**** BEGIN LOGGING AT Fri Apr 23 02:59:56 2010 Apr 23 05:03:37 weird... neither networkmanager nor wicd will connect with my rtl8187. Apr 23 05:03:58 That's unexpected. Apr 23 05:10:18 X11 connection rejected because of wrong authentication. Apr 23 05:10:28 Trying to run ssh-with-X. Apr 23 05:11:39 Hrm. It generally works for me if I ssh out of my X session into somewhere else, with the appropriate .ssh/config bits to pass X, and then run an X app. Apr 23 05:11:51 Dunno if that's precisely how you're doing it. Apr 23 05:13:28 I have forwardx11 and forwardx11trusted in .ssh/config Apr 23 05:13:33 (capitalized properly there.) Apr 23 05:17:25 Ought do it, really. Apr 23 05:19:34 hmm, it's not working for me. Weird. Apr 23 05:20:47 Indeed. Apr 23 05:22:05 oh, "no space left on device" Apr 23 05:22:13 No wonder. Apr 23 05:22:33 Would've been nice if ssh had told me that. Apr 23 05:23:44 Heh. That explains it. Apr 23 05:25:32 Need bigger SD card -- 2 gigs is tiny. Apr 23 05:25:54 We recommend 4G of available storage for a running system. Apr 23 05:25:58 (or more). Apr 23 05:26:10 It can (barely) fit in 2G, but one usually runs into issues. Apr 23 05:30:52 hmm, what read and write speeds is the beagle itself capable of? Apr 23 05:31:14 For example, why bother with a "class-10" card if it can't do class 10 speeds?... =þ Apr 23 05:39:33 weird... blueman won't let me use the pulseaudio plugin. Apr 23 05:40:57 ValueError: invalid literal for int() with base 10: '21-63-gd3efa-dirty' Apr 23 05:41:23 File a bug. Apr 23 05:41:42 That *should* just work. Apr 23 05:41:51 (and not working is a clear port failure) Apr 23 05:43:29 https://bugs.launchpad.net/ubuntu/+source/blueman/+bug/535446 Apr 23 05:43:32 Launchpad bug 535446 in blueman (Ubuntu) "[lucid] Blueman Pulseaudio plugin cannot be loaded (affects: 4) (heat: 24)" [Undecided,Confirmed] Apr 23 05:44:18 Does the patch work for you? Apr 23 05:45:53 Nifty. Looks like I can upload that. Tell me the patch works, and I'll put it in lucid. Apr 23 05:54:33 I just edited the file manually, rather than patching it. Apr 23 05:55:18 The patch seems to be for something different. Apr 23 05:55:28 Now I just need to remember which audio port was line-out on the beagle. Apr 23 05:57:56 I'm getting no audio on either port. Apr 23 05:58:42 But your traceback matches the one in that bug? Apr 23 05:59:02 Yes, the invalid literal for int() Apr 23 05:59:44 Oh, xax200's traceback? Apr 23 05:59:55 I think the patch is for mmbossoni's traceback Apr 23 06:00:11 No, I'm confused Apr 23 06:00:15 * persia reads the bug again Apr 23 06:01:18 Looks like the bug partly got highjacked? Apr 23 06:02:13 By the original reporter, which makes it complicated. Apr 23 06:02:33 "I noticed another problem" is the key phrase one never wants to see in a bug report. Apr 23 06:03:25 xax200 seems to recommend small changes to lines 115 and 218 (comments 3&4). Does that work for you? Apr 23 06:03:48 I just changed the first one mentioned. Apr 23 06:04:13 115. Apr 23 06:05:34 Right, which is a bit different than the patch in comment #5, which seems to be a different bug. Apr 23 06:07:39 hmm, is there any way to get blueman to run without an X server? Or a dummy X server? Apr 23 06:09:26 xvfb? Apr 23 06:09:44 Mind you, it might be tricky to press the buttons that way :) Apr 23 06:10:30 Or I suppose I could run a real X server with just a blueman-applet. Apr 23 06:12:44 argh, too dang many mixer controls... Apr 23 06:12:49 now is it headset, or earpiece? Apr 23 06:20:31 ah, maybe I do need that other bit of "int" change. Apr 23 06:31:56 argh, pavucontrol over ssh controls the client's pulseaudio, not the server's. Apr 23 06:32:46 heh. pavucontrol is too smart :) Apr 23 06:33:19 Dag-blast it, I need this PPA in armel: Apr 23 06:33:20 https://launchpad.net/~blueman/+archive/blueman-dailies/+packages Apr 23 06:34:01 apt-get source, sbuild -d lucid source.dsc Apr 23 06:34:24 If that doesn't just work: `mk-sbuild lucid` should get you a nice clean buildd chroot. Apr 23 06:34:53 And since we never managed to fix the command-not-found bug, install ubuntu-dev-tools to get mk-sbuild. Apr 23 06:37:27 great, I'm tight on space as it is.. Apr 23 06:37:43 Well, you could build with qemu. Apr 23 06:37:53 Do you have a lucid x86 system available? Apr 23 06:37:59 (or can you make one) Apr 23 06:40:19 I do have lucid (amd64, rather). Apr 23 06:40:26 It's easier just to go out and get a bigger SD card. Apr 23 06:40:28 =þ Apr 23 06:40:32 IF you like. Apr 23 06:41:08 Installing ubuntu-dev-tools, sbuild, and qemu-extras-static on your amd64 machine lets you run `mk-sbuid --arch=armel lucid` which generates an emulated lucid chroot for building stuff. Apr 23 06:41:28 You can then run `sbuild -d lucid-armel foo.dsc` to generate lucid armel binaries on your host. Apr 23 06:41:47 I've had a couple issues with some mono packages, but most stuff seems to just work. Apr 23 06:50:45 oooh, I got bluez to segfault. Apr 23 06:51:16 what's a really, really tiny window manager (just needs to be able to change focus and move windows)? Apr 23 06:51:42 tinywm Apr 23 07:12:09 oh guawd, bluetooth audio streamed to the beagleboard works hilariously poorly. Apr 23 07:12:33 Picture taking an audio stream, chopping it into tiny bits, then chipmunkifying and looping those bits. Apr 23 07:12:52 Changing the resampler to speex-fixed-0 seems to have worked. Apr 23 07:13:54 Oh my. Apr 23 07:14:13 Interesting... changing it to speex-fixed-0 worked so well, in fact, that it works better than when I try to do the same between two x86 systems. Apr 23 07:14:23 On those, I get an assertion failure, and pulseaudio aborts. Apr 23 07:14:40 file a bug :) Apr 23 07:14:57 Can't do it with apport. =þ Apr 23 07:15:20 Why not? Apr 23 07:17:10 Upstream ticket, by the way: http://www.mail-archive.com/pulseaudio-tickets@mail.0pointer.de/msg02912.html Apr 23 07:18:17 Apport "does not support" reporting assertion failures. Apr 23 07:24:24 Ugh! Apr 23 07:25:14 oh yeah, better than speex-fixed-0 is src-linear Apr 23 07:26:07 ...maybe. Apr 23 07:26:17 You'll find it eventually :) Apr 23 07:32:26 Okay, linear is the only usable one. Apr 23 07:32:34 well, that, and fixed-0. Apr 23 07:33:26 https://bugs.launchpad.net/ubuntu/+source/pulseaudio/+bug/358831 Apr 23 07:33:28 old Apr 23 07:33:30 Launchpad bug 358831 in pulseaudio (Ubuntu) (and 1 other project) "[ARM] Pulse Audio eating up to much CPU" [High,Invalid] Apr 23 07:36:02 Are you running a distro kernel? If so, please file a new bug. Apr 23 07:36:24 interesting... speex-float-1 works. Apr 23 07:36:27 Surprisingly well. Apr 23 07:36:39 That old bug got marked invalid because someone filed with with random vendor HW and random vendor kernel, and it's not clear if the issue can be replicated in Ubuntu. Apr 23 07:36:53 Yeah, I figured as such. Apr 23 07:37:02 Wow, float-1 works really well. Apr 23 07:37:13 55% CPU usage, though. Apr 23 07:37:34 And granted, bluetooth is an extreme use case. =þ Apr 23 07:37:40 That needs a bug. Might or might not get a fix, but if it's replicable on known HW/kernel there's a chance that someone can work on it. Apr 23 07:37:53 Network streaming with pulseaudio native works really well with the default resampler. Apr 23 07:38:12 those resamplers would benefits from being rewritten for neon if they aren't already Apr 23 07:38:12 Not really. Lots of folk have made noises about running Ubuntu on an N900 or similar, where bluetooth is a common function. Apr 23 07:38:43 cwillu_at_work: Feeling bored today? Want to dig in? Apr 23 07:38:46 :) Apr 23 07:38:58 persia, I've been at work for 20 hours today :p Apr 23 07:39:14 That would be "no" then :) Apr 23 07:39:25 on the other hand, my sd cards _almost_ boot on c3, c4 beagles and overos :p Apr 23 07:39:38 What bit doesn't work? Apr 23 07:39:46 the bit where it boots :p Apr 23 07:39:50 * persia thought the standard images just worked on C3/C4 Apr 23 07:40:01 I'm not using standard