| Mirv | infinity: Riddell seemed ok to override the packages and I gave him the list, but then neither him or Scott had time to do that http://irclogs.ubuntu.com/2015/06/15/%23kubuntu-devel.html#t18:45 - so that's the current status | 04:44 |
|---|---|---|
| Mirv | I'm running Qt + KF5 + Plasma 5 from wily-proposed without issues | 04:51 |
| infinity | Mirv: Which packages are you trying to get to migrate? | 04:51 |
| infinity | Mirv: I won't override the KDE tests themselves, as that would cause the KDE packages to migrate, but I can force a pass on your packages if those tests are the only failing ones. | 04:52 |
| Mirv | infinity: Qt, since that affects non-Kubuntu too. full list is http://pastebin.ubuntu.com/11723237/ | 04:53 |
| Mirv | infinity: plasma-framework and frameworkintegration are part of the Qt transition since they use private Qt symbols. they are no-change rebuilds. | 04:54 |
| infinity | Mirv: Okay, some of those are waiting on broken boottests too. | 04:55 |
| Mirv | infinity: doh, fginther tried to fix qtmir-gles/qtubuntu-gles, they shouldn't be doing what they are doing (trying to install x86 packages on arm) | 04:55 |
| Mirv | he unblocked the previous uploads of those yesterday | 04:56 |
| infinity | *raise brow* | 04:56 |
| infinity | They're trying to do what now? :P | 04:56 |
| infinity | Mirv: Do you have a pointer to any public debugging of this oddity? | 04:57 |
| infinity | Mirv: Oh, lolz. I see. Those packages don't even EXIST on ARM, so the phone boottest is totally irrelevant. Check. | 04:58 |
| infinity | Mirv: Okay, let me try to unwind the rest of this and see what I can do. | 04:59 |
| Mirv | ok | 05:03 |
| Mirv | and yes, they are hard to install on the phone :) | 05:05 |
| infinity | Mirv: Yeah, I thought you meant it was trying to install the wrong arch's binaries, which would be super special, not that it was running tests for binaries that don't exist (which makes more sense). | 05:06 |
| infinity | Mirv: So, when I parse the mess to start looking, this is what I get. Care to verify the sanity of the list while I go looking at results for each to make sure this is reasonable? | 05:15 |
| infinity | Mirv: http://paste.ubuntu.com/11723303/ | 05:15 |
| infinity | Mirv: Hrm, on second thought, you're going to get stuck in the KDE transition anyway, due to frameworkintegration. If it needs to migrate. We'll see. | 05:23 |
| Mirv | infinity: the list looks sane. and yes, it seems even though plasma-framework is already in release pocket (the one before this rebuild), frameworkintegration in release pocket is the previous version. the old frameworkintegration depends on qtbase-abi-5-4-1 which would be going away with this Qt in favor of qtbase-abi-5-4-2 | 05:33 |
| Mirv | so frameworkintegration 5.10.0-0ubuntu2 is needed too | 05:34 |
| infinity | Mirv: Alright, that could get messy. We'll see. | 05:34 |
| infinity | Mirv: Maybe letting KDE migrate despite their broken tests will actually be easier, except it might be caught in the nettle transition. Not sure how much I feel like playing at midnight, but I'll look some. | 05:35 |
| Mirv | it might be that letting just frameworkintegration 5.10.0 through (and the plasma-framework 5.10.0 rebuild) would be enough. to me it'd look like packages depending on src:frameworkintegration would be fine with newer version too | 05:42 |
| Mirv | and frameworkintegration itself apparently is ok with older versions of kconfigwidgets and kiconthemes that are considered dependencies at update excuses | 05:44 |
| infinity | Mirv: If britney considers them deps, it's because of versioned deps... | 05:45 |
| infinity | Mirv: Or versioned build-deps. | 05:46 |
| infinity | Looks like it's the latter. | 05:46 |
| Mirv | ah.. it seems the dependency chain stops at kconfigwidgets and kiconthemes but then you'd be already in KDE territory | 05:47 |
| infinity | I'm okay with letthing those two pass and then letting the KDE folks sort the rest later. | 05:48 |
| infinity | But no guarantees that nettle wasn't in there somewhere too. I guess we'll find out. | 05:48 |
| Mirv | oh the same that was with CMake? uh.. | 05:49 |
