[00:02] <doko> slangasek, accepted libverto, do you want to delay promotion? otoh krb5 won't be built with it
[00:40] <slangasek> doko: I have no reason for or against delaying promotion
[00:40] <infinity> slangasek: If we only have one "port" at this point, I'd argue that we should move both armhf and powerpc (for the britney stuff), and do away with the distinction.
[00:40] <infinity> slangasek: But yeah, maybe waiting on libreoffice to not vomit all over the report would be nice.
[00:41] <slangasek> well, even if powerpc is the only port, if there *is* a difference in who's responsible for it I think the distinction ought to be kept
[00:41] <slangasek> especially since powerpc is also seemingly never at 0
[00:42] <infinity> slangasek: ppc was at zero until very recently, to be fair.
[00:42] <slangasek> ah
[00:42] <infinity> slangasek: And, I dunno.  Having it alone on another report leads to it being ignored.  It's a bit of a catch-22.
[00:42] <infinity> "Oh look, no one fixes problems with it, let's hide them, so no one notices there are problems to fix."
[00:42]  * infinity shrugs.
[00:43] <slangasek> well, I don't feel strongly either way
[00:43] <infinity> This has nothing to do, of course, with the mirroring distinction, or the "official support" tag in LP, just for the britney reports, I think it's not worth the separation.
[00:43] <slangasek> the plusoners may have more relevant opinions here
[00:44] <infinity> I think it made more sense when -ports was more than half the arches we had.
[00:44] <infinity> Now it's.. One.
[00:44] <doko> slangasek, either the third buildd appears or we should drop powerpc. could you escalate the ticket?
[00:44] <infinity> (And we list it on, say, the ftbfs reports and most of our home-grown archive reports)
[00:45] <slangasek> doko: what ticket is this?
[00:46] <doko> slangasek, infinity knows the number, I need to search ...
[00:47] <infinity> 48569
[00:48] <slangasek> ta
[00:48] <slangasek> I won't necessarily escalate it before I know what else is in lamont's queue ahead of it, however
[00:53] <doko> slangasek, and here is bug #907126 to reduce the build load as well
[00:53] <ubot4> Launchpad bug 907126 in launchpad "allow exclusion of powerpc builds from PPA's (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/907126
[00:55] <infinity> That's only for non-virt anyway.
[00:55] <infinity> Which people are meant to be careful with.
[00:55] <infinity> (In theory)
[00:55] <doko> practice beats theory
[00:55] <infinity> Heh.
[00:56] <infinity> Still, the third machine would make things fine.
[00:56] <infinity> Given that PPC always catches up, just sometimes get a bit backlogged.
[00:56] <infinity> One more box should make it happy.
[00:56] <infinity> And once I finish the livefs-in-soyuz thing, we get sulfur as an lp-buildd too.
[00:56] <doko> yes, but it does have an impact on workflow. and things like linaro-overlays have priority on the buildds
[00:57] <infinity> I might also cut build time when I'm rooting around in lp-buildd too.
[00:57] <lamont> slangasek: that ticket needs stakeholder escalation if you want it anytime soon... 'twas #460 in the queue earlier today (and it's not just lamont's queue, thanks)
[00:58] <infinity> We always meant to do something with the "uninstalling build-deps failed" return code from sbuild, but we never did.  So, given that no one cares anyway, we may as well skip the uninstall step.
[00:58] <infinity> Which shaves a ton of time off every build.
[01:00] <doko> slangasek, still see libev in component-mismatches
[01:01] <slangasek> doko: I don't - reload?
[01:01] <doko> ohh, not anymore. libverto promoted
[01:03] <doko> have to get used to the 30min intervals
[01:07]  * doko watches Riddell's kde mess