mercurial/verify.py
author Martin von Zweigbergk <martinvonz@google.com>
Tue, 06 Feb 2018 08:49:37 -0800
changeset 36339 a4d41ba4ad23
parent 36185 acc8e6e52af6
child 36577 d85ef895d5f6
permissions -rw-r--r--
verify: don't reimplement any() Differential Revision: https://phab.mercurial-scm.org/D2372
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
10263
25e572394f5c Update license to GPLv2+
Matt Mackall <mpm@selenic.com>
parents: 9690
diff changeset
     6
# GNU General Public License version 2 or any later version.
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
     7
25991
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
     8
from __future__ import absolute_import
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
     9
17860
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    10
import os
25991
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    11
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    12
from .i18n import _
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    13
from .node import (
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    14
    nullid,
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    15
    short,
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    16
)
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    17
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    18
from . import (
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    19
    error,
35585
35fb3367f72d py3: use pycompat.bytestr() instead of str()
Pulkit Goyal <7895pulkit@gmail.com>
parents: 33499
diff changeset
    20
    pycompat,
25991
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    21
    revlog,
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
    22
    scmutil,
25991
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    23
    util,
d21d1774c73b verify: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents: 25846
diff changeset
    24
)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    25
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
    26
def verify(repo):
27849
900d36a3e4dd with: use context manager in verify
Bryan O'Sullivan <bryano@fb.com>
parents: 27695
diff changeset
    27
    with repo.lock():
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    28
        return verifier(repo).verify()
4915
97b734fb9c6f Use try/finally pattern to cleanup locks and transactions
Matt Mackall <mpm@selenic.com>
parents: 4635
diff changeset
    29
17860
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    30
def _normpath(f):
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    31
    # under hg < 2.4, convert didn't sanitize paths properly, so a
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    32
    # converted repo may contain repeated slashes
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    33
    while '//' in f:
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    34
        f = f.replace('//', '/')
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    35
    return f
a45b33f12627 verify: fix all doubled-slash sites (issue3665)
Bryan O'Sullivan <bryano@fb.com>
parents: 17851
diff changeset
    36
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
    37
class verifier(object):
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
    38
    # The match argument is always None in hg core, but e.g. the narrowhg
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
    39
    # extension will pass in a matcher here.
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
    40
    def __init__(self, repo, match=None):
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    41
        self.repo = repo.unfiltered()
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    42
        self.ui = repo.ui
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
    43
        self.match = match or scmutil.matchall(repo)
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    44
        self.badrevs = set()
27453
8462d7f2c4fe verify: clean up weird error/warning lists
Matt Mackall <mpm@selenic.com>
parents: 27450
diff changeset
    45
        self.errors = 0
8462d7f2c4fe verify: clean up weird error/warning lists
Matt Mackall <mpm@selenic.com>
parents: 27450
diff changeset
    46
        self.warnings = 0
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    47
        self.havecl = len(repo.changelog) > 0
30375
11b8b740d54a manifest: remove last uses of repo.manifest
Durham Goode <durham@fb.com>
parents: 30295
diff changeset
    48
        self.havemf = len(repo.manifestlog._revlog) > 0
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    49
        self.revlogv1 = repo.changelog.version != revlog.REVLOGV0
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    50
        self.lrugetctx = util.lrucachefunc(repo.changectx)
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    51
        self.refersmf = False
27445
cc178057ab49 verify: move fncachewarned up to a class variable
Durham Goode <durham@fb.com>
parents: 27444
diff changeset
    52
        self.fncachewarned = False
32288
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
    53
        # developer config: verify.skipflags
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
    54
        self.skipflags = repo.ui.configint('verify', 'skipflags')
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
    55
27446
6b2c1a1871a6 verify: move warn() to a class level function
Durham Goode <durham@fb.com>
parents: 27445
diff changeset
    56
    def warn(self, msg):
6b2c1a1871a6 verify: move warn() to a class level function
Durham Goode <durham@fb.com>
parents: 27445
diff changeset
    57
        self.ui.warn(msg + "\n")
27453
8462d7f2c4fe verify: clean up weird error/warning lists
Matt Mackall <mpm@selenic.com>
parents: 27450
diff changeset
    58
        self.warnings += 1
27446
6b2c1a1871a6 verify: move warn() to a class level function
Durham Goode <durham@fb.com>
parents: 27445
diff changeset
    59
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    60
    def err(self, linkrev, msg, filename=None):
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    61
        if linkrev is not None:
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    62
            self.badrevs.add(linkrev)
36185
acc8e6e52af6 py3: use "%d" to convert integers to bytes
Pulkit Goyal <7895pulkit@gmail.com>
parents: 35585
diff changeset
    63
            linkrev = "%d" % linkrev
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    64
        else:
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    65
            linkrev = '?'
