##// END OF EJS Templates
verify: do not abort on fully corrupted revlog
Benoit Boissinot -
r7832:1fce19b9 default
parent child Browse files
Show More
@@ -1,239 +1,243 b''
1 1 # verify.py - repository integrity checking for Mercurial
2 2 #
3 3 # Copyright 2006, 2007 Matt Mackall <mpm@selenic.com>
4 4 #
5 5 # This software may be used and distributed according to the terms
6 6 # of the GNU General Public License, incorporated herein by reference.
7 7
8 8 from node import nullid, short
9 9 from i18n import _
10 import revlog, util
10 import revlog, util, error
11 11
12 12 def verify(repo):
13 13 lock = repo.lock()
14 14 try:
15 15 return _verify(repo)
16 16 finally:
17 17 del lock
18 18
19 19 def _verify(repo):
20 20 mflinkrevs = {}
21 21 filelinkrevs = {}
22 22 filenodes = {}
23 23 revisions = 0
24 24 badrevs = {}
25 25 errors = [0]
26 26 warnings = [0]
27 27 ui = repo.ui
28 28 cl = repo.changelog
29 29 mf = repo.manifest
30 30
31 31 if not repo.cancopy():
32 32 raise util.Abort(_("cannot verify bundle or remote repos"))
33 33
34 34 def err(linkrev, msg, filename=None):
35 35 if linkrev != None:
36 36 badrevs[linkrev] = True
37 37 else:
38 38 linkrev = '?'
39 39 msg = "%s: %s" % (linkrev, msg)
40 40 if filename:
41 41 msg = "%s@%s" % (filename, msg)
42 42 ui.warn(" " + msg + "\n")
43 43 errors[0] += 1
44 44
45 45 def exc(linkrev, msg, inst, filename=None):
46 46 if isinstance(inst, KeyboardInterrupt):
47 47 ui.warn(_("interrupted"))
48 48 raise
49 49 err(linkrev, "%s: %s" % (msg, inst), filename)
50 50
51 51 def warn(msg):
52 52 ui.warn(msg + "\n")
53 53 warnings[0] += 1
54 54
55 55 def checklog(obj, name):
56 56 if not len(obj) and (havecl or havemf):
57 57 err(0, _("empty or missing %s") % name)
58 58 return
59 59
60 60 d = obj.checksize()
61 61 if d[0]:
62 62 err(None, _("data length off by %d bytes") % d[0], name)
63 63 if d[1]:
64 64 err(None, _("index contains %d extra bytes") % d[1], name)
65 65
66 66 if obj.version != revlog.REVLOGV0:
67 67 if not revlogv1:
68 68 warn(_("warning: `%s' uses revlog format 1") % name)
69 69 elif revlogv1:
70 70 warn(_("warning: `%s' uses revlog format 0") % name)
71 71
72 72 def checkentry(obj, i, node, seen, linkrevs, f):
73 73 lr = obj.linkrev(obj.rev(node))
74 74 if lr < 0 or (havecl and lr not in linkrevs):
75 75 t = "unexpected"
76 76 if lr < 0 or lr >= len(cl):
77 77 t = "nonexistent"
78 78 err(None, _("rev %d point to %s changeset %d") % (i, t, lr), f)
79 79 if linkrevs:
80 80 warn(_(" (expected %s)") % " ".join(map(str,linkrevs)))
81 81 lr = None # can't be trusted
82 82
83 83 try:
84 84 p1, p2 = obj.parents(node)
85 85 if p1 not in seen and p1 != nullid:
86 86 err(lr, _("unknown parent 1 %s of %s") %
87 87 (short(p1), short(n)), f)
88 88 if p2 not in seen and p2 != nullid:
89 89 err(lr, _("unknown parent 2 %s of %s") %
90 90 (short(p2), short(p1)), f)
91 91 except Exception, inst:
92 92 exc(lr, _("checking parents of %s") % short(node), inst, f)
93 93
94 94 if node in seen:
95 95 err(lr, _("duplicate revision %d (%d)") % (i, seen[n]), f)
96 96 seen[n] = i
97 97 return lr
98 98
99 99 revlogv1 = cl.version != revlog.REVLOGV0
100 100 if ui.verbose or not revlogv1:
101 101 ui.status(_("repository uses revlog format %d\n") %
102 102 (revlogv1 and 1 or 0))
103 103
104 104 havecl = len(cl) > 0
105 105 havemf = len(mf) > 0
106 106
107 107 ui.status(_("checking changesets\n"))
108 108 seen = {}
109 109 checklog(cl, "changelog")
110 110 for i in repo:
111 111 n = cl.node(i)
112 112 checkentry(cl, i, n, seen, [i], "changelog")
113 113
114 114 try:
115 115 changes = cl.read(n)
116 116 mflinkrevs.setdefault(changes[0], []).append(i)
117 117 for f in changes[3]:
118 118 filelinkrevs.setdefault(f, []).append(i)
119 119 except Exception, inst:
120 120 exc(i, _("unpacking changeset %s") % short(n), inst)
121 121
122 122 ui.status(_("checking manifests\n"))
123 123 seen = {}
124 124 checklog(mf, "manifest")
125 125 for i in mf:
