**** BEGIN LOGGING AT Fri Sep 09 02:59:57 2011 Sep 09 06:42:02 errordeveloper: What email address should be taken for the patches? The new-synth.conf one or the Google Mail one? Sep 09 09:01:51 pb_: ping Sep 09 09:40:21 hi all Sep 09 10:06:41 hi ant_work Sep 09 10:06:44 hi bluelightning, all Sep 09 10:08:43 pb_: hi, I finally sent an RFC for klibc to oe-core. By chance, pls review. I suspect there could be a missing dep but my builds were fine. Sep 09 10:08:59 sorry for pinging/bothering again ;) Sep 09 10:11:35 righto, I'll take a look at it Sep 09 10:12:25 hi pb_ Sep 09 10:51:38 hi guys i getting checksum error Sep 09 10:51:59 binutils-cross_2.20.1.bb <---- this seems to be the error Sep 09 10:53:05 is there a error with that file? Sep 09 10:54:39 kevinsc: there is patch for that on ML Sep 09 10:55:07 does doing a git pull solve the problem? Sep 09 10:55:16 ML? Sep 09 10:59:49 ML=mailing list Sep 09 11:00:10 so no, you have to apply from patchwork manually if you want Sep 09 11:00:36 how do i patch it? Sep 09 11:01:11 contrib/patchwork/pw-am.sh PATCH_NUMBER Sep 09 11:01:21 or wget + git am Sep 09 11:01:45 i am kinda new to it Sep 09 11:01:51 mind teaching me? Sep 09 11:06:12 kevinsc: here is your patch http://patches.openembedded.org/patch/11245/ and contrib/patchwork/pw-am.sh 11245 in your checkout will apply it.. that's all Sep 09 11:06:22 kaylessa: ^^ Sep 09 11:06:24 kevinsc: sorry Sep 09 11:11:22 so i have to change the file for 2.20.1 and create a new file for 2.20.1a? Sep 09 11:26:28 is there other patch that i need to apply to avoid the checksum error cause by binutil 2.20.1 Sep 09 11:26:29 ? Sep 09 13:17:02 hi, I've that: Missing or unbuildable dependency chain was: ['task-sdk-bare'] when trying to build meta-toolchain-qte Sep 09 13:17:20 so...where can I find that recipe(task-sdk-bare) Sep 09 13:20:38 sorry I did an error Sep 09 14:01:21 * JaMa|Off hates Subject:.*binutils.*checksums.* Sep 09 14:03:01 :) Sep 09 14:03:41 ericben: yours were fine :) Sep 09 14:04:16 JaMa|Off: you show me the thing to do , that was yeasy ;-) Sep 09 14:18:23 * kergoth_ thinks whoever decided to use a symlink for .1/.1a was a moron. better to remove .1 entirely and force people to be aware of why it went away than to change the sources at a url Sep 09 14:20:04 it was removed entirely for a while, then someone "fixed it" Sep 09 14:20:39 that's the proof most people get the thing without verifying what they download Sep 09 14:21:20 or do git --pull from kernel.org.. :) Sep 09 14:22:04 don't know how many machine kernel.org have to reinstall Sep 09 14:22:58 but it seems linuxfundation should put money on the infrastructure side instead of marketing & co ;-) Sep 09 14:55:26 * JaMa|Off hates Subject:.*binutils.*checksums.* even more Sep 09 14:56:22 upstreams should really know better. you never just replace a tarball with a newer version, and a symlink essentially does that Sep 09 14:56:31 * kergoth_ rolls eyes and gets back to work Sep 09 15:07:02 PaulePanter: errordeveloper@gmail.com Sep 09 15:07:50 PaulePanter: I though I have fixed the issue with the local domain from showing up Sep 09 16:20:17 hi, in the qt4 nativesdk in oe.dev we had qmake -v => Using Qt version 4.7.1 but in /usr/local/angstrom/arm/lib but in oe-core there is: qmake -v => Using Qt version 4.7.3 in /home/denis/oe/angstrom-core/setup-scripts/build/tmp-angstrom_2010_x-eglibc/sysroots/x86_64-linux/usr/lib