rbb@covalent.net wrote:
>
> > > + if (strcmp(r->handler, "*.*")) {
> > > + return DECLINED;
> > > + }
> > > +
> >
> > As extensively discussed, this is _wrong_.
>
> Jon Travis pointed out to me yesterday that this is wrong in more ways
> than one. The problem is that before the change, Apache would run all the
> handlers that registered specific handler types, like DIR_MAGIC_TYPE or
> server-parsed-html before it ran any handler that registered generic
> handler types, ones with *'s in them.
>
> This behavior is now gone, which is a bad thing IMHO. This means that a
> module that matches multiple handler types, only gets one chance to
> matche them all. I didn't think this was important yesterday, but I woke
> up this morning, and I believe it is. Somebody tell me I'm wrong
> please. I just want somebody else to say I am wrong. :-)
No, you are right ... but ... read the @tip:
* @tip non-wildcard handlers should HOOK_MIDDLE, wildcard HOOK_LAST
Cheers,
Ben.
--
http://www.apache-ssl.org/ben.html
"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff
|