1
# Copyright (C) 2004, 2005 by Martin Pool
2
# Copyright (C) 2005 by Canonical Ltd
1
# Copyright (C) 2005, 2006 Canonical Ltd
4
3
# This program is free software; you can redistribute it and/or modify
5
4
# it under the terms of the GNU General Public License as published by
6
5
# the Free Software Foundation; either version 2 of the License, or
7
6
# (at your option) any later version.
9
8
# This program is distributed in the hope that it will be useful,
10
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
11
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12
11
# GNU General Public License for more details.
14
13
# You should have received a copy of the GNU General Public License
15
14
# along with this program; if not, write to the Free Software
16
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
28
27
# TODO: Get every revision in the revision-store even if they're not
29
28
# 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
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
40
from bzrlib.trace import log_error, note
39
42
class Check(object):
42
def __init__(self, branch):
43
"""Check a repository"""
45
# The Check object interacts with InventoryEntry.check, etc.
47
def __init__(self, repository):
48
self.repository = repository
44
49
self.checked_text_cnt = 0
45
50
self.checked_rev_cnt = 0
48
53
self.missing_parent_links = {}
49
54
self.missing_inventory_sha_cnt = 0
50
55
self.missing_revision_cnt = 0
51
# maps (file-id, version) -> sha1
56
# maps (file-id, version) -> sha1; used by InventoryFile._check
52
57
self.checked_texts = {}
53
58
self.checked_weaves = {}
59
self.unreferenced_versions = set()
60
self.inconsistent_parents = []
56
self.branch.lock_read()
57
self.progress = bzrlib.ui.ui_factory.progress_bar()
63
self.repository.lock_read()
64
self.progress = bzrlib.ui.ui_factory.nested_progress_bar()
59
self.progress.update('retrieving inventory', 0, 0)
66
self.progress.update('retrieving inventory', 0, 2)
60
67
# do not put in init, as it should be done with progess,
61
68
# and inside the lock.
62
self.inventory_weave = self.branch._get_inventory_weave()
63
self.history = self.branch.revision_history()
64
if not len(self.history):
69
self.inventory_weave = self.repository.get_inventory_weave()
70
self.progress.update('checking revision graph', 1)
71
self.check_revision_graph()
67
72
self.plan_revisions()
70
74
while revno < len(self.planned_revisions):
71
75
rev_id = self.planned_revisions[revno]
72
76
self.progress.update('checking revision', revno,
73
77
len(self.planned_revisions))
75
79
self.check_one_rev(rev_id)
80
# check_weaves is done after the revision scan so that
81
# revision index is known to be valid.
84
self.progress.finished()
85
self.repository.unlock()
87
def check_revision_graph(self):
88
if not self.repository.revision_graph_can_have_wrong_parents():
89
# This check is not necessary.
90
self.revs_with_bad_parents_in_index = None
92
bad_revisions = self.repository._find_inconsistent_revision_parents()
93
self.revs_with_bad_parents_in_index = list(bad_revisions)
80
95
def plan_revisions(self):
81
if not self.branch.revision_store.listable():
82
self.planned_revisions = self.branch.get_ancestry(self.history[-1])
83
self.planned_revisions.remove(None)
84
# FIXME progress bars should support this more nicely.
86
print ("Checking reachable history -"
87
" for a complete check use a local branch.")
90
self.planned_revisions = set(self.branch.revision_store)
91
inventoried = set(self.inventory_weave.names())
92
awol = self.planned_revisions - inventoried
96
repository = self.repository
97
self.planned_revisions = repository.all_revision_ids()
99
inventoried = set(self.inventory_weave.versions())
100
awol = set(self.planned_revisions) - inventoried
94
102
raise BzrCheckError('Stored revisions missing from inventory'
95
103
'{%s}' % ','.join([f for f in awol]))
96
self.planned_revisions = list(self.planned_revisions)
98
105
def report_results(self, verbose):
99
note('checked branch %s format %d',
101
self.branch._branch_format)
106
note('checked repository %s format %s',
107
self.repository.bzrdir.root_transport,
108
self.repository._format)
103
109
note('%6d revisions', self.checked_rev_cnt)
110
note('%6d file-ids', len(self.checked_weaves))
104
111
note('%6d unique file texts', self.checked_text_cnt)
105
112
note('%6d repeated file texts', self.repeated_text_cnt)
106
note('%6d weaves', len(self.checked_weaves))
113
note('%6d unreferenced text versions',
114
len(self.unreferenced_versions))
107
115
if self.missing_inventory_sha_cnt:
108
116
note('%6d revisions are missing inventory_sha1',
109
117
self.missing_inventory_sha_cnt)
116
124
for ghost in self.ghosts:
117
125
note(' %s', ghost)
118
126
if len(self.missing_parent_links):
119
note('%6d revisions missing parents in ancestry',
127
note('%6d revisions missing parents in ancestry',
120
128
len(self.missing_parent_links))
122
130
for link, linkers in self.missing_parent_links.items():
123
131
note(' %s should be in the ancestry for:', link)
124
132
for linker in linkers:
125
133
note(' * %s', linker)
135
for file_id, revision_id in self.unreferenced_versions:
136
log_error('unreferenced version: {%s} in %s', revision_id,
138
if len(self.inconsistent_parents):
139
note('%6d inconsistent parents', len(self.inconsistent_parents))
141
for info in self.inconsistent_parents:
142
revision_id, file_id, found_parents, correct_parents = info
143
note(' * %s version %s has parents %r '
145
% (file_id, revision_id, found_parents,
147
if self.revs_with_bad_parents_in_index:
148
note('%6d revisions have incorrect parents in the revision index',
149
len(self.revs_with_bad_parents_in_index))
151
for item in self.revs_with_bad_parents_in_index:
152
revision_id, index_parents, actual_parents = item
154
' %s has wrong parents in index: '
156
revision_id, index_parents, actual_parents)
127
158
def check_one_rev(self, rev_id):
128
159
"""Check one revision.
