**** BEGIN LOGGING AT Mon Oct 10 02:59:57 2011 Oct 10 06:55:39 good morning Oct 10 08:31:21 morning all Oct 10 11:30:12 what i the recommended way in creating a new kernel buid recipe ( e.g linux-3.0.4.bb with custom config ) Oct 10 11:30:46 s/what\ i/what\ is/ Oct 10 12:01:59 Hi, is there a way to exclude a package from an image, that was included via RRECOMMENDS from another package? Why are RRECOMMENDed packages included in an image anyway? Oct 10 12:03:02 there is BAD_RECOMMENDATIONS, but at least for me it doesn't work Oct 10 12:06:21 Thanks, i will try that. Oct 10 12:19:48 HI all Oct 10 12:20:05 while building bluez4 I am getting this error ERROR: QA Issue: non debug package contains .debug directory: bluez4 path /work/armv7a-vfp-neon-mgc-linux-gnueabi/bluez4-4.96-r2/packages-split/bluez4/lib/udev/.debug/hid2hci Oct 10 12:21:21 secondly I think bluez sources are not placed on kernel.org Oct 10 12:21:40 http://www.kernel.org/pub/linux/ does not have bluetooth dir now Oct 10 12:22:09 we need to update the bluez4 and bluez-hcidump recipes Oct 10 12:24:25 bluez.org still poinsto to kernel.org for downloads.. so I guess there still isn't whole pub mirror populated after that security breach Oct 10 12:42:08 hhhmmm that means we have to wait Oct 10 13:11:07 Hello Oct 10 13:11:23 I am having trouble to generate one image that works with getty Oct 10 13:11:38 my sysvint keeps saying it cannot execute getty Oct 10 13:11:44 any clue? Oct 10 13:34:21 otavio: and do you have getty installed in /sbin? Oct 10 13:34:36 * otavio checks Oct 10 13:34:40 otavio: you can have agetty only or in different path.. I had those too Oct 10 13:36:25 JaMa: no getty; only agetty Oct 10 13:36:32 JaMa: any idea why? Oct 10 13:36:55 depends on image and task Oct 10 13:37:38 JaMa: I didn't find where it use alternatives to the getty Oct 10 13:37:41 otavio: task-core-basic.bb: mingetty \ Oct 10 13:38:17 but I don't know if your image depends on task-core-initscripts Oct 10 13:38:21 i try to start an oe-build in quemu and it fails at init of the framebuffer (SDL). even i have a bunch of SDL installed, there is neither /dev/fb0 nor /dev/fb/0 (debian) where could i start to dig ? Oct 10 13:39:31 JaMa: I was depending on u-l-agetty Oct 10 13:40:10 then you have to alter inittab in your images Oct 10 13:40:32 or add u-a providing /sbin/getty to mingetty and u-l-agetty Oct 10 13:40:42 which is IMHO better Oct 10 13:41:13 JaMa: I am rebuilding with mingetty Oct 10 13:41:16 * otavio checks Oct 10 13:42:22 JaMa: ahh easy Oct 10 13:42:32 JaMa: it is on ready-only fs Oct 10 13:42:43 JaMa: so u-a cannot be done Oct 10 14:25:06 does sdl relay on a kernel-framebuffer ? Oct 10 14:36:02 lautriv: somewhat OT but it is capable of framebuffer output but can also use X or one of several other methods Oct 10 14:38:19 eFfeM_work: ping Oct 10 14:39:30 bluelightning, not absolute OT because i try to emulate oe in qemu :D Oct 10 14:40:33 lautriv: ah right, I see :) FYI I often view qemu output via ssh-forwarded X11 Oct 10 14:40:46 not the fastest, but adequate for testing Oct 10 14:41:06 and saves me from switching the keyboard/monitor over to my build box :) Oct 10 14:54:33 pff....disconnected Oct 10 14:54:35 bluelightning, in my case qemu runs on the right box and complains about missing /dev/fb* . because i don't need qemu and SDL for anything else, i have to fiddle with it the first time and now i assume i have to reboot/load some fb. Oct 10 14:55:03 lautriv: is DISPLAY set? Oct 10 14:55:41 lautriv: SDL also has some environment variables that control its output: http://www.libsdl.org/docs/html/sdlenvvars.html Oct 10 14:56:55 bluelightning, it runs all from scripts delivered by angstrom and because SDL will try