images Apr 23 07:40:09 Oh, what are you changing? Apr 23 07:40:25 * persia is wondering how much of the standard-image-build-stuff can be leveraged Apr 23 07:40:26 the overo is the source of my grief, and it looks like I fried something on the serial level shifter Apr 23 07:40:32 Ouch! Apr 23 07:40:35 persia: my use case is not streaming from ARM to a headset... but the other way around... Apr 23 07:40:43 streaming from one computer to the beagle over BT. Apr 23 07:40:44 I've got a hacked up rootstock, so it's basically just a standard deboostrap + extra scripts Apr 23 07:41:14 DanaG: Do you expect that pulse performance differs based on the direction of the stream? Apr 23 07:41:26 I can see the uboot prompts, but it sends garbage constantly, so I can't get _past_ the uboot while serial is plugged in Apr 23 07:42:00 I'm not sure... my only regular non-PC headset device is my FreePulse headset... and the power button on it has failed. Apr 23 07:42:01 And you can't see what you're doing when serial isn't attached. Yeah. That makes it hard. Apr 23 07:42:31 and I only have two sd card readers, and they're both tied up writing out btrfs images at ~50kb/s Apr 23 07:42:43 icantbelieveitsnotbtr? Apr 23 07:43:12 DanaG, already saved my ass a couple times; I really really hate sd cards, and I really really hate driving four hours to replace one Apr 23 07:43:24 checksumming + metadata mirroring ftw! Apr 23 07:43:33 Hmm, how stable is the on-disk format for btrfs? Apr 23 07:43:48 pretty much completely Apr 23 07:43:57 Can you convert from ext4 with extents? Apr 23 07:44:00 the occasional forward transition, but they're not mucking with it Apr 23 07:44:01 yep Apr 23 07:44:27 there's still a bunch of corner cases you have to be aware of though Apr 23 07:44:45 For now, I've been using ext4 with data=journal, since I value data integrity and the ability to recover from just plain unplugging... over the lifespan of SD cards. At least, that's true of this 2-gig card. Apr 23 07:45:02 i.e., determining how much free space is non-trivial, and it behaves fairly badly if you run out Apr 23 07:45:23 see, journalling doesn't get along with sd cards Apr 23 07:45:41 Hmm, I may stick with ext4 for now... though a good point: I should make a DD image, if nothing else, of the card. Apr 23 07:46:20 and let me assure you that fsck does more harm than good if the journal goes kabloiee Apr 23 07:47:08 hmm. Apr 23 07:47:13 So am I better off with no journal? Apr 23 07:47:43 you're better off with a keen understanding of how filesystems work and how crappy sd cards handle corner cases :) Apr 23 07:48:00 but yes, no journal is preferable Apr 23 07:48:31 At least, unless you're sure that your particular card behaves reasonably well regarding wear levelling (although they're all crappy, these aren't ssd drives) and power loss Apr 23 07:49:40 Ah. Hmm, then I need to find what tune2fs option to set... Apr 23 07:49:53 specifically, most/many/some cards handle things _really_ badly if they lose power while they were writing; it's not the clean case that a journal handles, you can lose data that you weren't even writing to Apr 23 07:50:29 Ugh. Apr 23 07:50:37 hence my current love of btrfs: it can tell you if things are corrupted and exactly how corrupted they are Apr 23 07:53:24 Login incorrect. Apr 23 07:53:24 Give root password for maintenance Apr 23 07:53:24 (or type Control-D to continue): Apr 23 07:53:36 I press ANYTHING ... even ctrl-d, it loops back to there. Apr 23 07:53:39 How did you get there? Apr 23 07:53:45 did you have an fsck ? Apr 23 07:53:47 Yeah, you have to have a password. Apr 23 07:53:56 I think I did have one. Apr 23 07:54:00 you guys haven't fixed that yet? Apr 23 07:54:16 cwillu_at_work, release managers are hard to convince sometimes :) Apr 23 07:54:17 I'd been trying to get the udisks-over-ssh working... but it kept giving "stdin: is not a tty" Apr 23 07:54:18 cwillu_at_work: Fixed which? Apr 23 07:54:24 still working on it Apr 23 07:54:31 persia, bug 563618 Apr 23 07:54:35 Launchpad bug 563618 in util-linux (Ubuntu Lucid) (and 3 other projects) "Ignoring a broken clock results in infinite reboots; not ignoring results in fsck failure; no solution to this problem (affects: 1) (heat: 14)" [High,Triaged] https://launchpad.net/bugs/563618 Apr 23 07:54:43 persia, the arm images having the root account locked rather than simply a deleted password like the rest of ubuntu's archs Apr 23 07:54:44 Oh, that one. Yeah. Apr 23 07:54:45 passwd -d root Apr 23 07:54:47 (it doesnt always result in reboots) Apr 23 07:54:59 cwillu_at_work, they dont Apr 23 07:55:00 DanaG, passwd -d root the next time you're in to make it work right :p Apr 23 07:55:14 Thanks. I'm just doing fsck on my host, for now. =þ Apr 23 07:55:23 cwillu_at_work, dont mix up "arm images" with rootstock built rootfses Apr 23 07:55:30 cwillu_at_work: I have disabled password on amd64/powerpc. Are you sure? Apr 23 07:55:30 ogra, my bad; Apr 23 07:55:34 ;) Apr 23 07:55:47 hence my hacked up rootstock :p Apr 23 07:55:57 cwillu_at_work, and you still havent filed a bug :) Apr 23 07:56:08 Take care with that: there's no guarantee that rootstock generates a policy-compliant install. Apr 23 07:56:11 * ogra will happily pull that into an SRU Apr 23 07:56:21 pretty sure I did actually :p Apr 23 07:57:06 https://bugs.edge.launchpad.net/project-rootstock/+bugs https://bugs.edge.launchpad.net/ubuntu/+source/rootstock ... neither has it Apr 23 07:57:21 http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=12454&prodSeriesId=4063703&prodNameId=4063704&swEnvOID=4030&swLang=13&mode=2&taskId=135&swItem=vc-78398-1 Apr 23 07:57:30 HP actually uses Debian on some Marvell thingy. Apr 23 07:57:50 * cwillu_at_work checks his bug list Apr 23 07:58:16 Doesn't surprise me much. HP used to ship a live Debian CD as their rescue environment back when I did server farm management. Apr 23 08:05:51 https://bugs.launchpad.net/ubuntu/+source/linux-ti-omap/+bug/566238 Apr 23 08:05:52 Launchpad bug 566238 in linux-ti-omap (Ubuntu) "wlan0 "Interface doesn't support scanning." -- CONFIG_CFG80211_WEXT is not set (affects: 1) (heat: 6)" [Undecided,New] Apr 23 08:06:54 Heat? Apr 23 08:09:17 hawt Apr 23 08:19:55 It's a new metric the LP guys added to help identify which bugs are interesting. So far, I haven't seen it provide interesting information, but I understand they are tuning the algorithm. Apr 23 08:22:06 It's the number of times the page is loaded Apr 23 08:22:18 It's a very often used metric amongst web developers Apr 23 08:22:34 The number of heats of a page gives a good idea of its popularity Apr 23 08:23:47 It's more complex than that. It takes into account *who* loads it, etc. Apr 23 08:23:55 And who files it, what status it is, etc. Apr 23 08:24:06 (or else someone gave me wrong information) Apr 23 08:25:07 persia: I was kidding actually :) Apr 23 08:25:10 hit versus heat Apr 23 08:25:21 I did fail at making it sound like a joke it seems! Apr 23 08:25:53 But it's one of those rare jokes that's funny *after* it's explained :) Apr 23 08:26:23 (but I'm in end-of-day mode, and so a bit humour impaired just now) Apr 23 08:29:24 * ogra_beagle waves from a successful netbook install Apr 23 08:29:35 \o/ Apr 23 08:29:44 only the clock issue left Apr 23 08:30:21 and htop shows a moderate 174M being used Apr 23 08:30:42 (only xchat, one terminal and desktop up though) Apr 23 08:31:30 which clock issue? Apr 23 08:31:46 * cwillu_at_work strongly recommends buying the lithium battery for the beagle Apr 23 08:32:17 cwillu_at_work, bug 563618 Apr 23 08:32:18 Launchpad bug 563618 in util-linux (Ubuntu Lucid) (and 3 other projects) "Ignoring a broken clock results in infinite reboots; not ignoring results in fsck failure; no solution to this problem (affects: 1) (heat: 14)" [High,Triaged] https://launchpad.net/bugs/563618 Apr 23 08:32:57 ogra, keep your computer plugged into the network :) Apr 23 08:33:10 cwillu_at_work, by