**** BEGIN LOGGING AT Tue Jan 25 02:59:57 2011 Jan 25 04:13:36 ohai lool Jan 25 04:13:37 er Jan 25 04:13:40 Lopi: Jan 25 04:14:12 hey lilstevie :D Jan 25 04:14:31 Lopi: you taken over iX by yourself now? Jan 25 04:14:58 lilstevie: yeah, been like that for awhile now Jan 25 04:15:19 ah, since my 3G blew its display driver i got left behind lol Jan 25 04:15:51 ah I see, yeah I'm not really sure what to do with the project anymore Jan 25 04:15:52 working on getting ubuntu on my android tablet now :p Jan 25 04:16:07 Lopi: once oib is ported to more devices :) Jan 25 04:16:33 Lopi: the SGX535 is a much more supported GPU so bring on the fun ;) Jan 25 04:17:03 lilstevie: yeah, I really want to play around with the ipad Jan 25 04:17:15 lilstevie: getting bored of my 3G tbh Jan 25 04:17:21 yerah Jan 25 04:17:23 yeah* Jan 25 04:17:28 i am bored of iDevices TBH Jan 25 04:17:47 I know what you mean, I want something open Jan 25 04:17:57 once oib works on the a4 i may get debian running on it though Jan 25 04:18:06 on my ATV* Jan 25 04:18:20 heh thats what i am loving with the galaxy tab at the moment Jan 25 04:18:37 the fact that without an exploit i can flash a custom kernel and fs Jan 25 04:19:27 I can only imagine how awesome that is ;p Jan 25 04:19:56 heh it is amazing Jan 25 04:20:07 and for being a candidate for ubuntu too Jan 25 04:20:16 1024*600 display resolution Jan 25 04:20:40 nice, going with unity interface then? Jan 25 04:21:25 I guess :p Jan 25 04:21:37 i need to get this damn kernel to compile Jan 25 04:21:55 the only bad thing with the tab is that the initramfs is compiled in to the kernel Jan 25 04:22:19 that kind of sucks ;p Jan 25 04:22:36 better then relying on an exploit though Jan 25 04:23:14 well the kernel is opensource as well :p Jan 25 04:24:43 the Nook Color pretty much has the same issues Jan 25 04:27:24 the most fun I have had with my 3G was making Mer work, packages are fairly limited on armv6 for Ubuntu Jan 25 04:32:03 lilstevie: Have any suggestions for something new to play with on my 3G? Jan 25 04:33:34 Lopi: none really Jan 25 04:33:52 lilstevie: bleh, that's what I was afraid of Jan 25 04:34:29 lilstevie: I suppose I'll release Ubuntu with LXDE then wait for the a4 ports to be finished :/ Jan 25 04:34:40 sad to say it but 3G has hit EOL Jan 25 04:34:57 thus all remaining armv6 devices are discontinued at apple Jan 25 04:36:57 some people in the iDroid community suggested making something more custom, but I don't see the point with the 3G hitting EOL Jan 25 04:39:18 I hate to let iX go inactive for so long, but I'm losing interest :/ Jan 25 07:24:24 hii Jan 25 07:24:58 can any know an efficient ARM emulator so that I can ARM based filesystem Jan 25 07:37:21 sanket, qemu is the only one I've used. Jan 25 07:40:41 blah Jan 25 07:40:49 the initramfs is corrupy Jan 25 07:40:51 corrupt* Jan 25 07:40:59 can u help me the installation process of qemu on ubuntu-linux... Jan 25 07:42:31 sanket, What host, and which guest? Jan 25 07:42:59 host is i386 and guest is ARM Jan 25 07:52:30 apt-get install qemu-kvm-extras should get you the emulator. From there, boot in your favourite kernel and rootfs. Jan 25 07:52:30 I believe that the current set only supports the versatile processor, although I know work has been ongoing for omap and a fast virtual variant. I don't happen to know the current status of those efforts as reflected in the Ubuntu packages. Jan 25 07:57:58 versatile goes well Jan 25 08:02:56 lilstevie, Doesn't it still need hacking to have more than 256MB memory? Jan 25 08:12:02 persia: i mean for qemu Jan 25 08:12:14 what device has less than that these days Jan 25 08:12:16 So do I. Jan 25 08:12:47 My understanding was that qemu only had 256MB emulating versatile, but could go up to 1GB emulating omap. Jan 25 08:12:54 I could be misinformed. Jan 25 08:12:58 ah Jan 25 08:13:05 i dont bother emulating more Jan 25 08:13:26 because an emu inside a vm is not fun Jan 25 09:56:45 I m getting an error while using qemu .... qemu: fatal: Trying to execute code outside RAM or ROM at 0x30008000 Jan 25 09:57:15 can anyone help me what is this error and how to solve it? Jan 25 11:24:51 i really wish rootstock would cache packages Jan 25 11:34:34 Why do you get like 70 console-kit-daemon processes? Jan 25 11:59:55 sveinse, are you looking at threads ? Jan 25 12:00:32 you normally dont get 70 of them but there can easily be 70 userland threads Jan 25 12:10:01 ogra, yes I think they are. They