[00:30] <garyhbaker> Have been trying various Netinstalls on BeagleBoard C4 w/Zippy.  Successful in past several times.  Today fails "Unable to install busybox-initramfs. Fails consistently.  Any ideas, or is something newly broken....
[01:11] <rcn-ee_lpt> garyhbaker, that can happen if it's not finding the root partition..
[01:12] <rcn-ee_lpt> are you using special bootargs, root on mmc or root on usb ?
[03:51] <garyhbaker> rcn-ee_lpt: I am using no special bootargs and root on mmc...
[03:52] <rcn-ee> garyhbaker, can you pastebin your serial boot log for me to take a look at...... (and in uboot type 'printenv')...
[04:25] <NCommander> ndec: ping?
[08:11] <hrw> moin
[09:14] <hrw> ogra: how goes thumb2 on omap3/4? reliable works?
[09:14] <te_> Hi, is there some repository problem currently? trying to run the rootstock script, but get the following warning: "Couldn't download package lsb-base"
[09:15] <ogra> hrw, seen no issues here
[09:16] <XorA> hrw: no issue here with ubuntu or angstrom with thumb2
[09:18] <hrw> thx guys
[10:16] <ndec> NCommander: pong
[10:25]  * XorA has screwed up his sudoers file
[10:58] <ndec> lool: hi, i have a question, maybe you can help: what do all the XB-Gstreamer-xxx fields mean ? I am making a new package for a gst plugin...
[11:12] <ogra_cmpc> XorA, use visudo thats why it exists :)
[11:14] <hrw> or in worst situation reboot to /bin/bash
[11:14] <ogra_cmpc> well, with single user mode rather
[11:14] <ogra_cmpc> (the visudo comment wasnt about fixing, just about never breaking it ;) )
[11:15] <XorA> heh, lucky with these embedded system just pull out SD card, fix on desktop and put it back in :-)
[11:17] <hrw> ogra: single user mode asks for password
[11:20] <ogra_cmpc> hrw, it doesnt
[11:20] <ogra_cmpc> unless you set a root passwprd
[11:20] <ogra_cmpc> if it does, thats clearly a bug
[11:21] <hrw> ok, did not checked ubuntu way
[11:21] <hrw> nasty it is then
[11:22] <ogra_cmpc> why ?
[12:10] <playya_> moin
[14:02] <ogra> hrm
[14:02] <ogra> mcopy seems to prevent MLO from working
[14:34]  * ogra sighs about CHS and partitioning
[14:35] <hrw> ;D
[14:37] <ogra> its funny, i can build images i dd to the SD but MLO isnt found
[14:37] <ogra> CHS and partitioning are 100% identical to a booting SD
[14:39] <vstehle> ogra: what about the 'bootable' flag?
[14:40] <ogra> its set
[14:40] <ogra> as i said, its 100% identical
[14:40] <vstehle> ogra: is it possible it comes from the way the filesystem was created, then?
[14:40] <ogra> bytesize, start and end of the partitions CHS values of the dd'ed image on the SD vs a natively created one etc
[14:41] <ogra> well, i'm trying mkfs.vfat vs mkdosfs now
[14:41] <vstehle> ogra: also, the "position" of MLO on the filesystem (copied first or not)
[14:41] <ogra> might make a difference but i dont really expect to
[14:41] <ogra> i *only* copy MLO :)
[14:41] <vstehle> ogra: it is known to make a difference on OMAP3, sadle
[14:41] <vstehle> *sadly
[14:41] <ogra> not even u-boot or anything
[14:42] <vstehle> ogra: hmm...
[14:42] <vstehle> ogra: I use 'mkfs.vfat -F 32' on OMAP3
[14:42] <vstehle> ogra: don't know if this helps...
[14:43] <ogra> thats what i used too here
[14:43] <ogra> note though thst i create an empty partitioned image and then dd the filesystems in place
[14:46] <ogra> hmm, no, mkdosfs didnt make any difference
[14:47] <hrw> use omap3-sdcard.sh script?
[14:49] <XorA> its been reported to me that once you copy the MLO 4/5 times for some reason omap3 stops seeing it
[14:49] <XorA> its not an issue Ive seen personally
[14:50] <XorA> I guess there is a bug in fat handling in omap3 somewhere
[14:51] <ojn> yeah. i've seen that in the past too but after more than 4 or 5 copies. I think there's some limitations as to where in the directory table the MLO has to be.
