steve 2006/11/16 16:16:30 GMT
Modified files:
exim-src/src eximstats.src
Log:
2005-07-26 V1.44 Steve Campbell
Use a glob alias rather than an array ref in the generated
parser. This improves both readability and performance.
2005-09-30 V1.45 Marco Gaiarin / Steve Campbell
Collect SpamAssassin and rejection statistics.
Don't display local sender or destination tables unless
there is data to show.
Added average volumes into the top table text output.
2006-02-07 V1.46 Steve Campbell
Collect data on the number of addresses (recipients)
as well as the number of messages.
2006-05-05 V1.47 Steve Campbell
Added 'Message too big' to the list of mail rejection
reasons (thanks to Marco Gaiarin).
2006-06-05 V1.48 Steve Campbell
Mainlog lines which have GMT offsets and are too short to
have a flag are now skipped.
2006-11-10 V1.49 Alain Williams
Added the -emptyok flag.
2006-11-16 V1.50 Steve Campbell
Fixes for obtaining the IP address from reject messages.
Revision Changes Path
1.11 +833 -450 exim/exim-src/src/eximstats.src
Index: eximstats.src
===================================================================
RCS file: /home/cvs/exim/exim-src/src/eximstats.src,v
retrieving revision 1.10
retrieving revision 1.11
diff -u -r1.10 -r1.11
--- eximstats.src 30 Jun 2005 10:12:42 -0000 1.10
+++ eximstats.src 16 Nov 2006 16:16:30 -0000 1.11
@@ -1,5 +1,5 @@
#!PERL_COMMAND -w
-# $Cambridge: exim/exim-src/src/eximstats.src,v 1.10 2005/06/30 10:12:42 steve Exp $
+# $Cambridge: exim/exim-src/src/eximstats.src,v 1.11 2006/11/16 16:16:30 steve Exp $
# Copyright (c) 2001 University of Cambridge.
# See the file NOTICE for conditions of use and distribution.
@@ -226,6 +226,35 @@
# 2005-06-30 V1.43 Steve Campbell
# Bug fix for V1.42 with -h0 specified. Spotted by Chris Lear.
#
+# 2005-07-26 V1.44 Steve Campbell
+# Use a glob alias rather than an array ref in the generated
+# parser. This improves both readability and performance.
+#
+# 2005-09-30 V1.45 Marco Gaiarin / Steve Campbell
+# Collect SpamAssassin and rejection statistics.
+# Don't display local sender or destination tables unless
+# there is data to show.
+# Added average volumes into the top table text output.
+#
+# 2006-02-07 V1.46 Steve Campbell
+# Collect data on the number of addresses (recipients)
+# as well as the number of messages.
+#
+# 2006-05-05 V1.47 Steve Campbell
+# Added 'Message too big' to the list of mail rejection
+# reasons (thanks to Marco Gaiarin).
+#
+# 2006-06-05 V1.48 Steve Campbell
+# Mainlog lines which have GMT offsets and are too short to
+# have a flag are now skipped.
+#
+# 2006-11-10 V1.49 Alain Williams
+# Added the -emptyok flag.
+#
+# 2006-11-16 V1.50 Steve Campbell
+# Fixes for obtaining the IP address from reject messages.
+#
+#
#
# For documentation on the logfile format, see
# http://www.exim.org/exim-html-4.50/doc/html/spec_48.html#IX2793
@@ -449,6 +478,11 @@
Specify the relative directory for the "img src=" tags from where to include
the charts
+=item B<-emptyok>
+
+Specify that it's OK to not find any valid log lines. Without this
+we will output an error message if we don't find any.
+
=item B<-d>
Debug flag. This outputs the eval()'d parser onto STDOUT which makes it
@@ -518,7 +552,7 @@
@days_per_month = (0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334);
$gig = 1024 * 1024 * 1024;
-$VERSION = '1.43';
+$VERSION = '1.50';
# How much space do we allow for the Hosts/Domains/Emails/Edomains column headers?
$COLUMN_WIDTHS = 8;
@@ -530,7 +564,7 @@
# Declare global variables.
use vars qw($total_received_data $total_received_data_gigs $total_received_count);
-use vars qw($total_delivered_data $total_delivered_data_gigs $total_delivered_count);
+use vars qw($total_delivered_data $total_delivered_data_gigs $total_delivered_messages $total_delivered_addresses);
use vars qw(%timestamp2time); #Hash of timestamp => time.
use vars qw($last_timestamp $last_time); #The last time convertion done.
use vars qw($last_date $date_seconds); #The last date convertion done.
@@ -540,10 +574,18 @@
use vars qw($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
-use vars qw($workbook $ws_global $ws_relayed $ws_top50 $ws_errors ); #For use in Speadsheed::WriteExcel
-use vars qw($row $col $row_hist $col_hist $row_league_table);
+
+# SpamAssassin variables
+use vars qw($spam_score $spam_score_gigs);
+use vars qw($ham_score $ham_score_gigs);
+use vars qw(%ham_count_by_ip %spam_count_by_ip);
+use vars qw(%rejected_count_by_ip %rejected_count_by_reason);
+
+#For use in Speadsheed::WriteExcel
+use vars qw($workbook $ws_global $ws_relayed $ws_errors);
+use vars qw($row $col $row_hist $col_hist);
use vars qw($run_hist);
-use vars qw($f_default $f_header1 $f_header2 $f_headertab $f_percent); #Format Header
+use vars qw($f_default $f_header1 $f_header2 $f_header2_m $f_headertab $f_percent); #Format Header
# Output FileHandles
use vars qw($txt_fh $htm_fh $xls_fh);
@@ -554,7 +596,7 @@
# set by command line switches:
use vars qw($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 $volume_rounding);
+use vars qw($hist_opt $hist_interval $hist_number $volume_rounding $emptyOK);
use vars qw($relay_pattern @queue_times @user_patterns @user_descriptions);
use vars qw(@rcpt_times @delivery_times);
use vars qw($include_original_destination);
@@ -567,11 +609,11 @@
# The following are modified in the parse() routine, and
# referred to in the print_*() routines.
use vars qw($delayed_count $relayed_unshown $begin $end);
-use vars qw(%messages $message_aref);
+use vars qw(%messages @message);
use vars qw(%received_count %received_data %received_data_gigs);
-use vars qw(%delivered_count %delivered_data %delivered_data_gigs);
+use vars qw(%delivered_messages %delivered_data %delivered_data_gigs %delivered_addresses);
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(%delivered_messages_user %delivered_addresses_user %delivered_data_user %delivered_data_gigs_user);
use vars qw(%transported_count %transported_data %transported_data_gigs);
use vars qw(%relayed %errors_count $message_errors);
use vars qw(@qt_all_bin @qt_remote_bin);
@@ -1005,14 +1047,10 @@
printf $txt_fh ("%s\n%s\n\n", $temp, "-" x length($temp)) if $txt_fh;
if ($htm_fh) {
print $htm_fh "<hr><a name=\"$title $message_type\"></a><h2>$temp</h2>\n";
- print $htm_fh "<table border=0 width=\"100%\">\n";
- print $htm_fh "<tr><td>\n";
- print $htm_fh "<table border=1>\n";
+ print $htm_fh "<table border=0 width=\"100%\"><tr><td><table border=1>\n";
print $htm_fh "<tr><th>Time</th><th>Messages</th><th>Percentage</th><th>Cumulative Percentage</th>\n";
}
-if ($xls_fh)
-{
-
+if ($xls_fh) {
$ws_global->write($row++, $col, "$title: ".$message_type, $f_header2);
my @content=("Time", "Messages", "Percentage", "Cumulative Percentage");
&set_worksheet_line($ws_global, $row++, 1, \@content, $f_headertab);
@@ -1082,8 +1120,7 @@
#printf("Unknown %6d\n", $queue_unknown) if $queue_unknown > 0;
if ($htm_fh) {
- print $htm_fh "</table>\n";
- print $htm_fh "</td><td>\n";
+ print $htm_fh "</table></td><td>";
if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals > 0)) {
my @data = (
@@ -1133,12 +1170,11 @@
my(@chartdatavals);
my($maxd) = 0;
-if (!$run_hist) # save first row of print_histogram for xls output
-{
+# save first row of print_histogram for xls output
+if (!$run_hist) {
$row_hist = $row;
}
-else
-{
+else {
$row = $row_hist;
}
@@ -1169,17 +1205,16 @@
print $htm_fh "<tr><td><pre>\n";
}
-if ($xls_fh)
-{
+if ($xls_fh) {
$title =~ s/Messages/Msg/ ;
+ $row += 2;
$ws_global->write($row++, $col_hist+1, $title, $f_headertab);
}
my $hour = 0;
my $minutes = 0;
-for ($i = 0; $i < $hist_number; $i++)
-{
+for ($i = 0; $i < $hist_number; $i++) {
my $c = $interval_count[$i];
