#!/usr/bin/perl
use strict;
use warnings;
use utf8;
use Encode;

#########################
# newsstat.pl version 0.4

############################################################################
# Collect statistics about a newsgroup (specified by first argument) in
# the local news spool. Check all articles in the last 30-day period.
# Rank posters by number of posts and by volume of posts, report on top and
# bottom 20 posters. Show their name, number of posts, size of posts,
# percentage of quoted lines. Rank user-agents used, by poster rather than
# by post. Rank top 20 threads. Rank top 10 cross-posted groups.
#
# (Numbers and paths can be configured below.  -- PE)
############################################################################

############################################################################
#                       RECENT CHANGES                                     #
# 2011-07-03  PE  - Use Encode to decode/encode MIME encodings
#                 - Use warnings, utf8 (just in case)
#                 - Documentation update
# N/A         NN  - Take newsgroup name as argument
# 2004-06-19  NN  - newsgroup name is $ARGV[0]
#                 - Allow command line flags for subtracting
#                   output if not pertinent for a group
# 2002-11-09  NN  - Put Garry's writedata() function back in.
#                 - added "rn" to my list of UA's
#                 - Started using %distinct_agent for both User agent
#                   sections
#                 - named it newsstat.pl version 0.3
# 2002-11-06  NN  - Fixed the earliest/latest file problem by using
#                   mtime rather than ctime, and simplifying the logic
# 2002-11-05  NN  - moved user configurations to the top
#                 - fixed the cross-posting section
#                 - introduced the $newsgroup_name variable which
#                   later becomes $news$group
#                 - changed $name to $agent_name in countagents()
#
# Contributors
# -------------
# NN  Nomen nominandum (name to be determined later)
# PE  Thomas 'PointedEars' Lahn <startrek@PointedEars.de>

########### TODO #############
# Commas in bottom section of report
# Show date the figures were compiled
# No. of HTML articles (Content-Type: text/html)
# No. of quoted sigs (/>\s*-- /)
# Per cent of top-posted articles
# Top 10 cross-posters
# Top 20 news posting hosts (from Path)
# Count of certain subject words: newbie, kde, burner, sendmail, etc.
# Count *all* User Agents that each poster uses
# What do we do about Bill Unruh's ] quote style?
# Change the way dates/times are checked
# include % share in posters by no. of arts
# include % share in posters by size
# Total, orig & quoted lines by user agent with per cent
# Take more arguments
#######################################################

###################### USER CONFIGURATIONS ############################

# The name of the group to do stats for
my $newsgroup_name = $ARGV[0];
$newsgroup_name or &usage;

# Check for removal flags
my $ix;
my $j;
my %skipSec;
my @skiplist;
my $args = @ARGV;
for ( $ix = 1 ; $ix < $args ; $ix++ )
{
  $j = $ix + 1;
  if ( $ARGV[$ix] eq "-x" )
  {
    @skiplist = split( ",", $ARGV[$j] );
  }
  elsif ( $ARGV[$ix] =~ /-x(\d.*)/ )
  {
    @skiplist = split( ",", $1 );
  }
}
foreach (@skiplist)
{
  $skipSec{$_} = 1;
}

