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
24
25
# TODO: Check revision, inventory and entry objects have all
27
# TODO: Get every revision in the revision-store even if they're not
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.errors import BzrCheckError
37
from bzrlib.trace import note
30
from bzrlib.trace import note, warning
31
from bzrlib.osutils import rename, sha_string, fingerprint_file
32
from bzrlib.trace import mutter
33
from bzrlib.errors import BzrCheckError, NoSuchRevision
34
from bzrlib.inventory import ROOT_ID
35
from bzrlib.branch import gen_root_id
39
38
class Check(object):
40
"""Check a repository"""
42
# The Check object interacts with InventoryEntry.check, etc.
44
def __init__(self, repository):
45
self.repository = repository
41
def __init__(self, branch):
46
43
self.checked_text_cnt = 0
47
44
self.checked_rev_cnt = 0
50
47
self.missing_parent_links = {}
51
48
self.missing_inventory_sha_cnt = 0
52
49
self.missing_revision_cnt = 0
53
# maps (file-id, version) -> sha1; used by InventoryFile._check
50
# maps (file-id, version) -> sha1
54
51
self.checked_texts = {}
55
self.checked_weaves = {}
58
self.repository.lock_read()
59
self.progress = bzrlib.ui.ui_factory.nested_progress_bar()
54
self.branch.lock_read()
61
self.progress.update('retrieving inventory', 0, 0)
62
# do not put in init, as it should be done with progess,
63
# and inside the lock.
64
self.inventory_weave = self.repository.get_inventory_weave()
56
self.history = self.branch.revision_history()
57
if not len(self.history):
60
self.planned_revisions = self.branch.get_ancestry(self.history[-1])
61
self.planned_revisions.remove(None)
64
self.progress = bzrlib.ui.ui_factory.progress_bar()
68
65
while revno < len(self.planned_revisions):
69
66
rev_id = self.planned_revisions[revno]
70
67
self.progress.update('checking revision', revno,
71
68
len(self.planned_revisions))
73
70
self.check_one_rev(rev_id)
75
self.progress.finished()
76
self.repository.unlock()
78
def plan_revisions(self):
79
repository = self.repository
80
self.planned_revisions = set(repository.all_revision_ids())
82
inventoried = set(self.inventory_weave.versions())
83
awol = self.planned_revisions - inventoried
85
raise BzrCheckError('Stored revisions missing from inventory'
86
'{%s}' % ','.join([f for f in awol]))
87
self.planned_revisions = list(self.planned_revisions)
89
75
def report_results(self, verbose):
90
note('checked repository %s format %s',
91
self.repository.bzrdir.root_transport,
92
self.repository._format)
76
note('checked branch %s format %d',
78
self.branch._branch_format)
93
80
note('%6d revisions', self.checked_rev_cnt)
94
81
note('%6d unique file texts', self.checked_text_cnt)
95
82
note('%6d repeated file texts', self.repeated_text_cnt)
96
note('%6d weaves', len(self.checked_weaves))
97
83
if self.missing_inventory_sha_cnt:
98
84
note('%6d revisions are missing inventory_sha1',
99
85
self.missing_inventory_sha_cnt)
118
104
"""Check one revision.
120
106
rev_id - the one to check
108
last_rev_id - the previous one on the mainline, if any.
122
rev = self.repository.get_revision(rev_id)
111
# mutter(' revision {%s}' % rev_id)
114
rev_history_position = self.history.index(rev_id)
116
rev_history_position = None
118
if rev_history_position:
119
rev = branch.get_revision(rev_id)
120
if rev_history_position > 0:
121
last_rev_id = self.history[rev_history_position - 1]
123
rev = branch.get_revision(rev_id)
124
125
if rev.revision_id != rev_id:
125
126
raise BzrCheckError('wrong internal revision id in revision {%s}'
128
for parent in rev.parent_ids:
129
if not parent in self.planned_revisions:
130
missing_links = self.missing_parent_links.get(parent, [])
131
missing_links.append(rev_id)
132
self.missing_parent_links[parent] = missing_links
133
# list based so somewhat slow,
134
# TODO have a planned_revisions list and set.
135
if self.repository.has_revision(parent):
136
missing_ancestry = self.repository.get_ancestry(parent)
137
for missing in missing_ancestry:
138
if (missing is not None
139
and missing not in self.planned_revisions):
140
self.planned_revisions.append(missing)
129
# check the previous history entry is a parent of this entry
131
if last_rev_id is None and rev_history_position is not None:
132
# what if the start is a ghost ? i.e. conceptually the
134
raise BzrCheckError("revision {%s} has %d parents, but is the "
135
"start of the branch"
136
% (rev_id, len(rev.parent_ids)))
137
if last_rev_id is not None:
138
for parent_id in rev.parent_ids:
139
if parent_id == last_rev_id:
142
self.ghosts.append(rev_id)
142
raise BzrCheckError("previous revision {%s} not listed among "
144
% (last_rev_id, rev_id))
145
for parent in rev.parent_ids:
146
if not parent in self.planned_revisions:
147
missing_links = self.missing_parent_links.get(parent, [])
148
missing_links.append(rev_id)
149
self.missing_parent_links[parent] = missing_links
150
# list based so slow, TODO have a planned_revisions list and set.
151
if self.branch.has_revision(parent):
152
missing_ancestry = self.branch.get_ancestry(parent)
153
for missing in missing_ancestry:
154
if (missing is not None
155
and missing not in self.planned_revisions):
156
self.planned_revisions.append(missing)
158
self.ghosts.append(rev_id)
160
raise BzrCheckError("revision {%s} has no parents listed "
161
"but preceded by {%s}"
162
% (rev_id, last_rev_id))
144
164
if rev.inventory_sha1:
145
inv_sha1 = self.repository.get_inventory_sha1(rev_id)
165
inv_sha1 = branch.get_inventory_sha1(rev_id)
146
166
if inv_sha1 != rev.inventory_sha1:
147
167
raise BzrCheckError('Inventory sha1 hash doesn\'t match'
148
168
' value in revision {%s}' % rev_id)
170
missing_inventory_sha_cnt += 1
171
mutter("no inventory_sha1 on revision {%s}" % rev_id)
149
172
self._check_revision_tree(rev_id)
150
173
self.checked_rev_cnt += 1
152
def check_weaves(self):
153
"""Check all the weaves we can get our hands on.
157
if self.repository.weave_store.listable():
158
weave_ids = list(self.repository.weave_store)
159
n_weaves = len(weave_ids)
160
self.progress.update('checking weave', 0, n_weaves)
161
self.inventory_weave.check(progress_bar=self.progress)
162
for i, weave_id in enumerate(weave_ids):
163
self.progress.update('checking weave', i, n_weaves)
164
w = self.repository.weave_store.get_weave(weave_id,
165
self.repository.get_transaction())
166
# No progress here, because it looks ugly.
168
self.checked_weaves[weave_id] = True
170
175
def _check_revision_tree(self, rev_id):
171
tree = self.repository.revision_tree(rev_id)
176
tree = self.branch.revision_tree(rev_id)
172
177
inv = tree.inventory
174
179
for file_id in inv:
added
removed
bzrlib/check.py
