| cjwatson | ScottK: The transition tracker's a good deal easier to follow for this. http://people.canonical.com/~ubuntu-archive/transitions/libav10.html | 02:46 |
|---|---|---|
| ScottK | cjwatson: Can you tell why nepomuk-core shows up as bad? AFAICT, it's OK (and has been rebuilt) | 02:47 |
| ScottK | Thanks | 02:47 |
| cjwatson | ScottK: That's I think the one false positive on the tracker, yes. | 02:48 |
| ScottK | OK. | 02:48 |
| cjwatson | ScottK: It's because there's a binary in utopic that's no longer built in utopic-proposed - nepomuk-core-ffmpegextractor | 02:51 |
| cjwatson | ScottK: proposed-migration recognises that that will go away, but the tracker doesn't | 02:51 |
| ScottK | I see. | 02:52 |
| cjwatson | Everything else there makes sense. I have most of the dvdstyler fix. | 02:53 |
| ScottK | As far as I can tell between libav, cups, and gnutls there's no way for us to get an updated desktop in right now. | 02:53 |
| cjwatson | And I was going to stare at opencv/arm64 ASAP (and mpv/ppc64el once I can get the hw I have access to to talk to ports.ubuntu.com again). | 02:53 |
| ScottK | Fun. | 02:53 |
| * ScottK needs to go find a wayard $child, so see you later. | 02:54 | |
| cjwatson | Yeah, it's a gnarly transition. There'll be a couple of things at the end we can remove/demote. | 02:54 |
| cjwatson | But I want to save that until the rest is done. | 02:54 |
| infinity | cjwatson: opencv looks like it just needs an s/ifneq/ifeq/ logic inversion. | 03:59 |
| infinity | cjwatson: Are you already playing with it, or shall I? | 03:59 |
| infinity | cjwatson: Test-building on am1 now. | 04:07 |
| rsalveti | can we unblock latest pulseaudio I pushed a few minutes ago? changes are touch specific (new pulseaudio droid module, that works on top of the audio hal) | 04:35 |
| rsalveti | built finished fine and tests were also good, just blocked now due freeze | 04:36 |
| ScottK | rsalveti: That lands on all the images. | 04:57 |
| rsalveti | ScottK: I know, that's why I'm asking if it'd be possible to approve it | 04:57 |
| ScottK | Sigh. | 04:58 |
| ScottK | So we should respin all the images so your thing doesn't sit in proposed for a day and a half? | 04:58 |
| rsalveti | did we already spin all the images? (sorry if I'm not following the release as usual here) | 04:59 |
| rsalveti | that's why I asked here, fine if not approved as well | 04:59 |
| rsalveti | doesn't hurt asking though | 04:59 |
| ScottK | Yes. | 04:59 |
| ScottK | We have candidate images. | 04:59 |
| ScottK | Not too much tested yet, but they're there. | 04:59 |
| ScottK | rsalveti: Are you doing much with pulseaudio? I think we will need http://cgit.freedesktop.org/pulseaudio/pulseaudio/commit/?id=2b85ae048970b7faa7505fd0cd4746541d1b09eb (although not for Alpha 1) | 05:00 |
| rsalveti | that seems fine, want me to push that? | 05:01 |
| hyperair | N.B. The ykpersonalize command will by default include -oappend-cr. This will append CRLF (or typing ENTER) to the OTP, so that the dialog or web form is immediately submitted. If you do not want to have the form submitted after touching your Yubikey (e.g. a Nano sitting in your laptop), leave out the -oappend-cr switch. | 06:29 |
| hyperair | crap middle click | 06:29 |
| hyperair | whoops | 06:29 |
| shadeslayer | mmm | 08:32 |
| shadeslayer | Kubuntu still has 6 packages from 4.13.0 on the ISO | 08:32 |
| shadeslayer | stgraber: ^^ | 08:32 |
| cjwatson | infinity: not as yet, had gone to bed | 09:18 |
| infinity | cjwatson: Well, done now. | 09:19 |
| cjwatson | Thanks | 09:19 |
