1
# Copyright (C) 2005, 2006 Canonical Ltd
1
# Copyright (C) 2004, 2005 by Martin Pool
2
# Copyright (C) 2005 by Canonical Ltd
3
4
# This program is free software; you can redistribute it and/or modify
4
5
# it under the terms of the GNU General Public License as published by
5
6
# the Free Software Foundation; either version 2 of the License, or
6
7
# (at your option) any later version.
8
9
# This program is distributed in the hope that it will be useful,
9
10
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
11
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
12
# GNU General Public License for more details.
13
14
# You should have received a copy of the GNU General Public License
14
15
# along with this program; if not, write to the Free Software
15
16
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
27
28
# TODO: Get every revision in the revision-store even if they're not
28
29
# referenced by history and make sure they're all valid.
30
# TODO: Perhaps have a way to record errors other than by raising exceptions;
31
# would perhaps be enough to accumulate exception objects in a list without
32
# raising them. If there's more than one exception it'd be good to see them
35
from bzrlib import errors
36
from bzrlib import repository as _mod_repository
37
from bzrlib import revision
38
from bzrlib.errors import BzrCheckError
40
from bzrlib.trace import log_error, note
32
from bzrlib.trace import note, warning
33
from bzrlib.osutils import rename, sha_string, fingerprint_file
34
from bzrlib.trace import mutter
35
from bzrlib.errors import BzrCheckError, NoSuchRevision
36
from bzrlib.inventory import ROOT_ID
42
39
class Check(object):
43
"""Check a repository"""
45
# The Check object interacts with InventoryEntry.check, etc.
47
def __init__(self, repository):
48
self.repository = repository
42
def __init__(self, branch):
44
self.repository = branch.repository
49
45
self.checked_text_cnt = 0
50
46
self.checked_rev_cnt = 0
53
49
self.missing_parent_links = {}
54
50
self.missing_inventory_sha_cnt = 0
55
51
self.missing_revision_cnt = 0
56
# maps (file-id, version) -> sha1; used by InventoryFile._check
52
# maps (file-id, version) -> sha1
57
53
self.checked_texts = {}
58
54
self.checked_weaves = {}
59
self.revision_versions = _mod_repository._RevisionTextVersionCache(
61
self.unreferenced_ancestors = set()
62
self.inconsistent_parents = []
63
self.dangling_versions = set()
66
self.repository.lock_read()
67
self.progress = bzrlib.ui.ui_factory.nested_progress_bar()
57
self.branch.lock_read()
58
self.progress = bzrlib.ui.ui_factory.progress_bar()
69
self.progress.update('retrieving inventory', 0, 2)
60
self.progress.update('retrieving inventory', 0, 0)
70
61
# do not put in init, as it should be done with progess,
71
62
# and inside the lock.
72
self.inventory_weave = self.repository.get_inventory_weave()
73
self.progress.update('checking revision graph', 1)
74
self.check_revision_graph()
63
self.inventory_weave = self.branch.repository.get_inventory_weave()
64
self.history = self.branch.revision_history()
65
if not len(self.history):
75
68
self.plan_revisions()
77
71
while revno < len(self.planned_revisions):
78
72
rev_id = self.planned_revisions[revno]
79
73
self.progress.update('checking revision', revno,
80
74
len(self.planned_revisions))
82
76
self.check_one_rev(rev_id)
83
# check_weaves is done after the revision scan so that
84
# revision_versions is pre-populated
87
self.progress.finished()
88
self.repository.unlock()
90
def check_revision_graph(self):
91
if not self.repository.revision_graph_can_have_wrong_parents():
92
# This check is not necessary.
93
self.revs_with_bad_parents_in_index = None
81
def plan_revisions(self):
82
repository = self.branch.repository
83
if not repository.revision_store.listable():
84
self.planned_revisions = repository.get_ancestry(self.history[-1])
85
self.planned_revisions.remove(None)
86
# FIXME progress bars should support this more nicely.
88
print ("Checking reachable history -"
89
" for a complete check use a local branch.")
