[18:31] <xnox_> ubiquity is borked at the moment, fixes are in lp:ubiquity, but i can't release due to localechooser FTBFS with eglibc 2.10
[18:31] <xnox_> 2.19 that is
[18:31] <xnox_> bug 1283152
[18:31] <ubot2> Launchpad bug 1283152 in ubiquity (Ubuntu) "es_VE localedef is broken" [Undecided,New] https://launchpad.net/bugs/1283152
[19:53] <infinity> xnox_: That's fixed now, BTW.
[19:54] <xnox_> infinity: yeah, i see that you took es_VE fix. thanks.
[19:54] <xnox_> infinity: looking at glibc, don't we need to mimic more? https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=7447ccd98ee3944a95247ae23284dfac1de6c2aa
[19:54] <infinity> xnox_: I plan to move to glibc locales and drop langpack-locales entirely in the coming weeks.
[19:55] <infinity> xnox_: But for now, were any other locales actually failing to build?
[19:56] <infinity> langpack-locales should have had a testsuite that tried to build all of them...
[19:57] <xnox_> infinity: i'd be happy if localechooser builds, so anything that d-i supports... cause that seems to generate it's own locales....
[19:58] <infinity> Bah, I should have tried building all of them.
[19:59] <infinity> xnox_: Found a few more.  Iterating through all of them now to see what's busted. :/
[19:59] <infinity> ar_SD, az_AZ...
[20:00] <xnox_> infinity: localechooser works! i'll upload ubiquity to unbreak daily images =)
[20:01] <infinity> Hrm, kay, if I've onblocked you, maybe I shouldn't worry about fixing other locales, since I plan to ditch them anyway.
[20:01] <infinity> s/onblocked/unblocked/
[20:04] <xnox_> infinity: yeah.
[20:05] <xnox_> infinity: btw, i was not aware of langpack-locales at all... and i guess i shouldn't worry about trying to understand it either =)
[20:06] <infinity> xnox_: Yeah, it was a mostly failed experiment.
[20:06] <infinity> xnox_: The idea was to allow for rapid iteration of locale changes without uploading glibc, but langpack-locales, in practice, gets updated far less often than glibc.
[20:07] <infinity> xnox_: So, I'm going to take our slight tooling changes, jam them back into glibc, and get us in sync with Debian, ish.
[20:07] <infinity> xnox_: Which has the added bonus that we'll get the "locales-all" package, and can drop a fair few build-dep deltas.
[20:07] <xnox_> \o/
[21:01] <cjwatson> locales-all> yay.  Although locale-gen has an incompatible interface in Debian and Ubuntu, which will take some effort to resolve
[21:01] <cjwatson> incompatible both ways IIRC
[21:13] <infinity> cjwatson: Yeah, I know.  I'll fix that.
[21:14] <infinity> Well, by "fix", I might just keep our interface for now, and sort out how to bring them to parity later.
[21:14] <infinity> But we'll see.
[21:14] <cjwatson> If you could figure out how to make it a superset of both interfaces somehow, that'd be awesome
[21:14] <cjwatson> It's probably not impossible
[21:14] <infinity> No, it should be doable.
[21:15] <infinity> And both need work under the hood anyway.  We don't do atomic writes, for instance.
[21:15] <cjwatson> Yeah, I think I remember wincing at that
[21:15] <infinity> So, anything actually using a locale in the brief period where we regenerate could violently explode.
[21:16] <infinity> To be fair, this is because upstream localdef didn't use to make that particularly easy.
[21:16] <infinity> But Carlos added some switches recently to resolve that.
[21:20] <infinity> Anyhow, need to rescue the mess I made of the ppc64el builders before I go looking at glibc again.