**** BEGIN LOGGING AT Tue Apr 09 02:59:58 2013 Apr 09 06:51:24 hii all Apr 09 06:52:19 when i'm building some modules i;m getting this error "ERROR: QA Issue: Architecture did not match (3 to 40)..." what might be the problem.. if i do make clean- and buil individually then it was building properly Apr 09 06:52:46 any body faced this issue before? Apr 09 07:30:55 morning all Apr 09 07:32:35 good morning Apr 09 07:44:45 good morning Apr 09 07:45:29 hi silvio__, mckoan Apr 09 08:00:47 hi apelete, mckoan Apr 09 08:30:47 morning all Apr 09 08:35:15 hen i'm building some modules i;m getting this error "ERROR: QA Issue: Architecture did not match (3 to 40)..." what might be the problem.. if i do make clean- and buil individually then it was building properly Apr 09 08:35:30 any body faced this issue before? Apr 09 08:37:02 Sj: usually that's incorrect flags being passed to the compiler Apr 09 08:37:15 Sj: when you say "module" what do you mean? Apr 09 08:37:39 soo is it problm with .bb file? Apr 09 08:37:53 could be yes Apr 09 08:38:20 ok i will check it/// thank you blue :-) Apr 09 08:38:42 it could be that the build scripts of whatever you are building are picking up flags from the host system Apr 09 08:46:15 hi bluelightning Apr 09 08:46:30 hi apelete Apr 09 08:51:11 it is not the problem with flags @blue.. beacuse i'm not compiling the module. it is prebuilt just i'm copying those shared libs.. there it is gicing error like Architecture did not match (3 to 40) for .so Apr 09 08:51:22 giving* Apr 09 08:51:45 i have made do_package_qa() dummy function in my bb file.. :P Apr 09 08:53:15 Sj: the correct way to do that is to add INSANE_SKIP_${PN} = "arch" Apr 09 08:53:27 assuming you know the warning is OK to ignore Apr 09 08:54:03 usually unless it's firmware it's indicative of a problem you need to fix Apr 09 08:56:09 if you do INSANE_SKIP_${PN} = "arch" then will do_package() function be executed or not?? Apr 09 08:56:33 it will yes Apr 09 08:56:33 bear with my question.. i'm lil bit new to thi OE- stuff Apr 09 08:56:40 ohh ok .. Apr 09 08:56:51 it will just disable the architecture check Apr 09 08:57:13 (for the ${PN} package) Apr 09 08:57:38 ohh ok.. shall i know what other checks i can skip.. i mean like some times bad RPATH, etc Apr 09 08:58:17 your first approach to any QA error should be "how can I fix this" rather than "how can I ignore this" ... Apr 09 09:02:43 having said that all of the qa check names can be found in meta/classes/insane.bbclass Apr 09 09:02:50 yeah but some bad RPATH errors i'm getting like "some files in /lib are linking to /use/lib" But i need the same way for my project.. so i'm doing totall qa() function dummy in the bb file.. " " Apr 09 09:03:12 hi bluelightning Apr 09 09:03:28 yeah insane.bb i have changes and removes some checks from the array :P Apr 09 09:03:35 Sj: oh right, well that one can probably be disabled globally; in fact I thought we did disable that one out of the box Apr 09 09:04:14 Sj: you don't need to edit the bbclass, you can set WARN_QA and ERROR_QA in your distro / local config Apr 09 09:04:35 Sj: you should avoid editing files under meta/ so that you can easily upgrade Apr 09 09:04:56 @blue So, You seem to be expert in OE-core.. even i heart u were very helpful to some of my collegues.. so good of you Apr 09 09:05:11 most of the time you don't need to, changes can be applied elsewhere (e.g. distro config, using bbappends in another layer, etc.) Apr 09 09:05:23 ohh ok.. Apr 09 09:05:37 bluelightning: speaking of QA issues, I have the following warning at the end of a build: Apr 09 09:05:37 WARNING: QA Issue: linux-jlime-ben-nanonote: Files/directories were installed but not shipped Apr 09 09:05:39 /lib/modules/2.6.36/kernel Apr 09 