**** BEGIN LOGGING AT Sun May 29 02:59:58 2011 May 29 08:52:54 03Martin Jansa  07master * rd4f0211e20 10openembedded.git/recipes/qt4/ (5 files): May 29 08:52:54 qt4: remove qte.sh from shared SRC_URI and PACKAGE_ARCH_palmpre May 29 08:52:54 Signed-off-by: Martin Jansa May 29 12:59:18 what is the lowest kernel "supported" by oe? May 29 12:59:23 because first I've udev issues May 29 12:59:26 ok I can solve that May 29 12:59:31 but then I've tslib issues May 29 12:59:35 my kernel is 2.6.27 May 29 12:59:51 tslib: Selected device uses a different version of the event protocol than tslib was compiled for May 29 13:00:06 http://patches.openembedded.org/patch/257/ gives some pointers May 29 13:02:43 I'll look what OLDEST_KERNEL is set for May 29 13:02:46 *to May 29 13:04:17 bitbake -e gives: May 29 13:04:18 OLDEST_KERNEL="2.6.16" May 29 13:04:24 mine is 2.6.27 May 29 13:04:30 I'll look what headers are used May 29 13:05:07 PREFERRED_VERSION_linux-libc-headers="2.6.37" May 29 13:05:12 hmmm May 29 13:05:21 normally it would have some fallbacks for older kernels May 29 13:05:44 but I really wonder what happened with theses fallbacks with evdev May 29 13:37:09 hey. does bbappend work in bitbake 1.12.0 ? May 29 13:37:38 i am unsure how to set up a bbappend recipe to search files in both the layer dir and the OE layer dir May 29 13:44:38 CMoH-notebook: http://sakrah.homelinux.org/blog/2010/11/bblayers-bbappend/ May 29 13:45:10 I'm unsure about the version, use master :) May 29 13:45:17 I mean version of bitbake May 29 13:45:28 ynezz, i've done that but it only finds files in my layer and skips those in OE May 29 13:45:47 or the other way round :( May 29 13:46:09 then you've done something wrong May 29 13:46:39 is THISDIR a predefined variable? May 29 13:46:54 i can't grep it neither in bitbake, nor in openembedded May 29 13:47:17 neither FILESEXTRAPATHS May 29 13:51:05 http://fpaste.org/T2kM/ May 29 13:52:07 oh, thanks a lot! May 29 13:52:19 let me check this new info out :) May 29 13:52:52 ts72xx/recipes/netbase/netbase/ts72xx/interfaces May 29 13:53:14 this file is for example then used instead the one in the main layer May 29 13:53:58 and the layer config http://fpaste.org/BB6a/ May 29 13:54:28 ty again May 29 13:54:53 the bblayers seemed to be correct, since it did find both recipes May 29 13:55:38 however, when using the .bbappend recipe like in the blog i could only make it either only use the files in the oe layer, or those in my private layer (meaning i had to copy all the needed files from the oe layer) May 29 13:56:33 that's wrong May 29 13:56:58 try another bitbake version, it's fast these days May 29 13:57:15 there's still one trick on my setup May 29 13:57:22 i don't have both layers under the same topdir May 29 13:59:36 i suppose i should check if that makes a difference, cos it still doesn't work :( May 29 14:05:23 should i clean the whole build dir after each change to bblayers.conf perhaps? May 29 14:12:24 dunno May 29 14:14:20 ty ynezz - something is still not working May 29 14:14:24 actually im May 29 14:14:37 i'm on the 2011.03-maintenance branch on OE... this might be a problem also... May 29 14:16:15 i've seen much of this path computation is done in .bbclass files May 29 14:25:15 well, i guess i'll have to fight this later May 29 14:25:25 ty for the suggesions, ynezz May 29 17:00:08 Quick question, what is the supported version of bitbake for the classic metadata these days (just looking at some non oe-core stuff and I notice that bb git barfs due to a lot of the mandatory error checking). May 29 17:00:23 ? May 29 17:00:30 maintaince 2011.03? May 29 17:00:36 master? May 29 17:02:09 woglinde_: master so oe.dev, meta-oe and oe-core are not far enough along for something I am working on. May 29 17:02:29 latest git works May 29 17:03:02 because I compiling right now something with it May 29 17:03:52 woglinde_: what do I need to do to slacken down the sanity checker? I get legacy stage and license errors all over the shop. May 29 17:04:14 ? May 29 17:04:27 what did you checked out? May 29 17:06:00 New built build box so as usual it takes time to settle things down. http://pastebin.com/s2U3zDqb for example, that is with master bitbake and oe.dev. git://git.openembedded.org/openembedded,master,HEAD and git://git.openembedded.net/bitbake,master,HEAD May 29 17:08:05 hm May 29 17:08:16 maybee I didnt hit a do_stage recipe yet May 29 17:08:19 but just fix it May 29 