mercurial/verify.py
author Matt Mackall <mpm@selenic.com>
Wed, 28 Oct 2009 13:17:03 -0500
changeset 9657 96c803e9018f
parent 9545 98feea5659d9
child 9690 b33d70849a20
permissions -rw-r--r--
verify: filter the candidate list for broken linkrevs
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
     1
# verify.py - repository integrity checking for Mercurial
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
     2
#
4635
63b9d2deed48 Updated copyright notices and add "and others" to "hg version"
Thomas Arendsen Hein <thomas@intevation.de>
parents: 4395
diff changeset
     3
# Copyright 2006, 2007 Matt Mackall <mpm@selenic.com>
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
     4
#
8225
46293a0c7e9f updated license to be explicit about GPL version 2
Martin Geisler <mg@lazybytes.net>
parents: 8209
diff changeset
     5
# This software may be used and distributed according to the terms of the
46293a0c7e9f updated license to be explicit about GPL version 2
Martin Geisler <mg@lazybytes.net>
parents: 8209
diff changeset
     6
# GNU General Public License version 2, incorporated herein by reference.
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
     7
6211
f89fd07fc51d Expand import * to allow Pyflakes to find problems
Joel Rosdahl <joel@rosdahl.net>
parents: 5541
diff changeset
     8
from node import nullid, short
3891
6b4127c7d52a Simplify i18n imports
Matt Mackall <mpm@selenic.com>
parents: 3744
diff changeset
     9
from i18n import _
7832
1fce19b9b011 verify: do not abort on fully corrupted revlog
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 7676
diff changeset
    10
import revlog, util, error
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    11
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    12
def verify(repo):
4915
97b734fb9c6f Use try/finally pattern to cleanup locks and transactions
Matt Mackall <mpm@selenic.com>
parents: 4635
diff changeset
    13
    lock = repo.lock()
97b734fb9c6f Use try/finally pattern to cleanup locks and transactions
Matt Mackall <mpm@selenic.com>
parents: 4635
diff changeset
    14
    try:
97b734fb9c6f Use try/finally pattern to cleanup locks and transactions
Matt Mackall <mpm@selenic.com>
parents: 4635
diff changeset
    15
        return _verify(repo)
97b734fb9c6f Use try/finally pattern to cleanup locks and transactions
Matt Mackall <mpm@selenic.com>
parents: 4635
diff changeset
    16
    finally:
8109
496ae1ea4698 switch lock releasing in the core from gc to explicit
Ronny Pfannschmidt <Ronny.Pfannschmidt@gmx.de>
parents: 7926
diff changeset
    17
        lock.release()
4915
97b734fb9c6f Use try/finally pattern to cleanup locks and transactions
Matt Mackall <mpm@selenic.com>
parents: 4635
diff changeset
    18
97b734fb9c6f Use try/finally pattern to cleanup locks and transactions
Matt Mackall <mpm@selenic.com>
parents: 4635
diff changeset
    19
def _verify(repo):
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    20
    mflinkrevs = {}
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    21
    filelinkrevs = {}
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    22
    filenodes = {}
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    23
    revisions = 0
8466
afb3e504b558 verify: use set instead of dict
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8394
diff changeset
    24
    badrevs = set()
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    25
    errors = [0]
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    26
    warnings = [0]
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
    27
    ui = repo.ui
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
    28
    cl = repo.changelog
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
    29
    mf = repo.manifest
9657
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    30
    lrugetctx = util.lrucachefunc(repo.changectx)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    31
7141
8d1bdaf842de issue 1144: prevent traceback on verify of bundles
John Mulligan <phlogistonjohn@asynchrono.us>
parents: 7004
diff changeset
    32
    if not repo.cancopy():
8d1bdaf842de issue 1144: prevent traceback on verify of bundles
John Mulligan <phlogistonjohn@asynchrono.us>
parents: 7004
diff changeset
    33
        raise util.Abort(_("cannot verify bundle or remote repos"))
8d1bdaf842de issue 1144: prevent traceback on verify of bundles
John Mulligan <phlogistonjohn@asynchrono.us>
parents: 7004
diff changeset
    34
5313
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    35
    def err(linkrev, msg, filename=None):
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    36
        if linkrev != None:
8466
afb3e504b558 verify: use set instead of dict
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8394
diff changeset
    37
            badrevs.add(linkrev)
5313
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    38
        else:
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    39
            linkrev = '?'
