**** BEGIN LOGGING AT Mon May 17 02:59:58 2010
May 17 03:34:50 <khem> dcordes: too low
May 17 03:35:08 <khem> dcordes: somehow it went wrong somewhere before
May 17 03:35:41 <dcordes> khem, while putting the addresses in the device specific memory map in kernel ?
May 17 03:35:52 <dcordes> khem, or you mean runtime problem
May 17 03:38:22 <dcordes> khem, it's very possible the memory layout is defined in a wrong way
May 17 03:38:36 <dcordes> khem, but I don't know if it's related to that wrong address- justa thought.
May 17 04:05:32 <IgorK> Hi, guys
May 17 04:06:09 <IgorK> WHere can I get a documentation about different bitbake commands?
May 17 04:06:37 <IgorK> I'm talking about "bitbake -c <cmd>"....
May 17 05:32:06 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r9ab15d3633 10openembedded.git/classes/cross.bbclass: cross bbclass: add extra sed for gcc 4.3.x adding -L${STAGING_LIBDIR_NATIVE} to cross libstdc++.la
May 17 05:37:43 <IgorK> or maybe there is a list of available command I can use for the <cmd> in the bitbake -c<cmd>?
May 17 06:07:56 <khem> IgorK: bitbake -c listtasks <recipe>
May 17 06:08:17 <khem> will give you list of individual tasks available for a recipe
May 17 06:13:15 <IgorK> khem: but it won't give me something like a "devshell" command... ;-)
May 17 06:17:53 <eFfeM_work> ls
May 17 06:22:24 <IgorK> khem: and how do I find out what task do what?
May 17 06:24:49 <khem> IgorK: read bitbake manual
May 17 06:27:40 <dcordes> khem, I did a setup where it loads an initrd with an initscript that mounts a fat partitaion from sd, creates a loopfile with mknod, loop mounts an ext3 loopfile from the fat, chroot to that fat, plays back 720 video with mplayer. zero segfaults.
May 17 06:27:55 <IgorK> khem: all I can find is this: http://bitbake.berlios.de/manual/ch04s02.html and this: http://www.uv-ac.de/openembedded/openembedded-3.html section 3.2
May 17 06:28:26 <IgorK> it's not really helpful
May 17 06:28:41 <dcordes> khem, I think that is pretty nice and I'm wondering why udevd segfaults
May 17 06:30:16 <IgorK> for example what is the difference between compile and build?
May 17 06:30:42 <dcordes> khem, somebody on the ml said there might be a problem with statically linked binaries but I can't see any being executed
May 17 06:32:07 <dcordes> IgorK, while running each of these tasks, bitbake emmits a log for them. they can be found in the workdir of each package. something like ~/oe/tmp/work/armv5te/vi/temp/log.do_compile
May 17 06:32:35 <dcordes> IgorK, if you look at them you might get an idea of what the referring tasks do
May 17 06:33:18 <dcordes> khem, do you have any idea how to debug it further?
May 17 06:34:17 <IgorK> Guys, I compiled wxwidgets from the devshell
May 17 06:35:04 <IgorK> Then I issued "build" "install" and "package"
May 17 06:35:50 <mckoan> good morning
May 17 06:36:29 <dcordes> IgorK, what's your aim? are you patching it? using different compiler?
May 17 06:36:38 <dcordes> morning mckoan
May 17 06:36:48 <IgorK> in that order, but I think I need staging, before putting it to the x11 image
May 17 06:37:19 <IgorK> dcordes: The task is to put the wxwidgets library into the x11-gpe-image
May 17 06:38:00 <dcordes> IgorK, you don't need to hack around then. you can just have OE adding it to the image
May 17 06:38:06 <IgorK> I don't need to patch it, it is precompiled from "devshell"
May 17 06:38:20 <IgorK> and it is the same compiler
May 17 06:39:18 <dcordes> IgorK, why are you compiling from devshell ?
May 17 06:40:04 <IgorK> dcordes: because I had to change the Makefile.in manually to fix the bug inside the wx build system
May 17 06:40:16 <IgorK> good morning to u as well
May 17 06:40:56 <IgorK> mckoan: is it Monday already over there
May 17 06:53:29 <mckoan> IgorK: unfortunately yes :-)
May 17 06:54:45 <IgorK> I hear u. its still Sunday here for another 5 min ;-)
May 17 06:55:22 <IgorK> dcordes: any suggestion?
May 17 06:55:53 <mckoan> IgorK: where are you from?
May 17 06:59:54 <IgorK> originally from Ukraine, now live in US, Silicon Valley, U?
May 17 07:01:38 <mckoan> IgorK: cool place 8-) I'm from Italy
May 17 07:03:28 <IgorK> well since u r using wx maybe u can help me with my problem? I recompiled wx correctly from devshell and now want to put it into the x11-gpe-image
May 17 07:06:46 <CIA-2> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * r22c1f60698 10openembedded.git/conf/distro/include/sane-srcrevs.inc: 
May 17 07:06:46 <CIA-2> EFL: bump SRCREV a bit more for batget module fix
May 17 07:06:46 <CIA-2> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 07:12:16 <hrw> morning
May 17 07:12:48 <CIA-2> 03Khem Raj <raj.khem@gmail.com> 07org.openembedded.dev * rb362ce8d23 10openembedded.git/recipes/gcc/ (5 files in 2 dirs): (log message trimmed)
May 17 07:12:48 <CIA-2> gcc-4.3.3: Fix spurious leak of build sysroot into .la files.
May 17 07:12:48 <CIA-2> * ldflags were being passed unnessarily so set LDFLAGS to nothing in
May 17 07:12:48 <CIA-2> Makefile.tpl and regenerate top level Makefile.in
May 17 07:12:48 <CIA-2> * Divide the patches individually earlier the patches were
May 17 07:12:49 <CIA-2> combined into one big patch
May 17 07:12:49 <CIA-2> * Shuffle the order of patches.
May 17 07:12:50 <ericben> hi
May 17 07:12:58 <CIA-2> 03Khem Raj <raj.khem@gmail.com> 07org.openembedded.dev * r15f7ad6212 10openembedded.git/recipes/qemu/ (12 files in 2 dirs): 
May 17 07:12:58 <CIA-2> qemu-0.12.4: Add recipes for 0.12.4
May 17 07:12:58 <CIA-2> * Tested by booting arm, ppc, mips, sh4, and x86 systems
May 17 07:12:58 <CIA-2> Signed-off-by: Khem Raj <raj.khem@gmail.com>
May 17 07:16:13 <mckoan> hi hrw welcome back
May 17 07:16:31 <mckoan> IgorK: simply add wxwidgets to your image recipe
May 17 07:18:41 <IgorK> mckoan: I had to workaround the bug in the wx build system, so it is copmpiled from the devshell
May 17 07:19:17 <khem> hmm so 2.6.34 is out 
May 17 07:19:30 * JaMa building 2.6.34 :)
May 17 07:19:35 <JaMa> but from _git recipes
May 17 07:19:37 <khem> http://lkml.org/lkml/2010/5/16/89
May 17 07:20:05 <khem> JaMa: cool. will you add the recipe too ?
May 17 07:20:21 <khem> I can move qemu machine over to .34 one by one
May 17 07:20:42 <JaMa> _git versions are already there.. I can add _2.6.34. but I can test it only on spitz
May 17 07:21:02 <dcordes> khem, see it?
May 17 07:22:30 <IgorK> mckoan: and so can't really use OE recipe directly
May 17 07:22:46 <hrw> nice 
May 17 07:23:03 <hrw> JaMa: just add it and then add machines support
May 17 07:30:14 <mckoan> IgorK: which wx version are you using?
May 17 07:31:25 <IgorK> mckoan: I built wx-2.8.10 from sources
May 17 07:32:20 <likewise> mckoan: Good morning.  Does kaeilos support openjdk for 9263?
May 17 07:32:55 <mckoan> likewise: never tried
May 17 07:33:11 <IgorK> the bug is in bakefile for my distribution (Linux Angstrom) and its confirmed with wx-dev so I had to build it from devshell
May 17 07:33:30 <dcordes> khem, any thoughts?
May 17 07:33:49 <dcordes> JaMa, you a zaurus maintainer?
May 17 07:34:03 <likewise> morning all
May 17 07:37:23 <khem> good night all
May 17 07:37:50 <hrw> n8 khem 
May 17 07:39:28 <likewise> nite
May 17 07:40:56 <mckoan> IgorK: hmmm strange, I'll try an angstrom build later
May 17 07:41:55 <IgorK> mckoan: Im using mini2440 OE tree
May 17 07:42:28 <mckoan> IgorK: I use at91sam9263ek + wx and it works
May 17 07:42:46 <mckoan> IgorK: with KeilOS of course
May 17 07:42:50 <IgorK> wx compiled without GUI support, so I had to tweak Makefile.in to confirm...
