**** BEGIN LOGGING AT Fri Jun 06 03:00:00 2014 Jun 06 06:14:27 morning Jun 06 06:15:01 hi there Jun 06 07:48:20 Hello. I'm during some build on each commit workflow and the case is that the old versions just pile up. What can I do to avoid that. Basically I just want the latest version.? Jun 06 08:42:42 otavio: for some reason, no longer works :S Jun 06 08:42:50 *pulls hair out* Jun 06 08:44:59 good morning Jun 06 11:48:44 rink_: heheh Jun 06 11:49:12 rink_: please paste your run.do_install script Jun 06 11:59:32 SorenHolm: lookup rm_old_work Jun 06 12:01:52 RP: Thanks for the hint. You mean RM_OLD_IMAGE right? Jun 06 12:02:46 RP: no .. you're right .... someone actually did write an rm_old_work class Jun 06 12:02:47 btw, short q - I need en_US.UTF-8 locale to be available - how do I tell Yocto this ? Jun 06 12:06:45 rink_: I've tried to do that also but no matter what I do I always end up with a iso8859 console. Even ssh into the target does not allow me to type local (in my case danish) characters. Jun 06 12:07:19 well, my code does an explicit setlocale to enforce en_US.UTF-8 Jun 06 12:07:26 but that fails because the locale is not available Jun 06 12:07:37 and that is correct since /usr/lib/locale doesn't even exist Jun 06 12:07:40 I wonder if I forgot something Jun 06 12:10:16 should I add libc-locales to DISTRO_FEATURES ? Jun 06 12:47:35 so.. I could use rm_old_work class instead of cleanup-workdir script ? Jun 06 12:49:15 rink_: Well I tried loading on all the locale and eglibc-gconv or whatever the package is called. But no luck. No UTF8 ;( Jun 06 12:52:11 hmm in my case I guess it worked Jun 06 12:53:06 in the distro's config, I added DISTRO_FEATURES_append = " libc-locales " Jun 06 12:53:12 and Jun 06 12:53:12 GLIBC_GENERATE_LOCALES = "en_GB.UTF-8 en_US.UTF-8" Jun 06 12:53:16 that works for me Jun 06 12:55:04 RP, would you accept a patch adding rm_old_work.bbclass into oe-core ? I see it's duplicated a few layers .. Jun 06 13:02:43 rink_: What is it that does not work for you then? Jun 06 13:26:58 without those lines, I did not have any locale at all Jun 06 13:29:50 afk, have a nice weekend Jun 06 14:24:00 how often is the Complete Documentation Set regenerated from the rest? I got a surprise yesterday when I got different text when looking at the bitbake as opposed to the complete set. Jun 06 14:57:39 also, I'm not sure why, but I'm not seeing messages I send to any of the lists anymore. I've gotten a few replies (on one of them only), but it's strange to not see the originals. Jun 06 15:02:42 blloyd: check the "receive your own posts" option on the list Jun 06 15:10:49 thanks rburton: I checked and they are set to yes. Jun 06 15:16:51 wow, do images now track state in master? I actually get a X of X didn't need to be rerun when building an image. Jun 06 15:20:55 yes, and 1.6 Jun 06 15:26:06 so I'll notice that in daisy too? Goodie. Jun 06 15:27:01 the fun bit is when it says "i've built that" and you realise that you've about 20 images in deploy Jun 06 15:28:25 is there a problem in identifying the one desired or something I haven't noticed yet? Jun 06 15:28:57 I just noticed the $MACHINE subdirectories under deploy too. Jun 06 15:29:16 when you've 20 core-image-sato-[timestamp] spanning two days, finding the one it says has already been built is non-trivial Jun 06 15:29:17 deploy/images/$MACHINE, that is Jun 06 15:30:07 RM_OLD_IMAGE = "1" is nice for keeping nbr of images in deploy down .. Jun 06 15:31:05 oh, so old ones are watched too. Wow. So sstate of each dated image is maintained. I can see where that can be fun if you do things like add items in local.conf for one build then build again without it later. Jun 06 15:51:53 used to, I could do a MACHINE=seldombuilt