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    66
        msg = "%s: %s" % (linkrev, msg)
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    67
        if filename:
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    68
            msg = "%s@%s" % (filename, msg)
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    69
        self.ui.warn(" " + msg + "\n")
27453
8462d7f2c4fe verify: clean up weird error/warning lists
Matt Mackall <mpm@selenic.com>
parents: 27450
diff changeset
    70
        self.errors += 1
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
    71
27448
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
    72
    def exc(self, linkrev, msg, inst, filename=None):
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
    73
        if not str(inst):
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
    74
            inst = repr(inst)
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
    75
        self.err(linkrev, "%s: %s" % (msg, inst), filename)
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
    76
27642
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    77
    def checklog(self, obj, name, linkrev):
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    78
        if not len(obj) and (self.havecl or self.havemf):
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    79
            self.err(linkrev, _("empty or missing %s") % name)
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    80
            return
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    81
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    82
        d = obj.checksize()
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    83
        if d[0]:
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    84
            self.err(None, _("data length off by %d bytes") % d[0], name)
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    85
        if d[1]:
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    86
            self.err(None, _("index contains %d extra bytes") % d[1], name)
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    87
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    88
        if obj.version != revlog.REVLOGV0:
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    89
            if not self.revlogv1:
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    90
                self.warn(_("warning: `%s' uses revlog format 1") % name)
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    91
        elif self.revlogv1:
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    92
            self.warn(_("warning: `%s' uses revlog format 0") % name)
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
    93
27643
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
    94
    def checkentry(self, obj, i, node, seen, linkrevs, f):
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
    95
        lr = obj.linkrev(obj.rev(node))
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
    96
        if lr < 0 or (self.havecl and lr not in linkrevs):
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
    97
            if lr < 0 or lr >= len(self.repo.changelog):
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
    98
                msg = _("rev %d points to nonexistent changeset %d")
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
    99
            else:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   100
                msg = _("rev %d points to unexpected changeset %d")
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   101
            self.err(None, msg % (i, lr), f)
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   102
            if linkrevs:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   103
                if f and len(linkrevs) > 1:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   104
                    try:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   105
                        # attempt to filter down to real linkrevs
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   106
                        linkrevs = [l for l in linkrevs
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   107
                                    if self.lrugetctx(l)[f].filenode() == node]
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   108
                    except Exception:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   109
                        pass
35585
35fb3367f72d py3: use pycompat.bytestr() instead of str()
Pulkit Goyal <7895pulkit@gmail.com>
parents: 33499
diff changeset
   110
                self.warn(_(" (expected %s)") % " ".join
35fb3367f72d py3: use pycompat.bytestr() instead of str()
Pulkit Goyal <7895pulkit@gmail.com>
parents: 33499
diff changeset
   111
                          (map(pycompat.bytestr, linkrevs)))
27643
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   112
            lr = None # can't be trusted
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   113
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   114
        try:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   115
            p1, p2 = obj.parents(node)
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   116
            if p1 not in seen and p1 != nullid:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   117
                self.err(lr, _("unknown parent 1 %s of %s") %
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   118
                    (short(p1), short(node)), f)
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   119
            if p2 not in seen and p2 != nullid:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   120
                self.err(lr, _("unknown parent 2 %s of %s") %
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   121
                    (short(p2), short(node)), f)
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   122
        except Exception as inst:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   123
            self.exc(lr, _("checking parents of %s") % short(node), inst, f)
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   124
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   125
        if node in seen:
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   126
            self.err(lr, _("duplicate revision %d (%d)") % (i, seen[node]), f)
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   127
        seen[node] = i
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   128
        return lr
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   129
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
   130
    def verify(self):
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
   131
        repo = self.repo
27648
e72e669dd51f verify: get rid of some unnecessary local variables
Durham Goode <durham@fb.com>
parents: 27647
diff changeset
   132
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   133
