**** BEGIN LOGGING AT Tue Nov 05 02:59:58 2013 Nov 05 06:33:42 hmm... Nov 05 08:38:05 hi, i'm trying to get libsdl2 (2.0.1) into my image. Version 2.0.0 works (official package). I then changed the name from libsdl2_2.0.0.bb to libsdl2_2.0.1.bb and updated the hashes inside. But libsdl has compile errors (EGL stuff missing) and i checked do_configure log and it shows "checking for EGL support... no". Anyone got an idea how to get a yes there? Nov 05 08:42:50 rayo: there's a virtual/egl package that you can depend on.. Should be resolved to the proper EGL package for your machine Nov 05 08:54:40 i already have virtual/egl in the DEPENDS list, still the same error. I tried eglinfo-x11 too and it fails already during configure: Checking for library EGL : yes \n Checking for header EGL/egl.h : not found Nov 05 08:55:03 so the egl library is there but not the header Nov 05 09:28:04 egl stuff should come from either mesa or some other machine-specific graphics lib... sounds like the virtual isn't resolving correctly on that machine... Nov 05 09:37:36 i have a freescale i.MX board and i'm using meta-fsl-arm layer. opengl is working on the device (glmark2-es2 runs fast). Somehow just the egl.h is missing during build (it's in the sysroot: tmp/sysroots/imx6qsabresd/usr/include/EGL/egl.h and the package gpu-viv-bin-mx6q has it as include file too. Nov 05 09:39:02 i guess you'll have to look at the sdk configure and find out why its not looking in the right place Nov 05 09:39:05 sdl even, not sdk Nov 05 09:40:43 where should it look? it the sysroot path? how does libsdl know it should not look in the host includes? Nov 05 09:42:32 that's the trick and why you need to look. Nov 05 09:48:13 how does it work with other libraries with yocto? is it done through the bb files? Nov 05 09:48:38 depends. if a configure uses pkgconfig then it just works. Nov 05 09:48:49 if it directly pokes and guesses then it will go wrong. Nov 05 09:51:36 morning all Nov 05 10:09:23 rburton: it looks like it tries to compile a simple main with just the EGL/egl.h include. And this also gives an error: tmp/sysroots/imx6qsabresd/usr/include/EGL/eglvivante.h:244:2: error: #error "Platform not recognized" Nov 05 10:09:33 looks like some defines are missing for my platform Nov 05 10:14:42 hi all i have used yocto for at91sam9x5ek arch,i have added a extra package openobex in yocto,but while compiling the package it shows the following issues http://pastebin.com/H65uTk7v can you tell me what is that issues Nov 05 12:16:23 is there a problem in the yocto manual? ubuntu 13.10 is not listed at http://www.yoctoproject.org/docs/1.5/ref-manual/ref-manual.html#detailed-supported-distros Nov 05 12:16:31 but it is listed here https://wiki.yoctoproject.org/wiki/Yocto_Project_v1.5_Status Nov 05 13:13:18 Hi everyone Nov 05 13:13:32 I'm having the same issue described here : http://comments.gmane.org/gmane.linux.embedded.yocto.meta-freescale/5321 Nov 05 13:13:40 please can someone give me a clue? Nov 05 13:22:59 Stygia: Hi, I was successful in adding .bbappend to my recipe . Thank You. Nov 05 13:25:02 zaif_, Fantastic! Sorry I wasn't online Saturday, I was a bit pre-coffe e when I said I'd be working, since obviously I wouldn't be on a Saturday. Nov 05 13:25:23 mrAlmond, A regular or raging clue? Nov 05 13:25:58 mrAlmond, What this means is somewhere in the recipe, there is an an include path (something like -L/usr/bin) that points to somewhere on your host, instead of on your target. Nov 05 13:26:23 mrAlmond, Which would suck if you're trying to compile something that works - Imagine building a program for your ARM target with some x86 libs. Nov 05 13:26:26 Stygia : so this is a recipe bug? Nov 