**** BEGIN LOGGING AT Sat Apr 13 02:59:58 2013 Apr 13 08:36:14 lol Apr 13 09:41:19 hmm will fixes be backported to the danny branch in meta-oe ? the libvpx softfp fix from master also works in danny, and without it, the recipe breaks if callconvention-hard is used Apr 13 09:45:44 dv_: they can be if you send them for backport (see README in danny) Apr 13 09:45:51 okay Apr 13 10:06:08  Apr 13 10:06:08  Apr 13 11:34:34 JaMa: shouldn't we be applying those meta-systemd patches for oe-core recipes in oe-core? Apr 13 13:19:08 bluelightning: yes someone should Apr 13 13:19:20 "someone" ? Apr 13 13:19:28 bluelightning: I'm doing meta-oe layers Apr 13 13:19:54 oe-core bbappens will probably stay at least with PRINC and upgrade-path fixes :/ Apr 13 13:20:35 until PV bumps for those who were using meta-systemd I suppose there is not a lot of choice Apr 13 13:20:39 so I have no motivation to move them Apr 13 13:21:01 well forever for upgrade-path (RPROVIDES/RREPLACES/RCONFLICTS combo) Apr 13 13:21:08 which was refused to go to oe-core Apr 13 13:21:08 for you yes Apr 13 13:21:25 nobody who never used meta-systemd needs that path Apr 13 13:21:46 yes but adding it to systemd.bbclass would cover 90% cases Apr 13 13:21:56 when PN-systemd was moved to PN Apr 13 13:22:49 well, you know what RP thinks about R* upgrade paths spread around various places Apr 13 13:23:15 fine if they're in an inc file that those who need it can include at the top level Apr 13 13:23:31 there are only 2 recipes where it's not in PN openssh and dhcp Apr 13 13:23:56 bluelightning: that's why I propose to add them in systemd.bbclass Apr 13 13:24:28 well, I wouldn't necessarily be opposed to that Apr 13 13:24:42 if it helps us remove the need for meta-systemd Apr 13 13:25:11 I need to make some time to test out our systemd integration Apr 13 13:25:34 good luck :) Apr 13 13:25:49 why? Apr 13 13:26:10 there is quite a lot of issues Apr 13 13:26:48 JaMa: what are the most pressing ones? Apr 13 13:26:59 failing postinst/prerm/postrm scripts Apr 13 13:27:26 this is with sysvinit removed ? Apr 13 13:27:30 and run-postinsts hanging sometimes on first boot Apr 13 13:27:33 (from DISTRO_FEATURES) Apr 13 13:27:35 yes Apr 13 13:27:56 * bluelightning starts a fresh build on another machine Apr 13 13:30:17 * JaMa sends another batch of meta-systemd removal patches Apr 13 13:30:56 FYI I should be able to send some meta-oe -> meta-networking / meta-webserver move patches out next week Apr 13 13:33:56 bluelightning: ok, then please move them with my meta-systemd patches applied Apr 13 13:34:02 so we don't get many conflicts Apr 13 13:36:11 now only 17 bbappends for oe-core recipes Apr 13 13:41:33 right, I will do that Apr 13 14:45:19 i cant for the love of it disable qemu-native while building my image Apr 13 14:47:30 is it required for building a root fs ? Apr 13 14:54:38 usually no Apr 13 14:54:56 only some recipes depend on it Apr 13 14:55:30 im trying your patch next, the one you showed woglinde yesterday Apr 13 15:02:16 hm right the patch Apr 13 15:02:29 I am building again and when the image works Apr 13 15:02:33 I will send it out Apr 13 15:03:11 its my first patch so I need to set the upstream Apr 13 15:06:20 status and so on Apr 13 15:06:23 for oe-core Apr 13 15:11:44 hehe :) Apr 13 15:11:56 that reminds me that llvm move Apr 13 15:12:14 haha Apr 13 15:12:48 jama do I need to bump pr now for new patches in oe-core? Apr 13 15:12:56 or is just adding the patch enough Apr 13 15:12:56 woglinde: no Apr 13 15:12:59 okay Apr 13 15:13:00 thanks Apr 13 15:13:04 was not sure Apr 13 15:14:53 JaMa: which postinsts are failing? Apr 13 15:20:30 e.g. dbus udev if you don't remove update-rc.d first Apr 13 15:20:46 woglinde: isn't llvm missing in some meta-java DEPENDS? Apr 13 15:21:17 woglinde: I don't see it required anywhere after moving it to meta-oe Apr 13 15:21:25 jama its disabled at the moment Apr 13 15:22:22 ok, but added to rpath if I read git grep right Apr 13 15:22:41 ah and then cleaned for _arm Apr 13 15:23:04 at least in openjdk-7 Apr 13 15:26:21 jkroon_: in recent master, qemu is used to attempt to