[DGD] Changing connect() (network package)

[Shentino]

On Dec 11, 2007 5:43 PM, Shentino <shentino at gmail.com> wrote:
>
> On Dec 11, 2007 4:04 PM, Felix A. Croes <felix at dworkin.nl> wrote:
> > Shentino <shentino at gmail.com> wrote:
> >
> > > Developing an asynchronous model would probably be mere coding.
> > >
> > > If you would implement a way to call LPC asynchronously from C, that
> > > would probably help.
> > >
> > > Anyway, I should clarify that I'm not looking to IMPROVE the net
> > > package, I'm looking to REPLACE it with something BRAND NEW.
> > >
> > > My version of networking would stay COMPLETELY out of the way of DGD's
> > > normal activities.  I'd have my own special objects for connections
> > > and ports, and they'd play nicely with the rest of DGD simply by
> > > staying the hell out of the way.
> >
> > You mean, add a way to start a LPC execution round from a separate
> > thread in DGD, which is completely managed by the extension package?
> >
> > That would be more or less equivalent to having an external program
> > which communicates with DGD on a normal socket, and performs all the
> > extra functions.  You could wrap it up in an LPC layer, and even
> > have your own "connection" objects, indistinguishable by the rest
> > of the mudlib from genuine connection objects.
> >
> > I must admit that the idea of having a thread support dependency in
> > vanilla DGD does not have me jumping for joy. :)
> >
> > Regards,
> > Dworkin
> >
> > ___________________________________________
> > https://mail.dworkin.nl/mailman/listinfo/dgd
> >
>
> "Do it in LPC, do it in LPC, do it in LPC..."
>
> Man, you just HATE extensions don't you?
> :P
>

One thing I would like to ask though...

If you have a lot of connections comin gin and out, how is this
redirect going to take advantage of DGD/MP's parallellization?