Error / warning count from error log can over-count

Lines 1590 to 1602 in 13ec507

    
           while ( my $logLi = <$fh> ) { 
        
               chomp $logLi; 
        
               $numLi++; 
        
               debugprint "$numLi: $logLi" 
        
                 if $logLi =~ /warning|error/i and $logLi !~ /Logging to/; 
        
               $nbErrLog++ 
        
                 if $logLi  =~ /error/i 
        
                 and $logLi !~ /(Logging to|\[Warning\].*ERROR_FOR_DIVISION_BY_ZERO)/; 
        
               $nbWarnLog++ if $logLi =~ /warning/i; 
        
               push @lastShutdowns, $logLi 
        
                 if $logLi =~ /Shutdown complete/ and $logLi !~ /Innodb/i; 
        
               push @lastStarts, $logLi if $logLi =~ /ready for connections/; 
        
           }

The regexes used to summarise the error log file can over-count e.g. a log line:
[Note] Error reading relay log event: slave SQL thread was killed
will get flagged as an error.

The 'level' of the events flagged in the error log appear to be (for MariaDB at least):

[ERROR]
[Warning]
[Note] (not used for analysis)

Updating the regexes to include the [ and ] will prevent false matches, and also remove the need for the match on line 1597.

I'm not sure whether the log formatting has changed over the years - leading to the simplified regexes currently in use?

	while ( my $logLi = <$fh> ) {
	chomp $logLi;
	$numLi++;
	debugprint "$numLi: $logLi"
	if $logLi =~ /warning\|error/i and $logLi !~ /Logging to/;
	$nbErrLog++
	if $logLi =~ /error/i
	and $logLi !~ /(Logging to\|\[Warning\].*ERROR_FOR_DIVISION_BY_ZERO)/;
	$nbWarnLog++ if $logLi =~ /warning/i;
	push @lastShutdowns, $logLi
	if $logLi =~ /Shutdown complete/ and $logLi !~ /Innodb/i;
	push @lastStarts, $logLi if $logLi =~ /ready for connections/;
	}