[DGD] Writing an external network daemon

[Shentino]

Which reminds me...

Having the external daemon also handling inbound connections could expose
network APIs that could fiddle with settings that DGD itself doesn't expose
to the LPC layer.



On Sun, Jul 26, 2009 at 5:50 PM, Blake Arnold <BArnold at cartonservice.com>wrote:

> My 2c, on the actual transmission of the data across the connection would
> be something along the lines of,
> A similar process to the DGD phraser, which would phrase any user->server
> input into local_processed input and external_processed_input. It would then
> be conveniently packaged into a packet of some sorts, and transported across
> the connection at which point a de-package and computation of the process is
> completed and the return is then packaged back up and transmitted back. The
> packaging of the code simply represents a segmentation that is readily
> recognized as a group of code so that it can be segmented from any other
> jargon that may also be transmitted indirectly across the connection, such
> as any potential underlying protocol used for QOS, routing across multiple
> tiers of connections and ect. Ect. The later mentioned topics are actually
> theoretically of no value since the support is provided via existing
> underlying protocol.
>
> -----Original Message-----
> From: dgd-bounces at dworkin.nl [mailto:dgd-bounces at dworkin.nl] On Behalf Of
> Petter Nyström
> Sent: Sunday, July 26, 2009 10:34 AM
> To: All about Dworkin's Game Driver
> Subject: Re: [DGD] Writing an external network daemon
>
> Ah.
>
> Yes, that makes sense. Thanks for clarifying. However, I am
> experimenting some now, and I'm not sure what I should be able to
> expect from DGD's buffering. I am using binary connections, and If I
> send the string "123" back to the MUD via my external daemon, I most
> of the times receive the string "123". But sometimes I get three
> consecutive calls to receive_message() with "1", "2" and "3",
> respectively. Or any variation thereof. I guess I should be using the
> telnet protocol if I want a more predictable behaviour?
>
> Also not using any multiplexing, either.
>
> Regards,
>
> Jimorie
>
> On Sun, Jul 26, 2009 at 4:23 PM, Felix A. Croes<felix at dworkin.nl> wrote:
> > <jimorie at gmail.com> wrote:
> >
> >>[...]
> >> But you don't have any thoughts on how you would control the flow
> >> going back to receive_message() of the connected objects? (If you did,
> >> I apologize and ask that you repeat them in simpler terms. :P)
> >
> > Easiest would be not to do any multiplexing.  Let the external
> > network thingy open a new connection to DGD for every "outbound"
> > connection.  Then those connections could be binary or telnet mode,
> > with the appropriate buffering handled automatically by DGD.
> >
> > Regards,
> > Felix Croes
> > ___________________________________________
> > https://mail.dworkin.nl/mailman/listinfo/dgd
> >
> ___________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd
> ___________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd
>