**** BEGIN LOGGING AT Mon Sep 07 02:59:58 2015 Sep 07 07:19:54 good morning Sep 07 07:20:08 morning mckoan Sep 07 08:33:59 morning all Sep 07 08:52:01 hi bluelightning, all Sep 07 08:57:14 hi mckoan Sep 07 09:26:55 FYI it does look like I'll be at Dublin for ELCE / OEDEM Sep 07 09:30:40 bluelightning: great! looking forward to meet you there Sep 07 09:31:02 indeed, it will be great to see you (and everyone else) there Sep 07 10:33:50 sadly I won't be making it this year, the price has made it too prohibitive Sep 07 10:48:02 Hopefully I'll be there as well. Sep 07 11:09:18 Hello people, I made a system image for imx6 processor. I've also Qt5 included from meta-qt5 layer. I am using now master branch, because i wanted to compile qt 5.5. Everything seems working on embedded devices, examples etc. But when i am builiding toolchan through -c populate_sdk task, in my toolchain there is not qmake. Do you know how to solve it? Sep 07 12:17:43 anybody familiar with qmake here specifically Qt Sep 07 12:17:54 need to fix one minor thing but cant get hold of it Sep 07 12:18:05 upstream might need its fix well Sep 07 12:23:49 spaszPL: you can use the native qmake in this case Sep 07 12:24:04 mckoan, bitbake it? Sep 07 12:24:50 mckoan, you mean to bake this recipe nativesdk-qtbase_git.bb ? Sep 07 12:28:21 spaszPL: nevermind, I did a mistake Sep 07 12:46:06 mckoan, Ok, as far as i remember it was branch problem. I used master meta-qt5 branch because of qt 5.5. but i had to change it to dizzy or fido, and then i have qmake in toolchain. Sep 07 13:01:58 spaszPL: unfortunately I haven't used qt-5.5 yet Sep 07 13:10:01 what is all this about SDK vs Ext SDK on the mailing lists Sep 07 13:10:17 I've never heard anyone mention the term Extensible SDK until a few weeks ago Sep 07 13:10:45 is it a Poky thing thats being moved over? or a replacement for our current sdk which has package management built in or something? Sep 07 13:15:39 the latter :) Sep 07 13:15:41 ish Sep 07 13:17:14 jackmitchell: it's something that I've been working on with help from a few others Sep 07 13:17:31 jackmitchell: not quite ready to replace the traditional SDK so it's available side-by-side with it Sep 07 13:18:32 basically the idea is you still get direct access to a toolchain, but additionally you can use devtool Sep 07 13:18:52 so you have a stepping-stone to adding things into the proper build Sep 07 13:19:49 ok, I'll wait for the manual write up, thanks! Sep 07 13:22:16 does anybody know where pkgConfigExecutable is defined Sep 07 13:22:30 it is giving me wrong info Sep 07 13:40:34 ok got it where it is defined Sep 07 13:42:38 strange, I have this in environment PKG_CONFIG_SYSROOT_DIR="/home/noor/SB/release_2015.12/build_i.mx6/tmp/sysroots/mx6q" Sep 07 13:43:25 but qt_functions.prf is getting its value "/home/noor/SB/release_2015.12/build_i.mx6/tmp/sysroots/mx6q/usr" Sep 07 13:43:28 usr is added Sep 07 13:43:43 Can anybody help me how this is added Sep 07 13:48:10 JaMa: otavio: ^ Sep 07 13:48:41 while building qtwebkit -isystem path is not being set for libxml2 Sep 07 13:48:51 extra usr is being added Sep 07 13:48:59 I am unable to see it in environment Sep 07 13:49:15 we have to add it eplicitly in recipe but I think this should be fixed Sep 07 13:49:23 need little help on this Sep 07 14:06:31 Noor: see ML it was discussed there and fix is already in master-next Sep 07 15:25:10 JaMa: thanks :) Sep 07 15:34:03 JaMa: are you talking about this thread qt5: add host paths to qt.conf Sep 07 15:37:23 "qtbase: Fix prefix substitution from OE to QT variable" Sep 07 15:41:05 JaMa: thanks .... Now I get it what was going on Sep 07 15:41:19 one less issue to deal with :) Sep 07 15:43:53 JaMa: I think qtwebkit need libxml2 dependency. I'll conform it tomorrow and send a patch Sep 07 15:44:33 I am not sure but I read soemwhere that libxml2 need icu