to access (nonexistent) /dev/fb* i guess that will not help. Oct 10 14:57:52 lautriv: well, it shouldn't care if /dev/fb* exist if it's forced to output to X11... Oct 10 14:58:45 that's true but changing the env is NOT encouraged. did you try this ? Oct 10 14:59:07 no I haven't, not with QEMU recently at any rate Oct 10 15:02:46 ok, changed --> export SDL_VIDEODRIVER="x11" gives me no avilable video-device. Oct 10 15:03:12 and DISPLAY is set? Oct 10 15:03:19 yes Oct 10 15:03:29 hmm, that is odd... Oct 10 15:03:40 echo $DISPLAY --> :0.0 Oct 10 15:03:52 try :0 Oct 10 15:05:51 nope Oct 10 15:07:03 strange anyway because i modprobed first uvesafb and then fb and both did not bringup any /dev/fb. Oct 10 15:07:32 vesa is for x86 Oct 10 15:08:08 GNUtoo, sure but i need at least a fake to get qemu catching any fb Oct 10 15:08:16 ok Oct 10 15:08:41 uvesafb is for userspace vesafb if I understoof well Oct 10 15:09:04 maybe try a less complicated setup with the old 100% kernel based vesafb Oct 10 15:09:22 <-- hates emulation/virtualisation and usually runs on the right hardware :-) Oct 10 15:10:01 ok Oct 10 15:11:37 so, the only thing what brings a /dev/fb up is vga16fb, lemme see what happens now :-) Oct 10 15:15:58 ok, looks like my nvidia quadro fx and the propietary drivers are the problem :-(( Oct 10 15:16:31 cbrake, hi are you the one in charge of adding ssh keys to the openembedded server? Oct 10 15:17:06 lautriv, nouveau doesn't work? Oct 10 15:17:28 so you run it on real hardware or in qemu? Oct 10 15:18:33 GNUtoo: yes, I can do that Oct 10 15:19:07 ok because I had a problem with my hard disk this summer and my ssh key was destroyed among all my data Oct 10 15:19:31 and I can't push to contrib/shr branch because of that Oct 10 15:19:45 GNUtoo, Nouveau targets mainly the GForce, Nvidia/Xorg runs on a x86_64 SMP system where qemu should emulate the arm, rootfs is bilt and mounts via NFS but SDL fails. Oct 10 15:20:02 *built Oct 10 15:20:19 lautriv, ah sdk of qemu, not sdl inside qemu Oct 10 15:20:26 no idea Oct 10 15:21:07 that is ........i'll do some steps with my dogs, brings new ideas (hopefully) Oct 10 15:21:14 ok Oct 10 15:24:01 cbrake, should I send you a new key trough mail? Oct 10 15:24:22 GNUtoo: yes, please Oct 10 15:24:29 ok thanks a lot Oct 10 15:28:23 done Oct 10 17:02:53 Have someone make dbus postinst runnable at rootfs building? Oct 10 17:09:39 otavio: looks like it would be at least partially possible, and worthwhile effort... Oct 10 17:12:32 bluelightning: I never played with the useradd class Oct 10 17:12:43 bluelightning: are you willing to take a look on that? Oct 10 17:13:40 otavio: I haven't used it yet either... I probably have too much to do in the time coming up to ELCE I'm afraid Oct 10 17:15:12 * otavio fail to see how mingetty is integrated onto the image Oct 10 17:15:24 mingetty doesn't call update-alternatives Oct 10 17:16:40 otavio: I wouldn't have thought you'd use alternatives for something like getty, you wouldn't expect both to be installed at once Oct 10 17:16:59 bluelightning: but there's something wrong Oct 10 17:17:18 sysvinit-inittab sets it to /sbin/getty but noone provides it Oct 10 17:54:59 GNUtoo, still around ? Oct 10 17:55:54 yes Oct 10 17:56:15 dude, you made my day, great thanks :-)))) Oct 10 17:59:28 GNUtoo, i was searching the error for my qemu/SDL problem and you mentioned nouveau. i was aware of this driver and perpetually had a look on it but my card (QuadroFX) was never supported. but after you mentioned it again, i thought "heh, it's kernel 3.0 maybe they improved so give it a shot" did so and success..........now running 3360x1080 on 2 24" Touchscreens very smooth :-) Oct 10 17:59:58 wow nice Oct 10 18:00:14 