**** BEGIN LOGGING AT Fri May 31 02:59:58 2013 May 31 03:34:23 I get "Cannot use the installed version of file (file-5.14) to cross-compile file 5.14. Please install file 5.14 locally first". any ideas? May 31 03:48:49 sgw1: looks like http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=3aaa4f78c609a47324da2db3b5c92a28f4f2ccff May 31 03:49:09 sgw1: inside devshell, file --version returns file-5.14 instead of file.real-5.14 now May 31 03:49:35 sgw1: so http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-devtools/file/file/fix_version_check.patch isn't needed anymore May 31 03:58:19 sgw1: patch submitted May 31 07:14:52 Net147: thanks, I merged it May 31 07:21:51 RP: could you also merge the additional documentation build fixes for findutils? May 31 07:22:56 RP: I wasn't able to build findutils without it on one of my systems May 31 07:23:08 Net147: done May 31 07:23:13 RP: thanks May 31 08:21:44 What could be the reason for "No recipe found for xxx.bbappend" if I can see the .bb file in another layer and that layer is included in BBLAYERS? May 31 08:26:17 Never mind. All of my errors had some kind of version mismatch. Must be tired because I was sure I saw matching bb files but apparently not... May 31 08:40:07 morning all May 31 08:47:28 moin May 31 08:47:53 hi Net147 May 31 08:48:09 hi bluelightning May 31 09:10:14 how come systemd core-image-minimal has libX11, libXau, libxcb, libXdmcp? May 31 09:10:28 compared to sysvinit-based image May 31 09:12:16 and has bash completions May 31 09:15:46 Net147: use your package manager to find out. i did know but have forgotten. May 31 09:15:58 oh, dbus? May 31 09:16:15 dbus probably pulls in the X stuff, so if you don't want anything X then turn off the x11 distro feature May 31 09:19:36 argh, I thought we fixed that..? May 31 09:20:04 Net147: buildhistory depgraphs should show you how those got in May 31 09:21:22 bluelightning: dbus respects the x11 feature, so if you have x11 then you get x11 autolaunch, otherwise not May 31 09:22:09 hmm.. pity that couldn't be runtime detected May 31 09:22:18 but I suppose it's fair May 31 09:22:47 yeah, would be nice if it dlopen'd libx11 May 31 09:31:49 is x11 included by default? May 31 09:32:20 I can try updating and building again. it's an old git clone. May 31 09:36:18 rburton: I can backport the patch, but it really is the only thing that has changed between the 0.6.0 and the revision I specified May 31 09:36:38 rburton: http://cgit.freedesktop.org/xorg/driver/xf86-video-modesetting/log/ May 31 09:37:59 Net147: i do prefer a patch rather than mystery git commit May 31 09:38:09 rburton: okay May 31 10:01:32 Net147: that fix is amazingly small :) May 31 10:19:28 rburton: yeah =) May 31 10:24:47 good morning May 31 10:31:22 what's up with all the fails? "rpmbuild: --buildroot: unknown option" May 31 10:32:55 is build history supposed to record removals when packages removed? May 31 10:37:46 it's doing "git add ." but doesn't specify to also include the files that were deleted when committing May 31 10:45:19 bluelightning: it seems the current git behavior is to not include the deleted files when doing git add . May 31 10:45:33 bluelightning: and the behavior will change in Git 2.0 May 31 10:48:36 Net147: hmm, you're right May 31 10:48:48 and apparently we never noticed... May 31 10:49:42 bluelightning: so the intention is to record the deleted files right? May 31 10:49:51 Net147: absolutely yes May 31 12:02:23 rburton: your mesa/wayland patchsets look good, I'm going to test it and add PACKAGECONFIGs for gallium and llvmpipe May 31 12:02:47 rburton: pity that both patchsets are in poky-contrib not oe-core May 31 12:13:00 JaMa: btw removal of linux-yocto_3.2 from oe-core is breaking meta-initramfs and meta-handheld. I'll have to upgrade all to 3.8 (w/out proper testing :/) May 31 12:19:52 ant_work: FWIW my last test with 3.8 worked on spitz IIRC May 31 12:20:53 JaMa: are you working with Net147 for llvmpipe, or are you both doing the same thing May 31 12:22:24 rburton: I have put it in hold for now May 31 12:23:10 rburton: I haven't communicated