09:05:48 Sj: I'm not an expert in every aspect but I am fairly knowledgeable and try to help out folks where I can Apr 09 09:06:03 what i have done is i have copied the insane bbclass into my lmeta-layer/classes.. so it is making use of mine Apr 09 09:06:21 bluelightning: is that something I should worry about ? (http://paste.debian.net/248401/) Apr 09 09:06:22 apelete: is that an empty directory? if so you can just rmdir it at the end of do_install Apr 09 09:07:07 Sj: you shouldn't need to do that, you should only need to set WARN_QA and ERROR_QA in your config Apr 09 09:07:30 ok.. i will try to do.. Apr 09 09:09:11 bluelightning: yes, tmp-eglibc/sysroots/ben-nanonote/lib/modules/2.6.36/kernel is empty indeed Apr 09 09:15:14 I guess that's what's causing the QA warning Apr 09 09:58:50 can anyone tell me if there is going to be a move to gst in the short term? I'm looking for a recipe before I go off and put together my own. Apr 09 09:58:59 *gst 1.0 Apr 09 10:04:24 adding it to 1.5 is under discussion. in the meantime, I wrote some recipes. Apr 09 10:04:25 https://github.com/dv1/meta-gstreamer1.0 Apr 09 10:12:03 hi all Apr 09 10:12:26 is this a good location to ask questions about meta-fsl-arm ? Apr 09 10:17:30 florian: did you change your kernelconcepts.de mail server recently? Apr 09 10:18:47 pb_: hmm... not really. we moved to a new machine a while ago... Apr 09 10:19:33 oh right, fair enough. you just got kicked from oe-core because your server is bouncing list mails with "550 Message headers fail syntax check". Apr 09 10:22:27 pb_: again? I remember I hat this issue some days ago and whondered why (because that's the only list with trouble like this). Apr 09 10:22:45 oh Apr 09 10:22:52 I see... really Apr 09 10:24:58 otavio, hmm, you may seem to be the person I am looking for :) -- have you ever figured out what causes https://lists.yoctoproject.org/pipermail/meta-freescale/2012-November/000036.html - I'm having the same problem.... Apr 09 10:25:58 pb_: thanks for the hint Apr 09 10:26:10 any time :-) Apr 09 10:46:16 rink_: sure, please ask your question Apr 09 10:46:33 ah, I see that you did, heh Apr 09 10:47:10 indeed otavio is probably the best person to respond to that Apr 09 10:49:15 During compilation on my toolchain I had this trouble: Apr 09 10:49:20 /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory Apr 09 10:50:14 looking online, i see that could be an error from ia32 if using a system at 64bit, but I m using a 32bit system Apr 09 10:51:16 someone encourred in a similar issue? Apr 09 11:08:26 bluelightning, thanks for fixing the libudev + sdk thing Apr 09 11:08:31 that was "weird" Apr 09 11:08:56 bluelightning, thanks, i'll /query him in case he misses it. no idea how busy this channel is :) Apr 09 11:23:05 rink_: yes Apr 09 11:23:40 rink_: this happens when the kernel module for the gpu is not loaded. In our config we set it as built-in to avoid loadng it byhand Apr 09 11:28:12 pb_: http://pastebin.com/4y9uT95G Apr 09 11:28:33 pb_: can you check if you understands it from commit log now? Apr 09 11:48:48 otavio, fwiw, I'm not exactly using OpenEmbedded - i'm using buildroot *hides*, but the code should be the same Apr 09 11:50:22 but modules, hmm, that's a pretty good idea. How on earth did you find this?! Apr 09 11:56:04 what's the name of the GPU driver anyway (i.e. what do i need to load?) Apr 09 12:20:24 dv_: thanks, getting some python errors with denzil, think it is down to spacing, will investigate. Apr 09 12:23:42 ah. Apr 09 12:24:02 right, spacing. thanks for bringing it up. I'll correct this Apr 09 12:32:06 rink_: hehe Apr 09 12:32:21 rink_: I don't recall of hand. Apr 09 12:32:24 rink_: let me find out Apr 09 12:33:08 rink_: CONFIG_MXC_GPU_VIV=y