# If the interval is an hour (the maximum) print the starting and
@@ -1187,26 +1222,23 @@
# minutes, which take up the same space.
my $temp;
- if ($hist_opt == 1)
- {
+ if ($hist_opt == 1) {
$temp = sprintf("%02d-%02d", $hour, $hour + 1);
print $txt_fh $temp if $txt_fh;
print $htm_fh $temp if $htm_fh;
- if ($xls_fh)
- {
- if ($run_hist==0) # only on first run
- {
- &set_worksheet_line($ws_global, $row, 0, [$temp], $f_default);
+ if ($xls_fh) {
+ if ($run_hist==0) {
+ # only on first run
+ $ws_global->write($row, 0, [$temp], $f_default);
}
}
push(@chartdatanames, $temp);
$hour++;
}
- else
- {
+ else {
if ($minutes == 0)
{ $temp = sprintf("%02d:%02d", $hour, $minutes) }
else
@@ -1214,28 +1246,24 @@
print $txt_fh $temp if $txt_fh;
print $htm_fh $temp if $htm_fh;
- if (($xls_fh) and ($run_hist==0)) # only on first run
- {
+ if (($xls_fh) and ($run_hist==0)) {
+ # only on first run
$temp = sprintf("%02d:%02d", $hour, $minutes);
- &set_worksheet_line($ws_global, $row, 0, [$temp], $f_default);
+ $ws_global->write($row, 0, [$temp], $f_default);
}
push(@chartdatanames, $temp);
$minutes += $hist_interval;
- if ($minutes >= 60)
- {
+ if ($minutes >= 60) {
$minutes = 0;
$hour++;
- }
+ }
}
push(@chartdatavals, $c);
printf $txt_fh (" %6d %s\n", $c, "." x ($c/$scale)) if $txt_fh;
printf $htm_fh (" %6d %s\n", $c, "." x ($c/$scale)) if $htm_fh;
- if ($xls_fh)
- {
- &set_worksheet_line($ws_global, $row++, $col_hist+1, [$c], $f_default);
- }
+ $ws_global->write($row++, $col_hist+1, [$c], $f_default) if $xls_fh;
} #end for
@@ -1285,234 +1313,244 @@
#######################################################################
# print_league_table();
#
-# print_league_table($league_table_type,\%message_count,\%message_data,\%message_data_gigs);
+# print_league_table($league_table_type,\%message_count,\%address_count,\%message_data,\%message_data_gigs, $spreadsheet, $row_sref);
#
-# 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).
+# Given hashes of message count, address 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 $htm_format;
-my $txt_format = "%7d %10s %s\n";
-
-# write header
-printf $txt_fh ("%s\n%s\n\n", $temp, "-" x length($temp)) if $txt_fh;
-if ($htm_fh) {
- print $htm_fh "<hr><a name=\"$text count\"></a><h2>$temp</h2>\n";
- print $htm_fh "<table border=0 width=\"100%\">\n";
- print $htm_fh "<tr><td>\n";
- print $htm_fh "<table border=1>\n";
- print $htm_fh "<tr><th>Messages</th><th>Bytes</th><th>Average</th><th>\u$text</th>\n";
+ my($text,$m_count,$a_count,$m_data,$m_data_gigs,$spreadsheet, $row_sref) = @_;
+ my($name) = ($topcount == 1)? "$text" : "$topcount ${text}s";
+ my($title) = "Top $name by message count";
+ my(@chartdatanames) = ();
+ my(@chartdatavals) = ();
+ my $chartotherval = 0;
+ $text = ucfirst($text);
# 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';
- $htm_format = "<tr><td align=\"right\">%d</td><td align=\"right\">%s</td><td align=\"right\">%s</td><td align=\"$align\" nowrap>%s</td>\n";
-}
-if ($xls_fh)
-{
- $ws_top50->write($row_league_table++, 0, $temp, $f_header2);
- &set_worksheet_line($ws_top50, $row_league_table++, 0, ["Messages", "Bytes", "Average", $text], $f_headertab );
-}
-
-# write content
-my($key,$htmlkey,$rounded_volume,$rounded_average,$count,$data,$gigs);
-foreach $key (top_n_sort($topcount,$m_count,$m_data_gigs,$m_data)) {
-
- # When displaying the average figures, we calculate the average of
- # the rounded data, as the user would calculate it. This reduces
- # the accuracy slightly, but we have to do it this way otherwise
- # when using -merge to convert results from text to HTML and
- # vice-versa discrepencies would occur.
- $rounded_volume = volume_rounded($$m_data{$key},$$m_data_gigs{$key});
- $data = $gigs = 0;
- un_round($rounded_volume,\$data,\$gigs);
- $count = $$m_count{$key};
- $rounded_average = volume_rounded($data/$count,$gigs/$count);
- my @content=( $count, $rounded_volume, $rounded_average);
- # write content
- # any reason not to include rounded_average in txt-output? -fh
- printf $txt_fh ($txt_format, $count, $rounded_volume, $key) if $txt_fh;
+ ################################################
+ # Generate the printf formats and table headers.
+ ################################################
+ my(@headers) = ('Messages');
+ push(@headers,'Addresses') if defined $a_count;
+ push(@headers,'Bytes','Average') if defined $m_data;
+
+ my $txt_format = "%10s " x @headers . " %s\n";
+ my $txt_col_headers = sprintf $txt_format, @headers, $text;
+ my $htm_format = "<tr>" . '<td align="right">%s</td>'x@headers . "<td align=\"$align\" nowrap>%s</td></tr>\n";
+ my $htm_col_headers = sprintf $htm_format, @headers, $text;
+ $htm_col_headers =~ s/(<\/?)td/$1th/g; #Convert <td>'s to <th>'s for the header.
+
+
+ ################################################
+ # Write the table headers
+ ################################################
+ printf $txt_fh ("%s\n%s\n%s", $title, "-" x length($title),$txt_col_headers) if $txt_fh;
if ($htm_fh) {
- $htmlkey = $key;
- $htmlkey =~ s/>/\>\;/g;
- $htmlkey =~ s/</\<\;/g;
- printf $htm_fh ($htm_format, @content, $htmlkey);
- }
- if ($xls_fh)
- {
- &set_worksheet_line($ws_top50, $row_league_table++, 0, [@content, $key], $f_default);
+ print $htm_fh <<EoText;
+<hr><a name="$text count"></a><h2>$title</h2>
+<table border=0 width="100%">
+<tr><td>
+<table border=1>
+EoText
+ print $htm_col_headers;
}
- if (scalar @chartdatanames < $ntopchart)
- {
- push(@chartdatanames, $key);
- push(@chartdatavals, $$m_count{$key});
- }
- else
- {
- $chartotherval += $$m_count{$key};
+ if ($xls_fh) {
+ $spreadsheet->write(${$row_sref}++, 0, $title, $f_header2);
+ $spreadsheet->write(${$row_sref}++, 0, [@headers, $text], $f_headertab);
}
-}
-push(@chartdatanames, "Other");
-push(@chartdatavals, $chartotherval);
-print $txt_fh "\n" if $txt_fh;
-if ($htm_fh)
-{
- print $htm_fh "</table>\n";
- print $htm_fh "</td><td>\n";
- if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals > 0))
- {
- # calculate the graph
- my @data = (
- \@chartdatanames,
- \@chartdatavals
- );
- my $graph = GD::Graph::pie->new(300, 300);
- $graph->set(
- x_label => 'Name',
- y_label => 'Amount',
- title => 'By count',
- );
- my $gd = $graph->plot(\@data) or warn($graph->error);
- if ($gd) {
- my $temp = $text;
- $temp =~ s/ /_/g;
- open(IMG, ">$chartdir/${temp}_count.png") or die "Could not write $chartdir/${temp}_count.png: $!\n";
- binmode IMG;
- print IMG $gd->png;
- close IMG;
- print $htm_fh "<img src=\"$chartrel/${temp}_count.png\">";
- }
- }
- print $htm_fh "</td><td>\n";
- print $htm_fh "</td></tr></table>\n\n";
-}
-if ($xls_fh)
-{
- $row_league_table++;
-}
+ # write content
+ foreach my $key (top_n_sort($topcount,$m_count,$m_data_gigs,$m_data)) {
+ # When displaying the average figures, we calculate the average of
+ # the rounded data, as the user would calculate it. This reduces
+ # the accuracy slightly, but we have to do it this way otherwise
+ # when using -merge to convert results from text to HTML and
+ # vice-versa discrepencies would occur.
+ my $messages = $$m_count{$key};
+ my @content = ($messages);
+ push(@content, $$a_count{$key}) if defined $a_count;
+ if (defined $m_data) {
+ my $rounded_volume = volume_rounded($$m_data{$key},$$m_data_gigs{$key});
+ my($data,$gigs) = (0,0);
+ un_round($rounded_volume,\$data,\$gigs);
+ my $rounded_average = volume_rounded($data/$messages,$gigs/$messages);
+ push(@content, $rounded_volume, $rounded_average);
+ }
-# write header
+ # write content
+ printf $txt_fh ($txt_format, @content, $key) if $txt_fh;
-$temp = "Top $name by volume";
+ if ($htm_fh) {
+ my $htmlkey = $key;
+ $htmlkey =~ s/>/\>\;/g;
+ $htmlkey =~ s/</\<\;/g;
+ printf $htm_fh ($htm_format, @content, $htmlkey);
+ }
+ $spreadsheet->write(${$row_sref}++, 0, [@content, $key], $f_default) if $xls_fh;
-printf $txt_fh ("%s\n%s\n\n", $temp, "-" x length($temp)) if $txt_fh;
-if ($htm_fh) {
- print $htm_fh "<hr><a name=\"$text volume\"></a><h2>$temp</h2>\n";
- print $htm_fh "<table border=0 width=\"100%\">\n";
- print $htm_fh "<tr><td>\n";
- print $htm_fh "<table border=1>\n";
- print $htm_fh "<tr><th>Messages</th><th>Bytes</th><th>Average</th><th>\u$text</th>\n";
-}
-if ($xls_fh)
-{
- $ws_top50->write($row_league_table++, 0, $temp, $f_header2);
- &set_worksheet_line($ws_top50, $row_league_table++, 0, ["Messages", "Bytes", "Average", $text], $f_headertab);
-}
+ if (scalar @chartdatanames < $ntopchart) {
+ push(@chartdatanames, $key);
+ push(@chartdatavals, $$m_count{$key});
+ }
+ else {
+ $chartotherval += $$m_count{$key};
+ }
+ }
-@chartdatanames = ();
-@chartdatavals = ();
-$chartotherval = 0;
-my $use_gig = 0;
-foreach $key (top_n_sort($topcount,$m_data_gigs,$m_data,$m_count)) {
- # The largest volume will be the first (top of the list).
- # If it has at least 1 gig, then just use gigabytes to avoid
- # risking an integer overflow when generating the pie charts.
- if ($$m_data_gigs{$key}) {
- $use_gig = 1;
- }
-
- $rounded_volume = volume_rounded($$m_data{$key},$$m_data_gigs{$key});
- $data = $gigs = 0;
- un_round($rounded_volume,\$data,\$gigs);
- $count = $$m_count{$key};
- $rounded_average = volume_rounded($data/$count,$gigs/$count);
- my @content=($count, $rounded_volume, $rounded_average );
+ push(@chartdatanames, "Other");
+ push(@chartdatavals, $chartotherval);
- # write content
- # any reasons for not including rounded_average in the txt-version?? -fh
- printf $txt_fh ($txt_format, $count, $rounded_volume, $key) if $txt_fh;
+ print $txt_fh "\n" if $txt_fh;
if ($htm_fh) {
- $htmlkey = $key;
- $htmlkey =~ s/>/\>\;/g;
- $htmlkey =~ s/</\<\;/g;
- printf $htm_fh ($htm_format, @content, $htmlkey);
- }
- if ($xls_fh)
- {
- &set_worksheet_line($ws_top50, $row_league_table++, 0, [@content, $key], $f_default);
+ print $htm_fh "</table>\n";
+ print $htm_fh "</td><td>\n";
+ if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals > 0))
+ {
+ # 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 "Could not write $chartdir/${temp}_count.png: $!\n";
+ binmode IMG;
+ print IMG $gd->png;
+ close IMG;
+ print $htm_fh "<img src=\"$chartrel/${temp}_count.png\">";
+ }
+ }
+ print $htm_fh "</td><td>\n";
+ print $htm_fh "</td></tr></table>\n\n";
}
+ ++${$row_sref} if $xls_fh;
+
+ if (defined $m_data) {
+ # write header
- if (scalar @chartdatanames < $ntopchart) {
- if ($use_gig) {
+ $title = "Top $name by volume";
+
+ printf $txt_fh ("%s\n%s\n%s", $title, "-" x length($title),$txt_col_headers) if $txt_fh;
+
+ if ($htm_fh) {
+ print $htm_fh <<EoText;
+<hr><a name="$text volume"></a><h2>$title</h2>
+<table border=0 width="100%">
+<tr><td>
+<table border=1>
+EoText
+ print $htm_col_headers;
+ }
+ if ($xls_fh) {
+ $spreadsheet->write(${$row_sref}++, 0, $title, $f_header2);
+ $spreadsheet->write(${$row_sref}++, 0, [@headers, $text], $f_headertab);
+ }
+
+ @chartdatanames = ();
+ @chartdatavals = ();
+ $chartotherval = 0;
+ my $use_gig = 0;
+ foreach my $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}) {
- push(@chartdatanames, $key);
- push(@chartdatavals, $$m_data_gigs{$key});
+ $use_gig = 1;
+ }
+
+ my $messages = $$m_count{$key};
+ my @content = ($messages);
+ push(@content, $$a_count{$key}) if defined $a_count;
+ my $rounded_volume = volume_rounded($$m_data{$key},$$m_data_gigs{$key});
+ my($data ,$gigs) = (0,0);
+ un_round($rounded_volume,\$data,\$gigs);
+ my $rounded_average = volume_rounded($data/$messages,$gigs/$messages);
+ push(@content, $rounded_volume, $rounded_average );
+
+ # write content
+ printf $txt_fh ($txt_format, @content, $key) if $txt_fh;
+ if ($htm_fh) {
+ my $htmlkey = $key;
+ $htmlkey =~ s/>/\>\;/g;
+ $htmlkey =~ s/</\<\;/g;
+ printf $htm_fh ($htm_format, @content, $htmlkey);
+ }
+ $spreadsheet->write(${$row_sref}++, 0, [@content, $key], $f_default) if $xls_fh;
+
+
+ if (scalar @chartdatanames < $ntopchart) {
+ if ($use_gig) {
+ if ($$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};
}
}
- 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);
+ push(@chartdatanames, "Other");
+ push(@chartdatavals, $chartotherval);
-print $txt_fh "\n" if $txt_fh;
-if ($htm_fh) {
- print $htm_fh "</table>\n";
- print $htm_fh "</td><td>\n";
- if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals > 0)) {
- # 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) {
- $temp = $text;
- $temp =~ s/ /_/g;
- open(IMG, ">$chartdir/${temp}_volume.png") or die "Could not write $chartdir/${temp}_volume.png: $!\n";
- binmode IMG;
- print IMG $gd->png;
- close IMG;
- print $htm_fh "<img src=\"$chartrel/${temp}_volume.png\">";
+ print $txt_fh "\n" if $txt_fh;
+ if ($htm_fh) {
+ print $htm_fh "</table>\n";
+ print $htm_fh "</td><td>\n";
+ if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals > 0)) {
+ # 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 "Could not write $chartdir/${temp}_volume.png: $!\n";
+ binmode IMG;
+ print IMG $gd->png;
+ close IMG;
+ print $htm_fh "<img src=\"$chartrel/${temp}_volume.png\">";
+ }
+ }
+ print $htm_fh "</td><td>\n";
+ print $htm_fh "</td></tr></table>\n\n";
}
+
+ ++${$row_sref} if $xls_fh;
}
- print $htm_fh "</td><td>\n";
- print $htm_fh "</td></tr></table>\n\n";
-}
-if ($xls_fh)
-{
- ++$row_league_table;
}
-}
#######################################################################
# top_n_sort();
@@ -1557,6 +1595,12 @@
my $n_minus_1 = $n - 1;
my $n_minus_2 = $n - 2;
+ # Create a dummy hash incase the user has not provided us with
+ # tiebreaker hashes.
+ my(%dummy_hash);
+ $href2 = \%dummy_hash unless defined $href2;
+ $href3 = \%dummy_hash unless defined $href3;
+
# Pick out the top $n keys.
my($key,$value1,$value2,$value3,$i,$comparison,$insert_position);
while (($key,$value1) = each %$href1) {
@@ -1564,7 +1608,18 @@
#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.
+ # top n keys that we're keeping. We test the main key first, because
+ # for the majority of cases we can skip creating dummy hash values
+ # should the user have not provided real tie-breaking hashes.
+ next unless $value1 >= $minimum_value1;
+
+ # Create a dummy hash entry for the key if required.
+ # Note that setting the dummy_hash value sets it for both href2 &
+ # href3. Also note that currently we are guarenteed to have a real
+ # value for href3 if a real value for href2 exists so don't need to
+ # test for it as well.
+ $dummy_hash{$key} = 0 unless exists $href2->{$key};
+
$comparison = $value1 <=> $minimum_value1 ||
$href2->{$key} <=> $minimum_value2 ||
$href3->{$key} <=> $minimum_value3 ||
@@ -1622,6 +1677,7 @@
}
+
#######################################################################
# html_header();
#
@@ -1718,6 +1774,8 @@
from where to include the charts in the html file
-chartdir and -chartrel default to '.'
+-emptyok It is OK if there is no valid input, don't print an error.
+
-d Debug mode - dump the eval'ed parser onto STDERR.
EoText
@@ -1745,8 +1803,9 @@
sub generate_parser {
my $parser = '
my($ip,$host,$email,$edomain,$domain,$thissize,$size,$old,$new);
- my($tod,$m_hour,$m_min,$id,$flag);
+ my($tod,$m_hour,$m_min,$id,$flag,$extra,$length);
my($seconds,$queued,$rcpt_time);
+ my $rej_id = 0;
while (<$fh>) {
# Convert syslog lines to mainlog format.
@@ -1754,15 +1813,34 @@
next unless s/^.*? exim\\b.*?: //;
}
- next if length($_) < 38;
+ $length = length($_);
+ 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;
+ # PH - watch for GMT offsets in the timestamp.
+ if (defined($4)) {
+ $extra = 6;
+ next if ($length < 44);
+ }
+ else {
+ $extra = 0;
+ }
+
$id = substr($_, 20 + $extra, 16);
$flag = substr($_, 37 + $extra, 2);
+
+ if ($flag !~ /^([<>=*-]+|SA)$/ && /rejected|refused|dropped/) {
+ $flag = "Re";
+ $extra -= 3;
+ }
+
+ # Rejects can have no MSGID...
+ if ($flag eq "Re" && $id !~ /^[-0-9a-zA-Z]+$/) {
+ $id = "reject:" . ++$rej_id;
+ $extra -= 17;
+ }
';
# Watch for user specified patterns.
@@ -1777,18 +1855,16 @@
}
$parser .= '
- next unless ($flag =~ /<=|=>|->|==|\\*\\*|Co/);
-
- #Strip away the timestamp, ID and flag (which could be "Com" for completed)
- #This speeds up the later pattern matches.
- # $_ = substr($_, 40);
+ next unless ($flag =~ /<=|=>|->|==|\\*\\*|Co|SA|Re/);
+ #Strip away the timestamp, ID and flag to speed up later pattern matches.