        ui = repo.ui
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   134
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   135
        if not repo.url().startswith('file:'):
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   136
            raise error.Abort(_("cannot verify bundle or remote repos"))
6752
e79a8f36c2a5 verify: lots of refactoring
Matt Mackall <mpm@selenic.com>
parents: 6751
diff changeset
   137
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   138
        if os.path.exists(repo.sjoin("journal")):
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   139
            ui.warn(_("abandoned transaction found - run hg recover\n"))
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   140
27648
e72e669dd51f verify: get rid of some unnecessary local variables
Durham Goode <durham@fb.com>
parents: 27647
diff changeset
   141
        if ui.verbose or not self.revlogv1:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   142
            ui.status(_("repository uses revlog format %d\n") %
27648
e72e669dd51f verify: get rid of some unnecessary local variables
Durham Goode <durham@fb.com>
parents: 27647
diff changeset
   143
                           (self.revlogv1 and 1 or 0))
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   144
27695
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   145
        mflinkrevs, filelinkrevs = self._verifychangelog()
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   146
27695
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   147
        filenodes = self._verifymanifest(mflinkrevs)
28111
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   148
        del mflinkrevs
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   149
28111
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   150
        self._crosscheckfiles(filelinkrevs, filenodes)
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   151
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   152
        totalfiles, filerevisions = self._verifyfiles(filenodes, filelinkrevs)
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   153
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   154
        ui.status(_("%d files, %d changesets, %d total revisions\n") %
27648
e72e669dd51f verify: get rid of some unnecessary local variables
Durham Goode <durham@fb.com>
parents: 27647
diff changeset
   155
                       (totalfiles, len(repo.changelog), filerevisions))
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   156
        if self.warnings:
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   157
            ui.warn(_("%d warnings encountered!\n") % self.warnings)
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   158
        if self.fncachewarned:
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   159
            ui.warn(_('hint: run "hg debugrebuildfncache" to recover from '
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   160
                      'corrupt fncache\n'))
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   161
        if self.errors:
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   162
            ui.warn(_("%d integrity errors encountered!\n") % self.errors)
27648
e72e669dd51f verify: get rid of some unnecessary local variables
Durham Goode <durham@fb.com>
parents: 27647
diff changeset
   163
            if self.badrevs:
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   164
                ui.warn(_("(first damaged changeset appears to be %d)\n")
27648
e72e669dd51f verify: get rid of some unnecessary local variables
Durham Goode <durham@fb.com>
parents: 27647
diff changeset
   165
                        % min(self.badrevs))
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   166
            return 1
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   167
27695
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   168
    def _verifychangelog(self):
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   169
        ui = self.ui
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   170
        repo = self.repo
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
   171
        match = self.match
27647
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   172
        cl = repo.changelog
2c2858f3c1bb verify: move changelog verificaiton to its own function
Durham Goode <durham@fb.com>
parents: 27646
diff changeset
   173
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   174
        ui.status(_("checking changesets\n"))
27695
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   175
        mflinkrevs = {}
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   176
        filelinkrevs = {}
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   177
        seen = {}
27642
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
   178
        self.checklog(cl, "changelog", 0)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   179
        total = len(repo)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   180
        for i in repo:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   181
            ui.progress(_('checking'), i, total=total, unit=_('changesets'))
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   182
            n = cl.node(i)
27643
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   183
            self.checkentry(cl, i, n, seen, [i], "changelog")
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   184
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   185
            try:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   186
                changes = cl.read(n)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   187
                if changes[0] != nullid:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   188
                    mflinkrevs.setdefault(changes[0], []).append(i)
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
   189
                    self.refersmf = True
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   190
                for f in changes[3]:
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
   191
                    if match(f):
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   192
                        filelinkrevs.setdefault(_normpath(f), []).append(i)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   193
            except Exception as inst:
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
   194
                self.refersmf = True
27448
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
   195
                self.exc(i, _("unpacking changeset %s") % short(n), inst)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   196
        ui.progress(_('checking'), None)
27695
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   197
        return mflinkrevs, filelinkrevs
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   198
28205
