[exim.git] / src / src / eximstats.src

#!PERL_COMMAND -w
# $Cambridge: exim/src/src/eximstats.src,v 1.4 2004/12/21 11:33:55 steve Exp $

# Copyright (c) 2001 University of Cambridge.
# See the file NOTICE for conditions of use and distribution.

# Perl script to generate statistics from one or more Exim log files.

# Usage: eximstats [<options>] <log file> <log file> ...

# 1996-05-21: Ignore lines not starting with valid date/time, just in case
#               these get into a log file.
# 1996-11-19: Add the -h option to control the size of the histogram,
#               and optionally turn it off.
#             Use some Perl 5 things; it should be everywhere by now.
#             Add the Perl -w option and rewrite so no warnings are given.
#             Add the -t option to control the length of the "top" listing.
#             Add the -ne, -nt options to turn off errors and transport
#               information.
#             Add information about length of time on queue, and -q<list> to
#               control the intervals and turn it off.
#             Add count and percentage of delayed messages to the Received
#               line.
#             Show total number of errors.
#             Add count and percentage of messages with errors to Received
#               line.
#             Add information about relaying and -nr to suppress it.
# 1997-02-03  Merged in some of the things Nigel Metheringham had done:
#               Re-worded headings
#               Added received histogram as well as delivered
#               Added local senders' league table
#               Added local recipients' league table
# 1997-03-10  Fixed typo "destinationss"
#             Allow for intermediate address between final and original
#               when testing for relaying
#             Give better message when no input
# 1997-04-24  Fixed bug in layout of error listing that was depending on
#               text length (output line got repeated).
# 1997-05-06  Bug in option decoding when only one option.
#             Overflow bug when handling very large volumes.
# 1997-10-28  Updated to handle revised log format that might show
#               HELO name as well as host name before IP number
# 1998-01-26  Bugs in the function for calculating the number of seconds
#               since 1970 from a log date
# 1998-02-02  Delivery to :blackhole: doesn't have a T= entry in the log
#               line; cope with this, thereby avoiding undefined problems
#             Very short log line gave substring error
# 1998-02-03  A routed delivery to a local transport may not have <> in the
#               log line; terminate the address at white space, not <
# 1998-09-07  If first line of input was a => line, $thissize was undefined;
#               ensure it is zero.
# 1998-12-21  Adding of $thissize from => line should have been adding $size.
#             Oops. Should have looked more closely when fixing the previous
#               bug!
# 1999-11-12  Increased the field widths for printed integers; numbers are
#               bigger than originally envisaged.
# 2001-03-21  Converted seconds() routine to use Time::Local, fixing a bug
#               whereby seconds($timestamp) - id_seconds($id) gave an
#               incorrect result.
#             Added POD documentation.
#             Moved usage instructions into help() subroutine.
#             Added 'use strict' and declared all global variables.
#             Added '-html' flag and resultant code.
#             Added '-cache' flag and resultant code.
#             Added add_volume() routine and converted all volume variables
#               to use it, fixing the overflow problems for individual hosts
#               on large sites.
#             Converted all volume output to GB/MB/KB as appropriate.
#             Don't store local user stats if -nfl is specified.
#             Modifications done by: Steve Campbell (<steve@computurn.com>)
# 2001-04-02  Added the -t_remote_users flag. Steve Campbell.
# 2001-10-15  Added the -domain flag. Steve Campbell.
# 2001-10-16  Accept files on STDIN or on the command line. Steve Campbell.
# 2001-10-21  Removed -domain flag and added -bydomain, -byhost, and -byemail.
#             We now generate our main parsing subroutine as an eval statement
#             which improves performance dramatically when not all the results
#             are required. We also cache the last timestamp to time convertion.
#
#             NOTE: 'Top 50 destinations by (message count|volume)' lines are
#             now 'Top N (host|email|domain) destinations by (message count|volume)'
#             where N is the topcount. Steve Campbell.
#
# 2001-10-30  V1.16 Joachim Wieland.
#            Fixed minor bugs in add_volume() when taking over this version
#               for use in Exim 4: -w gave uninitialized value warnings in
#               two situations: for the first addition to a counter, and if
#               there were never any gigabytes, thereby leaving the $gigs
#               value unset.
#             Initialized $last_timestamp to stop a -w uninitialized warning.
#             Minor layout tweak for grand totals (nitpicking).
#             Put the IP addresses for relaying stats in [] and separated by
#               a space from the domain name.
#             Removed the IPv4-specific address test when picking out addresses
#               for relaying. Anything inside [] is OK.
#
# 2002-07-02  Philip Hazel
#             Fixed "uninitialized variable" message that occurred for relay
#               messages that arrived from H=[1.2.3.4] hosts (no name shown).
#               This bug didn't affect the output.
#
# 2002-04-15  V1.17 Joachim Wieland.
#             Added -charts, -chartdir. -chartrel options which use
#             GD::Graph modules to create graphical charts of the statistics.
#
# 2002-04-15  V1.18 Steve Campbell.
#             Added a check for $domain to to stop a -w uninitialized warning.
#             Added -byemaildomain option.
#             Only print HTML header links to included tables!
#
# 2002-08-02  V1.19 Steve Campbell.
#             Changed the debug mode to dump the parser onto STDERR rather
#             than STDOUT. Documented the -d flag into the help().
#             Rejoined the divergent 2002-04-15 and 2002-07-02 releases.
#
# 2002-08-21  V1.20 Steve Campbell.
#             Added the '-merge' option to allow merging of previous reports.
#             Fixed a missing semicolon when doing -bydomain.
#             Make volume charts plot the data gigs and bytes rather than just bytes.
#             Only process log lines with $flag =~ /<=|=>|->|==|\*\*|Co/
#             Converted Emaildomain to Edomain - the column header was too wide!
#             This changes the text output slightly. You can revert to the old
#             column widths by changing $COLUMN_WIDTHS to 7;
#
# 2002-09-04  V1.21 Andreas J Mueller
#             Local deliveries domain now defaults to 'localdomain'.
#             Don't match F=<From> when looking for the user.
#
# 2002-09-05  V1.22 Steve Campbell
#             Fixed a perl 5.005 incompatibility problem ('our' variables).
#
# 2002-09-11  V1.23 Steve Campbell
#             Stopped -charts option from throwing errors on null data.
#             Don't print out 'Errors encountered' unless there are any.

# 2002-10-21  V1.23a Philip Hazel - patch from Tony Finch put in until
#               Steve's eximstats catches up.
#             Handle log files that include the timezone after the timestamp.
#             Switch to assuming that log timestamps are in local time, with
#               an option for UTC timestamps, as in Exim itself.
#
# 2003-02-05  V1.24 Steve Campbell
#             Added in Sergey Sholokh's code to convert '<' and '>' characters
#             in HTML output. Also added code to convert them back with -merge.
#             Fixed timestamp offsets to convert to seconds rather than minutes.
#             Updated -merge to work with output files using timezones.
#             Added cacheing to speed up the calculation of timezone offsets.
#
# 2003-02-07  V1.25 Steve Campbell
#             Optimised the usage of mktime() in the seconds subroutine.
#             Removed the now redundant '-cache' option.
#             html2txt() now explicitly matches HTML tags.
#             Implemented a new sorting algorithm - the top_n_sort() routine.
#             Added Danny Carroll's '-nvr' flag and code.
#
# 2003-03-13  V1.26 Steve Campbell
#             Implemented HTML compliance changes recommended by Bernard Massot.
#             Bug fix to allow top_n_sort() to handle null keys.
#             Convert all domains and edomains to lowercase.
#             Remove preceding dots from domains.
#
# 2003-03-13  V1.27 Steve Campbell
#             Replaced border attributes with 'border=1', as recommended by
#             Bernard Massot.
#
# 2003-06-03  V1.28 John Newman
#             Added in the ability to skip over the parsing and evaulation of
#             specific transports as passed to eximstats via the new "-nt/.../"
#             command line argument.  This new switch allows the viewing of
#             not more accurate statistics but more applicable statistics when
#             special transports are in use (ie; SpamAssassin).  We need to be
#             able to ignore transports such as this otherwise the resulting
#             local deliveries are significantly skewed (doubled)...
#
# 2003-11-06  V1.29 Steve Campbell
#             Added the '-pattern "Description" "/pattern/"' option.
#
# 2004-02-17  V1.30 Steve Campbell
#             Added warnings if required GD::Graph modules are not available or
#             insufficient -chart* options are specified.
#
# 2004-02-20  V1.31 Andrea Balzi 
#             Only show the Local Sender/Destination links if the tables exist.
#
# 2004-07-05  V1.32 Steve Campbell
#             Fix '-merge -h0' divide by zero error.
#
# 2004-07-15  V1.33 Steve Campbell
#             Documentation update - I've converted the subroutine
#             documentation from POD to comments.
#
# 2004-12-10  V1.34 Steve Campbell
#             Eximstats can now parse syslog lines as well as mainlog lines.
#
# 2004-12-20  V1.35 Wouter Verhelst
#	      Pie charts by volume were actually generated by count. Fixed.


=head1 NAME

eximstats - generates statistics from Exim mainlog or syslog files.

=head1 SYNOPSIS

 eximstats [Options] mainlog1 mainlog2 ... > report.txt
 eximstats -html [Options] mainlog1 mainlog2 ... > report.html
 eximstats -merge [Options] report.1.txt report.2.txt ... > weekly_report.txt

Options:

=over 4

=item B<-h>I<number>

histogram divisions per hour. The default is 1, and
0 suppresses histograms. Valid values are:

0, 1, 2, 3, 5, 10, 15, 20, 30 or 60.

=item B<-ne>

Don't display error information.

=item B<-nr>

Don't display relaying information.

=item B<-nr>I</pattern/>

Don't display relaying information that matches.

=item B<-nt>

Don't display transport information.

=item B<-nt>I</pattern/>

Don't display transport information that matches

=item B<-q>I<list>

List of times for queuing information single 0 item suppresses.

=item B<-t>I<number>

Display top <number> sources/destinations
default is 50, 0 suppresses top listing.

=item B<-tnl>

Omit local sources/destinations in top listing.

=item B<-t_remote_users>

Include remote users in the top source/destination listings.

=item B<-byhost>

Show results by sending host. This may be combined with
B<-bydomain> and/or B<-byemail> and/or B<-byedomain>. If none of these options
are specified, then B<-byhost> is assumed as a default.

=item B<-bydomain>

Show results by sending domain.
May be combined with B<-byhost> and/or B<-byemail> and/or B<-byedomain>.

=item B<-byemail>

Show results by sender's email address.
May be combined with B<-byhost> and/or B<-bydomain> and/or B<-byedomain>.

=item B<-byemaildomain> or B<-byedomain>

Show results by sender's email domain.
May be combined with B<-byhost> and/or B<-bydomain> and/or B<-byemail>.

=item B<-pattern> I<Description> I</Pattern/>

Look for the specified pattern and count the number of lines in which it appears.
This option can be specified multiple times. Eg:

 -pattern 'Refused connections' '/refused connection/'


=item B<-merge>

This option allows eximstats to merge old eximstat reports together. Eg:

 eximstats mainlog.sun > report.sun.txt
 eximstats mainlog.mon > report.mon.txt
 eximstats mainlog.tue > report.tue.txt
 eximstats mainlog.wed > report.web.txt
 eximstats mainlog.thu > report.thu.txt
 eximstats mainlog.fri > report.fri.txt
 eximstats mainlog.sat > report.sat.txt
 eximstats -merge       report.*.txt > weekly_report.txt
 eximstats -merge -html report.*.txt > weekly_report.html

=over 4

=item *

You can merge text or html reports and output the results as text or html.

=item *

You can use all the normal eximstat output options, but only data
included in the original reports can be shown!

=item *

When merging reports, some loss of accuracy may occur in the top I<n> lists.
This will be towards the ends of the lists.

=item *

The order of items in the top I<n> lists may vary when the data volumes
round to the same value.

=back

=item B<-html>

Output the results in HTML.

=item B<-charts>

Create graphical charts to be displayed in HTML output.

This requires the following modules which can be obtained
from http://www.cpan.org/modules/01modules.index.html

=over 4

=item GD

=item GDTextUtil

=item GDGraph

=back

To install these, download and unpack them, then use the normal perl installation procedure:

 perl Makefile.PL
 make
 make test
 make install

=item B<-chartdir>I <dir>

Create the charts in the directory <dir>

=item B<-chartrel>I <dir>

Specify the relative directory for the "img src=" tags from where to include
the charts

=item B<-d>

Debug flag. This outputs the eval()'d parser onto STDOUT which makes it
easier to trap errors in the eval section. Remember to add 1 to the line numbers to allow for the
title!

