs." % (self.repo,))
108
repo_reconciler = self.repo.reconcile_canonicalize_chks()
110
repo_reconciler = self.repo.reconcile(thorough=True)
111
self.inconsistent_parents = repo_reconciler.inconsistent_parents
112
self.garbage_inventories = repo_reconciler.garbage_inventories
113
if repo_reconciler.aborted:
115
'Reconcile aborted: revision index has inconsistent parents.')
117
'Run "bzr check" for more details.')
119
ui.ui_factory.note('Reconciliation complete.')
122
class BranchReconciler(object):
123
"""Reconciler that works on a branch."""
125
def __init__(self, a_branch, thorough=False):
126
self.fixed_history = None
127
self.thorough = thorough
128
self.branch = a_branch
131
operation = cleanup.OperationWithCleanups(self._reconcile)
132
self.add_cleanup = operation.add_cleanup
133
operation.run_simple()
135
def _reconcile(self):
136
self.branch.lock_write()
137
self.add_cleanup(self.branch.unlock)
138
self.pb = ui.ui_factory.nested_progress_bar()
139
self.add_cleanup(self.pb.finished)
140
self._reconcile_steps()
142
def _reconcile_steps(self):
143
self._reconcile_revision_history()
145
def _reconcile_revision_history(self):
146
repo = self.branch.repository
147
last_revno, last_revision_id = self.branch.last_revision_info()
150
for revid in repo.iter_reverse_revision_history(
152
real_history.append(revid)
153
except errors.RevisionNotPresent:
154
pass # Hit a ghost left hand parent
155
real_history.reverse()
156
if last_revno != len(real_history):
157
self.fixed_history = True
158
# Technically for Branch5 formats, it is more efficient to use
159
# set_revision_history, as this will regenerate it again.
160
# Not really worth a whole BranchReconciler class just for this,
162
ui.ui_factory.note('Fixing last revision info %s => %s' % (
163
last_revno, len(real_history)))
164
self.branch.set_last_revision_info(len(real_history),
167
self.fixed_history = False
168
ui.ui_factory.note('revision_history ok.')
171
class RepoReconciler(object):
172
"""Reconciler that reconciles a repository.
174
The goal of repository reconciliation is to make any derived data
175
consistent with the core data committed by a user. This can involve
176
reindexing, or removing unreferenced data if that can interfere with
177
queries in a given repository.
179
Currently this consists of an inventory reweave with revision cross-checks.
182
def __init__(self, repo, other=None, thorough=False):
183
"""Construct a RepoReconciler.
185
:param thorough: perform a thorough check which may take longer but
186
will correct non-data loss issues such as incorrect
189
self.garbage_inventories = 0
190
self.inconsistent_parents = 0
193
self.thorough = thorough
196
"""Perform reconciliation.
198
After reconciliation the following attributes document found issues:
199
inconsistent_parents: The number of revisions in the repository whose
200
ancestry was being reported incorrectly.
201
garbage_inventories: The number of inventory objects without revisions
202
that were garbage collected.
204
operation = cleanup.OperationWithCleanups(self._reconcile)
205
self.add_cleanup = operation.add_cleanup
206
operation.run_simple()
208
def _reconcile(self):
209
self.repo.lock_write()
210
self.add_cleanup(self.repo.unlock)
211
self.pb = ui.ui_factory.nested_progress_bar()
212
self.add_cleanup(self.pb.finished)
213
self._reconcile_steps()
215
def _reconcile_steps(self):
216
"""Perform the steps to reconcile this repository."""
217
self._reweave_inventory()
219
def _reweave_inventory(self):
220
"""Regenerate the inventory weave for the repository from scratch.
222
This is a smart function: it will only do the reweave if doing it
223
will correct data issues. The self.thorough flag controls whether
224
only data-loss causing issues (!self.thorough) or all issues
225
(self.thorough) are treated as requiring the reweave.