130
161
rev_id - the one to check
132
last_rev_id - the previous one on the mainline, if any.
135
# mutter(' revision {%s}', rev_id)
138
rev_history_position = self.history.index(rev_id)
140
rev_history_position = None
142
if rev_history_position:
143
rev = branch.get_revision(rev_id)
144
if rev_history_position > 0:
145
last_rev_id = self.history[rev_history_position - 1]
147
rev = branch.get_revision(rev_id)
163
rev = self.repository.get_revision(rev_id)
149
165
if rev.revision_id != rev_id:
150
166
raise BzrCheckError('wrong internal revision id in revision {%s}'
153
# check the previous history entry is a parent of this entry
155
if last_rev_id is not None:
156
for parent_id in rev.parent_ids:
157
if parent_id == last_rev_id:
169
for parent in rev.parent_ids:
170
if not parent in self.planned_revisions:
171
missing_links = self.missing_parent_links.get(parent, [])
172
missing_links.append(rev_id)
173
self.missing_parent_links[parent] = missing_links
174
# list based so somewhat slow,
175
# TODO have a planned_revisions list and set.
176
if self.repository.has_revision(parent):
177
missing_ancestry = self.repository.get_ancestry(parent)
178
for missing in missing_ancestry:
179
if (missing is not None
180
and missing not in self.planned_revisions):
181
self.planned_revisions.append(missing)
160
raise BzrCheckError("previous revision {%s} not listed among "
162
% (last_rev_id, rev_id))
163
for parent in rev.parent_ids:
164
if not parent in self.planned_revisions:
165
missing_links = self.missing_parent_links.get(parent, [])
166
missing_links.append(rev_id)
167
self.missing_parent_links[parent] = missing_links
168
# list based so somewhat slow,
169
# TODO have a planned_revisions list and set.
170
if self.branch.has_revision(parent):
171
missing_ancestry = self.branch.get_ancestry(parent)
172
for missing in missing_ancestry:
173
if (missing is not None
174
and missing not in self.planned_revisions):
175
self.planned_revisions.append(missing)
177
self.ghosts.append(rev_id)
179
raise BzrCheckError("revision {%s} has no parents listed "
180
"but preceded by {%s}"
181
% (rev_id, last_rev_id))
183
self.ghosts.append(rev_id)
183
185
if rev.inventory_sha1:
184
inv_sha1 = branch.get_inventory_sha1(rev_id)
186
inv_sha1 = self.repository.get_inventory_sha1(rev_id)
185
187
if inv_sha1 != rev.inventory_sha1:
186
188
raise BzrCheckError('Inventory sha1 hash doesn\'t match'
187
189
' value in revision {%s}' % rev_id)
189
missing_inventory_sha_cnt += 1
190
mutter("no inventory_sha1 on revision {%s}", rev_id)
191
190
self._check_revision_tree(rev_id)
192
191
self.checked_rev_cnt += 1
199
if self.branch.weave_store.listable():
200
weave_ids = list(self.branch.weave_store)
201
n_weaves = len(weave_ids)
202
self.progress.update('checking weave', 0, n_weaves)
198
if self.repository.weave_store.listable():
199
weave_ids = list(self.repository.weave_store)
200
n_weaves = len(weave_ids) + 1
201
self.progress.update('checking versionedfile', 0, n_weaves)
203
202
self.inventory_weave.check(progress_bar=self.progress)
203
files_in_revisions = {}
204
revisions_of_files = {}
205
weave_checker = self.repository.get_versioned_file_checker()
204
206
for i, weave_id in enumerate(weave_ids):
205
self.progress.update('checking weave', i, n_weaves)
206
w = self.branch.weave_store.get_weave(weave_id,
207
self.branch.get_transaction())
207
self.progress.update('checking versionedfile', i, n_weaves)
208
w = self.repository.weave_store.get_weave(weave_id,
209
self.repository.get_transaction())
208
210
# No progress here, because it looks ugly.
212
result = weave_checker.check_file_version_parents(w, weave_id,
213
self.planned_revisions)
214
bad_parents, unused_versions = result
215
bad_parents = bad_parents.items()
216
for revision_id, (weave_parents, correct_parents) in bad_parents:
217
self.inconsistent_parents.append(
218
(revision_id, weave_id, weave_parents, correct_parents))
219
for revision_id in unused_versions:
220
self.unreferenced_versions.add((weave_id, revision_id))
210
221
self.checked_weaves[weave_id] = True
212
223
def _check_revision_tree(self, rev_id):
213
tree = self.branch.revision_tree(rev_id)
224
tree = self.repository.revision_tree(rev_id)
214
225
inv = tree.inventory
216
227
for file_id in inv:
added
removed
bzrlib/check.py
