#title #ubuntu-meeting Meeting Meeting started by NCommander at 14:59:22 UTC. The full logs are available at http://ubottu.com/meetingology/logs/ubuntu-meeting/2011/ubuntu-meeting.2011-09-22-14.59.log.html . == Meeting summary == ''LINK:'' https://wiki.ubuntu.com/ARM/Meeting/2011/20110922 (NCommander, 14:59:52) ''LINK:'' http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm.html (NCommander, 15:00:35) ''LINK:'' http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm-ubuntu-11.10-beta-2.html (NCommander, 15:00:46) ''LINK:'' http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm-ubuntu-11.10.html - I typed the wrong page in my browser (NCommander, 15:02:56) *ARM Server Status (NCommander, Daviey) *Kernel Status (cooloney, ppisati) *ARM Porting/FTBFS status (NCommander, janimo) *ARM Image Status (ogra, NCommander) *QA Status (GrueMaster, mahmoh) *Linaro Updates (rsalveti) *AOB Meeting ended at 15:57:28 UTC. == Votes == == Action items == * (none) == People present (lines said) == * ogra_ (106) * rsalveti (51) * janimo (47) * infinity (39) * NCommander (31) * GrueMaster (27) * davidm (20) * ppisati (8) * skaet (3) * meetingology (3) * mahmoh (1) == Full Log == 14:59:22 #startmeeting 14:59:22 Meeting started Thu Sep 22 14:59:22 2011 UTC. The chair is NCommander. Information about MeetBot at http://wiki.ubuntu.com/AlanBell/mootbot. 14:59:22 14:59:22 Available commands: #accept #accepted #action #agree #agreed #chair #commands #endmeeting #endvote #halp #help #idea #info #link #lurk #meetingname #meetingtopic #nick #progress #rejected #replay #restrictlogs #save #startmeeting #subtopic #topic #unchair #undo #unlurk #vote #voters #votesrequired 14:59:36 hrm, the bot ene upgraded 14:59:36 * davidm waves 14:59:49 seems like 14:59:52 https://wiki.ubuntu.com/ARM/Meeting/2011/20110922 15:00:18 we have no action items from last meeting so 15:00:35 #link http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm.html 15:00:46 #link http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm-ubuntu-11.10-beta-2.html 15:00:47 hello 15:01:00 i guess we need to start the postponing fun 15:02:05 hrm, the tracker is only showing me and infinity 15:02:10 ogra_: fix it :-P 15:02:20 ?? 15:02:31 i wonder where you look at 15:02:38 i see the whole team 15:02:40 oh 15:02:43 nm 15:02:56 http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm-ubuntu-11.10.html - I typed the wrong page in my browser 15:03:16 GrueMaster, your update-initramfs test can be set to done if you made a succwessfull install ;) 15:03:39 workitem wise, we're not too bad off. I have a few stranglers which will die soon (and some I should have marked DONE already) 15:04:12 infinity, do we plan to still see the apt ship pool whatever changes ? 15:04:29 looks to me like thats P material 15:04:40 ogra_: Yeah, I'd like to land them early next week, I think. 15:04:43 (pool works as is) 15:04:45 But I could defer. 15:04:46 hmm 15:05:00 its really late and we should all concentrate on bugs from now on 15:05:05 Pool works, but tasks are broken, which is pretty annoying for server. 15:05:09 and given that it works atm ... 15:05:12 I'd prefer to postpone none critical items until next cycle 15:05:17 right 15:05:37 Lets clean up, get RC solid and be happy 15:05:44 right 15:05:44 * infinity shrugs. 15:05:50 Happy to defer, though that is a bug. :P 15:06:01 [topic] ARM Server Status (NCommander, Daviey) 15:06:09 infinity, is it big enough to SRU later? 15:06:31 GrueMaster, postzpone the preinstalled stuff for jasper too, unless you want to actually test preinstall commands on kernel cmdline (that works since forever) 15:06:41 davidm: SRUing it's not really an option, IMO. 15:06:46 w.r.t. to omap guided partitioning, I *finally* got partman-uboot ina state where it does what it should. I need to cook off a new partman-reciepes, but that shouldn't be too hard. I'll give GrueMaster a modifiednetinstall to test today or tomorrow. 