**** BEGIN LOGGING AT Thu Aug 21 02:59:59 2014 Aug 21 06:43:57 any idea how to fix this issue http://pastebin.com/5DRyXVQP with meta-qt5 when building wayland ? Aug 21 06:45:21 more precise qtwayland Aug 21 07:38:30 hello there! How are you? Aug 21 07:39:42 i have a question for you! Do someone know how to package local(/usr/lib/gconv/ISO8859-1.so) with yocto on my target embeded image? Aug 21 07:41:22 add the locale packages to your IMAGE_INSTALL Aug 21 07:50:34 hi LetoThe2nd, IMAGE_INSTALL is a variable or a file? it's in my eglibc recipe? Aug 21 07:51:51 erm, no. Aug 21 07:52:12 it describes what goes into your image, so it usually is set in your images recipe file Aug 21 07:52:24 but you can add things in local.conf, for example Aug 21 07:52:46 oh yes! guess i see what you mean! i try it right now! thanks! :) Aug 21 07:52:56 have a look into your deploy dir to find the name of the local package you want, and then add that to IMAGE_INSTALL Aug 21 07:59:51 in my environement, i guess what you tell means that IMAGE_INSTALL is my packagegroup-core-dev.bb and have to add "locale" in THIRDPARTY_APPS. It appear that "libc6" hold ISO8859-1.so, and libc6 is already packaged Aug 21 08:28:33 in the workdir, eglibc dosn't contains packages-split folder. Aug 21 08:29:11 morning all, every time I would like to change a patch, I first bitbake -c clean myRecipe and then bitbake myRecipe but my image isnt fully rebuild. Can someone explain me what happen? Look like the exec isnt really deleted Aug 21 08:30:29 TuTizz, do you have some sstate file? Aug 21 08:31:05 in those file, guess yocto didn't rebuild the package because it's marked as already build. can you do a -c clean sstate Aug 21 08:32:28 -c clean -c sstate Aug 21 08:32:33 yes I have a sstate-cache folder Aug 21 08:32:41 bitbake -c clean -c sstate myRecipe? Aug 21 08:32:59 you have to clean those sstate cache file Aug 21 08:33:16 (i'm not sure about the command line, but it's near to be those one) Aug 21 08:33:19 yes Aug 21 08:33:50 bitbake -c clean -c cleansstate myRecipe Aug 21 08:35:26 thanks you very much, I was bored to delete my entire folder every time I changed a single patch :) Aug 21 08:37:10 :p Aug 21 08:40:00 LetoThe2nd, in my eglibc deploy dir, ISO88951 is not in {D}/usr/lib/gconv and it's before the do_package() Aug 21 08:41:07 freaky_: thats probably because it gets stuffed into a seperate package Aug 21 08:42:00 yup, i guess libc6 i guess...but i see nothing to enable this locale.../me lost :) Aug 21 08:45:57 morning all Aug 21 08:48:11 i have fr_FR ISO-8859-1 in the SUPPORTED file :S Aug 21 08:48:16 morning bluelightning Aug 21 08:55:17 Hey Guys. I've build a image for an X86 machine. Is it possible to switch the keyboard layout to german? Aug 21 09:09:19 MasterCrumble, do you grep for de_DE string? Aug 21 09:29:53 freaky_ not now. In the image? Aug 21 09:41:57 RP: big thanks for "process: Further improve robustness against server shutdown" Aug 21 09:43:39 JaMa: I'm still not convinced that is the one you hit :/ Aug 21 09:47:31 MasterCrumble, \o/ i guess you jave to use xmodmap Aug 21 09:49:12 MasterCrumble, you may use somethings like xmodmap_1.0.8.bb? Aug 21 09:57:59 I'll try. Thank you Aug 21 10:24:32 RP: but lately it was hanging also from normal cli usage (not only when aborted jenkins) so this looks promissing Aug 21 10:25:59 RP: any idea about that ugly exception when you break the output stream (eg bitbake foo | some-typo)? Aug 21 10:27:53 http://pastebin.com/fMmeFkJR something like this Aug 21 11:21:10 JaMa: if you look at knotty, that can only happen if cacheprogress is None Aug 21 11:21:36 JaMa: in turn that can only happen if there is a bb.event.CacheLoadProgress out of sequence Aug 21 11:23:56 JaMa: ah, the first write will fail, that causes