1 |
## $Id: Checksum.pm,v 1.1 2003/02/09 05:10:13 joko Exp $ |
2 |
## |
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 |
## $Log: Checksum.pm,v $ |
9 |
## Revision 1.1 2003/02/09 05:10:13 joko |
10 |
## + initial commit |
11 |
## |
12 |
## ---------------------------------------------------------------------------------------- |
13 |
|
14 |
|
15 |
package Data::Transfer::Sync::Compare::Checksum; |
16 |
|
17 |
use strict; |
18 |
use warnings; |
19 |
|
20 |
use mixin::with qw( Data::Transfer::Sync ); |
21 |
|
22 |
|
23 |
# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - main |
24 |
|
25 |
use Data::Dumper; |
26 |
use Digest::MD5 qw(md5 md5_hex md5_base64); |
27 |
|
28 |
|
29 |
# get logger instance |
30 |
my $logger = Log::Dispatch::Config->instance; |
31 |
|
32 |
|
33 |
sub _calcChecksum { |
34 |
|
35 |
my $self = shift; |
36 |
my $descent = shift; |
37 |
my $specifier = shift; |
38 |
|
39 |
# calculate checksum for current object |
40 |
my $ident = $self->{node}->{$descent}->{ident}; |
41 |
|
42 |
# build dump of this node |
43 |
my $payload = $self->{node}->{$descent}->{payload}; |
44 |
#my $dump = $ident . "\n" . $item->quickdump(); |
45 |
#my $dump = $ident . "\n" . Dumper($item); |
46 |
my $dump = $ident . "\n" . $self->_dumpCompact($payload); |
47 |
|
48 |
# TODO: $logger->dump( ... ); |
49 |
#$logger->debug( __PACKAGE__ . ": " . $dump ); |
50 |
#$logger->dump( __PACKAGE__ . ": " . $dump ); |
51 |
|
52 |
# calculate checksum from dump |
53 |
# note: the 32-bit integer hash from DBI seems |
54 |
# to generate duplicates with small payloads already in ranges of hundreds of items/rows!!! |
55 |
# try to avoid to use it or try to use it only for payloads greater than, hmmm, let's say 30 chars? |
56 |
# (we had about 15 chars average per item (row)) |
57 |
|
58 |
# md5-based fingerprint, base64 encoded (from Digest::MD5) |
59 |
$self->{node}->{$descent}->{checksum} = md5_base64($dump) . '=='; |
60 |
# 32-bit integer "hash" value (maybe faster?) (from DBI) |
61 |
#$self->{node}->{$descent}->{checksum} = DBI::hash($dump, 1); |
62 |
|
63 |
# signal good |
64 |
return 1; |
65 |
|
66 |
} |
67 |
|
68 |
|
69 |
sub _readChecksum { |
70 |
my $self = shift; |
71 |
|
72 |
my $descent = shift; |
73 |
|
74 |
# signal checksum bad |
75 |
if (!$self->{node}->{$descent}) { |
76 |
return; |
77 |
} |
78 |
|
79 |
# trace |
80 |
#print "desc: $descent", "\n"; |
81 |
#print Dumper($self); |
82 |
#print Dumper($self->{meta}->{$descent}); |
83 |
#exit; |
84 |
|
85 |
# get checksum for current entry |
86 |
# TODO: don't have the checksum column/property hardcoded as "cs" here, make this configurable somehow |
87 |
if ($self->{meta}->{$descent}->{isChecksumAuthority}) { |
88 |
$self->_calcChecksum($descent); |
89 |
} else { |
90 |
$self->{node}->{$descent}->{checksum} = $self->{node}->{$descent}->{payload}->{cs}; |
91 |
} |
92 |
|
93 |
# signal checksum good |
94 |
return 1; |
95 |
|
96 |
} |
97 |
|
98 |
|
99 |
1; |