# Leafnode users will want /var/spool/news for this variable.
my $news = "/var/spool/news/";

# How many days are we doing statistics for?
my $numdays = 30;

# no. of agents we list
my $topagents = 10;

# no. of threads we want to know about
my $topthreads = 20;

# no. of top or bottom posters to show
my $topposters = 20;

# no. of cross-posted threads to show
my $topcrossposts = 10;

# no. of time zones to show
my $toptz = 10;

###################### DATA STRUCTURES ######################
my $group = $newsgroup_name;
$group =~ s!\.!/!g;
my %data;          # name, count, agent, total, orig, quoted
my %threads;       # subject, count
my %crossposts;    # group, count
my %tz;            # timezones by count
my %headers;       # holds header of current article
my %lcheader;      # holds lowercase headers
my @body;          # holds body of current article
my @sig;           # holds sig text;
my $totalposts;    # total no. of posts considered
my $filename;      # name of current article file
my $filesize;      # size of current article file
my $earliest;      # earliest article we have found
my $latest;        # latest article we have found
my $poster;        # poster we are dealing with
my $totsize   = 0; # holds total sizes of all files
my $totheader = 0; # total size of header material
my $totbody   = 0; # total size of body material
my $totsig    = 0; # total size of sig material
my $totorig   = 0; # total size of original material
my $totquoted = 0; # total size of quoted material
my $origposts = 0; # total no. of original posts
my $replies   = 0; # total no. of replies
my $i;             # general purpose
my %distinct_agent;
my %agents =       # used to hold counts of User Agents used
  (
  "KNode"                     => 0,
  "Pan"                       => 0,
  "Mozilla"                   => 0,
  "Sylpheed"                  => 0,
  "Gnus"                      => 0,
  "Forte Agent"               => 0,
  "Forte Free Agent"          => 0,
  "MicroPlanet Gravity"       => 0,
  "Microsoft Outlook Express" => 0,
  "Xnews"                     => 0,
  "slrn"                      => 0,
  "tin"                       => 0,
  "rn"                        => 0,
  "NN"                        => 0,
  "MacSOUP"                   => 0,
  "Foorum"                    => 0,
  "MT-NewsWatcher"            => 0,
  "News Rover"                => 0,
  "WebTV"                     => 0,
  "Compuserver"               => 0,
  "VSoup"                     => 0
  );

######################## MAIN CODE ########################
$! = 1;

chdir("$news$group") or die "Can't cd to $news$group: $!\n";
opendir( DIR, "." ) or die "Can't open $news$group directory: $!\n";
while ( defined( $filename = readdir(DIR) ) )
{
  %lcheader = ();
  next unless -f $filename;    # only want real files
  next if ( $filename eq ".overview" );    # real articles only
  next if ( -M $filename > $numdays );     # only want articles <= a certain age
  $earliest = ( stat $filename )[9] unless defined($earliest);
  $latest   = ( stat $filename )[9] unless defined($latest);
  &getarticle($filename);                  # read in the article
  &getdata;                                # grab the data from the article
  $totalposts++;                           # bump count of articles considered
}
closedir(DIR);                             # finished with the directory

# post-processing
&countagents;    # count agents, collapsing versions
&fixpercent;     # check percentages orig/total for posters

&writedata;

#################### DISPLAY RESULTS #####################
print "=" x 76, "\n";
printf "%s\n", &centred( "Analysis of posts to $newsgroup_name", 76 );
print "=" x 76, "\n";
printf "%s\n",
  &centred( "(stats compiled with a script by Garry Knight et al.)", 76 );
print "\n\n";
printf "Total posts considered: %s over %d days\n", commify($totalposts),
  $numdays;
printf "Earliest article: %s\n",               scalar localtime($earliest);
printf "Latest article:   %s\n",               scalar localtime($latest);
printf "Original articles: %s, replies: %s\n", commify($origposts),
  commify($replies);
printf "Total size of posts: %s bytes (%sK) (%.2fM)\n", commify($totsize),
  commify( int( $totsize / 1024 ) ), $totsize / 1048576;    #
printf "Average %s articles per day, %.2f MB per day, %s bytes per article\n",
  commify( int( $totalposts / $numdays ) ), $totsize / $numdays / 1048576,
  commify( int( $totsize / $totalposts ) );
my $count = keys %data;
printf "Total headers: %s KB  bodies: %s KB\n",
  commify( int( $totheader / 1024 ) ), commify( int( $totbody / 1024 ) );
printf "Body text - quoted: %s KB,  original: %s KB = %02.2f%%, sigs: %s KB\n",
  commify( int( $totquoted / 1024 ) ), commify( int( $totorig / 1024 ) ),
  ( $totorig * 100 ) / ( $totorig + $totquoted ),
  commify( int( $totsig / 1024 ) );
printf "Total number of posters: %s, average %s bytes per poster\n",
  commify($count), commify( int( $totsize / $count ) );     #/
$count = keys %threads;
printf "Total number of threads: %s, average %s bytes per thread\n",
  commify($count), commify( int( $totsize / $count ) );     #/
printf "Total number of User-Agents: %d\n", scalar keys %agents;
print "\n", "=" x 76, "\n";