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   199
    def _verifymanifest(self, mflinkrevs, dir="", storefiles=None,
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   200
                        progress=None):
27646
8f43793382c6 verify: move manifest verification to its own function
Durham Goode <durham@fb.com>
parents: 27645
diff changeset
   201
        repo = self.repo
8f43793382c6 verify: move manifest verification to its own function
Durham Goode <durham@fb.com>
parents: 27645
diff changeset
   202
        ui = self.ui
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
   203
        match = self.match
30295
f65faa4422c8 manifest: remove manifest.readshallowdelta
Durham Goode <durham@fb.com>
parents: 28467
diff changeset
   204
        mfl = self.repo.manifestlog
f65faa4422c8 manifest: remove manifest.readshallowdelta
Durham Goode <durham@fb.com>
parents: 28467
diff changeset
   205
        mf = mfl._revlog.dirlog(dir)
27646
8f43793382c6 verify: move manifest verification to its own function
Durham Goode <durham@fb.com>
parents: 27645
diff changeset
   206
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   207
        if not dir:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   208
            self.ui.status(_("checking manifests\n"))
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   209
27695
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   210
        filenodes = {}
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   211
        subdirnodes = {}
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   212
        seen = {}
28115
bd279da57e4b verify: extract "manifest" constant into variable
Martin von Zweigbergk <martinvonz@google.com>
parents: 28114
diff changeset
   213
        label = "manifest"
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   214
        if dir:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   215
            label = dir
28204
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   216
            revlogfiles = mf.files()
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   217
            storefiles.difference_update(revlogfiles)
28205
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   218
            if progress: # should be true since we're in a subdirectory
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   219
                progress()
27444
6647401858ab verify: move widely used variables into class members
Durham Goode <durham@fb.com>
parents: 27443
diff changeset
   220
        if self.refersmf:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   221
            # Do not check manifest if there are only changelog entries with
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   222
            # null manifests.
28115
bd279da57e4b verify: extract "manifest" constant into variable
Martin von Zweigbergk <martinvonz@google.com>
parents: 28114
diff changeset
   223
            self.checklog(mf, label, 0)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   224
        total = len(mf)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   225
        for i in mf:
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   226
            if not dir:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   227
                ui.progress(_('checking'), i, total=total, unit=_('manifests'))
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   228
            n = mf.node(i)
28115
bd279da57e4b verify: extract "manifest" constant into variable
Martin von Zweigbergk <martinvonz@google.com>
parents: 28114
diff changeset
   229
            lr = self.checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   230
            if n in mflinkrevs:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   231
                del mflinkrevs[n]
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   232
            elif dir:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   233
                self.err(lr, _("%s not in parent-directory manifest") %
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   234
                         short(n), label)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   235
            else:
28115
bd279da57e4b verify: extract "manifest" constant into variable
Martin von Zweigbergk <martinvonz@google.com>
parents: 28114
diff changeset
   236
                self.err(lr, _("%s not in changesets") % short(n), label)
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   237
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   238
            try:
30295
f65faa4422c8 manifest: remove manifest.readshallowdelta
Durham Goode <durham@fb.com>
parents: 28467
diff changeset
   239
                mfdelta = mfl.get(dir, n).readdelta(shallow=True)
f65faa4422c8 manifest: remove manifest.readshallowdelta
Durham Goode <durham@fb.com>
parents: 28467
diff changeset
   240
                for f, fn, fl in mfdelta.iterentries():
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   241
                    if not f:
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   242
                        self.err(lr, _("entry without name in manifest"))
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   243
                    elif f == "/dev/null":  # ignore this in very old repos
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   244
                        continue
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   245
                    fullpath = dir + _normpath(f)
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   246
                    if fl == 't':
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
   247
                        if not match.visitdir(fullpath):
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
   248
                            continue
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   249
                        subdirnodes.setdefault(fullpath + '/', {}).setdefault(
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   250
                            fn, []).append(lr)
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   251
                    else:
30866
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
   252
                        if not match(fullpath):
5249b6470de9 verify: replace _validpath() by matcher
Martin von Zweigbergk <martinvonz@google.com>
parents: 30393
diff changeset
   253
                            continue
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   254
                        filenodes.setdefault(fullpath, {}).setdefault(fn, lr)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   255
            except Exception as inst:
28115
bd279da57e4b verify: extract "manifest" constant into variable
Martin von Zweigbergk <martinvonz@google.com>
parents: 28114
diff changeset
   256
                self.exc(lr, _("reading delta %s") % short(n), inst, label)
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   257
        if not dir:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   258