+ #The flags include Co (Completed), Re (Rejected), and SA (SpamAssassin).
$_ = substr($_, 40 + $extra); # PH
- # Get a pointer to an array of information about the message.
+ # Alias @message to the array of information about the message.
# This minimises the number of calls to hash functions.
$messages{$id} = [] unless exists $messages{$id};
- $message_aref = $messages{$id};
+ *message = $messages{$id};
# JN - Skip over certain transports as specified via the "-nt/.../" command
@@ -1802,33 +1878,25 @@
#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;
+ # Host is local unless otherwise specified.
+ $ip = (/\\bH=.*?(\\[[^]]+\\])/) ? $1 : "local";
+ $host = (/\\bH=(\\S+)/) ? $1 : "local";
- #IFDEF ($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})
+ $domain = "localdomain"; #Domain is localdomain unless otherwise specified.
+ #IFDEF ($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 ($do_sender{Email})
#IFDEF ($include_original_destination)
@@ -1866,16 +1934,16 @@
if ($flag eq "<=") {
$thissize = (/\\sS=(\\d+)( |$)/) ? $1 : 0;
- $message_aref->[$SIZE] = $thissize;
- $message_aref->[$PROTOCOL] = (/ P=(\S+)/) ? $1 : undef;
+ $message[$SIZE] = $thissize;
+ $message[$PROTOCOL] = (/ P=(\S+)/) ? $1 : undef;
#IFDEF ($show_relay)
if ($host ne "local") {
# Save incoming information in case it becomes interesting
# later, when delivery lines are read.
my($from) = /^(\\S+)/;
- $message_aref->[$FROM_HOST] = "$host$ip";
- $message_aref->[$FROM_ADDRESS] = $from;
+ $message[$FROM_HOST] = "$host$ip";
+ $message[$FROM_ADDRESS] = $from;
}
#ENDIF ($show_relay)
@@ -1927,7 +1995,7 @@
add_volume(\\$total_received_data,\\$total_received_data_gigs,$thissize);
#IFDEF ($#queue_times >= 0 || $#rcpt_times >= 0)
- $message_aref->[$ARRIVAL_TIME] = $tod;
+ $message[$ARRIVAL_TIME] = $tod;
#ENDIF ($#queue_times >= 0 || $#rcpt_times >= 0)
#IFDEF ($hist_opt > 0)
@@ -1936,9 +2004,9 @@
}
elsif ($flag eq "=>") {
- $size = $message_aref->[$SIZE] || 0;
+ $size = $message[$SIZE] || 0;
if ($host ne "local") {
- $message_aref->[$REMOTE_DELIVERED] = 1;
+ $message[$REMOTE_DELIVERED] = 1;
#IFDEF ($show_relay)
@@ -1948,7 +2016,7 @@
# addresses, there may be a further address between the first
# and last.
- if (defined $message_aref->[$FROM_HOST]) {
+ if (defined $message[$FROM_HOST]) {
if (/^(\\S+)(?:\\s+\\([^)]\\))?\\s+<([^>]+)>/) {
($old,$new) = ($1,$2);
}
@@ -1958,7 +2026,7 @@
if ("\\L$new" eq "\\L$old") {
($old) = /^(\\S+)/ if $old eq "";
- my $key = "H=\\L$message_aref->[$FROM_HOST]\\E A=\\L$message_aref->[$FROM_ADDRESS]\\E => " .
+ my $key = "H=\\L$message[$FROM_HOST]\\E A=\\L$message[$FROM_ADDRESS]\\E => " .
"H=\\L$host\\E$ip A=\\L$old\\E";
if (!defined $relay_pattern || $key !~ /$relay_pattern/o) {
$relayed{$key} = 0 if !defined $relayed{$key};
@@ -1996,32 +2064,38 @@
my($parent) = $_ =~ /(<[^@]+@?[^>]*>)/;
$user = "$user $parent" if defined $parent;
}
- ++$delivered_count_user{$user};
+ ++$delivered_messages_user{$user};
+ ++$delivered_addresses_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}++;
+ $delivered_messages{Host}{$host}++;
+ $delivered_addresses{Host}{$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}{$domain};
+ ++$delivered_messages{Domain}{$domain};
+ ++$delivered_addresses{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};
+ ++$delivered_messages{Email}{$email};
+ ++$delivered_addresses{Email}{$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};
+ ++$delivered_messages{Edomain}{$edomain};
+ ++$delivered_addresses{Edomain}{$edomain};
add_volume(\\$delivered_data{Edomain}{$edomain},\\$delivered_data_gigs{Edomain}{$edomain},$size);
#ENDIF ($do_sender{Edomain})
- ++$total_delivered_count;
+ ++$total_delivered_messages;
+ ++$total_delivered_addresses;
add_volume(\\$total_delivered_data,\\$total_delivered_data_gigs,$size);
#IFDEF ($show_transport)
@@ -2040,29 +2114,77 @@
for ($i = 0; $i <= $#delivery_times; $i++) {
if ($seconds < $delivery_times[$i]) {
++$dt_all_bin[$i];
- ++$dt_remote_bin[$i] if $message_aref->[$REMOTE_DELIVERED];
+ ++$dt_remote_bin[$i] if $message[$REMOTE_DELIVERED];
last;
}
}
if ($i > $#delivery_times) {
++$dt_all_overflow;
- ++$dt_remote_overflow if $message_aref->[$REMOTE_DELIVERED];
+ ++$dt_remote_overflow if $message[$REMOTE_DELIVERED];
}
}
#ENDIF ($#delivery_times > 0)
}
- elsif ($flag eq "==" && defined($message_aref->[$SIZE]) && !defined($message_aref->[$DELAYED])) {
+ elsif ($flag eq "->") {
+
+ #IFDEF ($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</)? " <" : " ", $_)) {
+ #IFDEF ($include_original_destination)
+ {
+ #ENDIF ($include_original_destination)
+ #IFNDEF ($include_original_destination)
+ if ($user =~ /^[\\/|]/) {
+ #ENDIF ($include_original_destination)
+ my($parent) = $_ =~ /(<[^@]+@?[^>]*>)/;
+ $user = "$user $parent" if defined $parent;
+ }
+ ++$delivered_addresses_user{$user};
+ }
+ }
+ #ENDIF ($local_league_table || $include_remote_users)
+
+ #IFDEF ($do_sender{Host})
+ $delivered_addresses{Host}{$host}++;
+ #ENDIF ($do_sender{Host})
+ #IFDEF ($do_sender{Domain})
+ if ($domain) {
+ ++$delivered_addresses{Domain}{$domain};
+ }
+ #ENDIF ($do_sender{Domain})
+ #IFDEF ($do_sender{Email})
+ ++$delivered_addresses{Email}{$email};
+ #ENDIF ($do_sender{Email})
+ #IFDEF ($do_sender{Edomain})
+ ++$delivered_addresses{Edomain}{$edomain};
+ #ENDIF ($do_sender{Edomain})
+
+ ++$total_delivered_addresses;
+ }
+
+ elsif ($flag eq "==" && defined($message[$SIZE]) && !defined($message[$DELAYED])) {
++$delayed_count;
- $message_aref->[$DELAYED] = 1;
+ $message[$DELAYED] = 1;
}
elsif ($flag eq "**") {
- if (defined ($message_aref->[$SIZE])) {
- unless (defined $message_aref->[$HAD_ERROR]) {
+ if (defined ($message[$SIZE])) {
+ unless (defined $message[$HAD_ERROR]) {
++$message_errors;
- $message_aref->[$HAD_ERROR] = 1;
+ $message[$HAD_ERROR] = 1;
}
}
@@ -2075,18 +2197,18 @@
elsif ($flag eq "Co") {
#Completed?
#IFDEF ($#queue_times >= 0)
- $queued = queue_time($tod, $message_aref->[$ARRIVAL_TIME], $id);
+ $queued = queue_time($tod, $message[$ARRIVAL_TIME], $id);
for ($i = 0; $i <= $#queue_times; $i++) {
if ($queued < $queue_times[$i]) {
++$qt_all_bin[$i];
- ++$qt_remote_bin[$i] if $message_aref->[$REMOTE_DELIVERED];
+ ++$qt_remote_bin[$i] if $message[$REMOTE_DELIVERED];
last;
}
}
if ($i > $#queue_times) {
++$qt_all_overflow;
- ++$qt_remote_overflow if $message_aref->[$REMOTE_DELIVERED];
+ ++$qt_remote_overflow if $message[$REMOTE_DELIVERED];
}
#ENDIF ($#queue_times >= 0)
@@ -2095,13 +2217,13 @@
$seconds = wdhms_seconds($1);
#Calculate $queued if not previously calculated above.
