blockgrep.pl

blockgrep.pl

#!/usr/bin/perl
use strict;

=head1 SYNOPSIS

    grepblock.pl <start_of_block_regexp> <match_regexp> file1 file2 ... fileN

    Input files are split in blocks. Each block is delimited by `start_of_block_regexp`.
    Output those blocks that match `match_regexp` (/match_regexp/ms is used).

=cut

use Pod::Usage;
use constant MAX_BLOCK_SIZE => 1048576;

if (@ARGV < 3) {
    pod2usage(1);
    exit 1;
}

my $start_of_block_regexp = shift;
my $match_regexp          = shift;

my $state = 'bob';
my $block = <>;
while (<>) {
    my $line = $_;
    if ($state ne 'bob' and $line =~ /$start_of_block_regexp/s) {
        $state = 'bob';
        print $block if $block =~ /$match_regexp/ms;
        $block = $line;
    } else {
        $state = 'scanning';
        $block .= $line;
        die "Current block is getting too big, giving up..." if length($block) > MAX_BLOCK_SIZE;
    }
}

print $block if ($state ne 'bob' and $block =~ /$match_regexp/ms);