| infinity | Mirv: Yeah, the nettle transition is still two fixes short of being done, and I was busy all day with kernel security madness. :/ | 05:49 |
| infinity | Mirv: Looks like you win, your packages migrated. | 06:22 |
| infinity | Mirv: But britney made an uninstallability trade to make it happen. | 06:22 |
| infinity | Newly uninstallable packages in testing: | 06:22 |
| infinity | * amd64: kwin, kwin-addons, kwin-dbg, kwin-dev, oem-config-kde, ubiquity-frontend-kde | 06:22 |
| infinity | Mirv: Any chance you could double-check that once the mirrors have caught up and see why? | 06:22 |
| infinity | Mirv: Ahh, looks like kwin depends on qtbase-abi-5-4-1 | 06:23 |
| infinity | Mirv: You missed it in your transition, I guess? | 06:24 |
| infinity | Mirv: Ahh, it got superseded, so your rebuilt one never made it. | 06:26 |
| infinity | Mirv: A fresh -0ubuntu3 build would be appreciated. I can do that. | 06:26 |
| infinity | Mirv: Done. | 06:28 |
| infinity | cjwatson: Really unconvinced about britney's uninst trades being a good thing. | 06:29 |
| Mirv | infinity: !! hugely thanks, and I'm almost surprised it was even possible | 06:38 |
| Mirv | infinity: but regarding kwin, it seems there was a never git snapshot that then was deleted, so what could be done since I already uploaded ubuntu2? | 06:38 |
| Mirv | infinity: oh... | 06:38 |
| Mirv | infinity: there was also deleted ubuntu2... | 06:39 |
| Mirv | infinity: ok, thanks for the ubuntu3, I "uploaded" ubuntu2 yesterday via train and somehow the train didn't complain because the archive ubuntu2 was deleted... | 06:39 |
| infinity | Mirv: It would have been rejected. The train might fail miserably at rejects, since the emails go to la-la-land. :/ | 06:42 |
| infinity | Mirv: Okay. kwin sorted. Which, I think, fixed a bunch of autopkgtests too. Including the ones you wanted to skip. :P | 08:14 |
| infinity | Mirv: So, there's a lesson not to ignore tests... | 08:14 |
| cjwatson | infinity: If you ever figure out how to make it not do that in a way that doesn't entirely kill performance, be my guest | 08:14 |
| infinity | cjwatson: Oh, I thought it might just be a toggle, and it would magically stop. No such luck, I guess? :P | 08:16 |
| infinity | cjwatson: Of course, I guess the ideal way to toggle it off is to get to 0 uninst, so it has nothing to swap. | 08:17 |
| cjwatson | infinity: britney keeps a count of uninstallables at various stages, and just compares the number. It'd have to keep track of much more involved data structures in order to avoid swaps. | 08:19 |
| cjwatson | infinity: Right, that's more or less been my conclusion when looking at this in the past. :-/ | 08:19 |
| infinity | cjwatson: Ahh, so it's an O(shit) problem. | 08:19 |
| Mirv | infinity: yes it seems the kdeplasma-addons was because of kwin and not like the other kf5 failures that had actual unit test failures | 08:29 |
| infinity | Mirv: rocs started passing too. | 08:30 |
| infinity | Mirv: Anyhow, whatever. Live and learn. Be more careful next time, blah blah. It's all happy now, looks like. | 08:31 |
| Mirv | infinity: correct, rocs was also like that, started failing on i386 only now while it used to pass before. interestingly I don't see the passing run in jenkins. | 08:33 |
| infinity | Mirv: Oh, huh. I don't either. That might be a weird bug. :P | 08:34 |
| flexiondotorg | Laney, infinity I've seen this https://lists.ubuntu.com/archives/ubuntu-release/2015-June/003285.html | 09:58 |
| flexiondotorg | Laney, infinity And had a chat with the Xubuntu team. | 09:58 |
| flexiondotorg | Laney, infinity I have no idea what some of those tasks actually entail. Can give me a brief idea of what assistance is required? | 09:59 |
| flexiondotorg | I've agreed to work with the Xubuntu team but if I can muster the help I would be interested in making an Alpha 1 and Alpha 2 release. | 09:59 |
| flexiondotorg | I just need to understand if I can actually commit the time. | 09:59 |
| Laney | Find out who wants to participate. Make sure they do their testing in time and update the ISO tracker. Send a release announcement. | 10:00 |
| Laney | (If it's me doing the other side then I'll want to be released by 18:00 UK time on the Thursday) | 10:02 |