| xnox | vtk6 FTBFS on armhf is interesting - GL believes that GLdouble is double, yet Qt4 believes GLdouble is GLfloat. | 09:28 |
| cjwatson | https://wiki.ubuntu.com/ARM/FTBFS#OpenGL_and_Qt_combination and the one below | 09:29 |
| cjwatson | it's quite a common pattern | 09:29 |
| xnox | right. and also not a regression from upload that's in release already. bah. | 09:34 |
| infinity | cjwatson: That's the thing that was fixed in the infamous Qt5.2 ABI break, right? | 09:48 |
| cjwatson | AIUI | 09:48 |
| infinity | So, now we wait 5 years for everyone to port to Qt5, and we're saved. | 09:49 |
| shadeslayer | anyone have a clue what the heck "Back" is http://imgur.com/X5wBZGf | 09:49 |
| infinity | shadeslayer: If I had to guess, I'd say a bug. :) | 09:50 |
| shadeslayer | ^_^ | 09:50 |
| shadeslayer | infinity: what would I report that against btw? | 09:51 |
| infinity | gfxboot-something? cjwatson knows that bit better than I. | 09:52 |
| xnox | shadeslayer: infinity: well the bug is in new syslinux & the temporary fix we placed to get gfxboot working at all. I believe to many menus are now included by default hence "Back..." and "Help..." options which have not been there before. | 09:53 |
| shadeslayer | xnox: OEM mode is missing too | 09:53 |
| xnox | it's better that, than not having graphical boot at all. | 09:53 |
| shadeslayer | https://bugs.launchpad.net/ubuntu/+source/syslinux/+bug/1334189 | 09:53 |
| ubot93 | Launchpad bug 1334189 in syslinux (Ubuntu) "syslinux offers no OEM mode on Kubuntu Utopic Alpha 1" [Undecided,New] | 09:53 |
| cjwatson | gfxboot-theme-ubuntu | 09:53 |
| xnox | shadeslayer: i don't think it's a bug against syslinux though... | 09:53 |
| cjwatson | I did a quick fix last week, haven't yet had time to tidy up everything that's left | 09:54 |
| shadeslayer | ok, I've changed the affects | 09:54 |
| cjwatson | Is it critical for alpha 1? | 09:54 |
| infinity | I wouldn't imagine so. | 09:55 |
| infinity | Unless Kubuntu's really stepping up their Alpha 1 requirements. | 09:55 |
| xnox | shadeslayer: you should be able to still boot in OEM mode by manually typing the right cmdline arg... | 09:55 |
| shadeslayer | Hm, nope, should be fine, just better have it documented then | 09:55 |
| infinity | I think ours used to be "it boots, installs, and your computer doesn't set itself on fire and dive out the window". | 09:55 |
| shadeslayer | xnox: yeah | 09:55 |
| shadeslayer | infinity: words to live by | 09:55 |
| shadeslayer | :P | 09:55 |
| cjwatson | Well mostly I just wanted to know how rapidly I was dropping everything to investigate it :) | 09:56 |
| cjwatson | It's probably not too hard to fix but would be a respin-world | 09:56 |
| cjwatson | Now admittedly I'm sort of looking forward to finding out how fast we can do that now, but maybe shouldn't do it just 'cause :) | 09:57 |
| * shadeslayer points out that Kubuntu is even oversized at the moment | 09:57 | |
| cjwatson | Meh | 09:57 |
| cjwatson | The size limits are artificial now anyway :-/ | 09:57 |
| shadeslayer | cjwatson: hah, because of the new build infra ? :D | 09:57 |
| cjwatson | Right | 09:57 |
| shadeslayer | true :) | 09:57 |
| cjwatson | It's probably doable inside two hours, depending on buildd load | 09:58 |
| cjwatson | I'm working on the problems that cause occasional livefs failed-to-upload errors, BTW | 10:02 |
| cjwatson | They aren't new (they happen from time to time on package builds too) but the rate is high enough with livefses to be worth sorting out | 10:02 |
| infinity | cjwatson: Rate's pretty low with packages. Perhaps relating to filesizes/timeouts? | 10:04 |
