Recently came upon a similar block of code in a program which I've never had the need to fix in the last 11 years. That's the definition of legacy code, I guess.

I knew the guy who wrote the original. It had to be really frustrating to write all that to do a really simple thing. I know he and others were frustrated that perl was the only tool available to them and they didn't know it. But to take the time to write that, instead of investing that time in learning a much, much better way ... I just can't imagine that.

($hold1, $hold2, $hold3, $emailaddress) = (split /\|/,$linein)[0,6,7,8
+];
[download]

Ah yeah. Here's one of my own crimes. A quick&dirty way to allow for relative timestamps.

sub parseWebdate {
    my ($str) = @_;

    if($str =~ /^([+-])(\d+)(\w)/) {
        my %multipliers = (
            s   => 1,
            m   => 60,
            h   => 60*60,
            d   => 60*60*24,
            W   => 60*60*24*7,
            M   => 60*60*24*30,
            y   => 60*60*24*365,
        );
        my ($direction, $amount, $multiplier) = ($1, $2, $3);
        if(!defined($multipliers{$multiplier})) {
            croak("Undefined delta time multiplier $multiplier");
        }
        if($direction eq '-') {
            $amount = $amount * -1;
        }
        return time() + ($amount * $multipliers{$multiplier});
    }

    # normal time string
    return str2time($str);
}
[download]

When i wrote that, i told myself "It's only temporary, i'll clean it up next week". So far, it has been about 624 weeks on my Todo list and still going strong.

My tale of woe: Re: Re: There's Only One Way To Do It

Ah, the joys of legacy code. :) The "right" approach depends on how much code there is ... and what are the consequences if you break it. Millions of lines of crufty old business-critical code is the nightmare I've faced (and is faced daily by P5P ;-).

My mandatory list of references on this topic: Re: Strategies for maintenance of horrible code? (Legacy Code References)

s/Schultz/Schulz/g