[14:52] <ojn> omap4 is much better in those aspects. :)
[14:58] <ogra_cmpc> hrw, thats not possible with images
[15:01] <hrw> unless you will keep sdcard-header.bz2 somewhere
[15:01] <hrw> and such header will have proper partition table so you just need to append ext234 partition after it
[15:02] <ogra_cmpc> hrw, i need to generate it from scratch
[15:04] <ogra_cmpc> what we usually do is to determine the size of the partitions, add space for the MBR, create an empty image and add a partition table ...
[15:04] <ogra_cmpc> .... then dd the filesystems to the place in the image where they have to be
[15:05] <ogra_cmpc> that works fine for all images we build in ubuntu buut somehow chokes on thesee
[15:06] <ogra_cmpc> what i dont get is that what ends up on the SD is identical to what i create with omap3-sdcard.sh
[15:06] <ogra_cmpc> so there shouldnt be a reason they dont boot
[16:27] <ogra> WOHOO !
[16:27]  * ogra got it
[16:28] <ogra> how silly
[16:28] <vstehle> ogra: let me guess, no header ?
[16:29] <vstehle> (The 8 bytes in front of x-loader)
[16:29] <ogra> nah, apparently the blocksize isnt correct if you use dd to create the FS image, using mkdosfs -C gets me a working vfat
[16:29] <vstehle> ogra: oh, much more tricky then
[16:29] <ogra> i used to do dd, then mkdosfs, then using mcopy to get MLO in place
[16:29] <ogra> ignoring that mkdosfs vas a -C switch that creates the fat properly with the right blocksizes etc
[16:30] <ogra> s/vas/has/
[16:30] <ogra> so i'm one step closer to preinstalled ubuntu images :)
[16:30]  * ogra tries if that works on omap3 too 
[16:32] <ogra> hmm, how do i convince parted to show blocksizes
[16:34] <ogra> grmbl, parted has a unit command that doesnt offer blocks at all
[17:25] <lool> ndec: GStreamer fields are used to automatically discover packages
[17:26] <lool> ndec: Ultimately, this allows totem to offer installing this or that package with codecs when you're trying to play an unsupported media
[17:27] <lool> ndec: This involves two things: mapping a codec/MIME type to an element name (I think that's in app-install-data, but I'm not sure), and then mapping an element name to the package it's in (e.g. apt-cache show gstreamer0.10-esd should show Gstreamer-Elements: esdsink, so if you're trying to install the package with the esdsink element, you can install that package)
[17:29] <lool> ndec: In fact, app-install-data just provides some aggregated information found in all other packages
[17:53] <markos_> lool, ogra: hi guys, almost done with building ubuntu-desktop hardfp, using rootstock-native (using chroot not qemu), but got this problem http://paste.debian.net/76504/, any ideas how to fix this?
[17:53] <ogra> you shouldnt need hal
[17:53] <ogra> weird
[17:54] <ogra> is that lucid ?
[17:54] <ogra> hal is started by dbus now it doesnt come with an initscript
[17:55] <markos_> no karmic
[17:55] <ogra> ah
[17:55] <ogra> might be a chroot issue
[17:55] <markos_> next step is to build lucid, but I thought I'd at least complete base karmic install
[17:56] <markos_> most likely, as services don't start -ie the system doesn't "boot" from the chroot, I just chroot in the dir
[17:56] <ogra> hal had probs in the past running inside chroots, i know there were workarounds but i dont know if they are arch specific
[17:56] <cwillu> markos_, what all do you mount in the chroot?
[17:56] <markos_> proc, dev/pts, sys iirc
[17:56] <markos_> yes
[17:56] <ogra> yeah, you should definately have proc
[17:57] <cwillu> dev and dev/pts?
[17:57] <markos_> oh and -o bind /dev
[17:57] <markos_> hm, no sorry
[17:57] <markos_> just /dev/pts
[17:58] <markos_> (well, rootstock-native does)
[17:59] <zumbi> markos_: -o rbind /dev || -o bind dev/pts
[18:01] <zumbi> hrw|gone: just uploaded buildcross 0.0.4, it has an undocumented option (--bootstrap-debsrc) which tries to build a base from Debian sources, but quite does not work (that's why it is undocumented) ;-) Enjoy!
[18:19] <lool> markos_: So hal doesn't start and breaks your packages
[18:19] <lool> markos_: Is this under qemu-system-arm, or qemu syscall emulation?