the system to abort Aug 21 11:51:53 RP: I haven't looked at the code yet, it's just bad from UX pov, because the error from bash (about incorrect piped command) is first and most people will probably overlook it and start searching the cause of that python exception Aug 21 11:52:26 JaMa: agreed, I think I have a fix Aug 21 11:53:50 JaMa: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t222&id=a01dbf47f0e9f4fc3a91f02f2bcc5e75318cf334 Aug 21 11:54:00 JaMa: that ioerror handler is causing the problems Aug 21 11:54:32 * JaMa cherry-picking all 3 fixes to test Aug 21 11:55:21 JaMa: obviously I need to fix the commit message Aug 21 11:55:48 * RP is also chasing another hang with process pools which may well be our other mystery hang Aug 21 11:57:50 the patches on bitbake-devel are still from poky repo, right? (not complaining very loudly, because I like your fixes..) Aug 21 11:59:25 this is with all 3 included and it seems to hang http://pastebin.com/39PaFsF1 Aug 21 12:00:04 or at least parsing on background it seems (from htop) Aug 21 12:42:21 JaMa: It should have triggered an exit :/ Aug 21 12:45:14 JaMa: it works when the cache is hot but not when cold Aug 21 13:08:20 JaMa: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t222&id=2b49e1335d735f7cdf09afad1cb5cdc6c385dfae Aug 21 13:18:07 good morning Aug 21 13:18:22 good morning Aug 21 13:25:01 RP: cool, exited a lot quicker now Aug 21 13:28:24 JaMa: only 3 -world recipes failing? Pretty amazing Aug 21 13:28:41 (long blacklist ?) Aug 21 13:28:53 yes, even the qa.log isn't so bad now Aug 21 13:29:09 long list of fixes and medium-sized blacklist Aug 21 13:29:11 about the TEXTREL, I start to understand it better Aug 21 13:29:41 33 PNBLACKLISTs in meta-oe Aug 21 13:30:04 I see Aug 21 13:31:10 about textrel, I could verify that libgcrypt has been fixed (pending) Aug 21 13:31:36 I'll check libpostproc as next Aug 21 13:32:00 libav is a moving thing so there are hopes Aug 21 13:32:22 world contains few fixes for libav which weren't merged yet Aug 21 13:32:54 so you can see slightly different results Aug 21 13:53:00 someone working on fixing perf build with 3.16 kernel (e.g. linux-yocto-dev)? Aug 21 13:55:20 JaMa: perf built fine for me with 3.16. It was a condition of submission. Do you have a pointer to the error ? I can have a look. Aug 21 13:58:57 does pthread_lock_mutex and pthread_unlock_mutex have issues in yocto? in my app it never unlocks... which is annoying to say the least (but i just got rid of it ... only one small assignment is critical ... and if I get a junk value I just ignore it and it updates very frequently) Aug 21 14:06:13 zeddii: http://dpaste.com/3MMDN5E Aug 21 14:06:50 zeddii: also the log header looks like it's trying to autodetect some deps Aug 21 14:08:10 JaMa: thanks. Leave it with me, I'll have a look. I built and tested this for 3.16 already, so I can track down what has changed. Aug 21 14:09:29 zeddii: cool, thanks Aug 21 14:31:58 zeddii: are you testing the 3.16 headers as well? Aug 21 14:46:54 ant_work. I have a header update here, yep. Aug 21 14:47:11 still catching up from my 2 weeks off :) Aug 21 14:51:57 heh, good luck. There are some sadistic devs sending tons of patches in plain summer ;) Aug 21 14:59:11 hello, I am looking for some tips to configure pulseaudio. I think information in "the perfect setup" quickguide (from pulseaudio) are not valid for embedded system like OE dylan. After many search, I cannot find good way to launch pulseaudio. I found I need to export two var to have Dbus. I am not sure I have its daemon launched correctly when I do $pulseaudio --start (process "/usr/sbin/console-kit-daemon --no-daemon" is also laun Aug 21 14:59:47 Finaly I have the error : Failed to acquire org.PulseAudio1: org.freedesktop.DBus.Error.AccessDenied: Connection ":1.19" is not allowed to own the service "org.PulseAudio1" due to security policies in the configuration file. The file /etc/dbus-1/system.d/pulseaudio-system.conf has . I tried to modify without success. Aug 21 14:59:55 Does anyone an idea helping me in pulseaudio ? how to understand what is missing now ? looking in the source code ? Aug 21 15:11:20 why are you trying to start it yourself? does the package not already start it? Aug 21 15:11:34 it sounds pointless to be manually starting services when we ship service files Aug 21 15:13:36 kergoth : so something is mising. I configure my recipe adding pulseaudio-server that's all and there is no automatic launching after that. But it is probably an option somewhere... Aug 21 15:15:35 hmm, interesting. are you using systemd or sysvinit? Aug 21 15:21:57 kergoth : sysvinit if my understanding is correct. I have init task with pid 1 and used to launch application adding script in /etc/init.d/rc* Aug 21 15:30:14 What does pulseaudio-misc package or how to know it ? Aug 21 15:52:17 Q:? what has someone on my team done, so that every time we run bitbake, it re-parses every single recipe Aug 21 16:17:09 hi guys, anybody had any success with opengl app in qemu with the yocto rootfs? Aug 21 16:27:59 nobody? bummer Aug 21 16:28:37 make that opengl es in fact Aug 21 17:17:59 JaMa: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t222&id=66d5ff62d317e452579598a1f59e46e14c04e3ff and some of the preceeding patches are for the other bitbake hang I could reproduce Aug 21 17:18:14 * RP hates debugging multiprocess hangs Aug 21 17:20:06 RP: thanks, added to my test branch, I'll abort few jenkins jobs with it today :) Aug 21 17:23:41 JaMa: I've just realised I need to convert the rest of the multiprocess usage in OE-Core :/ Aug 21 17:32:23 Q:? what has someone on my team done, so that every time we run bitbake, it re-parses every single recipe aka "Parsing of 1563 .bb files complete (0 cached, 1563 parsed)" Aug 21 17:38:17 WarheadsSE: there is some global variable that is changing every time? Use of DATE/TIME somewhere problematic for example? Aug 21 17:38:38 JaMa: I've pushed an updated set of patches, sorry Aug 21 17:38:53 JaMa: still retesting but these ones seem more complete Aug 21 17:42:38 RP: all in the same branch? Aug 21 17:43:36 JaMa: yes Aug 21 17:44:12 RP: nasty feeling someone put a UUID global.. Aug 21 17:46:16 WarheadsSE: "bitbake -e > X; bitbake -e > Y; diff X Y" Aug 21 17:51:34 Looks like I might have actually found a culprit Aug 21 17:51:35 Hey everybody Aug 21 17:53:40 Can anyone help me with a screen blanking issue? Aug 21 18:03:16 My LCD is turning off after 15 minutes of activity and I'd like it to stay on Aug 21 18:03:33 setterm -blank 0 has no effect, and cat /sys/module/kernel/parameters/consoleblank always returns 900 Aug 21 18:41:34 RP: looking through all the diffs now. Thanks for the pointer. Aug 21 18:44:52 I am mainly seeing IMAGE_NAME and IMAGE_UUID, but also some items to appear to have a DATETIME in them as part of the env Aug 21 18:46:16 RP: refreshed from that branch + kept "process: Deal with infinite looping of the server" from ML Aug 21 18:52:24 I think this might be the insanity plea -- http://pastie.org/9492282 Aug 21 19:57:10 RP: and it worked fine on my jenkins-abort-hang case, thank you again Aug 21 20:27:57 JaMa: thanks for testing, that is useful data :) Aug 21 21:48:03 hi Aug 21 21:48:21 we install cronie on our device and add a file to /etc/cron.d as well as /etc/cron.hourly Aug 21 21:48:34 the cron.d file is executed. The cron.hourly is not. Aug 21 21:49:02 Bug or feature that the runparts command that run /etc/cron.{daily,hourly,monthly} is not installed? **** ENDING LOGGING AT Fri Aug 22 02:59:58 2014