filtering out "bad" regular expressions from user input

[Alex]

> We allow users to input regular expressions to our search engine.
> Problem is, when (and they inevitably do, so it's not "if") they input
> a bad regular expression, it sends the backend search engine into a
> tizzy (backend is MySQL, frontend is Python).  For example, "^" will
> cause it to match everything, as will "^.*$", "" and ".*".

Can you test the regexp against a few randomly generated strings?  If it
matches against any of them, it's going to be pretty promiscuous.

I guess you could also perform the tests in a thread, and kill the
thread if it takes too long.  That would go some way to locating really
demanding regexps.

Alex.

-- 
Speak softly but carry a big carrot.