=== chuck_ is now known as zul === robbiew1 is now known as robbiew === ogra is now known as Guest42038 === Guest42038 is now known as ogra_ === smspillaz|zzz is now known as smspillaz === jjohansen is now known as jj-afk [13:08] might I pester someone to process the overrides and review https://bugs.launchpad.net/ubuntu/+source/apt-clone/+bug/739808 today? I'd like to get a new ubiquity in to fix https://bugs.launchpad.net/ubuntu/natty/+source/ubiquity/+bug/738366 [13:08] Launchpad bug 739808 in apt-clone (Ubuntu) "[MIR] apt-clone (affects: 1) (heat: 8)" [Undecided,New] === jdstrand_ is now known as jdstrand [15:02] all, buildd turnaround for the next couple of hours may be a bit delayed as a fix for a security vulnerability is pushed through. Please contact me directly if this is causing you problems. [15:31] skaet: What's the difference between armel server and headless images? [15:36] ogra_, ^^ [15:38] ScottK, the list of packages that are available [15:38] headless only ships ubuntu-standard [15:38] OK. [15:38] they might turn into the base for server images in oneric [15:38] Right. [15:38] with pre-üpopulated apt cache or some such [15:40] ogra_: Would it make sense to swap the Kubuntu mx51 image that I'm currently volunteered to test for a headless image instead? Should be easier to test and more generally useful (sudo apt-get install kubuntu-desktop after install isn't very hard). [15:40] up to you, you own it :) [15:40] Should/Would [15:40] OK. I think it makes sense. [15:41] installing packages on headless definitely requires a working net connection [15:41] thats the only thing to regard [15:41] Ah. [15:41] and it defaults to serial all over for oem-config [15:41] That causes me to have second thoughts. [15:41] Let me consider it. [15:42] well, you can disable serial by dropping the preseed option from cmdline [15:42] and indeed oem-config offers netcfg ... but you still need a network to attach to [15:43] (on headless we also default to offering tasksel btw, so installing kubuntu-desktop should be very easy ... you could even preseed it) [15:43] ogra_: As it is now, if you install you don't have working video? [15:43] I'm trying to understand what defaulting to serial means. [15:43] you do, but oem-config does run on the serial tty by default [15:43] I see. [15:44] So if I drop that, it's like a server oem install, but fewer packages. [15:44] ? [15:44] we preseed debian-installer/framebuffer=false and set console=ttyO2,115200n8 [15:44] no packages [15:45] Not even the packages in ubuntu-standard? [15:45] they are installed but not shipped as debs [15:45] So it's a pre-installed image with ubuntu-standard, so after an install I have a bare ubuntu-standard system. [15:45] imagine a live image with only ubuntu-standard and defaulting to oem-config ... no archive or anything [15:46] yeah, right [15:46] So worst case with no network at install time, I get a network, reboot, and add stuff, right? [15:46] yes [15:46] you have a fully working cmdline system in any case [15:47] Why do you default oem-config to serial? [15:47] and we have tasksel so you could preselect kubuntu-desktop in a preseed cmdline option [15:47] because it is mainly designed as developer image [15:47] OK. [15:48] Would you be willing to help me get this set up? I can test it, but it's been ~a year and a half since I did anything with getting images set up and what little I knew about then is long forgotten. [15:49] I'm actually not that concerned about preseeding taskself for Kubuntu. [15:49] well, we need a flavour based hack for the image builder so your cmdline options are different [15:49] It would be just the standard headless less the serial preseed. [15:49] right [15:50] hmm, tricky [15:50] ENOCLUE on what needs to be done to the image builder. [15:51] well, looking at the code the imx51 images are 100% babbage board images [15:51] do you plan to use babbage ? [15:52] What I have is Efika MX Smarttop. [15:52] right [15:52] do you have a kernel and bootloader for that in the archive = [15:52] ?