**** BEGIN LOGGING AT Mon Jan 06 02:59:59 2014 Jan 06 05:51:17 hi all i have been using RTL8723AS wifi module,i compiled for arm at91sam9x5ek,it is working fine,but in the backround it throws me lot of logs,it disturb me while configuring the wifi,please see http://pastebin.com/BpdqcU3A is it possible to stop those logs? can you plese help me Jan 06 07:42:25 Hi yocto. Anyone succeeded with sstate-cache with nfs? I am following this guide but `https://wiki.yoctoproject.org/wiki/Enable_sstate_cache`. But still i cant. Any sugestions? Jan 06 08:39:58 morning all Jan 06 10:17:02 JaMa: your qt 5.2.0 branch seems to work. ok, still have a problem with input device detection, but that rather seems a qt or raspberry problem Jan 06 10:17:47 JaMa: ...it also seems a bit strange that you put the compiled examples into /usr/share/qt5 Jan 06 10:18:24 uh. and this AUTOINC thing is not working for me, resulting in strange version numers for your packages Jan 06 10:19:00 but guess for the AUTOINC i am just too stup^W inexperienced with yocto Jan 06 10:21:43 what you mean by AUTOINC? Jan 06 10:21:52 +gitrAUTOINC? Jan 06 10:21:56 yup Jan 06 10:22:12 these are replaced by real numbers from PR service Jan 06 10:22:24 it's to make git SHA-1s sortable Jan 06 10:22:57 e.g. +gitr1_abc12345 < +gitr2_989abc Jan 06 10:23:05 oh... checking the generated files it's really just bitbake that shows AUTOINC, instead of the sha Jan 06 10:23:14 the package files are all fine. so sorry on that. Jan 06 10:23:35 JaMa: which option was that for git when submitting patches the email only shows the differences between the revisions, not the whole change again? Jan 06 10:23:38 Was it -M? Jan 06 10:24:19 one more thing: the packages all are generated for armv6-vfp, although they contain machine specific platform bits Jan 06 10:24:39 hasselmm: yes, it's because it's replaced later in do_package task Jan 06 10:25:31 hasselmm: they don't contain any machine specific bits by default, if your .bbappend makes it machine specific, then you're responsible to set MACHINE_ARCH in qtbase and all recipes depending on it Jan 06 10:25:55 hasselmm: there is even small trick to set it automatically (merged in master few days ago) Jan 06 10:26:18 automatically for all, but you still have to say that you want to set it Jan 06 10:26:33 lpapp: if you mean to detect renames then yes -M is right one Jan 06 10:26:55 JaMa: oh. good point. Jan 06 10:27:24 might be worth to write down all those small things needed for building a machine specific qt with your layer Jan 06 10:32:48 hasselmm: it would be even nicer to find some way to keep it TUNE_PKGARCH :) Jan 06 10:33:09 I'll add note in https://github.com/meta-qt5/meta-qt5/wiki Jan 06 10:33:49 JaMa: do you also have qtquickcontrols? Jan 06 10:34:02 y Jan 06 10:40:52 hasselmm: https://github.com/meta-qt5/meta-qt5/wiki/Machine-arch Jan 06 10:41:53 JaMa: oh nice. thank you. Jan 06 10:42:38 JaMa: what is the current state of the Qt 5 SDK? Jan 06 10:48:25 lpapp: see wiki Jan 06 10:50:48 ok, will read later, thanks. Jan 06 10:51:28 JaMa: heh, my name is on the wiki. :) Jan 06 10:52:42 good to see the progress. Jan 06 11:46:57 JaMa: seems there is no package with the qt example's source code? Jan 06 11:49:31 hasselmm: it's possible, I don't use examples, maybe the FILES patttern doesn't match anymore? Jan 06 11:50:03 JaMa: they are useful for testing the image :-) Jan 06 11:51:35 does anyone know how bitbake knows to process python modules in openembedded-core/meta/lib/oe ? Jan 06 11:52:18 I'm trying to add our own signature handler and don't know if I need some special setup for bitbake to find it in our layer Jan 06 11:55:29 JaMa: /lib is added to the python path automatically for all layers Jan 06 11:55:58 I can't recall where the code to do that is, but I did find it while dealing with the runtime testing code Jan 06 11:57:43 bluelightning: ok, thanks, so now I need to find out why it didn't work for me :) Jan 06 11:58:46 note of course that if you put it in under lib/, the module would of course be referred to as . not just Jan 06 13:37:43 its in base.bbclass Jan 06 13:44:53 RP: Ah thanks sys.path is pythonpath I was thinking about PATH as os.path when I've noticed this line in git grep :) Jan 06 13:47:58 RP: can you please reconsider this patch http://lists.openembedded.org/pipermail/openembedded-core/2013-November/086571.html ? And this one didn't have any negative response, maybe it was just missed in icecc changes http://lists.openembedded.org/pipermail/openembedded-core/2013-November/087044.html Jan 06 13:49:31 JaMa: the icecc patch is fine and I can queue, its just got lost Jan 06 13:49:54 JaMa: for the first one, we did make some improvements in other areas which I hoped had removed the need for that? Jan 06 13:51:30 RP: this is the weird corner case with different toolchain for different MACHINEs and the need for this wasn't resolved AFAIK Jan 06 13:54:00 JaMa: is it just the set empty