[jflex-users] maximum number of rules

[Yuval O. <yu...@bl...>]

Hello,

I recently created a parser to extract domain names from host names, by
converting the public suffix list (http://publicsuffix.org) into a JFlex
file of about 4,000 rules. There are so many rules, because each country has
slightly different conventions (xxx.com vs. xxx.co.uk, etc). Running JFlex
on that file never finished, so I ended up having to partition it into
multiple JFlex files. It worked out in this case, as I was able to hash on
the last few characters of the domain name and pick the appropriate parser.

Is there a better general way to deal with that many rules, or is that a
limitation we just have to accept?

Thanks,
Yuval