**** BEGIN LOGGING AT Fri May 03 02:59:58 2013 May 03 04:48:18 okay, i built several rpi images with poky/master and they work nicely May 03 04:49:08 but there's almost nothing in /var/lib/opkg and "opkg list-installed" produces no output May 03 04:51:59 i have PACKAGE_CLASSES = "package_ipk" on local.conf and "package-management" in the included IMAGE_FEATURES May 03 04:52:21 s/on/in/ May 03 04:58:29 not sure what i'm missing... May 03 06:54:26 hi, I noticed patches.openembedded.org has debug enabled for Djano and dumps db passwords.. maybe you are aware of this already? May 03 07:00:18 is this why www.openembedded.org only shows a mysql error message? May 03 07:02:40 and i can update and install from my local build feed but opkg still thinks nothing is installed in my image except the few packages i just installed... May 03 07:11:24 ouuuu ... this is not good! i see passwords on patches.openembedded.org May 03 07:11:41 silviof: I know, I have emailed the appropriate folks May 03 07:12:20 just in case: ka6sox^ Crofton^ Crofton|work^ halstead^ May 03 07:12:54 bluelightning: good - I have searched but nothing found which person i should write a letter :-) May 03 07:12:57 thx for the names May 03 07:13:15 florian as well but he is not online May 03 07:13:46 nerdboy: which image did you build? May 03 07:15:07 my own image that includes a modified rpi-basic-image May 03 07:15:44 i have PACKAGE_CLASSES = "package_ipk" on local.conf and "package-management" in the included IMAGE_FEATURES May 03 07:15:49 nerdboy: ok... if you bitbake -e | grep ROOTFS_POSTPROCESS_COMMAND, does that have remove_packaging_data_files in it? May 03 07:16:10 or rather bitbake -e your-image | grep ^ROOTFS_POSTPROCESS_COMMAND= May 03 07:18:24 yup May 03 07:18:29 sneaky... May 03 07:18:56 is that part of core-image or something in features? May 03 07:23:26 nerdboy: I'm not sure how that would be getting in yet but what does this report exactly: bitbake -e your-image | grep ^IMAGE_FEATURES= May 03 07:26:28 it's getting in through a chain of image includes May 03 07:27:04 nerdboy: I see... does it include core-image-minimal somewhere perhaps? May 03 07:27:19 yup May 03 07:27:27 right, it should not do that May 03 07:27:51 if you want to create your own image, just do that, you don't need to include/require other image recipes May 03 07:28:04 image recipes are usually pretty simple May 03 07:29:32 brb May 03 07:29:37 bluelightning, what happened? May 03 07:30:11 ka6sox: no idea but if you go to http://www.openembedded.org/ or http://patches.openembedded.org/ you'll see May 03 07:30:43 mysql server is broken and django debug *must* be turned off for patches.openembedded.org May 03 07:30:59 oh and change the mysql db password after doing that since it's been exposed May 03 07:32:41 gotta run, back in an hour or so May 03 07:33:04 morning all May 03 07:33:15 morning all May 03 07:33:24 okay let me fix the mysql server May 03 07:33:25 brb May 03 07:36:01 good morning May 03 07:36:45 hi silviof, silvio May 03 07:36:53 hi apelete May 03 07:37:01 hi silviof May 03 07:37:23 hi apelete May 03 07:45:24 good morning May 03 07:56:14 morning mckoan May 03 07:57:35 hi silviof and mckoan and denis May 03 08:02:39 hi mckoan, hi woglinde May 03 08:04:05 hi apelete May 03 08:05:04 okay, i redid/rebuilt my image without the include May 03 08:06:01 will test it tomorrow... May 03 09:17:31 bluelightning, ping? May 03 09:17:53 hi ka6sox May 03 09:18:04 hiya...can you tell me if patches is all bettah? May 03 09:19:24 ka6sox: 504 gateway timeout May 03 09:23:35 ya, the patchwork backend isn't starting May 03 09:23:41 okay...lessee May 03 09:48:25 morning all May 03 09:49:15 hi pb_ May 03 09:49:29 hi bluelightning May 03 10:06:27 hi pb May 03 10:12:49 hi woglinde May 03 10:15:39 How can I disable the checksum check for a single recipy? May 03 10:16:50 BramSmulders: which checksum check do you mean? May 03 10:17:15 the MD5 and SHA256 checksums. May 03 10:17:40 I have a source package and it causes problems because it gives a different MD5 checksum each time I download it... May 03 10:21:59 BramSmulders: not sure... you're not trying to download a snapshot from a version control system by any chance are you? May 03 10:22:46 http://cgit.freedesktop.org/vaapi/xvba-driver/snapshot/xvba-driver-0.8.0.tar.gz May 03 10:24:51 it's the only link I have for this library... May 03 10:26:14 BramSmulders: I'd recommend using a git URL and pointing to the tag for the 0.8.0 release May 03 10:27:03 i.e. SRC_URI = "git://anongit.freedesktop.org/vaapi/xvba-driver" May 03 10:27:52 and SRCREV = "28e7d9ce0adc456e6d05884576a9b5eef7a484fa" (I think that's the