| shadeslayer | xnox: https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1334199 | 10:05 |
| ubot93 | Launchpad bug 1334199 in ubiquity (Ubuntu) "ubiquity-kde fails to install Kubuntu in Spanish" [Undecided,New] | 10:05 |
| * shadeslayer tries with other languages | 10:05 | |
| cjwatson | infinity: Probably, yes. We've seen three so far with livefses, which is high considering the build count. | 10:06 |
| cjwatson | infinity: But it's easy for the librarian client to retry. | 10:06 |
| xnox | uploaded libavg, zoneminder, mediatomb. Also uploaded gnome-media-player the last package to transition from xine-lib -> xine-lib-1.2, such that xine-lib can be removed from the archive. | 10:27 |
| xnox | fixed visp is in debian new. | 10:27 |
| cjwatson | xnox: what's wrong with the visp no-change rebuild I just uploaded? :) | 10:32 |
| xnox | cjwatson: haha =) | 10:32 |
| xnox | if that works, then we are golden | 10:33 |
| xnox | what's left? performous, dvdstyler, mpv/ppc64el, jitsi, mplayer? | 10:33 |
| cjwatson | ah, rockne-06 can talk to ports.ubuntu.com now, so I might have a look at mpv | 10:33 |
| cjwatson | I have dvdstyler in progress, mostly done, just need a bit more time to finish it | 10:34 |
| infinity | cjwatson: porter-ppc64el no good? | 10:34 |
| cjwatson | infinity: didn't want to potentially pollute its one chroot with -proposed | 10:34 |
| infinity | cjwatson: Too late for that, all the porters have proposed enabled. | 10:34 |
| cjwatson | ah, well, shrug | 10:34 |
| infinity | But man, need to find a round tuit to mangle dd-schroot stuff for our porters. | 10:34 |
| xnox | Ampelbein: can you please merge jitsi? or should I look into it? | 10:38 |
| Ampelbein | xnox: Please do it, I sadly don't have time currently. | 10:39 |
| xnox | Ampelbein: ok, thanks. | 10:39 |
| cjwatson | infinity: Does https://launchpadlibrarian.net/178169679/buildlog_ubuntu-utopic-ppc64el.mpv_0.3.11-1_FAILEDTOBUILD.txt.gz imply that __APPLE_ALTIVEC__ is just broken on ppc64el? It's causing bool to be defined to an altivec thing that apparently doesn't support being passed to the ! operator, which seems kind of busted | 10:41 |
| ScottK | shadeslayer: The 4.13.2 stuff that isn't on the ISO is stuck behind other packages that there's no way to get migrated without violence to the archive. | 10:43 |
| ScottK | You've got all you're going to get for Alpha 1. | 10:43 |
| shadeslayer | ack | 10:43 |
| infinity | cjwatson: Without looking, I'd have assumed that it implies an endian issue in disguise, but disabling Altivec and making IBM fix it is usually the simple way forward. :P | 10:43 |
| shadeslayer | all proposed migration output said was it was due to the migration block | 10:43 |
| shadeslayer | so I don't know why it really didn't migrate | 10:44 |
| ScottK | It was all blocked in britney before the block was in place. | 10:44 |
| shadeslayer | ScottK: right, but why? :) | 10:45 |
| infinity | cjwatson: The following angry errors with "bool" and "int" in the same sentence sort of point to endian issues, since turning ints into bools and then flipping them backwards explodes spectacularly. | 10:45 |
| ScottK | shadeslayer: Mostly due to the libav transition, but other not easily fixable stuff. | 10:46 |
| xnox | cjwatson: i've yet to see altivec work on ppc64el. All codes that i've tried, even with patches from ibm folks didn't not compile, or didn't link, or had further ELFv2 dislikes. | 10:46 |
| infinity | xnox: We have lots of altivec working on ppc64el. | 10:46 |
| ScottK | Trust me, I fixed everything that was this week fixable yesterday. | 10:46 |