227
transaction = self.repo.get_transaction()
228
self.pb.update('Reading inventory data')
229
self.inventory = self.repo.inventories
230
self.revisions = self.repo.revisions
231
# the total set of revisions to process
232
self.pending = set([key[-1] for key in self.revisions.keys()])
234
# mapping from revision_id to parents
236
# errors that we detect
237
self.inconsistent_parents = 0
238
# we need the revision id of each revision and its available parents list
239
self._setup_steps(len(self.pending))
240
for rev_id in self.pending:
241
# put a revision into the graph.
242
self._graph_revision(rev_id)
243
self._check_garbage_inventories()
244
# if there are no inconsistent_parents and
245
# (no garbage inventories or we are not doing a thorough check)
246
if (not self.inconsistent_parents and
247
(not self.garbage_inventories or not self.thorough)):
248
ui.ui_factory.note('Inventory ok.')
250
self.pb.update('Backing up inventory', 0, 0)
251
self.repo._backup_inventory()
252
ui.ui_factory.note('Backup inventory created.')
253
new_inventories = self.repo._temp_inventories()
255
# we have topological order of revisions and non ghost parents ready.
256
self._setup_steps(len(self._rev_graph))
257
revision_keys = [(rev_id,) for rev_id in topo_sort(self._rev_graph)]
258
stream = self._change_inv_parents(
259
self.inventory.get_record_stream(revision_keys, 'unordered', True),
260
self._new_inv_parents,
262
new_inventories.insert_record_stream(stream)
263
# if this worked, the set of new_inventories.keys should equal
265
if not (set(new_inventories.keys()) ==
266
set([(revid,) for revid in self.pending])):
267
raise AssertionError()
268
self.pb.update('Writing weave')
269
self.repo._activate_new_inventory()
270
self.inventory = None
271
ui.ui_factory.note('Inventory regenerated.')
273
def _new_inv_parents(self, revision_key):
274
"""Lookup ghost-filtered parents for revision_key."""
275
# Use the filtered ghostless parents list:
276
return tuple([(revid,) for revid in self._rev_graph[revision_key[-1]]])
278
def _change_inv_parents(self, stream, get_parents, all_revision_keys):
279
"""Adapt a record stream to reconcile the parents."""
280
for record in stream:
281
wanted_parents = get_parents(record.key)
282
if wanted_parents and wanted_parents[0] not in all_revision_keys:
283
# The check for the left most parent only handles knit
284
# compressors, but this code only applies to knit and weave
285
# repositories anyway.
286
bytes = record.get_bytes_as('fulltext')
287
yield FulltextContentFactory(record.key, wanted_parents, record.sha1, bytes)
289
adapted_record = AdapterFactory(record.key, wanted_parents, record)
291
self._reweave_step('adding inventories')
293
def _setup_steps(self, new_total):
294
"""Setup the markers we need to control the progress bar."""
295
self.total = new_total
298
def _graph_revision(self, rev_id):
299
"""Load a revision into the revision graph."""
300
# pick a random revision
301
# analyse revision id rev_id and put it in the stack.
302
self._reweave_step('loading revisions')
303
rev = self.repo.get_revision_reconcile(rev_id)
305
for parent in rev.parent_ids:
306
if self._parent_is_available(parent):
307
parents.append(parent)
309
mutter('found ghost %s', parent)
310
self._rev_graph[rev_id] = parents
312
def _check_garbage_inventories(self):
313
"""Check for garbage inventories which we cannot trust
315
We cant trust them because their pre-requisite file data may not
316
be present - all we know is that their revision was not installed.
318
if not self.thorough:
320
inventories = set(self.inventory.keys())
321
revisions = set(self.revisions.keys())
322
garbage = inventories.difference(revisions)
323
self.garbage_inventories = len(garbage)
324
for revision_key in garbage:
325
mutter('Garbage inventory {%s} found.', revision_key[-1])
327
def _parent_is_available(self, parent):
328
"""True if parent is a fully available revision
330
A fully available revision has a inventory and a revision object in the
333
if parent in self._rev_graph:
335
inv_present = (1 == len(self.inventory.get_parent_map([(parent,)])))
336
return (inv_present and self.repo.has_revision(parent))
338
def _reweave_step(self, message):
339
"""Mark a single step of regeneration complete."""