Sep 09 16:24:50 hmmm no one, I'll ask on #yocto Sep 09 17:01:52 JaMa|Off: around? Sep 09 17:04:26 y Sep 09 17:04:37 * JaMa|Off still @work Sep 09 17:05:02 JaMa|Off: as you are working at x11 update I am wondering if you notice any problem with xrandr use inverting mouse axis? Sep 09 17:05:44 otavio: with xorg-xserver or xf86-xserver-lite? Sep 09 17:18:28 JaMa|Off: in fact I am using xserver-xorg of meta-oe Sep 09 17:21:02 then there it should be fixed by http://git.openembedded.org/cgit.cgi/meta-openembedded/tree/meta-oe/recipes-graphics/xorg-xserver/xserver-xorg-1.10.2/randr-support.patch if you're using 1.10.2 Sep 09 17:21:22 1.9.4 is probably broken (see indexes for axis..) Sep 09 17:22:33 and 1.11.0 is using same patch as 1.10.2 so should be fine Sep 09 17:33:00 JaMa|Off: or maybe this is the cause of the problem Sep 09 17:33:31 JaMa|Off: in my case, when I change resolution the X and Y axis get inverted Sep 09 17:35:49 otavio: sorry cannot test on device, but you can try 1.11 and for 90% it's caused/can_be_resolved by this patch Sep 09 17:36:30 JaMa|Off: I will test my current X with this and without this patch to see if it changes something Sep 09 17:36:37 JaMa|Off: will tell you about it shortly Sep 09 17:37:36 JaMa|Off: I just fail to see what this patch tries to fix Sep 09 17:38:14 are you using xinput-calibrator? Sep 09 17:38:23 JaMa|Off: no, I am not Sep 09 17:39:22 otavio: this is to apply xinput changes after rotation properly Sep 09 17:43:19 03Richard Purdie  07master * r67e5e23034 10bitbake.git/lib/bb/runqueue.py: (log message trimmed) Sep 09 17:43:19 runqueue: Ensure task environment is correct Sep 09 17:43:19 This fixes two problems: Sep 09 17:43:19 a) Variables which were in the parent environment but not set as "export" Sep 09 17:43:19 variables in the datastore could end up in the task environment Sep 09 17:43:20 b) oe.environ.update() can't cope with the generator returned by Sep 09 17:43:21 bb.data.exported_vars() Sep 09 18:14:08 03Richard Purdie  07master * r459addf137 10bitbake.git/lib/bb/runqueue.py: Sep 09 18:14:08 runqueue.py: Ensure fakeroot variables are reflected in the datastore Sep 09 18:14:08 Without this, variables can be set to one thing in one part of the environment and something Sep 09 18:14:08 different in another part. This change ensures the datastore and the environment Sep 09 18:14:08 are consistent. Sep 09 18:14:08 Signed-off-by: Richard Purdie Sep 09 19:05:38 khem ping Sep 09 19:28:36 woglinde: yes whats up Sep 09 19:28:56 khem did you try uclibc with meta-oe lately? Sep 09 19:29:43 ummm may be 15-20days Sep 09 19:29:45 whats up Sep 09 19:30:00 I had some problems with libgcc Sep 09 19:31:20 hi khem Sep 09 19:32:18 woglinde: hmm Sep 09 19:32:22 not seen any Sep 09 19:32:33 woglinde: are u using git version or .32 Sep 09 19:32:38 GNUtoo|laptop: hey Sep 09 19:32:56 hi woglinde Sep 09 19:36:09 khem, at work I had an issue with the patchwork, I eared that you had something to do with the patchwork.... Sep 09 19:36:09 basically I suscribed, patchwork crashed with a trackback/trackball and then I can't: Sep 09 19:36:10 *register again (it fails with mail + acount already taken) Sep 09 19:36:10 *login(acount not activated) Sep 09 19:36:12 so now I'm polluting patchwork with my patches Sep 09 19:36:14 since I cannot tell they went in or were superseeded Sep 09 19:36:14 btw long