5313
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    40
        msg = "%s: %s" % (linkrev, msg)
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    41
        if filename:
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    42
            msg = "%s@%s" % (filename, msg)
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
    43
        ui.warn(" " + msg + "\n")
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    44
        errors[0] += 1
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    45
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    46
    def exc(linkrev, msg, inst, filename=None):
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    47
        if isinstance(inst, KeyboardInterrupt):
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    48
            ui.warn(_("interrupted"))
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    49
            raise
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    50
        err(linkrev, "%s: %s" % (msg, inst), filename)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    51
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    52
    def warn(msg):
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
    53
        ui.warn(msg + "\n")
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    54
        warnings[0] += 1
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    55
8292
29540554def8 verify: reference the correct linkrev when a filelog is missing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8291
diff changeset
    56
    def checklog(obj, name, linkrev):
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    57
        if not len(obj) and (havecl or havemf):
8292
29540554def8 verify: reference the correct linkrev when a filelog is missing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8291
diff changeset
    58
            err(linkrev, _("empty or missing %s") % name)
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    59
            return
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    60
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    61
        d = obj.checksize()
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    62
        if d[0]:
5313
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    63
            err(None, _("data length off by %d bytes") % d[0], name)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    64
        if d[1]:
5313
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
    65
            err(None, _("index contains %d extra bytes") % d[1], name)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    66
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    67
        if obj.version != revlog.REVLOGV0:
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    68
            if not revlogv1:
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    69
                warn(_("warning: `%s' uses revlog format 1") % name)
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    70
        elif revlogv1:
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    71
            warn(_("warning: `%s' uses revlog format 0") % name)
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    72
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    73
    def checkentry(obj, i, node, seen, linkrevs, f):
7361
9fe97eea5510 linkrev: take a revision number rather than a hash
Matt Mackall <mpm@selenic.com>
parents: 7141
diff changeset
    74
        lr = obj.linkrev(obj.rev(node))
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    75
        if lr < 0 or (havecl and lr not in linkrevs):
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    76
            if lr < 0 or lr >= len(cl):
7926
edf2d83a11aa verify, i18n: fix unmarked strings
Wagner Bruna <wbruna@yahoo.com>
parents: 7874
diff changeset
    77
                msg = _("rev %d points to nonexistent changeset %d")
edf2d83a11aa verify, i18n: fix unmarked strings
Wagner Bruna <wbruna@yahoo.com>
parents: 7874
diff changeset
    78
            else:
edf2d83a11aa verify, i18n: fix unmarked strings
Wagner Bruna <wbruna@yahoo.com>
parents: 7874
diff changeset
    79
                msg = _("rev %d points to unexpected changeset %d")
edf2d83a11aa verify, i18n: fix unmarked strings
Wagner Bruna <wbruna@yahoo.com>
parents: 7874
diff changeset
    80
            err(None, msg % (i, lr), f)
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    81
            if linkrevs:
9657
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    82
                if f and len(linkrevs) > 1:
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    83
                    try:
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    84
                        # attempt to filter down to real linkrevs
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    85
                        linkrevs = [l for l in linkrevs
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    86
                                    if lrugetctx(l)[f].filenode() == node]
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    87
                    except:
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
    88
                        pass
9198
061eeb602354 coding style: use a space after comma
Martin Geisler <mg@lazybytes.net>
parents: 9033
diff changeset
    89
                warn(_(" (expected %s)") % " ".join(map(str, linkrevs)))
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    90
            lr = None # can't be trusted
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    91
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    92
        try:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    93
            p1, p2 = obj.parents(node)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    94
            if p1 not in seen and p1 != nullid:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    95
                err(lr, _("unknown parent 1 %s of %s") %
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    96
                    (short(p1), short(n)), f)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    97
            if p2 not in seen and p2 != nullid:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    98
                err(lr, _("unknown parent 2 %s of %s") %
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
    99
                    (short(p2), short(p1)), f)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   100
        except Exception, inst:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   101
