=== Andre_Gondim-afk is now known as Andre_Gondim === Andre_Gondim is now known as Andre_Gondim-afk === ogra_ is now known as ogra === davmor2 is now known as davmor3 === davmor3 is now known as davmor2 === davmor2 is now known as bolt_ === bolt_ is now known as davmor2 [11:58] * persia peers about [11:58] G'day [11:59] * ogra coughs and will try to follow in his sick head [12:00] * lool waves [12:01] * lool shivers and sneezes [12:01] go away! you infected me ... [12:01] come on, we can't send the whole mobile team to hospital! [12:01] OK. Enough preamble, let's get started. [12:01] #startmeeting [12:01] Meeting started at 07:01. The chair is persia. [12:01] Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [12:01] * davmor2 runs clamtk to kill the virus [12:02] Agenda is https://wiki.ubuntu.com/MobileTeam/Meeting/2009/20090326 [12:02] First up, action item revew [12:02] [topic] persia to investigate 338148 (co) [12:02] New Topic: persia to investigate 338148 (co) [12:03] So, I've looked a bit more, and I now understand the reason it doesn't work, but I've been distracted by installer issues, and haven't tracked down a fix yet. [12:03] No NCommander? [12:03] [topic] persia & NCommander to review workflow for testing alternate images. [12:03] New Topic: persia & NCommander to review workflow for testing alternate images. [12:03] GrueMaster, plars: Around? :-) [12:03] lool: yup [12:04] So,we went through it, and did a lot of testing, and found some fixes that ended up as bug #345534 [12:04] Launchpad bug 345534 in flash-kernel "Please add support for the imx51 ARM subarchitecture." [High,Confirmed] https://launchpad.net/bugs/345534 [12:04] Most of those were just merged with trunk, and testing should be available post-beta. [12:04] i think lool has patches for that one [12:05] For the flash-kernel bit? [12:05] yes [12:05] it was waiting on redboot-tools [12:05] Cool! [12:05] I have a partial patch [12:05] It misses the udeb parts and the fconfig part [12:06] fconfig is a two liner if we work with the bin [12:06] lool, Is it somewhere viewable? I'm happy to help track down the udeb part. [12:06] persia: Will attach it [12:06] Great. [12:06] Moving on... [12:06] if we dont, fconfig needs to grow an init function [12:06] [topic] NCommander to post dmesg for jax10 to 280669 (co) [12:06] New Topic: NCommander to post dmesg for jax10 to 280669 (co) [12:06] NCommander appears to be absent. [12:07] The dmesg appears to have been attached according to the bug log: I'll mark the action as done. [12:07] [topic] lool to spec ec2-package-builder for jaunty+1 (co) [12:07] New Topic: lool to spec ec2-package-builder for jaunty+1 (co) [12:07] no progress due to beta and I don't expect progress next week, but I keep it in my TODO [12:08] OK. I'll keep carrying it over then :) [12:08] [topic] ogra to trim selection-of-arm-images to a smaller scope (co) [12:08] New Topic: ogra to trim selection-of-arm-images to a smaller scope (co) [12:08] the build-babbage-image.sh script on http://cdimage.ubuntu.com/custom/20090326-armel+imx51/ will serve as a base for the tool here [12:09] i didnt update the spec though [12:09] so co [12:09] OK. [12:09] same goes for the next two points [12:09] Right. [12:10] OK. That completes action items. [12:10] Let's move on to Roadmap [12:10] [topic] offline-installer [12:10] New Topic: offline-installer [12:10] * ogra only has to say something to one point there ... (babbage status) [12:10] [topic] mobile-setup-wizard [12:10] New Topic: mobile-setup-wizard [12:10] all other items are as they were [12:11] No change since last week. [12:11] [topic] arm-library-optimisation [12:11] New Topic: arm-library-optimisation [12:11] Ready to be pushed now that glibc is in [12:11] Excellent! [12:11] * NCommander coughs [12:11] pango changes ready to be pushed; persia on the other changes [12:12] persia: Feel free to revisit the other topics involving Michael [12:12] [ACTION] persia to chase VFP for cairo, ffmpeg, gtk+2.0 [12:12] ACTION received: persia to chase VFP for