**** BEGIN LOGGING AT Wed Apr 23 02:59:58 2014 Apr 23 06:37:47 with all these patching going on in OE, doesn't it kind of make development more difficult? lets say i make a patch, and then i want to make another change. should I apply another patch, or change the previous one? i realize it becomes less dependent on a specific SCM using patches, but it also takes away some of the "joy" having a SCM. Right now I'm only using git to actually create the patch itself, basically git init, make change, git commit and Apr 23 06:37:47 then git formatpatch Apr 23 06:50:04 i'm also not sure how one can use review systems such as gerrit if I only use "one-shot" git repos which I throw away after the patch is done Apr 23 07:27:30 moin Apr 23 08:24:34 is it possible to disable volatiles ? Apr 23 08:29:59 never mind Apr 23 08:58:18 morning all Apr 23 09:03:11 bluelightning, morning Apr 23 09:03:16 hi kroon Apr 23 09:28:05 khem: did I see somewhere that you had a GCC 4.9 branch? Apr 23 09:29:06 jackmitchell: yes, email sent on oe-core, oe-dev and yocto last night Apr 23 09:30:02 ndec: ah hah, my mail filtering is a touch buggy when all 3 lists are CC'ed it all ends up in the yocto folder which I don't check that often, thanks! Apr 23 09:30:13 ;-) Apr 23 09:30:28 jackmitchell: just ask me next time you need to find an email ;-) Apr 23 12:40:36 I need to send in my ntp patch Apr 23 12:40:44 motivate me people! Apr 23 12:41:46 Crofton|work: do it! do it now! Apr 23 12:41:49 ;) Apr 23 12:42:09 as soon as I update work about state Apr 23 12:42:24 going to ELC next week and I need to make sure all my work is on the clouds Apr 23 13:06:05 http://pastebin.com/ZKsFFgU7 Apr 23 13:06:48 jsut passing a link between machines ignore and it will expire :) Apr 23 13:10:10 OK, hopefully that is goof Apr 23 13:10:13 good Apr 23 13:32:59 do_foo () { Apr 23 13:33:07 uses 4 spaces or 8 spaces? Apr 23 13:33:13 oestylize says spcaes not tabs Apr 23 13:37:23 Crofton|work: 4 spaces in meta-oe, single tab in oe-core Apr 23 13:37:37 rofl Apr 23 13:37:47 no wonder I am confused Apr 23 13:40:02 so the only thing that really anoys me about oe-stylize is wants me to remove a blank line between d_foo entries Apr 23 13:40:20 I suspetc whoever thought of this wasn't thinnking of the ice recipe Apr 23 13:51:28 bitbake still explodes if run without a MACHINE set somewhere Apr 23 14:03:58 my image does not contain my kernel but i have a kernel ipk that is generated... what did i forget/erase ? Apr 23 14:08:03 afournier: iirc, kernel aren't automatically added to images, they are in the 'deploy' folder along with bootloaders and rootfs. Apr 23 14:08:17 yes they are Apr 23 14:08:33 what do you mean? Apr 23 14:09:35 kernel are in deploy folder, and i was hoping that kernel would also be in the image, do i need to use a different class than core-image ? Apr 23 14:10:02 yes, it's possible to do it, i just don't remember otoh Apr 23 14:10:07 :( Apr 23 14:13:34 afournier: Apr 23 14:13:35 # Install kernel and symlink in /boot Apr 23 14:13:36 MACHINE_ESSENTIAL_EXTRA_RDEPENDS_append = " kernel-base kernel-image" Apr 23 14:13:38 afournier: i had a similar issue with beaglebone, where the kernel needs to be installed since uboot reads uimage from /boot. see https://github.com/koenkooi/meta-beagleboard/commit/6004a6eae800f449538fb510b12b62f860b8afca Apr 23 14:13:58 same problem here with syslinux/extlinux Apr 23 14:14:09 ok thanks ant_work and ndec ! Apr 23 14:14:23 well, both solutions are slightly different. Apr 23 14:14:47 i dont use yocto for kernels is this related ? Apr 23 14:14:56 MACHINE_ESSENTIAL_EXTRA_RDEPENDS_append = " kernel-base kernel-image" sounds ok to me Apr 23 14:15:12 i think you meant you don't use poky ;-) Apr 23 14:15:13 there is a 3rd option: only modules Apr 23 14:15:15 see [yocto] How to include kernel module without including the kernel in the rootfs? Apr 23 14:15:55 no thanks :) Apr 23 14:16:13 is it ok to add kernel-base kernel-image directly to IMAGE_INSTALL ? Apr 23 14:17:07 well, yes, but this is normally MACHINE_ scope Apr 23 14:17:34 by their names kernel-base kernel-image don't look to MACHINE related Apr 23 14:17:38 s/to/too/ Apr 23 14:17:51 I mean, some machine do boot from a fixed mtd partition Apr 23 14:18:04 i see Apr 23 14:54:26 do we need checksums for patch files now? Apr 23 15:04:49 Crofton|work: err, no... Apr 23 15:05:07 not unless they aren't local Apr 23 15:07:23 WARNING: Unable to get checksum for nativesdk-z-style SRC_URI entry 0002-Modify-Makefile-for-cross-compile.patch: file could not be found Apr 23 15:08:26 someting funny is going on Apr 23 15:25:18 Crofton|work: something is wrong with how you've placed the file with respect to where bitbake is looking for it Apr 23 15:26:07 this is a checksum for the purposes of detecting modifications, not a SRC_URI checksum Apr 23 15:26:27 (detecting modifications and re-triggering do_fetch, that is) Apr 23 15:27:10 Crofton|work: it's nativesdk so it could be a PN vs BPN issue perhaps? Apr 23 15:31:08 Any idea why toolchain (from dylan release) that was build in Ubuntu 13.10 is broken after update to 14.04? For example cross compiler arm-angstrom-linux-gcc core dumps (sigsegv). I tried to rebuild everything from scratch, but it didn't help. Apr 23 15:32:13 dbfi: I don't know what the cause is, but from what I keep hearing I think it's an angstrom-specific problem Apr 23 15:32:42 assuming this isn't a new problem... Apr 23 15:33:15 Thank's, I'll try to dig deeper in angstrom Apr 23 15:39:20 * kergoth discovers that do_populate_sdk isn't rerun when you change SDK_POSTPROCESS_COMMAND, rolls eyes Apr 23 16:03:02 bluelightning, can you explain the BPN versus PN Apr 23 16:15:01 i think volatile got superseded by /run no ? Apr 23 16:25:22 Crofton|work: BPN is PN without any prefixes/suffixes so if PN is "nativesdk-z-style" then BPN will be "z-style" Apr 23 16:25:50 I'm trying to figure out how to use this to solve my problem Apr 23 16:27:21 Crofton|work: where have you placed this file? and are bbappends involved or not? Apr 23 16:27:35 no bbappends Apr 23 16:29:05 http://pastebin.com/dVxXMe73 Apr 23 18:08:03 hmm binutils-crossdk just blew up trying to do populate_sdk **** ENDING LOGGING AT Thu Apr 24 02:59:58 2014