default the board comes without, i want the images to work on a default config ... fsck needs fixing and will get fixed but that wont happen for lucid Apr 23 08:33:16 or do what I do (on the boards that don't yet have lithium batteries): Apr 23 08:33:23 network wont help Apr 23 08:33:25 use a filesystem that doesn't have a working fsck :) Apr 23 08:33:34 network isnt up when fsck is run Apr 23 08:34:15 nah, we will fix it properly Apr 23 08:34:33 for now it has to be a workaround though Apr 23 08:34:49 but i'm still waiting for release manager approval Apr 23 08:35:13 ted tso already agreed on fixing it right in fsck Apr 23 08:36:15 * ogra thinks he earned some breakfast before setting up the installation wikipages Apr 23 08:36:19 bbl Apr 23 09:00:45 ogra_cmpc: what address in the nand do you write the boot.scr, kernel, initrd to? (sd install) Apr 23 09:13:49 * cwillu_at_work jumps Apr 23 09:13:54 first image finished umounting! Apr 23 09:14:31 now if only I knew which reader was sdh and which was sdd so I don't panic the kernel removing the wrong one :p Apr 23 09:31:59 amitk, ?? Apr 23 09:32:09 amitk, i write to mtdblock devices Apr 23 09:32:23 no special addressing Apr 23 09:32:43 ogra: if i wanted to change the boot.scr, how would I do that? Apr 23 09:32:53 after install ? Apr 23 09:32:57 yes Apr 23 09:33:09 to get rid of the POS splash and get back serial console Apr 23 09:33:09 sudo fw_printenv and fw_setenv Apr 23 09:33:28 ogra: ah, nice Apr 23 09:33:29 sudo fw_setenv bootargs "your boot args" Apr 23 09:33:50 similar to redboot-cmdline Apr 23 09:34:22 nice integration! perhaps too simple :) Apr 23 09:40:20 heh Apr 23 09:41:12 asac, welcome to the "league of old farts" and ******** H A P P Y B I R T H D A Y ! ! ! ! ! ! ! ! ! ! ********* Apr 23 09:41:38 /dev/disk/by-id is useful. Apr 23 09:41:48 or /dev/disk/by-path Apr 23 09:43:52 eggonlea: you around? Apr 23 09:47:57 ogra: hi there Apr 23 09:48:01 hey Apr 23 09:48:27 Just saw your comment in https://bugs.launchpad.net/ubuntu/+source/util-linux/+bug/563618 Apr 23 09:48:29 Launchpad bug 563618 in util-linux (Ubuntu Lucid) (and 5 other projects) "Ignoring a broken clock results in infinite reboots; not ignoring results in fsck failure; no solution to this problem (affects: 1) (heat: 14)" [High,Triaged] Apr 23 09:48:33 yup Apr 23 09:48:40 its still not approved :/ Apr 23 09:48:51 How would I set up a branch with the extra code to test? Apr 23 09:48:59 and the longer it takes the less likely it will be we'll get it in Apr 23 09:49:38 branch from my branch, edit the script file i added, copy hook and script in place on your board, run update initramfs and try it Apr 23 09:50:06 (and indeed make sure the clock is wrong and mount time of the disk is in the future :) ) Apr 23 09:50:13 Er, what's a branch? Are we talking bzr or launchpad here, or something else? Apr 23 09:50:26 i linked a bzr branch to the bug Apr 23 09:50:46 lp:~ogra/initramfs-tools/initramfs-tools-fixrtc Apr 23 09:51:10 Unfortunately, I don't really know how to use bzr yet... Apr 23 09:51:14 My suggestions were "nice to have"— we could skip those if it's a problem. Apr 23 09:51:30 https://code.edge.launchpad.net/~ogra/initramfs-tools/initramfs-tools-fixrtc says at the top what to do Apr 23 09:51:41 bzr branch lp:~ogra/initramfs-tools/initramfs-tools-fixrtc Apr 23 09:52:40 cd into the newly created dir ... cd to scripts/local-premount ... edit fixrtc Apr 23 09:53:36 if the change is good: bzr commit -m 'your commit message'; bzr push lp:~dmart/initramfs-tools/initramfs-tools-fixrtc and add the branch to the bug Apr 23 09:54:11 Does the bzr push create a new branch owned by me? Apr 23 09:54:27 If you push to that location, yes. Apr 23 09:54:41 IF you push to an existing branch (to which you have write permission), no. Apr 23 09:54:50 so lp:~dmart/initramfs-tools/initramfs-tools-fixrtc gets made if it doesn't exist yet? Apr 23 09:54:54 the hook needs to go into /usr/share/initramfs-tools/hooks and the script into /usr/share/initramfs-tools/scripts/local-premount before calling update-initramfs -u (make sure both are executable) Apr 23 09:54:55 RIght. Apr 23 09:55:05 right Apr 23 09:55:11 ok... I will have a go Apr 23 09:56:39 oh, indeed ~dmart needs to be your LP id Apr 23 09:56:54 * ogra just noticed its not dmart :) Apr 23 09:57:29 I should maybe change it sometime, but yeah. I can probably work that one out :) Apr 23 09:57:35 :) Apr 23 09:57:58 No need to change it: just make sure your nick is registered on LP, and folks can find you from launchpad.net/people Apr 23 09:58:17 Also, once you start publishing branches and having a PPA, changing the name is hard. Apr 23 10:07:13 ogra: is scripts/local-premount/fixrtc sourced or exec'd ? Apr 23 10:07:29 exec'd Apr 23 10:07:39 (thats why it has a shebang) Apr 23 10:07:48 (doesn't always follow) Apr 23 10:08:24 Is it worth quitting the script if we didn't find root= ? Apr 23 10:08:41 heh Apr 23 10:08:46 (though it's unlikely) Apr 23 10:08:50 if we dont find root= we have bigger probs Apr 23 10:09:04 true - probably not worth it yet Apr 23 10:14:33 ogra: Do we have e2fsck.conf in the initramfs? Apr 23 10:14:42 nope Apr 23 10:14:47 and it wouldnt help anyway Apr 23 10:15:17 Just wondered if we could default to the fixrtc behaviour if the broken rtc option is set. Apr 23 10:15:22 no Apr 23 10:15:25 But this is a hack anyway... Apr 23 10:15:30 so it doesn't matter too much Apr 23 10:15:44 right, you would need to create an e2fsck.conf and set the parameter etc Apr 23 10:16:01 well, its the right fix once we have an actual option in fsck Apr 23 10:16:20 if thats there in maverick i'll make the installer create an e2fsck.conf Apr 23 10:16:34 but with the existing option not working its moot Apr 23 10:17:02 Can the e2fsck.conf be grabbed from /etc when making the initramfs? That way whatever the admin configures would be used Apr 23 10:17:17 i think thats happening already Apr 23 10:17:28 if not i'll make sure it happens in maverick Apr 23 10:18:04 in any case the option needs to be fixed first Apr 23 10:18:22 we'll have to wait for upstream to fix that Apr 23 10:23:18 What happens if this script terminates with non-zero exit status? Apr 23 10:23:50 Make it not do that. Use ||true if necessary. Apr 23 10:24:03 I want to use set -e, since if dumpe2fs fails or something else goes wrong, we may accidentally set the clock to junk... Apr 23 10:24:21 I believe it dumps when it fails. Apr 23 10:25:07 Is it possible to do sanity checking and set the clock to the epoch if we can't get useful data? Apr 23 10:25:31 depends what you mean by epoch :) Apr 23 10:25:50 What kind of sanity-check do you suggest? Apr 23 10:26:01 the beagle epoch is 01-01-2000 ... i'm not sure if you could even set the clock to an earlier date Apr 23 10:26:17 dmart: e.g. make sure the date is set to something between 1970 and 2038. Apr 23 10:26:48 dont put so much effort into that hack Apr 23 10:26:57 and dont make it to big Apr 23 10:27:05 * persia doesn't want set -e + the hack to make systems unbootable Apr 23 10:27:09 s/big/complex/ Apr 23 10:27:20 date --set will barf if the supplied textual date won't fit in time_t Apr 23 10:27:26 persia, they are unbootable already Apr 23 10:27:32 so we just give up (which may be the best thing) Apr 23 10:28:02 We shouldn't make any otherwise bootable cases unbootable Apr 23 10:28:02 and if you want to not use the fix, just drop fixrtc from the cmdline Apr 23 10:28:13 Fine. I don't like it because it's not elegant, but I agree it makes sense to do it right later. Apr 23 10:28:28 right, the hack is temporary anyway Apr 23 10:28:37 This is a bodge in advence of the real fix (which is to make e2fsck less picky about the clock) Apr 23 10:28:48 exactly Apr 23 10:28:58 and upstream is aware and agreed to fix it already Apr 23 10:28:59 The other way to work round the issues is mount -o remount,rw / ; mount -o remount,ro / Apr 23 10:29:15 (but this is obviously not