95
bad_revisions = self.repository._find_inconsistent_revision_parents()
96
self.revs_with_bad_parents_in_index = list(bad_revisions)
98
def plan_revisions(self):
99
repository = self.repository
100
self.planned_revisions = repository.all_revision_ids()
101
self.progress.clear()
102
inventoried = set(self.inventory_weave.versions())
103
awol = set(self.planned_revisions) - inventoried
92
self.planned_revisions = set(repository.revision_store)
93
inventoried = set(self.inventory_weave.names())
94
awol = self.planned_revisions - inventoried
105
96
raise BzrCheckError('Stored revisions missing from inventory'
106
97
'{%s}' % ','.join([f for f in awol]))
98
self.planned_revisions = list(self.planned_revisions)
108
100
def report_results(self, verbose):
109
note('checked repository %s format %s',
110
self.repository.bzrdir.root_transport,
111
self.repository._format)
101
note('checked branch %s format %s',
103
self.branch._branch_format)
112
105
note('%6d revisions', self.checked_rev_cnt)
113
note('%6d file-ids', len(self.checked_weaves))
114
106
note('%6d unique file texts', self.checked_text_cnt)
115
107
note('%6d repeated file texts', self.repeated_text_cnt)
116
note('%6d unreferenced text ancestors',
117
len(self.unreferenced_ancestors))
108
note('%6d weaves', len(self.checked_weaves))
118
109
if self.missing_inventory_sha_cnt:
119
110
note('%6d revisions are missing inventory_sha1',
120
111
self.missing_inventory_sha_cnt)
127
118
for ghost in self.ghosts:
128
119
note(' %s', ghost)
129
120
if len(self.missing_parent_links):
130
note('%6d revisions missing parents in ancestry',
121
note('%6d revisions missing parents in ancestry',
131
122
len(self.missing_parent_links))
133
124
for link, linkers in self.missing_parent_links.items():
134
125
note(' %s should be in the ancestry for:', link)
135
126
for linker in linkers:
136
127
note(' * %s', linker)
138
for file_id, revision_id in self.unreferenced_ancestors:
139
log_error('unreferenced ancestor: {%s} in %s', revision_id,
141
if len(self.inconsistent_parents):
142
note('%6d inconsistent parents', len(self.inconsistent_parents))
144
for info in self.inconsistent_parents:
145
revision_id, file_id, found_parents, correct_parents = info
146
note(' * %s version %s has parents %r '
148
% (file_id, revision_id, found_parents,
150
if self.revs_with_bad_parents_in_index:
151
note('%6d revisions have incorrect parents in the revision index',
152
len(self.revs_with_bad_parents_in_index))
154
for item in self.revs_with_bad_parents_in_index:
155
revision_id, index_parents, actual_parents = item
157
' %s has wrong parents in index: '
159
revision_id, index_parents, actual_parents)
160
if self.dangling_versions:
161
note('%6d file versions are not referenced by their inventory',
162
len(self.dangling_versions))
164
129
def check_one_rev(self, rev_id):
165
130
"""Check one revision.
167
132
rev_id - the one to check
134
last_rev_id - the previous one on the mainline, if any.
169
rev = self.repository.get_revision(rev_id)
137
# mutter(' revision {%s}', rev_id)
140
rev_history_position = self.history.index(rev_id)
142
rev_history_position = None
144
if rev_history_position:
145
rev = branch.repository.get_revision(rev_id)
146
if rev_history_position > 0:
147
last_rev_id = self.history[rev_history_position - 1]
149
rev = branch.repository.get_revision(rev_id)
171
151
if rev.revision_id != rev_id:
172
152
raise BzrCheckError('wrong internal revision id in revision {%s}'
175
for parent in rev.parent_ids:
176
if not parent in self.planned_revisions:
177
missing_links = self.missing_parent_links.get(parent, [])
178
missing_links.append(rev_id)