do font/icon cache creation on the host rather than first boot Apr 13 15:30:29 so lets see if it starts up Apr 13 15:32:01 hm looks good Apr 13 15:32:03 nice Apr 13 15:33:01 hm I wonder why we patch qemu for 384M and runqemu still uses 128 Apr 13 15:33:34 woglinde: hmm, maybe ask khem about that since IIRC he added that patch Apr 13 15:38:39 woglinde: fwiw I'm using different scripts then runqemu to run qemu Apr 13 15:38:45 so maybe he does the same Apr 13 15:39:46 hm runqemu is fine for me Apr 13 15:40:08 besides I need to send the fix for new kernel-image location Apr 13 15:40:48 hm whats the patch status for not sended upstream? Apr 13 15:41:02 woglinde: Pending Apr 13 15:41:37 uhm Apr 13 15:41:47 even if I did not send it upstream? Apr 13 15:41:55 woglinde: yes Apr 13 15:42:01 okay Apr 13 15:42:05 it means pending submission Apr 13 15:42:16 if you've submitted it it should be "Submitted" Apr 13 15:43:19 woglinde: FYI: https://lists.yoctoproject.org/pipermail/yocto/2011-April/003446.html Apr 13 15:43:20 bl is http://paste.debian.net/plain/249312 okay? Apr 13 15:43:58 woglinde: should say briefly what the patch fixes and have Signed-off-by Apr 13 15:44:11 hm Apr 13 15:44:17 its in the oe patch Apr 13 15:44:24 okay I will double it Apr 13 15:44:54 I made the qemu patch with quilt Apr 13 15:45:12 sure, just having the reference in the in-tree patch makes it easier when reviewing patches for a recipe on upgrade Apr 13 15:45:41 bbl Apr 13 15:45:43 (and when upstreaming) Apr 13 15:46:22 hm right Apr 13 16:09:37 bluelightning, can i switch to letting that target system do it instead on boot ? Apr 13 16:25:16 jkroon_: it'll do that anyway Apr 13 16:26:18 if running it on the host fails Apr 13 16:29:34 bluelightning, yeah, wouldnt that stop pulling in qemu-native from being built ? cause right now im struggling with getting an image I can download to the sd-card Apr 13 16:29:42 *but wouldn't* Apr 13 16:30:05 jkroon_: how is qemu-native preventing you from doing that? Apr 13 16:31:35 bluelightning, I can't build qemu-native in master, not sure why, but woglinde has an upcoming patch for it, and if i understand things correctly, I cant do bitbacke my-core-image without building qemu-native Apr 13 16:33:55 jkroon_: easiest thing would be just to apply his patch Apr 13 16:34:00 it's not straightforward to disable Apr 13 16:34:37 bluelightning, ok, thanks Apr 13 16:34:47 woglinde, do you have the patch published somewhere yet ? Apr 13 16:36:20 jkroon_: which host distro are you building on? Apr 13 16:36:27 bluelightning, Fedora 18 Apr 13 16:36:46 hmm, I wonder why qemu-native is failing to build then? Apr 13 16:37:24 bluelightning, something with fdt headers, maybe related to which kernel youre building, but im just guessing.. woglinde knows the gory details Apr 13 16:55:52 bluelightning, im using the meta-beagleboard layer, building 3.8.x kernels, dunno which is the default for qemuarm Apr 13 16:56:25 the kernel you're building should have nothing to do with whether qemu-native builds successfully or not Apr 13 16:56:40 qemu-native is intended for running on the host Apr 13 16:57:55 ok, i thought qemu maybe used some kernel-headers or something Apr 13 17:20:27 no I was shopping Apr 13 17:24:22 no Apr 13 17:24:28 for kernel headers Apr 13 18:35:35 so patch is sent Apr 13 18:42:10 woglinde, thanks Apr 13 18:54:21 woglinde: your patch doesn't explain how this issue comes up Apr 13 18:56:32 ???? Apr 13 18:59:03 woglinde: jkroon_ says this shows up on Fedora 18, and yet I just built an image on Fedora 18 for a qemu machine and ran it successfully Apr 13 18:59:57 I think woglinde and I are both using the meta-beagleboard layers, right ? Apr 13 18:59:59 it fails on every Apr 13 19:00:13 dtc/libdt is the problem Apr 13 19:00:48 the typedefed the stuff and qemu did not catched it up jet Apr 13 20:01:09 bah, qemu-native on qemuarm fails as well Apr 13 20:14:56 hm cleanuped patchwork too Apr 14 01:04:13 * nerdboy can't seem to find a recipe for daemontools... Apr 14 01:04:36 is there another/preferred solution for that? **** ENDING LOGGING AT Sun Apr 14 02:59:58 2013