May 17 07:43:12 <mckoan> IgorK: aha that must be the problem
May 17 07:43:23 <IgorK> I still have pastebin from config.log and do_compile log
May 17 07:43:55 <IgorK> the problem is that it is configured with it and config.log shows that everything is defined
May 17 07:44:34 <IgorK> but log.do_compile shows wxUSE_GUI=0
May 17 07:45:42 <IgorK> I guess I will just copy the files manually just to test it in the morning and if everything will work, then I will c how to put the libs in the image
May 17 07:47:38 <JaMa> hrw: adding it atm
May 17 07:47:45 <IgorK> mckoan: if you know how to put the precompiled libraries in the image, let me know when I show up here in the morning or during the day. TIA
May 17 07:48:07 <IgorK> good night all, thank u for u support
May 17 07:48:08 <JaMa> dcordes: not really, more like tester
May 17 07:54:29 <ant_work> JaMa: it looks LZMA for arm is not yet there :/
May 17 07:54:43 <ant_work> Russel said he has it 'in queue'
May 17 07:55:28 <JaMa> ant_work: but it's in recipe again.. with status=pending
May 17 07:56:58 <ant_work> JaMa: I'll write an RFC about removing some linux-kexecboot versions.
May 17 07:57:09 <ant_work> we could then merge in /linux
May 17 07:58:37 <JaMa> ant_work: that would be nice
May 17 07:59:31 <ant_work> I hope we can live with a 2.6.26 and a 2.6.34. Stop. 
May 17 08:03:05 <JaMa> ant_work: and _git :)
May 17 08:03:35 <ant_work> oh, yes
May 17 08:06:13 <pepere> hello, I've just started to work with oe. I experience issue using uImage generated for hawkboard board. uImage is tftp'ed but when kernel decompress than I got a Starting kernel and nothing else...
May 17 08:06:23 <pepere> does anybody get such sort of issue?
May 17 08:06:46 <CIA-2> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * r33eea80786 10openembedded.git/recipes/linux/ (3 files in 3 dirs): 
May 17 08:06:46 <CIA-2> linux: add 2.6.34 with spitz defconfig
May 17 08:06:46 <CIA-2> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 08:06:47 <CIA-2> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * rcd37fc611f 10openembedded.git/recipes/freesmartphone/libframeworkd-glib_git.bb: 
May 17 08:06:47 <CIA-2> libframeworkd-glib: bump SRCREV
May 17 08:06:47 <CIA-2> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 08:06:51 <CIA-2> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * r3c26dd6900 10openembedded.git/recipes/linux/ (linux/spitz/defconfig linux_git.bb): 
May 17 08:06:51 <CIA-2> linux_git: bump SRCREV for 2.6.34
May 17 08:06:51 <CIA-2> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 08:06:51 <CIA-2> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * r481fef91ed 10openembedded.git/recipes/kexecboot/ (linux-kexecboot/spitz/defconfig linux-kexecboot_git.bb): 
May 17 08:06:52 <CIA-2> linux-kexecboot_git: bump SRCREV for 2.6.34
May 17 08:06:52 <CIA-2> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 08:08:41 <pepere> I'm using remotes/origin/HEAD -> origin/org.openembedded.dev so it should support hawkboard as is...
May 17 08:09:39 <pepere> (and i've checked defconfig, some latest patch seems to having been applied, so I'm up to date with latest patches)
May 17 08:30:47 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r90fc51ebe1 10openembedded.git/recipes/ti/ti-devshell.bb: ti-devshell: fix method conflict
May 17 08:48:48 <florian> good morning
May 17 08:50:59 <likewise> gm
May 17 08:51:45 <likewise> can recipes inherit from other overlays, can recipes include/require from other overlays?  I cannot find the proper documentation on this anymore.
May 17 08:55:15 <ericben> likewise: it seems to be possible, in an overlay's linux recipe I have "require recipes/linux/linux.inc" and that works fine
May 17 08:57:27 <likewise> ericben: you are sure you do not have a copy of linux.inc in your overlay?
May 17 08:58:26 <ericben> likewise: yes
May 17 08:58:58 <ericben> likewise: I only have a linux_2.6.32.bb
May 17 08:59:12 <likewise> ericben: which bitbake version are you using?
May 17 08:59:24 <ericben> and the same in other recipes (at91bootstrap for example)
May 17 08:59:42 <ericben> likewise: 1.10.1
May 17 08:59:48 <ericben> 1.10.0 sorry 
May 17 09:00:53 <ericben> but that also worked with 1.8.19
May 17 09:14:22 <pb__> florian: good morning
May 17 09:14:26 <pb__> mickey|office: good morning
May 17 09:14:41 <Spyzer> hi, i do i build a zImage for an embedded device(mini2440) through bitbake??
May 17 09:14:53 <Spyzer> i mean how, sorry
May 17 09:15:25 <JaMa> Spyzer: bitbake virtual/kernel?
May 17 09:15:46 <Spyzer> pardon me, virtual/kernel??
May 17 09:18:29 <JaMa> Spyzer: I don't build for mini2440, but it's common target for all machines to build their preferred kernel
May 17 09:18:59 <Spyzer> hmm thanks
May 17 09:19:19 <Spyzer> so does the kernel gets build as a binary in the deploy directory??
May 17 09:19:51 <florian> Spyzer: yes
May 17 09:20:09 <Spyzer> k
May 17 09:21:51 <mickey|office> good morning pb_ 
May 17 09:51:09 <Spyzer> hi, i am trying to burn yaffs image on the mini2440 (i use minicom which shows the option of download root_yaffs image). But when i try to boot the machine i get kernel panic not syncing error, nothing can boot the the root yaffs image. 
May 17 09:51:29 <Spyzer> i am using a jffs2 image should i use a cramfs image instead
May 17 09:51:31 <Spyzer> ?
May 17 09:52:44 <ericben> Spyzer: you are using a jffs2 image and you say you are burning yaffs : isn't your problem here ?
May 17 09:54:14 <Spyzer> actually i don't know the difference, what is yaffs(yet another fileflash system , ibelieve) but in bitbake i cannot make a yaffs image now can I, so I though i should go for a jffs2 image. But what exactly should i go for if its not jffs
May 17 09:54:44 <Spyzer> should i go for cramfs
May 17 09:55:18 <Spyzer> ?
May 17 09:55:52 <ericben> just look how your bootloader can write a jffs2 into flash
May 17 10:01:31 <Spyzer> is there any way top create a yaffs image through bitbake??
May 17 10:02:37 <ericben> look in conf/machine/neuros-osd2.conf and conf/machine/include/neuros.inc : it seems to create yaffs2 images (do you need yaffs or yaffs2 ?)
May 17 10:05:35 <Spyzer> yaffs i believe
May 17 10:06:04 <Spyzer> i just saw a tutorial to create a yaffs image here http://processors.wiki.ti.com/index.php/Create_a_YAFFS_Target_Image
May 17 10:06:33 <Spyzer> kindly have a look at it and please specify that whether the image created by this the same as the .tar.gz image created by bitbake
May 17 10:07:20 <Spyzer> it seems to me that .tar.gz image itself can act as a yaffs image(as i can derive from the tutorial)
May 17 10:08:01 <ericben> Spyzer: they say : "However, it this document we will only cover creating a tarball image of a file system."
May 17 10:08:11 <ericben> yaffs != tar.gz ;-)
May 17 10:13:28 <Spyzer> No filesystem could mount root, tried:  ext3 cramfs vfat msdos romfs
May 17 10:13:33 <Spyzer> this is the sort of error i get
May 17 10:13:35 <Spyzer> :(
May 17 10:17:08 <ericben> Spyzer: first you need to be sure your kernel integrates the fs support you want, then that your bootloader can write this fs in flash, then that your flash is properly partitionned, then that you pass the right parameter to the kernel command line (root= and rootfstype= ) then that should boot
May 17 10:21:46 <Spyzer> the extension ".img" and i burned them and they run f9. Can i produce a .img image from bitbake then??
May 17 10:21:58 <Spyzer>  the images provided with the machine itself have
May 17 10:23:34 <Spyzer> the last line should be read before the first line
May 17 10:39:09 <pb_> mickey|office: I am looking for someone to help us with a small iphone project.  do you have any spare time at the moment?
May 17 10:42:22 <hrw> /home/hrw/devel/oe/build/tmp/work/x86_64-linux/guile-native-1.8.7-r0/sysroot-destdir///home/hrw/devel/oe/build/tmp/sysroots/x86_64-linux/usr/lib/libguile.so.17.3.1: new RPATH: $ORIGIN/../../usr/lib
May 17 10:42:39 <hrw> how to deal with it? I am new to many .dev changes ;(
May 17 10:42:52 <mickey|office> pb_: i'm buried in a project right now and it looks like it's not going to be finished before Mid-June... Is that too late?
