#!/usr/bin/perl -w

## The script is placed in twiki/tools dir

BEGIN {
    unshift @INC, '.';
    require '../bin/setlib.cfg';
}

my $web   = $ARGV[0] || die "Error: Web miss.(./findnontopics.pl WEB)";

use Socket;
use TWiki;
my $twiki = new TWiki();
my $store = $twiki->{store};
my @topics= $store->getTopicNames( $web );
my $twurl =$TWiki::cfg{ScriptUrlPath};

my $adim="SuperAdmin";
my $admpwd="passadmin";
my $logheader="username=$adim&password=$admpwd&origurl=";
my $twid;
my %counter;
my %out;
my $html;
my $url;
my $host   = "127.0.0.1";
my $port   = "80";
my $header = "Host: $host";

##log in
$url    = "GET  $twurl/login/TWiki/WebHome?origurl=";
$twid   = &getUrl($host,$port,$url,$header);
$twid=~/Set-Cookie: (TWIKISID=\w+)/;
$url    ="POST  $twurl/login/TWiki/WebHome";
$header ="Host: $host\r\nCookie: $1\r\nContent-Type: application/x-www-form-urlencoded\r\nContent-Length: ".length($logheader)."\r\n\r\n$logheader";
$twid = &getUrl($host,$port,$url,$header);

## loged sesssion
$header ="Host: $host\r\nCookie: $1";

##Colect data
foreach my $topic ( $store->getTopicNames( $web )) {
  $url    = "GET  $twurl/view/$web/$topic";
  next if $topic =~/^NonTopicsReport$|^Trash|^WebAtom$|^WebRss$|^WebPreferences$|^WebStatistics$/;
  $html   = &getUrl($host,$port,$url,$header);
  while ($html=~/<span class="twikiNewLink">([^<]*)</g){
    if (defined $out{$1} ){
      $counter{$1}+=1;
      $out{$1} .=", $topic" unless $out{$1} =~/\b$topic\b/;
    }else{
      $counter{$1}=1;
      $out{$1} = "$topic";
    }
  }
}

##Sub
sub getUrl
{
    my ( $theHost, $thePort, $theUrl, $theHeader ) = @_;
    my $result = '';
    my $req = "$theUrl HTTP/1.0\r\n$theHeader\r\n\r\n";
    my ( $iaddr, $paddr, $proto );
    $iaddr   = inet_aton( $theHost );
    $paddr   = sockaddr_in( $thePort, $iaddr );
    $proto   = getprotobyname( 'tcp' );
    socket( SOCK, PF_INET, SOCK_STREAM, $proto )  or die "socket: $!";
    connect( SOCK, $paddr ) or die "connect: $!";
    select SOCK; $| = 1;
    print SOCK $req;
    while( <SOCK> ) { $result .= $_; }
    close( SOCK )  or die "close: $!";
    select STDOUT;
    return $result;
}


##Report

my $DATE = `date +\%Y-\%m-\%d\\ \%H:\%M`;
print <<REPORT;
---+ List of non-existing pages
This page automagically lists all[*] pages which are referenced but undefined (i.e. which show with a '?' in TWiki.
This list is refreshed once a week .
The last time this list was generated was $DATE.

---++ Pages that are referenced but which which do not exist

%META:TOPICPARENT{name="WebHome"}%
REPORT

foreach my $key (sort keys %out){
  print "   * $key(".$counter{$key}.") - ".$out{$key}."\n";
}