            exc(lr, _("checking parents of %s") % short(node), inst, f)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   102
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   103
        if node in seen:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   104
            err(lr, _("duplicate revision %d (%d)") % (i, seen[n]), f)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   105
        seen[n] = i
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   106
        return lr
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   107
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   108
    revlogv1 = cl.version != revlog.REVLOGV0
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   109
    if ui.verbose or not revlogv1:
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   110
        ui.status(_("repository uses revlog format %d\n") %
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   111
                       (revlogv1 and 1 or 0))
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   112
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   113
    havecl = len(cl) > 0
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   114
    havemf = len(mf) > 0
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   115
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   116
    ui.status(_("checking changesets\n"))
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   117
    seen = {}
8292
29540554def8 verify: reference the correct linkrev when a filelog is missing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8291
diff changeset
   118
    checklog(cl, "changelog", 0)
6750
fb42030d79d6 add __len__ and __iter__ methods to repo and revlog
Matt Mackall <mpm@selenic.com>
parents: 6534
diff changeset
   119
    for i in repo:
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   120
        n = cl.node(i)
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   121
        checkentry(cl, i, n, seen, [i], "changelog")
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   122
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   123
        try:
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   124
            changes = cl.read(n)
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   125
            mflinkrevs.setdefault(changes[0], []).append(i)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   126
            for f in changes[3]:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   127
                filelinkrevs.setdefault(f, []).append(i)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   128
        except Exception, inst:
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   129
            exc(i, _("unpacking changeset %s") % short(n), inst)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   130
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   131
    ui.status(_("checking manifests\n"))
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   132
    seen = {}
8292
29540554def8 verify: reference the correct linkrev when a filelog is missing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8291
diff changeset
   133
    checklog(mf, "manifest", 0)
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   134
    for i in mf:
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   135
        n = mf.node(i)
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   136
        lr = checkentry(mf, i, n, seen, mflinkrevs.get(n, []), "manifest")
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   137
        if n in mflinkrevs:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   138
            del mflinkrevs[n]
8394
850b5a7c210d verify: detect manifest revs not in any changeset
Peter Arrenbrecht <peter.arrenbrecht@gmail.com>
parents: 8292
diff changeset
   139
        else:
850b5a7c210d verify: detect manifest revs not in any changeset
Peter Arrenbrecht <peter.arrenbrecht@gmail.com>
parents: 8292
diff changeset
   140
            err(lr, _("%s not in changesets") % short(n), "manifest")
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   141
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   142
        try:
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   143
            for f, fn in mf.readdelta(n).iteritems():
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   144
                if not f:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   145
                    err(lr, _("file without name in manifest"))
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   146
                elif f != "/dev/null":
9657
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
   147
                    filenodes.setdefault(f, {}).setdefault(fn, lr)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   148
        except Exception, inst:
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   149
            exc(lr, _("reading manifest delta %s") % short(n), inst)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   150
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   151
    ui.status(_("crosschecking files in changesets and manifests\n"))
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   152
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   153
    if havemf:
8209
a1a5a57efe90 replace util.sort with sorted built-in
Matt Mackall <mpm@selenic.com>
parents: 8164
diff changeset
   154
        for c,m in sorted([(c, m) for m in mflinkrevs for c in mflinkrevs[m]]):
5541
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   155
            err(c, _("changeset refers to unknown manifest %s") % short(m))
9033
98a5652bfed9 verify: fix scope issues with del statement
Alejandro Santos <alejolp@alejolp.com>
parents: 8993
diff changeset
   156
        mflinkrevs = None # del is bad here due to scope issues
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   157
8209
a1a5a57efe90 replace util.sort with sorted built-in
Matt Mackall <mpm@selenic.com>
parents: 8164
diff changeset
   158
        for f in sorted(filelinkrevs):
5541
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   159
            if f not in filenodes:
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   160
                lr = filelinkrevs[f][0]
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   161
                err(lr, _("in changeset but not in manifest"), f)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   162
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   163
    if havecl:
8209
a1a5a57efe90 replace util.sort with sorted built-in
Matt Mackall <mpm@selenic.com>
parents: 8164
diff changeset
   164