[18:19] <lool> markos_: I wouldn't trust the later to allow starting hal
[18:20] <markos_> none, chroot under native arm
[18:21] <zumbi> markos_: i remember start-stop-daemon (provided by dpkg) was masked by some script which did nothing, so it prevent hal to work
[18:21] <zumbi> the good one was renamed start-stop-daemon.REAL
[18:22] <zumbi> markos_: maybe you are affected by such
[18:22] <markos_> well, just read that hal doesnt' do anything if /sys is not mounted so disabled that mount, let's see...
[18:23] <zumbi> hrw|gone: i forgot to uncomment some code for the bootstrap stuff to work, so better wait to next release.
[18:26] <gregoiregentil> rcn-ee rcn-ee_lpt: Ping?
[18:31] <lool> markos_: Might be a feature in your kernel being missing for hal
[19:04] <rcn-ee> gregoiregentil, pong (lunch) how's it going..
[19:04] <gregoiregentil> rcn-ee: good. thanks.
[19:05] <gregoiregentil> I feel a little bit "frustrated" because I would like to do Lucid
[19:05] <gregoiregentil> but I'm really stuck
[19:05] <gregoiregentil> Compiling on the device doesn't work well for me. It segfaults all the time
[19:05] <gregoiregentil> where do we stand on the rootstock issues?
[19:06] <rcn-ee> we are all frustrated with lucid.. ;)  it's segfaulting on all my 64bit machines..  did you guys have any luck with the rootstock-on-arm tree i sent you?  It's pretty solid on my beagle running debian squeeze...
[19:10] <rcn-ee> gregoiregentil, just thought of something, what kernel are you using when you compile on your device.. (I'm assuming touchbook 2.6.29/2.6.32? CONFIG_ARM_ERRATA_430973=y has to be enabled, lucid userspace will kill the native machine kernel)
[19:11] <gregoiregentil> yes. I'm using this errata
[19:12] <gregoiregentil> the rootstock-on-arm tree doesn't work well for me. It segfaults on the device. I need to try more
[19:14] <cwillu_at_work> rcn-ee, I've had generally good luck with rootstock-on-chroot
[19:14] <rcn-ee> strange.. is there a specific package? or something.. this is your current list right? http://git.alwaysinnovating.com/cgit.cgi/ai.ubuntu/tree/packages
[19:22] <rcn-ee> gregoiregentil, i'm giving that package list a quick run on my beagle..
[19:29] <gregoiregentil> this is the current list indeed
[19:33] <rcn-ee> twittling thumbs as the beagle goes to work... ;) didn't we have a qemu guy kinda interested in this bug...
[19:42] <rcn-ee> gregoiregentil, it's made it to the main package install phase (apt-get install "touchbook") it'll take anouther 30mins before all the packages are downloaded..  I'm gonna drive back to work, but i'll monitor it and then email you in an hour or two on what happend...
[19:42] <gregoiregentil> Thanks
[19:46] <ogra_tb> gregoiregentil, hey
[19:46] <gregoiregentil> hello
[19:47] <ogra_tb> i'm using your .32 kernel, i sthere any way to convince it to charge the battery ?
[19:48] <cwillu_at_work> rcn-ee, am I going to have a 2.6.35rc2 kernel tonight? :)
[19:48] <ogra_tb> lucid netbook runs awesome  on it but its a bit annoying that i have to replace the sd and run angstrom to let the battery charge
[19:56] <DanaG1> I wish AMD would make an ARM thingy with a Radeon.
[19:57] <ogra_tb>  DanaG1 you mean copying nvidia ?
[19:57] <DanaG1> yeah, but being more open.
[19:58] <ogra_tb> well, wait for maverick the TI situation will become a lot better
[20:21] <cwillu_at_work> ogra_tb, that sounds pleasantly ominous :D
[20:22] <ogra_tb> hehe
[22:58] <armin76> zumbi: there's a strike tmrw, so i won't be at my workplace
[23:01] <slangasek> ogra: does https://blueprints.launchpad.net/ubuntu/+spec/mobile-m-image-builds-without-root need a spec?  I see that it already has workitems
[23:23] <cwillu_at_work> rcn-ee, you're not ignoring my cries are you? :)
[23:47] <rcn-ee> nope.. just wasn't home.... ;) 2.6.35-rc2 is building, it was queyed up last night.. http://rcn-ee.homeip.net:81/dl/farm/log/
[23:58]  * cwillu_at_work hits refresh... occasionally