**** BEGIN LOGGING AT Fri Apr 26 02:59:58 2013 Apr 26 07:16:38 is RCONFLICTS really working as documented? I found I had added some bash-depending recipes to a busybox image, added RCONFLICTS_$[PN} = "bash" to the busybox recipe and still got an image with both bash and busybpx Apr 26 07:17:21 I had expected an error Apr 26 07:49:17 Good morning Apr 26 07:56:17 morning apelete Apr 26 08:15:12 hi silviof Apr 26 08:51:49 hi Apr 26 08:53:07 will rootfs tar after fakeroot do_rootfs () function or before? Apr 26 08:55:52 bcoz whatever i'm writing in fakeroot do_rootfs () function is not reflecting in the tar file, but coming in the rootfs folder Apr 26 09:18:46 morning all Apr 26 09:52:02 hi pb_ Apr 26 10:36:06 In my custom distro conf I have IMAGE_FSTYPES = "tar.gz" Apr 26 10:36:24 however somewhere tar.xz gets into the image types Apr 26 10:36:48 should setting the above variable in a distro conf not be the final word? Apr 26 10:37:27 ah, found it Apr 26 10:37:34 common-bsp/conf/machine/beaglebone.conf:IMAGE_FSTYPES_append = " tar.xz " Apr 26 10:53:27 jackmitchell: imho machines should not add such generic ones Apr 26 10:54:11 hrw: yeah, I've pinged koen about it, will try at least to get IMAGE_FSTYPES += " tar.xz" so I can override it Apr 26 10:58:15 jackmitchell: IMAGE_FSTYPES_forcevariable = "tar.gz" maybe? Apr 26 10:58:58 hrw: I didn't know about the _forcevariable, thanks Apr 26 10:59:35 jackmitchell: I learnt from pb_ Apr 26 11:00:27 be aware that _forcevariable is really strong.. so in some exceptions you need then add constructs like: Apr 26 11:00:33 IMAGE_FSTYPES_forcevariable_pn-initramfs-kexecboot-klibc-image_spitz = "cpio.lzma" Apr 26 11:01:11 well _spitz is optional, but you get the idea Apr 26 11:02:37 I used it twice in external linaro toolchain config. to kill multilib and to remove languages from images Apr 26 11:05:16 well in public layer it's even worse, because then there is only few options left for end-user to override it from local.conf Apr 26 11:14:20 JaMa: when external toolchain is used there is no locale packages. so image will fail at do_rootfs() due to IMAGE_LINGUAS not being empty. - in this case I think that forcing lack of locales in images is fine Apr 26 11:16:02 JaMa: and second use was GCCMULTILIB which is usable only when gcc-cross is also multilib capable (like OE default one). Linaro cross is not multilib configured and gcc with --enable-multilib (default setup in OE) was not buildable Apr 26 11:30:27 hrw: I'm not saying that you shouldn't use it.. just that it should be considered as last resort option when nothing else works Apr 26 11:49:40 what happened to the shr branch in openembedded-core-contrib? Apr 26 11:57:59 rah: it was moved to https://github.com/shr-distribution Apr 26 11:58:45 rah: this way I can tag each shr staging feed without unwanted noise in openembedded-core-contrib Apr 26 11:59:24 I see Apr 26 12:00:03 JaMa: the documentation hasn't been updated Apr 26 12:01:21 rah: shr Makefile was and some documentation too, can you point me to outdated one or fix it yourself? Apr 26 12:04:40 JaMa: sure Apr 26 12:06:30 JaMa: http://shr-project.org/trac/wiki/Building%20SHR#BuildingSHRmanuallyfromscratch Apr 26 12:08:18 rah: hmm maybe we should drop this one and use the new one (marked as in progres) Apr 26 12:08:30 I've updated the new one only Apr 26 12:14:29 rah: update Apr 26 12:14:32 d Apr 26 13:54:54 ERROR: ParseError at /home/rah/phones/shr/shr-core-gta04/meta-smartphone/meta-samsung/recipes-kernel/linux/linux-samsung-tuna_git.bb:12: Could not inherit file classes/kernel_android.bbclass Apr 26 13:54:58 :-/ Apr 26 14:19:00 wtf is this kernel_android.bbclass? Apr 26 14:19:04 