17:08:53 Seemed that it is taking all the Yocto/Poky sanity rules, I have screwed up somewhere I am sure as based on these sanity rules oe.dev is not building any time soon ;). May 29 17:10:08 woglinde_: well you can't parse the tree even if you hack things about as the old failme.bb has no LICENSE field set and it is the 1st of many. Ho hum, I screwed up somewhere ;). May 29 17:10:20 and fixi aalib is easy May 29 17:10:29 just remove the do_install stuff May 29 17:10:36 ups do_stage May 29 17:11:35 woglinde_: fixing is easy enough, it's the volume of oddness, I am trying to get a handle on is the sanity settings are intentional as I have not seem commits such as 'add missing LICENSE settings' etc. ;). May 29 17:11:54 thats from core May 29 17:12:05 maybee you have some strange checkout May 29 17:12:42 woglinde_: yep, that was my take, ok, so it's me, that's all I wanted to know, I assume I have something stuffed up ;) May 29 17:51:57 khem!!!!!!!!!!! May 29 17:52:07 I just need you now May 29 17:52:27 khem do you know if I can run malta with framebuffer support on qemu? May 29 17:56:45 DJWillis: are you still working with omap3 platform ? May 29 17:57:27 woglinde_: I have not tried it myself but I have read it works May 29 17:57:42 hm kernel need to have enabled it May 29 17:58:07 dcordes: yep, well 3 and 4 a little and some other Atom and ARM11 stuff on the bench at the moment. May 29 17:58:18 khem any pointers? May 29 17:59:10 khem btw. in which layers should qemu machines go? May 29 17:59:14 meta-oe? May 29 17:59:16 core? May 29 17:59:42 woglinde_: here it is http://permalink.gmane.org/gmane.comp.emulators.qemu/69586 May 29 17:59:56 woglinde_: qemus are in core May 29 18:00:05 ah okay May 29 18:00:10 woglinde_: and qemumips is well supported May 29 18:00:27 but we need to add support for mipsel and mips64 May 29 18:01:17 DJWillis: I just remember the openpandora thing :) May 29 18:01:30 woglinde_: in oe-core qemu should just work May 29 18:01:36 they are tested everyday May 29 18:02:01 khem hm he didnt say where he put his patches May 29 18:03:27 woglinde_: hmm yeah I havent explored it **** BEGIN LOGGING AT Sun May 29 18:06:11 2011 May 29 18:07:25 dcordes: still hacking on that (well I am again), took time out of it but on the image/OE side the platform has just stagnated, nobody did anything really so trying to bring it back up together again. May 29 18:08:35 DJWillis: what have people using on the pandora instead of oe ? May 29 18:08:49 win ce May 29 18:08:51 haha May 29 18:09:35 :] May 29 18:09:45 woglinde_: how is ac100 coming along ? May 29 18:09:53 ynezz, still here? May 29 18:10:03 dcordes hm suspend and sound is missing May 29 18:10:07 in .37 kernel May 29 18:10:11 in oe I didnt do much May 29 18:10:18 but mr_nice has now an ac100 too May 29 18:10:41 I ordered one as well today :) May 29 18:11:15 planning on using oe rather than ubuntu for performance reasons May 29 18:11:22 at which price? May 29 18:11:38 dcordes ah and only xorg fbdev May 29 18:11:43 and not gles May 29 18:11:51 180 euros May 29 18:11:57 the hsdpa version May 29 18:12:08 hehe May 29 18:12:14 cyperport.de? May 29 18:12:22 no May 29 18:12:39 I'm not going to advertise for the shop :D May 29 18:16:33 dcordes: they have not really (well there is Debian and a few small projects), people just used the firmware as it shipped on the device mostly as I left the image ;). May 29 18:18:32 woglinde_: I think it is a very good deal for such a device May 29 18:18:42 woglinde_: how much did you pay ? May 29 18:19:19 dont ask May 29 18:19:26 370 but I had a voucher May 29 18:19:28 so it was 320 May 29 18:20:09 Ahhh, the AC100 appeals to hack on if the price is right May 29 18:42:53 hmmm I'm a bit frustrated about still using org.oe.dev and would like to move to oe-core. but I have no idea to start. is there a "oe core for people using classic oe setup" ? May 29 18:43:38 I'm really not interested in running magic-setup-script.sh because I will never understand the layer system May 29 18:44:52 dcordes, not that I've found May 29 18:49:37 DJWillis: what's your reason for not switching to oe-core ? May 29 18:51:20 dcordes: I am, it was an old project that had me switch back, oe-core and meta-oe and meta-angstrom <> openembedded.dev at the moment. The coverage is just not there in terms of recipe volume in oe-core, the quality is better but the volume is also lower ;) May 29 18:52:13 DJWillis: so you're