bitbake myimage Jun 06 15:52:24 now, doing that throws a sanity check failure. Should a build directory only do one machine type now? Jun 06 15:54:23 nope Jun 06 15:54:26 what's the error? Jun 06 15:55:47 it says MACHINE needs to be specified on command line or in local.conf. Come to find out, the problem really appears to be that I fat fingered the name. So it was set, but to an invalid machine. Jun 06 16:16:40 I have generated an image with yocto, but I cannot get the kernel to boot using a tftpboot and and NFS. I've tested my tftp and NFS and both work. Does someone have a clue? when I burn the .sdcard it does work. When the uImage is on the tftboot it does transfer but it got stuck at starting kernel without any further messages Jun 06 17:20:53 both daisy and master do this. Why would you want to keep all filesystems mounted during the bootup filesystem search? Let's say there are an sda, sdb, and sdc. All three get mounted and then filesystem is found on sdc. The image on sdc knows that sda should have been mounted as /home but since it's already it's busy.... Jun 06 17:21:45 mtetreault: how do you boot your kernel? Are you sure you want to boot a uImage? do you load it at the right address? Jun 06 17:56:13 zecke: I'm booting using bootm. I have tripple check the address and it's the right one. I'm not sure what you mean by if I,m sure if I want to load a uImage, what other option do I have as a kernel file? Jun 06 17:57:15 zecke: if it can help, I'm working with a freescale sabre lite(i.mx6) Jun 06 19:19:21 yocti, seen pidge? Jun 06 19:19:22 halstead: Error: 'pidge?' is not a valid nick. Jun 06 19:19:26 yocti, seen pidge Jun 06 19:19:26 halstead: pidge was last seen in #yocto 2 days, 1 hour, 48 minutes, and 52 seconds ago: Letothe2nd: for now, that's probably best until I or someone else gets a chance to sit down and hack this out. We're trying to maintain at least 2 years worth of backwards compatibility on the AB, so that's why you see all the layerversion code (image location changes/machine name changes, etc) Jun 06 19:20:05 halstead: nice! Jun 06 19:20:10 yocto, seen jama Jun 06 19:20:30 yocti, seen JaMa Jun 06 19:20:31 sgw_: JaMa was last seen in #yocto 3 days, 21 hours, 8 minutes, and 3 seconds ago: fwiw it's running on Ubuntu 12.04.1 LTS Jun 06 19:21:12 sgw_, pidge and I were going to co-work today but I don't know if she still has time with all the craziness. Jun 06 19:25:14 halstead: ah, she was around earlier Jun 06 19:27:54 sgw_: pong Jun 06 19:28:34 JaMa: was not thinking you were going to be around on a Friday evening, if you have something better to do it can wait! Jun 06 19:29:28 Anybody knows how to have 2 same recipe in one image that will generate 2 different binary? example psplash_git.bb + psplash-dektop_git.bb same recipe but different binary. Jun 06 19:29:48 I tried couple of things but always ended up with warning or error... Jun 06 19:30:17 sgw_: hehe, I'm thinking the same for last 4 hours but still working.. Jun 06 19:31:11 jama, sorry then I will ask chose not to answer! I am trying to build qt5 and getting a failure, in qtbase (sent email to you a few days ago about this0 Jun 06 19:31:24 kapare: what kind of error/warnings? Jun 06 19:31:51 kapare: 2 recipes with 2 different binaries should be OK Jun 06 19:32:15 JaMa: adding you to bug #6371 also Jun 06 19:32:16 sgw_, The recipe psplash is trying to install files into a shared area when those files already exist.... Jun 06 19:32:17 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=6371 critical, Medium, 1.6.1, saul.wold, NEEDINFO , Cross-compile with python native on 64bits host machine Jun 06 19:32:59 sgw_, trying to see if there is other solution then changing the Makefile to