show up in pstree, but not in ps Jan 25 12:10:11 right Jan 25 12:10:20 thats fine Jan 25 12:10:29 70 processes would be odd Jan 25 12:10:34 It's the sheer number of processes/threads I'm curious of Jan 25 12:11:19 I seem to have a HW problem linked to the serial port, as getty ttyO2 keeps respawning. Could it be related to it? Jan 25 12:11:31 every process in userspace is wrapped in CK Jan 25 12:11:34 no Jan 25 12:11:52 that rather smeppls like an issue with upstart Jan 25 12:11:56 *smells Jan 25 12:12:10 what kind of kernel do you use ? Jan 25 12:12:18 yeah, ok. I'll keep my eyes open then Jan 25 12:12:21 there are certain features udev and upstart need enabled Jan 25 12:12:26 ogra, custom since its a custom board Jan 25 12:12:30 right Jan 25 12:13:05 dont ask me which features that are (dont know from the top of my head) but if certain features are missing udev and upstart wont work correctly Jan 25 12:13:20 one i remember is signalfd Jan 25 12:13:42 ogra, which... hahaha, you got ahead of me. I'll take a look the the versatile kernel or something and diff against it. Thanks Jan 25 12:14:00 better take the omap3 or 4 configs Jan 25 12:14:12 which one is that? Jan 25 12:14:37 from the linux-image-...-omap3/4 packages Jan 25 12:15:09 we ship configs in /boot/config- in the packages Jan 25 12:15:57 i am having a bit of fail getting bootup in qemu :/ Jan 25 12:16:31 versatile isnt really cared for since lucid iirc ... it wasnt touched by anyone ... and it is likely to go away in natty Jan 25 12:17:01 hm Jan 25 12:17:28 ogra, Why is it going away? Does qemu do omap well enough now? Jan 25 12:17:47 there is work going on to merge omap in, yeah Jan 25 12:18:02 ogra, yes. I'll take a look at linux-image-2.6.35-24-omap.. Think that's closest to our HW. (we're running 2.6.37, so there's some other diffing issues) Jan 25 12:18:05 My memory of UDS was that the qemu guys didn't expect to have something ready for natty, and wanted to stick with versatile, and then go for something else for natty+1, but I haven't been following progress. Jan 25 12:18:18 sveinse, natty has a .37 omap kernel Jan 25 12:18:37 so for now is it better to just do everything in chroot? Jan 25 12:18:44 persia, afaik slangasek is working on merging it into our branch now, ask janimo_ for details Jan 25 12:19:04 Oh, cool. I'm glad to hear the work made it. Jan 25 12:19:05 lilstevie, ?? Jan 25 12:19:24 new qemu should be a lot better Jan 25 12:19:28 and work with omap 3 kernel Jan 25 12:19:33 well i need to set some things up in the image, that i would usually do while it is running in the emu Jan 25 12:21:00 lilstevie, well, qemu (at least the arm side) nor the kernel should have changed much since maverick Jan 25 12:21:23 it kinda screws up on boot and just halts Jan 25 12:21:54 if that would be the case rootstock wouldnt work at all Jan 25 12:22:01 and afaik that works Jan 25 12:22:17 so it must be something in your rootfs thats messed up i'd assume Jan 25 12:22:59 but i havent changed a thing in the rootfs Jan 25 12:23:37 using the instructions on the wiki, versatilepb and cortex-a8 Jan 25 12:23:44 well, file a bug, attach the logs Jan 25 12:24:16 A little off-track question: I created my rootstock branch yesterday, and commited some proposals to it. I planned on submitting a merge request, however I discovered a rather stupid error in the commit message. Would it be important to fix the commit message prior to merge, or doesn't it matter? Jan 25 12:24:34 doesnt really matter Jan 25 12:24:46 i make typos all the time in mine :P Jan 25 12:25:30 * ogra is famous for typing teh instaed of the Jan 25 12:25:40 * persia notes that bzr supports uncommit in case it's really something one is ashamed to have posted Jan 25 12:25:50 heh, and apparently for mistyping instead Jan 25 12:26:28 Well, that's typo. I wrote "added --apt-clean option to rootstock" when the real option is --apt-update. Somewhat more significant error IMHO Jan 25 12:26:48 well, as persia said ... you can uncommit Jan 25 12:27:07 thanks, I'll do that for the correctness of things Jan 25 12:27:12 yup, uncommit, commit again and force the push Jan 25 12:27:16 if you have pushed already you need to use --overwrite for the push command Jan 25 12:29:42 Heh. That was far easier than expected Jan 25 12:30:05 Uncommit in subversion isn't exactly trivial Jan 25 12:30:19 thats why we have bzr ;) Jan 25 12:34:09 persia, pm215 from linaro is workiing on selecting omap bits from qemu-maemo