cairo, ffmpeg, gtk+2.0 [12:12] lool, No need: both action items were completed from other evidence. [12:12] [topic] poulsbo-packaging [12:12] New Topic: poulsbo-packaging [12:12] I didn't look at this at all this week. [12:13] * persia skips general-resolution-for-touchscreen-handling [12:13] You were to look into poulsbo-packaging? [12:13] [topic] arm-softboot-loader [12:13] New Topic: arm-softboot-loader [12:13] No progress, carryover. [12:13] * persia skips selection-of-arm-images [12:13] [topic] lpia-versus-i386 [12:13] New Topic: lpia-versus-i386 [12:14] Nothing new to report [12:14] [topic] mobile-spec-cleanup [12:14] New Topic: mobile-spec-cleanup [12:14] Still being ignored. [12:15] [topic] bug #299847 [12:15] New Topic: bug #299847 [12:15] Launchpad bug 299847 in linux "Shared memory operations on very fast ARM hardware suffer from non-atomic operations and race conditions." [High,Triaged] https://launchpad.net/bugs/299847 [12:15] * ogra added two new bugs to roadmap, please reload [12:15] No progress [12:15] * persia plays browser caching games [12:16] [topic] bug #328167 [12:16] New Topic: bug #328167 [12:16] Launchpad bug 328167 in gnome-keyring "[arm] gnome-keyring-daemon eating 100% CPU at login in Jaunty" [High,Triaged] https://launchpad.net/bugs/328167 [12:16] post beta [12:16] [topic] bug #280699 [12:16] New Topic: bug #280699 [12:16] Launchpad bug 280699 in libcanberra "canberra-gtk-play crashed with SIGSEGV in malloc()" [Medium,Invalid] https://launchpad.net/bugs/280699 [12:16] Grr.... Not again! [12:16] [topic] bug #280669 [12:16] New Topic: bug #280669 [12:16] Launchpad bug 280669 in linux "DMA mode and driver jax10" [Low,Triaged] https://launchpad.net/bugs/280669 [12:17] carry over please. [12:18] I added a bug also just now [12:18] [topic] bug #338148 [12:18] New Topic: bug #338148 [12:18] Launchpad bug 338148 in vnc4 "Needs new version from Debian: fails to build with removal of mesa-swx11-source" [High,Triaged] https://launchpad.net/bugs/338148 [12:19] See progress in my action item [12:19] [topic] bug #348411 [12:19] New Topic: bug #348411 [12:19] Launchpad bug 348411 in partman-base "offers to install to source media in ubiquity if source media is SD (/dev/mmcblk0) card" [Medium,Fix committed] https://launchpad.net/bugs/348411 [12:20] fix committed? [12:20] nothing to discuss then? [12:20] Apparently. [12:21] there is one other bug with configuring apt persia and i both saw but neither of us reported or investigated yet [12:21] [topic] bug #348411 [12:21] Launchpad bug 348411 in partman-base "offers to install to source media in ubiquity if source media is SD (/dev/mmcblk0) card" [Medium,Fix committed] https://launchpad.net/bugs/348411 [12:21] New Topic: bug #348411 [12:22] We just had that one [12:22] * persia is confused. [12:22] yeah [12:23] Also the bug title doesn't match the bug number on the wiki page [12:23] Right. I found it, and will fix the Roadmap. [12:23] hmm, right, my fault [12:23] ogra: Fix the bug numbers or titles? [12:24] [topic] bug #348504 [12:24] New Topic: bug #348504 [12:24] Launchpad bug 348504 in linux "USB issues with linux-image-2.6.28-11-imx51 fill /var/log" [Undecided,New] https://launchpad.net/bugs/348504 [12:24] yeah, thts the one [12:24] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/348504 [12:24] Ubuntu bug 348504 in linux "USB issues with linux-image-2.6.28-11-imx51 fill /var/log" [Undecided,New] [12:24] Right [12:24] ogra: I am not sure why we need to track it in our meetings? [12:25] Kernel team said they didn't want to milestone this as it was dependent on a patch from FSL and they wouldn't fix it themselves [12:25] its a releasse blocker ? [12:25] And mobile is not going to fix the bug [12:25] It's a release blocker. [12:25] You can install to SD or to SATA and we can't do anything about it [12:25] SATA is USB [12:26] So documenting the workaround in release notes is the only thing to do I'm afraid [12:26] all target media is usb [12:26] I agree [12:26] ogra: Uh? [12:26] * NCommander didn't have this