1 |
jonen |
1.4 |
## $Id: Checksum.pm,v 1.3 2003/05/13 08:19:00 joko Exp $ |
2 |
joko |
1.1 |
## |
3 |
|
|
## Copyright (c) 2002 Andreas Motl <andreas.motl@ilo.de> |
4 |
|
|
## |
5 |
|
|
## See COPYRIGHT section in pod text below for usage and distribution rights. |
6 |
|
|
## |
7 |
|
|
## ---------------------------------------------------------------------------------------- |
8 |
joko |
1.2 |
## $Log: Checksum.pm,v $ |
9 |
jonen |
1.4 |
## Revision 1.3 2003/05/13 08:19:00 joko |
10 |
|
|
## switched to crc32 |
11 |
|
|
## |
12 |
joko |
1.3 |
## Revision 1.2 2003/02/11 09:53:07 joko |
13 |
|
|
## + metadata-structure-change, fixed some code here |
14 |
|
|
## |
15 |
joko |
1.2 |
## Revision 1.1 2003/02/09 05:10:13 joko |
16 |
|
|
## + initial commit |
17 |
|
|
## |
18 |
joko |
1.1 |
## ---------------------------------------------------------------------------------------- |
19 |
|
|
|
20 |
|
|
|
21 |
joko |
1.3 |
=pod |
22 |
|
|
|
23 |
|
|
=head1 Todo |
24 |
|
|
|
25 |
|
|
o Data::Transfer::Sync::Compare::Slot using Compare::Struct |
26 |
|
|
|
27 |
|
|
=cut |
28 |
|
|
|
29 |
joko |
1.1 |
package Data::Transfer::Sync::Compare::Checksum; |
30 |
|
|
|
31 |
|
|
use strict; |
32 |
|
|
use warnings; |
33 |
|
|
|
34 |
|
|
use mixin::with qw( Data::Transfer::Sync ); |
35 |
|
|
|
36 |
|
|
|
37 |
|
|
use Data::Dumper; |
38 |
|
|
|
39 |
joko |
1.3 |
# TODO: Load these appropriatly at runtime. |
40 |
|
|
use Digest::MD5 qw( md5 md5_hex md5_base64 ); |
41 |
|
|
use String::CRC32; |
42 |
jonen |
1.4 |
use File::RsyncP::Digest; |
43 |
joko |
1.1 |
|
44 |
|
|
# get logger instance |
45 |
|
|
my $logger = Log::Dispatch::Config->instance; |
46 |
|
|
|
47 |
|
|
|
48 |
joko |
1.3 |
# Maybe refactor to shortcuts::checksum? |
49 |
joko |
1.1 |
sub _calcChecksum { |
50 |
|
|
|
51 |
|
|
my $self = shift; |
52 |
|
|
my $descent = shift; |
53 |
|
|
my $specifier = shift; |
54 |
|
|
|
55 |
|
|
# calculate checksum for current object |
56 |
|
|
my $ident = $self->{node}->{$descent}->{ident}; |
57 |
|
|
|
58 |
|
|
# build dump of this node |
59 |
|
|
my $payload = $self->{node}->{$descent}->{payload}; |
60 |
|
|
#my $dump = $ident . "\n" . $item->quickdump(); |
61 |
|
|
#my $dump = $ident . "\n" . Dumper($item); |
62 |
|
|
my $dump = $ident . "\n" . $self->_dumpCompact($payload); |
63 |
|
|
|
64 |
|
|
# TODO: $logger->dump( ... ); |
65 |
|
|
#$logger->debug( __PACKAGE__ . ": " . $dump ); |
66 |
|
|
#$logger->dump( __PACKAGE__ . ": " . $dump ); |
67 |
|
|
|
68 |
|
|
# calculate checksum from dump |
69 |
joko |
1.3 |
|
70 |
|
|
# 1. md5-based fingerprint, base64 encoded (from Digest::MD5) |
71 |
|
|
#$self->{node}->{$descent}->{checksum} = md5_base64($dump) . '=='; |
72 |
|
|
|
73 |
|
|
# 2. 32-bit integer "hash" value (maybe faster?) (from DBI) |
74 |
|
|
# Note: The 32-bit integer hash from DBI seems to generate duplicates |
75 |
|
|
# with small payloads already in ranges of hundreds of items/rows!!! |
76 |
|
|
# Try to avoid it or try to use it only for payloads greater than, hmmm, let's say 30 chars? |
77 |
|
|
# (we had about 15 chars average per item (row)) |
78 |
|
|
# Possible (generic) solution: Just generate checksum, if length(checksum(payload)) < payload |
79 |
joko |
1.1 |
#$self->{node}->{$descent}->{checksum} = DBI::hash($dump, 1); |
80 |
joko |
1.3 |
|
81 |
|
|
# 3. good ol' crc32??? |
82 |
jonen |
1.4 |
#$self->{node}->{$descent}->{checksum} = crc32($dump); |
83 |
joko |
1.3 |
|
84 |
jonen |
1.4 |
# 4. File::RsyncP::Digest - Perl interface to rsync message digest algorithms |
85 |
|
|
my $rsDigest = new File::RsyncP::Digest; |
86 |
|
|
$rsDigest->add($dump); |
87 |
|
|
my $digest = $rsDigest->digest(); |
88 |
|
|
$self->{node}->{$descent}->{checksum} = unpack("H*", $digest); |
89 |
|
|
|
90 |
|
|
# 5. some more modern Digest::SHA1 or similar? |
91 |
joko |
1.1 |
|
92 |
|
|
# signal good |
93 |
|
|
return 1; |
94 |
|
|
|
95 |
|
|
} |
96 |
|
|
|
97 |
|
|
|
98 |
|
|
sub _readChecksum { |
99 |
|
|
my $self = shift; |
100 |
|
|
|
101 |
|
|
my $descent = shift; |
102 |
|
|
|
103 |
|
|
# signal checksum bad |
104 |
|
|
if (!$self->{node}->{$descent}) { |
105 |
|
|
return; |
106 |
|
|
} |
107 |
|
|
|
108 |
|
|
# trace |
109 |
|
|
#print "desc: $descent", "\n"; |
110 |
|
|
#print Dumper($self); |
111 |
joko |
1.2 |
#print Dumper($self->{meta}->{$descent}); |
112 |
joko |
1.1 |
#exit; |
113 |
|
|
|
114 |
|
|
# get checksum for current entry |
115 |
|
|
# TODO: don't have the checksum column/property hardcoded as "cs" here, make this configurable somehow |
116 |
joko |
1.2 |
if ($self->{meta}->{$descent}->{isChecksumAuthority}) { |
117 |
joko |
1.1 |
$self->_calcChecksum($descent); |
118 |
|
|
} else { |
119 |
|
|
$self->{node}->{$descent}->{checksum} = $self->{node}->{$descent}->{payload}->{cs}; |
120 |
|
|
} |
121 |
|
|
|
122 |
|
|
# signal checksum good |
123 |
|
|
return 1; |
124 |
|
|
|
125 |
|
|
} |
126 |
|
|
|
127 |
|
|
|
128 |
|
|
1; |
129 |
joko |
1.3 |
__END__ |