**** BEGIN LOGGING AT Thu Jul 28 02:59:57 2011 **** ENDING LOGGING AT Thu Jul 28 04:47:30 2011 **** BEGIN LOGGING AT Thu Jul 28 04:42:15 2011 Jul 28 06:59:18 morning all Jul 28 13:23:30 anyone have any ideas about what I should clean / rebuild to get rid of this: Jul 28 13:23:32 | Processing locale-base-en-us... Jul 28 13:23:32 | Unable to find package locale-base-en-us! Jul 28 13:23:51 eglibc-local Jul 28 13:24:00 * zeddii goes to find it himself as well. Jul 28 13:43:13 zeddii: this is at rootfs generation time? Jul 28 13:46:42 bluelightning, .. it is! Jul 28 13:47:17 I saw it once before months ago, I recall cleaning a lot of packages and finally getting it working again. but I didn't take any notes. Jul 28 13:47:53 yeah I've definitely seen it before Jul 28 13:48:07 clearly bitbake thinks everything is completed but the packaging system disagrees Jul 28 13:48:18 but as to the cause I'm in the dark Jul 28 13:48:47 at times I have had suspicions that it has been to do with arch mismatches Jul 28 13:49:12 I'm guessing locale-base-en-us should be noarch/allarch though right? Jul 28 13:51:23 could be. Jul 28 13:51:36 this is a build that I've been cycling through qemuarm/qemux86,etc Jul 28 13:51:45 it was fine fore core-image-minimal Jul 28 13:51:47 for all. Jul 28 13:52:05 then when I did x86 sato, followed by ppc sato ... you get the error on rootfs assembly. Jul 28 13:52:25 and as a kernel guy, I just want to get booting ;) Jul 28 14:41:43 argh. I really don't want to build from scratch. Jul 28 14:41:45 * zeddii tries again Jul 28 17:05:59 khem: ping Jul 28 17:06:43 sgw: whats up Jul 28 17:07:48 khem: with RP__ away, we need to push the PACKAGE_EXTRA_ARCHS fix to unbreak oe-core, do you know who else has write perms to it? The other option is to have a temp master in oe-core-contrib Jul 28 17:08:04 I can push it Jul 28 17:08:14 point me to patch Jul 28 17:08:28 on patchwork Jul 28 17:08:50 khem: http://patches.openembedded.org/patch/8805/ Jul 28 17:08:55 khem, it's Koen's patch Jul 28 17:09:42 khem: sgw: http://patches.openembedded.org/patch/8779/ Jul 28 17:09:49 his is probably better Jul 28 17:10:04 I was looking for that, bluelightning., thanks Jul 28 17:10:05 bluelightning: I thought ?? was fixed in bitbake Jul 28 17:10:18 khem: I'm not sure if this was the issue rp fixed Jul 28 17:10:31 does this show up with bb master Jul 28 17:10:33 Did someelse make a make a fix in bitbake? Jul 28 17:10:52 I am chceking now Jul 28 17:11:10 sgw: 32fee2e650dfdd3aa9a7572dad1251e0c24ca34b Jul 28 17:11:18 that's the bitbake rev Jul 28 17:11:29 I'm not sure it's related though Jul 28 17:11:45 ok Jul 28 17:11:52 so it happens with bb master too ? Jul 28 17:12:01 khem, I will verify and let you know shortly Jul 28 17:12:06 ok Jul 28 17:12:11 I haven't tested Jul 28 17:12:28 I imagine it does though because that's what Koen will be using Jul 28 17:12:49 yes Jul 28 17:13:02 I will wait for sgw to vefify Jul 28 17:13:11 to be sure Jul 28 17:18:47 there are other problems Jul 28 17:18:52 qemuarm is broken too Jul 28 17:18:58 did that get fixed Jul 28 17:19:20 tune stuff has caused a bit of churn Jul 28 17:19:26 khem: is that the cortex fix from Phil Jul 28 17:19:54 sgw: no its because thumb becomes default for 't' arches Jul 28 17:20:19 khem: Ah the other mail thread that I have been avoiding! Jul 28 17:20:45 khem: confirmed I just tried oe-core with bitbake master and the problem persists Jul 28 17:20:51 can you try a minimal build for qemuarm Jul 28 17:21:01 sgw ok thx Jul 28 17:21:16 khem uclibc or eglibc? Jul 28 17:21:24 glibc Jul 28 17:21:27 eglibc Jul 28 17:21:35 uclibc works ok Jul 28 17:21:48 because uclibc can be compiled pure thumb mode Jul 28 17:21:55 but eglibc/glibc cannot Jul 28 17:22:38 Many moons ago I dealt with ARM stuff for pjava and ejava and their successors Jul 28 17:22:56 It's all drifted away, but I still have the manuals on the shelf behind me! Jul 28 17:25:52 khem, building minimal image for qemuarm now Jul 28 17:31:07 tomz: ping Jul 28 17:33:41 khem: is the qemuarm failure in eglibc? Jul 28 17:37:40 fray: ping Jul 28 18:08:28 sgw: pong Jul 28 18:09:21 tomz: on the phone, but wanted to check with you about grub Jul 28 18:09:55 sgw: haven't had a chance to dig into it beyond my last comment, will have to get to it tomorrow Jul 28 18:14:18 tomz: thanks, I also wanted to see about moving grub into oe-core Jul 28 18:14:45 sgw: sounds like a 1.2 thing Jul 28 18:15:32 tomz: why would it be a 1.2 thing? we can add the recipe now, I can just copy it over to oe-core. Jul 28 18:16:13 sgw: ok, if you think it's ready - I don't have time to dig into changes to it now Jul 28 18:16:39 other thank the bug fix ? Jul 28 18:17:09 sgw: oh, I thought you were talking about grub2, nevermind Jul 28 18:17:54 sgw: yeah, I don't