ik wroteI'm not sure what you are looking for.
I started binding a osip to FPC but i didn't had the time (and will)
to finish the binding.
Sorry, i will clarify what i want to do.
-
My firm is writing software for telco-carriers.
VoIP Solutions, billing solutions, firmware etc.
We traditionally programmed in delphi, fpc, c and c++. I have a carrier grade sip/rtp stack written in fpc for linux and windows. Lib osip is nice but has neither the features nor the performance needed here.
Now i have to answer a tender (rfq/rfi). If i win my team and i have to port the applications to an embedded box. (powerpc/netbsd embedded system)
I am not willing to port our work to c++.
(We are currently try to kill all old c firmware projects as object pascal / fpc gives us a clear market advantage through cut down turnaround times and less error prone code.)
If i win the tender our sip stack will maybe run around 100.000 Voip/Gateway/CSTA boxes. They have to be supported with a timeframe for around 5-8 years.
So my questions are in detail:
What is the state of a netbsd-powerpc port ?
What is done ? What has to be done ?
If we can use fpc for development my firm can support the fpc project by supporting a/the netbsd-powerpc port. Also we plan to release some basic datastructures libraries into public domain.
Thank you, helmut
Show quoted text 26 lines
On Tue, May 20, 2008 at 2:20 PM, Helmut Hartl
wrote:
> Hi to all,
>
> I want to do some programming for a embedded system:
> (Port our SIP/RTP Stack to some embedded system)
>
> System is:
> ----
> NetBSD 3.0 #1570: Tue Aug 15 15:08:31 CEST 2006 TG Board total
> memory = 65536 KB avail memory = 46160 KB plb0 (root) cpu0 at plb0:
> 200MHz 405EP (Revision 9.80) ----
>
> This is a VoIP Box running under NetBSD with a PPC Architecture.
> (ibm405 core embedded in Xilinx Virtex {2-Pro, 4 FX} series FPGA)
>
> Can you give me some directions ?
>
> Our sip stack is commercial/closed source at the moment, but if i win
> the project, i can do some porting of libraries/targets if necessary
> and give that work back of course ...
>
> Thank you,
>
> helmut