Re: testing 2010-10-04

[Stefan Schmidt <stefan@datenfreihafen.org>]

Hello.

On Tue, 2010-10-05 at 15:28, Koen Kooi wrote:
> On 05-10-10 14:32, Stefan Schmidt wrote:
> > Hello.
> > 
> > On Mon, 2010-10-04 at 14:05, Cliff Brake wrote:
> >> Testing run from last week:
> >> http://cgit.openembedded.org/cgit.cgi/openembedded/tag/?id=testing_2010-09-27
> >>
> >> testing-next is updated and ready for clean builds:
> >> http://wiki.openembedded.net/index.php/Testing
> > 
> > Openjdk-6 build fails on cacao:
> > http://tinderbox.openembedded.net/public/logs/task/8408323.txt
> 
> Openjdk fails in a lot of ways. Building it is a lottery. When changing
> nothing (host system and OE) I can do 2 builds from scratch and have one
> building openjdk and the other will fail. Or hang in icedtea compilation
> taking 0% cpu.

It was the first fail for all testing-next builds that have been done until now.

I agree that building it is not as stable as I would like it, but it is also not
as bad as described. The only problem i have been run into for a long time is a
segfault of the host javavm when bringing up other packages. Restarting the
build always gaves me a working image.

> So openjdk is not a good measure of passing judgement over a branch.

It failed in 1 out of 7 testing cycles for now. Not that bad imho. Surely it
gives no idea of the overall usability of a certain branch, but thats why we
have more images and distros to cover up more areas.

regards
Stefan Schmidt