15:07:15 ok 15:07:26 NCommander, you still have three open items too 15:07:48 server kernel is on the schedule for P, i think you can postpone that 15:08:08 the other two look pretty well, like paperwork 15:08:13 yup, will do. 15:08:18 infinity, OK then not that critical I suspect 15:08:32 paperwork fortunately isn't affected by freezes 15:08:37 ppisati, the kernel config WI is also covered by a UDS session 15:08:40 i'm unsure if thats a :-) or a :-( 15:08:45 I'm going to mark iSCSI as done. The only remaining issue was booting from iSCSI and that is being worked on in the server team. 15:08:47 ppisati, so i'd say postpone that 15:09:07 ogra_: WI? 15:09:10 * ogra_ guesses that were all WIs that were left 15:09:28 ah 15:09:29 i must say we did pretty well given the sheer amount of crao we had after UDS :) 15:09:30 work item 15:09:34 right 15:09:41 ppisati, see the link above 15:09:49 you have one open item that can be postponed 15:09:57 NCommander, move ? :) 15:10:06 [topic] Kernel Status (cooloney, ppisati) 15:10:23 nothing exciting this week 15:10:24 * janimo working on getting ac100 new kernel packaged and uploaded 15:10:29 on my side 15:10:32 woohoo 15:10:39 i'm working on the sd removal bug 15:10:54 (and i found the problem) 15:11:01 wow, that was fast 15:11:08 janimo: Didn't you have an mx5 kernel incoming for me too? 15:11:25 infinity, jcrigby is handling that 15:11:28 yep, but i don't know yet how to properly fix it (because that would be a layering violation) 15:11:31 janimo: Shiny. 15:11:39 his git tree and processes, I had no idea how to make even such a small change 15:11:55 yeah, its tricky if you are not used to it 15:12:23 which is sad. One line change and a bugfix that can be expressed clearly in a single sentence (turn on CONFIG_LDBAF) needs a lot of legwork 15:12:35 and even if you know how to do it ... you will fall flat on your face if it doesnt work right :) 15:12:53 sincer these scripts are pure mystery 15:12:56 *since 15:13:13 well, some of the docs help, but not everything is documented 15:13:18 yeah 15:13:23 worse there is more than one way to do it 15:13:29 hehe 15:13:44 linux is all about choice ;) 15:13:55 NCommander, ? 15:13:56 yes, and it is refelcted by the awesome market share 15:14:30 ogra_: ? 15:14:45 * ogra_ puts NCommander on wheels so he starts moving 15:14:47 :) 15:15:08 [topic] ARM Porting/FTBFS status (NCommander, janimo) 15:15:39 gobject-introspection fails :( 15:15:42 The rebuild test is finally done, we should be combing through the results as quickly as we can. 15:15:43 I tentatively looked at postgres failing. 15:15:55 and postgres might be intresting for the server side 15:16:05 right 15:16:07 (looking at the ftbfs list) 15:16:16 I'd say it is essential? 15:16:25 beyond that, main looks as good as universe looks bad 15:17:28 we should all try to paticipate in the 5 a day project cjwatson started ... ate least with one or two packages per week or so 15:17:38 _*at 15:17:47 more if you can :) 15:17:57 [topic] ARM Image Status (ogra, NCommander) 15:18:09 2 packages per week is not participation in 5 a day project :) 15:18:21 janimo, more than nothing :) 15:18:28 that is true 15:18:28 images look okayish 15:18:45 they *should* be better at this point of the release though 15:18:53 desktop really fell behind 15:19:01 Need to reenable swap, at least on omap. 15:19:03 but overall they are fine 15:19:07 yeah 15:19:18 i think infinity committed to fix that properly yesterday 15:19:28 mx5 is a bit in a weird state 15:19:37 GrueMaster: Did I get a bug report? 15:19:40 needs kernel to test properly 15:19:42 we know it works but it didnt survive the switch to ext4 15:20:00 ogra_: I suspect mx5 will be Just Fine with the new kernel, but hard to tell without. 