and this will be packages with slangasek's help into a new source Jan 25 12:34:36 which will rpvide binari packages for the current non-x86 binariesm leaveing current qemu-kvm x86 ony and be used by server team Jan 25 12:34:49 Excellent. last I heard was in October, when it was thought to not be possible in time for natty. Jan 25 12:34:57 In this case I'm *really* glad to be wrong. Jan 25 12:35:15 persia, indeed for natty it is still not possible to have omap in upstream - which would be ideal - so a temporary solution is anew source package Jan 25 12:35:17 pm215 is doing a quite good work with qemu Jan 25 12:35:31 lots of old bugs fixed already Jan 25 12:35:45 so we should have a more stable one for natty Jan 25 12:35:51 persia, but the last details were still beiong discussed at the rally so no wonder the october info is out of date Jan 25 12:36:16 new source package? Ugh. Oh well. I suppose, if that's considered less effort than maintaining versatile another cycle (not that versatile is all that useful, really) Jan 25 12:36:29 rsalveti, indeed he gets the fixes in very fast, the bottleneck and the boring part is getting the packaging (and possibly backports) right Jan 25 12:37:09 persia, versatile annoys kernel people as it adds extra build time Jan 25 12:37:29 and it would be better to have qemu support one of our official images - I am not sure versatile runs those? Jan 25 12:37:38 Nope. Jan 25 12:37:42 ok. Now I've created my very first merge request. I presume it's either orga or rsalveti who will review this. Jan 25 12:37:53 so omap3 as omap4 is not yet nearly functional in qemu Jan 25 12:39:15 sveinse: yup, thanks for creating it, will take a look later Jan 25 12:39:38 sure, np. Just wanted to see if I got things right Jan 25 13:09:26 is a rename to unity-2d planned? Jan 25 13:09:40 or a new metapackage of that name added? Jan 25 13:09:40 janimo_: hm, why? Jan 25 13:09:50 janimo_, yes, i'm working on that Jan 25 13:09:59 rsalveti, I seem to remember ogra mentioning it a while ago, and for clearer naming Jan 25 13:10:03 we have unity already Jan 25 13:10:06 (merging unity-2d-default-settings into the unity-2d source) Jan 25 13:10:18 having the default-settings package be the top level one seems unusual Jan 25 13:10:20 that makes sense, but I don't understand why rename unity-2d Jan 25 13:10:49 rsalveti,I meant rename an existing package to unity-2d Jan 25 13:10:49 if it's just this merge, then ok Jan 25 13:10:57 ok :-) Jan 25 13:11:45 so that gnome-session can Recommends gnome-panel | unity | unity-2d and do not pull in gnome-panel all the time Jan 25 13:11:58 The traditional reasoning behind the foo-default-settings packages is that they set defaults for various configuration subsystems that are unique to a selected interaction model, and may not always be appropriate for some specific package. If unity-2d-default-settings is doing this, it would be better not to merge. Jan 25 13:12:22 If it's just setting stuff for the unity-2d package, then it doesn't matter. Jan 25 13:12:35 persia, makes sense, espcially since those settings are ubuntu addons to upstreams. Jan 25 13:12:45 anyway a to plevel meta[pavkage with a short and sane name is desired Jan 25 13:12:46 :) Jan 25 13:12:56 Indeed. For comparison, consider xubuntu-default-settings, kubuntu-mobile-default-settings, etc. Jan 25 13:13:15 Maybe. Depends if it's a flavour or not. Jan 25 13:13:41 I'd claim that the right choice would be ubuntu-desktop or similar, to get that experience. Jan 25 13:13:44 unity on debian or fedora will probably not use our themes or autostart settingd Jan 25 13:13:59 Right, which is why -default-settings doesn't belong in that source. Jan 25 13:14:05 I agree Jan 25 13:14:17 I thikn the source should contain what the upciek team writes Jan 25 13:14:20 (assuming that we seek null packaging, where feasible, in pursuit of automation) Jan 25 13:14:38 Excellent. I'll let you argue with everyone else whilst I sleep then :) Jan 25 13:16:00 this is a tangent though to what I was askking - to have a clean and final name we can stick into gnome-session Recommends: field :) Jan 25 13:18:17 Create some virtual supplied by all unity interfaces, and then control which meets the requirement through seeding, defaulting to "unity" Jan 25 13:19:13 packages like flash-kernel that in addition to the package source have a bzr trunk in LP need to be changed in both places ? Jan 25 13:19:27 I did not see it satte that it is mainatined