            ui.progress(_('checking'), None)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   259
28111
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   260
        if self.havemf:
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   261
            for c, m in sorted([(c, m) for m in mflinkrevs
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   262
                        for c in mflinkrevs[m]]):
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   263
                if dir:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   264
                    self.err(c, _("parent-directory manifest refers to unknown "
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   265
                                  "revision %s") % short(m), label)
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   266
                else:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   267
                    self.err(c, _("changeset refers to unknown revision %s") %
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   268
                             short(m), label)
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   269
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   270
        if not dir and subdirnodes:
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   271
            self.ui.status(_("checking directory manifests\n"))
28204
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   272
            storefiles = set()
28205
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   273
            subdirs = set()
28204
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   274
            revlogv1 = self.revlogv1
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   275
            for f, f2, size in repo.store.datafiles():
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   276
                if not f:
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   277
                    self.err(None, _("cannot decode filename '%s'") % f2)
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   278
                elif (size > 0 or not revlogv1) and f.startswith('meta/'):
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   279
                    storefiles.add(_normpath(f))
28205
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   280
                    subdirs.add(os.path.dirname(f))
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   281
            subdircount = len(subdirs)
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   282
            currentsubdir = [0]
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   283
            def progress():
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   284
                currentsubdir[0] += 1
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   285
                ui.progress(_('checking'), currentsubdir[0], total=subdircount,
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   286
                            unit=_('manifests'))
28204
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   287
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   288
        for subdir, linkrevs in subdirnodes.iteritems():
28205
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   289
            subdirfilenodes = self._verifymanifest(linkrevs, subdir, storefiles,
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   290
                                                   progress)
28203
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   291
            for f, onefilenodes in subdirfilenodes.iteritems():
7297e9e13a8a verify: check directory manifests
Martin von Zweigbergk <martinvonz@google.com>
parents: 28115
diff changeset
   292
                filenodes.setdefault(f, {}).update(onefilenodes)
28111
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   293
28204
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   294
        if not dir and subdirnodes:
28205
53f42c8d5f71 verify: show progress while verifying dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28204
diff changeset
   295
            ui.progress(_('checking'), None)
28204
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   296
            for f in sorted(storefiles):
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   297
                self.warn(_("warning: orphan revlog '%s'") % f)
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   298
27695
fb0cc863d172 verify: replace "output parameters" by return values
Martin von Zweigbergk <martinvonz@google.com>
parents: 27648
diff changeset
   299
        return filenodes
27645
df8973e1fb45 verify: move file cross checking to its own function
Durham Goode <durham@fb.com>
parents: 27644
diff changeset
   300
28111
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   301
    def _crosscheckfiles(self, filelinkrevs, filenodes):
27645
df8973e1fb45 verify: move file cross checking to its own function
Durham Goode <durham@fb.com>
parents: 27644
diff changeset
   302
        repo = self.repo
df8973e1fb45 verify: move file cross checking to its own function
Durham Goode <durham@fb.com>
parents: 27644
diff changeset
   303
        ui = self.ui
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   304
        ui.status(_("crosschecking files in changesets and manifests\n"))
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   305
28111
06205989264b verify: move cross-checking of changeset/manifest out of _crosscheckfiles()
Martin von Zweigbergk <martinvonz@google.com>
parents: 28007
diff changeset
   306
        total = len(filelinkrevs) + len(filenodes)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   307
        count = 0
27645
df8973e1fb45 verify: move file cross checking to its own function
Durham Goode <durham@fb.com>
parents: 27644
diff changeset
   308
        if self.havemf:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   309
            for f in sorted(filelinkrevs):
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   310
                count += 1
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   311
                ui.progress(_('crosschecking'), count, total=total)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   312
                if f not in filenodes:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   313
                    lr = filelinkrevs[f][0]
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   314
                    self.err(lr, _("in changeset but not in manifest"), f)
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   315
27645
df8973e1fb45 verify: move file cross checking to its own function
Durham Goode <durham@fb.com>
parents: 27644
diff changeset
   316
        if self.havecl:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   317
            for f in sorted(filenodes):
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   318
                count += 1
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   319