and with qemu on SDL Oct 10 18:01:10 just need to find the angstrom way of keyboard settings and can compile native code. Oct 10 18:49:54 GNUtoo: module-autoload still works as before just double check if its g_ether or g-ether in used kernel Oct 10 18:51:07 ok btw I want to make 2 changes in meta-smartphone Oct 10 18:51:36 1)add bluetooth support for om-gta02 defconfig => do I have to add it for om-gta01 too, I cannot test Oct 10 18:51:45 2)add bluetooth to modules autoload Oct 10 18:51:52 bluetooth is composed of at least 3 modules Oct 10 18:52:04 hidp,bnep,rfcomm Oct 10 18:52:10 and btusb that was already there Oct 10 18:52:55 again for 2) I've no gta01 to test Oct 10 18:53:21 JaMaOff, ^^^ Oct 10 19:27:13 hmm, gcc-runtime-nativesdk 4.6 fails to build, and 4.5.1 fails due to the use of -nostdlib++ Oct 10 19:27:20 khem: tried building the nativesdk stuff before? Oct 10 19:31:49 kergoth: no I did not try it in long time Oct 10 19:32:01 what errors do u get Oct 10 21:04:23 otavio: hey, do you have a recent asterisk pbx recipe? Oct 10 21:04:53 zecke: hey Oct 10 21:04:57 zecke: not really Oct 10 21:08:52 otavio: thanks. :) Oct 10 21:11:14 zecke, you updating the asterisk recipe? Oct 10 21:15:09 perhaps there could be a meta-telephony layer...? Oct 10 21:16:08 I am having a hard time to get image-live booting ... crap ... Oct 10 21:23:45 bluelightning, too many layers ..... Oct 10 21:28:14 Crofton: maybe not... not a big fan of asterisk... i need it for a quick hack. Oct 10 21:28:35 Crofton|work: I think more layers are fine as long as they are maintained and reasonably focused Oct 10 21:29:10 heading to bed. Oct 10 21:29:14 cheers Oct 10 21:36:46 yeah, I am just having a hard tiem finding stuff Oct 10 21:37:34 Crofton|work: isn't there a wiki page to track what layers are where now? that's a start, anyway.. would be nice to have a search engine for them, though.. Oct 10 21:37:35 heh Oct 10 22:10:40 khem: | work/i686-nativesdk-oesdk-linux/gcc-runtime-nativesdk-4.6.1+svnr178924-r16/gcc-4_6-branch/build.i686-oesdk-linux.i686-oesdk-linux/libstdc++-v3/include/tuple:97:29: error: parameter 'std::__future_base::_State_base* const& __h' is not of literal type Oct 10 22:31:11 kergoth, we have a huge problem with stale wiki pages Oct 10 22:31:31 true, but this one in particular was updated quite recently, afaik Oct 10 23:31:31 kergoth: Whats your build host ? and what version of gcc is installed Oct 10 23:32:39 Crofton|work: How is your oe-core port going on Oct 10 23:32:49 ready with 7 layer burrito yet :) Oct 10 23:35:39 I've started to id missing bits Oct 10 23:35:42 which is good Oct 10 23:39:20 khem: old :) I do a fair number of builds on an old centos 5.4 box, mostly so the sstate packages are most likely to be reusable on as many other hosts as possible Oct 10 23:39:29 i'll give it a shot on ubuntu 11.04 Oct 10 23:44:43 kergoth: usually it should not be a problem but I have to understand the interaction a bit here Oct 10 23:44:59 kergoth: can you post the complete compile log somewhere Oct 10 23:45:06 I didn't even consider that it might be the build machine's toolchain that was the problem Oct 10 23:45:22 going to verify that it doesn't happen on another machine, then will wrap up a log Oct 10 23:47:13 ok Oct 10 23:47:42 mostly, its like that libstdc++ libgcc are tightly coupled to gcc versions Oct 10 23:47:58 but then it should have used gcc 4.6 to compile the file in question Oct 10 23:48:05 if not then we have to figure out why Oct 10 23:49:06 * kergoth_ nods Oct 10 23:49:17 okay, thanks for the help, will probably come back with more info tomorrow Oct 10 23:51:13 ok Oct 10 23:51:13 ok Oct 10 23:52:35 crappy wireless **** ENDING LOGGING AT Tue Oct 11 02:59:57 2011