        for f in sorted(filenodes):
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   165
            if f not in filelinkrevs:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   166
                try:
7361
9fe97eea5510 linkrev: take a revision number rather than a hash
Matt Mackall <mpm@selenic.com>
parents: 7141
diff changeset
   167
                    fl = repo.file(f)
9fe97eea5510 linkrev: take a revision number rather than a hash
Matt Mackall <mpm@selenic.com>
parents: 7141
diff changeset
   168
                    lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]])
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   169
                except:
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   170
                    lr = None
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   171
                err(lr, _("in manifest but not in changeset"), f)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   172
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   173
    ui.status(_("checking files\n"))
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   174
8466
afb3e504b558 verify: use set instead of dict
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8394
diff changeset
   175
    storefiles = set()
6900
def492d1b592 store: change handling of decoding errors
Matt Mackall <mpm@selenic.com>
parents: 6892
diff changeset
   176
    for f, f2, size in repo.store.datafiles():
def492d1b592 store: change handling of decoding errors
Matt Mackall <mpm@selenic.com>
parents: 6892
diff changeset
   177
        if not f:
def492d1b592 store: change handling of decoding errors
Matt Mackall <mpm@selenic.com>
parents: 6892
diff changeset
   178
            err(None, _("cannot decode filename '%s'") % f2)
def492d1b592 store: change handling of decoding errors
Matt Mackall <mpm@selenic.com>
parents: 6892
diff changeset
   179
        elif size > 0:
8466
afb3e504b558 verify: use set instead of dict
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8394
diff changeset
   180
            storefiles.add(f)
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   181
8209
a1a5a57efe90 replace util.sort with sorted built-in
Matt Mackall <mpm@selenic.com>
parents: 8164
diff changeset
   182
    files = sorted(set(filenodes) | set(filelinkrevs))
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   183
    for f in files:
8291
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   184
        try:
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   185
            linkrevs = filelinkrevs[f]
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   186
        except KeyError:
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   187
            # in manifest but not in changelog
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   188
            linkrevs = []
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   189
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   190
        if linkrevs:
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   191
            lr = linkrevs[0]
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   192
        else:
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   193
            lr = None
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   194
7832
1fce19b9b011 verify: do not abort on fully corrupted revlog
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 7676
diff changeset
   195
        try:
1fce19b9b011 verify: do not abort on fully corrupted revlog
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 7676
diff changeset
   196
            fl = repo.file(f)
1fce19b9b011 verify: do not abort on fully corrupted revlog
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 7676
diff changeset
   197
        except error.RevlogError, e:
7833
794def2fe232 verify: find correct first corrupted cset for missing/corrupted revlogs
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 7832
diff changeset
   198
            err(lr, _("broken revlog! (%s)") % e, f)
7832
1fce19b9b011 verify: do not abort on fully corrupted revlog
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 7676
diff changeset
   199
            continue
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   200
6900
def492d1b592 store: change handling of decoding errors
Matt Mackall <mpm@selenic.com>
parents: 6892
diff changeset
   201
        for ff in fl.files():
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   202
            try:
8466
afb3e504b558 verify: use set instead of dict
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8394
diff changeset
   203
                storefiles.remove(ff)
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   204
            except KeyError:
7833
794def2fe232 verify: find correct first corrupted cset for missing/corrupted revlogs
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 7832
diff changeset
   205
                err(lr, _("missing revlog!"), ff)
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   206
8292
29540554def8 verify: reference the correct linkrev when a filelog is missing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 8291
diff changeset
   207
        checklog(fl, f, lr)
5313
29be4228303b verify: report first bad changeset
Matt Mackall <mpm@selenic.com>
parents: 5179
diff changeset
   208
        seen = {}
6750
fb42030d79d6 add __len__ and __iter__ methods to repo and revlog
Matt Mackall <mpm@selenic.com>
parents: 6534
diff changeset
   209
        for i in fl:
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   210
            revisions += 1
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   211
            n = fl.node(i)
8291
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   212
            lr = checkentry(fl, i, n, seen, linkrevs, f)
5541
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   213
            if f in filenodes:
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   214
                if havemf and n not in filenodes[f]:
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   215
                    err(lr, _("%s not in manifests") % (short(n)), f)
5541
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   216
                else:
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   217
                    del filenodes[f][n]
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   218
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   219
            # verify contents
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   220
            try:
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   221