###############################
# show posters by article count  Sec 1;
###############################
unless ( $skipSec{1} )
{
  if ( keys %data < $topposters )
  {
    $count = keys %data;
  }
  else
  {
    $count = $topposters;
  }
  printf "%s\n", &centred( "Top $count posters by number of articles", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach $poster ( sort { $data{$b}{count} <=> $data{$a}{count} } keys %data )
  {
    my $name = substr( $poster, 0, 65 );
    printf "%2d: %-63s : %6d\n", $i + 1, rpad( $poster, 63, "." ),
      $data{$poster}{count};
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

################################
# show posters by size in Kbytes Sec 2;
################################
unless ( $skipSec{2} )
{
  if ( keys %data < $topposters )
  {
    $count = keys %data;
  }
  else
  {
    $count = $topposters;
  }
  printf "%s\n", &centred( "Top $count posters by article size in Kbytes", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach $poster ( sort { $data{$b}{size} <=> $data{$a}{size} } keys %data )
  {
    my $name = substr( $poster, 0, 62 );
    printf "%2d: %-63s : %6d\n", $i + 1, rpad( $poster, 63, "." ),
      $data{$poster}{size} / 1024;    #/
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

####################################
# show top posters for original text
####################################
unless ( $skipSec{3} )
{
  if ( keys %data < $topposters )
  {
    $count = keys %data;
  }
  else
  {
    $count = $topposters;
  }
  printf "%s\n",
    &centred( "Top $count responders by original text (> 5 posts)", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach $poster ( sort { $data{$b}{percent} <=> $data{$a}{percent} }
    keys %data )
  {
    next if $data{$poster}{quoted} == 0;
    next if $data{$poster}{count} < 5;
    my $name = substr( $poster, 0, 63 );
    printf "%2d: %-63s : %02.2f%%\n", $i + 1, rpad( $poster, 63, "." ),
      $data{$poster}{percent};
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

#######################################
# show bottom posters for original text
#######################################
unless ( $skipSec{4} )
{
  if ( keys %data < $topposters )
  {
    $count = keys %data;
  }
  else
  {
    $count = $topposters;
  }
  printf "%s\n",
    &centred( "Bottom $count responders by original text  (> 5 posts)", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach $poster ( sort { $data{$a}{percent} <=> $data{$b}{percent} }
    keys %data )
  {
    next if $data{$poster}{quoted} == 0;
    next if $data{$poster}{count} < 5;
    my $name = substr( $poster, 0, 63 );
    printf "%2d: %-63s : %02.2f%%\n", $i + 1, rpad( $poster, 63, "." ),
      $data{$poster}{percent};
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

####################################
# show threads by number of articles
####################################
unless ( $skipSec{5} )
{
  if ( keys %threads < $topthreads )
  {
    $count = keys %threads;
  }
  else
  {
    $count = $topthreads;
  }
  printf "%s\n", &centred( "Top $count threads by no. of articles", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach my $thread ( sort { $threads{$b}{count} <=> $threads{$a}{count} }
    keys %threads )
  {
    my $name = substr( $thread, 0, 65 );
    printf "%2d: %-63s : %6d\n", $i + 1, rpad( $name, 63, "." ),
      $threads{$thread}{count};
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}
################################
# show threads by size in Kbytes
################################
unless ( $skipSec{6} )
{
  if ( keys %threads < $topthreads )
  {
    $count = keys %threads;
  }
  else
  {
    $count = $topthreads;
  }
  printf "%s\n", &centred( "Top $count threads by size in KB", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach my $thread ( sort { $threads{$b}{size} <=> $threads{$a}{size} }
    keys %threads )
  {
    my $name = substr( $thread, 0, 65 );
    printf "%2d: %-63s : %6d\n", $i + 1, rpad( $name, 63, "." ),
      $threads{$thread}{size} / 1024;    #/
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