#IFNDEF ($#queue_times >= 0)
- $queued = queue_time($tod, $message_aref->[$ARRIVAL_TIME], $id);
+ $queued = queue_time($tod, $message[$ARRIVAL_TIME], $id);
#ENDIF ($#queue_times >= 0)
$rcpt_time = $seconds - $queued;
my($protocol);
- if (defined $message_aref->[$PROTOCOL]) {
- $protocol = $message_aref->[$PROTOCOL];
+ if (defined $message[$PROTOCOL]) {
+ $protocol = $message[$PROTOCOL];
# Create the bin if its not already defined.
unless (exists $rcpt_times_bin{$protocol}) {
@@ -2127,6 +2249,97 @@
delete($messages{$id});
}
+ elsif ($flag eq "SA") {
+ $ip = (/From.*?(\\[[^]]+\\])/ || /\\((local)\\)/) ? $1 : "";
+ #SpamAssassin message
+ if (/Action: ((permanently|temporarily) rejected message|flagged as Spam but accepted): score=(\d+\.\d)/) {
+ #add_volume(\\$spam_score,\\$spam_score_gigs,$3);
+ ++$spam_count_by_ip{$ip};
+ } elsif (/Action: scanned but message isn\'t spam: score=(-?\d+\.\d)/) {
+ #add_volume(\\$ham_score,\\$ham_score_gigs,$1);
+ ++$ham_count_by_ip{$ip};
+ } elsif (/(Not running SA because SAEximRunCond expanded to false|check skipped due to message size)/) {
+ ++$ham_count_by_ip{$ip};
+ }
+ }
+
+ # Look for Reject messages or blackholed messages (deliveries
+ # without a transport)
+ if ($flag eq "Re" || ($flag eq "=>" && ! /\\sT=\\S+/)) {
+ # Correct the IP address for rejects:
+ # rejected EHLO from my.test.net [10.0.0.5]: syntactically invalid argument(s):
+ $ip = $1 if ($ip eq "local" && /^rejected [HE][HE]LO from .*? (\[.+?\]):/);
+ ++$rejected_count_by_ip{$ip};
+ if (
+ /(listed at [^ ]+)/ ||
+ /(Forged IP detected in HELO)/ ||
+ /(Invalid domain or IP given in HELO\/EHLO)/ ||
+ /(unqualified recipient rejected)/ ||
+ /(closed connection (after|in response) .*?)\s*$/ ||
+ /(sender rejected)/ ||
+ # 2005-09-23 15:07:49 1EInHJ-0007Ex-Au H=(a.b.c) [10.0.0.1] F=<> rejected after DATA: This message contains a virus: (Eicar-Test-Signature) please scan your system.
+ # 2005-10-06 10:50:07 1ENRS3-0000Nr-Kt => blackhole (DATA ACL discarded recipients): This message contains a virus: (Worm.SomeFool.P) please scan your system.
+ / rejected after DATA: (.*)/ ||
+ /.DATA ACL discarded recipients.: (.*)/ ||
+ /rejected after DATA: (unqualified address not permitted)/ ||
+ /(VRFY rejected)/ ||
+# /(sender verify (defer|fail))/i ||
+ /(too many recipients)/ ||
+ /(refused relay.*?) to/ ||
+ /(rejected by non-SMTP ACL: .*)/ ||
+ /(rejected by local_scan.*)/ ||
+ # SMTP call from %s dropped: too many syntax or protocol errors (last command was "%s"
+ # SMTP call from %s dropped: too many nonmail commands
+ /(dropped: too many ((nonmail|unrecognized) commands|syntax or protocol errors))/ ||
+
+ # local_scan() function crashed with signal %d - message temporarily rejected
+ # local_scan() function timed out - message temporarily rejected
+ /(local_scan.. function .* - message temporarily rejected)/ ||
+ /(temporarily refused connection)/ ||
+ # SMTP protocol synchronization error (input sent without waiting for greeting): rejected connection from %s
+ /(SMTP protocol .*?(error|violation))/ ||
+ /(message too big)/
+ ) {
+ ++$rejected_count_by_reason{"\u$1"};
+ }
+ elsif (/rejected [HE][HE]LO from [^:]*: syntactically invalid argument/) {
+ ++$rejected_count_by_reason{"Rejected HELO/EHLO: syntactically invalid argument"};
+ }
+ elsif (/response to "RCPT TO.*? was: (.*)/) {
+ ++$rejected_count_by_reason{"Response to RCPT TO was: $1"};
+ }
+ elsif (
+ /(lookup of host )\S+ (failed)/ ||
+
+ # rejected from <%s>%s%s%s%s: message too big:
+ /(rejected [A-Z]*) .*?(: .*?)(:|\s*$)/ ||
+ # refused connection from %s (host_reject_connection)
+ # refused connection from %s (tcp wrappers)
+ /(refused connection )from.*? (\(.*)/ ||
+
+ # error from remote mailer after RCPT TO:<a@???>: host a.b.c [10.0.0.1]: 450 <a@???>: Recipient address rejected: Greylisted for 60 seconds
+ # error from remote mailer after MAIL FROM:<> SIZE=3468: host a.b.c [10.0.0.1]: 421 a.b.c has refused your connection because your server did not have a PTR record.
+ /(error from remote mailer after .*?:).*(: .*?)(:|\s*$)/ ||
+
+ # a.b.c F=<a@???> rejected after DATA: "@" or "." expected after "Undisclosed-Recipient": failing address in "To" header is: <Undisclosed-Recipient:;>
+ /rejected after DATA: ("." or "." expected).*?(: failing address in .*? header)/ ||
+
+ # connection from %s refused load average = %.2f
+ /(Connection )from.*? (refused: load average)/ ||
+ # connection from %s refused (IP options)
+ # Connection from %s refused: too many connections
+ # connection from %s refused
+ /([Cc]onnection )from.*? (refused.*)/ ||
+ # [10.0.0.1]: connection refused
+ /: (Connection refused)()/
+ ) {
+ ++$rejected_count_by_reason{"\u$1$2"};
+ }
+ else {
+ ++$rejected_count_by_reason{Unknown};
+ print STDERR "Unknown rejection: $_" if $debug;
+ }
+ }
}';
# We now do a 'C preprocessor style operation on our parser
@@ -2224,12 +2437,13 @@
print $htm_fh "<li><a href=\"#Relayed messages\">Relayed messages</a>\n" if $show_relay;
if ($topcount) {
+ print $htm_fh "<li><a href=\"#mail rejection reason count\">Top $topcount mail rejection reasons by message count</a>\n" if %rejected_count_by_reason;
foreach ('Host','Domain','Email','Edomain') {
next unless $do_sender{$_};
print $htm_fh "<li><a href=\"#sending \l$_ count\">Top $topcount sending \l${_}s by message count</a>\n";
print $htm_fh "<li><a href=\"#sending \l$_ volume\">Top $topcount sending \l${_}s by volume</a>\n";
}
- if ($local_league_table || $include_remote_users) {
+ if (($local_league_table || $include_remote_users) && %received_count_user) {
print $htm_fh "<li><a href=\"#local sender count\">Top $topcount local senders by message count</a>\n";
print $htm_fh "<li><a href=\"#local sender volume\">Top $topcount local senders by volume</a>\n";
}
@@ -2238,10 +2452,14 @@
print $htm_fh "<li><a href=\"#\l$_ destination count\">Top $topcount \l$_ destinations by message count</a>\n";
print $htm_fh "<li><a href=\"#\l$_ destination volume\">Top $topcount \l$_ destinations by volume</a>\n";
}
- if ($local_league_table || $include_remote_users) {
+ if (($local_league_table || $include_remote_users) && %delivered_messages_user) {
print $htm_fh "<li><a href=\"#local destination count\">Top $topcount local destinations by message count</a>\n";
print $htm_fh "<li><a href=\"#local destination volume\">Top $topcount local destinations by volume</a>\n";
}
+
+ print $htm_fh "<li><a href=\"#rejected ip count\">Top $topcount rejected ips by message count</a>\n" if %rejected_count_by_ip;
+ print $htm_fh "<li><a href=\"#non-rejected spamming ip count\">Top $topcount non-rejected spamming ips by message count</a>\n" if %spam_count_by_ip;
+
}
print $htm_fh "<li><a href=\"#errors\">List of errors</a>\n" if %errors_count;
print $htm_fh "</ul>\n<hr>\n";
