>
> > In the quick_handler, returning HTTP_mumble will result in that status
> > being returned in a real request, but it will fallback to the real
> > handler in a subrequest.  Is that intentional?
>
> I don't know, but it sounds completely incorrect.  Bill, can you shed
> any light here?

The primary use of quick_handler is for a url keyed cache, so it must be able to handle
returning an HTTP_NOT_MODIFIED.  If mod_cache elects to handle the request and for some
reason the ap_pass_brigade fails (ie, quick_handler returns with something other than 
OK
or DECLINED) we need to handle that case as well. Conclusion... quick_handler needs be
able to throw HTTP_mumble status responses on mainline requests.

I don't think we have tested mod_cache enough with subrequests to know for certain 
what is
most useful. I am inclined to think that quick_handler on subrequests should always 
return
either OK or DECLINED.

Bill

Reply via email to