=back

=head1 DESCRIPTION

Eximstats parses exim mainlog and syslog files to output a statistical
analysis of the messages processed. By default, a text
analysis is generated, but you can request an html analysis
by using the B<-html> flag. See the help (B<-help>) to learn
about how to create charts from the tables.

=head1 AUTHOR

There is a web site at http://www.exim.org - this contains details of the
mailing list exim-users@exim.org.

=head1 TO DO

This program does not perfectly handle messages whose received
and delivered log lines are in different files, which can happen
when you have multiple mail servers and a message cannot be
immeadiately delivered. Fixing this could be tricky...

=cut

use integer;
use strict;

# use Time::Local;  # PH/FANF
use POSIX;

use vars qw($HAVE_GD_Graph_pie $HAVE_GD_Graph_linespoints);
eval { require GD::Graph::pie; };
$HAVE_GD_Graph_pie = $@ ? 0 : 1;
eval { require GD::Graph::linespoints; };
$HAVE_GD_Graph_linespoints = $@ ? 0 : 1;


##################################################
#             Static data                        #
##################################################
# 'use vars' instead of 'our' as perl5.005 is still in use out there!
use vars qw(@tab62 @days_per_month $gig);
use vars qw($VERSION);
use vars qw($COLUMN_WIDTHS);


@tab62 =
  (0,1,2,3,4,5,6,7,8,9,0,0,0,0,0,0,     # 0-9
   0,10,11,12,13,14,15,16,17,18,19,20,  # A-K
  21,22,23,24,25,26,27,28,29,30,31,32,  # L-W
  33,34,35, 0, 0, 0, 0, 0,              # X-Z
   0,36,37,38,39,40,41,42,43,44,45,46,  # a-k
  47,48,49,50,51,52,53,54,55,56,57,58,  # l-w
  59,60,61);                            # x-z

@days_per_month = (0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334);
$gig     = 1024 * 1024 * 1024;
$VERSION = '1.35';

# How much space do we allow for the Hosts/Domains/Emails/Edomains column headers?
$COLUMN_WIDTHS = 8;

# Declare global variables.
use vars qw($total_received_data  $total_received_data_gigs  $total_received_count);
use vars qw($total_delivered_data $total_delivered_data_gigs $total_delivered_count);
use vars qw(%arrival_time %size %from_host %from_address);
use vars qw(%timestamp2time);			#Hash of timestamp => time.
use vars qw($last_timestamp $last_time);	#The last time convertion done.
use vars qw($last_date $date_seconds);		#The last date convertion done.
use vars qw($last_offset $offset_seconds);	#The last time offset convertion done.
use vars qw($localtime_offset);
use vars qw($i);				#General loop counter.
use vars qw($debug);				#Debug mode?
use vars qw($ntopchart);			#How many entries should make it into the chart?
use vars qw($gddirectory);			#Where to put files from GD::Graph

$ntopchart = 5;

# The following are parameters whose values are
# set by command line switches:
use vars qw($show_errors $show_relay $show_transport $transport_pattern);
use vars qw($topcount $local_league_table $include_remote_users);
use vars qw($hist_opt $hist_interval $hist_number $volume_rounding);
use vars qw($relay_pattern @queue_times $html @user_patterns @user_descriptions);

use vars qw(%do_sender);                #Do sender by Host, Domain, Email, and/or Edomain tables.
use vars qw($charts $chartrel $chartdir $charts_option_specified);
use vars qw($merge_reports);			#Merge old reports ?

# The following are modified in the parse() routine, and
# referred to in the print_*() routines.
use vars qw($queue_more_than $delayed_count $relayed_unshown $begin $end);
use vars qw(%received_count       %received_data       %received_data_gigs);
use vars qw(%delivered_count      %delivered_data      %delivered_data_gigs);
use vars qw(%received_count_user  %received_data_user  %received_data_gigs_user);
use vars qw(%delivered_count_user %delivered_data_user %delivered_data_gigs_user);
use vars qw(%transported_count    %transported_data    %transported_data_gigs);
use vars qw(%remote_delivered %relayed %delayed %had_error %errors_count);
use vars qw(@queue_bin @remote_queue_bin @received_interval_count @delivered_interval_count);
use vars qw(@user_pattern_totals);

use vars qw(%report_totals);


##################################################
#                   Subroutines                  #
##################################################


#######################################################################
# volume_rounded();
#
# $rounded_volume = volume_rounded($bytes,$gigabytes);
#
# Given a data size in bytes, round it to KB, MB, or GB
# as appropriate.
#
# Eg 12000 => 12KB, 15000000 => 14GB, etc.
#
# Note: I've experimented with Math::BigInt and it results in a 33%
# performance degredation as opposed to storing numbers split into
# bytes and gigabytes.
#######################################################################
sub volume_rounded {
  my($x,$g) = @_;
  $x = 0 unless $x;
  $g = 0 unless $g;
  my($rounded);

  while ($x > $gig) {
    $g++;
    $x -= $gig;
  }

  if ($volume_rounding) {
    # Values < 1 GB
    if ($g <= 0) {
      if ($x < 10000) {
        $rounded = sprintf("%6d", $x);
      }
      elsif ($x < 10000000) {
        $rounded = sprintf("%4dKB", ($x + 512)/1024);
      }
      else {
        $rounded = sprintf("%4dMB", ($x + 512*1024)/(1024*1024));
      }
    }
    # Values between 1GB and 10GB are printed in MB
    elsif ($g < 10) {
      $rounded = sprintf("%4dMB", ($g * 1024) + ($x + 512*1024)/(1024*1024));
    }
    else {
      # Handle values over 10GB
      $rounded = sprintf("%4dGB", $g + ($x + $gig/2)/$gig);
    }
  }
  else {
    # We don't want any rounding to be done.
    $rounded = sprintf("%4d", ($g * $gig) + $x);
  }

  return $rounded;
}


#######################################################################
# un_round();
# 
#  un_round($rounded_volume,\$bytes,\$gigabytes);
# 
# Given a volume in KB, MB or GB, as generated by volume_rounded(),
# do the reverse transformation and convert it back into Bytes and Gigabytes.
# These are added to the $bytes and $gigabytes parameters.
# 
# Given a data size in bytes, round it to KB, MB, or GB
# as appropriate.
# 
# EG: 500 => (500,0), 14GB => (0,14), etc.
#######################################################################
sub un_round {
  my($rounded,$bytes_sref,$gigabytes_sref) = @_;

  if ($rounded =~ /(\d+)GB/) {
    $$gigabytes_sref += $1;
  }
  elsif ($rounded =~ /(\d+)MB/) {
    $$gigabytes_sref +=   $1 / 1024;
    $$bytes_sref     += (($1 % 1024 ) * 1024 * 1024);
  }
  elsif ($rounded =~ /(\d+)KB/) {
    $$gigabytes_sref +=  $1 / (1024 * 1024);
    $$bytes_sref     += ($1 % (1024 * 1024) * 1024);
  }
  elsif ($rounded =~ /(\d+)/) {
    $$gigabytes_sref += $1 / $gig;
    $$bytes_sref     += $1 % $gig;
  }

  #Now reduce the bytes down to less than 1GB.
  add_volume($bytes_sref,$gigabytes_sref,0) if ($$bytes_sref > $gig);
}


#######################################################################
# add_volume();
# 
#   add_volume(\$bytes,\$gigs,$size);
# 
# Add $size to $bytes/$gigs where this is a number split into
# bytes ($bytes) and gigabytes ($gigs). This is significantly
# faster than using Math::BigInt.
#######################################################################
sub add_volume {
  my($bytes_ref,$gigs_ref,$size) = @_;
  $$bytes_ref = 0 if ! defined $$bytes_ref;
  $$gigs_ref = 0 if ! defined $$gigs_ref;
  $$bytes_ref += $size;
  while ($$bytes_ref > $gig) {
    $$gigs_ref++;
    $$bytes_ref -= $gig;
  }
}


#######################################################################
# format_time();
# 
#  $formatted_time = format_time($seconds);
# 
# Given a time in seconds, break it down into
# weeks, days, hours, minutes, and seconds.
# 
# Eg 12005 => 3h20m5s
#######################################################################
sub format_time {
my($t) = pop @_;
my($s) = $t % 60;
$t /= 60;
my($m) = $t % 60;
$t /= 60;
my($h) = $t % 24;
$t /= 24;
my($d) = $t % 7;
my($w) = $t/7;
my($p) = "";
$p .= "$w"."w" if $w > 0;
$p .= "$d"."d" if $d > 0;
$p .= "$h"."h" if $h > 0;
$p .= "$m"."m" if $m > 0;
$p .= "$s"."s" if $s > 0 || $p eq "";
$p;
}


#######################################################################
#  unformat_time();
# 
#  $seconds = unformat_time($formatted_time);
# 
# Given a time in weeks, days, hours, minutes, or seconds, convert it to seconds.
# 
# Eg 3h20m5s => 12005
#######################################################################
sub unformat_time {
  my($formated_time) = pop @_;
  my $time = 0;

  while ($formated_time =~ s/^(\d+)([wdhms]?)//) {
    $time +=  $1 if ($2 eq '' || $2 eq 's');
    $time +=  $1 * 60 if ($2 eq 'm');
    $time +=  $1 * 60 * 60 if ($2 eq 'h');
    $time +=  $1 * 60 * 60 * 24 if ($2 eq 'd');
    $time +=  $1 * 60 * 60 * 24  * 7 if ($2 eq 'w');
  }
  $time;
}


#######################################################################
# seconds();
# 
#  $time = seconds($timestamp);
# 
# Given a time-of-day timestamp, convert it into a time() value using
# POSIX::mktime.  We expect the timestamp to be of the form
# "$year-$mon-$day $hour:$min:$sec", with month going from 1 to 12,
# and the year to be absolute (we do the necessary conversions). The
# timestamp may be followed with an offset from UTC like "+$hh$mm"; if the
# offset is not present, and we have not been told that the log is in UTC
# (with the -utc option), then we adjust the time by the current local
# time offset so that it can be compared with the time recorded in message
# IDs, which is UTC.
# 
# To improve performance, we only use mktime on the date ($year-$mon-$day),
# and only calculate it if the date is different to the previous time we
# came here. We then add on seconds for the '$hour:$min:$sec'.
# 
# We also store the results of the last conversion done, and only
# recalculate if the date is different.
# 
# We used to have the '-cache' flag which would store the results of the
# mktime() call. However, the current way of just using mktime() on the
# date obsoletes this.
#######################################################################
sub seconds {
  my($timestamp) = @_;

  # Is the timestamp the same as the last one?
  return $last_time if ($last_timestamp eq $timestamp);

  return 0 unless ($timestamp =~ /^((\d{4})\-(\d\d)-(\d\d))\s(\d\d):(\d\d):(\d\d)( ([+-])(\d\d)(\d\d))?/o);

  unless ($last_date eq $1) {
    $last_date = $1;
    my(@timestamp) = (0,0,0,$4,$3,$2);
    $timestamp[5] -= 1900;
    $timestamp[4]--;
    $date_seconds = mktime(@timestamp);
  }
  my $time = $date_seconds + ($5 * 3600) + ($6 * 60) + $7;

  # SC. Use cacheing. Also note we want seconds not minutes.
  #my($this_offset) = ($10 * 60 + $11) * ($9 . "1") if defined $8;
  if (defined $8 && ($8 ne $last_offset)) {
    $last_offset = $8;
    $offset_seconds = ($10 * 60 + $11) * 60;
    $offset_seconds = -$offset_seconds if ($9 eq '-');
  }


  if (defined $7) {
    #$time -= $this_offset;
    $time -= $offset_seconds;
  } elsif (defined $localtime_offset) {
    $time -= $localtime_offset;
  }

  # Store the last timestamp received.
  $last_timestamp = $timestamp;
  $last_time      = $time;

  $time;
}


#######################################################################
#  id_seconds();
# 
#  $time = id_seconds($message_id);
# 
# Given a message ID, convert it into a time() value.
#######################################################################
sub id_seconds {
my($sub_id) = substr((pop @_), 0, 6);
my($s) = 0;
my(@c) = split(//, $sub_id);
while($#c >= 0) { $s = $s * 62 + $tab62[ord(shift @c) - ord('0')] }
$s;
}