issue with the raw FSL 2.6.28 kernel ... [12:26] lool, the SATA is juts a usb bridge [12:26] I thought this SoC was special in that it was *not* using USB for SATA/PATA [12:26] lool, the bug makes the image hardlock after some minutes [12:27] ogra: So how do we fix it? [12:27] stopping syslog is the only way atm [12:27] acording to persia cooloney had a fix [12:27] lool, pata is not USB sata is [12:27] Or editing syslog.conf to drop all kernel messages. [12:27] davidm: hmm ok [12:27] ogra, No, not a fix. Was aware of the issue. [12:28] oh [12:28] then i misunderstood you [12:28] ogra: My understanding was that kernel team would either merge fixes for final or as a SRU [12:28] we cant release witjh it though [12:28] In all cases the mobile team can only document the workaround and that's done already [12:28] ogra: It's on the kernel team's plate to review potential fixes to it and they requested not to milestone it [12:29] well, we wont get any installer debugging without syslog [12:29] You want to discuss improvements to the workaround? [12:29] i dont see improvements [12:29] Let's set this aside. This doesn't seem like productive discussion. The bug isn't blocking me reviewing install issues. [12:30] Moving on. [12:30] Well I'd like to decide whether we're reviewing it [12:30] syslog, debug log and kern log get filled with 100s of MB [12:30] I don't see the point in having this bug every week [12:30] that makes oom kick in [12:30] lool, fair point. [12:30] So let's document how to setup syslog properly [12:30] Instead of just commenting out kern, you could drop kern.debug to /dev/null or something like that [12:31] and syslog ? [12:31] Add a kern.none? [12:31] hrm [12:31] So, who's going to document the improved workaround? [12:31] I'm not using USB for my installs happy to leave it to one of you [12:31] Let's keep this bug on our list, until we have a workaround that means it's only a kernel issue. [12:31] lool, huh ? what are you using ? [12:32] SD [12:32] how do you do that when testing the live image ? [12:32] I don't use the live image yet; I intend to add a partition after the live image for install [12:32] ah [12:33] makes sense ... but you will still get the errors as long as usb is active [12:34] lool, btw, as soon as bug 348411 is fixed your way wont work [12:34] Launchpad bug 348411 in partman-base "offers to install to source media in ubiquity if source media is SD (/dev/mmcblk0) card" [Medium,Fix committed] https://launchpad.net/bugs/348411 [12:34] persia, please reload I added more bugs [12:34] ogra, Are you up for documenting how to configure syslog to be quiet? [12:34] sigh [12:35] seems i have to [12:35] i have to redo the builder script as well for the ubiquity bug [12:35] before beta gets out [12:35] [action] ogra to document workaround for 348504 [12:35] ACTION received: ogra to document workaround for 348504 [12:36] and i'm massively sick [12:36] and have to test nslu2 which nobody else can apparently [12:36] lool, did you go visist ogra this week? ;-P [12:36] [topic] bug #348660 [12:36] New Topic: bug #348660 [12:36] Launchpad bug 348660 in ubiquity "ubiquity unsets ubiquity/install_bootloader=false at some point during installation" [High,Confirmed] https://launchpad.net/bugs/348660 [12:36] We had a phone call, that must be it [12:36] davidm, we're working together day and night since 10 days without break ... [12:37] So I put together a quick hack that hides this issue. [12:37] i guess the intense closeness is the cause :) [12:37] ogra, I know, just bad joke about physical visit [12:37] ogra: Give me more blanket [12:37] I've a roadmap to actually fix it. [12:37] lool, heh [12:38] an electronic human bug, who would have thunk ;-P [12:38] * persia bangs the gavel [12:38] persia: As long as u-i folks are happy with it, let's use it [12:38] lool, use which? [12:38] persia: I guess it shows we need flash-kernel ubiquity integration pretty quick afterwards though [12:38] We do need to fix this bug in the installed system [12:38] persia: Your quick hack [12:38] Or else it will continue to fill up the syslogs. [12:38] lool, The quick hack is already committed to ubiquity. The real fix is needed also. [12:38] NCommander: syslog.conf gets copied over [12:39] Moving on. [12:39] Ew [12:39] [topic] bug #347350 [12:39] New Topic: bug #347350 [12:39] Launchpad bug 347350 in oem-config "oem-config-firstboot does not fit in 1024x600" [High,Triaged] https://launchpad.net/bugs/347350 [12:39] This issue is long standing and known since before UDS and was discussed there: not sure what we can do now [12:39] This is one for StevenK to track and perhaps fix if possible it's quite nasty [12:40] Well, there's been massive improvements to ubiquity UI during the jaunty cycle. For some other installation cases, the ubiquity UI was moved to oem-setup, addressing the issue. [12:40] We picked up a slew of bugs in yesterdays testing round on the UNR [12:40] Except in cases where ubiquity doesn't fit, I suspect this ought be fairly straightforward. [12:40] yeah, with some of these resolution bugs you can workaround by alt-clicking and moving the window about, but not with this one [12:41] I'm personally pressing enter :-) [12:41] not in oem-config i think [12:41] since it runs fullsceen [12:41] lool: which works, so long as you don't want to change anything [12:41] I pressed enter all the way through OEM confg and made it to complete install [12:42] I was able to set timezone, enter my name password etc. just had to hit enter for each screen to go forward [12:42] davidm, That's for UNR, I presume? [12:42] Yes [12:42] davidm: You did? I can't press enter on the import from other OS screen, which I reported [12:42] All the last bugs I added are UNR [12:42] perhaps that was fixed in the mean time hmm [12:42] Right. MID still has an issue that it can't complete oem-install interactively because of kourou. [12:42] The problem is we are past UI freeze ... [12:43] StevenK: exactly [12:43] StevenK: That's why I fear we can't do much at this point [12:43] Not only is it risky to change the UI of all people installing ubuntu, but it breaks doc & screenshots :-/ [12:44] (and the fixes aren't trivial in my eyes either) [12:46] OK. We seem stalled. [12:46] davidm: So how important is it when compared to other issues? [12:46] So, Do we want to pursue the oem-config UI alignment with ubiquity at this point? [12:47] We need to release note it I think [12:47] this class of issues is only worth fixing if we fix all screens, the changes don't seem trivial and will require a lot of our cycles and a horde of exceptions [12:47] You can install it, the OEM team will have to fix this for sure for their builds, so we will have to get a fix in a some point. [12:48] persia: I saw the problem in ubiquity, I don't think oem versus ubiquity is the problem [12:48] We are now responsible for UNR fitting on screens or getting someone to fix them [12:48] It's our headach [12:48] Hmm theres' something which we could do for timezone which wouldn't be too insane: don't show the map widget when it doesn't fit [12:49] lool, "The same problem"? I've been running ubiquity installs at 800x600, and can see the buttons. This sounds like something more severe. [12:49] persia, skip Bug 346589 I just saw a post from cjwatson about it, [12:49] Launchpad bug 346589 in ubiquity "[Jaunty] Misleading information when installing with mounted partitions" [High,Fix committed] https://launchpad.net/bugs/346589 [12:49] persia: I ran it on my EeePC 701SD in UNR and saw this issue [12:49] it's in hand [12:50] The netbooks have smaller vertical res that removes the buttons from the screen completely [12:50] 576 I think [12:50] 600 - twp panels (= 48px) [12:50] *two [12:50] just enough to remove the buttons from the screen [12:50] Oh, Yeah. There's little enough space at 600 that I can imagine 576 just doesn't fit at all. [12:50] or is that the new glass ? [12:51] That is the new glass size [12:51] ah [12:51] VERY popular [12:51] but only one panel in UNR [12:51] persia: eeepc is 800×480 [12:51] so 576-24 [12:51] Yea [12:51] So 701 is even smaller [12:51] So it's 480 minus one panel on UNR [12:52] lool, Ah. Yeah, nothing we have fits in 480. [12:52] Anyway, it needs to be tracked and fixed at some point [12:52] UNR works well on the thing however [12:52] We don't have a hard goal at the moment, do we? [12:52] We asked for that at UDS, and got a *lot* of improvements, but it's not enough for 480. [12:52] definately not [12:52] I thought the hard goal was currently at 600 vertical pixels. [12:52] lool, we own the UNR now and the issues around it. [12:52] * ogra thought 576-panel [12:53] We could move it to 576, but that's less. 480 sounds too low. [12:53] We have have to at least deal with 576 panels [12:53] davidm: Let me more explicit: I have a 100x100 device, will UNR work on it? [12:53] heh [12:53] lool, no [12:53] davidm: What's the limit? [12:53] eee PC 701 [12:53] * persia notes there are only 7 minutes left, and lots to cover. [12:54] Lets move on for now [12:54] [topic] bug #344449 [12:54] New Topic: bug #344449 [12:54] Launchpad bug 344449 in netbook-remix "Installer puts up confusing message about partitions in use (dup-of: 346589)" [Undecided,New] https://launchpad.net/bugs/344449 [12:54] Launchpad bug 346589 in ubiquity "[Jaunty] Misleading information when installing with mounted partitions" [High,Fix committed] https://launchpad.net/bugs/346589 [12:54] So, the fix to make the message less confusing has been committed. [12:55] Yes just happened :-) [12:55] The underlying issue that one may or may not want to install on the install media is a bit murky. [12:55] there's still the fact that it appears at all, but I posted an extensive proposal to one of Matt's bugs [12:55] I think it will be tractable for final [12:55] cjwatson, thanks [12:55] you can't actually install on the installation medium; that won't be fixable [12:55] well, ish [12:56] if the installation medium has other partitions already set up then you can install on those as long as you don't change the partition table [12:56] cjwatson: Is it ok to install on a partition of that medium which isn't used by the installer? [12:56] ^- [12:56] Ok; thanks [12:56] That's actually what I want to do [12:56] bear in mind that's a historically awkward case [12:56] But others are installing on another medium [12:56] cjwatson: I was just curious and like that setup; but it's personal [12:56] because the partition table is locked, the kernel will refuse to reread it if you change any partitions at all in non-trivial ways [12:57] persia: Next bug? [12:57] [topic] bug #319825 [12:57] New Topic: bug #319825 [12:57] Launchpad bug 319825 in linux "acer_wmi in Jaunty on Aspire One exposes non-functional (always disabled) rfkill device" [Medium,In progress] https://launchpad.net/bugs/319825 [12:57] This looks like a kernel bug to me. [12:57] (cjwatson: Happy to read whether you think it's possible to fit ubiquity and oem-setup in 480 screens) [12:58] Again we need to track this [12:58] Do we need to discuss it each week? [12:58] make sure it's resolved [12:58] lool: 600 yes, 480 ambitious [13:00] davidm: ^ [13:00] although 600 still probably not trivial [13:00] cjwatson, is 576 within your opinion of 600? [13:00] 576 [13:00] davidm: I don't think it's worthwhile to track it in the mobile meetings; we could poke the kernel team weekly about it via email or in their meeting perhaps? [13:00] davidm: not sure [13:00] cjwatson, 600 works in ubiquity today, with the latest UI changes. [13:00] lool, OK on 319825 [13:00] davidm: There are already kernel bugs hotlists; I think that would be my preferred way [13:00] lool, OK by me [13:00] persia: oh good. oem-config is still awkward due to the annoying language screen [13:00] cjwatson, Right. [13:00] might be possible to make better use of horizontal space there though [13:00] doent work on 600 for me on the babbage without moving the window [13:00] *doesnt [13:00] ogra, Hrm. Dunno why you and I see different results. Let's compare after the meeting. [13:00] the buttons are hidden