179
self.missing_parent_links[parent] = missing_links
180
# list based so somewhat slow,
181
# TODO have a planned_revisions list and set.
182
if self.repository.has_revision(parent):
183
missing_ancestry = self.repository.get_ancestry(parent)
184
for missing in missing_ancestry:
185
if (missing is not None
186
and missing not in self.planned_revisions):
187
self.planned_revisions.append(missing)
155
# check the previous history entry is a parent of this entry
157
if last_rev_id is not None:
158
for parent_id in rev.parent_ids:
159
if parent_id == last_rev_id:
189
self.ghosts.append(rev_id)
162
raise BzrCheckError("previous revision {%s} not listed among "
164
% (last_rev_id, rev_id))
165
for parent in rev.parent_ids:
166
if not parent in self.planned_revisions:
167
missing_links = self.missing_parent_links.get(parent, [])
168
missing_links.append(rev_id)
169
self.missing_parent_links[parent] = missing_links
170
# list based so somewhat slow,
171
# TODO have a planned_revisions list and set.
172
if self.branch.has_revision(parent):
173
missing_ancestry = self.repository.get_ancestry(parent)
174
for missing in missing_ancestry:
175
if (missing is not None
176
and missing not in self.planned_revisions):
177
self.planned_revisions.append(missing)
179
self.ghosts.append(rev_id)
181
raise BzrCheckError("revision {%s} has no parents listed "
182
"but preceded by {%s}"
183
% (rev_id, last_rev_id))
191
185
if rev.inventory_sha1:
192
inv_sha1 = self.repository.get_inventory_sha1(rev_id)
186
inv_sha1 = branch.repository.get_inventory_sha1(rev_id)
193
187
if inv_sha1 != rev.inventory_sha1:
194
188
raise BzrCheckError('Inventory sha1 hash doesn\'t match'
195
189
' value in revision {%s}' % rev_id)
191
missing_inventory_sha_cnt += 1
192
mutter("no inventory_sha1 on revision {%s}", rev_id)
196
193
self._check_revision_tree(rev_id)
197
194
self.checked_rev_cnt += 1
204
if self.repository.weave_store.listable():
205
weave_ids = list(self.repository.weave_store)
206
n_weaves = len(weave_ids) + 1
207
self.progress.update('checking versionedfile', 0, n_weaves)
201
if self.branch.repository.weave_store.listable():
202
weave_ids = list(self.branch.repository.weave_store)
203
n_weaves = len(weave_ids)
204
self.progress.update('checking weave', 0, n_weaves)
208
205
self.inventory_weave.check(progress_bar=self.progress)
209
files_in_revisions = {}
210
revisions_of_files = {}
211
206
for i, weave_id in enumerate(weave_ids):
212
self.progress.update('checking versionedfile', i, n_weaves)
213
w = self.repository.weave_store.get_weave(weave_id,
214
self.repository.get_transaction())
207
self.progress.update('checking weave', i, n_weaves)
208
w = self.branch.repository.weave_store.get_weave(weave_id,
209
self.branch.repository.get_transaction())
215
210
# No progress here, because it looks ugly.
218
weave_checker = self.repository.get_versioned_file_checker(
219
self.planned_revisions, self.revision_versions)
220
result = weave_checker.check_file_version_parents(w, weave_id)
221
bad_parents, dangling_versions = result
222
bad_parents = bad_parents.items()
223
for revision_id, (weave_parents,correct_parents) in bad_parents:
224
self.inconsistent_parents.append(
225
(revision_id, weave_id, weave_parents, correct_parents))
226
if weave_parents is None:
228
unreferenced_parents = set(weave_parents)-set(correct_parents)
229
for unreferenced_parent in unreferenced_parents:
230
self.unreferenced_ancestors.add(
231
(weave_id, unreferenced_parent))
232
self.dangling_versions.update(dangling_versions)
233
212
self.checked_weaves[weave_id] = True
235
214
def _check_revision_tree(self, rev_id):
236
tree = self.repository.revision_tree(rev_id)
237
self.revision_versions.add_revision_text_versions(tree)
215
tree = self.branch.repository.revision_tree(rev_id)
238
216
inv = tree.inventory
240
218
for file_id in inv:
added
removed
bzrlib/check.py