                ui.progress(_('crosschecking'), count, total=total)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   320
                if f not in filelinkrevs:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   321
                    try:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   322
                        fl = repo.file(f)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   323
                        lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]])
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   324
                    except Exception:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   325
                        lr = None
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   326
                    self.err(lr, _("in manifest but not in changeset"), f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   327
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   328
        ui.progress(_('crosschecking'), None)
8291
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   329
27644
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   330
    def _verifyfiles(self, filenodes, filelinkrevs):
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   331
        repo = self.repo
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   332
        ui = self.ui
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   333
        lrugetctx = self.lrugetctx
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   334
        revlogv1 = self.revlogv1
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   335
        havemf = self.havemf
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   336
        ui.status(_("checking files\n"))
8291
f5c1a9094e41 verify: avoid exception on missing file revlog
Henrik Stuart <hg@hstuart.dk>
parents: 8225
diff changeset
   337
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   338
        storefiles = set()
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   339
        for f, f2, size in repo.store.datafiles():
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   340
            if not f:
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   341
                self.err(None, _("cannot decode filename '%s'") % f2)
28007
fb92927f9775 treemanifests: fix streaming clone
Martin von Zweigbergk <martinvonz@google.com>
parents: 27964
diff changeset
   342
            elif (size > 0 or not revlogv1) and f.startswith('data/'):
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   343
                storefiles.add(_normpath(f))
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   344
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   345
        files = sorted(set(filenodes) | set(filelinkrevs))
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   346
        total = len(files)
27644
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   347
        revisions = 0
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   348
        for i, f in enumerate(files):
28467
bd37f0d53a49 verify: specify unit for ui.progress when checking files
Anton Shestakov <av6@dwimlabs.net>
parents: 28205
diff changeset
   349
            ui.progress(_('checking'), i, item=f, total=total, unit=_('files'))
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   350
            try:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   351
                linkrevs = filelinkrevs[f]
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   352
            except KeyError:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   353
                # in manifest but not in changelog
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   354
                linkrevs = []
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   355
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   356
            if linkrevs:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   357
                lr = linkrevs[0]
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   358
            else:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   359
                lr = None
2778
fdc232d8a193 Move repo.verify
Matt Mackall <mpm@selenic.com>
parents:
diff changeset
   360
3744
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   361
            try:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   362
                fl = repo.file(f)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   363
            except error.RevlogError as e:
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   364
                self.err(lr, _("broken revlog! (%s)") % e, f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   365
                continue
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   366
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   367
            for ff in fl.files():
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   368
                try:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   369
                    storefiles.remove(ff)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   370
                except KeyError:
27446
6b2c1a1871a6 verify: move warn() to a class level function
Durham Goode <durham@fb.com>
parents: 27445
diff changeset
   371
                    self.warn(_(" warning: revlog '%s' not in fncache!") % ff)
27445
cc178057ab49 verify: move fncachewarned up to a class variable
Durham Goode <durham@fb.com>
parents: 27444
diff changeset
   372
                    self.fncachewarned = True
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   373
27642
f6457349985b verify: move checklog() onto class
Durham Goode <durham@fb.com>
parents: 27453
diff changeset
   374
            self.checklog(fl, f, lr)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   375
            seen = {}
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   376
            rp = None
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   377
            for i in fl:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   378
                revisions += 1
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   379
                n = fl.node(i)
27643
62ce86fcfd06 verify: move checkentry() to be a class function
Durham Goode <durham@fb.com>
parents: 27642
diff changeset
   380
                lr = self.checkentry(fl, i, n, seen, linkrevs, f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   381
                if f in filenodes:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   382
                    if havemf and n not in filenodes[f]:
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   383
                        self.err(lr, _("%s not in manifests") % (short(n)), f)
6534
9b35a9f34675 verify: check copy source revlog and nodeid
Patrick Mezard <pmezard@gmail.com>
parents: 6211
diff changeset
   384
                    else:
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   385
                        del filenodes[f][n]