behind the bottom panel [13:00] persia, head for last bug please [13:00] i have a 1px line for each of them i can click on [13:00] [topic] bug #346078 [13:00] New Topic: bug #346078 [13:00] Launchpad bug 346078 in jockey "do not announce ath_pci on systems where ath5k is running" [Medium,In progress] https://launchpad.net/bugs/346078 [13:01] I had that happen to me, we need to track again unless someone has better idea [13:01] So, pitti seems to have this on-track. Shall we just coordinate with him, as with the kernel bug above? [13:01] Yes [13:01] That's on pitti's radar already [13:01] Right [13:01] RIght. Let's try to drop them from the roadmap for next week. [13:02] Personally, I think things like this are better added as discussion items for the meeting than roadmap items. [13:02] Anyway. [13:02] [topic] NSLU2 enablement [13:02] New Topic: NSLU2 enablement [13:02] let's skip that: image killed for beta [13:02] finishing the beta install here [13:02] Oh no sorry, I'm confused [13:02] OK. [13:02] I mixed it up with versatile [13:02] i'm at 97% [13:02] ogra, Nothing new to report? [13:03] persia, seems to have run fine up to here [13:03] [topic] Babbage enablement [13:03] New Topic: Babbage enablement [13:03] persia: Don't think we need to visit it every week anymore; ogra is doing beta testing using the already done enablement [13:03] and i dont plan to touch it agaiun if nobody reports breakage [13:03] persia: (I mean on NSLU2) [13:03] ogra, IF it's done, remove it from the roadmap :p [13:04] feel free [13:04] Babbage is a good topic though: NCommander what needs to be done for d-i support after beta still? [13:04] On ubiquity, we need a flash-kernel frontend IIUC what persia explained me [13:04] http://cdimage.ubuntu.com/custom/20090326-armel+imx51/ is there, i will add the fixes fdor syslog and ubiquity today [13:04] LINK received: http://cdimage.ubuntu.com/custom/20090326-armel+imx51/ is there, i will add the fixes fdor syslog and ubiquity today [13:04] lool, its pending cjwatson's merge [13:04] NCommander: It's done i think [13:04] NCommander, That's done. [13:04] which is committed [13:04] the dl page points to https://wiki.ubuntu.com/BabbageJauntyBetaInstall [13:04] NCommander: But what needs to be done from our side still? [13:04] I'll upload right after beta [13:04] There's also an outstanding choose-mirror branch to make armel use ports. [13:04] lool: there's still a flash-kernel task on that bug [13:05] persia: oh, there is? I'll hunt that down [13:05] which should get you going with a working install after i changed the two issues [13:05] cjwatson: Yeah, I have a flash-kernel update for the regular updates [13:05] But not for the udeb part [13:05] lool, then we need to hook up the cdimage code to spit out an image. The only thing that doesn't work is kernel installation due to the post-install on the kernel not doing anything AFAIK [13:05] And it doesn't fix the bug with initramfs size [13:05] cjwatson, I'll pass you the URL [13:05] and not for fconfig yet [13:05] (I assigned that task on the bug to you lool) [13:05] persia: it's ok, if it's on LP I'll find it easily [13:06] cjwatson: I'd like to ask you for help on the issue with flash-kernel and fconfig and bootscript with initramfs size after the meeting [13:06] NCommander: Oh ok [13:06] lool: ok [13:06] all in all i think we got a lot futher with babbage than lool and i belived last week ... [13:06] persia: Perhaps we can also discuss the -installer bit moving to a separate script [13:06] yeah [13:06] redboot-install or something [13:06] ogra, well, we have live images now, and alternate ones will be here soon AFAIK [13:07] lool, Doesn't need to happen. There's another way it can be done, which makes it easier. [13:07] persia: Hmm ok [13:07] See the #ubuntu-devel backscroll from about 75 minutes ago. [13:07] ok [13:07] Anything else on babbage enablement? [13:07] nope [13:07] There are some hardware and kernel issues, but not on our plate [13:07] Alright then. [13:07] [topic] ARM Benchmarking [13:07]