time no seen Sep 09 19:36:34 GNUtoo|laptop: ok let me fix it Sep 09 19:37:44 thanks a lot!!! Sep 09 19:38:03 GNUtoo|laptop: whats your id Sep 09 19:38:09 you registered Sep 09 19:38:09 denis.eukrea Sep 09 19:38:58 GNUtoo id is still active is that right ? Sep 09 19:39:17 * mario-goulart parsed eureka Sep 09 19:39:31 denis@eukrea.com Sep 09 19:39:31 yes it's a different id Sep 09 19:39:31 one is for work Sep 09 19:39:32 one is for personal free software contributions Sep 09 19:39:53 it's on oe-core Sep 09 19:39:58 ok should be fine now Sep 09 19:40:18 thanks a lot!!! Sep 09 19:41:01 now it works, thanks again Sep 09 19:43:06 we get lot of spam and its hard to prevent it Sep 09 19:43:27 so add your name etc so it sounds legitimate Sep 09 19:43:57 ok Sep 09 19:43:59 how? Sep 09 19:44:15 khem I still have it Sep 09 19:44:18 configure: error: cannot compute suffix of object files: cannot compile Sep 09 19:44:20 in profile? Sep 09 19:44:34 onfigure: error: in `/devel/arm/git/setup-scripts/build/tmp-angstrom_2010_x-uclibc/work/armv7a-angstrom-linux-uclibceabi/gcc-cross-initial-4.5-r40+svnr176640/gcc-4_5-branch/build.i686-linux.arm-angstrom-linux-uclibceabi/arm-angstrom-linux-uclibceabi/libgcc Sep 09 19:50:26 GNUtoo|laptop: yes Sep 09 19:50:35 lot of .ru address spam Sep 09 19:50:54 woglinde: hmm thats strange point me to full log somewhere Sep 09 19:51:27 ok Sep 09 19:51:34 but I don't see how to change the name there Sep 09 19:51:38 only how to add mails Sep 09 19:52:41 khem can you test at your side? Sep 09 19:53:26 woglinde: is that angstrom 2010 ? Sep 09 19:53:33 yes Sep 09 19:53:40 tmp-angstrom_2010_x-uclibc Sep 09 19:54:01 ok no promises since I am deep burried in shit these days Sep 09 19:54:12 uhm? Sep 09 19:55:01 what happened? Sep 09 19:55:24 (if you don't want to tell it's ok too) Sep 09 20:00:23 GNUtoo|laptop: oh its work Sep 09 20:00:32 and customers sitting on top of my head Sep 09 20:00:49 ah ok Sep 09 20:01:08 I feared something worse Sep 09 20:01:20 kerne.org is down and build fails Sep 09 20:01:28 since its tryingto do some ls-remote crap Sep 09 20:01:44 I guess there should be a way to keep goin Sep 09 20:02:24 yes we have that problem too at work Sep 09 20:04:17 basically we workarround modifying recipes Sep 09 20:04:51 JaMa|Off: yes; it does fix the issue if I remove this patch Sep 09 20:13:55 khem ah found it Sep 09 20:13:57 error: invalid floating point option: -mfpu=vfp-16 Sep 09 20:14:19 damn why I didnt looked earlier into the config.log Sep 09 20:18:45 hi guys. i'm a newby to arm based platforms and have a problem compiling a i386 based image on arm. i believe its on obvious mistake, but i can't get the point. Sep 09 20:19:32 o.O Sep 09 20:19:42 why you have to compile i386 on arm? Sep 09 20:20:07 woglinde: sorry, no. Sep 09 20:20:22 i can build my image for i386 and it works fine. Sep 09 20:20:33 now i wanna cross compile it formy beagleboard Sep 09 20:20:50 okay Sep 09 20:21:01 because this was a bit missleading -> "have a problem compiling a i386 based image on arm" Sep 09 20:21:26 yeah sorry. that would be even too crazy for me. Sep 09 20:21:28 paepke what do you use oe.dev oe-maintaince or oe-core Sep 09 20:22:03 and why is narcissus not an option for you` Sep 09 20:22:04 ? Sep 09 20:22:16 http://narcissus.angstrom-distribution.org/ Sep 09 20:22:51 woglinde: i use 2011.3-maintenance Sep 09 20:23:46 woglinde: i can