Chris Mair wrote:
> talk about monitoring the buildfarm members...
> Can/should we do something about emu and guppy?
> Michael Meskes concluded that strtod on OpenBSD/x86 is broken,
> hence the persistent failures on make ecpg check for these two
I'm in active discussions on the issue with the OpenBSD people and they
are working on a fix though the issue is rather complex(compiler
optimisation issue and affects OpenBSD only on x86 and m68k - all other
archs are fine).
However it is highly unlikely that they will backport the fix to the old
branches (and maybe not even to 4.0 which is just around the corner).
Maybe we should either have a variant regression test file for those
openbsd-archs or a way to mark some failures in the buildfarm as
"recognized and understood" (maybe OK with a * or so).