` [15:53] I think we do. [15:53] I've been procrastinating investigating this. [15:53] i know the linaor one doesnt work for one of the efikas [15:53] We have a kernel. [15:53] err [15:53] i think the netbook though [15:53] I'm told the linaro one should work now. [15:53] I think it's the netbook that doesn't work. [15:54] The smarttop, I think, works. [15:54] i was told yesterday that it only works on one of the models [15:54] yeah [15:54] i think someone said efikasb doesnt work [15:55] who was supposed to implement that on the builder [15:55] If you'd look into what it would take to get the headless image while I verify we have something that works on my hardware, I'd appreciate it. [15:55] Dunno. [15:55] it's several days of work to implement a new subarch [15:55] Sigh. [15:56] Is it a new subarch? [15:56] and especially harder if you dont have the HW to test on while developing [15:56] Right. [15:56] Let me go investigate a bit and see where we are. [15:56] we cant drop imx51 i think [15:56] so it would likely be a new subarch [15:57] (the builder needs to be able to roll images for older releases still so i wouldnt like to drop imx51) [15:57] Does it? [15:57] (not that we usually do that for armel) [15:58] I thought Karmic was the last mx51 release? [15:58] It's about to go away. [15:58] i thought it was lucid [15:58] * ScottK looks [15:59] lucid should be dove and imx51 plus an unsupported omap3 image [15:59] You're right. [15:59] nevermind [16:00] anyway, we can probably drop it if nobody complains ... its just not the usual habit [16:01] It would be nice to have something going forward. I think that's more important than the last 6 months of support on Lucid (since it's non-LTS) [16:01] but still, someone needs to implement it, the arm team is swamped in pre beta work atm :( [16:01] Let me see what I can figure out. [16:02] we should have identified someone from the cdimage team to implement it during UDS [16:02] lets note that for next time :) [16:03] I shouldn't have assumed someone had been. === jj-afk is now known as jjohansen [17:32] skaet: just a heads up, I'm likely going to need a beta freeze exception to do a no-change kernel upload to force a rebuild due to the recent gcc upload. [17:32] ogasawara, when do you expect the rebuild to happen? [17:33] ogasawara, thanks for the heads up, btw. ;) [17:34] skaet: gcc should finish building by tomorrow and I'll plan to upload the kernel after that. The arm build takes the longest for the kernel (~19hrs) so it should finish by fri morning your time. [17:35] skaet: but that's a day past beta freeze [17:35] ogasawara, thanks for letting me know. Should all work out ok. [17:35] skaet: I might need a beta freeze exception for gmusicbrowser (it's really small) for xubuntu (default music player), latest it would be ready would be midday tomorrow [17:36] micahg, betafreeze goes into effect at 2300 UTC. Let me know if you can't make that deadline tomorrow. [17:37] skaet: 2300UTC tomorrow? [17:37] yup. [17:37] skaet: shouldn't be a problem then, thanks :) [18:07] ogasawara: do you need to wait for gcc to finish on arm before uploading, in fact? Are there any out-of-tree modules that we need to worry about building for arm? [18:08] could I get an archive admin to poke at the linux-linaro-* kernels in binary NEW? [18:08] would like to get that metapackage uploaded today [18:09] slangasek: I was talking with tgardner and we really don't have any out of tree modules we really care about in Natty that isn't handled by a DKMS package. [18:10] ogasawara: well, AIUI dkms doesn't protect anyone from the issue with checking compiler versions, as that's in the upstream kernel build rules [18:10] I guess my question is, do we care if dkms works on armel [18:10] for beta anyway [18:14] slangasek: I don't think we really care, but since skaet said it should all work out ok to upload the kernel a day late, I'll just do it. [18:15] ok === bjf is now known as bjf[afk] === bjf[afk] is now known as bjf [19:50] ScottK: I'd like to get a FFe to upload the latest version of bikeshed into Natty universe; it adds small/useful shell script and manpage; the changes are visible at: http://bazaar.launchpad.net/~bikeshed/bikeshed/trunk/revision/84 [19:51] kirkland: If the only feature change is a new script (I consider the manpage a bug fix), please go ahead. [19:54] ScottK: that's all; thanks. === barry is now known as make-it-stop === make-it-stop is now known as barry === barry is now known as make-it-stop === make-it-stop is now known as yrrab === yrrab is now known as barry