05 13:26:32 mrAlmond, Is it a default recipe? Nov 05 13:26:46 yes...it comes from the official repository Nov 05 13:27:31 Stygia: No problem Sir. Nov 05 13:28:21 mrAlmond, Then yes. Nov 05 13:28:52 Stygia : ok, it should be something introduced recently Nov 05 13:29:16 Zaif, That's herr Oberstgruppenführer to you. Nov 05 13:29:22 mrAlmond, Presumably. Nov 05 13:30:47 Stygia : thank you also for this http://www.urbandictionary.com/define.php?term=raging%20clue :-D Nov 05 13:30:59 mrAlmond, Hah, yup. Nov 05 13:31:08 it's a pleasure to learn such things :-D Nov 05 13:31:21 Yup. Nov 05 13:31:31 Zaif, Icncase you missed it, don't actually call me that, it's a nazi joke. Nov 05 13:32:50 Stygia: Aha, ok. Generally in my college we call our seniors as Sir. It was that. No worries. Nov 05 13:33:12 rburton: hi ! how are you ? any news for weston/wayland 1.3 ? Nov 05 13:33:13 mrAlmond, You could probably fix it if you run do_configure, find the path, and either alter it with a configure argument, or sed it in. Nov 05 13:33:27 Zaif, Yes, I know, I have had some Pakistani/Indian friends. Nov 05 13:33:43 Zaif, I'm only 22, though, so very likely that I'm not your elder at all. Nov 05 13:34:19 Stygia : thank you for the (raging) clue Nov 05 13:34:46 mrAlmond, Now I feel violated. Nov 05 13:35:04 omg sorry Nov 05 13:39:05 mrAlmond, Again, kidding. Nov 05 13:39:14 Stygia me too Nov 05 13:39:28 mrAlmond, What's the difference between a spare tire and a pile of dead babies? Nov 05 13:39:42 let me know Nov 05 13:40:35 I don't have a spare tire in my trunk. Nov 05 13:42:03 mrAlmond, If you're trying to fix this I can give you an example of a recipe I have where I fixed an issue like that one. Nov 05 13:42:25 great...better than the dead babies stuff Nov 05 13:43:12 mrAlmond, I happen to like nasty jokes. Nov 05 13:43:23 My favourite joke involves blasphemy, buttsexs and nuns. Nov 05 13:43:45 so you should read Johnny Ryan comics Nov 05 13:45:32 mrAlmond, http://pastebin.com/VLtApV5q Nov 05 13:45:40 A few things to note, which I do that are abnormal. Nov 05 13:45:56 There is a default recipe for rrdtool, but it disables the perl bindings, which I need. Things failed once I enabled them. Nov 05 13:46:06 in do_configure_append, I remove an RPATH=... thing. Nov 05 13:46:21 Because this got set during do_compile, was almost impossible to alter, and was useless. Nov 05 13:47:10 It set the RPATH (runtime library path) to somewhere on HOST, which failed somewhere in QA checks. Nov 05 13:47:20 So, I remove it entirely, let it use the other provided params. Nov 05 13:47:42 Stygia : let me have a look at your paste Nov 05 13:47:44 Then, I recompile a certain file which had some HOST paths compiled in (I found no way to affect this as the config was created during do_compile -> a mess. Nov 05 13:48:00 The rest (everything in do_compile_append) has to do with this. Nov 05 13:48:12 The part you're interested in, I think, is do_configure_append Nov 05 14:00:12 Stygia : is there a way I can execute manually do_package_qa for that recipe and obtain a verbose output of what's going on? Nov 05 14:00:58 I'm not an expert of yocto... Nov 05 14:03:08 mrAlmond, Me neither. Nov 05 14:03:13 mrAlmond, But, not that I know. Nov 05 14:03:23 What's it saying, exactly? Nov 05 14:05:54 http://pastebin.com/VCh6HP7d Nov 05 14:07:30 You're compiling python? Nov 05 14:07:34 Default recipe? Nov 05 14:08:37 I have the same vernsion as you, and I don't have this problem. Nov 05 14:08:46 Try cleanall on it? Nov 05 14:12:42 mrAlmond, Nov 05 14:13:07 Stygia, yes, python from devtools default recipe Nov 05 14:13:31 Stygia : I think it could be an host problem Nov 05 14:13:48 I try to cleanall it