340
self.pb.update(message, self.count, self.total)
344
class KnitReconciler(RepoReconciler):
345
"""Reconciler that reconciles a knit format repository.
347
This will detect garbage inventories and remove them in thorough mode.
350
def _reconcile_steps(self):
351
"""Perform the steps to reconcile this repository."""
355
except errors.BzrCheckError:
358
# knits never suffer this
360
self._fix_text_parents()
362
def _load_indexes(self):
363
"""Load indexes for the reconciliation."""
364
self.transaction = self.repo.get_transaction()
365
self.pb.update('Reading indexes', 0, 2)
366
self.inventory = self.repo.inventories
367
self.pb.update('Reading indexes', 1, 2)
368
self.repo._check_for_inconsistent_revision_parents()
369
self.revisions = self.repo.revisions
370
self.pb.update('Reading indexes', 2, 2)
372
def _gc_inventory(self):
373
"""Remove inventories that are not referenced from the revision store."""
374
self.pb.update('Checking unused inventories', 0, 1)
375
self._check_garbage_inventories()
376
self.pb.update('Checking unused inventories', 1, 3)
377
if not self.garbage_inventories:
378
ui.ui_factory.note('Inventory ok.')
380
self.pb.update('Backing up inventory', 0, 0)
381
self.repo._backup_inventory()
382
ui.ui_factory.note('Backup Inventory created')
383
# asking for '' should never return a non-empty weave
384
new_inventories = self.repo._temp_inventories()
385
# we have topological order of revisions and non ghost parents ready.
386
graph = self.revisions.get_parent_map(self.revisions.keys())
387
revision_keys = topo_sort(graph)
388
revision_ids = [key[-1] for key in revision_keys]
389
self._setup_steps(len(revision_keys))
390
stream = self._change_inv_parents(
391
self.inventory.get_record_stream(revision_keys, 'unordered', True),
394
new_inventories.insert_record_stream(stream)
395
# if this worked, the set of new_inventory_vf.names should equal
396
# the revisionds list
397
if not(set(new_inventories.keys()) == set(revision_keys)):
398
raise AssertionError()
399
self.pb.update('Writing weave')
400
self.repo._activate_new_inventory()
401
self.inventory = None
402
ui.ui_factory.note('Inventory regenerated.')
404
def _fix_text_parents(self):
405
"""Fix bad versionedfile parent entries.
407
It is possible for the parents entry in a versionedfile entry to be
408
inconsistent with the values in the revision and inventory.
410
This method finds entries with such inconsistencies, corrects their
411
parent lists, and replaces the versionedfile with a corrected version.
413
transaction = self.repo.get_transaction()
414
versions = [key[-1] for key in self.revisions.keys()]
415
mutter('Prepopulating revision text cache with %d revisions',
417
vf_checker = self.repo._get_versioned_file_checker()
418
bad_parents, unused_versions = vf_checker.check_file_version_parents(
419
self.repo.texts, self.pb)
420
text_index = vf_checker.text_index
421
per_id_bad_parents = {}
422
for key in unused_versions:
423
# Ensure that every file with unused versions gets rewritten.
424
# NB: This is really not needed, reconcile != pack.
425
per_id_bad_parents[key[0]] = {}
426
# Generate per-knit/weave data.
427
for key, details in bad_parents.iteritems():
430
knit_parents = tuple([parent[-1] for parent in details[0]])
431
correct_parents = tuple([parent[-1] for parent in details[1]])
432
file_details = per_id_bad_parents.setdefault(file_id, {})
433
file_details[rev_id] = (knit_parents, correct_parents)
434
file_id_versions = {}
435
for text_key in text_index:
436
versions_list = file_id_versions.setdefault(text_key[0], [])
437
versions_list.append(text_key[1])
438
# Do the reconcile of individual weaves.
439
for num, file_id in enumerate(per_id_bad_parents):
440
self.pb.update('Fixing text parents', num,
441
len(per_id_bad_parents))
442
versions_with_bad_parents = per_id_bad_parents[file_id]
443
id_unused_versions = set(key[-1] for key in unused_versions
444
if key[0] == file_id)
445
if file_id in file_id_versions:
446
file_versions = file_id_versions[file_id]
448
# This id was present in the disk store but is not referenced
449
# by any revision at all.
451
self._fix_text_parent(file_id, versions_with_bad_parents,
452
id_unused_versions, file_versions)
454
def _fix_text_parent(self, file_id, versions_with_bad_parents,
455
unused_versions, all_versions):
456
"""Fix bad versionedfile entries in a single versioned file."""