#######################################################################
#  calculate_localtime_offset();
# 
#  $localtime_offset = calculate_localtime_offset();
# 
# Calculate the the localtime offset from gmtime in seconds.
# 
#  $localtime = time() + $localtime_offset.
# 
# These are the same semantics as ISO 8601 and RFC 2822 timezone offsets.
# (West is negative, East is positive.)
#######################################################################

# $localtime = gmtime() + $localtime_offset.  OLD COMMENT
# This subroutine commented out as it's not currently in use.

#sub calculate_localtime_offset {
#  # Pick an arbitrary date, convert it to localtime & gmtime, and return the difference.
#  my (@sample_date) = (0,0,0,5,5,100);
#  my $localtime = timelocal(@sample_date);
#  my $gmtime    = timegm(@sample_date);
#  my $offset = $localtime - $gmtime;
#  return $offset;
#}

sub calculate_localtime_offset {
  # Assume that the offset at the moment is valid across the whole
  # period covered by the logs that we're analysing. This may not
  # be true around the time the clocks change in spring or autumn.
  my $utc = time;
  # mktime works on local time and gmtime works in UTC
  my $local = mktime(gmtime($utc));
  return $local - $utc;
}


#######################################################################
# print_queue_times();
# 
#  $time = print_queue_times($message_type,\@queue_times,$queue_more_than);
# 
# Given the type of messages being output, the array of message queue times,
# and the number of messages which exceeded the queue times, print out
# a table.
#######################################################################
sub print_queue_times {
no integer;
my($string,$array,$queue_more_than) = @_;
my(@chartdatanames);
my(@chartdatavals);

my $printed_one = 0;
my $cumulative_percent = 0;
#$queue_unknown += keys %arrival_time;

my $queue_total = $queue_more_than;
for ($i = 0; $i <= $#queue_times; $i++) { $queue_total += $$array[$i] }

my $temp = "Time spent on the queue: $string";

my($format);
if ($html) {
  print "<hr><a name=\"$string time\"></a><h2>$temp</h2>\n";
  print "<table border=0 width=\"100%\">\n";
  print "<tr><td>\n";
  print "<table border=1>\n";
  print "<tr><th>Time</th><th>Messages</th><th>Percentage</th><th>Cumulative Percentage</th>\n";
  $format = "<tr><td align=\"right\">%s %s</td><td align=\"right\">%d</td><td align=\"right\">%5.1f%%</td><td align=\"right\">%5.1f%%</td>\n";
}
else
{
  printf("%s\n%s\n\n", $temp, "-" x length($temp));
  $format = "%5s %4s   %6d %5.1f%%  %5.1f%%\n";
}

for ($i = 0; $i <= $#queue_times; $i++) {
  if ($$array[$i] > 0)
    {
    my $percent = ($$array[$i] * 100)/$queue_total;
    $cumulative_percent += $percent;
    printf($format,
      $printed_one? "     " : "Under",
      format_time($queue_times[$i]),
      $$array[$i], $percent, $cumulative_percent);
    if (!defined($queue_times[$i])) {
      print "Not defined";
    }
    push(@chartdatanames,
      ($printed_one? "" : "Under") . format_time($queue_times[$i]));
    push(@chartdatavals, $$array[$i]);
    $printed_one = 1;
  }
}

if ($queue_more_than > 0) {
  my $percent = ($queue_more_than * 100)/$queue_total;
  $cumulative_percent += $percent;
  printf($format,
    "Over ",
    format_time($queue_times[$#queue_times]),
    $queue_more_than, $percent, $cumulative_percent);
}
push(@chartdatanames, "Over " . format_time($queue_times[$#queue_times]));
push(@chartdatavals, $queue_more_than);

#printf("Unknown   %6d\n", $queue_unknown) if $queue_unknown > 0;
if ($html) {
  print "</table>\n";
  print "</td><td>\n";

  if ($HAVE_GD_Graph_pie && $charts) {
    my @data = (
       \@chartdatanames,
       \@chartdatavals
    );
    my $graph = GD::Graph::pie->new(200, 200);
    my $pngname;
    my $title;
    if ($string =~ /all/) { $pngname = "queue_all.png"; $title = "Queue (all)"; }
    if ($string =~ /remote/) { $pngname = "queue_rem.png"; $title = "Queue (remote)"; }
    $graph->set(
        title             => $title,
    );
    my $gd = $graph->plot(\@data) or warn($graph->error);
    if ($gd) {
      open(IMG, ">$chartdir/$pngname") or die $!;
      binmode IMG;
      print IMG $gd->png;
      close IMG;
      print "<img src=\"$chartrel/$pngname\">";
    }
  }
  print "</td></tr></table>\n";
}
print "\n";
}


#######################################################################
# print_histogram();
# 
#  print_histogram('Deliverieds|Messages received',@interval_count);
# 
# Print a histogram of the messages delivered/received per time slot
# (hour by default).
#######################################################################
sub print_histogram {
my($text) = shift;
my(@interval_count) = @_;
my(@chartdatanames);
my(@chartdatavals);
my($maxd) = 0;
for ($i = 0; $i < $hist_number; $i++)
  { $maxd = $interval_count[$i] if $interval_count[$i] > $maxd; }

my $scale = int(($maxd + 25)/50);
$scale = 1 if $scale == 0;

my($type);
if ($text eq "Deliveries")
  {
  $type = ($scale == 1)? "delivery" : "deliveries";
  }
else
  {
  $type = ($scale == 1)? "message" : "messages";
  }

my($title) = sprintf("$text per %s (each dot is $scale $type)",
  ($hist_interval == 60)? "hour" :
  ($hist_interval == 1)?  "minute" : "$hist_interval minutes");

if ($html) {
  print "<hr><a name=\"$text\"></a><h2>$title</h2>\n";
  print "<table border=0 width=\"100%\">\n";
  print "<tr><td><pre>\n";
}
else {
  printf("%s\n%s\n\n", $title, "-" x length($title));
}

my $hour = 0;
my $minutes = 0;
for ($i = 0; $i < $hist_number; $i++)
  {
  my $c = $interval_count[$i];

  # If the interval is an hour (the maximum) print the starting and
  # ending hours as a label. Otherwise print the starting hour and
  # minutes, which take up the same space.

  my $temp;
  if ($hist_opt == 1)
    {
    $temp = sprintf("%02d-%02d", $hour, $hour + 1);
    print $temp;
    push(@chartdatanames, $temp);
    $hour++;
    }
  else
    {
    if ($minutes == 0)
      { $temp = sprintf("%02d:%02d", $hour, $minutes) }
    else
      { $temp = sprintf("  :%02d", $minutes) }
    print $temp;
    push(@chartdatanames, $temp);
    $minutes += $hist_interval;
    if ($minutes >= 60)
      {
      $minutes = 0;
      $hour++;
      }
    }
  push(@chartdatavals, $c);
  printf(" %6d %s\n", $c, "." x ($c/$scale));
  }
print "\n";
if ($html)
  {
  print "</pre>\n";
  print "</td><td>\n";
  if ($HAVE_GD_Graph_linespoints && $charts) {
    # calculate the graph
    my @data = (
       \@chartdatanames,
       \@chartdatavals
    );
    my $graph = GD::Graph::linespoints->new(300, 300);
    $graph->set(
        x_label           => 'Time',
        y_label           => 'Amount',
        title             => $text,
        x_labels_vertical => 1
    );
    my($pngname);
    if ($text =~ /Deliveries/) { $pngname = "histogram_del.png"; }
    if ($text =~ /Messages/)   { $pngname = "histogram_mes.png"; }
    my $gd = $graph->plot(\@data) or warn($graph->error);
    if ($gd) {
      open(IMG, ">$chartdir/$pngname") or die $!;
      binmode IMG;
      print IMG $gd->png;
      close IMG;
      print "<img src=\"$chartrel/$pngname\">";
    }
  }
  print "</td></tr></table>\n";
}
}


#######################################################################
# print_league_table();
# 
#  print_league_table($league_table_type,\%message_count,\%message_data,\%message_data_gigs);
# 
# Given hashes of message count and message data, which are keyed by
# the table type (eg by the sending host), print a league table
# showing the top $topcount (defaults to 50).
#######################################################################
sub print_league_table {
my($text,$m_count,$m_data,$m_data_gigs) = @_;
my($name) = ($topcount == 1)? "$text" : "$topcount ${text}s";
my($temp) = "Top $name by message count";
my(@chartdatanames) = ();
my(@chartdatavals) = ();
my $chartotherval = 0;

my($format);
if ($html) {
  print "<hr><a name=\"$text count\"></a><h2>$temp</h2>\n";
  print "<table border=0 width=\"100%\">\n";
  print "<tr><td>\n";
  print "<table border=1>\n";
  print "<tr><th>Messages</th><th>Bytes</th><th>\u$text</th>\n";

  # Align non-local addresses to the right (so all the .com's line up).
  # Local addresses are aligned on the left as they are userids.
  my $align = ($text !~ /local/i) ? 'right' : 'left';
  $format = "<tr><td align=\"right\">%d</td><td align=\"right\">%s</td><td align=\"$align\" nowrap>%s</td>\n";
}
else {
  printf("%s\n%s\n\n", $temp, "-" x length($temp));
  $format = "%7d %10s   %s\n";
}

my($key,$htmlkey);
foreach $key (top_n_sort($topcount,$m_count,$m_data_gigs,$m_data)) {
  if ($html) {
    $htmlkey = $key;
    $htmlkey =~ s/>/\&gt\;/g;
    $htmlkey =~ s/</\&lt\;/g;
    printf($format, $$m_count{$key}, volume_rounded($$m_data{$key},$$m_data_gigs{$key}), $htmlkey);
  }
  else {
    printf($format, $$m_count{$key}, volume_rounded($$m_data{$key},$$m_data_gigs{$key}), $key);
  }
  if (scalar @chartdatanames < $ntopchart)
    {
    push(@chartdatanames, $key);
    push(@chartdatavals, $$m_count{$key});
    }
  else
    {
    $chartotherval += $$m_count{$key};
    }
  }
push(@chartdatanames, "Other");
push(@chartdatavals, $chartotherval);

if ($html)
  {
  print "</table>\n";
  print "</td><td>\n";
  if ($HAVE_GD_Graph_pie && $charts)
    {
    # calculate the graph
    my @data = (
       \@chartdatanames,
       \@chartdatavals
    );
    my $graph = GD::Graph::pie->new(300, 300);
    $graph->set(
        x_label           => 'Name',
        y_label           => 'Amount',
        title             => 'By count',
    );
    my $gd = $graph->plot(\@data) or warn($graph->error);
    if ($gd) {
      my $temp = $text;
      $temp =~ s/ /_/g;
      open(IMG, ">$chartdir/${temp}_count.png") or die $!;
      binmode IMG;
      print IMG $gd->png;
      close IMG;
      print "<img src=\"$chartrel/${temp}_count.png\">";
    }
  }
  print "</td><td>\n";
  print "</td></tr></table>\n";
}
print "\n";

$temp = "Top $name by volume";
if ($html) {
  print "<hr><a name=\"$text volume\"></a><h2>$temp</h2>\n";
  print "<table border=0 width=\"100%\">\n";
  print "<tr><td>\n";
  print "<table border=1>\n";
  print "<tr><th>Messages</th><th>Bytes</th><th>\u$text</th>\n";
}
else {
  printf("%s\n%s\n\n", $temp, "-" x length($temp));
}

