From: Ben Tilly <ben_tilly@...>
Date: 2001-02-24T03:28:00+09:00
Subject: [ruby-talk:11400] Suggestion: Slow RE mode

I would like to see an optional way to force all RE
matches to fail before they succeed.

This seems like an absurd request, but bear with me.

It is easy to write an RE that looks like, say,
/^(\s*yada\s*)*$/ which will match very quickly if
it can match, but take seemingly forever to fail.
Auditing code to verify that it has no dangerous
REs is very difficult.  However this is exactly the
kind of mistake you would like to be able to test
for.

So the idea would be that whenever you compile an RE
you would compile 2 of them.  The first is guaranteed
to fail, and is run just to show you how slow the
failure mode of the RE is.  The second is the original
RE.

So in the slow mode you would first run:

  /^(\s*yada\s*)*$(?:(?!a)a)/

and deliberately not set $~, $`, etc, and then you
would try normally with:

  /^(\s*yada\s*)*$/

Routinely try the slow mode in your testing suites, and
you can have some confidence that your program does not
have a disasterous RE waiting to trip you up.

Cheers,
Ben
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com