Nov 05 14:13:59 mrAlmond, Hmm, possible. Nov 05 14:14:14 Either way, it's weird. Try to cleanall it, then do_configure, then devshell to check it out? Nov 05 14:17:47 Stygia : please can you tell me how to exec do_configure only...and what do you mean for devshell? Nov 05 14:19:24 mrAlmond, You can execute individual steps with bitbake -c task Nov 05 14:19:45 mrAlmond, Such as, bitbake -c cleanall python; bitbake -c configure python; Nov 05 14:20:03 Devshell is a special task, where you open up a shell in the working dir for the recipe, with it's env variables and such already set. Nov 05 14:20:04 you'd do bitbake -c devshell python Nov 05 14:20:23 If you're working over SSH, you'd need to do ssh -X to do X forwarding to is can open a shell on your box. Nov 05 14:21:43 if I make a custom layer above distro (Arago in this case) layer, do I have to define DISTRO_FEATURES and DISTRO_FEATURES_LIBC again, or is it possible to inherit them somehow? Or more generally, is there something like bbappend for .conf files? Nov 05 14:21:54 Stygia : tnx Nov 05 14:22:45 Stygia : after the cleanall the package compiled without problems... Nov 05 14:22:55 this is weird Nov 05 14:23:01 mrAlmond, You said earlier you'd cleanall'd it. :) Nov 05 14:23:16 mrAlmond, Somehow it was affected after it's do_unpack, so it was prevented from working due to some local change. Nov 05 14:23:26 That happens, it pays to do cleanall before you fight it too much. Nov 05 14:24:11 Stygia : the solution is always the easiest Nov 05 14:24:15 mrAlmond, True. Nov 05 14:24:18 mrAlmond, However. Nov 05 14:24:24 Did you run all of the task? Or just do_configure? Nov 05 14:24:35 If you did -c cleanall python; bitbake -c configure python and it "works" it might not. Nov 05 14:24:37 all of them....just run bitbake python Nov 05 14:24:39 You have only done do_configure. Nov 05 14:24:42 Right, good. Nov 05 14:55:29 elbc: back from conference/leave, will be sending shortly Nov 05 15:10:32 hi all. I have build linux kerenel, and I have one message: Summary: There was 1 WARNING message shown. Nov 05 15:10:54 How to see this warning? I.e. - where to look for the log? Nov 05 15:13:32 drasko, Is that all you see? Nov 05 15:13:38 drasko: a linux-yocto kernel? Nov 05 15:13:46 yes Nov 05 15:13:50 ant_work, no Nov 05 15:14:19 Stygia, yes Nov 05 15:14:39 I did not compile with -v option Nov 05 15:14:55 So I do not see now where log is stored Nov 05 15:15:25 try find tmp/ -name *.log | grep -i RECIPE NAME Nov 05 15:15:31 maybe *\do_compile* Nov 05 15:15:39 sorry, *do_complie* Nov 05 15:15:47 ... that except spelled correctly. Nov 05 15:17:43 rburton: ok thanks ! Nov 05 15:55:22 The Yocto Project Tech Meeting Con-Call will be starting at the top of the hour Nov 05 15:55:22 Dial-in number: 1.972.995.7777 / Participant passcode: 42001078 Nov 05 15:55:22 This call is open to all and the channel remains open to discuss any topic Nov 05 15:58:32 i have grub, and i'm booting directly into a yocto-kernel from it. the kernel does not like me passing in rootflags. Nov 05 15:59:24 YPTM: Saul is on Nov 05 15:59:31 YPTM: Matthew is on Nov 05 15:59:59 YPTM: AlexG on the call Nov 05 15:59:59 YPTM: Cristiana on the call Nov 05 16:00:10 YPTM: Björn is on the call Nov 05 16:00:28 YPTM: Beth is on the call Nov 05 16:01:00 switching lines Nov 05 16:01:02 * nitink is on the call Nov 05 16:01:40 YPTM: Michael present Nov 05 16:01:57 YPTM: Scott Rifenbark is on the call Nov 05 16:01:59 YPTM: Denys is here Nov 05 16:02:00 YPTM: Richard is on the call Nov 05 16:02:16 YPTM: Welcome to the technical team meeting, please let me know who's on the call. Thanks! Nov 05 16:02:21 YPTM: Jeff is here Nov 05 16:02:24 YPTM: Tom Z on Nov 