May 17 10:43:12 <pb_> mickey|office: we might be able to live with that.  I'll send you an email with some details and you can see what you think.
May 17 10:44:47 <mickey|office> pb_: good, thanks
May 17 10:44:55 <mickey|office> pb_: can you please send it to late@app-developers.de?
May 17 10:45:00 <mickey|office> (that's my small iphone team)
May 17 10:45:05 <pb_> sure, will do
May 17 10:45:07 <mickey|office> thanks
May 17 10:53:43 <Spyzer> ok setting rootfstype=jffs2 in the machine atleast solved the cannot mount fs error, but now its saying that there is no init scripts try passing init= option, while i have already done that. I bitbaked a micro-base-image
May 17 10:54:10 <Spyzer> please lend me a hand...
May 17 10:54:49 <Spyzer> i have already passed init=/linuxrc, but it says can't find linuxrc
May 17 10:57:27 <Spyzer> Warning: unable to open an initial console.
May 17 10:57:27 <Spyzer> Failed to execute /linuxrc.  Attempting defaults...
May 17 10:57:27 <Spyzer> Kernel panic - not syncing: No init found.  Try passing init= option to kernel.
May 17 10:57:40 <hrw> arm?
May 17 10:57:48 <Spyzer> yes
May 17 10:57:56 <hrw> vendor kernel instead of OE one?
May 17 10:58:11 <Spyzer> no i am using an OE kernel
May 17 10:58:30 <Spyzer> should i use the vendor one?
May 17 10:59:00 <hrw> no, you should check does your kernel supports EABI and THUMB
May 17 10:59:18 <hrw> Spyzer: which machine it is? and which kernel recipe is uses?
May 17 11:00:36 <Spyzer> well its a mini2440 and i just a bitbake micro-base-image and but i do not remeber the kernel recipe
May 17 11:01:13 <hrw> mini2440? I would expect it to be working
May 17 11:02:12 <Spyzer> can the problem be possibly related to downloading the images to the mini2440 through usb, i use minicom??
May 17 11:02:54 <Spyzer> and i follow the general minicom terminologies, there is one tutorial on the net which does it through inbuilt capabilities of superviv
May 17 11:03:22 <Spyzer> but i don't go in that manner, so can it be the case that the fs gets written to a wrong block or something
May 17 11:04:13 <hrw> check with console-image? I do not know what is in micro-base-image
May 17 11:05:49 <Spyzer> ok will do
May 17 11:18:48 <gnutoo_> finally: NOTE: Running task 1122 of 6844 (ID: 3363, virtual:native:/home/embedded/oe/org.openembedded.dev/recipes/gtk+/gtk+_2.20.0.bb, do_install)
May 17 11:19:00 <gnutoo_> I just did git pull
May 17 11:40:12 <CIA-2> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * r98bed2876d 10openembedded.git/recipes/linux/ (5 files in 2 dirs): 
May 17 11:40:12 <CIA-2> linux-openmoko-2.6.32: move to 2.6.32.13, add 2 more patches
May 17 11:40:12 <CIA-2> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 12:10:06 <CIA-2> 03Steffen Sledz <sledz@dresearch.de> 07org.openembedded.dev * r543d057b7d 10openembedded.git/recipes/ (5 files in 3 dirs): 
May 17 12:10:06 <CIA-2> busybox/syslogs: use update-alternatives for syslog selection
May 17 12:10:06 <CIA-2> * u-a for init scripts of busybox-syslog, sysklogd, and rsyslog
May 17 12:10:06 <CIA-2> * syslog-ng needs to be integrated too!
May 17 12:10:06 <CIA-2> Signed-off-by: Steffen Sledz <sledz@dresearch.de>
May 17 12:10:06 <CIA-2> Acked-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 12:12:20 <siji>  'eglInitialize' returned egl error 'EGL_BAD_ALLOC' (0x3003)
May 17 12:22:16 <supjigator> Hi.  Any plans on upgrading dosbox?  I have a recipe for the current 74_r1 that works.
May 17 12:47:29 <v3gard> can anyone here point me to litterature that explains the pros of customizing a kernel (for an embedded system) versus using a generic one?
May 17 12:48:12 <ant_work> three words: "generic wont boot"
May 17 12:48:37 <ant_work> in the worst case
May 17 12:49:38 <v3gard> thanks for the input, but I actually require a reference of some sort for a paper i'm writing :)
May 17 12:50:49 <ant_work> well, then google for 'shrink the size' and 'boot time'
May 17 12:51:38 <ant_work> and search in elinux / CELF archives
May 17 12:51:55 <ant_work> lot of papers there
May 17 12:54:25 <v3gard> great!
May 17 12:54:26 <v3gard> thanks
May 17 13:11:18 <ant_work> any dfu-tools master here?
May 17 13:11:58 <ant_work> I have a friend trying to use the OpenMoko OSX tools but these appear to be broken :/
May 17 13:17:59 <ant_work> stefan_schmidt: ping
May 17 13:19:19 <stefan_schmidt> ant_work: pong
May 17 13:19:29 <ant_work> hello there
May 17 13:19:36 <ant_work> just wondering about dfu-utils
May 17 13:19:52 <stefan_schmidt> ant_work: tested the release candidate?
May 17 13:20:00 <ant_work> no, different question
May 17 13:20:08 <stefan_schmidt> ok, go ahead
May 17 13:20:18 <ant_work> a friend is trying to flash a BT dongle from OSX
May 17 13:20:30 <ant_work> he tried to compile the sources
May 17 13:20:48 <ant_work> but even using the OpenMoko binary there are no devices shown
May 17 13:20:53 <ant_work> (on OSX)
May 17 13:20:58 <stefan_schmidt> ant_work: hmm
May 17 13:21:04 <ant_work> then I found your git tree
May 17 13:21:05 <ant_work> ...
May 17 13:21:14 <stefan_schmidt> Never used OSX myself
May 17 13:21:22 <stefan_schmidt> ant_work: The openmoko binaries are for linux
May 17 13:21:27 <ant_work> first question: must one put the dongle in DFU mode somehow?
May 17 13:21:30 <stefan_schmidt> ant_work: Did it compile?
May 17 13:21:57 <ant_work> he followed
May 17 13:21:59 <ant_work> http://wiki.openmoko.org/wiki/User:SNMoore
May 17 13:22:03 <stefan_schmidt> ant_work: That depends on the dongle of course. Some seem to have a dfu descriptor in runtime as well
May 17 13:22:40 * stefan_schmidt reads
May 17 13:23:22 <stefan_schmidt> Ah, someone tried to get it working under OSX
May 17 13:23:38 <stefan_schmidt> That's good. I'm glad to get patches for this.
May 17 13:24:09 <stefan_schmidt> ant_work: For the USB dongle. It really depends on the dingle I think.
May 17 13:24:10 <ant_work> you got bad words until now....
May 17 13:24:14 <ant_work> ;)
May 17 13:24:44 <stefan_schmidt> ant_work: The one I have was not recognized at all. Perhaps it is not able to upgrade the firmware.
May 17 13:24:53 <stefan_schmidt> ant_work: bad words? care to explain?
May 17 13:24:58 <ant_work> ok, I did the upgrade 5-6 years ago
May 17 13:25:09 <ant_work> I extracted the fw from ther Apple 1.2 updater
May 17 13:25:18 <ant_work> this one had the HID proxy feature
May 17 13:25:30 <ant_work> so on boot you can branch BT kb + mouse
May 17 13:25:58 <ant_work> he's trying to do it now with recent dongles
May 17 13:25:59 <stefan_schmidt> The upgrade on the dongle you are talking about or another one?
May 17 13:26:14 <ant_work> I suppose he has BT 2.0 EDR at least
May 17 13:26:31 <ant_work> we are talking about CSR dongles with Ext memory, btw
May 17 13:27:00 <ant_work> stefan_schmidt: ''.. currently this dfu stuff doesn't see any bluetooth devices.''
May 17 13:27:01 <stefan_schmidt> Does it offer a DFU descriptor?
May 17 13:27:18 <ant_work> how to tell? it seems OSX lacks the basic bluez utils
May 17 13:27:29 <stefan_schmidt> ant_work: lsusb?
May 17 13:27:30 <ant_work> dfutool is enough?
May 17 13:27:41 <ant_work> yep, I made him compile lsusb :)
May 17 13:27:51 <stefan_schmidt> ant_work: They can send me patches for the usb dongles and OSX support :)
May 17 13:27:56 <ant_work> do youneed some output?
May 17 13:28:34 <stefan_schmidt> That or a crystal ball :)
May 17 13:28:51 <stefan_schmidt> For now the most testing was done with the openmoko devices
May 17 13:28:52 <ant_work> ok, mom
May 17 13:29:18 <stefan_schmidt> bluez should also have a dfu implementation which might find the device
May 17 13:29:33 <ant_work> I'm sure on Linux there would be less issues...