| flexiondotorg | Laney, I can commit to doing the testing and updating the iso tracker in the time required. | 11:40 |
| flexiondotorg | Laney, Is that all is required from a flavour? Is there anything else that needs to be done? | 11:40 |
| Laney | flexiondotorg: You don't have to do it all yourself, just make sure that everyone is marked ready at the appropriate time. | 11:47 |
| Laney | I think that's all. | 11:47 |
| flexiondotorg | Laney, well I have a few people who help with the testing and know how the ISO tracker works. | 11:48 |
| flexiondotorg | Laney, So if Ubuntu MATE would like a 15.10 Alpha 1 do I just need to enter myself as QA Contact? | 11:48 |
| flexiondotorg | Laney, Basically, given that Ubuntu MATE is still fresh into Ubuntu and I am introducing MATE 1.10 this cycle I really need some test releases to feel confident I've not busted everything horribly. | 11:49 |
| Laney | What we need is someone to do the things I outlined, which are for *all* flavours participating in the milestone. | 11:50 |
| Laney | Flavours always do their own QA | 11:50 |
| flexiondotorg | Laney, right understood. | 11:55 |
| flexiondotorg | So this is an all or nothing situation? | 11:55 |
| Laney | I don't think it is much work | 11:56 |
| Laney | Mail ubuntu-release the week before to see who wants in, then on the Wednesday and Thursday morning see who is left to mark as ready and find out how far off they are. Then after the engineer pushes out the images you can send a release announcement to ubuntu-devel-announce. | 11:57 |
| Laney | But yes, we need someone to volunteer. Last cycle wxl gave a hand - he could probably brief/reassure you. | 11:58 |
| Laney | (Also note that stgraber has already volunteered for Alpha 1 so that is covered.) | 11:59 |
| === mzanetti is now known as mzanetti|run | ||
| === pgraner-dr is now known as pgraner | ||
| === mzanetti|run is now known as mzanetti | ||
| === pete-woods1 is now known as pete-woods | ||
| tedg | I can't figure out which seed results in this manifest: http://cdimage.ubuntu.com/ubuntu-core/releases/15.04/release/ubuntu-core-15.04-core-amd64.manifest | 13:39 |
| tedg | It seems like it should be obvious, but I am yet confused. | 13:39 |
| tedg | Can someone give me a couple pointers there? | 13:39 |
| davmor2 | tedg: ←→ are those any use? | 13:45 |
| * tedg tries dereferencing them | 13:45 | |
| slangasek | tedg: it's the ubuntu.vivid 'system-image' seed. The mapping of image names to seeds lies in the arcana of livecd-rootfs. | 14:53 |
| tedg | Ah, okay. So what is "core-libs" and "core-libs-devel" ? | 14:54 |
| ogra_ | slangasek, there doesnt seem to be a metapackage for this though | 14:55 |
| slangasek | tedg: "SDK", basically | 14:55 |
| ogra_ | (which makes changing anything in the PPA a bit tricky ... as we would usually do that manually in -meta in case we change something for a released release) | 14:56 |
| tedg | How can it be an SDK if those libs aren't included in the system image itself? | 14:56 |
| slangasek | tedg: I'm not sure that the core-libs* aren't completely obsolete | 14:57 |
| tedg | Okay, that would perhaps explain how I can't find relationships with them to other stuff :-) | 14:58 |
| tedg | (and a lot of my confusion) | 14:58 |
| tedg | So system image requires "boot" but not "required" or "minimal" | 15:01 |
| tedg | That's confusing. | 15:01 |
| arges | can another SRU memeber accept qemu in trusty? | 15:06 |
| bzoltan | tedg: is there any desktop application on the system image what would use the UITK? I remember about two years ago it was an issue that the whole Qt stack did not fit to the image... but by then the KUbuntu used 4.8 and we used 5.0. Now we are in sync. | 15:32 |
| tedg | bzoltan, Not on system image, there is now GUI stuff there. | 15:33 |
| tedg | bzoltan, Do you mean the touch image? | 15:33 |
| tedg | But, really, long term. I wouldn't expect that the UITK to be on any image, it should be included by the apps that use it. | 15:33 |
| bzoltan | tedg: yeps, i agree | 15:35 |
| wxl | Laney: flexiondotorg: happy to help. not sure i can dedicate my own time, unfortunately, as i'll be travelling. | 17:18 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!