@@ -2266,9 +2484,10 @@
# Get the sender by headings and results. This is complicated as we can have
# different numbers of columns.
- my($sender_txt_header,$sender_html_header,$sender_txt_format,$sender_html_format);
+ my($sender_txt_header,$sender_txt_format,$sender_html_format);
my(@received_totals,@delivered_totals);
my($row_tablehead, $row_max);
+ my(@col_headers) = ('TOTAL', 'Volume', 'Messages', 'Addresses');
foreach ('Host','Domain','Email','Edomain') {
next unless $do_sender{$_};
@@ -2280,15 +2499,15 @@
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";
+ push(@col_headers,"${_}s");
}
- my $txt_format1 = " %-16s %9s %6d $sender_txt_format";
+ my $txt_format1 = " %-16s %9s %6d %6s $sender_txt_format";
my $txt_format2 = " %6d %4.1f%% %6d %4.1f%%",
- my $htm_format1 = "<tr><td>%s</td><td align=\"right\">%s</td>$sender_html_format<td align=\"right\">%d</td>";
+ my $htm_format1 = "<tr><td>%s</td><td align=\"right\">%s</td><td align=\"right\">%s</td><td align=\"right\">%s</td>$sender_html_format";
my $htm_format2 = "<td align=\"right\">%d</td><td align=\"right\">%4.1f%%</td><td align=\"right\">%d</td><td align=\"right\">%4.1f%%</td>";
if ($txt_fh) {
@@ -2296,43 +2515,24 @@
print $txt_fh "\n";
print $txt_fh "Grand total summary\n";
print $txt_fh "-------------------\n";
- print $txt_fh " $sender_spaces At least one address\n";
- print $txt_fh " TOTAL Volume Messages $sender_txt_header Delayed Failed\n";
+ print $txt_fh " $sender_spaces At least one address\n";
+ print $txt_fh " TOTAL Volume Messages Addresses $sender_txt_header Delayed Failed\n";
}
if ($htm_fh) {
print $htm_fh "<a name=\"grandtotal\"></a>\n";
print $htm_fh "<h2>Grand total summary</h2>\n";
print $htm_fh "<table border=1>\n";
- print $htm_fh "<tr><th>TOTAL</th><th>Volume</th><th>Messages</th>$sender_html_header<th colspan=2>At least one addr<br>Delayed</th><th colspan=2>At least one addr<br>Failed</th>\n";
+ print $htm_fh "<tr><th>" . join('</th><th>',@col_headers) . "</th><th colspan=2>At least one addr<br>Delayed</th><th colspan=2>At least one addr<br>Failed</th>\n";
}
- if ($xls_fh)
- {
- $ws_global->write($row++, $col, "Grand total summary", $f_header2);
-
- $row_tablehead = $row+1; # header-row of TOTALS table
-
- &set_worksheet_line($ws_global, $row_tablehead, 0, ['Received', 'Delivered', 'TOTAL'], $f_headertab);
-
- my @content= (
- "Volume",
- "Messages",
- $sender_txt_header,
- "At least one address Delayed (Total)",
- "At least one address Delayed (Percent)",
- "At least one address Failed (Total)",
- "At least one address Failed (Percent)"
- );
-
- for (my $i=0; $i < scalar(@content); $i++)
- {
- $ws_global->write($row_tablehead+$i+1, 2, $content[$i], $f_default);
- $row++;
- }
- $row_max = $row_tablehead+scalar(@content)+2; # continue from this row
+ if ($xls_fh) {
+ $ws_global->write($row++, 0, "Grand total summary", $f_header2);
+ $ws_global->write($row, 0, \@col_headers, $f_header2);
+ $ws_global->merge_range($row, scalar(@col_headers), $row, scalar(@col_headers)+1, "At least one addr Delayed", $f_header2_m);
+ $ws_global->merge_range($row, scalar(@col_headers)+2, $row, scalar(@col_headers)+3, "At least one addr Failed", $f_header2_m);
+ #$ws_global->write(++$row, scalar(@col_headers), ['Total','Percent','Total','Percent'], $f_header2);
}
-
my($volume,$failed_count);
if ($merge_reports) {
$volume = volume_rounded($report_totals{Received}{Volume}, $report_totals{Received}{'Volume-gigs'});
@@ -2349,7 +2549,7 @@
no integer;
my @content=(
- $volume,$total_received_count,
+ $volume,$total_received_count,'',
@received_totals,
$delayed_count,
($total_received_count) ? ($delayed_count*100/$total_received_count) : 0,
@@ -2359,42 +2559,73 @@
printf $txt_fh ("$txt_format1$txt_format2\n", 'Received', @content) if $txt_fh;
printf $htm_fh ("$htm_format1$htm_format2\n", 'Received', @content) if $htm_fh;
- if ($xls_fh)
- {
- $row = $row_tablehead+1;
- for (my $i=0; $i < scalar(@content); $i++)
- {
+ if ($xls_fh) {
+ $ws_global->write(++$row, 0, 'Received', $f_default);
+ for (my $i=0; $i < scalar(@content); $i++) {
if ($i == 4 || $i == 6) {
- $ws_global->write($row+$i, 0, $content[$i]/100, $f_percent);
+ $ws_global->write($row, $i+1, $content[$i]/100, $f_percent);
}
else {
- $ws_global->write($row+$i, 0, $content[$i], $f_default);
+ $ws_global->write($row, $i+1, $content[$i], $f_default);
}
}
}
}
+
if ($merge_reports) {
$volume = volume_rounded($report_totals{Delivered}{Volume}, $report_totals{Delivered}{'Volume-gigs'});
- $total_delivered_count = get_report_total($report_totals{Delivered},'Messages');
+ $total_delivered_messages = get_report_total($report_totals{Delivered},'Messages');
+ $total_delivered_addresses = get_report_total($report_totals{Delivered},'Addresses');
}
else {
$volume = volume_rounded($total_delivered_data, $total_delivered_data_gigs);
}
- my @content=($volume, $total_delivered_count, @delivered_totals);
- printf $txt_fh ("$txt_format1\n\n", 'Delivered', @content) if $txt_fh;
- printf $htm_fh ("$htm_format1\n\n", 'Delivered', @content) if $htm_fh;
- printf $htm_fh "</table>\n" if $htm_fh;
- if ($xls_fh)
- {
+ my @content=($volume, $total_delivered_messages, $total_delivered_addresses, @delivered_totals);
+ printf $txt_fh ("$txt_format1\n", 'Delivered', @content) if $txt_fh;
+ printf $htm_fh ("$htm_format1\n", 'Delivered', @content) if $htm_fh;
- $row = $row_tablehead+1;
- for (my $i=0; $i < scalar(@content); $i++)
- {
- $ws_global->write($row+$i, 1, $content[$i], $f_default);
+ if ($xls_fh) {
+ $ws_global->write(++$row, 0, 'Delivered', $f_default);
+ for (my $i=0; $i < scalar(@content); $i++) {
+ $ws_global->write($row, $i+1, $content[$i], $f_default);
+ }
+ }
+
+ if ($merge_reports) {
+ foreach ('Rejects', 'Ham', 'Spam') {
+ my $messages = get_report_total($report_totals{$_},'Messages');
+ my $addresses = get_report_total($report_totals{$_},'Addresses');
+ if ($messages) {
+ @content = ($_, '', $messages, '');
+ push(@content,get_report_total($report_totals{$_},'Hosts')) if $do_sender{Host};
+ printf $txt_fh ("$txt_format1\n", @content) if $txt_fh;
+ printf $htm_fh ("$htm_format1\n", @content) if $htm_fh;
+ $ws_global->write(++$row, 0, \@content) if $xls_fh;
+ }
+ }
+ }
+ else {
+ foreach my $total_aref (['Rejects',\%rejected_count_by_ip],
+ ['Ham',\%ham_count_by_ip],
+ ['Spam',\%spam_count_by_ip]) {
+ my $messages = 0;
+ map {$messages += $_} values %{$total_aref->[1]};
+
+ if ($messages > 0) {
+ @content = ($total_aref->[0], '', $messages, '');
+ push(@content,scalar(keys %{$total_aref->[1]})) if $do_sender{Host};
+
+ printf $txt_fh ("$txt_format1\n", @content) if $txt_fh;
+ printf $htm_fh ("$htm_format1\n", @content) if $htm_fh;
+ $ws_global->write(++$row, 0, \@content) if $xls_fh;
}
- $row = $row_max;
+ }
}
+
+ printf $txt_fh "\n" if $txt_fh;
+ printf $htm_fh "</table>\n" if $htm_fh;
+ ++$row;
}
@@ -2446,10 +2677,7 @@
foreach $key (@user_descriptions) {
printf $txt_fh ("$txt_format1\n",$key,$user_pattern_totals[$user_pattern_index]) if $txt_fh;
printf $htm_fh ("$htm_format1\n",$key,$user_pattern_totals[$user_pattern_index]) if $htm_fh;
- if ($xls_fh)
- {
- &set_worksheet_line($ws_global, $row++, 0, [$key,$user_pattern_totals[$user_pattern_index]]);
- }
+ $ws_global->write($row++, 0, [$key,$user_pattern_totals[$user_pattern_index]]) if $xls_fh;
$user_pattern_index++;
}
}
@@ -2469,6 +2697,79 @@
}
}
+#######################################################################
+# print_rejects()
+#
+# print_rejects();
+#
+# Print statistics about rejected mail.