I can't find it anywhere Apr 26 14:19:20 including searching the web Apr 26 14:19:43 rah, hi Apr 26 14:19:57 it's probably in meta-smartphone Apr 26 14:20:32 it was probably used for SHR kernels for the phones comming from android Apr 26 14:23:28 hmm Apr 26 14:23:29 you're right Apr 26 14:23:35 $ find meta-smartphone/ -name "*android.bbclass" Apr 26 14:23:35 meta-smartphone/meta-android/classes/kernel_android.bbclass Apr 26 14:23:48 why isn't the build picking it up I wonder? Apr 26 14:25:03 ffs Apr 26 14:25:10 JaMa's gone now :-/ Apr 26 14:25:14 http://shr-project.org/trac/wiki/Building_SHR_without_chroot Apr 26 14:25:18 those isntructions are wrong Apr 26 14:25:56 $ ls common/conf/shr-core Apr 26 14:25:56 ls: cannot access common/conf/shr-core: No such file or directory Apr 26 14:27:27 rah, ok Apr 26 14:45:01 argh... who made libexecdir dependent on ${PN} and non-multilib compliant? Apr 26 14:47:16 openssh installs sftp-server as ${libexecdir}/sftp-server; how can this be referenced from other packages (dropbear) now? Apr 26 14:49:40 ensc|w: saul wold, apparently, since you ask. 406bd38b4232f9f399ef5ffe0b4fac72ed605a23 Apr 26 14:56:08 we actually put it back to the old value in our distro config, since we have too many things that break with the new layout. Apr 26 14:56:41 in the particular case of sftp-server, arguably it ought to be installed in ${sbindir} or something rather than going in libexecdir at all. Apr 26 14:57:06 I think the theory is that files in ${libexecdir} are meant to be private to the package in question and not used by anything else. Apr 26 14:58:49 JaMa: http://shr-project.org/trac/wiki/Building_SHR_without_chroot Apr 26 14:58:54 JaMa: those isntructions are wrong Apr 26 14:58:57 $ ls common/conf/shr-core Apr 26 14:58:59 ls: cannot access common/conf/shr-core: No such file or directory Apr 26 15:01:35 rah: it's wiki, please update it Apr 26 15:05:08 JaMa: to what? Apr 26 15:05:17 JaMa: I don't know what was in that directory in the past Apr 26 15:05:18 pb_: theory behind libexecdir is "The definition of ‘libexecdir’ is the same for all packages" (http://www.gnu.org/prep/standards/html_node/Directory-Variables.html) Apr 26 15:09:15 rah: I've removed '/shr-core' to match https://github.com/shr-distribution/shr-makefile/commit/efc4af409206c6f83480f147c685fd195f7847a7 Apr 26 15:11:48 ensc|w: yeah, unfortunately the linux FHS doesn't appear to have embraced libexec as a concept. I imagine that the reason for the oe-core change was trying to get the FHS compliance badge. Apr 26 15:11:58 JaMa: thanks Apr 26 15:17:03 you mean unlock the FHA achievement Apr 26 15:17:07 you mean unlock the FHS achievement Apr 26 15:17:09 damn it Apr 26 15:17:12 hm, this intel integrated graphics is teh suck Apr 26 15:17:32 * pb_ tosses another motherboard out of the window Apr 26 15:18:37 how do I get bitbake to tell me where the .bb file is for some package? Apr 26 15:19:38 bitbake -e | grep FILE Apr 26 15:20:03 maybe grep ^FILE= for a bit less extraneous spew Apr 26 15:20:39 pb_: that doesn't seem to work Apr 26 15:21:29 $ bitbake -e | grep gst-plugins-ugly Apr 26 15:21:31 $ Apr 26 15:21:42 er, no, that's not what I meant Apr 26 15:21:53 bitbake -e gst-plugins-ugly | grep ^FILE= Apr 26 15:22:04 you need to type FILE literally, not replace it with the name of the package Apr 26 15:22:13 I see Apr 26 15:22:27 i think it might be ^FILE_DIRNAME= Apr 26 15:22:40 I get neither Apr 26 15:22:51 ERROR: Nothing PROVIDES 'gst-plugins-ugly' Apr 26 15:23:03 well, in that case, there is no .bb file for that recipe Apr 26 15:23:10 so, the answer to your original question is "nowhere" Apr 26 15:23:29 but I get: Apr 26 15:23:30 ERROR: gst-plugins-ugly was skipped: because it has a restricted license not whitelisted in LICENSE_FLAGS_WHITELIST Apr 26 15:23:32 ah sorry pb_ i see the ^FILE= now Apr 26 15:24:05 oh, hm, I see. yeah, that is a bit annoying. Apr 26 15:24:27 there must be a .bb file otherwise how does bitbake know what the license is? Apr 26 15:24:54 yeah, there is. unfortunately if it decides that the .bb file is unwholesome then it doesn't fully parse it and doesn't enter it into the datastore. Apr 26 15:25:08 which is why "bitbake -e" doesn't admit to knowing anything about it. Apr 26 15:25:15 so Apr 26 15:25:19 how do I get bitbake to tell me where the .bb file is for some package? :-) Apr 26 15:25:20 in that case, I think your best option is probably "find . -name gst-plugins-ugly*.bb" Apr 26 15:25:39 rah: I can't think of any way to get bitbake to tell you under those circumstances. Apr 26 15:26:01 that sucks :-/ Apr 26 15:26:06 yes it does Apr 26 15:26:19 try -DDD | grep gst-plugins-ugly Apr 26 15:26:24 it's particularly unhelpful that it doesn't even tell you what the "restricted license" is so that you can whitelist it Apr 26 15:26:48 where is the whitelist anyway? Apr 26 15:26:55 you might like to file bug reports for both those issues, maybe the h4x0rs will fix them Apr 26 15:27:40 it's a variable, LICENSE_FLAGS_WHITELIST or something Apr 26 15:27:44 my liking to do that is.. possible :-) Apr 26 15:27:52 where would such a variable get set? Apr 26 15:27:59 there doesn't appear to be much documentation for how the license stuff works, though maybe it's in the yocto wiki somewhere Apr 26 15:28:07 distro.conf, local.conf, site.conf, any of those places Apr 26 15:32:02 how can I find the value that this variable has? Apr 26 15:32:27 rah: wasn't it already in local.conf from shr-makefile? https://github.com/shr-distribution/shr-makefile/commit/c48b8344461b69a03710d7f855add4fd3d6e30b1 Apr 26 15:33:42 JaMa: no, I didn't overwrite my existing local.conf Apr 26 15:34:06 * rah fixes Apr 26 15:35:40 (I thought I'd checked that) Apr 26 16:22:14 I want to append to a recipe's PACKAGECONFIG in my local.conf Apr 26 16:22:35 but this line replaces , it does not append: PACKAGECONFIG_pn-foo_append = " bar" Apr 26 16:22:48 bitbake -e shows PACKAGECONFIG="bar" Apr 26 16:22:58 PACKAGECONFIG_append_pn-foo Apr 26 16:23:11 is what you want Apr 26 16:24:11 ah. thanks, this works. Apr 26 16:41:57 moin Apr 26 17:03:04 hey. how can i disable the QA test about a package having a .so symlink? Apr 26 17:03:16 which needs to be distributed Apr 26 17:03:34 CMoH: search for INSANE_SKIP Apr 26 17:03:40 CMoH: and dev-so Apr 26 17:03:40 ty Apr 26 17:21:22 I'm trying to write recipe for gnome-themes-standard v3.4.2. Compiling fails because glib-compile-resources doesnt recognize .svg-files.. Apr 26 18:19:42 I need to make gdk-pixbuf build the svg-loader.. Apr 26 18:25:55 but how Apr 26 18:46:13 so if I have recipes for version 1.0.1 of a software, and want to update the recipes to 1.0.2 , do I keep the 1.0.1 ones around, or can I rename them? Apr 26 18:47:13 usually rename Apr 26 18:58:55 For the gnome-themes-standard_3.4.2 bb recipe, I added "librsvg-gtk-native" to DEPENDS, however I dont see gdk-pixbuf-query-loaders being run to update the loaders.cache file.. Apr 26 18:59:27 but at least it builds now, if I manually update the cache from a devshell Apr 26 19:00:45 hi jkroon Apr 26 19:00:53 hey woglinde Apr 26 19:19:41 * waynr is a little confused by the bitbake client/server code Apr 26 19:53:38 and meta-oe is broken again .. :/ Apr 26 19:53:48 QA Issue: non -staticdev package contains static .a library: python-numpy **** ENDING LOGGING AT Sat Apr 27 02:59:58 2013