May 17 13:29:47 <ant_work> he's posting the lo, mom
May 17 13:29:51 <ant_work> *log
May 17 13:29:52 <stefan_schmidt> As I said my bt dongle was not detected either and I haven't looked into this yet
May 17 13:30:16 <ant_work> it seems there are smaller dongles without ext memory, rom-only
May 17 13:30:16 <stefan_schmidt> Is there no update tool from CSR for OSX?
May 17 13:30:23 <ant_work> these cannot be upgraded
May 17 13:30:41 <ant_work> doh, no idea. I know about fw updater
May 17 13:30:45 <stefan_schmidt> indeed, so far I suspected mine is one of these. Or maybe dfu-util just does something wrong
May 17 13:30:49 <ant_work> http://pastebin.ca/M1cSEUz9
May 17 13:32:44 <stefan_schmidt> hmm, pastebin is slow...
May 17 13:33:01 <ant_work> yes...retrying..
May 17 13:33:49 <ant_work> http://pastebin.com/4wMtUVwM
May 17 13:36:57 <stefan_schmidt> ant_work: ok, let me consult the DFU spec
May 17 13:38:14 <ant_work> thx
May 17 13:43:37 <ant_work> hello BuildSmart 
May 17 13:43:55 <BuildSmart> stephan_schmidt: can I bother you for a moment?
May 17 13:44:45 <BuildSmart> stefan_schmidt: can I bother you for a moment?
May 17 13:45:08 <stefan_schmidt> BuildSmart: give me 10 minutes
May 17 13:45:21 <BuildSmart> no problem, thank you.
May 17 13:45:36 <ant_work> stefan_schmidt: ^^ he's the friend on OSX
May 17 13:46:23 <stefan_schmidt> ant_work: According to the DFU 1.0 spec a DFU run-time interface descriptor would need to have bNumEndpoint 0
May 17 13:46:41 <stefan_schmidt> I can't find such a descriptor in the output for the bt dongle
May 17 13:46:47 <stefan_schmidt> ant_work, BuildSmart: Ah, hi
May 17 13:47:43 <stefan_schmidt> It has   bInterfaceSubClass 0 but   bInterfaceProtocol never 0
May 17 13:47:53 <stefan_schmidt> No wonder dfu-util does not discover it
May 17 13:48:03 <stefan_schmidt> Are you guys sure that it can be upgraded?
May 17 13:48:26 <stefan_schmidt> It seems not to have a option to toggle from run-time to dfu mode
May 17 13:48:42 <stefan_schmidt> (also states hci mode in the name already)
May 17 13:48:43 <ant_work> it seems CSR confirmed the devices can be upgraded
May 17 13:48:51 <stefan_schmidt> strange
May 17 13:48:52 <ant_work> BuildSmart: ?
May 17 13:49:00 <BuildSmart> stefan, yes, CSR has assured me the devices are reprogrammable.
May 17 13:49:35 <ant_work> stefan_schmidt: straight question: which PS key to change in order to enable HID Proxy ?
May 17 13:49:42 <stefan_schmidt> One a first glance I can't see how you would switch into dfu mode
May 17 13:49:43 <ant_work> (if it is soo simple..)
May 17 13:49:52 <stefan_schmidt> ant_work: no idea
May 17 13:49:59 <ant_work> then flash.. :/+
May 17 13:51:11 <ant_work> Probably disassembling the generic csr fw provided by apple could reveal smthg
May 17 13:51:36 <stefan_schmidt> That or trying to find a fw updater from CSR for OSX?
May 17 13:51:52 <ant_work> well, the idea is to enable HID Proxy
May 17 13:51:53 <stefan_schmidt> It could good be that dfu-util is missing some magic here.
May 17 13:52:10 <BuildSmart> the OS can be windows, OSX or linux (I'm flexible)
May 17 13:52:17 <stefan_schmidt> But I still don't get how you would switch into dfu mode. The normal way over  runtime descriptor seems not be available
May 17 13:52:44 <stefan_schmidt> ant_work: Why that? What does HID proxy buy you?
May 17 13:52:52 <BuildSmart> that might be an OSX driver related issue?
May 17 13:52:59 <ant_work> it gives BT keyb + mouse on bios screen
May 17 13:53:03 <stefan_schmidt> BuildSmart: no idea, never used OSX
May 17 13:53:17 <stefan_schmidt> ant_work: yeah, and what does it bring for flashing it?
May 17 13:53:22 <BuildSmart> no problem... does linux have a tool to put the device in DFU mode?
May 17 13:53:39 <stefan_schmidt> ant_work: A PS key for switching in dfu mode would make more sense
May 17 13:53:56 <stefan_schmidt> BuildSmart: That is the question
May 17 13:54:19 * stefan_schmidt suggest to look into the small dfu util in bluez. The might have such a trick implemented already
May 17 13:54:31 <stefan_schmidt> They wrote it for bt dongles anyway. :)
May 17 13:54:59 <ant_work> I've seen in the bluez sources (dfutool.c)
May 17 13:55:05 <stefan_schmidt> Your dingle behaves like the one I have. Maybe it is really just a PS key to switch the modes
May 17 13:55:12 <stefan_schmidt> ant_work: yeah, thats the one
May 17 13:55:30 <BuildSmart> dfu-util is also supplied with OpenMoko (for OS X) but it reports no devices and the GUI makes absolutely no sense since it shows no devices.
May 17 13:55:44 <stefan_schmidt> I'll look into getting mine to work, but it has to wait until the weekend.
May 17 13:55:55 <stefan_schmidt> BuildSmart: GUI?
May 17 13:56:02 <stefan_schmidt> dfu-util does only have a commandline
May 17 13:56:04 <ant_work> I can test with USB and UART 
May 17 13:56:14 <BuildSmart> OpenMoko has a GUI for OS X
May 17 13:56:33 <BuildSmart> I have dfutil for OSX but no clue how to use it.
May 17 13:56:34 <stefan_schmidt> BuildSmart: well, then you are not using the official dfu-util but some fork
May 17 13:56:47 <stefan_schmidt> I can't help you then
May 17 13:56:56 <BuildSmart> I downloaded the source myself from svn as well.
May 17 13:57:04 <ant_work> stefan_schmidt: is your git the fresh one around?
May 17 13:57:16 <stefan_schmidt> ant_work: as fresh as possible
May 17 13:57:28 <stefan_schmidt> BuildSmart: From which svn?
May 17 13:58:16 <BuildSmart> svn.openmoko.org ehich may not be the best source I guess.
May 17 13:58:38 <stefan_schmidt> BuildSmart: And how did you get a GUI from there?
May 17 13:59:04 <BuildSmart> it's listed on thw wiki with a link to download.
May 17 13:59:28 <stefan_schmidt> https://svn.openmoko.org/trunk/src/host/dfu-util/
May 17 13:59:36 <stefan_schmidt> That is the svn. And there is no UI
May 17 13:59:53 <stefan_schmidt> BuildSmart: Did you build it from there or downloaded a binary for OSX?
May 17 14:00:25 <BuildSmart> again... I got the source form the svn and built it, also downloaded and tried the binary, neother show any devices.
May 17 14:00:30 <BuildSmart> neither
May 17 14:00:30 <stefan_schmidt> Point is there is no GUI in dfu-util and if you have one you are running a fork, a GUI on top or something different
May 17 14:00:44 <BuildSmart> it's just a wrapper.
May 17 14:00:59 <BuildSmart> don't get hung up on the GUI
May 17 14:01:58 <stefan_schmidt> Well, if you are using something else my help would be useless and I should work right now normally...
May 17 14:02:18 <BuildSmart> the binary is form the svn, ignore any reference to the GUI.
May 17 14:02:45 <BuildSmart> I have built source and downloaded the prebuilt, both list no devices.
May 17 14:02:46 <stefan_schmidt> And it does not detect the dongle? To be expected from what I see from the lsusb output
May 17 14:02:58 <stefan_schmidt> no surprise
May 17 14:03:28 <stefan_schmidt> If it has another method then the run-time descriptor to switch into dfu mode dfu-util does not support it yet
May 17 14:03:30 <BuildSmart> well I can boot ubuntu but am then lost without presupplied utils, same for windows.
May 17 14:03:48 <stefan_schmidt> No need to, it will be the same under linux
May 17 14:03:59 <BuildSmart> if somehting works in linux then I can make it work in OSX.
May 17 14:04:15 <stefan_schmidt> it will not work there
May 17 14:04:35 <BuildSmart> under linux how do you put a device into DFU mode?
May 17 14:04:57 <stefan_schmidt> it depends on the _device_ not on the host system
May 17 14:05:07 <stefan_schmidt> Anybody ready what I write?