Apr 09 12:36:00 ah yeah, I have that as Y but... I have a static /dev, and it seems the code tries to make a device Apr 09 12:36:19 so it can't find the device :) Apr 09 12:38:04 rink_: :P Apr 09 12:38:35 rink_: except if you have a real need to not do to so, use udev/mdev Apr 09 13:24:16 otavio, yeppee, using udev resolves the problem. Apr 09 13:27:29 thanks! Apr 09 13:29:25 I am including gst-meta-base in my image .... rootfs complains about following Apr 09 13:29:27 satisfy_dependencies_for: Cannot satisfy the following dependencies for packagegroup-machine-base: Apr 09 13:29:27 | * gst-plugins-good-souphttpsrc * Apr 09 13:29:30 I am danny branch Apr 09 13:29:50 there is not package like this in gst-plugins-good Apr 09 13:46:34 rink_: ;-) Apr 09 13:54:24 OWayne_: try again now Apr 09 14:06:34 how do I set the kernel load address in linux-yocto-custom Apr 09 14:15:21 dv_: same issue, error is in populate_packages Apr 09 14:15:39 dv_: i'll try from here Apr 09 14:16:21 dv_: recipes affected are base, good, ugly and bad Apr 09 14:16:42 what distro are you using? Apr 09 14:17:05 angstrom Apr 09 14:17:32 I think I know what is happening. but it should then happen with 0.10 gstreamer as well Apr 09 14:18:33 i did read somewhere some thing to do with changes to indentation in python script between denzil and danny Apr 09 14:19:01 in recipes-multimedia/gstreamer/gstreamer1.0-plugins-package.inc , change the indentation from 4 spaces to tabs Apr 09 14:19:12 I have been trying to change this one to a prefunc, but it does not work properly yet Apr 09 14:19:58 ah, perhaps the equivalent oe-core gstreamer 0.10 include uses tabs in denzil then? Apr 09 14:21:30 http://cgit.openembedded.org/openembedded-core/tree/meta/recipes-multimedia/gstreamer/gst-plugins-package.inc?h=danny indeed, switching the branch to denzil shows tabs instead of spaces Apr 09 14:27:25 dv_: ahh ok, I'll fork branch for now and test against denzil, don't want you wasting your time on older builds when you are trying to move forward :) Apr 09 15:16:52 dv_: glib too old now, maybe I need to look moving to danny. Apr 09 15:17:09 oh, yes. you need glib 2.32 at least Apr 09 15:18:29 is it possible to write something like DEPENDS = "glib>=2.32" ? Apr 09 16:43:18 otavio: can you post a patch to get the busybox bbappend in meta-oe into OE-Core? khem suggests that you get segfaults without it... Apr 09 16:46:50 or JaMa Apr 09 16:47:00 I'd do it but I'm not close enough to it to understand what it's doing Apr 09 16:49:49 bluelightning: I think it is JaMa Apr 09 16:50:00 JaMa: am I right? Apr 09 16:56:42 is it possible to get git repo information to include in a variable in a configuration file? Apr 09 16:57:16 waynr: what kind of information are you looking for? Apr 09 16:57:22 for example, i am storing a distro layer in a git repository and would like to tie the resulting filesystem image to the specific commit under which it was built Apr 09 16:57:52 waynr: ah ok.. I think we do have that, one sec Apr 09 16:58:24 so DISTRO_VERSION = "somethingsomething + ${GITCOMMIT}" Apr 09 16:58:54 even better if it could grab the current git tag and fall back to a commit hash if there is no current tag Apr 09 16:59:12 waynr: ah right; we do not have support out of the box for that kind of thing Apr 09 16:59:59 you would need to create a python function to fetch the hash/tag (possibly using our fetcher) and return it, and then use ${@function_name()} to get the value into DISTRO_VERSION Apr 09 17:00:21 okay Apr 09 17:42:57 re Apr 09 19:23:50 hello Apr 09 19:41:53 hello Apr 09 19:45:53 hi effem Apr 09 22:20:48 see it all the time if it was improperly assembled Apr 09 22:21:00 correct, you would **** ENDING LOGGING AT Wed Apr 10 02:59:58 2013