in bzr which makes it more ocnfusing Jan 25 13:58:08 janimo_, change it in the bzr tree and build the package from there Jan 25 13:59:25 change, commit (UNRELEASED), change debian/changelog to natty), debcommit -r, push, bzr builddeb (or manually bzr export) and upload Jan 25 14:01:22 ogra, thanks. I have not yet built from bzr before. Jan 25 14:06:55 * janimo_ misses almost everything from git when working with bzr Jan 25 14:07:54 * rsalveti understands janimo_ Jan 25 14:08:10 but I must say I'm quite used with bzr already again Jan 25 14:08:23 god how I hate to debug qt Jan 25 14:08:36 a simple make takes an incredible amount of time Jan 25 14:08:52 that touches just one lib Jan 25 14:09:18 I hope I'll get used to it again, switching to git 3 years ago was equally awkward I think Jan 25 14:11:58 janimo_: git-bzr-ng provides you one way to get bzr into git (push is broken) Jan 25 14:13:03 hrw, thanks, I can work with bzr just don;t like it that much anymore Jan 25 14:13:33 my fingers also type git when I work with bzr Jan 25 14:27:47 persia, could you take a look at the control file at http://paste.ubuntu.com/558125/ ? Jan 25 14:29:00 purpose is to merge the -default-settings package into unity-2d source and create a metapackage ... i'm unsure about the conflicts/replaces and also aboutnot the fact if i need a transitional package or Jan 25 15:32:21 Does anyone know how to install ubuntu natively on an arm MID? Jan 25 15:33:39 there is no such thing as "an ARM MID" Jan 25 15:33:43 :) Jan 25 15:33:52 you need to be more specific Jan 25 15:33:58 (SoC model etc) Jan 25 15:34:52 The haipad m701 is an MID that claims it runs android 2.2. I want to be rid of android completely and install ubuntu on it natively, if that's possible. Jan 25 15:35:57 Here, I'm looking up the specifics... Jan 25 15:35:59 that's a Telechip TCC8902, one of my buddies has it.. i think it's arm9.. Jan 25 15:36:19 ah, no recent ubuntu then Jan 25 15:38:56 Is there an older version, say 9.04 or something, that would work on it? Jan 25 15:39:25 9.04 would work buut it is unsupported (EOL) Jan 25 15:39:58 wow, it's actually arm11... (armv6) Jan 25 15:40:05 so karmic.. Jan 25 15:40:20 yeah, that should work then Jan 25 15:40:38 and karmic is actually still not EOL until april Jan 25 15:40:51 so you have three wonderful months :) Jan 25 15:40:53 SaidKLE, was that the one on woot? about a week ago? Jan 25 15:41:30 SaidKLE: hint: ubuntu has a relatively conservative older brother which is still supporting arm9. :-) Jan 25 15:42:02 hey, we strill support arm9 (for three months) :P Jan 25 15:42:06 LetoThe2nd: I thought squeeze dropped arm for armel? Jan 25 15:42:19 Amaranth, they did Jan 25 15:42:36 but they didnt go -march=armv7 Jan 25 15:42:58 * LetoThe2nd is running an openrd here on deb5.0. :-) Jan 25 15:43:10 i think they still support armv5te Jan 25 15:43:18 even with armel Jan 25 15:43:22 ogra: Oh, I thought that was the point of changing the port name :) Jan 25 15:43:42 well the old arm port did armv4.. Jan 25 15:43:53 right Jan 25 15:43:57 and no thumb Jan 25 15:44:02 So, I should probably run karmic? I'm fine with that, but how does one go about actually getting ubuntu onto the device? Jan 25 15:44:09 for openmoko compatibility ;) Jan 25 15:44:27 SaidKLE, you should know everything about your bootloader and kernel Jan 25 15:44:46 then just make them use an ubuntu rootfs (see topic) Jan 25 15:44:55 I don't. I've tried to figure that out, but haven't managed it yet. Jan 25 15:45:09 SaidKLE: there's a manual for the openrd providing a guide to flashing ubuntu. read it, combine it with what ogra said and there you go :-) Jan 25 15:45:10 right, thats the first step as ubuntu doesnt have either for you Jan 25 15:45:31 okay. I'll be looking that up... Jan 25 15:45:48 SaidKLE: What HW are you sitting on? Jan 25 15:45:51 your kernel needs to have all options builtin for the udev used in karmic Jan 25 15:45:59 once I know that though, do I just put a certain ubuntu file where the boot loader expects to find it? Jan 25 15:46:01 else it wont work Jan 25 15:46:40 Haipad m701 telechip tcc8902 Jan 25 15:46:48 mind you, its usually not a beginner task to get another OS on such devices Jan 25 15:47:05 SaidKLE: nö, its first about getting your bootloader and kernel right, including config and layout on bare flash. Jan 25 15:47:12 and you will likely have to rebuild your kernel etc Jan 25 15:47:38 and know about flashing or how to run the OS off an