thh/newsstats
1
0
Fork 0
Code Issues 3 Pull requests Packages Projects Releases Activity

Add more hosts and special cases.

Browse source 
Signed-off-by: Thomas Hochstein <thh@thh.name>
This commit is contained in:
Thomas Hochstein 2025-05-11 11:11:52 +02:00
parent f7485561dd
commit e40e96a1e2
1 changed files with 11 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
bin/gatherstats.pl
View file

@ -133,11 +133,11 @@ foreach my $Month (&ListMonth($Period)) {
### HostStats ### HostStats
if ($OptStatsType eq 'all' or $OptStatsType eq 'hosts') { if ($OptStatsType eq 'all' or $OptStatsType eq 'hosts') {
# define known hosts using subdomains # define known hosts using subdomains
my @KnownHosts = qw(aioe.org arcor-online.net arcor-ip.de news.astraweb.com read.cnntp.org easynews.com my @KnownHosts = qw(abavia.com aioe.org arcor-online.net arcor-ip.de news.astraweb.com read.cnntp.org
eternal-september.org euro.net fernuni-hagen.de free.fr newsread.freenet.ag easynews.com eternal-september.org euro.net fernuni-hagen.de free.fr newsread.freenet.ag
googlegroups.com news.neostrada.pl newsdawg.com newscene.com news-service.com googlegroups.com heirich.name news.neostrada.pl netcologne.de newsdawg.com newscene.com
octanews.com wieslauf.sub.de highway.telekom.at united-newsserver.de xsnews.nl news-service.com octanews.com .readnews.com wieslauf.sub.de highway.telekom.at
news.xs4all.nl); united-newsserver.de xennanews.com xlned.com xsnews.nl news.xs4all.nl);
&HostStats($DBHandle,$DBRaw,$DBHosts,$Month,$OptMID,$OptTest,$OptDebug,@KnownHosts); &HostStats($DBHandle,$DBRaw,$DBHosts,$Month,$OptMID,$OptTest,$OptDebug,@KnownHosts);
}; };
}; };
@ -315,14 +315,15 @@ sub HostStats {
} }
} }
# special cases
$Host = 'news.highwinds-media.com' if $Host =~ /fx\d\d\.\S{3}\.POSTED/
or $Host =~ /newsfe\d+\.(iad|ams2)/;
$Host = 'newshosting.com' if $Host =~ /post\d*\.iad/;
# trailing .POSTED # trailing .POSTED
($Host) = $Host =~ /(\S+)\.POSTED$/ if $Host =~ /\.POSTED$/; ($Host) = $Host =~ /(\S+)\.POSTED$/ if $Host =~ /\.POSTED$/;
# special cases
$Host = 'news.highwinds-media.com' if $Host =~ /f(e|x)\d\d\.\S{3}\d?$/
or $Host =~ /(newsfe|fed)\d+\.(iad|ams2)$/;
$Host = 'newshosting.com' if $Host =~ /post\d*\.iad$/;
$Host = 'eternal-september.org' if $Host =~ /dont-email\.me$/;
# normalize hosts # normalize hosts
foreach (@KnownHosts) { foreach (@KnownHosts) {
if ($Host =~ /\.$_$/) { if ($Host =~ /\.$_$/) {