waiting for someone else to do migration of those recipes you need for you? :) May 29 18:53:57 JaMa: no, that's not really fair ;), as I said, I am using it for what I need, and I plan to migrate more. Sometimes it's just the need to balance a release and stable metadata against hacking time ;). May 29 18:54:49 ok.. just wanted to point that it's usually faster and not so hard to migrate what you need May 29 18:56:24 JaMa: the one thing is core/meta that supprised me was the lack of task-propper-tools ;), no, migrating is not hard as all (ok, cheery-picking from dev is a slow process over copy/paste ;)). May 29 18:57:23 oe-core/meta is SO much nicer than the monolithic setup :) May 29 18:57:38 agreed May 29 18:58:45 dcordes use the setup scripts and than look at the resulting files May 29 19:00:03 JaMa: I started to move over all the OpenPandora layer bits that are not mainline-able to OE core/meta layers, that is my current 'pet project' ;) - https://github.com/openpandora - Just never enough free time. May 29 19:02:16 DJWillis: nice.. meta-openpandora looks nice May 29 19:02:38 JaMa: thanks, nice to know it is going on the right lines. May 29 19:04:26 damnit May 29 19:04:37 how does fb works on qemumipsel May 29 19:06:20 One question on the oe-core front, where did things like sdk.bbclass go? have they just not come over or is there a reason (Yocto/Poky) that keeps them out? May 29 19:08:04 DJWillis: nativesdk May 29 19:13:12 JaMa: thanks May 29 19:14:43 woglinde_: where canI find those ? May 29 19:15:45 dcordes what? May 29 19:17:29 dcordes: setup scripts for oe-core? http://gitorious.org/angstrom/angstrom-setup-scripts oe-core branch should get you started I think. May 29 19:24:21 DJWillis: thanks a lot May 29 19:24:33 woglinde_: exactly, something to get me started with oe-core May 29 19:56:22 JaMa, hi I've a question: May 29 19:56:44 I've a 2.6.27 kernel with tslib interface May 29 19:56:55 you did that commit: May 29 19:57:27 http://patches.openembedded.org/patch/257/ May 29 19:58:01 the angstrom linux-libc-headers are 2.6.37 May 29 19:58:12 and finally: May 29 19:58:23 my device prints that: May 29 19:58:24 tslib: Selected device uses a different version of the event protocol than tslib was compiled for May 29 19:58:30 should I switch to evdev? May 29 19:58:36 or is there a better solution? May 29 19:59:19 note that while the patch is quite small for switching the kernel to evdev, the calibration can be problematic May 29 19:59:40 xinput_calibrator expect it to be somewhat initially calibrated May 29 19:59:43 what kernel version do you have in runtime? is it that 2.6.27 or something like that? May 29 19:59:47 yes May 29 19:59:49 2.6.27 May 29 19:59:57 this patch won't help iirc May 29 20:00:02 udev doesn't even work May 29 20:00:10 I've to add compat for udev May 29 20:00:17 it was when you had 2.6.38 in runtime but built with 2.6.37 headers May 29 20:00:20 and to tell -lt 27 May 29 20:00:24 instead of -lt 28 May 29 20:00:52 for the if comparing the kenrel version and that uses udev compat if the kernel is < 2.6.28 May 29 20:01:13 yes I know May 29 20:01:27 I asked what to do since you know that stuff at least a bit May 29 20:01:32 if it works with evdev I would switch May 29 20:01:46 xinput_calibrator doesn't expect precalibrated ts May 29 20:02:10 ok May 29 20:02:21 then I'll switch and hope I don't broke the screen this time May 29 20:02:48 (bug1:if you press too hard or that you open your screen it breaks it) May 29 20:03:24 thanks a lot for the advise May 29 20:03:40 it doesn't work yet with evdev but I know how to switch it May 29 20:03:47 I already did it for dream: May 29 20:04:00 evdev-multitouch -> tslib -> evdev May 29 20:05:35 for udev I've a question too: May 29 20:05:52 it says that a kenrel at least 2.6.27 is required May 29 20:05:56 but mine doesn't work May 29 20:06:04 so there are 2 possibles causes: May 29 20:06:12 *my kernel is not a real 2.6.27 May 29 20:06:28 *it requires a kernel at least 2.6.26 May 29 20:06:33 oops May 29 20:06:45 s/2.6.26/2.6.28 May 29 20:07:13 I wonder if it's likely to have a fake 2.6.27 kernel May 29 20:08:57 maybe I must ask to buglabs May 29 20:09:06 but I don't know what to ask May 29 20:09:10 or what syscalls to check May 29 22:26:21 does OE support creating read-only root file systems ? opkg-cl configure is pretty unhappy when the rootfs is mounted read-only **** ENDING LOGGING AT Mon May 30 02:59:59 2011