@chartdatanames = ();
@chartdatavals = ();
$chartotherval = 0;
my $use_gig = 0;
foreach $key (top_n_sort($topcount,$m_data_gigs,$m_data,$m_count)) {
  # The largest volume will be the first (top of the list).
  # If it has at least 1 gig, then just use gigabytes to avoid
  # risking an integer overflow when generating the pie charts. 
  if ($$m_data_gigs{$key}) {
    $use_gig = 1;
  }
  if ($html) {
    $htmlkey = $key;
    $htmlkey =~ s/>/\&gt\;/g;
    $htmlkey =~ s/</\&lt\;/g;
    printf($format, $$m_count{$key}, volume_rounded($$m_data{$key},$$m_data_gigs{$key}), $htmlkey);
  }
  else {
    printf($format, $$m_count{$key}, volume_rounded($$m_data{$key},$$m_data_gigs{$key}), $key);
  }

  if (scalar @chartdatanames < $ntopchart) {
    if ($use_gig) {
      if ($$m_data_gigs{$key}) {
	push(@chartdatanames, $key);
	push(@chartdatavals, $$m_data_gigs{$key});
      }
    }
    else {
      push(@chartdatanames, $key);
      push(@chartdatavals, $$m_data{$key});
    }
  }
  else {
    $chartotherval += ($use_gig) ? $$m_data_gigs{$key} : $$m_data{$key};
  }
}
push(@chartdatanames, "Other");
push(@chartdatavals, $chartotherval);

if ($html) {
  print "</table>\n";
  print "</td><td>\n";
  if ($HAVE_GD_Graph_pie && $charts) {
    # calculate the graph
    my @data = (
       \@chartdatanames,
       \@chartdatavals
    );
    my $graph = GD::Graph::pie->new(300, 300);
    $graph->set(
        x_label           => 'Name',
        y_label           => 'Volume' ,
        title             => 'By Volume',
    );
    my $gd = $graph->plot(\@data) or warn($graph->error);
    if ($gd) {
      my $temp = $text;
      $temp =~ s/ /_/g;
      open(IMG, ">$chartdir/${temp}_volume.png") or die $!;
      binmode IMG;
      print IMG $gd->png;
      close IMG;
      print "<img src=\"$chartrel/${temp}_volume.png\">";
    }
  }
  print "</td><td>\n";
  print "</td></tr></table>\n";
}

print "\n";
}


#######################################################################
# top_n_sort();
# 
#   @sorted_keys = top_n_sort($n,$href1,$href2,$href3);
# 
# Given a hash which has numerical values, return the sorted $n keys which
# point to the top values. The second and third hashes are used as
# tiebreakers. They all must have the same keys.
# 
# The idea behind this routine is that when you only want to see the
# top n members of a set, rather than sorting the entire set and then
# plucking off the top n, sort through the stack as you go, discarding
# any member which is lower than your current n'th highest member.
# 
# This proves to be an order of magnitude faster for large hashes.
# On 200,000 lines of mainlog it benchmarked 9 times faster.
# On 700,000 lines of mainlog it benchmarked 13.8 times faster.
# 
# We assume the values are > 0.
#######################################################################
sub top_n_sort {
  my($n,$href1,$href2,$href3) = @_;

  # PH's original sort was:
  #
  # foreach $key (sort
  #               {
  #               $$m_count{$b}     <=> $$m_count{$a} ||
  #               $$m_data_gigs{$b} <=> $$m_data_gigs{$a}  ||
  #               $$m_data{$b}      <=> $$m_data{$a}  ||
  #               $a cmp $b
  #               }
  #             keys %{$m_count})
  #

  #We use a key of '_' to represent non-existant values, as null keys are valid.
  #'_' is not a valid domain, edomain, host, or email.
  my(@top_n_keys) = ('_') x $n;
  my($minimum_value1,$minimum_value2,$minimum_value3) = (0,0,0);
  my $top_n_key = '';
  my $n_minus_1 = $n - 1;
  my $n_minus_2 = $n - 2;

  # Pick out the top $n keys. 
  my($key,$value1,$value2,$value3,$i,$comparison,$insert_position);
  while (($key,$value1) = each %$href1) {

    #print STDERR "key $key ($value1,",$href2->{$key},",",$href3->{$key},") <=> ($minimum_value1,$minimum_value2,$minimum_value3)\n";
    
    # Check to see that the new value is bigger than the lowest of the
    # top n keys that we're keeping.
    $comparison = $value1        <=> $minimum_value1 || 
		  $href2->{$key} <=> $minimum_value2 ||
		  $href3->{$key} <=> $minimum_value3 ||
		  $top_n_key cmp $key;
    next unless ($comparison == 1);

    # As we will be using these values a few times, extract them into scalars.
    $value2 = $href2->{$key};
    $value3 = $href3->{$key};

    # This key is bigger than the bottom n key, so the lowest position we
    # will insert it into is $n minus 1 (the bottom of the list).
    $insert_position = $n_minus_1;

    # Now go through the list, stopping when we find a key that we're
    # bigger than, or we come to the penultimate position - we've
    # already tested bigger than the last.
    #
    # Note: we go top down as the list starts off empty.
    # Note: stepping through the list in this way benchmarks nearly
    # three times faster than doing a sort() on the reduced list.
    # I assume this is because the list is already in order, and
    # we get a performance boost from not having to do hash lookups
    # on the new key.
    for ($i = 0; $i < $n_minus_1; $i++) {
      $top_n_key = $top_n_keys[$i];
      if ( ($top_n_key eq '_') ||
	   ( ($value1 <=> $href1->{$top_n_key} || 
              $value2 <=> $href2->{$top_n_key} ||
	      $value3 <=> $href3->{$top_n_key} ||
	      $top_n_key cmp $key) == 1
	   )
	 ) {
	$insert_position = $i;
	last;
      }
    }

    # Remove the last element, then insert the new one.
    $#top_n_keys = $n_minus_2;
    splice(@top_n_keys,$insert_position,0,$key);

    # Extract our new minimum values.
    $top_n_key = $top_n_keys[$n_minus_1];
    if ($top_n_key ne '_') {
      $minimum_value1 = $href1->{$top_n_key};
      $minimum_value2 = $href2->{$top_n_key};
      $minimum_value3 = $href3->{$top_n_key};
    }
  }

  # Return the top n list, grepping out non-existant values, just in case
  # we didn't have that many values.
  return(grep(!/^_$/,@top_n_keys));
}


#######################################################################
# html_header();
# 
#  $header = html_header($title);
# 
# Print our HTML header and start the <body> block.
#######################################################################
sub html_header {
  my($title) = @_;
  my $text = << "EoText";
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-15">
<title>$title</title>
</head>
<body bgcolor="white">
<h1>$title</h1>
EoText
  return $text;
}


#######################################################################
# help();
# 
#  help();
# 
# Display usage instructions and exit.
#######################################################################
sub help {
  print << "EoText";

eximstats Version $VERSION

Usage: eximstats [Options] mainlog1 mainlog2 ... > report.txt
       eximstats -html  [Options] mainlog1 mainlog2 ... > report.html
       eximstats -merge [Options] report.1.txt report.2.txt ... > weekly_rep.txt
       eximstats -merge -html [Options] report.1.html ... > weekly_rep.html

Parses exim mainlog or syslog files and generates a statistical analysis
of the messages processed. Valid options are:

-h<number>      histogram divisions per hour. The default is 1, and
                0 suppresses histograms. Other valid values are:
		2, 3, 5, 10, 15, 20, 30 or 60.
-ne             don't display error information
-nr             don't display relaying information
-nr/pattern/    don't display relaying information that matches
-nt             don't display transport information
-nt/pattern/    don't display transport information that matches
-nvr		don't do volume rounding. Display in bytes, not KB/MB/GB.
-q<list>        list of times for queuing information
                single 0 item suppresses
-t<number>      display top <number> sources/destinations
                default is 50, 0 suppresses top listing
-tnl            omit local sources/destinations in top listing
-t_remote_users show top user sources/destinations from non-local domains

-byhost		show results by sending host (default unless bydomain or
                byemail is specified)
-bydomain	show results by sending domain.
-byemail	show results by sender's email address
-byedomain	show results by sender's email domain

-pattern "Description" /pattern/
                Count lines matching specified patterns and show them in
		the results. It can be specified multiple times. Eg:
		-pattern 'Refused connections' '/refused connection/'

-merge          merge previously generated reports into a new report

-html           output the results in HTML
-charts         Create charts (this requires the GD::Graph modules)
-chartdir <dir> Create the charts' png files in the directory <dir>
-chartrel <dir> Specify the relative directory for the "img src=" tags
                from where to include the charts in the html file
		-chartdir and -chartrel default to '.'

-d		Debug mode - dump the eval'ed parser onto STDERR.

EoText

  exit 1;
}


#######################################################################
# generate_parser();
# 
#  $parser = generate_parser();
# 
# This subroutine generates the parsing routine which will be
# used to parse the mainlog. We take the base operation, and remove bits not in use.
# This improves performance depending on what bits you take out or add.
# 
# I've tested using study(), but this does not improve performance.
# 
# We store our parsing routing in a variable, and process it looking for #IFDEF (Expression)
# or #IFNDEF (Expression) statements and corresponding #ENDIF (Expression) statements. If
# the expression evaluates to true, then it is included/excluded accordingly.
#######################################################################
sub generate_parser {
  my $parser = '
  my($ip,$host,$email,$edomain,$domain,$thissize,$size,$old,$new);
  my($tod,$m_hour,$m_min,$id,$flag);
  while (<$fh>) {

    # Convert syslog lines to mainlog format.
    if (! /^\\d{4}/) {
      next unless s/^.*? exim\\b.*?: //;
    }

    next if length($_) < 38;
    next unless /^(\\d{4}\\-\\d\\d-\\d\\d\\s(\\d\\d):(\\d\\d):\\d\\d( [-+]\\d\\d\\d\\d)?)/o;

    ($tod,$m_hour,$m_min) = ($1,$2,$3);

    # PH
    my($extra) = defined($4)? 6 : 0;
    $id   = substr($_, 20 + $extra, 16);
    $flag = substr($_, 37 + $extra, 2);
';

  # Watch for user specified patterns.
  my $user_pattern_index = 0;
  foreach (@user_patterns) {
    $user_pattern_totals[$user_pattern_index] = 0;
    $parser .= "  \$user_pattern_totals[$user_pattern_index]++ if $_;\n";
    $user_pattern_index++;
  }

  $parser .= '
    next unless ($flag =~ /<=|=>|->|==|\\*\\*|Co/);

    #Strip away the timestamp, ID and flag (which could be "Com" for completed)
    #This speeds up the later pattern matches.
    # $_ = substr($_, 40);

    $_ = substr($_, 40 + $extra);  # PH

    # JN - Skip over certain transports as specified via the "-nt/.../" command
    # line switch (where ... is a perl style regular expression).  This is
    # required so that transports that skew stats such as SpamAssassin can be
    # ignored.
    #IFDEF ($transport_pattern)
    if (/\\sT=(\\S+)/) {
       next if ($1 =~ /$transport_pattern/o) ;
    }
    #ENDIF ($transport_pattern)


    $host = "local";		#Host is local unless otherwise specified.
    $domain = "localdomain";	#Domain is localdomain unless otherwise specified.


    # Do some pattern matches to get the host and IP address.
    # We expect lines to be of the form "H=[IpAddr]" or "H=Host [IpAddr]" or
    # "H=Host (UnverifiedHost) [IpAddr]" or "H=(UnverifiedHost) [IpAddr]".
    # We do 2 separate matches to keep the matches simple and fast.
    if (/\\sH=(\\S+)/) {
      $host = $1;

      ($ip) = /\\sH=.*?(\\s\\[[^]]+\\])/;
      # If there is only an IP address, it will be in $host and $ip will be
      # unset. That is OK, because we only use $ip in conjunction with $host
      # below. But make it empty to avoid warning messages.
      $ip = "" if !defined $ip;

      #IFDEF ($do_sender{Domain})
      if ($host !~ /^\\[/ && $host =~ /^(\\(?)[^\\.]+\\.([^\\.]+\\..*)/) {
	# Remove the host portion from the DNS name. We ensure that we end up with
	# at least xxx.yyy. $host can be "(x.y.z)" or  "x.y.z".
	$domain = lc("$1.$2");
	$domain =~ s/^\\.//;		#Remove preceding dot.
      }
      #ENDIF ($do_sender{Domain})

    }

    #IFDEF ($do_sender{Email})
    $email = (/^(\S+)/) ? $1 : "";
    #ENDIF ($do_sender{Email})

    #IFDEF ($do_sender{Edomain})
    $edomain = (/^\S*?\\@(\S+)/) ? lc($1) : "";
    #ENDIF ($do_sender{Edomain})

    if ($tod lt $begin) {
      $begin = $tod;
    }
    elsif ($tod gt $end) {
      $end   = $tod;
    }


    if ($flag eq "<=") {
      $thissize = (/\\sS=(\\d+)( |$)/) ? $1 : 0;
      $size{$id} = $thissize;

      #IFDEF ($show_relay)
      if ($host ne "local") {
	# Save incoming information in case it becomes interesting
	# later, when delivery lines are read.
	my($from) = /^(\\S+)/;
	$from_host{$id} = "$host$ip";
	$from_address{$id} = $from;
      }
      #ENDIF ($show_relay)

      #IFDEF ($local_league_table || $include_remote_users)
	if (/\sU=(\\S+)/) {
	  my $user = $1;

	  #IFDEF ($local_league_table && $include_remote_users)
	  {				#Store both local and remote users.
	  #ENDIF ($local_league_table && $include_remote_users)

	  #IFDEF ($local_league_table && ! $include_remote_users)
	  if ($host eq "local") {		#Store local users only.
	  #ENDIF ($local_league_table && ! $include_remote_users)