SD card Jan 25 15:47:39 from past experience, if you have a bootloader and kernel (config) available, then its a lot easier to get ubuntu up and running Jan 25 15:48:06 SaidKLE: if you're a beginner and don't have access to advanced debricking tools (i.e. a good jtag interface) i guess it's advisable to go either hunting for some existing port or leave the thing as is. Jan 25 15:48:27 all boards I have worked with in the past had a linux demo kit with kernel which I could use with ubuntu Jan 25 15:48:36 at the moment we don't even know the bootloader type... :-/ Jan 25 15:48:52 nor what kind of kernel is used Jan 25 15:49:07 yeah, basically we know - nothing. Jan 25 15:51:22 I know this is a wrong place for the question, but what is console-kit? If I'm building a system with no interactive uses (only services), can I live without it? Jan 25 15:53:25 my system is not going to run X -- only a Qt QWS application, and without DBUS Jan 25 16:15:42 persia, updated paste at http://paste.ubuntu.com/558148/ Jan 25 20:17:14 ogra_, I don't know precisely how the pieces fit together, but I'm opposed to merging a -default-settings package into anything that has code, and don't see the point of metapackages except as expressions of flavours. Jan 25 20:23:24 We may have a problem with images next week. X is getting updated Friday. May cause turbulence. Jan 25 20:26:02 ogra_, Looking at your changes to control: Please don't call something a "metapackage" if it has content (and isn't in section: metapackages). The mechanics of replacement look acceptable, although take care with versions (there's no changelog entry in the diff, so I don't see that). Jan 25 20:35:05 persia, well, how should i call it ? Jan 25 20:35:39 it depends on the rest of unity-2d but indeed ships gconf bits and postinst Jan 25 20:36:33 Hmm. Jan 25 20:36:50 Well, there's two ways to do it. the one I like, and the one you seem to be pursuing. Jan 25 20:36:59 effectively it does the same a metapackage does Jan 25 20:37:12 i only do what upstream will do anyway Jan 25 20:37:18 The common way is to have a metapackage (which is empty by definition) on a per-flavour basis, and have that pull -default-settings. Jan 25 20:37:27 i noticed there is a omap4 ubuntu server daily build now Jan 25 20:37:36 jhobbs, Does it work for you? Jan 25 20:37:36 where does the package list that goes into that come from? Jan 25 20:37:45 well, thats exactly what upstream wants to get rid of Jan 25 20:37:47 i haven't tried it yet, but i'll let you know Jan 25 20:37:56 jhobbs, It's the server stuff from the ubuntu.natty seed Jan 25 20:38:01 ogra_, Why? Jan 25 20:38:10 we are supposed to have a unity-2d as we have a unity package in the end Jan 25 20:38:26 the prob is that the binaries in unity-2d can be used separately Jan 25 20:38:48 so there are four binary packages and we need some toplevel that has to be called unity-2d Jan 25 20:39:01 which depends on the other binaries Jan 25 20:39:12 In that case, don't claim to be a metapackage. unity is "Unity is a graphical interface designed for Ubuntu Netbook Edition". That's unfortunate, as I heard that Ubuntu Netbook and Ubuntu Desktop were merging, but it's close. Jan 25 20:39:23 the toplevel is supposed to ship the session and config Jan 25 20:39:51 Just have the description say "Unity 2D is a graphical interface providing the Unity experience, but relying only on 2D acceleration" or similar. Jan 25 20:40:18 Just stay away from "metapackage", because that means something specific, which this isn't. Jan 25 20:49:37 btw: The une-efl session bug still exists in today's image. Jan 25 20:49:55 * GrueMaster thought it was fixed. Jan 25 20:49:57 Which bug is that? Jan 25 20:50:49 bug 707014 Jan 25 20:50:50 Launchpad bug 707014 in netbook-launcher-efl "Ubuntu Netbook Edition 2D fails to launch from gdm by giving "No valid session found"." [Undecided,Fix released] https://launchpad.net/bugs/707014 Jan 25 20:51:24 Which version of netbook-launcher-efl do you have? Jan 25 20:52:28 0.3.2-0ubuntu5. It doesn't appear to have been properly Fix Released. I don't see a commited entry in the bug. Jan 25 20:53:01 It claims to have been fixed in 0.3.2-0ubuntu6, so that's expected behaviour (based on https://launchpad.net/ubuntu/+source/netbook-launcher-efl/+changelog ) Jan 25 20:53:21 rsalveti forgot to add LP: foo in the changelog, that's all. Jan 25 20:53:39 Still, it should be in the pool. Jan 25 20:54:00 checking. Jan 25 20:54:03 It is, from