with JaMa regarding it May 31 12:23:47 Net147: how far are you with it? May 31 12:24:23 I've .bbappend in our layer to enable it with danny release May 31 12:24:53 so if you're working on proper PACKAGECONFIGs in master I can leave it for you May 31 12:25:20 JaMa: I haven't done any work in PACKAGECONFIGs. I was just adding a bbappend in my own layer May 31 12:25:28 are you Jonathan Liu? May 31 12:25:31 JaMa: yes May 31 12:25:52 llvm patches look good to me BTW May 31 12:26:30 JaMa: the last issues I had. 1. Mesa was using llvm-config from native sysroot bin, 2. swrast was failing to load with a missing llvm ostream symbol May 31 12:27:06 JaMa: maybe the latter can be solved by building LLVM with --enable-shared and --enable-shared-llvm in Mesa May 31 12:27:57 I had the same issue with llvm-config (now testing your patch) May 31 12:28:05 and I haven't tried it in runtime yet May 31 12:28:37 I was just thinking about testing and solving it in master and then just backporting needed changes to .bbappend for danny May 31 12:28:37 JaMa: my patch just fixes installing llvm-config into crossscripts directory May 31 12:29:36 JaMa: I am wondering if part of the solution to solving first issue would be to reorder the PATH so crossscripts comes before native sysroot bin if building for target. if building native, crossscripts would either not be present or would appear after native sysroot bin in PATH. May 31 12:30:13 JaMa: what do we really use the native llvm for? May 31 12:31:34 I'm checking mesa configure to see if there is option to pass required llvm-config May 31 12:31:53 llvm-native was imported with target llvm from meta-java May 31 12:32:08 not sure how much we need from it May 31 12:32:59 JaMa: LLVM 3.2 (and 3.3 when it is released) would be nice May 31 12:33:15 JaMa: atm I'm not thinking about runtime errors, just bitbake parsing May 31 12:33:43 is anyone else seeing the "rpmbuild: --buildroot: unknown option" errors? I have a fresh clone and see it, and it looks like a lot of autobuilders fail as well. May 31 12:37:08 erbo: yes, ross just sent a patch for it May 31 12:37:20 erbo: also https://bugzilla.yoctoproject.org/show_bug.cgi?id=4581 May 31 12:37:21 Bug 4581: normal, Undecided, ---, ross.burton, IN PROGRESS REVIEW , [Autobuilder] build failures because of unrecognized '--buildroot' option May 31 12:38:05 bluelightning: thanks May 31 12:45:18 JaMa: I think it would be better to solve the llvm-config issue properly so that it works for both native and cross builds instead of having to modify every recipe that depends on llvm-config May 31 13:09:59 starting work on llvm-config now May 31 13:53:52 llvm-config should work now May 31 13:54:29 back to Mesa now May 31 14:02:07 JaMa: do you know why we keep multiple LLVM versions around? May 31 14:28:50 Net147: openjdk being picky about version and 2.8 was more tested when I'was merging it from meta-java May 31 14:33:30 JaMa: okay. I think llvm-common.bb isn't installing llvm-config to crossscripts directory correctly May 31 14:38:50 JaMa: did you have any luck with llvmpipe? May 31 17:13:26 would there be any objection to pulling in a pre-gplv3 binutils recipe, either into oe-core or into toolchain-layer? perf requires binutils bits on the target, and oprofile can use them to find segment locations in vmlinux, so doing so will let you build/use perf & oprofile on non-GPLv3 images May 31 17:13:44 for now we have the recipe (patchset pulled from oe-classic) in our own layer, which works for now May 31 17:13:55 but figured it was possible someone else might find it useful May 31 17:19:09 Question, is the convention to use cherry-pick's -x option for yocto release branches, or not to? May 31 17:19:19 * kergoth is a fan of it, but wasn't sure if others tended to use it May 31 17:29:30 kergoth: I just change "oe-core rev" to "oe-core master rev" in my cherry-pick pull requests May 31 17:29:55 kergoth: re gplv2 binutils, I think that would be fine in OE-Core, we maintain other gplv2 recipes May 31 17:33:18 * kergoth nods May 31 17:33:20 okay, thanks **** ENDING LOGGING AT Sat Jun 01 02:59:58 2013