← Back to branch summary

~bzr-pqm/bzr/bzr.dev

« back to all changes in this revision

Viewing changes to bzrlib/reconcile.py

  • Committer: Andrew Bennetts
  • Date: 2009-07-27 08:02:52 UTC
  • mto: This revision was merged to the branch mainline in revision 4573.
  • Revision ID: andrew.bennetts@canonical.com-20090727080252-1r4s9oqwlkzgywx7
Fix trivial bug in _vfs_set_tags_bytes.

Show diffs side-by-side

added added

removed removed

Lines of Context:
bzrlib/reconcile.py
1
 
# Copyright (C) 2006-2010 Canonical Ltd
 
1
# Copyright (C) 2005, 2006 Canonical Ltd
2
2
#
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
27
27
 
28
28
 
29
29
from bzrlib import (
30
 
    cleanup,
31
30
    errors,
32
 
    revision as _mod_revision,
33
31
    ui,
 
32
    repository,
 
33
    repofmt,
34
34
    )
35
 
from bzrlib.trace import mutter
36
 
from bzrlib.tsort import topo_sort
 
35
from bzrlib.trace import mutter, note
 
36
from bzrlib.tsort import TopoSorter
37
37
from bzrlib.versionedfile import AdapterFactory, FulltextContentFactory
38
38
 
39
39
 
40
 
def reconcile(dir, canonicalize_chks=False):
 
40
def reconcile(dir, other=None):
41
41
    """Reconcile the data in dir.
42
42
 
43
43
    Currently this is limited to a inventory 'reweave'.
47
47
    Directly using Reconciler is recommended for library users that
48
48
    desire fine grained control or analysis of the found issues.
49
49
 
50
 
    :param canonicalize_chks: Make sure CHKs are in canonical form.
 
50
    :param other: another bzrdir to reconcile against.
51
51
    """
52
 
    reconciler = Reconciler(dir, canonicalize_chks=canonicalize_chks)
 
52
    reconciler = Reconciler(dir, other=other)
53
53
    reconciler.reconcile()
54
54
 
55
55
 
56
56
class Reconciler(object):
57
57
    """Reconcilers are used to reconcile existing data."""
58
58
 
59
 
    def __init__(self, dir, other=None, canonicalize_chks=False):
 
59
    def __init__(self, dir, other=None):
60
60
        """Create a Reconciler."""
61
61
        self.bzrdir = dir
62
 
        self.canonicalize_chks = canonicalize_chks
63
62
 
64
63
    def reconcile(self):
65
64
        """Perform reconciliation.
66
65
 
67
66
        After reconciliation the following attributes document found issues:
68
 
 
69
 
        * `inconsistent_parents`: The number of revisions in the repository
70
 
          whose ancestry was being reported incorrectly.
71
 
        * `garbage_inventories`: The number of inventory objects without
72
 
          revisions that were garbage collected.
73
 
        * `fixed_branch_history`: None if there was no branch, False if the
74
 
          branch history was correct, True if the branch history needed to be
75
 
          re-normalized.
 
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.
76
74
        """
77
75
        self.pb = ui.ui_factory.nested_progress_bar()
78
76
        try:
92
90
            # Nothing to check here
93
91
            self.fixed_branch_history = None
94
92
            return
95
 
        ui.ui_factory.note('Reconciling branch %s' % self.branch.base)
 
93
        self.pb.note('Reconciling branch %s',
 
94
                     self.branch.base)
96
95
        branch_reconciler = self.branch.reconcile(thorough=True)
97
96
        self.fixed_branch_history = branch_reconciler.fixed_history
98
97
 
99
98
    def _reconcile_repository(self):
100
99
        self.repo = self.bzrdir.find_repository()
101
 
        ui.ui_factory.note('Reconciling repository %s' %
102
 
            self.repo.user_url)
 
100
        self.pb.note('Reconciling repository %s',
 
101
                     self.repo.bzrdir.root_transport.base)
103
102
        self.pb.update("Reconciling repository", 0, 1)
104
 
        if self.canonicalize_chks:
105
 
            try:
106
 
                self.repo.reconcile_canonicalize_chks
107
 
            except AttributeError:
108
 
                raise errors.BzrError(
109
 
                    "%s cannot canonicalize CHKs." % (self.repo,))
110
 
            repo_reconciler = self.repo.reconcile_canonicalize_chks()
111
 
        else:
112
 
            repo_reconciler = self.repo.reconcile(thorough=True)
 
103
        repo_reconciler = self.repo.reconcile(thorough=True)
113
104
        self.inconsistent_parents = repo_reconciler.inconsistent_parents
114
105
        self.garbage_inventories = repo_reconciler.garbage_inventories
115
106
        if repo_reconciler.aborted:
116
 
            ui.ui_factory.note(
 
107
            self.pb.note(
117
108
                'Reconcile aborted: revision index has inconsistent parents.')
118
 
            ui.ui_factory.note(
 
109
            self.pb.note(
119
110
                'Run "bzr check" for more details.')
120
111
        else:
121
 
            ui.ui_factory.note('Reconciliation complete.')
 