enabled Sep 07 15:44:45 need to confirm these 2 things tomorrow Sep 07 15:46:27 JaMa: http://webhamster.ru/mytetrashare/index/mtb4/1422011844ww5q8upnm1 Sep 07 15:57:34 recipes-qt/qt5/qtbase_git.bb:# be aware that you need to add icu to build qtwebkit, default Sep 07 15:57:37 recipes-qt/qt5/qtbase_git.bb:PACKAGECONFIG[icu] = "-icu,-no-icu,icu" Sep 07 15:58:08 I'm already enabling icu in world builds Sep 07 15:59:17 JaMa: yeah we have enabled it in our builds Sep 07 16:00:56 but I thik qtwebkit need libxml by default Sep 07 16:01:01 its hard dependency Sep 07 16:01:13 so we have to add it in DEPENDS Sep 07 19:27:40 Hey, folks. I was not recompiling my project for a while and now when I do this from scratch I got a weird isuue about arm-angstrom-linux-gnueabi-ar: https://paste.kde.org/pumd6qcce#line-642 Sep 07 19:27:53 Any ideas what can be wrong ? Sep 07 19:43:16 it seems no right option for ar is passed Sep 08 00:17:19 ERROR: Task 545 (virtual:native:/oe/setup-scripts/sources/openembedded-core/meta/recipes-support/gmp/gmp_5.1.0.bb, do_compile) failed with exit code '1' Sep 08 00:17:36 seems to be trying to build native stuff for x86_64, but my system is x86_*32* **** BEGIN LOGGING AT Tue Sep 08 01:09:44 2015 Sep 08 02:10:15 Luke-Jr: why :) it should work but you're building lots of code, a bigger 64 bit machine is useful. even 8 GB of RAM would help and isn't that expensive these days. Sep 08 02:10:36 vmeson: I have 16 GB of RAM, but my OS is 32-bit Sep 08 02:10:57 vmeson: it seems bitbake is incorrectly using uname -m to determine build architecture Sep 08 02:13:18 Luke-Jr: I thought that uname -m and arch would give the same result. If not , can you send a patch? Sep 08 02:14:04 uname -m and arch agree on my 32bit rhel-6 box. Sep 08 02:15:56 arch is also incorrect Sep 08 02:16:10 uname -m and arch give the kernel architecture, not the system Sep 08 02:16:10 Luke-Jr: see: https://en.wikipedia.org/wiki/Physical_Address_Extension <-- you're running a 'retro' distro. :) Sep 08 02:16:17 no, I'm running a normal 64-bit kernel Sep 08 02:17:14 64-bit kernel + 32-bit system Sep 08 02:17:56 Luke-Jr: I see. Again I wonder why but patches to make this work might be accepted. Sep 08 02:18:23 it's a fairly standard configuration these days :p (lower memory use than x86_64) Sep 08 02:19:44 yeah, I'm spoilt with my 16x2 core, 64 GB build server.. :) Sep 08 02:19:49 :P Sep 08 02:19:56 gcc -dumpmachine => i686-pc-linux-gnu Sep 08 02:20:21 I *think* this is the best way to determine it, but it's not the same format as uname -m Sep 08 02:21:25 what's the best repo to build bitbake merge reqs on? Sep 08 02:21:31 my old rhel632 gcc says: i686-redhat-linux Sep 08 02:22:03 Luke-Jr: git://git.openembedded.org/bitbake Sep 08 02:22:10 master. Sep 08 02:22:17 and best way to submit a merge req? Sep 08 02:23:10 hmm... http://www.openembedded.org/wiki/How_to_submit_a_patch_to_OpenEmbedded but that doesn't mention bitbake directly. Sep 08 02:25:24 Luke-Jr: subject: [bitbake-devel] [PATCH 0/6] Fix handling of "default" project Sep 08 02:25:34 to bitbake-devel@lists.openembedded.org Sep 08 02:25:39 * vmeson goes to sleep... Sep 08 02:26:08 thanks Sep 08 02:26:34 thanks for doing the right thing and hopefully sending a patch, Luke-Jr. Sep 08 02:27:10 hmm, quite a few calls to uname -m ; is there a place I could sanely put a helper script for this? Sep 08 02:28:09 is sources/openembedded-core/scripts/ likely to be in $PATH ? Sep 08 02:32:47 Luke-Jr: 'openembedded-core/scripts/' will be on PATH Sep 08 02:34:02 hmm. just realised I've left bitbake in this grep Sep 08 02:34:52 * Luke-Jr ponders where build arch is actually being set Sep 08 02:35:45 sources/openembedded-core/meta/conf/bitbake.conf:BUILD_ARCH = "${@os.uname()[4]}" Sep 08 02:35:49 seem probable? Sep 08 02:36:21 yep thats the one **** ENDING LOGGING AT Tue Sep 08 02:59:58 2015