X-Git-Url: http://dxcluster.org/gitweb/gitweb.cgi?a=blobdiff_plain;f=perl%2FAnnTalk.pm;h=f6b4653d28d497b9a854de43a206dc470eb86a58;hb=refs%2Fheads%2Fip_address;hp=63f2512b697356dd60c5f989a87556c7889446e2;hpb=50c5cbb346de7d1b5026e8a205b5cf1e04e6cf05;p=spider.git diff --git a/perl/AnnTalk.pm b/perl/AnnTalk.pm index 63f2512b..f6b4653d 100644 --- a/perl/AnnTalk.pm +++ b/perl/AnnTalk.pm @@ -3,7 +3,7 @@ # # Copyright (c) 2000 Dirk Koopman # -# $Id$ +# # package AnnTalk; @@ -17,8 +17,8 @@ use DXVars; use vars qw(%dup $duplth $dupage $filterdef); -$duplth = 60; # the length of text to use in the deduping -$dupage = 5*24*3600; # the length of time to hold spot dups +$duplth = 30; # the length of text to use in the deduping +$dupage = 5*3600; # the length of time to hold ann dups $filterdef = bless ([ # tag, sort, field, priv, special parser ['by', 'c', 0], @@ -38,12 +38,6 @@ $filterdef = bless ([ ['origin_state', 'nz', 14], ], 'Filter::Cmd'); -use vars qw($VERSION $BRANCH); -$VERSION = sprintf( "%d.%03d", q$Revision$ =~ /(\d+)\.(\d+)/ ); -$BRANCH = sprintf( "%d.%03d", q$Revision$ =~ /\d+\.\d+\.(\d+)\.(\d+)/ || (0,0)); -$main::build += $VERSION; -$main::branch += $BRANCH; - # enter the spot for dup checking and return true if it is already a dup sub dup { @@ -53,10 +47,10 @@ sub dup chomp $text; unpad($text); $text =~ s/\%([0-9A-F][0-9A-F])/chr(hex($1))/eg; - $text = pack("C*", map {$_ & 127} unpack("C*", $text)); +# $text = Encode::encode("iso-8859-1", $text) if $main::can_encode && Encode::is_utf8($text, 1); $text =~ s/[^\#a-zA-Z0-9]//g; $text = substr($text, 0, $duplth) if length $text > $duplth; - my $dupkey = "A$to|\L$text"; + my $dupkey = "A$call|$to|\L$text"; return DXDupe::check($dupkey, $t); }