May 17 14:05:56 <BuildSmart> ok and how would I know under linux????
May 17 14:07:57 <BuildSmart> do you have a BT device that shows DFU capability?
May 17 14:08:07 <stefan_schmidt> I give up here. Read the backlog. I explained what I suspect more then once and also gave you a hint about the bluez version.
May 17 14:09:45 <BuildSmart> I see so basically there is no BT devie with DFU mode is what you are saying.
May 17 14:10:00 <BuildSmart> reading the backlog tells me nothing.
May 17 14:10:46 <BuildSmart> the software may be broken in OSX so trying in another OS makes sense but I'd like to see some out put that shows what I should be looking for....
May 17 14:11:28 <BuildSmart> telling me that output exists but giving me nothing to compare it to is not helpful....
May 17 14:14:16 <ant_work> stefan_schmidt: thx, I'll try on Linux later
May 17 14:14:56 <stefan_schmidt> ant_work: The best tip is the dfuutils from bluez
May 17 14:15:13 <stefan_schmidt> ant_work: I'm sure it will not work from linux either with dfu-util. As I already stated.
May 17 14:15:28 <ant_work> I have still that old 1.2 dongle somewhere
May 17 14:15:29 <stefan_schmidt> ant_work: It seems to use something else to switch to DFU mode
May 17 14:15:36 <ant_work> I updated years ago
May 17 14:15:39 <BuildSmart> stefan, tried, they do not have an option to list any devices
May 17 14:15:42 <stefan_schmidt> ant_work: Did that work with dfu-util?
May 17 14:15:52 <ant_work> I'll have yet to see
May 17 14:16:06 <ant_work> ah, misunderstoood
May 17 14:16:25 <ant_work> yep, I extracted the apple fw from dmg and flashed on my CSR dongle
May 17 14:16:36 <stefan_schmidt> with _dfu-util_?
May 17 14:16:51 <ant_work> I'd say yes, back in 2005 
May 17 14:17:06 <ant_work> but at the time I was playing with windows too
May 17 14:17:19 <ant_work> (dfu-tools of Win)
May 17 14:17:43 <ant_work> it was old Bluecore2 chipset iirc
May 17 14:19:03 <BuildSmart> so far all I can confirm is that dfu-util does not recognize any devices
May 17 14:19:04 <stefan_schmidt> I wonder if df-util even exists in 2005
May 17 14:19:18 <stefan_schmidt> BuildSmart: Any bluetooth devices you mean
May 17 14:20:25 <BuildSmart> dfu-util -l shows no devies of any kind.
May 17 14:20:30 <BuildSmart> *  BlueZ - Bluetooth protocol stack for Linux
May 17 14:20:32 <woglinde> jo
May 17 14:20:47 <woglinde> BuildSmart did you run dfu-util with root?
May 17 14:20:50 <BuildSmart>  *  Copyright (C) 2003-2007  Marcel Holtmann <marcel-kz+m5ild9QBg9hUCZPvPmw@public.gmane.org>
May 17 14:21:28 <stefan_schmidt> What should me tell this copyright notice?
May 17 14:22:50 <BuildSmart> I've tried the bluez software, it has no options to list any devices
May 17 14:23:01 <woglinde> hi stefan
May 17 14:23:07 <stefan_schmidt> hi woglinde 
May 17 14:23:46 <stefan_schmidt> BuildSmart: Then it maybe can't detect it either.
May 17 14:23:54 <stefan_schmidt> Anyway, not my problem. I have to work
May 17 14:24:02 <woglinde> hm 2.6.34 is out
May 17 14:24:18 <BuildSmart> setfan, the bluez software has no utility or options to list available devices.
May 17 14:24:38 <stefan_schmidt> ant_work: Sorry, I really wanted to do you a favour but I can't work this way
May 17 14:24:38 <BuildSmart> this means you need to know the device you wish to interact with.
May 17 14:25:11 <stefan_schmidt> BuildSmart: Use the prdout and device ID
May 17 14:25:19 <ant_work> stefan_schmidt: thx again
May 17 14:35:57 <woglinde> jo kergoth
May 17 14:36:11 <woglinde> hm
May 17 14:36:18 <woglinde> till later
May 17 14:37:43 <kergoth_> hey
May 17 14:45:51 <pb_> g'day kergoth
May 17 15:32:17 <woglinde> re
May 17 15:32:18 <woglinde> hi pb
May 17 15:35:10 <pb_> hi woglinde 
May 17 16:03:01 <aalonso> Hi, I'm working on adding OE support for microblaze arch, and reach the point where gcc-cross-initial fails to compile showing the next error "configure: error: cannot compute suffix of object files: cannot compile"
May 17 16:03:31 <aalonso> has some one deal with this type of error? 
May 17 16:05:28 <aalonso> An error log can be found at http://pastebin.com/kqC8AP1s 
May 17 16:36:42 <woglinde> re zecke
May 17 17:24:55 <florian> re
May 17 17:33:38 <khem> gm all
May 17 17:34:27 <GNUtoo> khem, hi, my issue was solved with a git pull
May 17 17:34:43 <khem> GNUtoo: nice
May 17 17:34:47 <GNUtoo> thanks a lot for the help!
May 17 17:34:51 <khem> np
May 17 17:45:36 <Spyzer> hi, i am trying to burn a jffs2 image on mini2440. I even set the kernel parameter: root=/dev/mtdblock3 and yet it says that VFS: Cannot open root device "mtdblock3" or unknown-block(31,3)
May 17 17:45:37 <Spyzer> Please append a correct "root=" boot option; here are the available partitions:
May 17 17:45:37 <Spyzer> 1f00             256 mtdblock0 (driver?)
May 17 17:45:37 <Spyzer> 1f01             128 mtdblock1 (driver?)
May 17 17:45:37 <Spyzer> 1f02            5120 mtdblock2 (driver?)
May 17 17:45:38 <Spyzer> 1f03          256640 mtdblock3 (driver?)
May 17 17:45:57 <Spyzer> as u can see mtdblock3 holds the jffs2 file system
May 17 17:46:05 <Spyzer> yet the kernel is not recognising it
May 17 17:46:17 <Spyzer> i bitbaked  a console-image for mini2440
May 17 17:46:24 <Spyzer> please help...
May 17 17:47:32 <Spyzer> if some one may tell me as how do i create a .img (yaffs)file system through a tar.gz fs that can also help me a lot
May 17 17:47:35 <khem> Spyzer: have you enabled jffs2 support in kernel ?
May 17 17:47:54 <Spyzer> yes i did pass the option rootfstype=jffs2
May 17 17:48:03 <Spyzer> in the kernel parameter boot options
May 17 17:48:12 <khem> in .config
May 17 17:48:38 <khem> look for CONFIG_JFFS2_FS
May 17 17:48:45 <Spyzer> ohh but i built the kernel from the bitbake virtual/kernel is there any recipe which i should change??
May 17 17:49:04 <khem> go into the kernel build tree and grok the .config
May 17 17:49:44 <Spyzer> kernel build tree?? but bitbake directly extracts the source while making the kernel??
May 17 17:50:01 <Spyzer> can't this setting be changed in any kind of recipe??
May 17 17:50:06 <khem> Spyzer: yes where it built for you in your tmpdir
May 17 17:50:24 <Spyzer> ok and then after changing it i rebuild it??
May 17 17:52:25 <khem> first see id its unset 
May 17 17:52:34 <khem> if it is set then problem is something else
May 17 17:53:38 <Spyzer> there is one more thing I have also added INHERIT+="rm_work" so i believe that kernel tree must have been removed after the image ahs been deployed
May 17 17:54:11 <Spyzer> khem: do you know how do i convert a tar.gz fs image to a .img(yaffs2) fs image
May 17 17:54:26 <Spyzer> that can also solve the provblem i believe
May 17 17:55:48 <khem> Spyzer: hmmm
May 17 17:56:11 <khem> Spyzer: you can create an img 
May 17 17:56:40 <khem> remove rm_work
May 17 17:56:55 <khem> and then bitbake the virtual/kernel once again
May 17 17:59:39 <khem> Spyzer you have to do few steps
May 17 18:00:42 * khem is googling to avoid typing
May 17 18:00:45 <khem> http://processors.wiki.ti.com/index.php/Create_a_YAFFS_Target_Image
May 17 18:01:01 <khem> ah spyzer is gone
May 17 18:02:04 <Spyzer> i got disconnected
May 17 18:02:23 <Spyzer> :(
May 17 18:16:33 <Spyzer> i am sorry that i keep on getting disconnected even if someone is helping me from there i just keep ruining that up. :(
May 17 18:17:25 <Spyzer> in any case is can anyone tell me as to what is the default rootfstype that is supported by the kernel built from bitbake virtual/kernel ??