what I see. Jan 25 20:54:06 GrueMaster: Failed to upload on anacardiaceae (armel) Jan 25 20:54:13 that's the error I saw here Jan 25 20:54:20 Ah, source is in pool. Binary failed. Jan 25 20:54:27 rsalveti, Have you asked the LP folks about that yet? Jan 25 20:54:29 not to build, but to upload Jan 25 20:54:35 persia: not yet, just saw it Jan 25 20:55:01 persia: should I ask for help at the lp channel? Jan 25 20:55:05 Fail to Upload usually indicates something wonky with soyuz. It's almost always worth asking in #launchpad Jan 25 20:55:08 or is more related with ubuntu itself? Jan 25 20:55:16 No, it's #launchpad. Jan 25 20:55:25 Let me check if an update will pull it. Jan 25 20:55:30 * rsalveti asking Jan 25 20:56:00 I'm not seeing anything on lp that indicates fail to publish. Jan 25 20:56:50 Looks like it was just posted after the image build started. No worries. Jan 25 20:57:03 well, I got an email with it Jan 25 20:57:21 apt-get update;apt-get install netbook-launcher-efl has found the updated package. Jan 25 20:57:23 Odd, indeed. Jan 25 20:57:53 It'll be in tomorrows image (assuming pool churn doesn't clobber the image build). Jan 25 20:58:40 I am a little worried about A2. We have a new QT, X stack, and now I find we are getting 2.6.38 kernel. Jan 25 20:58:47 All for A2. Jan 25 20:58:47 probably it failed but then soyuz tried to publish it again later on Jan 25 20:58:55 Possible. Jan 25 20:59:25 haha, probably lots of new bugs :D Jan 25 20:59:36 yea. Jan 25 21:00:46 GrueMaster: did you try latest omap 3 image already? Jan 25 21:00:59 I still didn't check to see if the x-loader got into the image correctly Jan 25 21:01:28 btw: rsalveti: when you post a bug fix, add (LP: #) to the changelog. Then LP will handle marking the bug as Fix Released. Jan 25 21:01:45 I know, I was going to do that after you tested the bug Jan 25 21:01:54 And no, I haven't tried the omap 3 image. Beagle is put away, and I haven't gotten an XM yet. Jan 25 21:01:56 but then janimo pushed my debdiff Jan 25 21:02:09 oops. Jan 25 21:02:12 as I created the debdiff before actually opening the bug Jan 25 21:02:19 Ah. Jan 25 21:02:32 was planning to update it after the test :-) Jan 25 21:02:34 bass ackwards. Jan 25 21:02:39 :P Jan 25 21:03:03 ok, will download latest omap 3 image then Jan 25 21:03:56 whee. 46 upgrades since today's image was built. Jan 25 21:04:00 !ohmy > GrueMaster Jan 25 21:04:03 GrueMaster, please see my private message Jan 25 21:45:12 well, omap 3 image seems to be working again Jan 25 21:45:17 at least going into resize Jan 25 22:17:43 GrueMaster: awesome, omap 3 image seems to be working quite well Jan 25 22:18:01 time to activate the console and see the boot messages Jan 25 22:18:28 has anyone heard of anybody getting ubuntu-arm booting native on a nook color? Jan 25 22:41:46 yup, no issues and no kernel traces Jan 25 22:41:54 GrueMaster: so it seems ok, finally :-) Jan 25 22:54:28 ka6sox, Are you able to boot arbitrary kernel against arbitrary rootfs yet? Jan 25 23:01:40 persia, arbitrary? I have a rebuilt kernel from Android (have the source and have rebuilt it) Jan 25 23:02:09 and also have the openembedded.org reference distribution called Angstrom booting Jan 25 23:03:02 so thats pretty random Jan 25 23:03:15 ka6sox, where does it stop booting? Jan 25 23:03:17 Well, hardly random, but at least one of your choice. Jan 25 23:03:18 if you already have a common distro running on it should be easy to run ubuntu Jan 25 23:03:24 Have you tried with an Ubuntu rootfs? Jan 25 23:04:23 rcn-ee_at_work, it boots all the way up Jan 25 23:04:44 persia, I didn't know about the maverick-omap till later Jan 25 23:05:26 ka6sox, So it works for you? Jan 25 23:05:34 How is the screen refresh? Is X behaving nicely? Jan 25 23:05:37 rsalveti, the main issue is that we used a no-initrd setup on Angstrom Jan 25 23:06:09 persia, x isn't quite there...mostly because they don't seem to have the latest sgx running but we are using FB. Jan 25 23:06:19 hm, it'd be good to have initrd to boot ubuntu, but I believe you can make it work with current setup Jan 25 23:06:34 Does the nook not support an initrd? Jan 25 23:06:48 if you have access to the bootloader, it should be fine Jan 25 23:06:52 (funny because the guy who does Angstrom was the guy who worked on beagleboard. Jan 25 23:07:07 rsalveti, we don't really... Jan 25 23:07:15 its kind of kludgy... Jan 25 23:07:35 ka6sox: how are you flashing a newer kernel and booting it? Jan 25 23:07:47 