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   386
31761
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   387
                # Verify contents. 4 cases to care about:
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   388
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   389
                #   common: the most common case
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   390
                #   rename: with a rename
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   391
                #   meta: file content starts with b'\1\n', the metadata
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   392
                #         header defined in filelog.py, but without a rename
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   393
                #   ext: content stored externally
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   394
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   395
                # More formally, their differences are shown below:
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   396
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   397
                #                       | common | rename | meta  | ext
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   398
                #  -------------------------------------------------------
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   399
                #   flags()             | 0      | 0      | 0     | not 0
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   400
                #   renamed()           | False  | True   | False | ?
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   401
                #   rawtext[0:2]=='\1\n'| False  | True   | True  | ?
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   402
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   403
                # "rawtext" means the raw text stored in revlog data, which
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   404
                # could be retrieved by "revision(rev, raw=True)". "text"
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   405
                # mentioned below is "revision(rev, raw=False)".
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   406
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   407
                # There are 3 different lengths stored physically:
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   408
                #  1. L1: rawsize, stored in revlog index
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   409
                #  2. L2: len(rawtext), stored in revlog data
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   410
                #  3. L3: len(text), stored in revlog data if flags==0, or
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   411
                #     possibly somewhere else if flags!=0
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   412
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   413
                # L1 should be equal to L2. L3 could be different from them.
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   414
                # "text" may or may not affect commit hash depending on flag
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   415
                # processors (see revlog.addflagprocessor).
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   416
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   417
                #              | common  | rename | meta  | ext
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   418
                # -------------------------------------------------
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   419
                #    rawsize() | L1      | L1     | L1    | L1
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   420
                #       size() | L1      | L2-LM  | L1(*) | L1 (?)
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   421
                # len(rawtext) | L2      | L2     | L2    | L2
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   422
                #    len(text) | L2      | L2     | L2    | L3
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   423
                #  len(read()) | L2      | L2-LM  | L2-LM | L3 (?)
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   424
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   425
                # LM:  length of metadata, depending on rawtext
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   426
                # (*): not ideal, see comment in filelog.size
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   427
                # (?): could be "- len(meta)" if the resolved content has
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   428
                #      rename metadata
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   429
                #
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   430
                # Checks needed to be done:
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   431
                #  1. length check: L1 == L2, in all cases.
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   432
                #  2. hash check: depending on flag processor, we may need to
b044c339c06d verify: document corner cases
Jun Wu <quark@fb.com>
parents: 30866
diff changeset
   433
                #     use either "text" (external), or "rawtext" (in revlog).
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   434
                try:
32288
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
   435
                    skipflags = self.skipflags
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
   436
                    if skipflags:
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
   437
                        skipflags &= fl.flags(i)
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
   438
                    if not skipflags:
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
   439
                        fl.read(n) # side effect: read content and do checkhash
a2ab9ebcd85b verify: add a config option to skip certain flag processors
Jun Wu <quark@fb.com>
parents: 32250
diff changeset
   440
                        rp = fl.renamed(n)
32250
8a137ef6e5da verify: always check rawsize