May 17 18:17:48 <Tartarus> The supported ones are in the kernel config file
May 17 18:19:33 <Spyzer> ok, now where will i find that, i just have the tar.gz sources in my sources directory and I have done INHERIT+="rm_work" so built stuff gets deleted
May 17 18:19:47 <Spyzer> i mean is there any other location which i am missing
May 17 18:19:47 <Spyzer> ?
May 17 18:19:48 <Tartarus> Well, openembedded/recipes/linux/linux-VER/MACHINE/defconfig
May 17 18:19:53 <Tartarus> Or so
May 17 18:20:02 <Spyzer> ohh thnks
May 17 18:21:48 <Spyzer> i don't have any openembedded/recipes/linux/linux-VER??
May 17 18:21:58 <Spyzer> i have many others in openembedded/recipes/linux/
May 17 18:22:04 <Spyzer> but not linux-VER ??
May 17 18:24:13 <Spyzer> really i am serious
May 17 18:25:25 <Tartarus> Yes, not a litteral VER
May 17 18:25:28 <Tartarus> but whatever version you built
May 17 18:25:34 <Tartarus> And for whatever machine you're talking about
May 17 18:26:07 <Spyzer> ohh i am so sorry for showing such lameness :'(
May 17 18:26:29 <Tartarus> No worries :)\
May 17 18:33:15 <Spyzer> oh man, i saw the config file for linux kernel and its thew support for all ext3, jffs2 and others which bitbake builds. Now its beyond me as to why the kernel is unable to mount the fs in mini2440
May 17 18:33:51 <Spyzer> i send the fs through minicom in which there is only one option regarding the sending of fs and thats is download a yaffs image
May 17 18:34:17 <Spyzer> but does it matter if i send a jffs2 image and write the kernel boot option rootfstype=jffs2??
May 17 18:35:16 <khem> Spyzer try the number 1f03
May 17 18:35:50 <Spyzer> umm sorry i didn't quite get u, i mean how and where should i try it
May 17 18:36:16 <khem> root=1f03
May 17 18:37:28 <Spyzer> umm ok, but how did u get this number??
May 17 18:37:57 <khem> from what you typed earlier 
May 17 18:46:51 <Spyzer> i did root=1f03 but am getting the same error, i can describe the error a little bit more . First "uncorrectable error:[\n]" keeps on printing for about 100 times with great speed and then i finally get 
May 17 18:46:53 <Spyzer> Empty flash at 0x0017d130 ends at 0x0017d180
May 17 18:46:54 <Spyzer> Node at 0x0017ffd4 with length 0x00000244 would run over the end of the erase block
May 17 18:46:54 <Spyzer> Perhaps the file system was created with the wrong erase size?
May 17 18:46:54 <Spyzer> Cowardly refusing to erase blocks on filesystem with no valid JFFS2 nodes
May 17 18:46:54 <Spyzer> empty_blocks 1888, bad_blocks 115, c->nr_blocks 2005
May 17 18:46:55 <Spyzer> VFS: Cannot open root device "1f03" or unknown-block(31,3)
May 17 18:46:57 <Spyzer> Please append a correct "root=" boot option; here are the available partitions:
May 17 18:46:59 <Spyzer> 1f00             256 mtdblock0 (driver?)
May 17 18:47:01 <Spyzer> 1f01             128 mtdblock1 (driver?)
May 17 18:47:03 <Spyzer> 1f02            5120 mtdblock2 (driver?)
May 17 18:47:05 <Spyzer> 1f03          256640 mtdblock3 (driver?)
May 17 18:47:07 <Spyzer> Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(31,3)
May 17 18:53:42 <khem> Spyzer:  root=/dev/mtdblock3 rootfstype=jffs2 rw
May 17 18:53:45 <khem> does that work
May 17 18:54:15 <khem> Spyzer: is it NAND or NOR
May 17 18:54:25 <Spyzer> i have trying that only since quite a time but no success :(
May 17 18:54:34 <Spyzer> it is NAND i am trying to write
May 17 18:54:48 <Spyzer> but an doing the writing stuff through NOR
May 17 18:55:07 <khem> CONFIG_MTD_NAND is it set ?
May 17 18:55:52 <khem> look at CONFIG_MTD_* options 
May 17 18:58:15 <Spyzer> khem: yes it is set to 'y'
May 17 19:08:56 <Tartarus> khem: Yes, someone might wanna play with some of the other options w/ -g3 to get the size down a tad, so long as they're really not otherwise harmful
May 17 19:09:13 * Tartarus suspects where Koen is going with this, but doesn't have time to confirm himself
May 17 19:10:08 <khem> Tartarus: thats ok. I stopped using angstrom because of this nuisance 
May 17 19:10:19 <khem> My laptop partition for lnx is 64gb
May 17 19:10:32 <khem> and I can only afford one angstrom console-image build
May 17 19:10:45 <khem> where as three minimal distro builds 
May 17 19:13:31 <Spyzer> khem: may u please tell me as to where should i look to resolve the error
May 17 19:14:18 <khem> Spyzer: whatever I poked seems to be in place
May 17 19:14:25 <Spyzer> also let me tell u that i try to burn the vendors provided kernel and their pre built images everything works just f9
May 17 19:14:33 <Spyzer> yeah :(
May 17 19:15:10 <khem> Spyzer: ok try to boot same root file system as you have with vendor kernel
May 17 19:15:26 <Spyzer> yeah i was also thinking so
May 17 19:15:29 <khem> and check what it says about mtdpartitions 
May 17 19:15:30 <Spyzer> will try it
May 17 19:15:35 <Spyzer> ok
May 17 19:15:51 <khem> the ? after driver is strange
May 17 19:16:35 <Spyzer> although right now am trying to burn an image created through mkyaff2img utility with which i converted the tar.gz image generated by bitbake
May 17 19:16:39 <Spyzer> yeah
May 17 19:22:45 <CIA-2> 03Klaus Kurzmann <mok@fluxnetz.de> 07org.openembedded.dev * r7c393dcee8 10openembedded.git/recipes/freesmartphone/fsonetworkd_git.bb: 
May 17 19:22:45 <CIA-2> fsonetworkd_git.bb: add RDEPENDS for iptables
May 17 19:22:45 <CIA-2> Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>
May 17 19:25:25 <CIA-2> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * r14e1ed7036 10openembedded.git/recipes/linux/ (31 files in 2 dirs): 
May 17 19:25:25 <CIA-2> linux-openmoko-2.6.32: add patches from gdrm-for-merging instead of gdrm-2.6.32 branch and rebase
May 17 19:25:25 <CIA-2> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
May 17 19:26:55 <Spyzer> khem: vendor provided me qtopia."img" and am trying to burn ".jffs2" can, that create problems even after i set rootfstype=jffs2 in the kernel options??
May 17 19:28:15 <khem> Spyzer: hope the partions are formatted for jffs2
May 17 19:28:43 <khem> Spyzer: do you know what file system is this qtopia.imp
May 17 19:29:19 <Spyzer> as far as i know it is yaffs because when i boot into NOR through minicom it says "Download a yaffs image"
May 17 19:30:00 <khem> then generate a yaffs image from OE too
May 17 19:32:45 <Spyzer> how??
May 17 19:33:18 <Spyzer> as far as i know the only available ones are jffs2 cramsfs tar.gz ect3 ext2
May 17 19:33:24 <Spyzer> *ect3 ext3
May 17 19:33:33 <Spyzer> and squashfs
May 17 19:33:37 <Spyzer> ??
May 17 19:33:50 <Spyzer> khem: is there any way of making yaffs2 image from oe??
May 17 19:33:55 <Spyzer> or yaffs image
May 17 19:33:57 <Spyzer> anyone
May 17 19:35:09 <GNUtoo> mmm
May 17 19:36:08 <khem> Spyzer: IMAGE_FSTYPES += "yaffs2"
May 17 19:36:15 <khem> in your local.conf
May 17 19:36:48 <lgopher> Hi everyone, is there some at least short description what the new and old staging style means ? I'm trying to understand why my package complains that it's doing the legacy style stage and looking at the patches to convert to new style didn't bring me understanding.
May 17 19:38:15 <Tartarus> lgopher: old style == recipe has do_stage, new style == do_install is complete and places everythere where it needs to be so that a package can be installed into the staging directory
May 17 19:38:25 <Tartarus> rather than by hand copying files over in do_stage
May 17 19:40:33 <lgopher> Tartarus: this is exactly what i do i have only do_install and no do_stage. after install step I have in WORKDIR/image/...../usr/bin/slice2cppe. This the only file which I require. And still I got the old style complain. And log shows that do_stage is calling make install anyway.
May 17 19:41:31 <lgopher> Tartarus: I must say that this no autotools but native package. does it make difference ?