457
mutter('fixing text parent: %r (%d versions)', file_id,
458
len(versions_with_bad_parents))
459
mutter('(%d are unused)', len(unused_versions))
460
new_file_id = 'temp:%s' % file_id
463
for version in all_versions:
464
if version in unused_versions:
466
elif version in versions_with_bad_parents:
467
parents = versions_with_bad_parents[version][1]
469
pmap = self.repo.texts.get_parent_map([(file_id, version)])
470
parents = [key[-1] for key in pmap[(file_id, version)]]
471
new_parents[(new_file_id, version)] = [
472
(new_file_id, parent) for parent in parents]
473
needed_keys.add((file_id, version))
474
def fix_parents(stream):
475
for record in stream:
476
bytes = record.get_bytes_as('fulltext')
477
new_key = (new_file_id, record.key[-1])
478
parents = new_parents[new_key]
479
yield FulltextContentFactory(new_key, parents, record.sha1, bytes)
480
stream = self.repo.texts.get_record_stream(needed_keys, 'topological', True)
481
self.repo._remove_file_id(new_file_id)
482
self.repo.texts.insert_record_stream(fix_parents(stream))
483
self.repo._remove_file_id(file_id)
485
self.repo._move_file_id(new_file_id, file_id)
488
class PackReconciler(RepoReconciler):
489
"""Reconciler that reconciles a pack based repository.
491
Garbage inventories do not affect ancestry queries, and removal is
492
considerably more expensive as there is no separate versioned file for
493
them, so they are not cleaned. In short it is currently a no-op.
495
In future this may be a good place to hook in annotation cache checking,
496
index recreation etc.
499
# XXX: The index corruption that _fix_text_parents performs is needed for
500
# packs, but not yet implemented. The basic approach is to:
501
# - lock the names list
502
# - perform a customised pack() that regenerates data as needed
503
# - unlock the names list
504
# https://bugs.launchpad.net/bzr/+bug/154173
506
def __init__(self, repo, other=None, thorough=False,
507
canonicalize_chks=False):
508
super(PackReconciler, self).__init__(repo, other=other,
510
self.canonicalize_chks = canonicalize_chks
512
def _reconcile_steps(self):
513
"""Perform the steps to reconcile this repository."""
514
if not self.thorough:
516
collection = self.repo._pack_collection
517
collection.ensure_loaded()
518
collection.lock_names()
519
self.add_cleanup(collection._unlock_names)
520
packs = collection.all_packs()
521
all_revisions = self.repo.all_revision_ids()
522
total_inventories = len(list(
523
collection.inventory_index.combined_index.iter_all_entries()))
524
if len(all_revisions):
525
if self.canonicalize_chks:
526
reconcile_meth = self.repo._canonicalize_chks_pack
528
reconcile_meth = self.repo._reconcile_pack
529
new_pack = reconcile_meth(collection, packs, ".reconcile",
530
all_revisions, self.pb)
531
if new_pack is not None:
532
self._discard_and_save(packs)
534
# only make a new pack when there is data to copy.
535
self._discard_and_save(packs)
536
self.garbage_inventories = total_inventories - len(list(
537
collection.inventory_index.combined_index.iter_all_entries()))
539
def _discard_and_save(self, packs):
540
"""Discard some packs from the repository.
542
This removes them from the memory index, saves the in-memory index
543
which makes the newly reconciled pack visible and hides the packs to be
544
discarded, and finally renames the packs being discarded into the
545
obsolete packs directory.
547
:param packs: The packs to discard.
550
self.repo._pack_collection._remove_pack_from_memory(pack)
551
self.repo._pack_collection._save_pack_names()
552
self.repo._pack_collection._obsolete_packs(packs)