Erlang and Aculab

[Per Bergqvist]

Hi,

> 
> call events still come in via a blocking ioctl()... :/
> 

Then you have two alternatives:
1) External port program
2) Multithreaded linked in driver with a separate wait thread

   handling the board communication.
> Did you have no problems with wild pointers in their code? Early
> versions of my interface sometimes broke with obscure errors hinting
> at the library shooting holes in the program. It got a lot better
when
> the interface was broken down to lots of processes with their own
> memory space.
> 

We had similiar problems, but as I said this was ages ago.

I have implemented erlang interfaces for Aculab, NMS and Dialogic
hardware.
In my opinion Dialogic (now Intel) was the best vendor but that was
absolutely not without flaws. It is possible to write a linked in
driver for dialogic (but one need to find the multiplex fd by
scanning).

My favourite nowadays for CT is Corelatus (http://www.corelatus.com)
and their GTH box. It is great stuff if you roll your own call control
(or use third party). 

/Per

>       Goodbyte, Gerd.
> -- 
> Gerd Flaig                     Technik               
gerd@REDACTED
> Bei Schlund + Partner AG       Brauerstraße 48      D-76135
Karlsruhe
>  Physics is like sex: sure, it may give some practical results,
>  but that's not why we do it. -- Richard Feynman
> 
=========================================================
Per Bergqvist
Synapse Systems AB
Phone: +46 709 686 685
Email: per@REDACTED