In message <3C97D633.1010000@imperitek.com>,
`David Corbin <dcorbin@imperitek.com>' wrote:
> GOTOU Yuuzou wrote:
> > In message <3C967342.6040204@imperitek.com>,
> > `David Corbin <dcorbin@imperitek.com>' wrote:
> > > a) enables RequestDispatcher to support include/forward. There is
> > > probably much more than can be done here, like providing support for
> > > POST on the forward. I'm also sure that there are better ways to manage
> > > the response(s), but I think this will cover the 80% case.
> >
> > Hmm, RequestDispacher class seems to be overdone
> > considering its work.
>
> I'm not sure I understand what you mean by this. You mean more
> complicated than necessary?
Yes, and now I cannot judge that 80% is enough or not.
> Today, but it would be possible in the future to have RequestDispatcher
> be 'broader in scope', including forwards to other servers, back to the
> actual web server (if it's running in an "app-server mode), etc. Then
> RequestDispatcher could have the smarts to do this.
I know it helps especially JServlet users.
But it's not interesting for me to incline toward the
model which JServlet offers. I think WEBrick is just
a toolkit to handle HTTP but not a framework. I hope
it helps to build frameworks.
BTW, I checked in following changes:
* GenericServer#[] is added. You can refer the config parameters
through this method like "svr[:Port]".
* HTTPServer passes itsetf to get_instance() instead of @config.
And the first argument of HTTPServlet::AbstractServelet#initialize()
is a HTTPServer. It is set to @server. (@config remains for
backword comaptibility.)
* HTTPRequest#attributes is added.
--
gotoyuzo