**** BEGIN LOGGING AT Fri Apr 18 02:59:59 2014 Apr 18 08:40:01 is there a way to tell bitbake to continue its work even if a task is failing, so it would stop when there is no work available because all remaining tasks depends on failing tasks ? Apr 18 08:40:53 this would makes my coffee management process easier :) Apr 18 08:40:57 -k Apr 18 08:41:09 thanks ! Apr 18 08:41:13 provided you do 'bitbake -k target1 target2' Apr 18 08:41:26 failure in target1 won't impact target2 Apr 18 08:41:35 it's not 100% what you're looking for, though Apr 18 08:42:04 bitbake -k my-image would stop even if a depency fails ? Apr 18 08:42:11 yes Apr 18 08:42:29 so no coffee for me until every recipe works Apr 18 08:43:00 right Apr 18 08:43:11 but having coffee first tends to help things Apr 18 08:44:44 sure Apr 18 08:48:30 :) Apr 18 09:28:27 ka6sox: git+ssh:// seems down on git.oe.org Apr 18 09:44:39 close Building recipe: xfce4-panel (1705/1776) Apr 18 11:24:34 ka6sox: nvm, git+ssh:// is working again Apr 18 11:25:40 JaMa: found any new issue wrt meta-initramfs / meta-handheld? Apr 18 11:35:14 ant_work: I'm still not sure what deleting the meta-initramfs layers buys you Apr 18 11:35:43 then again, I'm grumpy because a number of BSPs cleaned up their recipes and delete bbappends which made PR go backwards Apr 18 11:35:48 so ignore me :) Apr 18 11:36:29 koen: I know it is handy to have an intermediate layer to base the .bbappends upon Apr 18 11:38:21 I have nothing against letting all in place as it is, was just wondering about some recipes needing regularly maintainance (the linux-yocto-tiny-kexecboot kernel in this case) Apr 18 11:39:37 it must match an existing kernel in oe-core...when 3.10 will be removed the recipe will break Apr 18 11:39:41 ant_work: don't know yet, test-dependencies script is still running, probably will finish today, then I'll send report Apr 18 11:39:52 JaMa: ok Apr 18 11:39:54 thx Apr 18 11:40:22 ant_work: fwiw I kind of agree with koen about meta-initramfs Apr 18 11:40:31 koen: and the point is, people only do grep in oe-core when updating/removing stuff so any extra layer is in danger... Apr 18 11:40:56 but that is true for all layers Apr 18 11:41:05 ant_work: it shouldn't be too hard to make the kernel recipe independent of the core one Apr 18 11:41:18 and still we're not going back to monolitic repo Apr 18 11:41:54 I'm not complaining, just observing what's happened in the past Apr 18 11:41:58 ant_work: the solution to that problem is probably more regular testing of meta-handheld / meta-initramfs Apr 18 11:42:09 koen: for my BSPs I've sent corresponding PR bumps to oe-core, problem is when these 2 layers are upgraded independently :/ Apr 18 11:42:20 JaMa: I know Apr 18 11:42:39 JaMa: and it's extra annoying because dropping bbappends is the right thing Apr 18 11:42:46 surely that is a temporary problem? we shouldn't ever be in that situation again Apr 18 11:43:15 in the 1.6 branch for angstrom I still have a lot of PRINC warnings left Apr 18 11:43:22 once those are gone, we should be safe Apr 18 11:43:34 so, what would you guys suggest do with that linux-kexecboot recipe? we have a working 3.10 and I don't think we need updates any soon Apr 18 11:44:07 sometimes I would use something like DISTRO_PE if only PE isn't too ugly Apr 18 11:44:27 we had DISTRO_PR in oe classic Apr 18 11:44:48 I know, but sometimes you need even bigger hammer :) Apr 18 11:44:54 indeed Apr 18 11:45:57 when changing DEFAULTTUNEs in SHR I was incrementing all values in PR serv db Apr 18 11:46:08 has anyone written something like debian-installer for OE produced distros? Apr 18 11:46:14 that's relatively simple implementation of DISTRO_PR :) Apr 18 12:14:22 must submit make ntp do something patch Apr 18 14:29:05 why does udev have so many exotic keymaps in /lib/udev/keymaps/ ? Apr 18 18:04:17 this is eveil Apr 18 18:04:38 I need to change an env var in a recipe, depending on native or target build Apr 18 18:07:00 use the _class-native or _class-target override. quite common Apr 18 18:07:20 export ICE_HOME = "${STAGING_DIR_NATIVE}/usr" Apr 18 18:07:28 export ICE_HOME_class-native ? = "${STAGING_DIR_NATIVE}/usr" Apr 18 18:07:34 export ICE_HOME_class-native = "${STAGING_DIR_NATIVE}/usr" ? Apr 18 18:07:35 der Apr 18 21:00:22 yup Apr 18 21:01:59 it looks like there is something missing in the name, bzImage--2.6.32.27-r0-alix-20140418153235.bin what is it ? Apr 18 21:02:33 yeah, my bbb images are coming out that way too Apr 18 21:02:43 not rpi, just bbb Apr 18 21:03:47 modules--3.8.13-43-r1b-beaglebone-20140417181059.tgz Apr 18 21:03:52 so it's machine related Apr 18 21:03:57 yeah modules tool Apr 18 21:04:01 looks like just the kernel recipe for me Apr 18 21:04:08 modules and uImage Apr 18 21:04:10 what is there normaly ? Apr 18 21:04:18 not sure Apr 18 21:04:47 i take that back, same thing in rpi now Apr 18 21:04:58 ok Apr 18 21:05:09 * nerdboy just started a clean rpi build Apr 18 21:06:14 modules--3.10.30+git26d1de2993c81c01f91ebc93a1fad8957c9ece17-r0.2-raspberrypi-20140418162520.tgz Apr 18 21:06:56 KERNEL_IMAGE_BASE_NAME ?= "${KERNEL_IMAGETYPE}-${PKGE}-${PKGV}-${PKGR}-${MACHINE}-${DATETIME}" Apr 18 21:07:23 PKGE is missing... so the question is : what is PKGE ? Apr 18 21:09:31 http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#1.3-linux-kernel-naming Apr 18 21:10:00 PE = the epoch of the recipe. By default, this variable is unset. The field is used to make upgrades possible when the versioning scheme changes in some backwards incompatible way. Apr 18 21:11:53 what does a package epoch even look like? an int? Apr 18 21:12:06 no idea, but it looks like everything is normal and under control :) Apr 18 21:13:04 i think it should be a int if it's related to versionning and backward compat Apr 18 21:13:26 but i don't see any usecases Apr 18 21:15:29 could be usefull when downgrading a kernel Apr 18 21:59:05 JaMa, i thought I'd have a go at one of those floating dependencies in your mail. So "console-tools/console-tools/latest lost dependency on flex", does this mean that when console-tools is built in "world"-populated sysroot, it picks up a dependency on flex, but not when its built in a minimal sysroot ? Apr 18 22:23:57 someone pointed out to me that http://cgit.openembedded.org/openembedded-core/tree/README Apr 18 22:24:07 does not tell you how to send in patches Apr 18 22:24:42 crofton send a patch Apr 18 22:24:49 rofl Apr 18 22:25:38 I wonder how RP would take that so close to the release Apr 18 22:27:07 good nite and happy eastern **** ENDING LOGGING AT Sat Apr 19 02:59:58 2014