+#######################################################################
+sub print_rejects {
+ my($format1,$reason);
+
+ my $txt_format1 = " %-40s %6d";
+ my $htm_format1 = "<tr><td>%s</td><td align=\"right\">%d</td>";
+
+ if ($txt_fh) {
+ print $txt_fh "Rejected mail by reason\n";
+ print $txt_fh "-----------------------";
+ print $txt_fh "\n Total\n";
+ }
+ if ($htm_fh) {
+ print $htm_fh "<hr><a name=\"patterns\"></a><h2>Rejected mail by reason</h2>\n";
+ print $htm_fh "<table border=0 width=\"100%\"><tr><td><table border=1>\n";
+ print $htm_fh "<tr><th> </th><th>Total</th>\n";
+ }
+ if ($xls_fh) {
+ $ws_global->write($row++, $col, "Rejected mail by reason", $f_header2);
+ &set_worksheet_line($ws_global, $row++, 1, ["Total"], $f_headertab);
+ }
+
+
+ my $href = ($merge_reports) ? $report_totals{rejected_mail_by_reason} : \%rejected_count_by_reason;
+ my(@chartdatanames, @chartdatavals_count);
+
+ foreach $reason (top_n_sort($topcount, $href, undef, undef)) {
+ printf $txt_fh ("$txt_format1\n",$reason,$href->{$reason}) if $txt_fh;
+ printf $htm_fh ("$htm_format1\n",$reason,$href->{$reason}) if $htm_fh;
+ set_worksheet_line($ws_global, $row++, 0, [$reason,$href->{$reason}], $f_default) if $xls_fh;
+ push(@chartdatanames, $reason);
+ push(@chartdatavals_count, $href->{$reason});
+ }
+
+ $row++ if $xls_fh;
+ print $txt_fh "\n" if $txt_fh;
+
+ if ($htm_fh) {
+ print $htm_fh "</tr></table></td><td>";
+ if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals_count > 0)) {
+ # calculate the graph
+ my @data = (
+ \@chartdatanames,
+ \@chartdatavals_count
+ );
+ my $graph = GD::Graph::pie->new(200, 200);
+ $graph->set(
+ x_label => 'Rejection Reasons',
+ y_label => 'Messages',
+ title => 'By count',
+ );
+ my $gd = $graph->plot(\@data) or warn($graph->error);
+ if ($gd) {
+ open(IMG, ">$chartdir/rejections_count.png") or die "Could not write $chartdir/rejections_count.png: $!\n";
+ binmode IMG;
+ print IMG $gd->png;
+ close IMG;
+ print $htm_fh "<img src=\"$chartrel/rejections_count.png\">";
+ }
+ }
+ print $htm_fh "</td></tr></table>\n\n";
+ }
+}
+
+
+
+
#######################################################################
# print_transport();
@@ -2493,14 +2794,12 @@
}
if ($htm_fh) {
print $htm_fh "<hr><a name=\"transport\"></a><h2>Deliveries by Transport</h2>\n";
- print $htm_fh "<table border=0 width=\"100%\">\n";
- print $htm_fh "<tr><td>\n";
- print $htm_fh "<table border=1>\n";
+ print $htm_fh "<table border=0 width=\"100%\"><tr><td><table border=1>\n";
print $htm_fh "<tr><th> </th><th>Volume</th><th>Messages</th>\n";
}
if ($xls_fh) {
- $ws_global->write($row++, $col, "Deliveries by transport", $f_header2);
- &set_worksheet_line($ws_global, $row++, 1, ["Volume", "Messages"], $f_headertab);
+ $ws_global->write(++$row, $col, "Deliveries by transport", $f_header2);
+ $ws_global->write(++$row, 1, ["Volume", "Messages"], $f_headertab);
}
my($key);
@@ -2515,9 +2814,7 @@
push(@chartdatavals_vol, $report_totals{transport}{$key}{'Volume-gigs'}*$gig + $report_totals{transport}{$key}{Volume} );
printf $txt_fh ("$txt_format1\n", @content) if $txt_fh;
printf $htm_fh ("$htm_format1\n", @content) if $htm_fh;
- if ($xls_fh) {
- &set_worksheet_line($ws_global, $row++, 0, \@content, $f_default);
- }
+ $ws_global->write(++$row, 0, \@content) if $xls_fh;
}
}
else {
@@ -2530,15 +2827,13 @@
push(@chartdatavals_vol, $transported_data_gigs{$key}*$gig + $transported_data{$key});
printf $txt_fh ("$txt_format1\n", @content) if $txt_fh;
printf $htm_fh ("$htm_format1\n", @content) if $htm_fh;
- if ($xls_fh) {
- &set_worksheet_line($ws_global, $row++, 0, \@content);
- }
+ $ws_global->write(++$row, 0, \@content) if $xls_fh;
}
}
print $txt_fh "\n" if $txt_fh;
if ($htm_fh) {
- print $htm_fh "</table>\n";
- print $htm_fh "</td><td>\n";
+ print $htm_fh "</tr></table></td><td>";
+
if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals_count > 0))
{
# calculate the graph
@@ -2561,7 +2856,7 @@
print $htm_fh "<img src=\"$chartrel/transports_count.png\">";
}
}
- print $htm_fh "</td><td>\n";
+ print $htm_fh "</td><td>";
if ($HAVE_GD_Graph_pie && $charts && ($#chartdatavals_vol > 0)) {
my @data = (
@@ -2574,19 +2869,16 @@
);
my $gd = $graph->plot(\@data) or warn($graph->error);
if ($gd) {
- open(IMG, ">$chartdir/transports_vol.png") or die "Could not write $chartdir/transports_count.png: $!\n";
+ open(IMG, ">$chartdir/transports_vol.png") or die "Could not write $chartdir/transports_vol.png: $!\n";
binmode IMG;
print IMG $gd->png;
close IMG;
print $htm_fh "<img src=\"$chartrel/transports_vol.png\">";
}
}
+
print $htm_fh "</td></tr></table>\n\n";
}
- if ($xls_fh) {
- $row++;
- }
-
}
@@ -2790,25 +3082,50 @@
$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);
+ # Fill in $report_totals{Received|Delivered}{Volume|Messages|Addresses|Hosts|Domains|...|Delayed|DelayedPercent|Failed|FailedPercent}
+ my(@fields, @delivered_fields);
+ my $doing_table = 0;
while (<$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
+# TOTAL Volume Messages Addresses Hosts Domains Delayed Failed
+# Received 26MB 237 177 23 8 3.4% 28 11.8%
+# Delivered 13MB 233 250 99 88
if (/TOTAL\s+(.*?)\s*$/) {
- @fields = split(/\s+/,$1);
+ $doing_table = 1;
+ @delivered_fields = split(/\s+/,$1);
+
#Delayed and Failed have two columns each, so add the extra field names in.
- splice(@fields,-1,1,'DelayedPercent','Failed','FailedPercent');
+ splice(@delivered_fields,-1,1,'DelayedPercent','Failed','FailedPercent');
+
+ # Addresses only figure in the Delivered row, so remove them from the
+ # normal fields.
+ @fields = grep !/Addresses/, @delivered_fields;
}
- elsif (/(Received|Delivered)\s+(.*?)\s*$/) {
+ elsif (/(Received)\s+(.*?)\s*$/) {
print STDERR "Parsing $_" if $debug;
add_to_totals($report_totals{$1},\@fields,$2);
}
- last if (/Delivered/); #Last line of this section.
+ elsif (/(Delivered)\s+(.*?)\s*$/) {
+ print STDERR "Parsing $_" if $debug;
+ add_to_totals($report_totals{$1},\@delivered_fields,$2);
+ my $data = $2;
+ # If we're merging an old report which doesn't include addresses,
+ # then use the Messages field instead.
+ unless (grep(/Addresses/, @delivered_fields)) {
+ my %tmp;
+ line_to_hash(\%tmp,\@delivered_fields,$data);
+ add_to_totals($report_totals{Delivered},['Addresses'],$tmp{Messages});
+ }
+ }
+ elsif (/(Rejects|Ham|Spam)\s+(.*?)\s*$/) {
+ print STDERR "Parsing $_" if $debug;
+ add_to_totals($report_totals{$1},['Messages','Hosts'],$2);
+ }
+ else {
+ last if $doing_table;
+ }
}
}
@@ -2909,10 +3226,15 @@
}
- my $reached_table = 0;
+ my ($blank_lines, $reached_table) = (0,0);
while (<$fh>) {
$_ = html2txt($_); #Convert general HTML markup to text.
- $reached_table = 1 if (/^\s*Under/);
+ # The table is preceded by one blank line, and has one blank line
+ # following it. As the table may be empty, the best way to determine
+ # that we've finished it is to look for the second blank line.
+ ++$blank_lines if /^\s*$/;
+ last if ($blank_lines >=2); #Finished the table ?
+ $reached_table = 1 if (/\d/);
next unless $reached_table;
my $previous_seconds_on_queue = 0;
if (/^\s*(Under|Over|)\s+(\d+[smhdw])\s+(\d+)/) {
@@ -2932,9 +3254,6 @@
$$overflow_sref += $count if ($i > $#$times_aref);
}
- else {
- last; #Finished the table ?
- }
}
}
@@ -2992,73 +3311,76 @@
#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);
+ my($messages_href,$addresses_href,$data_href,$data_gigs_href);
if ($category =~ /local sender/) {
- $count_href = \%received_count_user;
+ $messages_href = \%received_count_user;
+ $addresses_href = undef;
$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"}};
+ $messages_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;
+ $messages_href = \%delivered_messages_user;
+ $addresses_href = \%delivered_addresses_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"}};
+ $messages_href = \%{$delivered_messages{"\u$1"}};
+ $addresses_href = \%{$delivered_addresses{"\u$1"}};
$data_href = \%{$delivered_data{"\u$1"}};
$data_gigs_href = \%{$delivered_data_gigs{"\u$1"}};
}
+ elsif ($category =~ /rejected ips/) {
+ $messages_href = \%rejected_count_by_ip;
+ }
+ elsif ($category =~ /non-rejected spamming ips/) {
+ $messages_href = \%spam_count_by_ip;
+ }
+ elsif ($category =~ /mail rejection reasons/) {
+ $messages_href = \%rejected_count_by_reason;
+ }
my $reached_table = 0;
+ my $row_re;
while (<$fh>) {
# Watch out for empty tables.
- goto PARSE_OLD_REPORT_LINE if (/<h2>/ or /^[a-zA-Z]/);
+ goto PARSE_OLD_REPORT_LINE if (/<h2>/ or (/^\s*[a-zA-Z]/ && !/^\s*Messages/));
$_ = html2txt($_); #Convert general HTML markup to text.
-
- $reached_table = 1 if (/^\s*\d/);
+ # Messages Addresses Bytes Average
+ if (/^\s*Messages/) {
+ my $pattern = '^\s*(\d+)';
+ $pattern .= (/Addresses/) ? '\s+(\d+)' : '()';
+ $pattern .= (/Bytes/) ? '\s+([\dKMGB]+)' : '()';
+ $pattern .= (/Average/) ? '\s+[\dKMGB]+' : '';
+ $pattern .= '\s+(.*?)\s*$';
+ $row_re = qr/$pattern/;
+ $reached_table = 1;
+ next;
+ }
next unless $reached_table;
- # Remove optional 'average value' column.
- s/^\s*(\d+)\s+(\S+)\s+(\d+(KB|MB|GB|\b)\s+)/$1 $2 /;
-
- 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;
- $$data_gigs_href{$entry} = 0;
- $league_table_value_was_zero{$entry} = 1;
- }
+ my($messages, $addresses, $rounded_volume, $entry);
- $$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 $debug;
- }
+ if (/$row_re/) {
+ ($messages, $addresses, $rounded_volume, $entry) = ($1, $2, $3, $4);
}
- else { #Finished the table ?