correct hash from the repository) May 03 10:28:46 Ok, thanks. I will try that May 03 11:19:57 JaMa: I'm looking at the jansa/test-20130429 branch of meta-qt5, and the Makefile sed stuff done in do_install() in qt5.inc would also be needed when a third party recipe inherits qmake5 and build e.g. a qt-plugin May 03 11:25:22 erbo: yes that's true May 03 11:25:37 erbo: I'm still debugging some issues with it May 03 11:26:13 erbo: I've added check that nothing exists in $D$STAGING_DIR_TARGET after make install and some modules !sometimes still install something there May 03 11:28:06 kudos to you for doing a great job btw May 03 11:29:30 thx May 03 12:05:52 JaMa: ok if I add meta-qt5 to the layer index? May 03 12:27:27 bluelightning: I would like to fix few more issues and add more recipes before that May 03 12:28:00 JaMa: surely you want people to be able to find it at the moment if they'd like to test/contribute to it? it doesn't have to be perfect May 03 12:28:40 I'd like to try to avoid people doing the same work as others because they just don't know that something already exists May 03 12:30:38 well before changes I've merged yesterday it was very broken, now it should work, but because I have a lot more recipes to add (already existing) recipes I would like to publish it with them May 03 12:30:51 because otherwise people can send pull-requests with modules I already have May 03 12:31:02 it shouldn't take long to finish it.. I hope this weekend May 03 12:32:55 JaMa: ok, I'll hold off for now then May 03 13:32:33 bluelightning, hi about layer index I've 2 not-yet ready layers May 03 13:32:41 I saw the thread on the ml May 03 13:34:02 I guess I'll wait for them to be ready before adding them to the layer-index May 03 13:48:55 well, I dunno, it seems to me that waiting for a layer to be "ready" is just inviting other people to duplicate the same work unintentionally May 03 13:49:47 I would have thought it would be feasible to add some sort of marker to the layer index to say whether the layer is "ready" or not, or create a README inside the layer's own git repository with that information. May 03 13:51:20 pb better in the index May 03 13:51:28 otherwise you have to checkout it May 03 13:51:37 or looking at the webpage May 03 13:52:12 yeah. looking at the readme in cgit doesn't seem like such a big hurdle to jump, but putting it in the index sounds like a fine plan May 03 13:53:19 If a layer really isn't ready, you could do something to fake a bitbake warning stating the layer isn't production ready yet May 03 13:56:39 jackmitchell yes that could help to May 03 13:56:54 but for quick glimps it should be in the layer index too May 03 14:07:36 hmm, perf didn't build for me May 03 14:07:38 | util/evsel.c:18:44: fatal error: ../../include/linux/perf_event.h: No such file or directory May 03 14:09:05 ieehks May 03 14:09:14 ../../ includes are bad May 03 14:09:24 curse the programmer May 03 14:10:33 * woglinde curses meta-ti for not beeing updated May 03 14:12:32 I shoudl update .... May 03 14:17:00 denisATeukrea: I'd agree with the other comments above, better to add it to the index if it's already reasonably started May 03 14:18:51 we have a "layer note" feature that allows you to add a note at the top for "this layer isn't ready for production yet" or "this layer is deprecated" type things that will catch the reader's attention May 03 14:22:50 ok May 03 14:23:59 oh May 03 15:42:37 and perf really does fail to build May 03 15:42:48 what failure? May 03 15:43:01 | util/evsel.c:18:44: fatal error: ../../include/linux/perf_event.h: No such file or directory May 03 15:43:19 I saw it earlier, so i updated everything just in case May 03 15:43:31 this is for a bsp I am updating to oe-core May 03 15:48:15 huh. old kernel? May 03 15:52:05 hmm May 03 15:52:06 sort of May 03 15:52:31 3.5 era May 03 15:52:39 is that old now? May 03 15:53:15 hmm, not sure May 03 15:54:24 the file exists in the kenel source May 03 16:13:29 morning kergoth May 03 16:13:50 hi kergoth May 03 16:14:20 hey guys May 03 16:37:56 quit May 03 17:11:35 Here I have May 03 17:11:35 $ bitbake -e | grep -E '^(PARALLEL_MAKE|BB_NUMBER_OF_THREADS)' May 03 17:11:36 PARALLEL_MAKEINST="-j 8" May 03 17:11:39 PARALLEL_MAKE="-j 8" May 03 17:11:42 BB_NUMBER_OF_THREADS="8" May 03 17:12:05 but I'm always getting one task running. May 03 17:12:47 Where should I start debugging? May 04 00:23:37 loving my new gmail "OE/Yocto Core Patches" saved search, not perfect, but nice: label:patches {label:projects-oe label:projects-yocto} -subject:{hob eclipse} {bbclass /bitbake/ bitbake.conf} -/recipes **** ENDING LOGGING AT Sat May 04 02:59:58 2013