	  #IFDEF ($include_remote_users && ! $local_league_table)
	  if ($host ne "local") {		#Store remote users only.
	  #ENDIF ($include_remote_users && ! $local_league_table)

	    $received_count_user{$user}++;
	    add_volume(\\$received_data_user{$user},\\$received_data_gigs_user{$user},$thissize);
          }
	}
      #ENDIF ($local_league_table || $include_remote_users)

      #IFDEF ($do_sender{Host})
	$received_count{Host}{$host}++;
	add_volume(\\$received_data{Host}{$host},\\$received_data_gigs{Host}{$host},$thissize);
      #ENDIF ($do_sender{Host})

      #IFDEF ($do_sender{Domain})
        if ($domain) {
	  $received_count{Domain}{$domain}++;
	  add_volume(\\$received_data{Domain}{$domain},\\$received_data_gigs{Domain}{$domain},$thissize);
	}
      #ENDIF ($do_sender{Domain})

      #IFDEF ($do_sender{Email})
	$received_count{Email}{$email}++;
	add_volume(\\$received_data{Email}{$email},\\$received_data_gigs{Email}{$email},$thissize);
      #ENDIF ($do_sender{Email})

      #IFDEF ($do_sender{Edomain})
	$received_count{Edomain}{$edomain}++;
	add_volume(\\$received_data{Edomain}{$edomain},\\$received_data_gigs{Edomain}{$edomain},$thissize);
      #ENDIF ($do_sender{Edomain})

      $total_received_count++;
      add_volume(\\$total_received_data,\\$total_received_data_gigs,$thissize);

      #IFDEF ($#queue_times >= 0)
	$arrival_time{$id} = $tod;
      #ENDIF ($#queue_times >= 0)

      #IFDEF ($hist_opt > 0)
	$received_interval_count[($m_hour*60 + $m_min)/$hist_interval]++;
      #ENDIF ($hist_opt > 0)
    }

    elsif ($flag eq "=>") {
      $size = $size{$id} || 0;
      if ($host ne "local") {
        $remote_delivered{$id} = 1;


        #IFDEF ($show_relay)
        # Determine relaying address if either only one address listed,
        # or two the same. If they are different, it implies a forwarding
        # or aliasing, which is not relaying. Note that for multi-aliased
        # addresses, there may be a further address between the first
        # and last.

        if (defined $from_host{$id}) {
          if (/^(\\S+)(?:\\s+\\([^)]\\))?\\s+<([^>]+)>/) {
            ($old,$new) = ($1,$2);
	  }
          else {
	    $old = $new = "";
	  }

          if ("\\L$new" eq "\\L$old") {
            ($old) = /^(\\S+)/ if $old eq "";
            my $key = "H=\\L$from_host{$id}\\E A=\\L$from_address{$id}\\E => " .
              "H=\\L$host\\E$ip A=\\L$old\\E";
            if (!defined $relay_pattern || $key !~ /$relay_pattern/o) {
              $relayed{$key} = 0 if !defined $relayed{$key};
              $relayed{$key}++;
	    }
            else {
	      $relayed_unshown++
            }
          }
        }
        #ENDIF ($show_relay)

      }

      #IFDEF ($local_league_table || $include_remote_users)
	#IFDEF ($local_league_table && $include_remote_users)
	{				#Store both local and remote users.
	#ENDIF ($local_league_table && $include_remote_users)

	#IFDEF ($local_league_table && ! $include_remote_users)
	if ($host eq "local") {		#Store local users only.
	#ENDIF ($local_league_table && ! $include_remote_users)

	#IFDEF ($include_remote_users && ! $local_league_table)
	if ($host ne "local") {		#Store remote users only.
	#ENDIF ($include_remote_users && ! $local_league_table)

	  if (my($user) = split((/\\s</)? " <" : " ", $_)) {
	    if ($user =~ /^[\\/|]/) {
	      my($parent) = $_ =~ /(<[^@]+@?[^>]*>)/;
	      $user = "$user $parent" if defined $parent;
	    }
	    $delivered_count_user{$user}++;
	    add_volume(\\$delivered_data_user{$user},\\$delivered_data_gigs_user{$user},$size);
	  }
	}
      #ENDIF ($local_league_table || $include_remote_users)

      #IFDEF ($do_sender{Host})
	$delivered_count{Host}{$host}++;
	add_volume(\\$delivered_data{Host}{$host},\\$delivered_data_gigs{Host}{$host},$size);
      #ENDIF ($do_sender{Host})
      #IFDEF ($do_sender{Domain})
        if ($domain) {
	  $delivered_count{Domain}{$domain}++;
	  add_volume(\\$delivered_data{Domain}{$domain},\\$delivered_data_gigs{Domain}{$domain},$size);
	}
      #ENDIF ($do_sender{Domain})
      #IFDEF ($do_sender{Email})
	$delivered_count{Email}{$email}++;
	add_volume(\\$delivered_data{Email}{$email},\\$delivered_data_gigs{Email}{$email},$size);
      #ENDIF ($do_sender{Email})
      #IFDEF ($do_sender{Edomain})
	$delivered_count{Edomain}{$edomain}++;
	add_volume(\\$delivered_data{Edomain}{$edomain},\\$delivered_data_gigs{Edomain}{$edomain},$size);
      #ENDIF ($do_sender{Edomain})

      $total_delivered_count++;
      add_volume(\\$total_delivered_data,\\$total_delivered_data_gigs,$size);

      #IFDEF ($show_transport)
        my $transport = (/\\sT=(\\S+)/) ? $1 : ":blackhole:";
        $transported_count{$transport}++;
        add_volume(\\$transported_data{$transport},\\$transported_data_gigs{$transport},$size);
      #ENDIF ($show_transport)

      #IFDEF ($hist_opt > 0)
        $delivered_interval_count[($m_hour*60 + $m_min)/$hist_interval]++;
      #ENDIF ($hist_opt > 0)

    }

    elsif ($flag eq "==" && defined($size{$id}) && !defined($delayed{$id})) {
      $delayed_count++;
      $delayed{$id} = 1;
    }

    elsif ($flag eq "**") {
      $had_error{$id} = 1 if defined ($size{$id});

      #IFDEF ($show_errors)
        $errors_count{$_}++;
      #ENDIF ($show_errors)

    }

    elsif ($flag eq "Co") {
      #Completed?
      #IFDEF ($#queue_times >= 0)
        #Note: id_seconds() benchmarks as 42% slower than seconds() and computing
        #the time accounts for a significant portion of the run time.
        my($queued);
        if (defined $arrival_time{$id}) {
          $queued = seconds($tod) - seconds($arrival_time{$id});
	  delete($arrival_time{$id});
        }
        else {
	  $queued = seconds($tod) - id_seconds($id);
        }

        for ($i = 0; $i <= $#queue_times; $i++) {
          if ($queued < $queue_times[$i]) {
            $queue_bin[$i]++;
            $remote_queue_bin[$i]++ if $remote_delivered{$id};
            last;
	  }
	}
        $queue_more_than++ if $i > $#queue_times;
      #ENDIF ($#queue_times >= 0)

      #IFDEF ($show_relay)
        delete($from_host{$id});
        delete($from_address{$id});
      #ENDIF ($show_relay)

    }
  }';

  # We now do a 'C preprocessor style operation on our parser
  # to remove bits not in use.
  my(%defines_in_operation,$removing_lines,$processed_parser);
  foreach (split (/\n/,$parser)) {
    if ((/^\s*#\s*IFDEF\s*\((.*?)\)/i  && ! eval $1) ||
	(/^\s*#\s*IFNDEF\s*\((.*?)\)/i &&   eval $1)	) {
      $defines_in_operation{$1} = 1;
      $removing_lines = 1;
    }

    $processed_parser .= $_."\n" unless $removing_lines;

    if (/^\s*#\s*ENDIF\s*\((.*?)\)/i) {
      delete $defines_in_operation{$1};
      unless (keys %defines_in_operation) {
	$removing_lines = 0;
      }
    }
  }
  print STDERR "# START OF PARSER:\n$processed_parser\n# END OF PARSER\n\n" if $debug;

  return $processed_parser;
}


#######################################################################
# parse();
# 
#  parse($parser,\*FILEHANDLE);
# 
# This subroutine accepts a parser and a filehandle from main and parses each
# line. We store the results into global variables.
#######################################################################
sub parse {
  my($parser,$fh) = @_;

  if ($merge_reports) {
    parse_old_eximstat_reports($fh);
  }
  else {
    eval $parser;
    die ($@) if $@;
  }

}


#######################################################################
# print_header();
# 
#  print_header();
# 
# Print our headers and contents.
#######################################################################
sub print_header {

  my $title = "Exim statistics from $begin to $end";

  if ($html) {
    print html_header($title);
    print "<ul>\n";
    print "<li><a href=\"#grandtotal\">Grand total summary</a>\n";
    print "<li><a href=\"#patterns\">User Specified Patterns</a>\n" if @user_patterns;
    print "<li><a href=\"#transport\">Deliveries by Transport</a>\n" if $show_transport;
    if ($hist_opt) {
      print "<li><a href=\"#Messages received\">Messages received per hour</a>\n";
      print "<li><a href=\"#Deliveries\">Deliveries per hour</a>\n";
    }
    if ($#queue_times >= 0) {
      print "<li><a href=\"#all messages time\">Time spent on the queue: all messages</a>\n";
      print "<li><a href=\"#messages with at least one remote delivery time\">Time spent on the queue: messages with at least one remote delivery</a>\n";
    }
    print "<li><a href=\"#Relayed messages\">Relayed messages</a>\n" if $show_relay;
    if ($topcount) {
      foreach ('Host','Domain','Email','Edomain') {
	next unless $do_sender{$_};
	print "<li><a href=\"#sending \l$_ count\">Top $topcount sending \l${_}s by message count</a>\n";
	print "<li><a href=\"#sending \l$_ volume\">Top $topcount sending \l${_}s by volume</a>\n";
      }
      if ($local_league_table || $include_remote_users) {
	print "<li><a href=\"#local sender count\">Top $topcount local senders by message count</a>\n";
	print "<li><a href=\"#local sender volume\">Top $topcount local senders by volume</a>\n";
      }
      foreach ('Host','Domain','Email','Edomain') {
	next unless $do_sender{$_};
	print "<li><a href=\"#\l$_ destination count\">Top $topcount \l$_ destinations by message count</a>\n";
	print "<li><a href=\"#\l$_ destination volume\">Top $topcount \l$_ destinations by volume</a>\n";
      }
      if ($local_league_table || $include_remote_users) {
	print "<li><a href=\"#local destination count\">Top $topcount local destinations by message count</a>\n";
	print "<li><a href=\"#local destination volume\">Top $topcount local destinations by volume</a>\n";
      }
    }
    print "<li><a href=\"#errors\">List of errors</a>\n" if %errors_count;
    print "</ul>\n<hr>\n";

  }
  else {
    print "\n$title\n";
  }
}


#######################################################################
# print_grandtotals();
# 
#  print_grandtotals();
# 
# Print the grand totals.
#######################################################################
sub print_grandtotals {

  # Get the sender by headings and results. This is complicated as we can have
  # different numbers of columns.
  my($sender_txt_header,$sender_html_header,$sender_txt_format,$sender_html_format);
  my(@received_totals,@delivered_totals);
  foreach ('Host','Domain','Email','Edomain') {
    next unless $do_sender{$_};
    if ($merge_reports) {
      push(@received_totals, get_report_total($report_totals{Received},"${_}s"));
      push(@delivered_totals,get_report_total($report_totals{Delivered},"${_}s"));
    }
    else {
      push(@received_totals,scalar(keys %{$received_data{$_}}));
      push(@delivered_totals,scalar(keys %{$delivered_data{$_}}));
    }
    $sender_html_header .= "<th>${_}s</th>";
    $sender_txt_header  .= " " x ($COLUMN_WIDTHS - length($_)) . $_ . 's';
    $sender_html_format .= "<td align=\"right\">%d</td>";
    $sender_txt_format  .= " " x ($COLUMN_WIDTHS - 5) . "%6d";
  }

  my($format1,$format2);
  if ($html) {
    print << "EoText";
<a name="grandtotal"></a>
<h2>Grand total summary</h2>
<table border=1>
<tr><th>TOTAL</th><th>Volume</th><th>Messages</th>$sender_html_header<th colspan=2>At least one addr<br>Delayed</th><th colspan=2>At least one addr<br>Failed</th>
EoText