| shadeslayer | aha | 10:46 |
| cjwatson | infinity: wouldn't those be runtime errors? | 10:46 |
| infinity | cjwatson: Well, not the incompat type error. But yeah, the exclamation mark one is definitely a new one on me too. ;) | 10:47 |
| cjwatson | I don't *think* it's endianness, it basically seems to be using BYTE_ORDER | 10:48 |
| * infinity grabs the source. | 10:48 | |
| infinity | cjwatson: Where is this bool->altivec definition you speak of? | 10:49 |
| cjwatson | infinity: /usr/lib/gcc/powerpc64le-linux-gnu/4.8/include/altivec.h | 10:50 |
| xnox | infinity: hm, yeah looks like most things got fixed that i was stumbling upon. | 10:50 |
| xnox | e.g. libav is build with altivec now, et.al. | 10:50 |
| infinity | cjwatson: Oh! | 10:50 |
| cjwatson | At least I think that's what it is; can't see what else would be inventing __vector __bool int | 10:50 |
| shadeslayer | ScottK: do we still want to release the Active ISO's | 11:00 |
| cjwatson | infinity: -mno-altivec does make it build, so let me know if you think I should just do that ... | 11:06 |
| infinity | cjwatson: Go ahead. I'm trying to figure out the weirdness here with a simple testcase, but my simple testcase might be too simple. :P | 11:06 |
| xnox | Ampelbein: jitsi uploaded. | 11:07 |
| infinity | cjwatson: Obviously something nastier at play here than just the altivec/bool combo, since this works fine: http://paste.ubuntu.com/7699872/ | 11:09 |
| cjwatson | infinity: Right, maybe try with SDL | 11:09 |
| cjwatson | (uploaded) | 11:09 |
| infinity | cjwatson: Although, suspiciously, the disassembly of building that with or without altivec is the same. | 11:11 |
| * infinity scratches his head and thinks this isn't a rathole he should dive down right now. | 11:11 | |
| cjwatson | Oh, I see why vtk6 is showing up on the tracker. Disentangling. | 11:14 |
| cjwatson | (libvtk6 -> libvtk6.1 sub-transition, not cleared out of -proposed) | 11:18 |
| cjwatson | dvdstyler fixed for libav 10 et al. Judging from the Debian bug we'll probably end up demoting performous. | 12:09 |
| cjwatson | mplayer's been removed from Debian, and mplayer2 Provides: mplayer, but we'll need to do something about the dependencies on mencoder | 12:13 |
| infinity | cjwatson: A provides isn't good enough to force upgrades either. Does mplayer2 not give me a transitional package? | 12:16 |
| cjwatson | infinity: haven't checked, don't think so though | 12:18 |
| infinity | cjwatson: Right, we'll want to fix that, I guess. | 12:18 |
| cjwatson | I've taken to using mpv myself, but ... | 12:18 |
| infinity | I use mplayer. I wonder how much worse mplayer2 is. I guess I'll find out really soon. :/ | 12:19 |
| infinity | I seriously wish this ffmpeg/libav battle of egos could be resolved. | 12:19 |
| infinity | cjwatson: There's a d-i for you. I reviewed your minimal changes in bzr, and they seem fine to me, if you want to review the rest. | 12:21 |
| cjwatson | I assume the trusty stuff is mostly copy-and-pasted | 12:22 |
| infinity | cjwatson: Some 'find | xargs sed $saucy > $trusty' sort of action, yeah. | 12:23 |
| cjwatson | done | 12:23 |
| infinity | (The actual shell was hideous, we'll just pretend the above covers it) | 12:23 |
| xnox | bug #1334262 | 12:25 |
| ubot93 | bug 1334262 in xine-lib (Ubuntu) "RM: xine-lib superseeded by xine-lib-1.2" [High,Confirmed] https://launchpad.net/bugs/1334262 | 12:25 |
| infinity | cjwatson: Hrm. precise isn't building with PROPOSED=1 ... Is that because it's still broken, or because no one turned it on? | 12:25 |