                t = fl.read(n)
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   222
                rp = fl.renamed(n)
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   223
                if len(t) != fl.size(i):
7675
011e69b96c69 verify: don't trip over binary files starting with 01 0a
Matt Mackall <mpm@selenic.com>
parents: 7361
diff changeset
   224
                    if len(fl.revision(n)) != fl.size(i):
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   225
                        err(lr, _("unpacked size is %s, %s expected") %
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   226
                            (len(t), fl.size(i)), f)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   227
            except Exception, inst:
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   228
                exc(lr, _("unpacking %s") % short(n), inst, f)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   229
3744
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   230
            # check renames
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   231
            try:
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   232
                if rp:
9545
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   233
                    if lr is not None and ui.verbose:
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   234
                        ctx = lrugetctx(lr)
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   235
                        found = False
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   236
                        for pctx in ctx.parents():
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   237
                            if rp[0] in pctx:
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   238
                                found = True
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   239
                                break
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   240
                        if not found:
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   241
                            warn(_("warning: copy source of '%s' not"
98feea5659d9 verify: detect file copy sources not in parents with --verbose
Patrick Mezard <pmezard@gmail.com>
parents: 9198
diff changeset
   242
                                   " in parents of %s") % (f, ctx))
3744
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   243
                    fl2 = repo.file(rp[0])
6750
fb42030d79d6 add __len__ and __iter__ methods to repo and revlog
Matt Mackall <mpm@selenic.com>
parents: 6534
diff changeset
   244
                    if not len(fl2):
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   245
                        err(lr, _("empty or missing copy source revlog %s:%s")
6534
9b35a9f34675 verify: check copy source revlog and nodeid
Patrick Mezard <pmezard@gmail.com>
parents: 6211
diff changeset
   246
                            % (rp[0], short(rp[1])), f)
9b35a9f34675 verify: check copy source revlog and nodeid
Patrick Mezard <pmezard@gmail.com>
parents: 6211
diff changeset
   247
                    elif rp[1] == nullid:
8993
46441934c585 verify: demote warning about nullid in copy to note
Matt Mackall <mpm@selenic.com>
parents: 8466
diff changeset
   248
                        ui.note(_("warning: %s@%s: copy source"
46441934c585 verify: demote warning about nullid in copy to note
Matt Mackall <mpm@selenic.com>
parents: 8466
diff changeset
   249
                                  " revision is nullid %s:%s\n")
7004
90227c42b5f6 c0bd7d8b69ef uses err() instead of warn() but prototype doesn't match
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 6900
diff changeset
   250
                            % (f, lr, rp[0], short(rp[1])))
6534
9b35a9f34675 verify: check copy source revlog and nodeid
Patrick Mezard <pmezard@gmail.com>
parents: 6211
diff changeset
   251
                    else:
7874
d812029cda85 cleanup: drop variables for unused return values
Peter Arrenbrecht <peter.arrenbrecht@gmail.com>
parents: 7833
diff changeset
   252
                        fl2.rev(rp[1])
3744
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   253
            except Exception, inst:
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   254
                exc(lr, _("checking rename of %s") % short(n), inst, f)
3744
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   255
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   256
        # cross-check
5541
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   257
        if f in filenodes:
9657
96c803e9018f verify: filter the candidate list for broken linkrevs
Matt Mackall <mpm@selenic.com>
parents: 9545
diff changeset
   258
            fns = [(lr, n) for n,lr in filenodes[f].iteritems()]
8209
a1a5a57efe90 replace util.sort with sorted built-in
Matt Mackall <mpm@selenic.com>
parents: 8164
diff changeset
   259
            for lr, node in sorted(fns):
5541
ceaa752fa316 verify: improve handling of empty or missing files
Matt Mackall <mpm@selenic.com>
parents: 5313
diff changeset
   260
                err(lr, _("%s in manifests not found") % short(node), f)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   261
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   262
    for f in storefiles:
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   263
        warn(_("warning: orphan revlog '%s'") % f)
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   264
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   265
    ui.status(_("%d files, %d changesets, %d total revisions\n") %
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   266
                   (len(files), len(cl), revisions))
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   267
    if warnings[0]:
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   268
        ui.warn(_("%d warnings encountered!\n") % warnings[0])
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   269
    if errors[0]:
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   270
        ui.warn(_("%d integrity errors encountered!\n") % errors[0])
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   271
        if badrevs:
6751
7424a75f919a verify: add some local variables
Matt Mackall <mpm@selenic.com>
parents: 6750
diff changeset
   272
            ui.warn(_("(first damaged changeset appears to be %d)\n")
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   273
                    % min(badrevs))
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   274
        return 1