TARGET_PREFIX part you need or all of it? Jan 06 13:55:00 all, because without INHIBIT_DEFAULT_DEPS they also depend on toolchain signatures Jan 06 13:55:41 JaMa: this is where I get confused since INHIBIT_DEFAULT_DEPS is set whenever TARGET_PREFIX is cleared Jan 06 13:56:30 Ah I'm confusing you again, this reply is more accurate than what I've just said :) http://lists.openembedded.org/pipermail/openembedded-core/2013-November/086574.html Jan 06 13:58:24 I'll retest and send v2 to clear this up Jan 06 13:58:49 JaMa: Either all of the class should be applied unconditionally or none of it should, mixing the two up sounds wrong Jan 06 13:59:07 JaMa: I need to stare at the code to better explain that but something doesn't feel right :/ Jan 06 14:06:50 RP: I'm retesting with TARGET_PREFIX only at least to better explain in commit message why the other part is needed Jan 06 14:09:07 JaMa: ok, thanks Jan 06 14:57:11 is there some known reason for linux-yocto to rebuild in every image build in dora branch? I thought that those issues were all only in master and already fixed Jan 06 14:57:34 the signatures are identical (no DATETIME dependency or something like that) Jan 06 15:05:42 is there a way to blacklist a bbappend file? I want to replace the qt5 bbappend file in meta-fsl-arm with my own. Jan 06 15:05:48 JaMa: the initramfs task nostamp? Jan 06 15:26:16 gjohnson: that sounds like a bad idea, but its possible BBMASK may work for that. Jan 06 15:30:34 RP: yes, do_bundle_initramfs[nostamp] = "1" is in dora's kernel.bbclass Jan 06 15:30:57 JaMa: I suspect that is the issue Jan 06 15:31:22 I'm looking @ 55989cb509340bd265d0ce0d8bfe849681be4616 in master which removed that Jan 06 15:31:48 but I remember that there was longer discussion about this problem and don't remember if it had some other required changes or not in the end Jan 06 15:33:44 it can be cherry-picked to dora after cherry-picking 81831db1c32afa3346f3ed9f4325ad280e5bb005 Jan 06 15:36:31 anyone know if there is a nice way to get around install -s in a makefile? is the answer patch the makefile? Jan 06 15:36:44 or is there an option that can be passed in a recipe? Jan 06 15:37:06 kergoth: Thanks, I will look into that. The problem I have is the qt5 append file in the freescale layer doesn't work well with the way I want to configure qt. Jan 06 15:37:17 at the moment the wrong strip is being used Jan 06 15:46:13 ajtag: if it's in extra variable then you can use EXTRA_OEMAKE, otherwise some recipes are patching makefile or calling sed to deal with extra -s Jan 06 15:47:07 ok, thanks, there are no extra variables at the moment, so I will build a patch. thanks Jan 06 15:59:10 Hi all, I know this is not yocto issue but just to get some suggestions, Any idea how to solve this "connection reset by peer" error while fetching opencv from github---> http://pastebin.com/hcWBRdFi Jan 06 16:21:44 hi there, there is new minicom, version 2.7 Jan 06 16:22:46 is it going to appear magicaly in master, or the answer is 'if you bump recipe and test it will...' ? :) Jan 06 16:23:08 hmm... there are two of me in here Jan 06 16:23:15 i don't think there is anything magic ;-) Jan 06 16:23:25 Xz: abc Jan 06 16:23:29 somebody has to submit the patch for it ;-) Jan 06 16:23:32 Xz_: Cristian Iorga is the maintainer for that recipe, I think it will be on his list to update Jan 06 16:23:39 Xz_: unless you get there first of course... Jan 06 16:23:46 you see, there is a bit of magic :) Jan 06 16:24:24 FYI, recipe maintainers for the core recipes are listed in meta-yocto/conf/distro/include/maintainers.inc Jan 06 16:24:59 but of course doing the upgrade means you need to test it properly Jan 06 16:26:09 bluelightning: does it mean when I add a new recipe I'm automacially becoming that recipe maintainer? Jan 06 16:26:40 Xz_: well, not necessarily Jan 06 16:27:00 but it is the default option ;) Jan 06 16:27:10 bluelightning: got it Jan 06 16:36:37 Hi there Jan 06 16:37:00 hey otavio Jan 06 16:37:19 :) Jan 06 16:37:22 Happy 2014 Jan 06 16:37:24 :D Jan 06 16:37:30 otavio: and to you :) Jan 06 18:04:24 RP: sgw_: I am back to work today Jan 06 18:04:37 I am looking at build failures in fsl-arm Jan 06 18:07:12 otavio: Happy New Year and thanks Jan 06 18:07:18 :) Jan 06 18:07:43 I have some already queued; I am building it in my builder and will push Jan 06 18:10:04 sgw_, every blown commit, you buy/bring a keg for the next conf :) Jan 06 18:10:19 and HNY Jan 06 18:11:03 Crofton|work: HNY to you, a beer yes, keg only if I can drive it there! Jan 06 18:11:22 :) Jan 06 18:11:40 sounds like funny dice Jan 06 18:11:50 doh Jan 06 18:17:43 someone say beer? Jan 06 18:18:02 great idea Jan 06 18:18:06 Crofton|work: keg seems a bit much :) Jan 06 18:26:48 Crofton|work: beer? Jan 06 18:27:09 oh, hey, wb to the office otavio Jan 06 19:10:29 RP: will you have by any chance some time to take a look at the bitbake patches? **** ENDING LOGGING AT Tue Jan 07 02:59:58 2014