15:20:11 infinity: I was asleep when you responded, and still on my first cup of coffee, so...not yet. 15:20:15 infinity, i thought janimo did a local ext3 test 15:20:19 GrueMaster: Kay. :) 15:20:46 ogra_, GrueMaster did a test, and so did I 15:20:52 k 15:20:56 that should be enough 15:20:57 looked reasonable 15:21:15 we just need to convince skaet that it is ;) 15:22:21 * skaet is a bit worried about mx5 for good reason.... :P 15:22:32 Any chance of getting a new kernel and a new image today to make release? 15:22:45 skaet, its all fine, but we only discovered during milestone testing that ext4 support is missing 15:22:58 GrueMaster, no, builds to long 15:23:07 k 15:23:22 Broken kernel configs are easy to fix. I'm not terribly concerned. Yet. :) 15:23:28 skaet: The image was fine Monday prior to EXT4 switch (and ubiquity churn). 15:23:46 thanks GrueMaster. :) 15:23:53 skaet, we will have to skip B2 for it, but from monday lates iot shoudl be fine, its a "community supported" image anyway 15:24:03 *latest 15:24:16 save for archive skew :) 15:24:22 hehe 15:24:36 since the upload floodgates are open again for the desktop team :) 15:24:42 * ogra_ has nothing else for images ... if NCommander wants to move ... 15:24:59 [topic] QA Status (GrueMaster, mahmoh) 15:25:14 * GrueMaster wakes up to respond. 15:25:19 * janimo was conned into signing in to iso.qa 15:25:35 Release testing went...Well it happened. 15:26:17 I really do not feel as good about this release as I have in prior releases this late in the cycle, due to lack of dep testing. 15:26:24 *deep 15:27:19 agreed 15:27:25 My focus next week will be dedicated to desktop deep testing to see what I can squeeze out. We don't have a lot of time left to shake out any bugs, so stay in touch. 15:27:25 that needs fixing in P 15:28:07 Beyond that, I don't have anything else. mahmoh? 15:28:18 nothing here GrueMaster 15:29:11 moving on.... 15:30:31 * GrueMaster listens to the silence. Somewhere in the distance, a cricket chirps. 15:30:38 [topic] Linaro Updates (rsalveti) 15:30:43 hey 15:31:23 first is the request to help sponsoring the libjpeg-turbo package 15:31:49 package at revu already, and skaet agreed that it would be good to publish it at least to universe, if possible 15:32:04 during last week at the release meeting 15:32:17 What is it, exactly? 15:32:20 rsalveti: what is it and why do we want to uploadless than 15 days from release? 15:32:21 if anyone with time to review it, please :-) 15:32:59 An ABI-compatible accelerated replacement for libjpeg is how I'd read it from just the name. 15:33:15 rsalveti, speaking of which , there should be an organized linaro sponsoring process. As it is now the requests ar ead-hoc and left to the developers themselves, who often do not dare asking or know where to do it, and are content with the many ppas 15:33:16 libjpeg62 replacement, with simd support 15:33:18 NCommander, its universe ... there will likely still be plenty of packages :) 15:33:29 s/packages/new packages/ 15:33:45 neon and such, for those who want to try at arm devices 15:33:46 I'd say lets go for it if skaet is OK with it 15:33:54 indeed 15:33:55 skaet already gave OK 15:33:58 simd support is very good 15:34:06 rsalveti: Too late now, of course, but why aren't the changes in libjpeg itself? :P 15:34:07 was discussed during last release meeting 15:34:22 I have plenty of jpg images to put it through some paces. 15:34:22 infinity: long history 15:34:39 * skaet nods, its universe, so low risk. 15:34:39 default at firefox, chromium and fedora already 15:34:45 infinity, this is an upstream fork not a linaro one, linaro contributed though lately 15:34:45 GrueMaster: Porn is not a test framework. 15:34:56 yup 15:35:04 libjpeg-turbo is quite old already 15:35:07 aww. 15:35:27 janimo: yeah, planning to discuss that at UDS 15:35:37 ok 15:35:42 there was some discussion over email already, will see how it goes 15:36:13 also worked to get 4460 working with our images 15:36:23 support is quite broken atm, need kernel and u-boot fixes 15:36:30 * ogra_ wishes we could test that already :P 15:36:39 rsalveti, but no new packages, they still go to omap4 right? 