#################################
# show top 10 cross-posted groups
#################################
unless ( $skipSec{7} )
{
  delete $crossposts{"$newsgroup_name"};    # don't include ours
  if ( keys %crossposts < $topcrossposts )
  {
    $count = keys %crossposts;
  }
  else
  {
    $count = $topcrossposts;
  }
  printf "%s\n", &centred( "Top $count cross-posted groups", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach
    my $name ( sort { $crossposts{$b} <=> $crossposts{$a} } keys %crossposts )
  {
    printf "%2d: %-63s : %6d\n", $i + 1, rpad( $name, 63, "." ),
      $crossposts{$name};
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}
#######################
#show agents and counts
#######################
unless ( $skipSec{8} )
{
  if ( keys %agents < $topagents )
  {
    $count = keys %agents;
  }
  else
  {
    $count = $topagents;
  }
  printf "%s\n", &centred( "Top $count User Agents by poster", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach my $agent ( sort { $agents{$b} <=> $agents{$a} } keys %agents )
  {
    printf "%2d: %-63s : %6d\n", $i + 1, rpad( $agent, 63, "." ),
      $agents{$agent};
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

#######################
#show distinct agents
#######################
unless ( $skipSec{9} )
{
  if ( keys %distinct_agent < $topagents )
  {
    $count = keys %distinct_agent;
  }
  else
  {
    $count = $topagents;
  }
  printf "%s\n", &centred( "Top $count User Agents by number of posts", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach my $agent ( sort { $distinct_agent{$b} <=> $distinct_agent{$a} }
    keys %distinct_agent )
  {
    printf "%2d: %-58s : %5d (%2.f%%)\n", $i + 1, rpad( $agent, 58, "." ),
      $distinct_agent{$agent},
      ( ( $distinct_agent{$agent} / $totalposts ) * 100 );
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

##########################
#show timezones and counts
##########################
unless ( $skipSec{10} )
{
  if ( keys %tz < $toptz )
  {
    $count = keys %tz;
  }
  else
  {
    $count = $toptz;
  }
  printf "%s\n", &centred( "Top 10 time zones", 76 );
  print "=" x 76, "\n";
  $i = 0;
  foreach my $zone ( sort { $tz{$b} <=> $tz{$a} } keys %tz )
  {
    printf "%2d: %-63s : %6d\n", $i + 1, rpad( $zone, 63, "." ), $tz{$zone};
    last if ( ++$i == $count );
  }
  print "\n", "=" x 76, "\n";
}

################################ SUBROUTINES ################################

#######################################
# get current article's header and body
#######################################
sub getarticle
{
  %headers = ();    # dump old headers
  my $filename = shift;    # get the name of the file

  # get stats about the file itself
  $filesize = -s $filename;    # get total size of file
  $totsize += $filesize;       # bump total sizes of all files

  my $mtime = ( stat $filename )[9];
  if ( $mtime < $earliest )
  {
    $earliest = $mtime;
  }
  elsif ( $mtime > $latest )
  {
    $latest = $mtime;
  }

  # now read the file
  open( FILE, $filename ) or die "Can't open $filename: $!\n";
  while (<FILE>)
  {
    $totheader += length($_);    # bump total header size
    last if (/^\s*$/);           # end of header?
    if (/^([^:\s]*):\s+(.*)/)
    {
      my ( $key, $val ) = ( $1, $2 );
      $headers{$key} = decode( 'MIME-Header', $val );
      $lcheader{ clean( lc($key) ) } = clean($val);
    }
  }
  @body = <FILE>;                # slurp up body
  close(FILE);
}    # getarticle

###################################
# get data from the current article
###################################
sub getdata
{
#### First, analyse header fields ####

  # Set up this poster if not defined, get counts, sizes
  $poster = encode( 'UTF-8', $headers{From} );    # get the poster's name
  if ( !defined( $data{$poster} ) )
  {                                               # seen this one before?
    $data{$poster}{agent}  = 'Unknown';           # comes after For: field
    $data{$poster}{orig}   = 0;
    $data{$poster}{quoted} = 0;
  }
  $data{$poster}{count}++;                        # bump count for this poster
  $data{$poster}{size} += $filesize;              # total size of file

  # The User-Agent and/or X-Newsreader fields
  # for User-Agent by poster
  if ( defined $lcheader{"user-agent"} )
  {
    $data{$poster}{agent} = $lcheader{"user-agent"};
  }
  if ( defined $lcheader{"x-newsreader"} )
  {
    $data{$poster}{agent} = $lcheader{"x-newsreader"};
  }