May 17 19:42:05 <khem> lgopher: NATIVE_INSTALL_WORKS = "1"
May 17 19:42:13 <khem> add that to recipe
May 17 19:42:55 <lgopher> khem: thanks for the tip, I'm trying this right now
May 17 19:44:20 * khem goes out for lunch hunt
May 17 19:46:36 * GNUtoo tought he already tried the IMAGE_FSTYPES 
May 17 19:46:36 <denix> should I generally bump the PR when changing the LICENSE? i.e. from GPL to GPLv2?
May 17 19:46:50 <woglinde> denix yes
May 17 19:46:59 <GNUtoo> denix,I'd say yes because it changes packages output
May 17 19:47:20 <GNUtoo> hi woglinde 
May 17 19:47:34 <GNUtoo> git pull unbroke my strange g_realloc_n issue
May 17 19:49:34 <Spyzer> khem: i added IMAGE_FSTYPES += "yaffs2" but there was no .img file created??
May 17 19:49:36 <lgopher> khem: for the record, it did fixed the old staging, and another problem with "rpath is too short", which I was complaining yesterday is gone too. Thanks !
May 17 19:49:53 <Spyzer> should i delete the entire tmp and rebuild the console image??
May 17 19:50:04 <denix> I know it changes the output. I was worried some people would not like to rebuild most of their packages when some core pkg changed its license from GPL to GPLv2... it's still GPL :)
May 17 19:50:40 <woglinde> gnutoo lol
May 17 19:50:50 <woglinde> about the g-realoc_n problem
May 17 19:51:28 <woglinde> Spyzer try bitbake -c clean your-image
May 17 19:51:39 <woglinde> and bitbake your-image
May 17 19:55:49 <lgopher> another question, I have a problem with PARALLEL_MAKE. it doesn't work, I see only one core compiling. In the log.do_compile I see in the beginning "NOTE: make -j3 -e MAKEFLAGS=" and this MAKEFLAGS= which makes it run with one core. Where is it set ?
May 17 19:57:30 <GNUtoo> ok
May 17 19:58:46 <woglinde> lgopher seems to come from the original Makefile
May 17 19:59:23 <lgopher> woglinde: not really, I see this for all packages I'm compiling.
May 17 20:00:25 <woglinde> check your env
May 17 20:00:26 <lgopher> woglinde: I can reproduce this in command line perfectly. make -j3 -e -> 3 cores, make -j3 -e MAKEFLAGS= -> 1 core.
May 17 20:00:45 <woglinde> but oe should wipe out it
May 17 20:01:34 <woglinde> sure
May 17 20:01:36 <woglinde> -e, --environment-overrides
May 17 20:01:50 <woglinde> MAKEFLAGS="" overrides
May 17 20:01:53 <woglinde> the -j4
May 17 20:03:32 <lgopher> woglinde: I checked, no, the MAKEFLAGS is not set in environment. the bitbake 1.8.18, debian lenny x86_64, angstrom-2008.1,  if this helps
May 17 20:05:10 <lgopher>  woglinde: BB_NUMBER_OF_THREADS works just fine, so it's not that big problem, but still I get really more than 50% of usage.
May 17 20:05:38 <woglinde> sorryI dont know where -e MAKEFLAGS="" comes from
May 17 20:05:41 <woglinde> never saw it here
May 17 20:09:24 <lgopher> woglinde: no problem, then I have another question :) How do I split package. I would like to have similar package structure as in debian and there one of the packages split in two, namely lib-dev and everything which in the share with completely different name. 
May 17 20:09:46 <woglinde> ??????
May 17 20:09:54 <woglinde> splitting is done automagicly
May 17 20:10:46 <woglinde> otherwise grep the recpies dir for "_files +="
May 17 20:10:47 <lgopher> yep I get lib, lib-dev, doc and I want to package part of the files in separate package.
May 17 20:10:56 <woglinde> or "packages +="
May 17 20:12:03 <lgopher> ok, thanks looking into this.
May 17 20:12:48 <woglinde> lgoper most stuff is also in the manual
May 17 20:12:53 <woglinde> try consider to read it
May 17 20:12:59 <woglinde> not all stuff is update
May 17 20:13:08 <woglinde> but most thinks
May 17 20:13:53 <lgopher> did it already, several times. I'm asking here only the questions which in the manual says to do.
May 17 20:14:22 <lgopher> or maybe I just don't understand what it's trying to say me.
May 17 20:17:12 <CIA-2> 03Denys Dmytriyenko <denis@denix.org> 07org.openembedded.dev * r9e3ba6b92a 10openembedded.git/recipes/bluez/bluez-hcidump_1.42.bb: bluez-hcidump: set the LICENSE to more specific GPLv2
May 17 20:17:13 <CIA-2> 03Denys Dmytriyenko <denis@denix.org> 07org.openembedded.dev * r8a18828d82 10openembedded.git/recipes/bluez/bluez4.inc: bluez4: set the LICENSE to more specific GPLv2/LGPLv2.1
May 17 20:17:17 <CIA-2> 03Denys Dmytriyenko <denis@denix.org> 07org.openembedded.dev * r4b2cf232f8 10openembedded.git/recipes/openobex/openobex_1.5.bb: openobex: set the LICENSE to more specific GPLv2/LGPLv2.2
May 17 20:17:18 <CIA-2> 03Denys Dmytriyenko <denis@denix.org> 07org.openembedded.dev * r2d4704ab46 10openembedded.git/recipes/fbset/fbset-modes.bb: fbset-modes: set the LICENSE to MIT by default
May 17 20:17:19 <CIA-2> 03Denys Dmytriyenko <denis@denix.org> 07org.openembedded.dev * r3e6d9e878c 10openembedded.git/recipes/obexftp/obexftp_0.23.bb: obexftp: set the LICENSE to more specific GPLv2/LGPLv2 (not 2.1?)
May 17 20:17:20 <CIA-2> 03Denys Dmytriyenko <denis@denix.org> 07org.openembedded.dev * re8f9f82923 10openembedded.git/recipes/pointercal/pointercal_0.0.bb: pointercal: set the LICENSE to MIT by default
May 17 20:17:22 <CIA-2> 03Denys Dmytriyenko <denis@denix.org> 07org.openembedded.dev * ra58f2237cb 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev
May 17 20:20:12 <IgorK> Hi, ALL
May 17 20:20:16 <denix> oops, the openobex description says LGPLv2.2 :)
May 17 20:20:53 <IgorK> so is there any way to add the precompiled libraries to the image?
May 17 20:47:35 <Tartarus> hey denix, how do those GPLv2/LGPLv2 work?
May 17 20:47:41 <Tartarus> Is it some package is one part, some is another?
May 17 20:47:45 <Tartarus> Or does the main pkg contain both?
May 17 20:49:59 <woglinde_> Tartarus most packages with this combination has the following
May 17 20:50:04 <woglinde_> lib is lgplv2
May 17 20:50:19 <woglinde_> gui stuff or other stuff based on the lib is gplv2
May 17 20:50:43 <Tartarus> Right
May 17 20:50:48 <Tartarus> And both end up in the same ipk?
May 17 20:50:50 <Tartarus> I assume yes
May 17 20:51:02 <woglinde_> hm ah
May 17 20:51:08 <woglinde_> now I understand your problem
May 17 20:51:09 <kergoth_> that depends on the recipe
May 17 20:51:11 <Tartarus> ie in the gnutls/gnutls-extra case we set LICENSE_${PN}-extra
May 17 20:51:20 <woglinde_> yes
May 17 20:51:21 <Tartarus> kergoth_: exactly
May 17 20:51:24 <woglinde_> we have to fix it
May 17 20:51:26 <woglinde_> okay
May 17 20:51:30 <Tartarus> And hoping denix did that kind of check :)
May 17 20:51:38 <khem> Probably we should packages keeping license in view
May 17 20:51:52 <lgopher> woglinde: I found the reason for make, it's in openembedded/conf/bitbake.conf: EXTRA_OEMAKE = "-e MAKEFLAGS="
May 17 20:52:16 <woglinde_> lgopher hm
May 17 20:52:18 <kergoth_> heh, we should think about killing that default
May 17 20:52:36 <woglinde_> kergoth yes
May 17 20:52:41 <kergoth_> yes, it ensures things obey our vars by default, but it encourages users to not actually read the m akefiles and grasp how to pass things in
May 17 20:52:44 <woglinde_> because we are killing the env anyway
May 17 20:52:46 <kergoth_> which leads to problems later on
May 17 20:52:53 <kergoth_> not sure what you mean.
May 17 20:53:04 <kergoth_> -e ensures our CFLAGS, LDFLAGS, etc override the toplevel makefile's
May 17 20:53:15 <kergoth_> well, all makefiles, but we override MAKEFLAGS so it doesn't get passed to submakes
May 17 20:53:25 <woglinde_> hm okay
May 17 20:53:35 <woglinde_> than I understood the man page wrong
May 17 20:53:37 <kergoth_> -e says "environment overrides file"
May 17 20:53:45 <lgopher> and kills my performance as well
May 17 20:53:54 <kergoth_> ?