15:36:46 do we want/need to make the support for the release? 15:36:54 janimo: no new package, just bugfixies 15:36:57 if possible sure 15:37:03 rsalveti, depends on the impact 15:37:10 Want, sure, need, no. 15:37:11 otherwise the borad will appear and there will be no ubuntu support for them till april 15:37:16 And yes, depending on impact. 15:37:31 For 4460? If we can get it in before release, it would be awesome. 15:37:32 kernel should be fine, as it's specific to a board revision 15:37:33 is this quad-core or merely dual? 15:37:33 below 200 LOC might be ok :) 15:37:36 u-boot should be tricky 15:37:45 or we can point folks at Linaro bootloads with instructions on how to fix image 15:37:54 janimo, still A9 15:37:57 but will get it working here first, and open the bugs for them 15:37:59 ogra_, even if obfuscated and 80 columns wide ascii art 200 LOC ? 15:38:13 then you can decide based on the amount of required changes 15:38:15 davidm: That gets ugly. I am already doing that for Maverick and Natty. 15:38:19 janimo, i dont care,i dont have to review it *g* 15:38:24 we can always point to linaro images, as they are usually more up to date, true 15:38:38 janimo, hmm ? 15:38:50 how can they be more up to dayte if they are a release behind ? 15:39:04 we're also creating oneiric based images now 15:39:10 and hope to switch to it next month 15:39:15 we'll see how it goes 15:39:20 they are based on a release behind but ahead with platform support stuff 15:39:23 anyway, that's most of ti 15:39:29 old evince, new kernel uboot X 15:39:32 we're preparing the release, so everybody is busy atm 15:39:45 I know I'd pick the second set if only that works on my board :) 15:40:02 depends what you want :) 15:40:11 ogra GrueMaster do you have 4460 boards yet 15:40:16 nope 15:40:24 not yet. 15:40:25 Yea, thats what I thought 15:40:27 nicolas wanted to ask around once again 15:40:41 i'll ask him tomorrow again :) 15:40:43 I got an email asking for shipping addresses which I provided 15:40:48 ah 15:40:53 but nothing so far then 15:40:55 Ooo, new toys? 15:41:05 same toys, new heart :) 15:41:06 Apparently not .... 15:41:10 infinity: you should join the ti call ;-) 15:41:14 yeah 15:41:24 if you like to get up in the middle of the night :) 15:41:31 rsalveti: That would involve one more meeting a week. 15:41:37 (And that) 15:41:40 first issue is that the led gpio is now connect at the mpu o_O 15:41:50 so once you boot the board, and the led is on, it'll basically turn itself off 15:42:02 fun ! 15:42:12 fixed already, but was fun :-) 15:42:47 anyway, I'm done 15:42:49 sorry about that 15:42:57 my laptop suddenly decided that it didn't have a wifi card 15:43:07 NCommander, we didnt notice 15:43:20 but yuo can move ;) 15:43:26 *you even 15:44:47 rsalveti, do linaro have beagle images too? 15:44:56 omap3 I mean 15:45:00 NCommander, moving on 15:45:07 [topic] AOB 15:45:12 janimo: yes 15:45:23 janimo: igep, overo, beagle, panda, and so on 15:45:39 no pandas this cycle for armel builds? We carry the timed out builds to P 15:45:52 ?? 15:46:00 we are already building on pandas 15:46:10 partitally at least ... until infinity needs them 15:46:14 one question I had, are we really planning to go just with armhf for next cycle? 15:46:17 half/half? 15:46:25 half half 15:46:28 just because it'll be a lts 15:46:30 janimo: We have a bunch. Though once armhf toolchain changes land, they'll get repurposed to armhf. 15:46:30 was always the plan 15:46:43 but was initially for this cycle 15:46:47 we have to pick which one is the default 15:46:52 rsalveti: "Hope" is a better word than "Plan" with al the delays we've had. :/ 15:46:53 and have only armfh for next cycpe 15:47:08 infinity: yeah =\ that's why I'm asking 15:47:09 no, you need at least one release where you have them in parallel 15:47:16 so was getting all the pandas in the build farm online. 