112
            self.pb.note('Reconciliation complete.')
122
113
 
123
114
 
124
115
class BranchReconciler(object):
130
121
        self.branch = a_branch
131
122
 
132
123
    def reconcile(self):
133
 
        operation = cleanup.OperationWithCleanups(self._reconcile)
134
 
        self.add_cleanup = operation.add_cleanup
135
 
        operation.run_simple()
136
 
 
137
 
    def _reconcile(self):
138
124
        self.branch.lock_write()
139
 
        self.add_cleanup(self.branch.unlock)
140
 
        self.pb = ui.ui_factory.nested_progress_bar()
141
 
        self.add_cleanup(self.pb.finished)
142
 
        self._reconcile_steps()
 
125
        try:
 
126
            self.pb = ui.ui_factory.nested_progress_bar()
 
127
            try:
 
128
                self._reconcile_steps()
 
129
            finally:
 
130
                self.pb.finished()
 
131
        finally:
 
132
            self.branch.unlock()
143
133
 
144
134
    def _reconcile_steps(self):
145
135
        self._reconcile_revision_history()
146
136
 
147
137
    def _reconcile_revision_history(self):
 
138
        repo = self.branch.repository
148
139
        last_revno, last_revision_id = self.branch.last_revision_info()
149
140
        real_history = []
150
 
        graph = self.branch.repository.get_graph()
151
141
        try:
152
 
            for revid in graph.iter_lefthand_ancestry(
153
 
                    last_revision_id, (_mod_revision.NULL_REVISION,)):
 
142
            for revid in repo.iter_reverse_revision_history(
 
143
                    last_revision_id):
154
144
                real_history.append(revid)
155
145
        except errors.RevisionNotPresent:
156
146
            pass # Hit a ghost left hand parent
161
151
            # set_revision_history, as this will regenerate it again.
162
152
            # Not really worth a whole BranchReconciler class just for this,
163
153
            # though.
164
 
            ui.ui_factory.note('Fixing last revision info %s => %s' % (
165
 
                 last_revno, len(real_history)))
 
154
            self.pb.note('Fixing last revision info %s => %s',
 
155
                         last_revno, len(real_history))
166
156
            self.branch.set_last_revision_info(len(real_history),
167
157
                                               last_revision_id)
168
158
        else:
169
159
            self.fixed_history = False
170
 
            ui.ui_factory.note('revision_history ok.')
 
160
            self.pb.note('revision_history ok.')
171
161
 
172
162
 
173
163
class RepoReconciler(object):
198
188
        """Perform reconciliation.
199
189
 
200
190
        After reconciliation the following attributes document found issues:
201
 
 
202
 
        * `inconsistent_parents`: The number of revisions in the repository
203
 
          whose ancestry was being reported incorrectly.
204
 
        * `garbage_inventories`: The number of inventory objects without
205
 
          revisions that were garbage collected.
 
191
        inconsistent_parents: The number of revisions in the repository whose
 
192
                              ancestry was being reported incorrectly.
 
193
        garbage_inventories: The number of inventory objects without revisions
 
194
                             that were garbage collected.
206
195
        """
207
 
        operation = cleanup.OperationWithCleanups(self._reconcile)
208
 
        self.add_cleanup = operation.add_cleanup
209
 
        operation.run_simple()
210
 
 
211
 
    def _reconcile(self):
212
196
        self.repo.lock_write()
213
 
        self.add_cleanup(self.repo.unlock)
214
 
        self.pb = ui.ui_factory.nested_progress_bar()
215
 
        self.add_cleanup(self.pb.finished)
216
 
        self._reconcile_steps()
 
197
        try:
 
198
            self.pb = ui.ui_factory.nested_progress_bar()
 
199
            try:
 
200
                self._reconcile_steps()
 
201
            finally:
 
202
                self.pb.finished()
 
203
        finally:
 
204
            self.repo.unlock()
217
205
 
218
206
    def _reconcile_steps(self):
219
207
        """Perform the steps to reconcile this repository."""
227
215
        only data-loss causing issues (!self.thorough) or all issues
228
216
        (self.thorough) are treated as requiring the reweave.
229
217
        """
 
218
        # local because needing to know about WeaveFile is a wart we want to hide
 
219
        from bzrlib.weave import WeaveFile, Weave
230
220
        transaction = self.repo.get_transaction()
231
221
        self.pb.update('Reading inventory data')
232
222
        self.inventory = self.repo.inventories
248
238
        # (no garbage inventories or we are not doing a thorough check)
249
239
        if (not self.inconsistent_parents and
250
240
            (not self.garbage_inventories or not self.thorough)):
251
 
            ui.ui_factory.note('Inventory ok.')
 
241
            self.pb.note('Inventory ok.')
252
242
            return
253
243
        self.pb.update('Backing up inventory', 0, 0)
254
244
        self.repo._backup_inventory()
255
 
        ui.ui_factory.note('Backup inventory created.')
 
