**** BEGIN LOGGING AT Mon Nov 25 02:59:58 2013 Nov 25 06:46:12 Been a while since I checked, but I suspect OSX is not a supported hostsystem? Nov 25 06:46:16 And, good morning :) Nov 25 06:50:10 JaMa: ping. Nov 25 06:51:09 JaMa: our gui-guy want to try Qt 5.2, and I see you have a working branch for it. is it functional? Nov 25 07:57:00 good morning Nov 25 09:46:42 morning all Nov 25 09:48:27 hi bluelightning, all Nov 25 09:48:34 hi mckoan Nov 25 10:05:37 I see wildcard bbappend support has been merged in bitbake master Nov 25 10:06:03 http://cgit.openembedded.org/bitbake/commit/?id=31bc9af9cd56e7b318924869970e850993fafc5f Nov 25 10:59:20 hi all Nov 25 10:59:52 i want to know whether we can build any module whichis having "xyz.mk" file instead of makefile using oe-core Nov 25 10:59:54 ? Nov 25 11:02:16 otherwise any hope in Oe-core to convert .mk to Makefile?? any use of autotools? Nov 25 11:06:20 Sj__: yes, you can. there is a 'default' do_compile() function, but you can have your own in your recipe. Nov 25 11:06:50 you can have a hook in your recipe that converts .mk to Makefile too, if you prefer Nov 25 11:07:00 ohh then manually there isno need for me to conver mk files to Makefiles right? Nov 25 11:07:07 convert* Nov 25 11:07:54 Actually i don;t need Makefile from .mk but my doubt is whether oe-core can parse .mk or not!! Nov 25 11:08:41 Sj__: just have your own do_compile() function in your recipe. Nov 25 11:09:08 no it's huge project! with many mk files! Nov 25 11:09:15 the 'base' or 'autotools' classes are convenient most of the time. but sometimes, you have no choice than making your own functions... Nov 25 11:09:31 hmm. then make your own class ;-) Nov 25 11:09:39 you are asking me to write what ever ther in mk in my do_compile() function right? Nov 25 11:09:46 btw, it's not Android, right ;-) Nov 25 11:10:04 hmm class file to convert all .mk to Makefile? Nov 25 11:10:12 isit android module only :-( Nov 25 11:10:24 it is* Nov 25 11:10:28 Sj__: no, you just need a do_compile() that calls your Makefile. you don't need to reimplement the Makefile Nov 25 11:11:14 but i have .mk file right !! do_compile() any way i can you make. ! Nov 25 11:11:38 i don't understand what you just said. Nov 25 11:11:43 is there any specific reason why you have asked this "btw, it's not Android, right ;-)" Nov 25 11:11:53 that was a joke.. Nov 25 11:12:11 e.g. building Android with OE... Nov 25 11:12:17 i have a module with lot of .mk files!! i need to build them using Oe-core bitbakes this ismy requirement Nov 25 11:12:31 hmm one of the android module ! Nov 25 11:12:35 modules* Nov 25 11:12:50 how do you build today, without OE? Nov 25 11:13:47 fyi, the 'default' do_compile which is used is here http://cgit.openembedded.org/openembedded-core/tree/meta/classes/base.bbclass#n327 Nov 25 11:13:53 some different build structure i'm not aware of that fully!! today i have been asked to check the compatability in OE for .mk files Nov 25 11:14:00 so do_compile simply ends up calling 'make' command Nov 25 11:14:30 Sj__: answer is simple: OE doesn't care about your makefile name. so you can do it. Nov 25 11:14:36 make needs MAKEFILE right!! but we have .mk! sory i have little knowledge on these things! Nov 25 11:14:45 no... Nov 25 11:14:48 make can parse .mk too? Nov 25 11:14:57 make -f foo.bar would work. Nov 25 11:15:23 make uses Makefile 'by default', unless if another script is provided. Nov 25 11:15:51 ohh ok then make -f xyz.mk woudl work right Nov 25 11:15:56 yep. Nov 25 11:16:08 sorry, i didn't realize that your confusion was about make. Nov 25 11:16:09 thanq :-) then my work gonna be easy Nov 25 11:16:25 well, btw, if make wouldn't work with xy.mk, such files wouldn't exist in the first place ;-) Nov 25 11:16:57 uh! Nov 25 11:17:20 i didn;t get you Nov 25 11:49:46 morning all