2 partitions on an SD Jan 25 23:08:00 1 vfat with kernel initrd and MLO Jan 25 23:08:08 1 ext2 with rootfs Jan 25 23:08:21 shove it in and away you go Jan 25 23:08:54 so you alread have an initrd Jan 25 23:09:06 an Android one Jan 25 23:09:08 ubly Jan 25 23:09:11 ugly Jan 25 23:09:18 if it's possible to load a valid kernel and initrd, then you can regenerate one for ubuntu Jan 25 23:09:39 but then it could have size restrictions Jan 25 23:09:41 they are using 2.6.29 for the existingkernel. Jan 25 23:09:50 urgh Jan 25 23:10:13 I was trying *not* to have to redo the kernel...but I can Jan 25 23:10:15 weird that they are still way behind, as it's basically an omap3 Jan 25 23:10:33 vendor choice as to which version you use...they are stuck on 2.1 Jan 25 23:10:42 but still, you can generate an initrd for it Jan 25 23:10:57 I have a working chroot of maverick on it. Jan 25 23:11:10 well, you have the sources, then you can check what they changed for it Jan 25 23:11:17 lot of work probably, but still something Jan 25 23:11:20 Redoing the kernel is part of the point, I'd think, so that you could have a distributed sourceful kernel package for reliable bugfixing. Jan 25 23:11:35 ka6sox: the image for ac100 is also using 2.6.29 Jan 25 23:12:01 once you got the correct kernel and modules, copy them to the ubuntu rootfs, regenerate the initrd and it should be fine Jan 25 23:12:08 if you're booting with correct cmd args Jan 25 23:12:09 the biggest problem is that they didn't send nice "patches" or even tell us what they started with before they started *their* patches. Jan 25 23:12:22 hehe, usual Jan 25 23:12:31 source == big and ugly tarball Jan 25 23:12:56 rsalveti, generated on a windows machine (Documents and Files dir was there) Jan 25 23:13:05 hehe Jan 25 23:14:04 What VCS tools need is a "discover" function, where they compare past revisions against a given source, find the set of commits that appear to have been applied, and construct a virtual history including prior commits (split point + cherrypicks) and then a massive commit including the rest of the diff, which can then be picked apart at leisure. Jan 25 23:14:07 okay so what I'll do is unpack the uImage of the kernel and get the modules out of the initrd thats there and put them in maverick chroot...then regenerate teh initramfs. Jan 25 23:14:35 yup, or even build the kernel on your own Jan 25 23:14:45 * ka6sox reprogrammes the LF servers that check for GPL compliance to do that :D Jan 25 23:14:45 and install the modules at a custom path, then you can easily copy the modules Jan 25 23:15:07 or just deb-pkg (iirc) that will generate a kernel deb file Jan 25 23:15:15 that you can install at your arm chroot Jan 25 23:15:18 and regenerate the initrd Jan 25 23:15:51 * persia notes that lots of extra points accrue for creating a source package that generates the appropriate binary packages and uploading to the archive so everyone can share Jan 25 23:16:01 I've never tried building cross with make-kpkg kernel_image :D Jan 25 23:16:18 Could build native. Would take a bit on the Nook, but not forever. Jan 25 23:16:21 make -j 6 ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- CONFIG_DEBUG_SECTION_MISMATCH=y deb-pkg Jan 25 23:16:25 cross compiling Jan 25 23:16:33 with any kernel source tree Jan 25 23:16:38 just need the proper .config file in place Jan 25 23:16:56 okay I have the defconfig that they used and the patched source Jan 25 23:17:00 if you're running maverick, just install the ubuntu/linaro cross compiler Jan 25 23:17:21 still on lucid Jan 25 23:17:45 I'm an infra guy so I generally stay with LTS Jan 25 23:18:05 On lucid, you should be able to run `mk-sbuild maverick` to generate a maverick schroot, which you can then use to install and run the cross-compiler. Jan 25 23:18:15 kk Jan 25 23:18:17 ka6sox: check https://launchpad.net/~linaro-maintainers/+archive/toolchain then Jan 25 23:18:24 the backported linaro/ubuntu cross compiler Jan 25 23:18:30 okay good. Jan 25 23:18:33 Unless you have another reason to install all of ubuntu-dev-tools, you might want to just cherrypick that script from it. Jan 25 23:19:09 okay so I'll have to generate the kpkg and have to bust it out though to create the uImage. Jan 25 23:19:33 I don't recall mkuImage being able to read a .deb Jan 25 23:19:37 generally I just install it, regenerate the initrd and then use mkimage to create the uImage and uInitrd Jan 25 23:19:51 all that inside a valid arm chroot Jan 25 