safe regex variables

[oconnelm]

Hi,

I have a pretty ugly text-processing question. I'm trying to match a string denoted as a boundary by MIME headers then use it as a variable to annotate the different sections of a multipart message. This would be fine, except for the fact that the boundary strings conatin lots of regular expression quantifiers and tend to look like

EB-002(.BhNa.RALxMe_ZRHLIBVcyn_4sgz+G(v

I tried backslashing the variable name for the boundary string like:

if (/^-+\$boundary/){<br>

but this never matched the following instances, although it did let me use the string without a bad regex error. Just for kicks, I also tried using a backslash on the input, since it has the same problematic characters, but it doesn't get evaluated as a regular expression, so that unsurprisingly had no effect.

My question then is "Is there a way to safely (and efectively) use strings that contain special regex characters as patterns to be matched, without knowing ahead of time when, where and which special characters will occur in the pattern?"

Thanks,
Mike O'Connell

[japhy]

You want quotemeta() or \Q...\E:

$safe = quotemeta $unsafe;
$text =~ /$safe/;

# or

$text =~ /\Q$unsafe\E/;



japhy -- Perl and Regex Hacker

[oconnelm]

Aha! Exactly what I needed but couldn't find. Thanks very much. -Mike

[MeowChow]

Your question has already been answered, but I have the feeling that you'll save yourself a lot of time if you check out MIME::tools, or perhaps MIME::Lite, if you haven't already. As we love to say around here, don't reinvent the wheel, use CPAN.

   MeowChow                                   
               s aamecha.s a..a\u$&owag.print

[srawls]

Japhy already answered your question, but I thought I'd add that when you write this:

if (/^-+\$boundary/){

It means to match a literal $ sign, then the string 'boundary.' To do what you want, just do what japhy said and use:

if (/^-+\Q$boundary\E/){

(or you can use quotemeta on $boundary)

The 15 year old, freshman programmer,
Stephen Rawls