Jun Wu <quark@fb.com>
parents: 31762
diff changeset
   441
                    # the "L1 == L2" check
8a137ef6e5da verify: always check rawsize
Jun Wu <quark@fb.com>
parents: 31762
diff changeset
   442
                    l1 = fl.rawsize(i)
8a137ef6e5da verify: always check rawsize
Jun Wu <quark@fb.com>
parents: 31762
diff changeset
   443
                    l2 = len(fl.revision(n, raw=True))
8a137ef6e5da verify: always check rawsize
Jun Wu <quark@fb.com>
parents: 31762
diff changeset
   444
                    if l1 != l2:
8a137ef6e5da verify: always check rawsize
Jun Wu <quark@fb.com>
parents: 31762
diff changeset
   445
                        self.err(lr, _("unpacked size is %s, %s expected") %
8a137ef6e5da verify: always check rawsize
Jun Wu <quark@fb.com>
parents: 31762
diff changeset
   446
                                 (l2, l1), f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   447
                except error.CensoredNodeError:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   448
                    # experimental config: censor.policy
33499
0407a51b9d8c codemod: register core configitems using a script
Jun Wu <quark@fb.com>
parents: 32288
diff changeset
   449
                    if ui.config("censor", "policy") == "abort":
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   450
                        self.err(lr, _("censored file data"), f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   451
                except Exception as inst:
27448
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
   452
                    self.exc(lr, _("unpacking %s") % short(n), inst, f)
3744
d626fc9e3985 verify: add rename link checking
Matt Mackall <mpm@selenic.com>
parents: 3473
diff changeset
   453
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   454
                # check renames
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   455
                try:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   456
                    if rp:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   457
                        if lr is not None and ui.verbose:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   458
                            ctx = lrugetctx(lr)
36339
a4d41ba4ad23 verify: don't reimplement any()
Martin von Zweigbergk <martinvonz@google.com>
parents: 36185
diff changeset
   459
                            if not any(rp[0] in pctx for pctx in ctx.parents()):
27446
6b2c1a1871a6 verify: move warn() to a class level function
Durham Goode <durham@fb.com>
parents: 27445
diff changeset
   460
                                self.warn(_("warning: copy source of '%s' not"
6b2c1a1871a6 verify: move warn() to a class level function
Durham Goode <durham@fb.com>
parents: 27445
diff changeset
   461
                                            " in parents of %s") % (f, ctx))
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   462
                        fl2 = repo.file(rp[0])
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   463
                        if not len(fl2):
27447
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   464
                            self.err(lr, _("empty or missing copy source "
d1b91c10ce70 verify: move err() to be a class function
Durham Goode <durham@fb.com>
parents: 27446
diff changeset
   465
                                     "revlog %s:%s") % (rp[0], short(rp[1])), f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   466
                        elif rp[1] == nullid:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   467
                            ui.note(_("warning: %s@%s: copy source"
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   468
                                      " revision is nullid %s:%s\n")
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   469
                                % (f, lr, rp[0], short(rp[1])))
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   470
                        else:
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   471
                            fl2.rev(rp[1])
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   472
                except Exception as inst:
27448
f4f2179077cb verify: move exc() function onto class
Durham Goode <durham@fb.com>
parents: 27447
diff changeset
   473
                    self.exc(lr, _("checking rename of %s") % short(n), inst, f)
6892
dab95717058d verify: check repo.store
Adrian Buehlmann <adrian@cadifra.com>
parents: 6889
diff changeset
   474
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   475
            # cross-check
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   476
            if f in filenodes:
30393
b667b78099eb verify: avoid shadowing two variables with a list comprehension
Augie Fackler <augie@google.com>
parents: 30375
diff changeset
   477
                fns = [(v, k) for k, v in filenodes[f].iteritems()]
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   478
                for lr, node in sorted(fns):
28114
2a03a365f645 verify: use similar language for missing manifest and file revisions
Martin von Zweigbergk <martinvonz@google.com>
parents: 28113
diff changeset
   479
                    self.err(lr, _("manifest refers to unknown revision %s") %
2a03a365f645 verify: use similar language for missing manifest and file revisions
Martin von Zweigbergk <martinvonz@google.com>
parents: 28113
diff changeset
   480
                             short(node), f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   481
        ui.progress(_('checking'), None)
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   482
28204
962921c330b0 verify: check for orphaned dirlogs
Martin von Zweigbergk <martinvonz@google.com>
parents: 28203
diff changeset
   483
        for f in sorted(storefiles):
27446
6b2c1a1871a6 verify: move warn() to a class level function
Durham Goode <durham@fb.com>
parents: 27445
diff changeset
   484
            self.warn(_("warning: orphan revlog '%s'") % f)
27443
937e73a6e4ff verify: move verify logic into a class
Durham Goode <durham@fb.com>
parents: 26900
diff changeset
   485
27644
331e5c28f5f0 verify: move filelog verification to its own function
Durham Goode <durham@fb.com>
parents: 27643
diff changeset
   486
        return len(files), revisions