have another binary name or any other way? Jun 06 19:33:14 sgw_: Sorry, I've replied all I know atm Jun 06 19:34:02 sgw_: I was building qt-5.3.0 many times since then and also haven't seen it.. I didn't do any ppc build though Jun 06 19:34:21 sgw_, this was the warning the error is Jun 06 19:34:23 JaMa: OK, I will try non-ppc build then Jun 06 19:34:25 sgw_, Recipe psplash-desktop is trying to create package psplash-default which was already written by recipe psplash. This will cause corruption, please resolve this and only provide the package from one recipe or the other or only build one of the recipes. Jun 06 19:35:10 sgw_: or just re-enable fpermissive if you just want to reproduce his issue Jun 06 19:36:03 EXTRA_QMAKEVARS_PRE_append = " QMAKE_CXXFLAGS+=-fpermissive" Jun 06 19:37:17 kapare: have you modified the original psplash recipe? the default psplash does not currently provide a psplash-default package Jun 06 19:37:34 JaMa: I assume in local.conf Jun 06 19:39:09 then with _pn-qtbase as well Jun 06 19:41:46 time for new kernel.. bbl Jun 06 19:44:20 Hi everyone Jun 06 19:44:58 Is anyone able to help me with an i.MX6 X11 problem? Jun 06 19:45:47 marlon_: not an iMZ6 person, but I can try Jun 06 19:46:07 Thanks, it might not be i.MX6 specific Jun 06 19:46:34 I'm building the fsl-image-gui (or a slightly modified version) and running a Qt app in X Jun 06 19:47:08 But if I use the vivante drivers in xorg.conf I get a black screen, if I set it to fbdev instead I get screen corruption Jun 06 19:47:22 sgw_, for now I tried something like that. change recipe name and remove stuff I don't need, http://pastebin.com/86EkHLdf Jun 06 19:50:47 marlon_: I know it's not related to your question, but did you get Hob working? I'm also working with an i.mx6 board, and I cannot get hob working properly without breaking my config Jun 06 19:53:01 mtetreault: Sorry, I haven't used Hob. I've been authoring my bitbake files manually, although I'm very much a beginner Jun 06 19:56:31 kapare: I guess I am still not following, the changes you made remove the update-alternative code, this will still generate a psplash binary, I think the error message above is confusing, both your recipes are creating a psplash binary Jun 06 19:56:55 marlon_: Did you try the freescale repo? I think it's pre-configure to use vivante drivers an mesa Jun 06 19:59:16 sgw_, sorry, so both recipe are the same except one have different patches so it generate the same binary. My question was is there way to change the name of my binary in my recipe ? Or best way simply change Makefile to generate binary with different name? Jun 06 20:02:16 mtetreault: I'm using Freescale's dora repo. It looks like it should work; when it boots it starts a demo GUI, but it's glitchy Jun 06 20:02:55 If I was using the wrong kernel version would that be a problem? I assume that the vivante binaries in the filesystem do not change, so perhaps certain kernels would work better than others? Jun 06 20:05:20 kapare, now that I better understand, you want to mv your binary in the do_install_append(), and keep the update-alternative code and FILES_${PN} bin, but use the renamed binary in the ALTERNATIVE_LINK_NAME, but you also need ALTNATIVE_TARGET[psplash] = ${bindir}/psplash-default Jun 06 20:05:27 I think that should do it Jun 06 20:30:20 sgw_, hmm not sure to understand it but will try to play with that thx! Jun 06 20:33:40 kapare: look at the update-alternatives.bbclass and associated documentation, this sounds like what you are trying todo. 2 packages providing the "same" binary, but it's done via links. Jun 06 22:38:19 is there an easy way to have an image not have auto mounting of disks? **** ENDING LOGGING AT Sat Jun 07 02:59:59 2014