build that angstrom image without problems. Sep 09 20:23:53 okay Sep 09 20:24:40 but where is the problem? Sep 09 20:24:41 I'm from the thinstation.org project which has a thinclient build environment based on self-developed scripts. Sep 09 20:25:28 woglinde: some years ago someone recreated that stuff in oe. Sep 09 20:25:46 ? Sep 09 20:26:11 ok but you said you had a problem compiling at the beginning, and then you say you can compile an angstrom image without problems, so where's the problem? Sep 09 20:26:37 yeah I am puzzled too Sep 09 20:26:43 and oe is over 10 years old Sep 09 20:27:22 * GNUtoo|laptop remembers monotone times Sep 09 20:27:28 or whatever there was before git Sep 09 20:27:43 monotone wasn't very practical Sep 09 20:28:00 GNUtoo|laptop, woglinde: the problem is: That thinstation "distribution" based on oe builds well for i386. now i wanna build it on arm. Sep 09 20:28:25 ok Sep 09 20:28:47 so I'll re-tell my question Sep 09 20:28:47 where does it fail or what's the problem Sep 09 20:28:59 i used angstrom to test if i have something broken on oe itself. it builds, so it doesn't look like i broke the oe or bitbake. Sep 09 20:29:25 GNUtoo|laptop: ok. let me describe. it builds up to glibc where i get an error. Sep 09 20:29:51 o.O Sep 09 20:30:02 oe is toolchain Sep 09 20:30:05 if ups glinbc Sep 09 20:30:16 if glibc dont builds nothing else will Sep 09 20:30:25 besides native packages Sep 09 20:30:33 and binutils and gcc-cross Sep 09 20:30:46 woglinde: ok. sorry if i misuse the terms. Sep 09 20:31:11 woglinde: yes, nothing else builds. i'm at the beginning. Sep 09 20:31:21 paepke, ok then what's the error while building glibc Sep 09 20:31:24 so oe is not building Sep 09 20:31:25 i spent already hours. but did not get the point Sep 09 20:31:37 pastebin is ok i assume? Sep 09 20:31:42 yes Sep 09 20:31:58 but you should pastebin your command Sep 09 20:32:05 and your local.conf too Sep 09 20:32:57 yes Sep 09 20:32:59 but also paste your distro configs Sep 09 20:33:08 log.do_configure: http://pastebin.com/HNj1AYTj Sep 09 20:34:56 distro conf: http://pastebin.com/yD5N7XAT Sep 09 20:35:11 btw: my appologies if there will be a lot of mistakes in it. Sep 09 20:35:28 do what the log said Sep 09 20:35:29 that is to say: Sep 09 20:35:29 See `config.log' for more details. Sep 09 20:35:41 hi GNUtoo|laptop you got in time for the train ? Sep 09 20:35:54 distro features Sep 09 20:36:05 ericben, yes I catched the train Sep 09 20:36:19 local.conf http://pastebin.com/5xGNmkhB Sep 09 20:36:36 distro features seem highly insufficent Sep 09 20:36:42 you lack the glibc features Sep 09 20:36:42 good. still fighting with qmake, I think I'll post to oe-core ml with details incase someone has an idea Sep 09 20:36:46 and similar things Sep 09 20:36:49 ok Sep 09 20:37:12 GNUtoo|laptop: ok. RTFM will be ok if you can point me to some docs Sep 09 20:37:25 paepke, grep will be better Sep 09 20:37:30 grep what the other distro do Sep 09 20:37:59 GNUtoo|laptop: ok. reading the code of others? Sep 09 20:38:15 where should i take a deeper look? angstrom? Sep 09 20:38:20 what would you suggest Sep 09 20:38:28 angstrom distro config is good Sep 09 20:39:09 but usually the question to ask is : Sep 09 20:39:17 do you really need to make your own distro? Sep 09 20:39:30 isn't angstrom ok with some minimal rebranding? Sep 09 20:40:04 that would be good for a start