| infinity | (Noted because I was trying to sort out if I should change the seeds or not, and I guess the answer is 'not') | 12:26 |
| cjwatson | infinity: It sure is | 12:26 |
| cjwatson | I turned it on yesterday | 12:26 |
| cjwatson | We might not have had a full build cycle yet | 12:26 |
| infinity | Err, wait. I re-learn this every point release, I think. It's in a config file, not in crontab, isn't it? :P | 12:27 |
| cjwatson | It's in cdimage/etc/config, yes | 12:27 |
| infinity | Yeah, just found it with greppery. | 12:27 |
| infinity | I don't know why I forget that every time. | 12:27 |
| cjwatson | Worked better there than in crontab, in practice, but it used to be in the crontab | 12:27 |
| infinity | cjwatson: Yeahp, makes more sense in a conffile, I'm just really crap at learning new tricks (says the man who still types -rfakeroot) | 12:28 |
| cjwatson | :) | 12:28 |
| infinity | cjwatson: Of course, I probably should have testbuilt that before I uploaded. Want to take bets on if it'll fail on image size constraints or something? | 12:31 |
| cjwatson | Pass :) | 12:31 |
| infinity | wget-udeb surely adds a bit too. I guess I'll play my favourite video game (watching build logs) and see. :P | 12:32 |
| cjwatson | One of these days maybe I'll convert build logs to longpoll | 12:34 |
| cjwatson | Although the "how many tabs can you have longpoll in" game might need to be fixed first ... | 12:35 |
| cjwatson | (Which is a bit hard for me) | 12:35 |
| ScottK | shadeslayer: No idea. | 13:01 |
| shadeslayer | ScottK: kdelibs uploaded with security fix for trusty | 13:05 |
| cjwatson | xine-lib removed | 13:19 |
| stgraber | shadeslayer: not sure what you want me to do ... | 13:33 |
| infinity | cjwatson: You should have taken the bet. | 13:52 |
| infinity | mcopy -i./tmp/trusty-netboot/boot.img ./tmp/trusty-netboot/initrd.gz ::initrd.gz | 13:52 |
| infinity | Disk full | 13:52 |
| infinity | And I should have test-built, obviously. *sigh* | 13:52 |
| cjwatson | libavg will be synced; performous should be demoted to -proposed once we're otherwise ready; vtk6 sorted; nepomuk-core false positive; so it's just sorting out mplayer now, I think | 13:54 |
| infinity | cjwatson: I'll fix and actually test-build this time. | 13:54 |
| cjwatson | ta | 13:54 |
| infinity | cjwatson: Hrm. Should we be concerned about the python traceback vomit from help-to-gfxboot.py in those build logs? | 14:17 |
| cjwatson | Slightly but not hugely. | 14:17 |
| cjwatson | Translation breakage, but it should tolerate it. | 14:18 |
| infinity | Oh, I guess it's been happening pretty much forever. | 14:18 |
| cjwatson | Yeah | 14:18 |
| === barry` is now known as barry_ | ||
| === barry_ is now known as barry | ||
| cjwatson | Ah, my NBS removals have caused vtk6 and wxsvg to show up across the board on the libav10 tracker now. They're both false positives for the same reason as nepomuk-core. | 14:32 |
| === zequence_ is now known as zequence | ||
| infinity | cjwatson: Fixed d-i in the queue. | 15:25 |
| infinity | -rw-rw-r-- 1 adconrad adconrad 1.7G Jun 25 09:24 debian-installer-images_20101020ubuntu136.19_amd64.tar.gz | 15:29 |
| infinity | That's getting impressively large. :/ | 15:29 |
| infinity | slangasek: Want to review that 2-line debian-installer fix in the precise queue? | 16:34 |
| cjwatson | infinity: I'll do it in a moment, sorry | 16:35 |
| infinity | cjwatson: Oh, thanks. Wasn't sure if you'd disappeared for the day. | 16:36 |
| cjwatson | No, just some calls | 16:36 |
| === oSoMoN_ is now known as oSoMoN | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!