Update: After posting version 1 the problem of overlapping matches has been bugging me so much I've actually found a solution. Instead of using straight regex substitution I now do all the matching first, then apply the colors afterwards. I've tested it quite a bit but it's still experimental.
1. I would love for someone to adopt this and put it on CPAN so myself and others can get easy access to it#!/usr/bin/perl use strict; use warnings; use Term::ANSIColor; my @rules = @ARGV; while (my $line = <STDIN>) { print rewrite($line, @rules); } exit; # Process a single line sub rewrite { my $line = shift; my @rules = @_; my @marks = (); # Process each rule and find areas to mark while (@rules) { my $regex = shift @rules; my $color = shift @rules || 'bold yellow'; $color = color('reset').color($color); while ($line =~ /$regex/ig) { my $reset = undef; # Scan match area to find last color foreach my $i (reverse $-[0] .. $+[0]) { if (defined $marks[$i]) { $reset = $marks[$i] unless defined $reset; $marks[$i] = undef; # Cancel previous color } } # If necessary, keep scanning to beginning of line unless (defined $reset) { foreach my $i (reverse 0 .. $-[0]) { if (defined $marks[$i]) { $reset = $marks[$i]; last; } } } # Mark area $marks[$-[0]] = $color; $marks[$+[0]] = $reset || color('reset'); } } # Apply color codes to the string foreach my $i (reverse 0 .. $#marks) { substr($line, $i, 0, $marks[$i]) if defined $marks[$i]; } return $line; } =pod =head1 NAME em - console emphasis tool version 2 =head1 DESCRIPTION em is a command line tool for visually emphasizing text in log files e +tc. by colorizing the output matching regular expressions. =head1 SYNOPSIS em REGEX1 [COLOR1] [REGEX2 [COLOR2]] ... [REGEXn [COLORn]] =head1 USAGE REGEX is any regular expression recognized by Perl. For some shells this must be enclosed in double quotes ("") to prevent the shell from interpolating special characters like * or ?. COLOR is any ANSI color string accepted by Term::ANSIColor, such as 'green' or 'bold red'. Any number of REGEX-COLOR pairs may be specified. If the number of arg +uments is odd (i.e. no COLOR is specified for the last REGEX) em will use 'bo +ld yellow'. Overlapping rules are supported. For characters that match multiple ru +les, only the last rule will be applied. =head1 EXAMPLES In a system log, emphasize the words "error" and "ok": =over tail -f /var/log/messages | em error red ok green =back In a mail server log, show all email addresses between <> in white, su +ccesses in green: =over tail -f /var/log/maillog | em "(?<=\<)[\w\-\.]+?\@[\w\-\.]+?(?=\>)" "b +old white" "stored message|delivered ok" "bold green" =back In a web server log, show all URIs in yellow: =over tail -f /var/log/httpd/access_log | em "(?<=\"get).+?\s" =back =head1 BUGS AND LIMITATIONS Multi-line matching is not implemented. All regular expressions are matched without case sensitivity. =head1 AUTHOR Andreas Lund <floyd@atc.no> =head1 COPYRIGHT AND LICENSE Copyright 2009-2013 Andreas Lund <floyd@atc.no>. This program is free +software; you may redistribute it and/or modify it under the same terms as Perl +itself. =cut
Update: There is one other cool way to use this tool, and that's regex testing. Simply type "em" and the regex you want to test. Example:
em "0x[0-9a-f]+"
Now input your test strings one by one, and "em" will show you exactly what matches and what doesn't. Hit Ctrl+D (EOF) to exit.
In reply to "em" - Emphasize text using regular expressions by FloydATC
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |