Re: Sources of noise on test.winehq.org?

[Austin English]

On Thu, Apr 9, 2009 at 2:24 PM, Dan Kegel  wrote:
> The problem with the current arrangement is that when machines pop in and
> out, any failures that are more likely on those machines also pop in and
> out, so error counts fluctuate, which obscures the smaller changes due to
> wine fixes or regressions.
> I like the current output, and don't want to change it; I just want to add
> another view for people who want to judge progress.
>
> Setting up a view with a fixed set of machines would do, too.  Which is
> easier?

FWIW, Paul's windows machines and my test results for Wine (32 bit and
64 bit) should be pretty stable. You'd probably want to remove the
BSD/Solaris runs though.