23:19:59 okay great...thanks. Jan 25 23:20:25 the omap3 image should be closest to what we have. Jan 25 23:20:37 3621 Jan 25 23:21:04 yup, I'd get first a minimal ubuntu rootfs (maverick) Jan 25 23:21:09 and try to make that work Jan 25 23:21:32 ya, I need rndis and ssh support mostly Jan 25 23:21:51 http://people.canonical.com/~rsalveti/lamont/beagle/ Jan 25 23:21:55 this image should work for you Jan 25 23:21:59 is a minimal maverick image Jan 25 23:22:06 that I generated for beagle xm Jan 25 23:22:18 okay thanks. I'll use that Jan 25 23:22:28 guess same one we're using at our xm builders Jan 25 23:23:19 I've been playing for years on things like the Slug and Pre but first time I"ve "ported" an OS on a device Jan 25 23:24:10 :-) Jan 25 23:24:26 would like to buy a nook color to play, but kind of expensive around here Jan 25 23:24:36 oh? Jan 25 23:24:43 out of country I suppose Jan 25 23:25:50 yup Jan 25 23:25:52 brazil Jan 25 23:26:07 appreciate the help...I'll stop by later with a report... Jan 25 23:26:39 we ported what we call "optware" (linux apps running on Android) Jan 25 23:26:51 so we have the tools we need to port this. Jan 25 23:26:55 (like chroot) Jan 25 23:27:34 later..thanks. Jan 25 23:28:03 ka6sox, Good luck. Jan 25 23:29:18 cool, later Jan 25 23:57:56 ka6sox-away: do you have a chroot version of init that works under bionic? Jan 26 00:49:11 tmzt: if init? Jan 26 00:49:20 as in sysV init? Jan 26 00:49:20 yes Jan 26 00:49:51 yeah Jan 26 00:50:18 yes, it works Jan 26 00:50:28 this is what I started on but I couldn't get it to work https://github.com/tmzt/nativenetbook-nativeinit Jan 26 00:51:40 tmzt, have you heard of Optware? Jan 26 00:51:52 yeah Jan 26 00:52:10 what happened is we ported optware to work on android Jan 26 00:52:56 https://github.com/optware4android/nook-color Jan 26 00:53:35 after a few iterations we had something that works. Jan 26 00:53:58 so that gave me the ability to use things like ltrace and strace etc to look at porting. Jan 26 00:54:14 and chroot Jan 26 00:54:35 ka6sox: I'm trying to get my X server to work with existing X sessions Jan 26 00:54:46 but xinit wants to start the server itself and mine runs outside of the chroot Jan 26 00:55:36 then only "env" I bring into my chroots are /proc /sys and /dev as bind mounts Jan 26 00:55:36 tmzt, Do you mean that you want stuff launched in the chroot to be clients of an X server outside the chroot? Jan 26 00:55:43 yes Jan 26 00:55:47 so nothing outside of that is available Jan 26 00:55:52 my X server androix.org runs as a native bionic aplication Jan 26 00:56:10 er, android application linked against bionic Jan 26 00:56:40 tmzt, I don't know much about android architecture, but you can usually get that to work if you can connect to the X server over the network or you bind-mount /proc: just export your desired DISPLAY Jan 26 00:56:51 riht Jan 26 00:56:55 er right Jan 26 00:57:23 I know, but the issue is I want to launch Xsession and also need things like a dbus system bus Jan 26 00:57:46 when I tried to do it it failed because it lost the controlling tty or other wise the forked children failed to run Jan 26 00:57:59 tmzt, well...we have optware running *alongside* of android. Jan 26 00:58:07 so somethings like that are possible. Jan 26 00:58:24 I'll look at it, I don't think most of the X clients are packaged though Jan 26 00:58:26 tmzt, If you need dbus, try bind-mounting /var/run Jan 26 00:58:48 okay back to porting. Jan 26 00:58:50 android doesn't use dbus, so dbus would be running inside of the chroot Jan 26 00:59:03 anyway, thanks I'll look at this nook stuff and see if there's anything I can use Jan 26 01:00:10 if nothing else it shows that you can run linux things with linux libs alongsiide of the android/bionic stuff. Jan 26 01:00:40 hardcoded paths are a pain (like /lib/ld-linux.so Jan 26 01:00:41 ) Jan 26 01:01:22 but since their /lib doesn't contain that it doesnt matter. Jan 26 01:36:55 ka6sox: ah, so you aren't using a chroot? Jan 26 01:37:17 for Ubuntu, yes, but optware is NOT a chroot Jan 26 01:37:25 it all goes in /opt/ Jan 26 01:37:34 and we bind mount to the rootfs Jan 26 01:39:12 the only thing in the original rootfs is a few things that *have* to be in /lib and a small script that calls sysV init to get the optware services running Jan 26 01:39:45 that's in the git repo? Jan 26 01:39:56 the one I pointed you to..yes Jan 26 01:39:59 including sources **** ENDING LOGGING AT Wed Jan 26 02:59:57 2011