splitting &preparse

Hello Terje, others,

Can you please look at the following:

I just did a few CVS commits to split &preparse into
two separate subroutines (now that it's called twice):

&preparse_meta: extracts 'charset' from <meta>
&preparse_doctype: extracts doctype and related stuff

I have then moved &preparse_meta close to other charset-
related code, and have simplified both functions to the
extent I was able to.

The goal of this is to isolate charset-related stuff,
in preparation for modularization.

There is still some more simplification possible in
both subs, but I think it's better if that is done
by somebody who is more familliar with HTML::Parser.

If this was a really bad idea, or if I missed some details,
please tell me.


Regards,    Martin.