Nov 25 12:06:57 hey pb_ Nov 25 12:14:56 hi florian Nov 25 12:29:23 tasslehoff: haven't tried that in runtime, but builds fine except qtwebkit-examples Nov 25 14:11:10 When compiling with the proposed qt5-sdk patches in meta-qt5 I get: Failed to process makespec for platform 'linux-oe-g++' Nov 25 14:11:38 the mkspec seems to be the same as the one from the qtbase recipe. any ideas? Nov 25 14:14:32 tasslehoff: AFAIK with Yocto now you need arm-poky-linux-gnueabi-g++ makespec Nov 25 14:18:51 mckoan: will try that then. thanks Nov 25 14:24:20 mckoan: that doesn't sound right to me Nov 25 14:24:44 at least, with Qt4 we create a linux-oe-g++ mkspec on the fly to use Nov 25 14:24:47 not sure about Qt5 Nov 25 14:25:16 hmm, my imgae lacks the certs needed for https? Nov 25 14:28:58 is imgae like gmae for the internet? Nov 25 14:29:05 sounds eleet. Nov 25 14:30:25 bluelightning: my mistake, I tested it with Yocto, not OE Nov 25 14:31:08 heh Nov 25 14:31:18 stop making fun of my typing Nov 25 14:31:48 Crofton|work: install ca-certificates Nov 25 14:31:56 I did, no help Nov 25 14:31:59 something weird Nov 25 14:32:09 export GIT_SSL_NO_VERIFY=1 Nov 25 14:32:10 :) Nov 25 14:36:05 Crofton|work: in case of github, check your git version Nov 25 14:36:16 hmm Nov 25 14:36:20 https://help.github.com/articles/https-cloning-errors Nov 25 14:36:25 I used whatever OE built Nov 25 14:36:33 thanks Nov 25 14:36:57 you can also s/https:/git:/ in the uri to make it work Nov 25 15:41:13 Hi all. I have one local package - not present on the network. How to configure SRC_URI? Nov 25 16:16:06 drasko: SRC_URI = "file:///path/to/your/tar" Nov 25 16:34:32 khem, very strange that this is not noted in the documentation: https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#var-SRC_URI Nov 25 16:45:07 What would bitbake say? Nov 25 19:07:33 bluelightning: the on-the-fly makespec seems to work for qtbase, but not for the native sdk (which is a work in progress) Nov 25 19:12:37 tasslehoff: hmm, ok... AFAIK we use the same method for Qt4 native/nativesdk Nov 25 19:14:37 bluelightning: odd. more verbose output here if you (or others) want to take a look: http://pastebin.com/RapgME9H Nov 25 19:16:28 ah, there is a difference from the qtbase-recipe Nov 25 19:17:49 tasslehoff: are you testing https://github.com/meta-qt5/meta-qt5/pull/33 or your own version? Nov 25 19:18:18 the one you linked to Nov 25 19:19:06 cherry-picked into dora, to make sure I get some issues ;) Nov 25 19:20:17 did the colleague already try 5.2.0-beta1? Nov 25 19:21:22 JaMa: haven't gotten around to that yet. started out with the sdk. will probably test 5.2.0-beta1 tomorrow. worked from home today. Nov 25 19:24:57 good Nov 25 19:25:02 thanks Nov 25 19:25:41 JaMa: have you tried compiling with the sdk-patches? Nov 25 19:43:26 tasslehoff: not yet Nov 25 20:02:12 JaMa: ok. too tired to investigate the compile error now. will resume tomorrow. Nov 26 01:43:47 hi Nov 26 01:44:49 WARNING: QA Issue: ELF binary Nov 26 01:44:56 how to remove this error Nov 26 01:45:27 WARNING: QA Issue: No GNU_HASH in the elf binary: Nov 26 01:49:07 boost12: pass LDFLAGS from OE to build system of that recipe Nov 26 01:49:38 boost12: it means that LDFLAGS in OE have --hash-style=gnu (or both), but your binary doesn't have it Nov 26 01:50:11 so how to pass it ...i am begginer for this Nov 26 01:50:28 lately gcc is built with option to pass --hash-style=gnu by default, so every binary built with OE has GNU_HASH by default Nov 26 01:50:55 boost12: depends on the recipe, most recipes with this issue are calling gcc directly Nov 26 01:51:10 or have very stupid Makefile and don't read LDFLAGS from environment Nov 26 01:51:37 ok Nov 26 01:52:25 btw, do I know you? :) (it's possible that I've assigned that task to you) **** ENDING LOGGING AT Tue Nov 26 02:59:58 2013