Re: Pthread assumption + starting real port

[Johan Rydberg]

OKUJI Yoshinori wrote:
> 
> From: address@hidden (Niels Möller)
> Subject: Re: Pthread assumption + starting real port
> Date: 16 Nov 2000 13:19:13 +0100
> 
> > Have anybody looked into what IDL and stub generator should be used?
> > Is it doable to hack some L4 backend into mig, or is it better to use
> > something else?
> 
>   You may be interested in IDL4 which is used in SawMill. A paper on
> IDL4 is available from the LA-KA web site. I don't think it is a good
> paper, because they didn't present what algorithms were used in IDL4
> or the L4 version of Flick backend code at all (i.e. they just argued
> that IDL4 was better), but it is worth reading.


Maybe we could use Flick from Utah.

"Flick is not just optimizing: it is also extremely flexible. Flick 
 currently supports the CORBA, ONC RPC (Sun RPC), and MIG IDLs.
Interfaces 
 written in any of these languages can be implemented by CORBA-, ONC
RPC-, 
 or MIG-style C language ``stubs'' communicating via CORBA IIOP,
 ONC/TCP, Mach 3 ports, Trapeze, or Fluke IPC. Flick also generates
optimized
 CORBA C++ stubs that work with TAO, the real-time CORBA ORB. Finally,
because 
 Flick is a ``kit'' of components, it can be extended to support new
IDLs, m
 essage data formats, and transport mechanisms."

   http://www.cs.utah.edu/flux/flick/

-- 
Johan Rydberg, Net Insight AB, Sweden, +46-8-685 04 00

$ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") -
             THEN EXCUSE/OBJECT=ME