know where the bug is yet in any case, just started happening out of the blue, like so many other things Jul 28 18:19:47 tomz, sgw I've been wondering if it makes sense to even keep grub at all Jul 28 18:20:19 It's "great" for a desktop where you have multiple kernels, maybe chainloading, etc - but for embedded platforms it seems like overkill Jul 28 18:20:30 souldn't syslinux be fine for all our needs? Jul 28 18:20:39 s/^s/w/ Jul 28 18:21:50 dvhart: hmm, I can see wanting to have multiboot capabilities, e.g some of the ecg use cases Jul 28 18:22:17 I don't know that syslinux can't do that Jul 28 18:22:22 tomz, you mean linux + windows? Jul 28 18:23:11 dvhart: no, more like using a machine as a packet processor and as a general-purpose machine at other times e.g. romley + dpdp Jul 28 18:23:36 tomz, syslinux can do that just fine Jul 28 18:24:11 it can do multiple boot targets with different arguments and the UI allows for changing arguments to those labels Jul 28 18:24:13 dvhart: ok, well, then whatever's simplest and covers everything, you know, occam's razor and all that Jul 28 18:24:28 We could drop grub and grub2 and just use syslinux Jul 28 18:24:38 and we already use it for the live images Jul 28 18:24:51 would really cut down the x86 boot loader complexity in yocto Jul 28 18:25:48 Let's look at that for 1.2 then, I would want to check with other to ensure no one was planning on enabling grub for other arches. Jul 28 18:25:50 dvhart: sounds like it might be a good idea Jul 28 18:26:05 sgw, yup, that's fine Jul 28 18:26:17 dvhart: my bbmatrix is *almost* done -- looks like there's a max of about 10 min variation Jul 28 18:26:24 how long did you say a build took you? Jul 28 18:26:53 which image? -sato? Jul 28 18:27:02 or... no minimal right? Jul 28 18:27:15 iirc minimal takes me 57 minutes Jul 28 18:28:36 does DEPENDS = "intltool" look usefull to anyone? intltool-native for sure, but is target intltool usefull in buildtime? Jul 28 18:29:01 polkit and libfm have it.. I have patch for both, but newer used libfm Jul 28 18:32:47 dvhart: interesting, slower than mine even with 12 cores :-) Jul 28 18:34:29 guess I better check .. maybe that was for sato... Jul 28 18:34:47 hrm yes, that was for sato Jul 28 18:35:34 jefferai, which commit id are you building and for what machine? Jul 28 18:39:16 JaMa: I think your right, sounds good. Jul 28 18:40:00 jefferai, kicked off a minimal build, will let you know the results Jul 28 18:44:35 sgw: ok will send both Jul 28 18:51:50 dvhart: doh, that was for sato? :-) Jul 28 18:52:05 I haven't done sato yet on this box but sato took me > 24 hours on my previous one Jul 28 18:52:07 hehe Jul 28 18:52:13 haha Jul 28 18:52:32 you should be able to do sato in <120 minutes on a quadcore Jul 28 18:52:50 this is with pseudo already built? Jul 28 18:52:53 no Jul 28 18:52:53 or without? Jul 28 18:52:58 clean start Jul 28 18:53:00 hm Jul 28 18:53:03 I guess we'll see Jul 28 18:53:18 >50% task count on minimal build Jul 28 18:53:19 I can do minimal in about 50 minutes Jul 28 18:56:30 bbl Jul 28 19:08:38 khem: ping Jul 28 19:19:10 jefferai, real 38m22.485s Jul 28 20:16:09 fray: ping Jul 28 21:57:24 okay, either I didn't know this existed, or I forgot Jul 28 21:57:27 * kergoth__ rolls eyes at self Jul 28 22:07:21 sgw: yes Jul 28 22:07:32 today is a crazy day for meetings Jul 28 22:23:29 khem: if you can push the PACKAGE_EXTRA_ARCHS to oe-core that would be great Jul 28 22:31:55 sgw: I think khem already did? Jul 28 22:33:24 * incandescant does a pull and confirms Jul 28 23:15:33 sgw: how did minimum build for qemuarm go ? Jul 28 23:16:20 khem failed in eglibc Jul 28 23:17:38 sgw: yes. Is fray aware of that ? Jul 28 23:17:45 or is there any patch out there Jul 28 23:18:00 I havent had any time to look in the stuff Jul 28 23:18:12 I tried to get fray's attention today, but he is not around. I have been fighting other fires Jul 28 23:20:16 "More than one recipe type class inherited in binutils-crosssdk-native: native, cross, crosssdk" .. sheesh Jul 28 23:38:10 sgw: how about other arches are they building ok ? Jul 28 23:48:34 khem, world builds seem to be ok for the other arches, but ppc seems to have rootfs issues still Jul 28 23:49:44 do we know the root cause Jul 28 23:51:49 khem looks similar to the PACKAGE_EXTRA_ARCHS, but I have not dug into it yet, I have had other issues today, I am not sure if Kumar's posts address this or other ppc issues. Jul 28 23:55:41 Khem, I will start looking at now also. Jul 28 23:56:15 sgw: ok. I would like the arm issue to be resolved as well. Jul 28 23:56:29 So if there is any patch posted try to test it out Jul 28 23:59:27 khem: I have not tested any of the patches at this point, will looking at that shortly. **** ENDING LOGGING AT Fri Jul 29 02:59:56 2011