15:47:17 infinity: what's the current status on that? 15:47:52 rsalveti: Need to land a GCC change, test it, re-bootstrap, and go. 15:48:03 * janimo wonders if giving back packages that fail after 10 hours hoping they land on a panda this time is a good use of build resources 15:48:08 a gcc change after b2 ? 15:48:08 Well, can probably skip most of re-bootstrapping with a compat symlink for a while. 15:48:09 ok, cool 15:48:19 ogra_: One that only affects a new architecture. 15:48:24 ah, k 15:48:46 janimo: No. If you want that to happen, talk to me, and we'll force it. 15:49:16 Much as I personally hate to say it, I think we will need to make both armel and armhf images for testing next cycle. 15:49:28 yes 15:49:29 infinity, good. Last time you said it is only possible if everythng is on manual, which is a hassle. I'll check the list to see what the latest status is 15:49:30 We already knew that. 15:49:35 but we will only release one of them 15:49:42 We have to test both until some arbitrary cutoff point where we choose one for release. 15:49:45 and will only test one of them :D 15:49:51 and are we supporting both as LTS? 15:49:52 janimo, nope 15:49:57 or just armel 15:50:01 janimo: It *is* only possible with everything on manual. :P 15:50:05 rsalveti, unlikely that we support both 15:50:06 janimo: But that's not rocket science either. 15:50:17 guess a good discussion for uds 15:50:31 we cant destabilize server 15:50:48 so hf needs to be rock solid before we switch 15:50:50 rsalveti: We won't really know by UDS. But we'll support the one that doesn't suck. 15:50:53 * janimo thinks of the churn in debian packaing to replace/amend armel ifdefs with armhf 15:51:12 hehe 15:51:16 janimo, the debian hf port should have a good bunch already 15:51:24 infinity, what are the suckage risks? Toolchain issues? 15:51:44 janimo: Honestly, my only concern is having it all built in a couple of months. 15:52:08 Whenever the armel buildds are idle, I intend to steal most of them for armhf to try to speed that process up. 15:52:20 We'll see how that works out. 15:53:16 rsalveti, we'll make at call at freature freeze which will be what we go out with for ARM 15:53:21 is buying setting up 20 or more boards to help such an obstacle? 15:53:34 once we choose, we will stop testing images on the other arch 15:53:44 davidm: fair enough 15:53:53 we seem to spend more in man hours trying to cater for that deficiency 15:54:12 janimo, getting pandas is VERY hard again 15:54:25 imx53 is a *lot* easier now 15:54:25 janimo: We have systems that aren't online yet. It isn't due only to a lack of boards. 15:54:29 and with sata 15:54:29 I'm hoping something else will pop up soonish 15:54:38 I'd happily take a bunch of mx53s. 15:54:40 GrueMaster, hence my 'setting up' in the question 15:54:43 rsalveti, well, it hads to proove stable first :) 15:54:46 *has 15:55:08 janimo: Manpower is also a big limiting factor. 15:55:11 indeed, lots of mx53s are good to, it;s not like we used dual cores so fa 15:55:12 r 15:55:13 with things like ext4 support missing in the kernel i wouldnt go for a buildd on it yet :) 15:55:14 well, at least you'll have support from linaro :-) 15:55:21 GrueMaster: Other than the master/slave in the DC being fiddled with for the "PPA cluster", we have others not online? 15:55:44 GrueMaster, right but seems a lot of manpower - albeit that of different men - is wasted to work around such issues, and to twddle thumbs while waiting 15:55:52 iirc, we have 5 in a sudo state. 15:56:05 We can discuss offline. 15:56:05 infinity, GrueMaster lets take this off-line and talk to the folks that know whats going on 15:56:11 * infinity nods. 15:56:29 GrueMaster, sudo state - the one in which a person gets you a sandwich without opposition? 15:56:34 like a trance you mean? 15:57:12 NCommander: Close? 15:57:21 closing in 3 15:57:23 2 15:57:25 1 15:57:28 #endmeeting Generated by MeetBot 0.1.5 (http://wiki.ubuntu.com/AlanBell/mootbot)