Re: Using DejaGnu in a regression farm

[Sean Cavanaugh]

On Monday 15 August 2005 17:57, you wrote:

>    Actually the original design conditions were to support a testing
> farm. Back at Cygnus we had a room full of embedded boards and weird
> unix machines, and would build and test GCC/GDB/Binutils in an automated
> fashion 24/7. 

That sounds quite interesting.  Thanks for the information, I'll definately be 
looking further into building this with DejaGnu then. 

> If worked pretty good, but once we had a few dozen
> systems, maintaining it became close to a full-time job.

We plan to have about half a dozen boards, so hopefully I can avoid having my 
job description changed on me. ;)

>    It would be good to add support for this in a more generic fashion. I
> imagine quite a few other people have similar testing farms up and
> running as well. For any application that already has DejaGnu style test
> suites, testing it remotely would be reasonably easy. Otherwise you'd
> need to write your own test suites.

In that case, I will try to coordinate our development so that at the end of 
the day we have a generic, customizable framework for automating the testing. 
I will also be glad contribute documentation both about the framework we'll be 
developing and also other aspects of DejaGnu (for instance I already have a 
short introduction to DejaGnu I'm writing for internal use that I'll share 
when it's done).

Thanks for the quick reply and the info!

--
Sean Cavanaugh
Firmware Design Engineer
PMC-Sierra Saskatoon