Comments on: Readable Regular Expressions Revisited

By: William Wegerson

William Wegerson — Mon, 28 Dec 2009 23:19:02 +0000

As much as I love regex, as you do no doubt, sometimes the patterns get intense. Hence commenting as you pointed out can be a time saver in trying to decipher the intent of the pattern not only for someone else but ourselves as we go back to older patterns. Good article.

By: Siderite

Siderite — Mon, 21 Dec 2009 09:04:46 +0000

You might want to check my project http://regexconverter.codeplex.com/ that isomorphically transforms regex expressions into XML. It was my way of transforming regular expressions into something that is more humanly readable and, eventually, can be XSLTed into another regex.

I also think Benjamin’s solution is valid, although I could never make it work in a usable way. I would create subpatterns and use them in a regular expression like this:
var $int=@”-?\d+”;
var match=@”{$int/hour}-{$int/minute}-{$int/second}”. But it still isn’t modular enough for me.

By: Benjamin Smith

Benjamin Smith — Sun, 20 Dec 2009 00:29:35 +0000

I have a take: break up the string into variables that describe their intention. For example, parsing a log file might look like this:

$year = “([0-9]{4})”;
$month=”([0-9]{2})”;
$day=”([0-9]{2})”;
$message=”(.*)”;
$match=”/$year\.$month\.$day $message/”;

This produces something that’s easy to understand, and breaks up the expression into chunks that can be easily added/removed from $match so that diagnosing a pattern failure is easy.