=== sturmflut_ is now known as sturmflut [14:12] infinity: hey, I just remember now. did you contact design about updating the slideshow? === doko_ is now known as doko [14:53] cyphermox: Nope. That needs doing. [14:55] infinity: k. I will contact people. [15:00] o/ [15:00] infinity: kickinz1 has done some testing and prepared a summary of our proposed docker 1.5 bump in bug 1430760 that I think answers your questions. Could you review please? [15:00] doko: ^^ any comments? [15:00] bug 1430760 in docker.io (Ubuntu) "[FFE] Bump up docker.io version to 1.5 in Vivid" [High,Triaged] https://launchpad.net/bugs/1430760 [15:02] cyphermox: you managed to get anywhere with the halt after removing install medium in vivid at all? [15:07] elfy: not yet :( [15:07] I did nail the oem-config though [15:07] okey doke [15:09] I'm getting back to getting plymouth to catch input right after my meetings [15:10] cyphermox: Ta. [15:10] rbasak: I can look after this meeting. [15:10] Thank you! [15:11] rbasak, are all build-deps in vivid? [15:12] doko, infinity: oh, good point, sorry. I forgot about that. [15:12] We had to sync two packages from Debian. [15:12] They'd be NEW in Vivid. [15:12] I will update the bug. [15:14] rbasak, gccgo-go [powerpc] drop this, not necessary anymore [15:15] doko: OK, no problem. I didn't know why that was needed - your changelog entry didn't say. [15:15] kickinz1: ^^ [15:15] Bug updated noting the two new build deps. [15:16] rbasak, OK, was updating bug also... [15:16] rbasak, that was working around issue in gccgo-5 on powerpc, not needed anymore [15:17] rbasak, any idea why arm64 fails to build? [15:17] doko: no, but I didn't pay much attention because I'm told it was broken previously, so there's no regression there. [15:18] I'm treating arm64 as a separate effort that I haven't taken on right now. [15:18] +ifeq ($(DEB_BUILD_ARCH),ppc64el) [15:18] + DOCKER_BUILD_TARGET = dyngccgo [15:18] +else [15:18] + DOCKER_BUILD_TARGET = dynbinary [15:18] +endif [15:18] + [15:18] this looks wrong ... [15:19] The patches add a new build target, so we use it on ppc64el. Why is it wrong? [15:19] should be ifneq (,$(filter $(DEB_HOST_ARCH), arm64 powerpc ppc64el)) [15:20] And armhf maybe? But it builds and works on armhf with golang-go it seems? [15:20] yes, unless you want to use gccgo on armhf [15:20] rbasak: yes, builds and seems to work on armhf [15:20] OK. I have no objection to changing it to your conditional then. [15:22] rbasak, do you have a build log for arm64? [15:23] doko, I'll get it [15:24] doko: OOI, why DEB_HOST_ARCH instead of DEB_BUILD_ARCH? [15:25] Shouldn't the choice of compiler be based on the target architecture? [15:25] cyphermox, That reference to "plymouth catching input" is that to to fix the issue with entering the full disk encryption passphrase? [15:26] BUILD is the arch you build on, HOST the arch you build for [15:27] rbasak, ^^^ [15:27] flexiondotorg: no [15:27] it's on shutdown [15:28] rbasak: HOST is the architecture that will be hosting the binaries, BUILD is the arch that's building them. And yes, we all know it's confusing, and we're a couple decades too late to fix it. :P [15:30] Ah. Thanks. I should have read the docs more carefully instead of assuming. [15:30] and don't confuse HOST and TARGET ;-P [15:35] doko: And then throw all the assumptions out the window when dealing with the kernel, where they renamed everything to make sense, but completely conflict with compiler tradition. [16:08] doko, sorry was in a meeting. [16:26] infinity: looks like update-manager is particularly unhappy with debconf passthrough failures from kernel postinsts [16:28] slangasek: Like, somehow more unhappy than it is with libc6? :P [16:28] slangasek: Or the same brand of unhappy? [16:28] infinity: I never had to run 'apt-get -f install' for libc6 [16:28] and I'm now getting very bizarre backtraces from u-m [16:28] slangasek: Oh, but that's just pure luck because libc6 debconfs in preinst. [16:29] actually no, because while I said postinst, looking closely I see that linux-image is also failing in preinst [16:30] Bleh. I guess I'll bump "figuring out HTF to debug that" further up the list. [16:33] slangasek, is that our breakage, or just breakage and we're the victim [16:33] apw: The latter. [16:33] * slangasek nods [16:35] I actually dreamt that I fixed the bug the other night, and woke up remembering just enough of the dream to realize that my subconscious is a very, very bad programmer. [16:35] "No, like, really, if you just insert a banana in this loop here, everything works." [16:40] heh [16:40] "bug in debconf passthrough; ate more fiber" [16:40] doko, rbasak: just updated bug logs with arm64 builds (both golang and