  # The User Agent for User-Agent by number of posts
  my $UA = "unknown";
  foreach my $keys ( keys %lcheader )
  {
    if ( defined $lcheader{'user-agent'} )
    {
      $UA = $lcheader{'user-agent'};
    }
    elsif ( defined $lcheader{"x-newsreader"} )
    {
      $UA = $lcheader{"x-newsreader"};
    }
    elsif ( defined $lcheader{'x-mailer'} )
    {
      $UA = $lcheader{'x-mailer'};
    }
    elsif (
      ( defined $lcheader{'organization'} )
      && ( $lcheader{'organization'} =~
        /groups\.google|AOL|Supernews|WebTV|compuserve/ )
      )
    {
      $UA = $lcheader{'organization'};
    }
    elsif ( $lcheader{'message-id'} =~ /pine/i )
    {
      $UA = "Pine";
    }    ## Hopefully found UA, else set to unknown
  }

  $UA = clean($UA);
  $UA = get_agent($UA);

  sub get_agent
  {
    my $raw   = shift;
    my $agent = $raw;

    ## strip http
    if ( $raw =~ /.*http.*/ )
    {
      $raw =~ s!posted via!!i;
      $raw =~ s!http://!!g;
      $raw =~ s!/!!g;
      $raw =~ s! !!g;
    }

    ## Fix Outlook from Mac
    if ( $raw =~ /^microsoft/i ) { $raw =~ s/-/ /g; }

    ## Pick out the popular agents
    if ( $raw =~ /(outlook express)/i
      || $raw =~ /(microplanet gravity)/i
      || $raw =~ /(news rover)/i
      || $raw =~ /(forte agent)/i
      || $raw =~ /(forte free agent)/i )
    {
      $agent = $1;
    }
    elsif (
      $raw =~ /^(
        pan
       |sylpheed
       |slrn
       |mozilla
       |knode
       |tin
       |hamster
       |xrn
       |xnews
       |aol
       |gnus
       |krn
       |macsoup
       |messenger
       |openxp
       |pine
       |thoth
       |turnpike
       |winvn
       |vsoup
       |google
       |supernews
       |nn
       |rn
       |007
       |webtv
       |compuserve
       )/ix
      )
    {
      $agent = $1;
    }
    else
    {
      ## Clean up unknown agents
      if ( $raw =~ m!^(.*?)/! )
      {
        $agent = $1;
      }
      elsif ( $raw =~ /^(\w*)\d.*/ )
      {
        $agent = $1;
      }
    }

    $distinct_agent{$agent}++;
    return $agent;
  }

  # Get all cross-posted newsgroups
  for ( split /,/, $headers{"Newsgroups"} )
  {
    $crossposts{$_}++;    # bump count for each
  }

  # Get threads
  my $thread = encode( 'UTF-8', $headers{"Subject"} );
  $thread =~ s/^re: //i;    # Remove Re: or re: at start
  $thread =~ s/\s+/ /g;     # collapse whitespace
  $threads{$thread}{count} += 1;            # bump count of this subject
  $threads{$thread}{size}  += $filesize;    # bump bytes for this thread

  # Is this an original post or a reply?
  if ( defined $headers{"References"} )
  {
    $replies++;
  }
  else
  {
    $origposts++;
  }

  # Get the time zone
  $_ = $headers{"Date"};
  my ($tz) = /\d\d:\d\d:\d\d\s+(.*)/;
  if ( ( $tz =~ /UTC/ ) or ( $tz =~ /GMT/ ) or ( $tz =~ /0000/ ) )
  {
    $tz = "UTC";
  }
  $tz{$tz}++;

#### Now analyse the body text ####
  my $insig = 0;
  for (@body)
  {
    $totbody += length($_);    # bump total body size
    next if (/^$>/);           # don't count blank lines in body
    if ( $insig == 1 )
    {
      $totsig += length($_);    # bump total sig size

      # Bill Unruh uses ] quotes, and another poster uses ::
    }
    elsif ( /^\s*[>\]]/ or /^\s*::/ )
    {                           # are we in a quote line?
      $data{$poster}{quoted} += length($_);    # bump count of quoted chrs
      $totquoted             += length($_);
    }
    elsif (/-- /)
    {
      $insig = 1;
    }
    else
    {