a great idea if the root fs might be damaged) Apr 23 10:29:19 i think that will slow down your boot significantly Apr 23 10:29:20 Not "less picky", but rather "handle the case where the clock has been reset" Apr 23 10:30:24 i was already hitting a race with setting the date to the proper time and needed to add a minute if you mount and remount you will likely have to add sleeps Apr 23 10:30:41 or at least wait until mount is done Apr 23 10:30:41 iiuc, ext[2-4] fs filesystem integrity does not depend on the integrity of the clock. I think e2fsck has some clock-based heuristics for sorting out filesystem errors, but that's only an issue if the filesystem is actually broken Apr 23 10:31:06 right, the prob is that fsck doesnt handle the case properly Apr 23 10:31:14 dmart: but we should let upstream have the final say, of course Apr 23 10:31:30 * dmart is talking to himself now... Apr 23 10:31:46 heh Apr 23 10:39:14 dpkg-buildpackage: host architecture armel Apr 23 10:39:14 /usr/bin/fakeroot debian/rules clean Apr 23 10:39:14 debian/rules:39: *** atlas doesn't build on arm, not terminating on the buildd. Stop. Apr 23 10:39:14 dpkg-buildpackage: error: /usr/bin/fakeroot debian/rules clean gave error exit status 2 Apr 23 10:39:20 * ogra shakes his head Apr 23 10:39:55 i wonder if the packager every heard of the Yrchitecture field :P Apr 23 10:40:05 *Architecture Apr 23 10:40:46 ogra: can you take a look at http://pastebin.ubuntu.com/420963/ Apr 23 10:42:01 dmart, looks fine, tested already ? Apr 23 10:42:40 I have tested "by diff" -- replacing $(cat /proc/cmdline) with $(echo) and prefixing the date and hwclock commands with echo to check that they are run at the right times and with the expected arguments Apr 23 10:42:45 is that enough? Apr 23 10:43:22 otherwise, how to I set up the initramfs? Apr 23 10:43:26 well, i'd like to hear that a boot succeeds with different root= cmdline options on a system that exposes the issue Apr 23 10:43:42 the hook needs to go into /usr/share/initramfs-tools/hooks and the script into /usr/share/initramfs-tools/scripts/local-premount before calling update-initramfs -u (make sure both are executable) Apr 23 10:44:00 OK, I'll try it Apr 23 10:44:00 (and you need to set fixrtc on the cmdline indeed) Apr 23 10:44:05 yeah Apr 23 10:44:18 thanks, I probably would have forgotten ;) Apr 23 10:44:23 heh Apr 23 10:52:13 blaargh, guess what? Apr 23 10:53:27 (forgot fixrtc) Apr 23 10:57:48 lol Apr 23 11:22:06 ogra: hi there Apr 23 11:22:15 hey Apr 23 11:22:17 works ? Apr 23 11:22:20 I think it mostly works Apr 23 11:22:26 mostly ? Apr 23 11:22:32 But dumpe2fs and its libraries appear to be missing from the initramfs Apr 23 11:22:51 did you copy the hook in place before running update-initramfs -u ? Apr 23 11:23:09 needs to go to /usr/share/initramfs-tools/hooks and needs to be executable Apr 23 11:23:40 Ah, hold on a minute --- what is "the hook"? Apr 23 11:24:21 http://bazaar.launchpad.net/~ogra/initramfs-tools/initramfs-tools-fixrtc/revision/182 Apr 23 11:24:31 hooks/fixrtc Apr 23 11:25:11 ah, oops Apr 23 11:25:58 * ogra kicks the wiki Apr 23 11:26:45 right, updating the initramfs again... Apr 23 11:28:06 ah, better Apr 23 11:30:22 Well, we still have the issue that if the clock is only intermittently corrupted across boot, we always break it (by resetting it based on the last mount time) Apr 23 11:30:35 But since this is only for broken systems anyway, that's probably OK Apr 23 11:30:44 The clock should never be seen to go backwards Apr 23 11:30:55 And ntp should patch things up Apr 23 11:32:23 right Apr 23 11:32:38 and you wont need it at all if you never pull the power plug :) Apr 23 11:33:17 ok, if it works fine, please note that on the bug and point to your branch (and link the branch to the bug (top right there is an option)) Apr 23 11:37:18 When bzr committing, how to I get the committer name and email right? Apr 23 11:37:27 I'm working on a dev board, so those are garbage Apr 23 11:37:36 ...or will launchpad patch it up? Apr 23 11:39:02 bzr launchpad-login will help Apr 23 11:39:08 bzr whoami will do the rest of it Apr 23 11:39:42 LP won't get it right based on your ssh keys, for complex reasons (you might be pushing someone else's commits, etc.) Apr 23 11:53:57 ok, done Apr 23 11:54:06 * dmart 's ignorance has decreased Apr 23 11:54:22 (slightly) Apr 23 11:54:27 thanks for the pointers Apr 23 11:59:38 That's why we're here. Thanks for helping work on the bug. Apr 23 12:01:00 I'll know for next time... lp+bzr is still rather a mystery for me ;) Are there any good docs floating aroung? Apr 23 12:01:05 around Apr 23 12:01:48 ogra: done (in case you didn't see) Apr 23 12:01:55 i did Apr 23 12:01:58 Docs? Apr 23 12:02:04 * persia digs some up Apr 23 12:02:06 thanks a lot ! Apr 23 12:02:57 dmart: https://help.launchpad.net/Code isn't a bad place to start. Apr 23 12:03:30 cool, thanks Apr 23 12:08:04 https://wiki.ubuntu.com/ARM/Beagle Apr 23 12:08:08 finished :) Apr 23 12:08:20 anyone who wants, feel free to improve/enhance Apr 23 12:09:19 You know we have a Qt/plasma variant as well, right? Apr 23 12:09:54 Has anyone installed that? Does it work OK on the Beagle? Apr 23 12:10:01 no idea Apr 23 12:10:08 i dont think we build it Apr 23 12:10:29 quick, glance it looks good ogra... Apr 23 12:10:36 http://cdimage.ubuntu.com/kubuntu-netbook/ports/daily-live/current/lucid-netbook-armel+omap.img Apr 23 12:10:37 oh, we do build it Apr 23 12:10:46 yeah, just saw that Apr 23 12:10:53 no idea, i wont have the time Apr 23 12:11:07 s/have/take/ Apr 23 12:11:22 * persia tries to find out if it passed RC testing Apr 23 12:11:38 i'll put info that it exists in the blog post i'll write soon Apr 23 12:11:56 probably someone in the community wants to test it Apr 23 12:12:16 (if the KDE ubiquity doesnt OOM anyway that is :) ) Apr 23 12:12:19 Yeah, didn't get into the RC testing, so there's no record if anyone tried it. Apr 23 12:12:26 That's my thought. Apr 23 12:12:34 omap didnt get into the RC testing at all Apr 23 12:12:37 I know it works for dove/imx51 (it passed RC testing) Apr 23 12:12:45 But those have more RAM. Apr 23 12:13:00 its fully voluntary even for release for 10.04 Apr 23 12:13:41 I didn't think it was even schedued for formal release, but just that we'd have cdimages. Apr 23 12:13:48 right Apr 23 12:14:20 Anyway, I don't have a Beagle, so I can't know. Anyone with a Beagle up for seeing if that image boots? Apr 23 12:14:30 given there is no EFL launcher for kubuntu-netbook i really doubt it will work Apr 23 12:14:48 qt-ubiquity will probably run though Apr 23 12:15:14 plasma-netbook is designed to be lightweight though. Apr 23 12:15:27 not as light as efl though Apr 23 12:15:47 and even that is close to the edge on beagle Apr 23 12:15:55 * persia refuses to argue about toolkits Apr 23 12:16:11 heh, that wasnt my intention :) Apr 23 12:16:19 xaw3d xaw3d! Apr 23 12:16:32 yeah ! Apr 23 12:16:43 thats a 3d toolkit that works everywhere at least ! Apr 23 12:17:22 Isn't Beagle specs similar to n810? Apr 23 12:17:34 persia, well, kubuntu is still at 20.1 that wont work anyway Apr 23 12:17:41 beagle is more powerful than n810 Apr 23 12:17:44 it needs a build from 22 Apr 23 12:17:55 else the flash-kernel fixes are missing Apr 23 12:18:16 Oh. Apr 23 12:18:31 Are image builders still on manual? They should probably get turned on for the weekend. Apr 23 12:18:55 i think they might stay on manual Apr 23 12:19:00 ask steve Apr 23 12:23:43 lool: why'd you upload kexec-tools to ~canonical-arm-dev/marvell-dove-public? Apr 23 12:23:50 NCommander: for testing Apr 23 12:24:00 NCommander: Because I needed public armel builds before pushing to the arcihve Apr 23 12:24:06 NCommander: can be dropped now Apr 23 12:24:18 lool: fair enough. I just didn't expect to see it there Apr 23 12:24:19 NCommander: I checked with other people before doing it -- note that it affected dove too Apr 23 12:24:29 lool: not a problem, was just