Sep 09 20:40:41 I would also suggest to use angstrom as distro since it's a good distro, and stay away from micro or uclibc at first Sep 09 20:40:48 GNUtoo|laptop: the concept of thinstation is to have a central configuration wich defines config files and system behaviour. thats why someone in the past decides to create an own distro. Sep 09 20:41:25 GNUtoo|laptop: and the i386 version builds well. and works well. Sep 09 20:42:00 primarly i wanna have a look its possible to convert it to another platform like arm. Sep 09 20:43:02 thats why i bougt an beagleboard about a year ago. Sep 09 20:44:37 ok Sep 09 20:44:42 look at the libc features Sep 09 20:44:54 I wont use .37 for beagle on maitance Sep 09 20:44:59 use 32-psp Sep 09 20:45:02 GNUtoo|laptop: ok. i'll take a look with the libc and distro features. Sep 09 20:45:02 or distro features that look like they are from libc Sep 09 20:45:04 kernel Sep 09 20:45:37 woglinde: ok. thx. i just copied that 37 kernel from angstrom.conf Sep 09 20:48:25 woglinde: is it ok to use 2011.3. thought it is a good idea cause its stable. Sep 09 20:51:00 not all is stable in maintaince Sep 09 20:51:56 good nite Sep 09 20:54:30 good night Sep 09 21:09:38 GNUtoo|laptop: ok compiling again... another question. is it still needed if you want the full graphic performance to get a driver/source from TI ? Sep 09 21:09:53 i read about it long time ago... Sep 09 21:10:26 define full graphic performance Sep 09 21:10:30 because: Sep 09 21:10:44 1) 3d is not mandatory to get full decent graphic performance Sep 09 21:11:03 xf86-video-omapfb even have xv acceleration Sep 09 21:11:29 2) 3d is now freely redistributable for the imagination technology powervr devices Sep 09 21:12:08 I would advise to try xf86-video-omapfb Sep 09 21:12:20 and not to use non-free 3d Sep 09 21:12:21 GNUtoo|laptop: 1) 3d isn't necessary. but 2d of course. Sep 09 21:12:30 GNUtoo|laptop: 2) nice to hear. Sep 09 21:13:03 GNUtoo|laptop: xf86-video-omapfb) great to hear. Sep 09 21:14:33 GNUtoo|laptop: i was impressed by the libav guys on linuxtag 2011. but did not have time to talk to them. Sep 09 21:15:14 in qmake_base.bbclass, there is addtask generate_qt_config_file after do_patch before do_configure Sep 09 21:15:46 how can I have a different generate_qt_config_file for virtclass-nativesdk ? Sep 09 21:15:54 hmm Sep 09 21:16:03 I tried do_generate_qt_config_file_virtclass-nativesdk() { but that doesn't work Sep 09 21:16:14 it seems to always take do_generate_qt_config_file Sep 09 21:16:53 that sounds worrisome Sep 09 21:18:03 at least I learned a word ;-) Sep 09 21:18:42 this is with angstrom setup script and oe-core Sep 09 21:19:38 Well, lets try a hack Sep 09 21:19:54 If you do real_d_gen_qt_…() { .. current … } Sep 09 21:20:13 real_d_gen_qt_virtclass-native() { … right for native … } Sep 09 21:20:25 and just have the function call real_do_gen_... Sep 09 21:20:33 sorry Sep 09 21:20:34 does it get the right stuff Sep 09 21:20:55 seems qmake_base.bbclass is present in both meta-openembedded AND openembedded-core Sep 09 21:21:01 Oh that one Sep 09 21:21:07 yes, that's also annoying at times :( Sep 09 21:21:08 so it must be taking the one of meta, while I'm changing core Sep 09 21:21:22 I don't understand why there are so much duplicates with oe core Sep 09 21:21:38 the number of linux.inc for example Sep 09 21:21:53 each bsp has his one copied on