+ else {
+ #Else we have finished the table and we may need to do some
+ #kludging to retain the order of the entries.
+
if ($by_count_or_volume =~ /volume/) {
#Add a few bytes to appropriate entries to preserve the order.
-
- my($rounded_volume);
foreach $rounded_volume (keys %table_order) {
#For each rounded volume, we want to create a list which has things
#ordered from the volume table at the front, and additional things
@@ -3082,9 +3404,37 @@
}
}
}
-
last;
}
+
+ # Store a new table entry.
+
+ # Add the entry into the %table_order hash if it has a rounded
+ # volume (KB/MB/GB).
+ push(@{$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 ($$messages_href{$entry}) {
+ $$messages_href{$entry} = 0;
+ $$addresses_href{$entry} = 0;
+ $$data_href{$entry} = 0;
+ $$data_gigs_href{$entry} = 0;
+ $league_table_value_was_zero{$entry} = 1;
+ }
+
+ $$messages_href{$entry} += $messages;
+
+ # When adding the addresses, be aware that we could be merging
+ # an old report which does not include addresses. In this case,
+ # we add the messages instead.
+ $$addresses_href{$entry} += ($addresses) ? $addresses : $messages;
+
+ #Add the rounded value to the data and data_gigs hashes.
+ un_round($rounded_volume,\$$data_href{$entry},\$$data_gigs_href{$entry}) if $rounded_volume;
+ print STDERR "$category by $by_count_or_volume: added $messages,$rounded_volume to $entry\n" if $debug;
+ }
+
}
}
elsif (/List of errors/) {
@@ -3202,22 +3552,39 @@
sub add_to_totals {
my($totals_href,$keys_aref,$values) = @_;
my(@values) = split(/\s+/,$values);
- my(@keys) = @$keys_aref; #Make a copy as we destroy the one we use.
- my($value);
- foreach $value (@values) {
- my $key = shift(@keys) or next;
- if ($value =~ /%/) {
- $$totals_href{$key} = $value;
+
+ for(my $i = 0; $i < @values && $i < @$keys_aref; ++$i) {
+ my $key = $keys_aref->[$i];
+ if ($values[$i] =~ /%/) {
+ $$totals_href{$key} = $values[$i];
}
else {
$$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} , " . $$totals_href{"$key-gigs"} . "GB.\n" if $debug;
+ un_round($values[$i], \$$totals_href{$key}, \$$totals_href{"$key-gigs"});
+ print STDERR "Added $values[$i] to $key - $$totals_href{$key} , " . $$totals_href{"$key-gigs"} . "GB.\n" if $debug;
}
}
}
+
+#######################################################################
+# line_to_hash();
+#
+# line_to_hash(\%hash,\@keys,$line);
+#
+# Given a line of space seperated values, set them into the provided hash
+# using @keys as the hash keys.
+#######################################################################
+sub line_to_hash {
+ my($href,$keys_aref,$values) = @_;
+ my(@values) = split(/\s+/,$values);
+ for(my $i = 0; $i < @values && $i < @$keys_aref; ++$i) {
+ $$href{$keys_aref->[$i]} = $values[$i];
+ }
+}
+
+
#######################################################################
# get_report_total();
#
@@ -3377,7 +3744,6 @@
$offset_seconds = 0;
$row=1;
-$row_league_table=1;
$col=0;
$col_hist=0;
$run_hist=0;
@@ -3424,6 +3790,7 @@
elsif ($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] =~ /^-emptyok$/) { $emptyOK = 1 }
elsif ($ARGV[0] =~ /^-nvr$/) { $volume_rounding = 0 }
elsif ($ARGV[0] =~ /^-show_rt([,\d\+\-\*\/]+)?$/) { @rcpt_times = parse_time_list($1) }
elsif ($ARGV[0] =~ /^-show_dt([,\d\+\-\*\/]+)?$/) { @delivery_times = parse_time_list($1) }
@@ -3460,8 +3827,7 @@
$do_sender{Host} = 1 unless ($do_sender{Domain} || $do_sender{Email} || $do_sender{Edomain});
# prepare xls Excel Workbook
- if (defined $xls_fh)
- {
+ if (defined $xls_fh) {
# Create a new Excel workbook
$workbook = Spreadsheet::WriteExcel->new($xls_fh);
@@ -3476,9 +3842,6 @@
$ws_relayed = $workbook->addworksheet('Relayed Messages');
$ws_relayed->set_column(1, 2, 80);
}
- if ($topcount) {
- $ws_top50 = $workbook->addworksheet('Deliveries');
- }
if ($show_errors) {
$ws_errors = $workbook->addworksheet('Errors');
}
@@ -3506,6 +3869,13 @@
$f_header2->set_valign();
# $ws_global->write($row++, 2, "Testing Headers 2", $f_header2);
+ # Create another header2 for use in merged cells.
+ $f_header2_m = $workbook->add_format();
+ $f_header2_m->set_bold();
+ $f_header2_m->set_size('8');
+ $f_header2_m->set_valign();
+ $f_header2_m->set_align('center');
+
$f_percent = $workbook->add_format();
$f_percent->set_num_format('0.0%');
@@ -3557,7 +3927,8 @@
$total_delivered_data = 0;
$total_delivered_data_gigs = 0;
-$total_delivered_count = 0;
+$total_delivered_messages = 0;
+$total_delivered_addresses = 0;
$qt_all_overflow = 0;
$qt_remote_overflow = 0;
@@ -3569,7 +3940,7 @@
$begin = "9999-99-99 99:99:99";
$end = "0000-00-00 00:00:00";
my($section,$type);
-foreach $section ('Received','Delivered') {
+foreach $section ('Received','Delivered','Rejects','Ham','Spam') {
foreach $type ('Volume','Messages','Delayed','Failed','Hosts','Domains','Emails','Edomains') {
$report_totals{$section}{$type} = 0;
}
@@ -3612,7 +3983,7 @@
}
-if ($begin eq "9999-99-99 99:99:99") {
+if ($begin eq "9999-99-99 99:99:99" && ! $emptyOK) {
print STDERR "**** No valid log lines read\n";
exit 1;
}
@@ -3624,6 +3995,9 @@
# Print counts of user specified patterns if required.
print_user_patterns() if @user_patterns;
+# Print rejection reasons.
+# print_rejects();
+
# Print totals by transport if required.
print_transport() if $show_transport;
@@ -3658,19 +4032,31 @@
# Print the league tables, if topcount isn't zero.
if ($topcount > 0) {
+ my($ws_rej, $ws_top50, $ws_rej_row, $ws_top50_row);
+ $ws_rej_row = $ws_top50_row = 0;
+ if ($xls_fh) {
+ $ws_top50 = $workbook->addworksheet('Deliveries');
+ $ws_rej = $workbook->addworksheet('Rejections') if (%rejected_count_by_reason || %rejected_count_by_ip || %spam_count_by_ip);
+ }
+
+ print_league_table("mail rejection reason", \%rejected_count_by_reason, undef, undef, undef, $ws_rej, \$ws_rej_row) if %rejected_count_by_reason;
+
foreach ('Host','Domain','Email','Edomain') {
next unless $do_sender{$_};
- print_league_table("sending \l$_", $received_count{$_}, $received_data{$_},$received_data_gigs{$_});
+ print_league_table("sending \l$_", $received_count{$_}, undef, $received_data{$_},$received_data_gigs{$_}, $ws_top50, \$ws_top50_row);
}
- print_league_table("local sender", \%received_count_user,
- \%received_data_user,\%received_data_gigs_user) if ($local_league_table || $include_remote_users);
+ print_league_table("local sender", \%received_count_user, undef,
+ \%received_data_user,\%received_data_gigs_user, $ws_top50, \$ws_top50_row) if (($local_league_table || $include_remote_users) && %received_count_user);
foreach ('Host','Domain','Email','Edomain') {
next unless $do_sender{$_};
- print_league_table("\l$_ destination", $delivered_count{$_}, $delivered_data{$_},$delivered_data_gigs{$_});
+ print_league_table("\l$_ destination", $delivered_messages{$_}, $delivered_addresses{$_}, $delivered_data{$_},$delivered_data_gigs{$_}, $ws_top50, \$ws_top50_row);
}
- print_league_table("local destination", \%delivered_count_user,
- \%delivered_data_user,\%delivered_data_gigs_user) if ($local_league_table || $include_remote_users);
+ print_league_table("local destination", \%delivered_messages_user, \%delivered_addresses_user, \%delivered_data_user,\%delivered_data_gigs_user, $ws_top50, \$ws_top50_row) if (($local_league_table || $include_remote_users) && %delivered_messages_user);
+
+ print_league_table("rejected ip", \%rejected_count_by_ip, undef, undef, undef, $ws_rej, \$ws_rej_row) if %rejected_count_by_ip;
+ print_league_table("non-rejected spamming ip", \%spam_count_by_ip, undef, undef, undef, $ws_rej, \$ws_rej_row) if %spam_count_by_ip;
+
}
# Print the error statistics if required.
@@ -3679,7 +4065,7 @@
print $htm_fh "</body>\n</html>\n" if $htm_fh;
-$txt_fh->close if $txt_fh;
+$txt_fh->close if $txt_fh && ref $txt_fh;
$htm_fh->close if $htm_fh;
if ($xls_fh) {
@@ -3692,6 +4078,3 @@
# End of eximstats
-
-
-# FIXME: Doku