curious Apr 23 12:25:00 removed Apr 23 12:25:25 lool, https://wiki.ubuntu.com/ARM/Beagle Apr 23 12:25:32 feel free to test :) Apr 23 12:26:18 Favoriten Apr 23 12:26:19 lol Apr 23 12:26:24 Büro! Apr 23 12:26:38 yeah, i need to add an english screenie once i have an english install :) Apr 23 12:26:48 will do that before final :) Apr 23 12:26:49 Should be Software-Zentrum! Apr 23 12:26:54 LOL ! Apr 23 12:27:25 ogra: You want the bootloaders links I guess Apr 23 12:27:39 i havent written that yet Apr 23 12:27:55 you mean how to install uboot to nand etc, right ? Apr 23 12:28:02 or put MLO on your fat Apr 23 12:28:23 i'll do that over the weekend, for now i wanted to have the general pages ready Apr 23 12:28:37 since i want to blog it to get some attention Apr 23 12:28:38 ogra: I mean links to working bootloaders Apr 23 12:28:44 ogra: how to flash them etc. Apr 23 12:28:49 yes Apr 23 12:28:54 ogra: because people might be stuck if they don't load boot.scr Apr 23 12:28:55 havent got that yet Apr 23 12:29:18 its all on the elinux wiki and i assume beagle users know that Apr 23 12:29:29 they will survive over the weekend :) Apr 23 12:29:39 or come here and complain ;) Apr 23 12:29:49 lool, boot.scr's are pretty much the default on the beagle elinux land .. ;) Apr 23 12:30:24 ogra: netbook install points at server image Apr 23 12:30:30 "Get the lucid-server-armel+omap.img image" Apr 23 12:30:41 rcn-ee, but you could have changed your uboot setup (which is why i have the "how to load a boot.scr from serial prompt" part in all the installation pages) Apr 23 12:30:53 fixed Apr 23 12:31:02 thanks :) Apr 23 12:31:03 oops Apr 23 12:31:12 luckily the link was ok Apr 23 12:31:23 yeap.. as long as you go over how to flash u-boot and change things, that'll cover those issues... Apr 23 12:31:35 right Apr 23 12:31:46 i still have a week to shake out all the docs Apr 23 12:31:55 ogra: Ideally, but together a SD card image with just x-loader + u-boot and a default config which just flashes the board Apr 23 12:32:09 So that people basically boot with it, and it flashes their NAND as appropriate Apr 23 12:32:12 thats just a start, i'm sure questions will come up that show wheer we need additional documentation Apr 23 12:32:23 lool, like https://code.launchpad.net/~beagleboard-kernel/+junk/omap-flasher ;) Apr 23 12:32:29 haha Apr 23 12:32:44 rcn-ee, i'll link that on the flasher page :) Apr 23 12:33:02 ogra: https://wiki.ubuntu.com/ARM/BeagleEditBootscr is awful Apr 23 12:33:05 oh great... i got tired of doing the commands my self.. so i scripted it.. Apr 23 12:33:17 ogra: can't we just ship boot.script in the image? Apr 23 12:33:17 lool, what dont you like about it ? Apr 23 12:33:23 the dd Apr 23 12:33:29 lool, i wont change debian-cd anymore now Apr 23 12:33:50 yeah the dd is different.. i never thought about doing that.. but it works... Apr 23 12:33:51 we can ship it in 10.++ Apr 23 12:34:24 lool, there is a lot of awful stuff about that image that needs to improve Apr 23 12:34:31 its a first hit Apr 23 12:34:44 created in a week Apr 23 12:34:54 dont expect shiny :) Apr 23 12:35:33 Rather, expect shiny, because that's interface, and shared, but don't peer under the sleek shiny cover too much. Apr 23 12:35:34 It look's fine and will work for most people... the others always do what they want anyways.. Apr 23 12:36:08 rcn-ee, there are a lot of things i could have done better and i know lool will die if he looks into details :) Apr 23 12:36:24 but there is always 10.++ Apr 23 12:36:33 which will fix that Apr 23 12:36:53 laughs, i just hope he doesn't look into my Netinstall hack.. Probally cringe and never look at any of patches.. ;) Apr 23 13:04:59 and we'll have the XM for 10.10 ;) Apr 23 13:05:37 Don't get too excited: that just means we have that much more work to do to create a boot script that can somehow boot on more platforms. Apr 23 13:07:05 well my script, currently boots Bx, Cx and my proto Xm.. ;) if only the kenrel didn't bomb on the xm... I'm hoping enough kernel bits for the Xm hit mainline that omap can share the ubuntu kernel.. Apr 23 13:07:55 Well, there's currently a special omap kernel in Ubuntu, but yeah, extra points for getting into the "normal" ports kernels. Apr 23 13:09:24 just for reference, the XM uses a new omap3 similar core, DM3730 but the kernel bits are in a ti staging repo... Apr 23 13:09:35 rcn-ee: public? Apr 23 13:10:07 yeap.. as i dig for the link... Apr 23 13:10:24 rcn-ee: it's good Apr 23 13:10:27 just knowing it's public is enough Apr 23 13:10:28 http://arago-project.org/git/people/sriram/ti-psp-omap.git?p=people/sriram/ti-psp-omap.git Apr 23 13:11:40 *okay heads to work, but will be back, i enjoy half fridays.. ;) Apr 23 13:22:38 http://ograblog.wordpress.com/2010/04/23/unleash-the-beagles/ Apr 23 13:23:00 * ogra had some issues to get the utf-8 right on lool's name :) Apr 23 13:52:20 rcn-ee: Do you know what comes with the digikey beagleboard? cant figure whether it has power adapter Apr 23 14:04:23 ndec, http://ograblog.wordpress.com/2010/04/23/unleash-the-beagles/ Apr 23 14:30:27 hello ogra Apr 23 14:30:37 gduarte, hey Apr 23 14:31:31 :D ogra, do you know where i can download (if exist) a pre-compiled qemu-arm image to use directly with qemu Apr 23 14:31:34 ? Apr 23 14:31:57 I'm trying for 3 days without success Apr 23 14:32:10 nope, but if you follow the https://wiki.ubuntu.com/ARM/RootfsFromScratch page you can create one Apr 23 14:32:20 yep Apr 23 14:32:37 I've tried it, no success Apr 23 14:32:46 what was the issue ? Apr 23 14:32:49 I'm trying it right now http://people.canonical.com/~ogra/arm/qemu/README Apr 23 14:33:05 oh, thats ancient Apr 23 14:33:08 but don't know if it is the same thing Apr 23 14:33:09 dont use that Apr 23 14:33:16 ahhh Apr 23 14:33:25 * ogra needs to clean up his homedir Apr 23 14:33:44 what did not work with https://wiki.ubuntu.com/ARM/RootfsFromScratch ? Apr 23 14:33:56 my problem is when I try to log in into my emulated system Apr 23 14:34:10 i type my login, but it's no recognized Apr 23 14:34:16 never Apr 23 14:34:33 and followed by filesystem issues Apr 23 14:34:37 what is your host system ? Apr 23 14:34:56 Ubuntu 9.10 amd64 Apr 23 14:35:04 hmm, ok Apr 23 14:35:23 you can install qemu-arm-static there and chroot into your image to fix it up Apr 23 14:35:53 ok :D I'll be able to compile my programs there? Apr 23 14:36:09 yes, but it is as slow as a VM Apr 23 14:36:40 no problem ;) thank you ogra Apr 23 14:37:04 after you installed qemu-arm-static, loop mount the image and cp the qemu-arm-static binary into /usr/bin/ Apr 23 14:37:12 then you can just chroot Apr 23 14:37:34 ok ok, I'll do it Apr 23 14:40:42 GrueMaster: Did you get a chance to test kubuntu-netbook on imx51/dove? Apr 23 14:41:13 I'm just now loading it on imx51. Already hit a few minor issues. Apr 23 14:44:09 ogra: plars: GrueMaster: persia: NCommander: https://wiki.ubuntu.com/MobileTeam/ReleaseStatus/Lucid Apr 23 14:44:12 dyfet: ^ Apr 23 14:44:15 OK. freeflying and Riddell were wondering if it got tested in #kubuntu-devel, and I misread "netboot" as "netbook" on the kubuntu page before, so had the wrong information :( Apr 23 14:44:21 maybe check out if i missed something in the summary that you want to have there Apr 23 14:44:21 ian_brasil : are a ARM developer from Brasil? Apr 23 14:44:21 **are you Apr 23 14:44:39 asac: looks good. Apr 23 14:44:41 fpc moved to SRU :( Apr 23 14:44:48 well i am from Brasil :) Apr 23 14:44:49 persia: sorry, I didn't get to that >.