      # we must be processing an original line
      $data{$poster}{orig} += length($_);      # bump count of original chrs
      $totorig             += length($_);
    }
  }    # end for (@body)

}    # getdata

########################################
# Count the User-Agents used, collapsing
# different versions into one per agent.
########################################
sub countagents
{
POSTER:
  foreach $poster ( keys %data )
  {
    foreach my $agent_name ( keys %distinct_agent )
    {    # check against known ones
      if ( $data{$poster}{agent} =~ /\Q$agent_name\E/ )
      {
        $agents{$agent_name}++;
        next POSTER;
      }
    }
    $agents{ $data{$poster}{agent} }++;
  }
}    # countagents

############################################
# set orig/total percentages for all posters
############################################
sub fixpercent
{
  foreach $poster ( keys %data )
  {
    my $percent = 100;
    if ( ( $data{$poster}{orig} != 0 ) and ( $data{$poster}{quoted} != 0 ) )
    {
      $percent = $data{$poster}{orig} * 100 /
        ( $data{$poster}{quoted} + $data{$poster}{orig} );    #/
    }
    elsif ( $data{$poster}{orig} == 0 )
    {
      $percent = 0;
    }
    $data{$poster}{percent} = $percent;
  }
}

##############################
# right pad a string with '.'s
##############################
sub rpad
{

  # get text to pad, length to pad, pad chr
  my ( $text, $pad_len, $pad_chr ) = @_;
  if ( length($text) > $pad_len )
  {
    $text = substr( $text, 0, $pad_len );
  }
  my $padded = $text . $pad_chr x ( $pad_len - length($text) );
  return $padded;
}

#################
# centre a string
#################
sub centred
{
  my ( $text, $width ) = @_;    # text to centre, size of field to centre in
  my $pad_len = ( $width - length($text) ) / 2;    #/
  my $centred = " " x $pad_len . $text;
  return $centred;
}

##########################
# put commas into a number
##########################
sub commify
{
  $_ = shift;
  1 while s/^(-?\d+)(\d{3})/$1,$2/;
  return $_;
}

#########################
# clean
#########################
sub clean
{
  my $dirty = shift;
  my $clean = $dirty;
  $clean =~ s/^\s*//;
  $clean =~ s/\s*$//;

  return $clean;
}

sub usage
{

  print "usage: newstat.pl newsgroupname\n";
  exit 1;
}

###################################
# Write data structures to a file #
###################################
sub writedata
{
  open OUTF, ">/tmp/XDATA" or die "Can't create XDATA: $!\n";
  print OUTF "Data collected from alt.os.linux.mandrake\n\n";
  print OUTF
    "Poster Data\nname : agent : count : size: orig : quoted : per cent\n";
  foreach my $name ( keys %data )
  {
    print OUTF
"$name : $data{$name}{agent} : $data{$name}{count} : $data{$name}{size} : $data{$name}{orig} : $data{$name}{quoted} : $data{$name}{percent}\n";
  }
  print OUTF
"============================================================================\n";
  print OUTF "Thread subjects\n";
  print OUTF
"----------------------------------------------------------------------------\n";
  foreach my $thread ( sort { "\L$a" cmp "\L$b" } keys %threads )
  {
    print OUTF "$thread : $threads{$thread}{count} : $threads{$thread}{size}\n";
  }
  print OUTF
"============================================================================\n";
  print OUTF "Cross-posts\n";
  print OUTF
"----------------------------------------------------------------------------\n";
  foreach my $name ( sort keys %crossposts )
  {
    print OUTF "$name : $crossposts{$name}\n";
  }
  print OUTF print OUTF
"============================================================================\n";
  print OUTF "User agents\n";
  print OUTF
"----------------------------------------------------------------------------\n";
  foreach my $name ( sort keys %agents )
  {
    print OUTF "$name : $agents{$name}\n";
  }
  print OUTF
"============================================================================\n";
  print OUTF "Time zones\n";
  print OUTF
"----------------------------------------------------------------------------\n";
  foreach my $name ( sort keys %tz )
  {
    print OUTF "$name : $tz{$name}\n";
  }
  close OUTF;
}    # writedata