an other one and hacked for its own purpose Sep 09 21:22:07 sorry for the noise, I'm going to change the other one an try again Sep 09 21:22:14 oe.dev was not easy for beginer Sep 09 21:22:33 oe-core is good to help consultant earn their life :-D Sep 09 21:25:14 Well, the point is that higher levels can say "no, this is right" and override things Sep 09 21:25:39 But with classes, it gets tricky since we (and RP has convinced me before) don't want to allow a kind of bbclassappend Sep 09 21:26:10 And then yes, there's still some sorting to be done for getting the kernels all sorted out again Sep 09 21:27:08 I understand that but it make things harder to debug Sep 09 21:27:55 ok changing in the right file did the trick Sep 09 21:28:52 yeah, I also agree there, it's tricky to get the right balance Sep 09 21:28:57 that fun how asking on IRC after searching for a few hours can help to find the stupid thing that was missed during these search hours ;-) Sep 09 21:29:06 fwiw, I'd been doing just oe-core only builds when debugging / fixing something in the core Sep 09 21:29:08 to avoid that Sep 09 21:29:29 that seems a good solution Sep 09 21:29:47 to isolate the problem Sep 09 21:30:04 we started on oe-core only last week so we are in the learning curve Sep 09 21:30:32 but our first BSP is nearly ready, only this meta-toolchain-qte problem to solve Sep 09 21:33:04 ericben, how did you get started? Sep 09 21:34:12 with angstrom scripts and GNUtoo|laptop 's work Sep 09 21:35:03 first goal is to get at least what we had in oe-dev Sep 09 21:35:16 where is GNUtoo|laptop 's work? Sep 09 21:36:00 Crofton|work: it's the time he spent to work in my company since 1st of this month ;-) Sep 09 21:36:08 :) Sep 09 21:36:08 ok Sep 09 21:36:40 some patch will follow to add things we were missing Sep 09 21:37:02 and I hope we will be able to write a little bit of doc Sep 09 21:37:40 thanks Sep 09 21:37:50 I need to get started also "soon" Sep 09 21:37:52 but angstrom's scripts are a very good starting point Sep 09 21:38:05 I'll be at a conference most of next week though :( Sep 09 21:38:18 are you going to elce? Sep 09 21:38:27 as our first build without angstrom's script generated a non working qt (and we don't yes know why) Sep 09 21:39:31 I'll know only at the last minute, that depends of many variables that I still have to fill ;-) Sep 09 21:39:53 I just registered, so i hope to meet you there Sep 09 21:40:25 but I'll do my best to go there to meet OE community Sep 09 21:42:53 hey, does anyone know of a replacement/mirror for the linux-omap repo that was hosted on git.kernel.org? Sep 09 21:45:23 notespace: git://gitorious.org/chase-linux/linux-omap.git Sep 09 21:45:48 I happened to run across http://mirror.nexcess.net/kernel.org/scm/linux/ , but I don't know how current it is or anything Sep 09 21:46:04 ah great! Sep 09 21:46:35 I think I might have found one here also http://github.com/tmlind/linux as well Sep 09 21:46:43 notespace: is was synced on 29st of august Sep 09 21:47:12 ok cool, thanks Sep 09 21:48:16 the github one is the one of a kernel maintainer so if you want current work you should take this one Sep 09 21:48:32 (kernel maintainer for omap) Sep 09 21:49:20 yeah I saw that, thanks Sep 10 02:53:35 hi anyknow what i should bitbake if i just wan a cross-compiler that create .bin for my target board? **** ENDING LOGGING AT Sat Sep 10 02:59:58 2011