<; Apr 23 14:45:16 well strictly speaking I am not Brazilian but i live in Brazil Apr 23 14:45:20 persia: thats a really old bug. no progress. if we had a fix today we can resurrect it for final maybe Apr 23 14:45:25 NCommander: No worries. THe issue is that it only doesn't work for the folks that are able to bootstrap. Works for the folks that wanted it to happen :) Apr 23 14:45:51 ah, he left Apr 23 14:45:53 asac: It's an old bug because doko decided to hijack the powerpc bootstrap bug to bootstrap armel. Apr 23 14:46:11 (not that I'm complaining, because it made powerpc get bootstrapped, but ... ) Apr 23 14:46:47 asac, looks all good to me Apr 23 14:46:53 asac: You might want to beg again for bug #538736 if you don't like the current answer Apr 23 14:46:55 Launchpad bug 538736 in vm "Filename encoding for external viewer inconsistent (affects: 1) (heat: 6)" [Undecided,Incomplete] https://launchpad.net/bugs/538736 Apr 23 14:47:02 NO. Apr 23 14:47:16 bug #568736 Apr 23 14:47:18 Launchpad bug 568736 in netbook-meta (Ubuntu Lucid) (and 1 other project) "Having Evolution installed along with Desktop-Email is pointlessly redundant (affects: 1) (heat: 12)" [High,Won't fix] https://launchpad.net/bugs/568736 Apr 23 14:47:37 asac, slangasek promised to review bug 563618 right after the meeting, we should make sure to tickle him so he doesnt forget again :) Apr 23 14:47:39 Launchpad bug 563618 in util-linux (Ubuntu Lucid) (and 5 other projects) "Ignoring a broken clock results in infinite reboots; not ignoring results in fsck failure; no solution to this problem (affects: 1) (heat: 14)" [High,Triaged] https://launchpad.net/bugs/563618 Apr 23 14:51:51 ogra: its on the list Apr 23 14:51:58 ogra: e.g. on the list of still final targetted bugs Apr 23 14:52:23 yeah, saw that Apr 23 15:07:38 rcn-ee, poke poek Apr 23 15:07:43 question for ya :) Apr 23 15:08:36 http://bazaar.launchpad.net/~beagleboard-kernel/%2Bjunk/2.6-stable/revision/57 Apr 23 15:08:46 which of your kernels is that patch in? Apr 23 15:14:27 ogra: sorry for annoying you again Apr 23 15:14:28 but Apr 23 15:14:38 shoot Apr 23 15:14:45 I'm having this problem Apr 23 15:15:01 I: Creating temporary Image Apr 23 15:15:02 I: Mounting temporary Image Apr 23 15:15:02 I: Running first stage Apr 23 15:15:02 sudo: unable to cache gid, already exists Apr 23 15:15:02 I: First stage install done Apr 23 15:15:02 /usr/bin/rootstock: 572: cannot create /tmp/tmp.x3pBJdpxDE/tmpmount/etc/fstab: Directory nonexistent Apr 23 15:15:19 in another machine Apr 23 15:15:39 can you pastebin the logfile somewhere ? Apr 23 15:15:53 it should have cerated one on exit and tell you where that lives Apr 23 15:16:43 ok Apr 23 15:17:36 and the version of rootstock to use as well as the complete command you call to run it would also help Apr 23 15:17:37 it's not creating a log file Apr 23 15:17:47 s/to use/you use/ Apr 23 15:18:29 sudo rootstock --fqdn ubuntu-arm-dev --login ubuntu --password ubuntu --imagesize 2G --seed xubuntu-desktop Apr 23 15:18:35 this is the command Apr 23 15:18:46 that should be fine Apr 23 15:19:06 could it be that your disk is full ? Apr 23 15:19:08 yes, but isn't working Apr 23 15:19:15 no, i'ts not full Apr 23 15:19:40 gduarte, do you have a million loop mounts showing up if you type "mount"? Apr 23 15:19:53 right, that was my next question Apr 23 15:19:58 I'll check Apr 23 15:20:38 no, no loop devices mounted Apr 23 15:21:17 and its directly going from "I: Running first stage" to "I: First stage install done" just with that one line inbetween ? Apr 23 15:22:50 no Apr 23 15:23:09 I'ill past at pastbin Apr 23 15:23:17 I'll* Apr 23 15:24:08 **paste Apr 23 15:26:09 http://pastebin.com/Su3gmYEB Apr 23 15:26:13 there Apr 23 15:30:10 saw? Apr 23 15:30:14 gduarte, so how did you install rootstock Apr 23 15:30:24 apt-get Apr 23 15:30:34 sudo apt-get install rootstock Apr 23 15:30:41 hmm, weird Apr 23 15:30:44 this way Apr 23 15:32:15 :( Apr 23 15:32:57 is debootstrap installed ? Apr 23 15:33:06 i dont get why it doesnt seem to run it at all Apr 23 15:33:40 ye, it's installed, is downloaded together with rootstock when run apt-get Apr 23 15:33:48 right Apr 23 15:33:57 but it isnt executed at all Apr 23 15:34:03 thast very weird Apr 23 15:34:09 *that's Apr 23 15:34:47 unlucky Apr 23 15:35:01 qemu-kvm is installed as well ? Apr 23 15:35:53 yes, it's also installed together with rootstock Apr 23 15:36:29 right, and qemu-img seems to run fine since it shows I: Mounting temporary Image Apr 23 15:37:00 but your first stage does definately not run Apr 23 15:37:11 (which is debootstrap) Apr 23 15:37:19 :S Apr 23 15:37:46 I really need it, because I need to compile and package to ARM Apr 23 15:38:00 might be an amd64 issue in karmic Apr 23 15:38:33 it's my exam to be hired to canonical :( Apr 23 15:39:32 try the following: "sudo debootsrap --arch=i386 lucid $HOME/lucid-i386" that will give you a lucid i386 environment Apr 23 15:40:00 ok Apr 23 15:40:01 chroot into that and you shopuld be able to install rootstock and run it under i386 Apr 23 15:41:46 it's retrieving packages Apr 23 15:41:50 yeah Apr 23 15:42:43 would be nice if would exist a default image to everybody download and modify Apr 23 15:43:08 we might provide something like that in maverick Apr 23 15:43:18 though the tools got a lot better in lucid already Apr 23 15:43:31 i know there were qemu related issues in karmic Apr 23 15:43:38 on amd64 Apr 23 15:43:55 :( Apr 23 15:44:16 hey all... I was thinking about registering a blueprint for a Mobile Ubuntu user interface (mo-buntu). The main idea would be creating a better mobile experience for "phone like" devices. Does anyone know if something like this is in the works for lucid? -- any feedback? Apr 23 15:45:07 orbarron, there is some QT based work ian_brasil works on iirc (not sure how much phone based that is) Apr 23 15:45:39 might be somewhere between netbook and phone rather Apr 23 15:45:57 beyond that i dont know whats planned for UDS Apr 23 15:46:04 we will integrate ophono Apr 23 15:46:10 so it will be phone based Apr 23 15:47:41 obarron..look at https://blueprints.launchpad.net/ubuntu/+spec/mobile-liquid Apr 23 15:47:51 ahh thanks... Apr 23 15:49:57 orbarron: there are likely to be some discussions at UDS about this - how to sanely break down kdelibs mainly if you are interested Apr 23 15:51:24 ian_brasil: ofono :) Apr 23 15:51:31 haha Apr 23 15:51:41 * ogra read that typo somewhere today already Apr 23 15:51:48 yep, asac Apr 23 15:53:34 ian_brasil: cool. can we have a call on your ofono experiences etc.? Apr 23 15:54:23 not today. but next week? Apr 23 15:54:36 or we can chat, but my hands are getting lamer every day i continue to do that ;) Apr 23 15:54:50 asac, thats the raising age :) Apr 23 15:55:25 (i wished you a happy b-day this morning in case you havent seen it) Apr 23 15:55:43 ogra: i saw ... i tried to ignore it for obvious reasons. thanks for spreading it to the world ;) Apr 23 15:55:52 lol Apr 23 15:56:05 but its over soon enough ;) Apr 23 15:56:11 asac: let me ping some people then Apr 23 15:56:14 and /me will get off now for a bit anyway Apr 23 15:56:22 ian_brasil: cool. lets talk on monday? Apr 23 15:56:25 yeah, enjoy your evening Apr 23 15:56:26 (i am mostly out) Apr 23 15:56:28 thanks!!!! Apr 23 15:56:33 and stay away from the kbd ! Apr 23 15:56:38 i will i will Apr 23 15:56:46 ogra: remember to submit your blueprints against ubuntu-arm ;) Apr 23 15:57:07 if i could rememebr the blueprints i had done it already ... i need to look them up Apr 23 15:57:09 asac: cool Apr 23 15:57:15 but i'm waiting for slangasek anyway Apr 23 16:07:22 ogra: I got it!!!!! Apr 23 16:07:32 wonderful Apr 23 16:08:21 it's basic, no gcc, no X, some few FS issues but works Apr 23 16:08:32 good Apr 23 16:08:37 I can make it avaliable to others? Apr 23 16:08:56 you can do what you like with it :) Apr 23 16:09:01 i know Apr 23 16:09:35 but, to other that do not want to download and buid it, get a pre-compiled image Apr 23 16:09:48 we could make it avaliable at that page Apr 23 16:10:52 sure, if you have space to host it you can make it available Apr 23 16:13:16 ok Apr 23 16:14:07 I'll improve that, install gcc and some other dev stuff and make it avaliable if i could Apr 23 16:14:19 gret Apr 23 16:14:22 *great even Apr 23 16:14:23 :D Apr 23 16:16:24 :D Apr 23 16:16:40 thank you for helping me!!!! Apr 23 16:17:09 welcome, if you have other issues, dont hesitate to ask here (even if i'm not around there are surely people who can help) Apr 23 16:18:16 surely ;) Apr 23 16:19:09 thank you again! Apr 23 16:30:42 http://www.bit-tech.net/news/hardware/2010/04/23/samsung-plans-own-netbook-chips/1 Apr 23 16:52:40 hi, I need to add FireWire support to the Ubuntu-ARM linux kernel. do i must recompile the whole kernel for that ? Apr 23 17:02:43 No doubt this has been asked before so apologies in advance. I am looking at the arm9 and wondering if Ubuntu is available for it. Apr 23 17:17:06 Anybody there? Apr 23 17:23:04 tractor, Apr 23 17:23:21 http://wiki.ubuntu.com/ Apr 23 17:23:37 Excellent, thanks. Let me check that. Apr 23 17:23:50 err.. Apr 23 17:24:02 wth is wrong with my clipboard -.