126 126 n = mf.node(i)
127 127 lr = checkentry(mf, i, n, seen, mflinkrevs.get(n, []), "manifest")
128 128 if n in mflinkrevs:
129 129 del mflinkrevs[n]
130 130
131 131 try:
132 132 for f, fn in mf.readdelta(n).iteritems():
133 133 if not f:
134 134 err(lr, _("file without name in manifest"))
135 135 elif f != "/dev/null":
136 136 fns = filenodes.setdefault(f, {})
137 137 if fn not in fns:
138 138 fns[fn] = i
139 139 except Exception, inst:
140 140 exc(lr, _("reading manifest delta %s") % short(n), inst)
141 141
142 142 ui.status(_("crosschecking files in changesets and manifests\n"))
143 143
144 144 if havemf:
145 145 for c, m in util.sort([(c, m) for m in mflinkrevs for c in mflinkrevs[m]]):
146 146 err(c, _("changeset refers to unknown manifest %s") % short(m))
147 147 del mflinkrevs
148 148
149 149 for f in util.sort(filelinkrevs):
150 150 if f not in filenodes:
151 151 lr = filelinkrevs[f][0]
152 152 err(lr, _("in changeset but not in manifest"), f)
153 153
154 154 if havecl:
155 155 for f in util.sort(filenodes):
156 156 if f not in filelinkrevs:
157 157 try:
158 158 fl = repo.file(f)
159 159 lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]])
160 160 except:
161 161 lr = None
162 162 err(lr, _("in manifest but not in changeset"), f)
163 163
164 164 ui.status(_("checking files\n"))
165 165
166 166 storefiles = {}
167 167 for f, f2, size in repo.store.datafiles():
168 168 if not f:
169 169 err(None, _("cannot decode filename '%s'") % f2)
170 170 elif size > 0:
171 171 storefiles[f] = True
172 172
173 173 files = util.sort(util.unique(filenodes.keys() + filelinkrevs.keys()))
174 174 for f in files:
175 fl = repo.file(f)
175 try:
176 fl = repo.file(f)
177 except error.RevlogError, e:
178 err(0, _("broken revlog! (%s)") % e, f)
179 continue
176 180
177 181 for ff in fl.files():
178 182 try:
179 183 del storefiles[ff]
180 184 except KeyError:
181 185 err(0, _("missing revlog!"), ff)
182 186
183 187 checklog(fl, f)
184 188 seen = {}
185 189 for i in fl:
186 190 revisions += 1
187 191 n = fl.node(i)
188 192 lr = checkentry(fl, i, n, seen, filelinkrevs.get(f, []), f)
189 193 if f in filenodes:
190 194 if havemf and n not in filenodes[f]:
191 195 err(lr, _("%s not in manifests") % (short(n)), f)
192 196 else:
193 197 del filenodes[f][n]
194 198
195 199 # verify contents
196 200 try:
197 201 t = fl.read(n)
198 202 rp = fl.renamed(n)
199 203 if len(t) != fl.size(i):
200 204 if len(fl.revision(n)) != fl.size(i):
201 205 err(lr, _("unpacked size is %s, %s expected") %
202 206 (len(t), fl.size(i)), f)
203 207 except Exception, inst:
204 208 exc(lr, _("unpacking %s") % short(n), inst, f)
205 209
206 210 # check renames
207 211 try:
208 212 if rp:
209 213 fl2 = repo.file(rp[0])
210 214 if not len(fl2):
211 215 err(lr, _("empty or missing copy source revlog %s:%s")
212 216 % (rp[0], short(rp[1])), f)
213 217 elif rp[1] == nullid:
214 218 warn(_("warning: %s@%s: copy source revision is nullid %s:%s")
215 219 % (f, lr, rp[0], short(rp[1])))
216 220 else:
217 221 rev = fl2.rev(rp[1])
218 222 except Exception, inst:
219 223 exc(lr, _("checking rename of %s") % short(n), inst, f)
220 224
221 225 # cross-check
222 226 if f in filenodes:
223 227 fns = [(mf.linkrev(l), n) for n,l in filenodes[f].iteritems()]
224 228 for lr, node in util.sort(fns):
225 229 err(lr, _("%s in manifests not found") % short(node), f)
226 230
227 231 for f in storefiles:
228 232 warn(_("warning: orphan revlog '%s'") % f)
229 233
230 234 ui.status(_("%d files, %d changesets, %d total revisions\n") %
231 235 (len(files), len(cl), revisions))
232 236 if warnings[0]:
233 237 ui.warn(_("%d warnings encountered!\n") % warnings[0])
234 238 if errors[0]:
235 239 ui.warn(_("%d integrity errors encountered!\n") % errors[0])
236 240 if badrevs:
237 241 ui.warn(_("(first damaged changeset appears to be %d)\n")
238 242 % min(badrevs))
239 243 return 1
General Comments 0
You need to be logged in to leave comments. Login now