    $format1 = "<tr><td>%s</td><td align=\"right\">%s</td>$sender_html_format<td align=\"right\">%d</td>";
    $format2 = "<td align=\"right\">%d</td><td align=\"right\">%4.1f%%</td><td align=\"right\">%d</td><td align=\"right\">%4.1f%%</td>";
  }
  else {
    my $sender_spaces = " " x length($sender_txt_header);
    print << "EoText";

Grand total summary
-------------------
                                    $sender_spaces           At least one address
  TOTAL               Volume    Messages $sender_txt_header      Delayed       Failed
EoText
    $format1 = "  %-16s %9s      %6d $sender_txt_format";
    $format2 = "  %6d %4.1f%% %6d %4.1f%%",
  }

  my($volume,$failed_count);
  if ($merge_reports) {
    $volume = volume_rounded($report_totals{Received}{Volume}, $report_totals{Received}{'Volume-gigs'});
    $total_received_count = get_report_total($report_totals{Received},'Messages');
    $failed_count  = get_report_total($report_totals{Received},'Failed');
    $delayed_count = get_report_total($report_totals{Received},'Delayed');
  }
  else {
    $volume = volume_rounded($total_received_data, $total_received_data_gigs);
    $failed_count = keys %had_error;
  }

  {
    no integer;
    printf("$format1$format2\n",'Received',$volume,$total_received_count,
      @received_totals,$delayed_count,
      ($total_received_count) ? ($delayed_count*100/$total_received_count) : 0,
      $failed_count,
      ($total_received_count) ? ($failed_count*100/$total_received_count) : 0);
  }

  if ($merge_reports) {
    $volume = volume_rounded($report_totals{Delivered}{Volume}, $report_totals{Delivered}{'Volume-gigs'});
    $total_delivered_count = get_report_total($report_totals{Delivered},'Messages');
  }
  else {
    $volume = volume_rounded($total_delivered_data, $total_delivered_data_gigs);
  }
  printf("$format1\n\n",'Delivered',$volume,$total_delivered_count,@delivered_totals);
  print "</table>\n" if $html;
}


#######################################################################
# print_user_patterns()
# 
#  print_user_patterns();
# 
# Print the counts of user specified patterns.
#######################################################################
sub print_user_patterns {
  my($format1);

  if ($html) {
    print "<hr><a name=\"patterns\"></a><h2>User Specified Patterns</h2>\n";
    print "<table border=0 width=\"100%\">\n";
    print "<tr><td>\n";
    print "<table border=1>\n";
    print "<tr><th>&nbsp;</th><th>Total</th>\n";
    $format1 = "<tr><td>%s</td><td align=\"right\">%d</td>";
  }
  else {
    print "User Specified Patterns\n";
    print "-----------------------";
    print "\n                       Total\n";
    $format1 = "  %-18s  %6d";
  }

  my($key);
  if ($merge_reports) {
    # We are getting our data from previous reports.
    foreach $key (@user_descriptions) {
      my $count = get_report_total($report_totals{patterns}{$key},'Total');
      printf("$format1\n",$key,$count);
    }
  }
  else {
    # We are getting our data from mainlog files.
    my $user_pattern_index = 0;
    foreach $key (@user_descriptions) {
      printf("$format1\n",$key,$user_pattern_totals[$user_pattern_index]);
      $user_pattern_index++;
    }
  }
  if ($html) {
    print "</table>\n";
  }
  print "\n";
}


#######################################################################
# print_transport();
# 
#  print_transport();
# 
# Print totals by transport.
#######################################################################
sub print_transport {
  my($format1);
  my(@chartdatanames);
  my(@chartdatavals_count);
  my(@chartdatavals_vol);
  no integer;		#Lose this for charting the data.

  if ($html) {
    print "<hr><a name=\"transport\"></a><h2>Deliveries by Transport</h2>\n";
    print "<table border=0 width=\"100%\">\n";
    print "<tr><td>\n";
    print "<table border=1>\n";
    print "<tr><th>&nbsp;</th><th>Volume</th><th>Messages</th>\n";
    $format1 = "<tr><td>%s</td><td align=\"right\">%s</td><td align=\"right\">%d</td>";
  }
  else {
    print "Deliveries by transport\n";
    print "-----------------------";
    print "\n                      Volume    Messages\n";
    $format1 = "  %-18s  %6s      %6d";
  }

  my($key);
  if ($merge_reports) {
    # We are getting our data from previous reports.
    foreach $key (sort keys %{$report_totals{transport}}) {
      my $count = get_report_total($report_totals{transport}{$key},'Messages');
      printf("$format1\n",$key,
	volume_rounded($report_totals{transport}{$key}{Volume},$report_totals{transport}{$key}{'Volume-gigs'}),
	$count);
      push(@chartdatanames, $key);
      push(@chartdatavals_count, $count);
      push(@chartdatavals_vol, $report_totals{transport}{$key}{'Volume-gigs'}*$gig + $report_totals{transport}{$key}{Volume} );
    }
  }
  else {
    # We are getting our data from mainlog files.
    foreach $key (sort keys %transported_data) {
      printf("$format1\n",$key,
	volume_rounded($transported_data{$key},$transported_data_gigs{$key}),
	$transported_count{$key});
      push(@chartdatanames, $key);
      push(@chartdatavals_count, $transported_count{$key});
      push(@chartdatavals_vol, $transported_data_gigs{$key}*$gig + $transported_data{$key});
    }
  }
  if ($html) {
    print "</table>\n";
    print "</td><td>\n";
    if ($HAVE_GD_Graph_pie && $charts)
      {
      # calculate the graph
      my @data = (
         \@chartdatanames,
         \@chartdatavals_count
      );
      my $graph = GD::Graph::pie->new(200, 200);
      $graph->set(
          x_label           => 'Transport',
          y_label           => 'Messages',
          title             => 'By count',
      );
      my $gd = $graph->plot(\@data) or warn($graph->error);
      if ($gd) {
	open(IMG, ">$chartdir/transports_count.png") or die $!;
	binmode IMG;
	print IMG $gd->png;
	close IMG;
	print "<img src=\"$chartrel/transports_count.png\">";
      }
    }
    print "</td><td>\n";

    if ($HAVE_GD_Graph_pie && $charts) {
      my @data = (
         \@chartdatanames,
         \@chartdatavals_vol
      );
      my $graph = GD::Graph::pie->new(200, 200);
      $graph->set(
          title             => 'By volume',
      );
      my $gd = $graph->plot(\@data) or warn($graph->error);
      if ($gd) {
	open(IMG, ">$chartdir/transports_vol.png") or die $!;
	binmode IMG;
	print IMG $gd->png;
	close IMG;
	print "<img src=\"$chartrel/transports_vol.png\">";
      }
    }
    print "</td></tr></table>\n";
  }
  print "\n";
}


#######################################################################
# print_relay();
# 
#  print_relay();
# 
# Print our totals by relay.
#######################################################################
sub print_relay {
  my $temp = "Relayed messages";
  print "<hr><a name=\"$temp\"></a><h2>$temp</h2>\n" if $html;
  if (scalar(keys %relayed) > 0 || $relayed_unshown > 0) {
    my $shown = 0;
    my $spacing = "";
    my($format);

    if ($html) {
      print "<table border=1>\n";
      print "<tr><th>Count</th><th>From</th><th>To</th>\n";
      $format = "<tr><td align=\"right\">%d</td><td>%s</td><td>%s</td>\n";
    }
    else {
      printf("%s\n%s\n\n", $temp, "-" x length($temp));
      $format = "%7d %s\n      => %s\n";
    }

    my($key);
    foreach $key (sort keys %relayed) {
      my $count = $relayed{$key};
      $shown += $count;
      $key =~ s/[HA]=//g;
      my($one,$two) = split(/=> /, $key);
      printf($format, $count, $one, $two);
      $spacing = "\n";
    }
    print "</table>\n<p>\n" if $html;
    print "${spacing}Total: $shown (plus $relayed_unshown unshown)\n";
  }
  else {
    print "No relayed messages\n";
    print "-------------------\n" unless $html;
  }
  print "\n";
}


#######################################################################
# print_errors();
# 
#  print_errors();
# 
# Print our errors. In HTML, we display them as a list rather than a table -
# Netscape doesn't like large tables!
#######################################################################
sub print_errors {
  my $total_errors = 0;

  if (scalar(keys %errors_count) != 0) {
    my $temp = "List of errors";
    my($format);
    if ($html) {
      print "<hr><a name=\"errors\"></a><h2>$temp</h2>\n";
      print "<ul><li><b>Count - Error</b>\n";
      $format = "<li>%d - %s\n";
    }
    else {
      printf("%s\n%s\n\n", $temp, "-" x length($temp));
    }

    my($key);
    foreach $key (sort keys %errors_count) {
      my $text = $key;
      chomp($text);
      $text =~ s/\s\s+/ /g;	#Convert multiple spaces to a single space.
      $total_errors += $errors_count{$key};
      if ($html) {
	 
        #Translate HTML tag characters. Sergey Sholokh.
        $text =~ s/\</\&lt\;/g;
        $text =~ s/\>/\&gt\;/g;

	printf($format,$errors_count{$key},$text);
      }
      else {
	printf("%5d ", $errors_count{$key});
	while (length($text) > 65) {
	  my($first,$rest) = $text =~ /(.{50}\S*)\s+(.+)/;
	  last if !$first;
	  printf("%s\n	    ", $first);
	  $text = $rest;
	}
	printf("%s\n\n", $text);
      }
    }
    print "</ul>\n<p>\n" if $html;

    $temp = "Errors encountered: $total_errors";
    print $temp,"\n";
    print "-" x length($temp),"\n" unless $html;
  }

}


#######################################################################
# parse_old_eximstat_reports();
# 
#  parse_old_eximstat_reports($fh);
# 
# Parse old eximstat output so we can merge daily stats to weekly stats and weekly to monthly etc.
# 
# To test that the merging still works after changes, do something like the following.
# All the diffs should produce no output.
# 
#  options='-bydomain -byemail -byhost -byedomain'
#  options="$options -pattern 'Completed Messages' /Completed/"
#  options="$options -pattern 'Received Messages' /<=/"
# 
#  ./eximstats $options mainlog > mainlog.txt
#  ./eximstats $options -merge mainlog.txt > mainlog.2.txt
#  diff mainlog.txt mainlog.2.txt
# 
#  ./eximstats $options -html mainlog > mainlog.html
#  ./eximstats $options -merge -html mainlog.txt  > mainlog.2.html
#  diff mainlog.html mainlog.2.html
# 
#  ./eximstats $options -merge mainlog.html > mainlog.3.txt
#  diff mainlog.txt mainlog.3.txt
# 
#  ./eximstats $options -merge -html mainlog.html > mainlog.3.html
#  diff mainlog.html mainlog.3.html
# 
#  ./eximstats $options -nvr   mainlog > mainlog.nvr.txt
#  ./eximstats $options -merge mainlog.nvr.txt > mainlog.4.txt
#  diff mainlog.txt mainlog.4.txt
# 
#  # double_mainlog.txt should have twice the values that mainlog.txt has.
#  ./eximstats $options mainlog mainlog > double_mainlog.txt
#######################################################################
sub parse_old_eximstat_reports {
  my($fh) = @_;

  my(%league_table_value_entered, %league_table_value_was_zero, %table_order);

  while (<$fh>) {
    if (/Exim statistics from ([\d\-]+ [\d:]+(\s+[\+\-]\d+)?) to ([\d\-]+ [\d:]+(\s+[\+\-]\d+)?)/) {
      $begin = $1 if ($1 lt $begin);
      $end   = $3 if ($3 gt $end);
    }
    elsif (/Grand total summary/) {
      # Fill in $report_totals{Received|Delivered}{Volume|Messages|Hosts|Domains|...|Delayed|DelayedPercent|Failed|FailedPercent}
      my(@fields);
      while (<$fh>) {
	$_ = html2txt($_);              #Convert general HTML markup to text.
	s/At least one addr//g;		#Another part of the HTML output we don't want.