245
        self.pb.note('Backup inventory created.')
256
246
        new_inventories = self.repo._temp_inventories()
257
247
 
258
248
        # we have topological order of revisions and non ghost parents ready.
259
249
        self._setup_steps(len(self._rev_graph))
260
 
        revision_keys = [(rev_id,) for rev_id in topo_sort(self._rev_graph)]
 
250
        revision_keys = [(rev_id,) for rev_id in
 
251
            TopoSorter(self._rev_graph.items()).iter_topo_order()]
261
252
        stream = self._change_inv_parents(
262
253
            self.inventory.get_record_stream(revision_keys, 'unordered', True),
263
254
            self._new_inv_parents,
271
262
        self.pb.update('Writing weave')
272
263
        self.repo._activate_new_inventory()
273
264
        self.inventory = None
274
 
        ui.ui_factory.note('Inventory regenerated.')
 
265
        self.pb.note('Inventory regenerated.')
275
266
 
276
267
    def _new_inv_parents(self, revision_key):
277
268
        """Lookup ghost-filtered parents for revision_key."""
378
369
        self._check_garbage_inventories()
379
370
        self.pb.update('Checking unused inventories', 1, 3)
380
371
        if not self.garbage_inventories:
381
 
            ui.ui_factory.note('Inventory ok.')
 
372
            self.pb.note('Inventory ok.')
382
373
            return
383
374
        self.pb.update('Backing up inventory', 0, 0)
384
375
        self.repo._backup_inventory()
385
 
        ui.ui_factory.note('Backup Inventory created')
 
376
        self.pb.note('Backup Inventory created')
386
377
        # asking for '' should never return a non-empty weave
387
378
        new_inventories = self.repo._temp_inventories()
388
379
        # we have topological order of revisions and non ghost parents ready.
389
380
        graph = self.revisions.get_parent_map(self.revisions.keys())
390
 
        revision_keys = topo_sort(graph)
 
381
        revision_keys = list(TopoSorter(graph).iter_topo_order())
391
382
        revision_ids = [key[-1] for key in revision_keys]
392
383
        self._setup_steps(len(revision_keys))
393
384
        stream = self._change_inv_parents(
402
393
        self.pb.update('Writing weave')
403
394
        self.repo._activate_new_inventory()
404
395
        self.inventory = None
405
 
        ui.ui_factory.note('Inventory regenerated.')
 
396
        self.pb.note('Inventory regenerated.')
406
397
 
407
398
    def _fix_text_parents(self):
408
399
        """Fix bad versionedfile parent entries.
504
495
    #  - lock the names list
505
496
    #  - perform a customised pack() that regenerates data as needed
506
497
    #  - unlock the names list
507
 
    # https://bugs.launchpad.net/bzr/+bug/154173
508
 
 
509
 
    def __init__(self, repo, other=None, thorough=False,
510
 
            canonicalize_chks=False):
511
 
        super(PackReconciler, self).__init__(repo, other=other,
512
 
            thorough=thorough)
513
 
        self.canonicalize_chks = canonicalize_chks
 
498
    # https://bugs.edge.launchpad.net/bzr/+bug/154173
514
499
 
515
500
    def _reconcile_steps(self):
516
501
        """Perform the steps to reconcile this repository."""
519
504
        collection = self.repo._pack_collection
520
505
        collection.ensure_loaded()
521
506
        collection.lock_names()
522
 
        self.add_cleanup(collection._unlock_names)
523
 
        packs = collection.all_packs()
524
 
        all_revisions = self.repo.all_revision_ids()
525
 
        total_inventories = len(list(
526
 
            collection.inventory_index.combined_index.iter_all_entries()))
527
 
        if len(all_revisions):
528
 
            if self.canonicalize_chks:
529
 
                reconcile_meth = self.repo._canonicalize_chks_pack
 
507
        try:
 
508
            packs = collection.all_packs()
 
509
            all_revisions = self.repo.all_revision_ids()
 
510
            total_inventories = len(list(
 
511
                collection.inventory_index.combined_index.iter_all_entries()))
 
512
            if len(all_revisions):
 
513
                new_pack =  self.repo._reconcile_pack(collection, packs,
 
514
                    ".reconcile", all_revisions, self.pb)
 
515
                if new_pack is not None:
 
516
                    self._discard_and_save(packs)
530
517
            else:
531
 
                reconcile_meth = self.repo._reconcile_pack
532
 
            new_pack = reconcile_meth(collection, packs, ".reconcile",
533
 
                all_revisions, self.pb)
534
 
            if new_pack is not None:
 
518
                # only make a new pack when there is data to copy.
535
519
                self._discard_and_save(packs)
536
 
        else:
537
 
            # only make a new pack when there is data to copy.
538
 
            self._discard_and_save(packs)
539
 
        self.garbage_inventories = total_inventories - len(list(
540
 
            collection.inventory_index.combined_index.iter_all_entries()))
 
520
            self.garbage_inventories = total_inventories - len(list(
 
521
                collection.inventory_index.combined_index.iter_all_entries()))
 
522
        finally:
 
523
            collection._unlock_names()
541
524
 
542
525
    def _discard_and_save(self, packs):
543
526
        """Discard some packs from the repository.