- Apr 23 17:24:04 I only know it is an arm9 processor, I don't have any further details than that unfortunately Apr 23 17:24:26 I meant https://wiki.ubuntu.com/ARM Apr 23 17:24:47 Oh, I see, thanks for the correction Apr 23 17:25:17 Lucid probably doesn't work on arm9 machines Apr 23 17:25:45 hmm, not Karmic either is seems Apr 23 17:26:21 I have error while loading shared libraries libraw1394.so.8: cannot open shared object file problem ... though all my lib files are linked correctly Apr 23 17:26:32 any ideals? Apr 23 17:29:13 should I link them to usr/include ? Apr 23 17:34:13 Thanks Meizirkki, just looking through it now. Apr 23 18:22:42 cwillu_at_work, pong.... rev 57 would be anything 2.6.32.10-l10.0 ++++ Apr 23 18:23:11 and in the karmic kernels as well? Apr 23 18:23:47 [ 1.412597] omapfb omapfb: no displays Apr 23 18:23:47 [ 1.416442] omapfb omapfb: failed to setup omapfb Apr 23 18:23:47 [ 1.421203] omapfb: probe of omapfb failed with error -22 Apr 23 18:24:26 yeah, exactly... one user said it worked on their overo... looks like it's broken... Apr 23 18:25:30 does it have an associated config item? I'd like to prove to myself that it's actually included Apr 23 18:26:39 it relies on a combined defconfig, so the same defconfig as the beagle.. Apr 23 18:28:39 cwillu_at_work, i'd actually test the lasted in the 2.6.32 series, as i had to tweak that patch in rev 65 Apr 23 18:29:11 which would be 2.6.32.11-x13 Apr 23 18:29:45 http://rcn-ee.net/deb/kernel/beagle/karmic/v2.6.32.11-x13/ Apr 23 18:30:00 exactly.. Apr 23 18:31:00 in that diff, look for: diff --git a/arch/arm/mach-omap2/board-overo.c b/arch/arm/mach-omap2/board-overo.c is the dss2 patch Apr 23 18:31:32 basicly pulled from sakoman's overo tree Apr 23 18:42:27 I see a penguin Apr 23 18:48:46 should I see that change in http://rcn-ee.net/deb/karmic/v2.6.33.2-d8/patch-*.diff, for example? Apr 23 18:55:45 Nope, two different patchsets... 2.6.33 is still only in my 2.6-dev tree, (that's what the 'd' is) I've been stress testing it on my gcc test farm, I'm thinking of moving 2.6.33.3 to stable after esc... Apr 23 18:56:00 ah, that's my problem then Apr 23 18:56:17 I'm on 2.6.33 as btrfs is significant more stable there Apr 23 18:57:18 so now that I know what to beg for Apr 23 18:57:32 * cwillu_at_work begs for a 2.6.33 with the overo dss2 patches :p Apr 23 18:57:42 I could just build it myself, but where's the fun in that? Apr 23 18:57:54 actually it's rcn-ee, get your ass in gear and move 2.6.33 to stable. ;) Apr 23 19:17:57 rcn-ee, indeed, and while you're at it, start getting 2.6.34 ready too :p\ Apr 23 19:18:44 hehe.. actually it's in pretty good shape... (other then dss2 on overo..) ;) Apr 23 19:19:09 yep, I've been on .33 for a long'ish while now Apr 23 19:19:39 i know.. i moved my x86's to 2.6.34-rc's as 2.6.33 just fells too old.. ;) Apr 23 19:19:47 just got some old overo boards back that I want to get running more up-to-date images, and boom, no video :p Apr 23 19:20:21 I'm anxiously waiting 2.6.35, as there's a bunch of btrfs stability stuff that should be landing there Apr 23 20:02:41 hello Apr 23 20:03:12 I have a mobile phone with a ARM processor. Is there any way I can find out if it supports ubuntu? Apr 23 20:03:59 EvaLuaTe, maybe... which phone? Apr 23 20:04:34 rcn-ee: it's an LG GT500. It's originale OS sucks, so I'm just looking to see if there's an alternative :p Apr 23 20:06:09 can I compile the ubuntu arm kernel on the beagleboard ? Apr 23 20:06:51 samuel_Sayag, yes i do it all the time... with alll modules 5 hours.. ;) Apr 23 20:07:15 okay ... can you plase help me :) rcn-ee Apr 23 20:07:30 I'm looking for the right kernel to download Apr 23 20:08:07 define: right? are you looking for a specific feature? or just something that works? Apr 23 20:08:12 due the fact that the offical http://www.kernel.org/ doesnot get compiled :( Apr 23 20:08:36 I just need a kernel with firewire support Apr 23 20:09:06 to do it I need to compile a preprepared kernel for the arm ( i guess ) Apr 23 20:09:14 2.6.32 works just fine if you build omap3_beagleboard_defconfig.. Apr 23 20:09:27 dss2 support needs patches... Apr 23 20:10:18 okay ... how do i begin ? how can i find the first foot grip in all this ? Apr 23 20:10:34 btw... how are you adding firewire to the beagle? Apr 23 20:10:41 thank you very much by the way rcn-ee Apr 23 20:11:36 I don't, I have a wird cemra interface that need it Apr 23 20:11:37 http://tldp.org/HOWTO/libdc1394-HOWTO/install.html Apr 23 20:12:13 the cemra name is FireFly MV by PG Apr 23 20:12:30 very nice pice of hardware Apr 23 20:12:44 interesting... well the problem, the beagle only has usb2.0, that's why i ask... but to build something that just works, take a look at: https://code.launchpad.net/~beagleboard-kernel/+junk/2.6-stable it's my stable branch with scripts to build a stable workign kernel.. Apr 23 20:13:48 I been there :) (BTW) I didn't find how to add this flags 1. Apr 23 20:13:49 OHCI-1394 support Apr 23 20:13:49 2. Apr 23 20:13:49 OHCI-1394 Video Support Apr 23 20:13:49 3. Apr 23 20:13:51 OHCI-1394 DVI/O Support Apr 23 20:13:53 4. Apr 23 20:13:55 RAW IEEE1394 I/O Support Apr 23 20:14:28 they require hardware firewire... the beagle doesn't have it, so it never shows up on the menuconfig.. Apr 23 20:14:53 :( Apr 23 20:15:14 So I'm just lost :( Apr 23 20:15:21 rcn-ee: so, do you have any idea if my device might support ubuntu? :) I've read that it sports a 93 Mhz ARM processor with 128MB or RAM memory btw... Apr 23 20:16:51 EvaLuaTe, i've been searching for it.. I think it's an armv5 core which woudl mean jaunty only... it's kinda too slow, you might not like the experience.. Apr 23 20:17:40 hi all Apr 23 20:17:46 ohh. Seeing as the 'reflash' with the original firmware isn't a pain in the ass though, I might think about trying it out. Is it hard to install? Apr 23 20:18:27 I'm trying to get an arm ubuntu image working thanks to the rootstock script Apr 23 20:18:49 EvaLuaTe, it's not so much as reflash the rom with an ubuntu install, you need a Kernel for your device, the blob for your phone interface and tehn the rootfs... Apr 23 20:18:55 it works well with default settings, but I can't get the --seed option to work correctly Apr 23 20:19:12 alextisserant, what are you sending to '--seed' Apr 23 20:19:15 (with lucid dist) Apr 23 20:19:26 a file with a list of packages Apr 23 20:19:48 at first I had a bunch of them, but now I'm just trying with mousepad Apr 23 20:19:51 rcn-ee: so, if there isn't a kernel for my device, i'd have to compile one myself, right? also, I've no idea what a blob is :p Apr 23 20:20:14 then I do --seed `cat packages | tr '\n' ',' Apr 23 20:20:52 so actually the script is just "stuck" while "Unpacking ttf-dejavu-core" Apr 23 20:21:17 when I had my long list of additional packages, it was actually stuck on Unpacking libc6-dev Apr 23 20:21:29 EvaLuaTe, exactly, and a lot of these 3rd party phones it's almost impossible to find... Apr 23 20:21:46 my CPU is still working, but after a whole night of unpacking, I guess this is not totally normal :-) Apr 23 20:22:01 alextisserant, lucid? Apr 23 20:22:04 so the script does not die, but it just seems to loop