#  TOTAL               Volume    Messages    Hosts Domains      Delayed       Failed
#  Received              26MB         237      177      23       8  3.4%     28 11.8%
#  Delivered             13MB         233       99      88
	if (/TOTAL\s+(.*?)\s*$/) {
	  @fields = split(/\s+/,$1);
          #Delayed and Failed have two columns each, so add the extra field names in.
	  splice(@fields,-1,1,'DelayedPercent','Failed','FailedPercent');
	}
	elsif (/(Received|Delivered)\s+(.*?)\s*$/) {
	  print STDERR "Parsing $_" if $debug;
	  add_to_totals($report_totals{$1},\@fields,$2);
	}
	last if (/Delivered/);		#Last line of this section.
      }
    }

    elsif (/User Specified Patterns/i) {
#User Specified Patterns
#-----------------------
#                       Total
#  Description             85

      while (<$fh>) { last if (/Total/); }	#Wait until we get the table headers.
      while (<$fh>) {
	print STDERR "Parsing $_" if $debug;
	$_ = html2txt($_);              #Convert general HTML markup to text.
	if (/^\s*(.*?)\s+(\d+)\s*$/) {
	  $report_totals{patterns}{$1} = {} unless (defined $report_totals{patterns}{$1});
	  add_to_totals($report_totals{patterns}{$1},['Total'],$2);
	}
	last if (/^\s*$/);			#Finished if we have a blank line.
      }
    }

    elsif (/Deliveries by transport/i) {
#Deliveries by transport
#-----------------------
#                      Volume    Messages
#  :blackhole:           70KB          51
#  address_pipe         655KB           1
#  smtp                  11MB         151

      while (<$fh>) { last if (/Volume/); }	#Wait until we get the table headers.
      while (<$fh>) {
	print STDERR "Parsing $_" if $debug;
	$_ = html2txt($_);              #Convert general HTML markup to text.
	if (/(\S+)\s+(\d+\S*\s+\d+)/) {
	  $report_totals{transport}{$1} = {} unless (defined $report_totals{transport}{$1});
	  add_to_totals($report_totals{transport}{$1},['Volume','Messages'],$2);
	}
	last if (/^\s*$/);			#Finished if we have a blank line.
      }
    }
    elsif (/(Messages received|Deliveries) per/) {
#      Messages received per hour (each dot is 2 messages)
#---------------------------------------------------
#
#00-01    106 .....................................................
#01-02    103 ...................................................

      # Set a pointer to the interval array so we can use the same code
      # block for both messages received and delivered.
      my $interval_aref = ($1 eq 'Deliveries') ? \@delivered_interval_count : \@received_interval_count;
      my $reached_table = 0;
      while (<$fh>) {
	$reached_table = 1 if (/^00/);
	next unless $reached_table;
	print STDERR "Parsing $_" if $debug;
	if (/^(\d+):(\d+)\s+(\d+)/) {		#hh:mm start time format ?
	  $$interval_aref[($1*60 + $2)/$hist_interval] += $3 if $hist_opt;
	}
	elsif (/^(\d+)-(\d+)\s+(\d+)/) {	#hh-hh start-end time format ?
	  $$interval_aref[($1*60)/$hist_interval] += $3 if $hist_opt;
	}
	else {					#Finished the table ?
	  last;
	}
      }
    }

    elsif (/Time spent on the queue: (all messages|messages with at least one remote delivery)/) {
#Time spent on the queue: all messages
#-------------------------------------
#
#Under   1m      217  91.9%   91.9%
#        5m        2   0.8%   92.8%
#        3h        8   3.4%   96.2%
#        6h        7   3.0%   99.2%
#       12h        2   0.8%  100.0%

      # Set a pointer to the queue bin so we can use the same code
      # block for both all messages and remote deliveries.
      my $bin_aref = ($1 eq 'all messages') ? \@queue_bin : \@remote_queue_bin;
      my $reached_table = 0;
      while (<$fh>) {
	$_ = html2txt($_);              #Convert general HTML markup to text.
	$reached_table = 1 if (/^\s*Under/);
	next unless $reached_table;
	my $previous_seconds_on_queue = 0;
	if (/^\s*(Under|Over|)\s+(\d+[smhdw])\s+(\d+)/) {
	  print STDERR "Parsing $_" if $debug;
	  my($modifier,$formated_time,$count) = ($1,$2,$3);
	  my $seconds = unformat_time($formated_time);
	  my $time_on_queue = ($seconds + $previous_seconds_on_queue) / 2;
	  $previous_seconds_on_queue = $seconds;
	  $time_on_queue = $seconds * 2 if ($modifier eq 'Over');
	  my($i);
	  for ($i = 0; $i <= $#queue_times; $i++) {
	    if ($time_on_queue < $queue_times[$i]) {
	      $$bin_aref[$i] += $count;
	      last;
	    }
	  }
          # There's only one counter for messages going over the queue
          # times so make sure we only count it once.
          $queue_more_than += $count if (($bin_aref == \@queue_bin) && ($i > $#queue_times));
	}
	else {
	  last;					#Finished the table ?
	}
      }
    }

    elsif (/Relayed messages/) {
#Relayed messages
#----------------
#
#      1 addr.domain.com [1.2.3.4] a.user@domain.com
#      => addr2.domain2.com [5.6.7.8] a2.user2@domain2.com
#
#<tr><td align="right">1</td><td>addr.domain.com [1.2.3.4] a.user@domain.com </td><td>addr2.domain2.com [5.6.7.8] a2.user2@domain2.com</td>

      my $reached_table = 0;
      my($count,$sender);
      while (<$fh>) {
	unless ($reached_table) {
	  last if (/No relayed messages/);
	  $reached_table = 1 if (/^\s*\d/ || />\d+</);
	  next unless $reached_table;
	}
	if (/>(\d+)<.td><td>(.*?) ?<.td><td>(.*?)</) {
	  update_relayed($1,$2,$3);
	}
	elsif (/^\s*(\d+)\s+(.*?)\s*$/) {
	  ($count,$sender) = ($1,$2);
	}
	elsif (/=>\s+(.*?)\s*$/) {
	  update_relayed($count,$sender,$1);
	}
	else {
	  last;					#Finished the table ?
	}
      }
    }

    elsif (/Top (.*?) by (message count|volume)/) {
#Top 50 sending hosts by message count
#-------------------------------------
#
#     48     1468KB   local
      my($category,$by_count_or_volume) = ($1,$2);

      #As we show 2 views of each table (by count and by volume),
      #most (but not all) entries will appear in both tables.
      #Set up a hash to record which entries we have already seen
      #and one to record which ones we are seeing for the first time.
      if ($by_count_or_volume =~ /count/) {
	undef %league_table_value_entered;
	undef %league_table_value_was_zero;
	undef %table_order;
      }

      #As this section processes multiple different table categories,
      #set up pointers to the hashes to be updated.
      my($count_href,$data_href,$data_gigs_href);
      if ($category =~ /local sender/) {
        $count_href      = \%received_count_user;
	$data_href	 = \%received_data_user;
	$data_gigs_href  = \%received_data_gigs_user;
      }
      elsif ($category =~ /sending (\S+?)s?\b/) {
        #Top 50 sending (host|domain|email|edomain)s
        #Top sending (host|domain|email|edomain)
        $count_href      = \%{$received_count{"\u$1"}};
	$data_href	 = \%{$received_data{"\u$1"}};
	$data_gigs_href  = \%{$received_data_gigs{"\u$1"}};
      }
      elsif ($category =~ /local destination/) {
        $count_href      = \%delivered_count_user;
	$data_href	 = \%delivered_data_user;
	$data_gigs_href  = \%delivered_data_gigs_user;
      }
      elsif ($category =~ /(\S+) destination/) {
        #Top 50 (host|domain|email|edomain) destinations
        #Top (host|domain|email|edomain) destination
        $count_href      = \%{$delivered_count{"\u$1"}};
	$data_href	 = \%{$delivered_data{"\u$1"}};
	$data_gigs_href  = \%{$delivered_data_gigs{"\u$1"}};
      }

      my $reached_table = 0;
      while (<$fh>) {
	$_ = html2txt($_);              #Convert general HTML markup to text.
	$reached_table = 1 if (/^\s*\d/);
	next unless $reached_table;
	if (/^\s*(\d+)\s+(\S+)\s*(.*?)\s*$/) {
	  my($count,$rounded_volume,$entry) = ($1,$2,$3);
          #Note: $entry fields can be both null and can contain spaces.

	  #Add the entry into the %table_order hash if it has a rounded volume (KB/MB/GB).
	  push(@{$table_order{$rounded_volume}{$by_count_or_volume}},$entry) if ($rounded_volume =~ /\D/);

          unless ($league_table_value_entered{$entry}) {
	    $league_table_value_entered{$entry} = 1;
	    unless ($$count_href{$entry}) {
	      $$count_href{$entry}     = 0;
	      $$data_href{$entry}      = 0;
	      $$data_gigs_href{$entry} = 0;
	      $league_table_value_was_zero{$entry} = 1;
	    }

	    $$count_href{$entry} += $count;
            #Add the rounded value to the data and data_gigs hashes.
	    un_round($rounded_volume,\$$data_href{$entry},\$$data_gigs_href{$entry});
	    print STDERR "$category by $by_count_or_volume: added $count,$rounded_volume to $entry\n" if $debug;
	  }
	}
	else {		#Finished the table ?
	  if ($by_count_or_volume =~ /volume/) {
	    #Add a few bytes to appropriate entries to preserve the order.

	    my($rounded_volume);
	    foreach $rounded_volume (keys %table_order) {
	      #For each rounded volume, we want to create a list which has things
	      #ordered from the volume table at the front, and additional things
	      #from the count table ordered at the back.
	      @{$table_order{$rounded_volume}{volume}} = () unless defined $table_order{$rounded_volume}{volume};
	      @{$table_order{$rounded_volume}{'message count'}} = () unless defined $table_order{$rounded_volume}{'message count'};
	      my(@order,%mark);
	      map {$mark{$_} = 1} @{$table_order{$rounded_volume}{volume}};
	      @order = @{$table_order{$rounded_volume}{volume}};
	      map {push(@order,$_)} grep(!$mark{$_},@{$table_order{$rounded_volume}{'message count'}});

	      my $bonus_bytes = $#order;
	      $bonus_bytes = 511 if ($bonus_bytes > 511);	#Don't go over the half-K boundary!
	      while (@order and ($bonus_bytes > 0)) {
		my $entry = shift(@order);
		if ($league_table_value_was_zero{$entry}) {
		  $$data_href{$entry} += $bonus_bytes;
		  print STDERR "$category by $by_count_or_volume: added $bonus_bytes bonus bytes to $entry\n" if $debug;
		}
		$bonus_bytes--;
	      }
	    }
	  }

	  last;
	}
      }
    }
    elsif (/List of errors/) {
#List of errors
#--------------
#
#    1 07904931641@one2one.net R=external T=smtp: SMTP error
#            from remote mailer after RCPT TO:<07904931641@one2one.net>:
#            host mail.one2one.net [193.133.192.24]: 550 User unknown
#
#<li>1 - ally.dufc@dunbar.org.uk R=external T=smtp: SMTP error from remote mailer after RCPT TO:<ally.dufc@dunbar.org.uk>: host mail.dunbar.org.uk [216.167.89.88]: 550 Unknown local part ally.dufc in <ally.dufc@dunbar.org.uk>


      my $reached_table = 0;
      my($count,$error,$blanks);
      while (<$fh>) {
	$reached_table = 1 if (/^( *|<li>)(\d+)/);
	next unless $reached_table;

	s/^<li>(\d+) -/$1/;	#Convert an HTML line to a text line.
	$_ = html2txt($_);      #Convert general HTML markup to text.

	if (/\t\s*(.*)/) {
	  $error .= ' ' . $1;	#Join a multiline error.
	}
	elsif (/^\s*(\d+)\s+(.*)/) {
	  if ($error) {
            #Finished with a previous multiline error so save it.
	    $errors_count{$error} = 0 unless $errors_count{$error};
	    $errors_count{$error} += $count;
	  }
	  ($count,$error) = ($1,$2);
	}
	elsif (/Errors encountered/) {
	  if ($error) {
            #Finished the section, so save our stored last error.
	    $errors_count{$error} = 0 unless $errors_count{$error};
	    $errors_count{$error} += $count;
	  }
	  last;
	}
      }
    }

  }
}


#######################################################################
# update_relayed();
# 
#  update_relayed($count,$sender,$recipient);
# 
# Adds an entry into the %relayed hash. Currently only used when
# merging reports.
#######################################################################
sub update_relayed {
  my($count,$sender,$recipient) = @_;

  #When generating the key, put in the 'H=' and 'A=' which can be used
  #in searches.
  my $key = "H=$sender => H=$recipient";
  $key =~ s/ ([^=\s]+\@\S+|<>)/ A=$1/g;
  if (!defined $relay_pattern || $key !~ /$relay_pattern/o) {
    $relayed{$key} = 0 if !defined $relayed{$key};
    $relayed{$key} += $count;
  }
  else {
    $relayed_unshown += $count;
  }
}


