Failing windows test — some delays for 2.8

i blieve the windows test simply fails because it runs too slow. partest has a timeout of 14400 seconds. that's why the other day i wrote about the shootout benchmarks who got dramatically slower. if it's only partest who is slower, then that's not a blocker.
lukas

On Fri, Apr 30, 2010 at 15:32, Antonio Cunei <antonio.cunei@epfl.ch> wrote:

Our automated nightly tests on Windows are currently failing. That is probably due to the particular configuration of our testing environment, but we are doing additional testing in order to make sure there aren't any other issues. That will involve some delays, unfortunately, before we can safely release an RC2.
Toni

On Fri, Apr 30, 2010 at 04:34:43PM +0200, Lukas Rytz wrote:
> i blieve the windows test simply fails because it runs too slow.
> partest has a timeout of 14400 seconds. that's why the other day i
> wrote about the shootout benchmarks who got dramatically slower. if
> it's only partest who is slower, then that's not a blocker.

Yeah, sorry I hadn't gotten to that. I just gave it 2 more hours.

I'm assuming we're not treating the auxjvm failures as blockers. I feel
sure that is a jvm bug, and I would work around it if I could, but how
does one work around a bug one cannot specify and cannot test except by
observing the nightlies? At best "slowly and with protracted and tedious
effort" and it seems like a huge misallocation.

On Fri, Apr 30, 2010 at 4:44 PM, Paul Phillips <paulp@improving.org> wrote:

On Fri, Apr 30, 2010 at 04:34:43PM +0200, Lukas Rytz wrote:
> i blieve the windows test simply fails because it runs too slow.
> partest has a timeout of 14400 seconds. that's why the other day i
> wrote about the shootout benchmarks who got dramatically slower. if
> it's only partest who is slower, then that's not a blocker.

Yeah, sorry I hadn't gotten to that.  I just gave it 2 more hours.

I'm assuming we're not treating the auxjvm failures as blockers.  I feel
sure that is a jvm bug, and I would work around it if I could, but how
does one work around a bug one cannot specify and cannot test except by
observing the nightlies? At best "slowly and with protracted and tedious
effort" and it seems like a huge misallocation.

I agree. auxjvm is not a blocker. -- Martin