05 16:02:54 YPTM: Cristiana on the call Nov 05 16:03:05 YPTM: Bruce on the call. Nov 05 16:03:19 YPTM: Paul Eggleton, Beth Flanagan, Alex Damian are on Nov 05 16:03:34 YPTM: Darren is on Nov 05 16:03:47 YPTM: Any opens today? Nov 05 16:04:25 YPTM: Randy is on the call. Nov 05 16:04:40 YTPM: ross on the call Nov 05 16:05:08 YPTM: I guess i jhoined too early Nov 05 16:05:18 YPTM: AlexG here too, Nov 05 16:06:18 YPTM: jzhang on the call Nov 05 16:10:06 #vaguebooking Nov 05 16:10:54 anyone else getting lots of static on the call ? or is it my phone ? Nov 05 16:11:05 zeddii, it's pretty clear here Nov 05 16:11:10 zeddii: fine here Nov 05 16:11:13 clear here Nov 05 16:11:44 ok. so my phone. Nov 05 16:11:53 I'm getting about 1 of 3 words. and none clear at all. Nov 05 16:12:16 * zeddii will just nod wisely. Nov 05 16:12:59 zeddii: good call Nov 05 16:15:11 vmeson: I didn't realise your IRC nickname until now :) Nov 05 16:15:41 don't encourage him. ;) Nov 05 16:15:49 ... oh look a neutrino Nov 05 16:16:09 :) it's both physics and computers VMEson and vMeson. Nov 05 16:16:50 vmeson, quantum computing is the dream. We'll never see you again. Nov 05 16:17:14 * vmeson ignores zeddii and looks at his bug backlog... Nov 05 16:18:56 zeddii: neutrinos are hard to observe, mesons are much easier to detect Nov 05 16:20:21 * vmeson has been involved in detecting both... oh look another bug... Nov 05 16:21:12 * RP is pleased to see the autobuilder tweaks happening early in the cycle :) Nov 05 16:21:51 vmeson: I studied Physics at Uni but never got to do real experiments with particles... Nov 05 16:22:42 ouch Nov 05 16:22:51 vmeson, I was almost about to tease you about that. Nov 05 16:25:40 Song_Liu: clang Nov 05 16:26:06 vmeson, I'm also in touch re: the kernel patches from behan and friends. Nov 05 16:32:56 Hi again Nov 05 16:33:10 I'm experiencing another problem Nov 05 16:33:11 http://pastebin.com/LPcYeUp6 Nov 05 16:33:45 this seems to be about an illegar character and a missing description...how can it be? Nov 05 16:34:18 mrAlmond: sounds like fsl didn't test with the rpm package backend :/ Nov 05 16:35:06 RP: rpm is the default in my case...how can I fix it? Nov 05 16:35:16 mrAlmond: Is there a - in PR or PV ? Nov 05 16:35:32 let me check Nov 05 16:37:25 PE = "1" Nov 05 16:37:30 but Nov 05 16:38:08 PN = d.getVar("PN", True) PV = d.getVar("PV", True) Nov 05 16:38:25 they are got using python I guess Nov 05 16:39:10 mrAlmond: bitbake -e gpu-viv-bin-mx6q | grep ^PV= Nov 05 16:40:24 PV="3.10.9-1.0.0-hfp" Nov 05 16:40:36 there are two "-" Nov 05 16:40:49 mrAlmond: What is the recipe called? Nov 05 16:40:56 (the filename) Nov 05 16:41:14 mrAlmond: ah, its in your log: /gpu-viv-bin-mx6q_3.10.9-1.0.0-hfp.bb Nov 05 16:41:30 yes Nov 05 16:41:34 mrAlmond: you could try renaming that without the dashes although I don't know where else its used Nov 05 16:42:31 mrAlmond: or add something like PV := "${@d.getVar('PV', True).replace('-','')}" Nov 05 16:47:41 when I do 'bitbake myimage -c populate_sdk' does populate_sdk command scan IMAGE_INSTALL variable to look for libraries xcompiler toolchain should support? Nov 05 16:49:38 Krz-: yes Nov 05 16:49:46 RP : tnx Nov 05 16:50:23 mrAlmond: it may still cause problems if PV is used in SRC_URI for example and needs the dashes there Nov 05 16:50:45 are the dashes prohibited in recipes name? Nov 05 16:53:06 mrAlmond: in versions they cause problems Nov 05 16:53:12 mrAlmond: they're ok in PN Nov 05 16:53:54 * RP can't make up his mind about http://cgit.openembedded.org/openembedded-core-contrib/commit/?h=sgw/stage&id=8308e22a44a2dea7d1bbfb429b9df9c63714a649 Nov 05 16:54:22 on the one hand it fixes rm_work. On the other it