#######################################################################
# add_to_totals();
# 
#  add_to_totals(\%totals,\@keys,$values);
# 
# Given a line of space seperated values, add them into the provided hash using @keys
# as the hash keys.
# 
# If the value contains a '%', then the value is set rather than added. Otherwise, we
# convert the value to bytes and gigs. The gigs get added to I<Key>-gigs.
#######################################################################
sub add_to_totals {
  my($totals_href,$keys_aref,$values) = @_;
  my(@values) = split(/\s+/,$values);
  my(@keys) = @$keys_aref;		#Make a copy as we destroy the one we use.
  my($value);
  foreach $value (@values) {
    my $key = shift(@keys) or next;
    if ($value =~ /%/) {
      $$totals_href{$key} = $value;
    }
    else {
      $$totals_href{$key} = 0 unless ($$totals_href{$key});
      $$totals_href{"$key-gigs"} = 0 unless ($$totals_href{"$key-gigs"});
      un_round($value, \$$totals_href{$key}, \$$totals_href{"$key-gigs"});
      print STDERR "Added $value to $key - $$totals_href{$key} , " . $$totals_href{"$key-gigs"} . "GB.\n" if $debug;
    }
  }
}

#######################################################################
# get_report_total();
# 
#  $total = get_report_total(\%hash,$key);
# 
# If %hash contains values split into Units and Gigs, we calculate and return
# 
#   $hash{$key} + 1024*1024*1024 * $hash{"${key}-gigs"}
#######################################################################
sub get_report_total {
  no integer;
  my($hash_ref,$key) = @_;
  if ($$hash_ref{"${key}-gigs"}) {
    return $$hash_ref{$key} + $gig * $$hash_ref{"${key}-gigs"};
  }
  return $$hash_ref{$key} || 0;
}

#######################################################################
# html2txt();
# 
#  $text_line = html2txt($html_line);
# 
# Convert a line from html to text. Currently we just convert HTML tags to spaces
# and convert &gt;, &lt;, and &nbsp; tags back.
#######################################################################
sub html2txt {
  ($_) = @_;

  # Convert HTML tags to spacing. Note that the reports may contain <Userid> and
  # <Userid@Domain> words, so explicitly specify the HTML tags we will remove
  # (the ones used by this program). If someone is careless enough to have their
  # Userid the same as an HTML tag, there's not much we can do about it.
  s/<\/?(html|head|title|body|h\d|ul|li|a\s+|table|tr|td|th|pre|hr|p|br)\b.*?>/ /og;

  s/\&lt\;/\</og;             #Convert '&lt;' to '<'.
  s/\&gt\;/\>/og;             #Convert '&gt;' to '>'.
  s/\&nbsp\;/ /og;            #Convert '&nbsp;' to ' '.
  return($_);
}

#######################################################################
# get_next_arg();
# 
#  $arg = get_next_arg();
# 
# Because eximstats arguments are often passed as variables,
# we can't rely on shell parsing to deal with quotes. This
# subroutine returns $ARGV[1] and does a shift. If $ARGV[1]
# starts with a quote (' or "), and doesn't end in one, then
# we append the next argument to it and shift again. We repeat
# until we've got all of the argument.
# 
# This isn't perfect as all white space gets reduced to one space,
# but it's as good as we can get! If it's esential that spacing
# be preserved precisely, then you get that by not using shell
# variables.
#######################################################################
sub get_next_arg {
  my $arg = '';
  my $matched_pattern = 0;
  while ($ARGV[1]) {
    $arg .= ' ' if $arg;
    $arg .= $ARGV[1]; shift(@ARGV);
    if ($arg !~ /^['"]/) {
      $matched_pattern = 1;
      last;
    }
    if ($arg =~ s/^(['"])(.*)\1$/$2/) {
      $matched_pattern = 1;
      last;
    }
  }
  die "Mismatched argument quotes - <$arg>.\n" unless $matched_pattern;
  return $arg;
}


##################################################
#                 Main Program                   #
##################################################


$last_timestamp = '';
$last_date = '';
$show_errors = 1;
$show_relay = 1;
$show_transport = 1;
$topcount = 50;
$local_league_table = 1;
$include_remote_users = 0;
$hist_opt = 1;
$volume_rounding = 1;
$localtime_offset = calculate_localtime_offset();    # PH/FANF

$charts = 0;
$charts_option_specified = 0;
$chartrel = ".";
$chartdir = ".";

@queue_times = (60, 5*60, 15*60, 30*60, 60*60, 3*60*60, 6*60*60,
                12*60*60, 24*60*60);

$last_offset = '';
$offset_seconds = 0;

# Decode options

while (@ARGV > 0 && substr($ARGV[0], 0, 1) eq '-')
  {
  if    ($ARGV[0] =~ /^\-h(\d+)$/) { $hist_opt = $1 }
  elsif ($ARGV[0] =~ /^\-ne$/)     { $show_errors = 0 }
  elsif ($ARGV[0] =~ /^\-nr(.?)(.*)\1$/)
    {
    if ($1 eq "") { $show_relay = 0 } else { $relay_pattern = $2 }
    }
  elsif ($ARGV[0] =~ /^\-q([,\d\+\-\*\/]+)$/)
    {
    @queue_times = split(/,/, $1);
    my($q);
    foreach $q (@queue_times) { $q = eval($q) + 0 }
    @queue_times = sort { $a <=> $b } @queue_times;
    @queue_times = () if ($#queue_times == 0 && $queue_times[0] == 0);
    }
  elsif ($ARGV[0] =~ /^-nt$/)       { $show_transport = 0 }
  elsif ($ARGV[0] =~ /^\-nt(.?)(.*)\1$/)
    {
    if ($1 eq "") { $show_transport = 0 } else { $transport_pattern = $2 }
    }
  elsif ($ARGV[0] =~ /^-t(\d+)$/)   { $topcount = $1 }
  elsif ($ARGV[0] =~ /^-tnl$/)      { $local_league_table = 0 }
  elsif ($ARGV[0] =~ /^-html$/)     { $html = 1 }
  elsif ($ARGV[0] =~ /^-merge$/)    { $merge_reports = 1 }
  elsif ($ARGV[0] =~ /^-charts$/)   {
    $charts = 1;
    warn "WARNING: CPAN Module GD::Graph::pie not installed. Obtain from www.cpan.org\n" unless $HAVE_GD_Graph_pie;
    warn "WARNING: CPAN Module GD::Graph::linespoints not installed. Obtain from www.cpan.org\n" unless $HAVE_GD_Graph_linespoints;
  }
  elsif ($ARGV[0] =~ /^-chartdir$/) { $chartdir = $ARGV[1]; shift; $charts_option_specified = 1; }
  elsif ($ARGV[0] =~ /^-chartrel$/) { $chartrel = $ARGV[1]; shift; $charts_option_specified = 1; }
  elsif ($ARGV[0] =~ /^-cache$/)    { }	#Not currently used.
  elsif ($ARGV[0] =~ /^-byhost$/)   { $do_sender{Host} = 1 }
  elsif ($ARGV[0] =~ /^-bydomain$/) { $do_sender{Domain} = 1 }
  elsif ($ARGV[0] =~ /^-byemail$/)  { $do_sender{Email} = 1 }
  elsif ($ARGV[0] =~ /^-byemaildomain$/)  { $do_sender{Edomain} = 1 }
  elsif ($ARGV[0] =~ /^-byedomain$/)  { $do_sender{Edomain} = 1 }
  elsif ($ARGV[0] =~ /^-nvr$/)      { $volume_rounding = 0 }
  elsif ($ARGV[0] =~ /^-d$/)        { $debug = 1 }
  elsif ($ARGV[0] =~ /^--?h(elp)?$/){ help() }
  elsif ($ARGV[0] =~ /^-t_remote_users$/) { $include_remote_users = 1 }
  elsif ($ARGV[0] =~ /^-pattern$/)
    {
    push(@user_descriptions,get_next_arg());
    push(@user_patterns,get_next_arg());
    }
  elsif ($ARGV[0] =~ /^-utc$/)
    {
    # We don't need this value if the log is in UTC.
    $localtime_offset = undef;
    }
  else
    {
    print STDERR "Eximstats: Unknown or malformed option $ARGV[0]\n";
    help();
    }
  shift;
  }

  # Check that all the charts options are specified.
  warn "-charts option not specified. Use -help for help.\n" if ($charts_option_specified && ! $charts);

  # Default to display tables by sending Host.
  $do_sender{Host} = 1 unless ($do_sender{Domain} || $do_sender{Email} || $do_sender{Edomain});


for (my $i = 0; $i <= $#queue_times; $i++) {
  $queue_bin[$i] = 0;
  $remote_queue_bin[$i] = 0;
}

# Compute the number of slots for the histogram

if ($hist_opt > 0)
  {
  if ($hist_opt > 60 || 60 % $hist_opt != 0)
    {
    print "Eximstats: -h must specify a factor of 60\n";
    exit 1;
    }
  $hist_interval = 60/$hist_opt;		#Interval in minutes.
  $hist_number = (24*60)/$hist_interval;	#Number of intervals per day.
  @received_interval_count = (0) x $hist_number;
  @delivered_interval_count = (0) x $hist_number;
  }

#$queue_unknown = 0;

$total_received_data = 0;
$total_received_data_gigs = 0;
$total_received_count = 0;

$total_delivered_data = 0;
$total_delivered_data_gigs = 0;
$total_delivered_count = 0;

$queue_more_than = 0;
$delayed_count = 0;
$relayed_unshown = 0;
$begin = "9999-99-99 99:99:99";
$end = "0000-00-00 00:00:00";
my($section,$type);
foreach $section ('Received','Delivered') {
  foreach $type ('Volume','Messages','Delayed','Failed','Hosts','Domains','Emails','Edomains') {
    $report_totals{$section}{$type} = 0;
  }
}

# Generate our parser.
my $parser = generate_parser();


if (@ARGV) {
  # Scan the input files and collect the data
  foreach my $file (@ARGV) {
    if ($file =~ /\.gz/) {
      unless (open(FILE,"gunzip -c $file |")) {
	print STDERR "Failed to gunzip -c $file: $!";
	next;
      }
    }
    elsif ($file =~ /\.Z/) {
      unless (open(FILE,"uncompress -c $file |")) {
	print STDERR "Failed to uncompress -c $file: $!";
	next;
      }
    }
    else {
      unless (open(FILE,$file)) {
	print STDERR "Failed to read $file: $!";
	next;
      }
    }
    #Now parse the filehandle, updating the global variables.
    parse($parser,\*FILE);
    close FILE;
  }
}
else {
  #No files provided. Parse STDIN, updating the global variables.
  parse($parser,\*STDIN);
}


if ($begin eq "9999-99-99 99:99:99") {
  print "**** No valid log lines read\n";
  exit 1;
}

# Output our results.
print_header();
print_grandtotals();

# Print counts of user specified patterns if required.
print_user_patterns() if @user_patterns;

# Print totals by transport if required.
print_transport() if $show_transport;

# Print the deliveries per interval as a histogram, unless configured not to.
# First find the maximum in one interval and scale accordingly.
if ($hist_opt > 0) {
  print_histogram("Messages received", @received_interval_count);
  print_histogram("Deliveries", @delivered_interval_count);
}

# Print times on queue if required.
if ($#queue_times >= 0) {
  print_queue_times("all messages", \@queue_bin,$queue_more_than);
  print_queue_times("messages with at least one remote delivery",\@remote_queue_bin,$queue_more_than);
}

# Print relay information if required.
print_relay() if $show_relay;

# Print the league tables, if topcount isn't zero.
if ($topcount > 0) {
  foreach ('Host','Domain','Email','Edomain') {
    next unless $do_sender{$_};
    print_league_table("sending \l$_", $received_count{$_}, $received_data{$_},$received_data_gigs{$_});
  }

  print_league_table("local sender", \%received_count_user,
    \%received_data_user,\%received_data_gigs_user) if ($local_league_table || $include_remote_users);
  foreach ('Host','Domain','Email','Edomain') {
    next unless $do_sender{$_};
    print_league_table("\l$_ destination", $delivered_count{$_}, $delivered_data{$_},$delivered_data_gigs{$_});
  }
  print_league_table("local destination", \%delivered_count_user,
    \%delivered_data_user,\%delivered_data_gigs_user) if ($local_league_table || $include_remote_users);
}

# Print the error statistics if required.
print_errors() if $show_errors;

if ($html) {
  print "</body>\n</html>\n"
}

# End of eximstats