on this unpacking... Apr 23 20:22:06 yes Apr 23 20:22:22 welcome to bug 56639 Apr 23 20:22:24 Launchpad bug 56639 in acpi-support (Ubuntu) "Thinkpad R40 Fn-F7 should switch active display in software" [Undecided,Invalid] https://launchpad.net/bugs/56639 Apr 23 20:22:30 crap, bug 566639 Apr 23 20:22:31 Launchpad bug 566639 in apt-setup (Ubuntu) "omap install ends up with security.ubuntu.com urls in sources.list after install (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/566639 Apr 23 20:22:43 hmm Apr 23 20:22:52 I give up: this bug: https://bugs.launchpad.net/ubuntu/+source/qemu-kvm/+bug/532733 Apr 23 20:22:58 Launchpad bug 532733 in qemu-kvm (Ubuntu Lucid) (and 2 other projects) "apt/dpkg in qemu-system-arm hangs if a big task is installed (affects: 4) (dups: 1) (heat: 34)" [High,Incomplete] Apr 23 20:23:01 rcn-ee: ohh, ok then. You know any other nice OS for such a mobile device that might be easier to install/try? Apr 23 20:29:07 rcn-ee: did anyone try with other versions of qemu? Apr 23 20:30:39 I've tried Debian Squeeze & Sid's... Apr 23 20:31:05 Dustin and Loci are the ubuntu qemu guys.. Apr 23 20:31:33 ok Apr 23 20:32:45 It's one of those disapointing things... If you have a beagle, the best method to build an SD card is the NetInstall.. I'm planning to make that work for the overo and igepv2, but after ESC... Apr 23 20:35:17 yes, but I'm trying to get a ready-to-use final image for the Touch Book Apr 23 20:35:48 at least I can indeed do a netinst and pack the image afterwards Apr 23 20:37:27 okay, I see my camera under the lsusb :) Apr 23 20:37:52 alextisserant, that's exactly what i'm planning for my x11 images on rcn-ee.net for the beagle.. build native, copy... Apr 23 20:38:40 yup Apr 23 20:38:46 just need more time :) Apr 23 20:38:48 But still when I'm trying to take a photo I get "err libdc1394: Failed to initialize libdc1394" odd... Apr 24 00:03:19 Weeeee Apr 24 00:03:20 Hi! Apr 24 00:03:30 Ive got my hands on an old Skye SL Apr 24 00:03:36 Now big question Apr 24 00:03:41 How do i get ubuntu on it? Apr 24 00:04:33 I presume you don't mean http://brickwell.com/product/10-trek-skye-sl-57861-1.htm : do you have a link to specs? Apr 24 00:05:12 Somehow, I expect not Apr 24 00:05:20 Yepp, its a rather rare device. I found some info and also tried out his Linux Apr 24 00:05:25 it would be interesting to try to install ubuntu on gears on a device Apr 24 00:05:37 Ubuntu, bike edition Apr 24 00:05:41 http://scholbert.homelinux.org/SkeyePad.html Apr 24 00:05:44 there you go Apr 24 00:05:50 maybe its too classic Apr 24 00:05:54 but it would be just so cool Apr 24 00:06:05 that WinCE 3.0 is giving me shivers Apr 24 00:06:10 "skeye" is quite different Apr 24 00:06:13 checking Apr 24 00:06:25 no, ARM on that one is too old. You might be able to get debian on it Apr 24 00:06:53 Debian will run fine. Apr 24 00:07:03 It's an SA1100 .. should run Debian fine Apr 24 00:07:04 Jaunty might even install, but would run vey slowly. Apr 24 00:07:16 Jaunty won't install .. its an arm v4 Apr 24 00:07:30 Oh, right. Apr 24 00:07:52 and the memory map and driver map for that device is cwazy Apr 24 00:08:13 * persia is often confused by the various extensions in the SA1110 and needs to hardwire the brain to say "Cannot run Ubuntu" Apr 24 00:08:17 http://scholbert.homelinux.org/SkeyePad_stuff.html Apr 24 00:08:22 OOOOllld hardware Apr 24 00:09:08 aye Apr 24 00:09:18 but still a very robust build Apr 24 00:09:21 Gestahlt: So, you should have good kernel support, but you likely have to make your own kernel. Getting it to boot directly (rather than chain-boot out of WinCE) may be a bit tricky. Apr 24 00:09:21 i want to pimp it up Apr 24 00:09:41 I hate to say it, but it won't be very "pimp" able Apr 24 00:09:44 it has very little memory Apr 24 00:09:52 And when I say very little .. I mean a TINY amount of RAM Apr 24 00:10:07 Samsung SDRAM 2x32MB (K4S561632-TC75) @ 103MHz Apr 24 00:10:11 Computers wont need ever more than 640kb of RAM Apr 24 00:10:19 Isn't that 64MB? Apr 24 00:10:27 Should be fine, for careful use. Apr 24 00:10:34 Actually Apr 24 00:10:44 Worse, it has even less -storage- Apr 24 00:10:44 Intel NOR Flash 2x16MB (28F128J3A-150) Apr 24 00:10:46 i need it for RDP / VNC Streaming and Tux Paint Apr 24 00:10:50 32MB of storage. That's it ... Apr 24 00:10:52 and maybe seamonkey Apr 24 00:10:56 Well Apr 24 00:11:01 there is a CF card Apr 24 00:11:02 getting a basic linux install squeezed in there, is going to be a bitch Apr 24 00:11:06 Zaurus has 64MB, and later Zauri were known to run Jaunty. Apr 24 00:11:15 That said, I wouldn't try to compile boost on it :) Apr 24 00:11:21 Yes, Zaurus had more storage though Apr 24 00:11:40 Depends on the model. SLC-1000 has no NOR. Apr 24 00:11:46 Err, -3000. Apr 24 00:12:05 3000 *ONLY* had CF storage. Mind you, two CF, one pre-installed with a microdrive, but still. Apr 24 00:12:07 That's because it had a -hard disk- Apr 24 00:12:11 yeah Apr 24 00:12:12 So? Apr 24 00:12:27 Stick a microdrive or a bundle of flash in the CF slot, and who cares. Apr 24 00:12:37 CF-ATA is all sorts of well supported. Apr 24 00:12:47 true Apr 24 00:12:49 could be interesting Apr 24 00:13:56 Gestahlt: But, yeah, your first steps are figuring out how to build the right kernel, and getting it to boot that kernel. Once you can do that, putting a Debian 5.0 filesystem on the CF card shuold be trivial. Apr 24 00:14:11 Zaurus は常に(今まで)非常にしっかり造られたハードウェアでした Apr 24 00:14:40 * Martyn mangles japanese with the best of 'em Apr 24 00:14:40 Persia, that scholbert guy already build an running kernel (2.4.x but still better than ce 3.0) with GPE Apr 24 00:14:43 In the beginning, this was true. Not so much these days :) Apr 24 00:14:58 Gestahlt: You really want a 2.6 kernel these days :) Apr 24 00:15:06 Probably at least > 2.6.13, really. Apr 24 00:15:08 I know... Apr 24 00:27:52 persia: none of the zaurus had nor, all of them have nand Apr 24 00:30:15 Ah, right. My mistake. Apr 24 00:30:29 3000 still didn't have either :) Apr 24 00:30:38 persia: 3000 had 128M of NAND Apr 24 00:30:43 persia: so your still wrong :-) Apr 24 00:30:46 What? Apr 24 00:30:54 all zaurus had NAND Apr 24 00:31:05 I thought that was precisely the difference between the 3000 and the 3100, and specifically didn't get a 3000 because of this understanding. Apr 24 00:31:21 It was reputed to boot slow because of this. Apr 24 00:32:08 it only had 16M Apr 24 00:32:16 OK. That I can believe. Apr 24 00:32:23 3100 had 128M Apr 24 00:32:35 but the rom and bootloader was always in NAND Apr 24 00:32:55 Ah, right. So 3000 had kernel in NAND, and full FS on CF-ATA, whereas 3100 had the base FS in NAND also. Apr 24 00:32:58 * XorA|gone has spent too many years of life hacking on openzaurus Apr 24 00:33:31 * persia was only ever a Zaurus user, and no longer (the 3100 died, and the Netwalker has left the 3200 to gather dust) Apr 24 00:33:51 persia: yeah something like that, couldnt tell you the exact layout of a 3000 as I never owned one, but its flashing script is identicle to 3100 Apr 24 00:33:52 Earlier models have just given up (no longer charge or boot) Apr 24 00:34:41 Makes sense, and simplifies internals to have the 16M for booting. Apr 24 00:34:54 (that said, I'm still happy to have never been a 3000 owner) Apr 24 00:35:41 Ive had 5500, 5600, 6000, c860, 3200 Apr 24 00:35:55 only got 5500 left Apr 24 00:36:30 I've had J1M1,760,860,3100,3200, and only the 3200 still works. Apr 24 00:36:48 J1M1 isn't directly comparable, of course, running an entirely different OS :) Apr 24 00:36:59 anyway Im off to bed its well late here :-) Apr 24 00:37:07 Sleep well :) Apr 24 00:55:17 * NCommander needs Marvell :-/ Apr 24 02:10:01 ogra_cmpc: BTW, have you ever caught a stray cat before? Apr 24 02:10:28 * NCommander has one that recently showed up in town, and I'd like to trap it, get it neutered, and release/put up for adaption **** ENDING LOGGING AT Sat Apr 24 02:59:56 2010