On 24/08/14 23:40, Christoph Egger wrote: > In general -- post release. And as running kfreebsd-10 kernels with a > wheezy system ontop doesn't seem to be too good an idea. I'm surprised it doesn't cause problems already for stable updates and building security updates, even on Linux arches. The most reliable way I can think of is to run two sets of buildds, having oldstable+stable hosts until freeze, then stable+testing hosts during. Since physical machines have mostly become VMs the overhead of something like this ought to be much lower than it was. Wookey seems to be giving a talk on a similar topic at 19:00 your time. > I can of course do binary uploads at least on kfreebsd-amd64 (actually > doing that right now) -- a more automatic solution would b better of > course. Are there any more cases like that? Looks like you managed to binNMU this. I think that demonstrates it works now with mlock() on kfreebsd-10. libgnome-keyring maintainers might be persuaded to not FTBFS any more because of this test failure. I can't think of any other examples. If something like this happens (especially on a major port like Linux amd64) I think the packager would usually just try to work around it. Regards, -- Steven Chamberlain steven@pyro.eu.org
Attachment:
signature.asc
Description: OpenPGP digital signature