gccgo). [16:41] kickinz1, arm64 golang? [16:42] slangasek: Too far. [16:42] * infinity drafts an HR complaint. [16:44] doko, I just tried, didn't check. there seems to be a golang package available for arm64 as it is all arch. [16:46] infinity, but did you even try the banana thing ? probably your subconscious isnt actually that far off, you just belive it is [16:46] ogra_: My new laptop doesn't have a banana port. [16:46] ah, damn ... technical hurdles .. yeah, thats bad then [16:52] kickinz1, this is an old build log, copied from the debian log archive ... please don't do that. rbasak, if you have a fixed package, please test-build it on vivid/arm64 [16:52] have to run now [16:54] doko, no, I did it tueday, maybe I uploaded the bad one... [17:10] infinity: because it's an Apple? [17:13] doko: are you looking at the right file? [17:13] doko: I see the build log kickinz1 produced. Eg: I: Finished running '/home/ubuntu/docker_test/repo/add'. [17:13] (hack for sbuild to pick up new build deps not in the archive yet) === kickinz1 is now known as kickinz1|afk [18:22] cyphermox: Hrm, did you really want "isolate" and not "start" in oem-config-firstboot? [18:22] cyphermox: The reading of isolate implies that it'll stop anything that graphical.target doesn't depend on. [18:23] cyphermox: Which would include, say, mysql, sshd, whatever other random things might have been running but not necessary to the target you asked for. [18:23] infinity: isolate in systemd means "change runlevel" [18:23] cyphermox: Also, how does this work for non-graphical systems (oem-config on a server)? [18:23] which before oem-config sounds fine, tbh [18:24] server appears to be landing in graphical.target too [18:24] Weird... [18:25] it's also not actually starting anything in oem-config unless you reboot into it [18:25] Mmkay. [18:25] from which point once it's done, it will go in graphical.target (could be any other target we want, really), to really do the rest of the startup [18:26] the point is to avoid starting getty@tty1, or display-manager, since they kind of break the oem-config [18:28] I love this brave new, and super-intuitive world. [18:28] ahaha [18:28] I agree with "world" [18:29] "I world as well!" [18:29] and new. [18:29] much less about super-intuitive. [19:18] slangasek: https://code.launchpad.net/~brian-murray/ubuntu-archive-tools/strike-thru-dupes/+merge/255131 [19:22] So I've got a package that is in proposed, but migration is blocked on a boottest error that seems to be a known bug. [19:23] How do I move that along? [19:23] tedg: which package and what bug? [19:23] Bug: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1421009 [19:23] Launchpad bug 1421009 in qtbase-opensource-src (Ubuntu) "unity8 sometimes hangs on boot" [Critical,In progress] [19:23] tedg: We just retry the test. [19:23] Package: ubuntu-themes [19:24] infinity, Is that something I can do? [19:24] tedg: I believe so, but not sure exactly how the permissions there work. [19:24] tedg: people with the necessary jenkins access can; I also don't know what the acl is [19:25] tedg: I retried it, but you can see if http://d-jenkins.ubuntu-ci:8080/view/Vivid/view/BootTest/job/vivid-boottest-ubuntu-themes/ has a "Build Now" link on the left for you after you've logged in (top right). [19:25] but anyway, yeah, for a boottest failing due to an intermittent bug, we want to retry it and see it *not* fail to be sure the known bug isn't masking unknown bugs [19:25] infinity, I can see it, but I don't have Build Now [19:26] tedg: And you're logged in? [19:26] Yup [19:26] tedg: In that case, the answer to the ACL question is "no". :P [19:26] It should be improved to make sure people with upload rights can also twiddle tests, but I think this is one of the many warts that aren't likely to get fixed before a move away from Jenkins. [19:26] addendum: "ask the QA team" [19:27] Well I don't have upload rights [19:27] Oh, in that case, nevermind. [19:27] Even if it was working the way I'd think it should, you wouldn't have access. [19:27] So the set may be correct [19:28] * tedg should really apply, but there are so many nice people to help [19:30] hello, I have two packages marked with "Regression" labels in the update excuses page (telephony-service and telepathy-ofono) [19:30] do I need to do anything regarding those two packages? [19:48] boiko: They succeeded on a retry. [20:05] infinity: ah nice, thanks === alexabreu is now known as alex-abreu|off [22:53] slangase`: https://code.launchpad.net/~brian-murray/ubuntu-archive-tools/strike-thru-dupes/+merge/255131 === slangase` is now known as slangasek [22:53] bdmurray: ack, looking [22:54] bdmurray: merged [22:54] thanks === salem_ is now known as _salem