May 17 20:55:16 <lgopher> PARALLEL_MAKE=-j4 is overwritten by this, as in log.do_compile  i get "make -j4 -e MAKEFLAGS="  which runs only on one core.
May 17 20:55:23 <khem> Spyzer: it should not be .img but .yaffs2 I guess the extention hardly matters
May 17 20:55:31 <kergoth_> oh, yes
May 17 20:55:36 <kergoth_> we shouldn't be blindly setting MAKEFLAGS=
May 17 20:55:45 <kergoth_> we should ideally be stripping out the -e from the other args
May 17 20:56:01 <kergoth_> but really, as i said before, we should think about dropping teh default, the question is how many recipes would break :)
May 17 20:56:27 <woglinde_> case for TSC?
May 17 20:56:43 <kergoth_> yeah, good idea
May 17 20:57:05 <kergoth_> it does make the common case of creating a recipe slightly easier, but it encourages ignorance on teh part of the person creating the recipe, which isn't good either
May 17 21:25:06 <denix> Tartarus: got your point. some of the recipes I pushed don't provide separate pkgs for lib/app, but some do... I was in a hurry for those I pushed today :) Will try to fix later
May 17 21:31:18 <Tartarus> k
May 17 21:39:29 <woglinde> hm how can split python code over serval lines?
May 17 21:40:16 <khem> woglinde: \ doesnt do it ?
May 17 21:41:18 <Tartarus> \ should, but iirc, indent is still an issue
May 17 21:41:54 <khem> yes indent should be followed all the time no qestion abt that
May 17 21:46:02 <kergoth_> \ isn't necessary if you're within parens or brackets
May 17 21:46:36 <woglinde> yeah thanks
May 17 21:46:38 <woglinde> worked
May 17 21:47:07 <lgopher> woglinde: thanks for the tip, i made it with package split. Now I have next challenge. Is it possible to generate packages in two different architectures from one bb file, namely native and all ? I can make separate bb, but having it in one shot would be much better.
May 17 21:51:22 <woglinde> lgopher look at BBCLASSEXTEND
May 17 21:51:30 <woglinde> ups
May 17 21:51:36 <woglinde> BBCLASSEXTENED
May 17 21:52:08 <lgopher> thanks
May 17 23:20:46 <kergoth_> ugh, there's some really ugly quote usage in qt4.inc's do_install
May 17 23:21:07 <kergoth_> sed -i -e s#"moc_location=.*$"#"moc_location=${bindir}/moc4"# \
May 17 23:24:57 <kergoth_> hmm, don't think the shell handling of quotes within unquoted strings is portable
May 17 23:25:07 * kergoth_ double checks posix/sus
May 17 23:26:41 <Tartarus> Something that should be 's#"..."#"..."#' ?
May 17 23:26:56 <kergoth_> if the quotes are really meant to be part of the pattern adn replacement, yeah
May 17 23:27:02 <kergoth_> sh-4.1$ echo sed -i -e s#"moc_location=.*$"#"moc_location=${bindir}/moc4"#
May 17 23:27:03 <kergoth_> sed -i -e s#moc_location=.*$#moc_location=/moc4#
May 17 23:28:50 <kergoth_> heh, that stuff has been there since before march of 09.. lets see..
May 17 23:31:51 <kergoth_> make that before dec of 07
May 17 23:33:24 <kergoth_> ~seen zecke
May 17 23:33:27 <ibot> zecke <~ich@mobile-out.siminn.is> was last seen on IRC in channel #oe, 2d 14h 3m 30s ago, saying: 'ao2: it is supported.'.
May 17 23:33:47 <kergoth_> now to see if zecke remembers what the original intent was :)
May 17 23:42:39 <ant__> I've seen: grep: /oe/build/tmp/work/i686-linux/bzip2-full-native-1.0.5-r1/bzip2-1.0.5/.pc/bzip2-1.0.5-autoconfiscated.patch/configure.ac: Permission denied
May 17 23:42:51 <ant__> well, no surprise...
May 17 23:42:52 <ant__> http://pastebin.com/ckhvd8WA
May 17 23:43:12 <ant__> build is going on...
May 17 23:43:36 <khem> kergoth_: as it seems fixing broken .pc files  for staging path
May 17 23:44:22 <kergoth_> right, but i don't have one of those handy, so can't really say whether the .pc has quotes in it that need removing :)
May 17 23:44:27 * kergoth_ will kick off a build
May 17 23:45:43 <khem> kergoth_: I dont have one either
May 17 23:46:13 * kergoth_ would assume its behaving as expected, and the quotes were just a funky way to deal with spaces in paths, and shouldn't actually be part of the pattern or replacements, but..
May 17 23:46:28 <kergoth_> in which case its better to quote the entire s## argument, not pieces of it
May 17 23:46:32 <kergoth_> heh
May 17 23:50:47 <khem> kergoth # can be used as separator if there are / to be replaced in pattern
May 17 23:51:05 <khem> which would be the case with dirnames
May 17 23:51:42 <kergoth_> eh?
May 17 23:51:45 <kergoth_> i'm not talking about #
May 17 23:51:49 <kergoth_> i know perfectly well how sed works
May 17 23:51:55 <kergoth_> its using *quotes* inside a string
May 17 23:52:11 <kergoth_> sed -i -e s#"moc_location=.*$"#moc_location=${bindir}/moc4"# \
May 17 23:52:19 <kergoth_> note the quoting of pieces of the inside of a single argument
May 17 23:52:24 <khem> oh I misunderstood
May 17 23:53:10 <kergoth_> so the question was whether they actually thought the quotes were in the .pc and it would match them, or if its just an oddball way to handle spaces in bindir.. either way, the entire s## argument should be quoted as a whole
May 17 23:53:56 <kergoth_> it broke pysh, which apparently can't handle it, though posix may allow it, i can't recall offhand..
May 17 23:54:00 <kergoth_> anyway..
May 17 23:54:02 <kergoth_> ugly :)
May 17 23:54:08 <kergoth_> thankfully simple enough to resolve
May 17 23:55:05 <khem> I think .pc files dont have quotes
May 17 23:57:37 <khem> kergoth_: as you said I think its to encounter spaces in dirname
May 17 23:58:30 <Tartarus> And that blows up other things
May 17 23:58:46 <Tartarus> I think that might have just been funny quoting
May 17 23:59:00 <kergoth_> aye, most likely
May 18 00:00:26 <khem> Tartarus: or gcc commandline options
May 18 00:00:28 <ant__> anybody has workdir of bzip2--full-native to check the permissions?
May 18 00:00:44 <khem> they also need spaces
May 18 00:01:25 <khem> ant__: I do
May 18 00:01:31 <khem> ant__: what do you want to check
May 18 00:01:45 <ant__> http://pastebin.com/ckhvd8WA
May 18 00:02:32 <ant__> I've seen: grep: /oe/build/tmp/work/i686-linux/bzip2-full-native-1.0.5-r1/bzip2-1.0.5/.pc/bzip2-1.0.5-autoconfiscated.patch/configure.ac: Permission denied
May 18 00:02:41 <khem> ant__: mine have read permissions
May 18 00:02:52 <ant__> mine is 0b
May 18 00:03:03 <khem> ant__: you always have wierdest errors with OE I am sorry :)
May 18 00:03:07 <ant__> yes
May 18 00:03:15 <khem> ant__: is umask proper ?
May 18 00:03:22 <ant__> but the good news is I'm building with 4 threads
May 18 00:03:30 <ant__> and it did not stop yet...
May 18 00:03:36 <khem> heh ok
May 18 00:03:53 * khem has to go
May 18 00:04:00 <ant__> I'm pretty sure somebody talked about similar permission issues
May 18 00:04:06 <ant__> we'll see
May 18 00:04:19 <ant__> anyway, good night
May 18 00:15:10 <kergoth_> ah ha! this was a pysh bug.  it chokes on encountering an incomplete variable expansion inside of single quotes, which is a bug, since vars aren't expanded inside single quoted strings
May 18 00:15:26 <kergoth_> "sed -i -e 's,IP{,I${,g $pc" -> boom
May 18 00:15:43 <kergoth_> s/g/g'/
May 18 00:53:15 <dcordes> hi
May 18 00:53:22 <dcordes> what's good ?
May 18 02:05:36 <khem> Tartarus: how did the ppc thing go ?
May 18 02:05:38 <khem> did it boot
May 18 02:19:27 <kergoth_> finally, past qt4-x11-free
May 18 02:19:30 <kergoth_> whew
**** ENDING LOGGING AT Tue May 18 02:59:57 2010