adds things to the build target which may be unnecessary Nov 05 17:04:44 RP : I did a cleanall and then the package run successfully...another mistery... Nov 05 17:33:25 * mranostay yanws Nov 05 17:40:34 moin Nov 05 17:44:42 why is dropbear unhappy all of a sudden? Nov 05 17:57:08 mr_science: No idea but if you find out please let me know too! Nov 05 18:36:41 zeddii, welcome to the OE eV, I need your email for the members list Nov 05 18:47:07 Crofton: does he get a members jacket? Nov 05 18:47:43 yes Nov 05 18:49:39 * zeddii bets Crofton would pick: http://en.wikipedia.org/wiki/Straitjacket Nov 05 18:51:04 zeddii: we badly need a 3.12, -dev or plain yocto ;) Nov 05 18:52:17 ant_home, I pushed the full 3.12-dev last night. Nov 05 18:52:48 oh, I've missed it. great thx Nov 05 18:52:53 since no full release will go out with 3.12 as a version, I hadn't planned on forking out the recipe. but taking a snapshot now would be the same. Nov 05 18:53:02 ant_home, you know where to shout if anything is broken. Nov 05 18:53:19 heh Nov 05 18:58:06 zeddii: not broken but nasty, the yaffs frags spit some warnings Nov 05 18:58:29 hmm. I'll have a look. Nov 05 19:02:55 hi darknighte__ Nov 05 19:04:35 hey! Nov 05 22:10:42 man License is a pain. Why does the build system moan about not having a license field set. And struggling to find documentation to explain exactly what it expects from the recipe Nov 05 22:11:11 I'm guessing they'd prefer that yocto users not be sued Nov 05 22:11:33 complying with software licensing is impormtant Nov 05 22:11:36 important, even Nov 05 22:15:50 not at all. But I'm writing a recipe I'm not even sure about licenses as yet or where ${COMMON_LICENSE_DIR} is Nov 05 22:17:27 I mean I'm not sure of a license for the code I'm playing with. I'm not writing anything serious so it'll never have a user. But even my hello world must have a license. Nov 05 22:17:45 as a temporary workaround, you can set LICENSE = "CLOSED", which basically means closed source / proprietary Nov 05 22:17:50 but you won't want to leave it htat way :) Nov 05 22:18:04 COMMON_LICENSE_DIR isn't a factor unless i'ts a custom license yocto doesn't know about Nov 05 22:18:21 normally you just set LICENSE, set LIC_FILES_CHKSUM to point to an existing license file and/or copyright/license header(s) in the source tree Nov 05 22:21:12 I'm trying to find one as we speak ;-) Struck Gold meta/files/common-licenses/ don't have to search the web have a pile to choose from here. Nov 05 22:23:48 you should set LICENSE to what the actual source license is, or set to CLOSED to disable license bits. don't just pick a supported license at random :) Nov 05 22:25:25 Always been partial to the GPL but you're right as it's a hello world and if CLOSED stops the errors we'll go with that. Thanks for you help Nov 05 22:25:36 no problem Nov 05 22:26:19 everyone has tehir own license preferences :) I tend to prefer MIT or GPLv2, myself, depending on context Nov 05 22:40:36 pev: somehow i just started getting this bug today: http://code.google.com/p/rt-n56u/issues/detail?id=1033 Nov 05 22:40:55 and there are many more instances than just that one Nov 05 23:53:20 Huh. That's odd. Nov 05 23:54:25 The tuning "armv7a-neon" doesn't have the thumb feature, which means that it will never pass either -marm or -mthumb. Which I guess makes sense, sort of, although it just bit me for complicated reasons. Nov 05 23:55:03 But the reason I'm looking at armv7a-neon at all is that tune-cortexa9.inc does DEFAULTTUNE ?= "armv7a-neon". Which is a little surprising to me, because I would have sort of assumed that the default would have been some kind of cortexa9. **** ENDING LOGGING AT Wed Nov 06 02:59:58 2013