**** BEGIN LOGGING AT Fri Jul 15 02:59:57 2011 Jul 15 12:19:37 Tartarus: The images have specific test profiles. there is one defined for minimal and it should work Jul 15 14:19:18 RP__: ok, starting point. oe-core + core-image-sato, http://pastebin.com/ZLFd3u84 in my local.conf, still have to zap the root pw Jul 15 14:19:25 by hand in terminal, that is Jul 15 14:22:37 Tartarus: :/ Jul 15 14:22:50 Tartarus: I'd have to dive into it, I don't know why that isn't working Jul 15 15:04:36 * Tartarus tries on just poky Jul 15 15:36:18 morning all Jul 15 15:37:38 morning sgw Jul 15 15:37:42 hey sgw Jul 15 16:01:46 sgw: I imagine you want me to throw http://pastebin.com/QiPU1mBK into a bug Jul 15 16:04:50 Tartarus: probably not! Jul 15 16:05:55 Tartarus: you have been bit by the COMMERCIAL_LICENSE issue libmad non-"free" in commercial use Jul 15 16:06:15 This prevents world from parsing Jul 15 16:06:22 So... Jul 15 16:06:33 Tartarus: well, the bug to file would be bad error messages Jul 15 16:06:48 Tartarus: you need to set COMMERCIAL_LICENSE = "" Jul 15 16:07:06 Is that expected for world to build is my next question Jul 15 16:07:14 yes Jul 15 16:07:37 ok Jul 15 16:08:24 You also need to enable COMMERCIAL_AUDIO_PLUGINS and COMMERCIAL_VIDEO_PLUGINS Jul 15 16:08:40 COMMERCIAL_AUDIO_PLUGINS ?= "gst-plugins-ugly-mad gst-plugins-ugly-mpegaudioparse" Jul 15 16:08:40 COMMERCIAL_VIDEO_PLUGINS ?= "gst-plugins-ugly-mpeg2dec gst-plugins-ugly-mpegstream gst-plugins-bad-mpegvideoparse" Jul 15 16:09:11 All that stuff should get built anyhow, it's just having that installed, yes? Jul 15 16:09:24 I'm working on bringing our weekend builder up on oe-core Jul 15 16:09:32 So I'm not doing anything w/ the images, yes Jul 15 16:09:35 s/yes/yet/ Jul 15 16:09:53 Of course, crap, the qemu GL thing will bite me next Jul 15 16:09:58 Is there a way to opt out of that cleanly yet? Jul 15 16:15:20 Tartarus: which? Jul 15 16:15:28 the install GLU stuff Jul 15 16:16:14 hmm, sorry not following on that one, was there a problem there that I missed? Jul 15 16:16:58 # For our gl powered QEMU you need libGL and SDL headers Jul 15 16:17:12 I don't want a gl powered QEMU Jul 15 16:17:18 Is that possible, cleanly? Jul 15 16:17:19 right Jul 15 16:17:24 I know fray was annoyed about it before Jul 15 16:21:03 Tartarus: no clean way that I know. I think I understand your need, this is on a headless build server, no graphics stuff installed, right? Jul 15 16:21:13 Correct Jul 15 16:21:24 related to my oe-core question, it's a bare minimal build server Jul 15 16:23:51 yes, I saw that thread. I have no good answer for qemu right now, but I do understand the issue Jul 15 16:41:19 * Tartarus switches to building for u10.04 for a bit since that's more easily able to add in help2man and mercurial, sets VMs up as shutdown and revert to keep them clean for other jobs Jul 15 16:51:37 Whee, on to actual building time Jul 15 17:10:28 arg, does oe-core/yocto not have the "make zip files depend on unzip-native" stuff? :( Jul 15 17:19:54 * Tartarus sighs, kicks git am a bit, emails Jul 15 17:25:14 nitink, what's the preferred way to specify in a recipe that the compile should be done with an older version of gcc? Jul 15 17:25:32 I'm having issues trying to get Firefox 5.0 to build and I'm suspicious that gcc 4.6 may be making things difficult Jul 15 17:26:26 zenlinux_laptop: I have not tried switching compiler just for a recipe, I have done it for a machine Jul 15 17:26:58 zenlinux_laptop: the mozilla people generally discuss toolchain revs in a bug, should be easy enough to find Jul 15 17:27:17 walters, good idea Jul 15 17:28:25 zenlinux_laptop: you would need to see these 2 vars in the machine.conf or your local.conf GCCVERSION SDKGCCVERSION Jul 15 17:29:03 nitink: I think that will change gcc version for the whole image Jul 15 17:30:03 sgw: right, just changing for a recipe may bring weird issues Jul 15 17:30:04 ok...after perusing mozilla's bugzilla looking for gcc and the compile error I'm running into, I have the feeling it's not gcc's fault anyway Jul 15 22:07:43 * Tartarus sighs, goes off to cry in his 200char+ deep /paths, will start a branch on fixing world builds in that next Jul 15 22:12:06 Tartarus: what's going on with world builds? Jul 15 22:12:16 sgw: Well, this is a deep path problem Jul 15 22:12:32 | /bin/sh: /var/hudson/workspace/sb_oe-core_autobuilder/BBPKGS/world/label/sb-u1004/machines/qemuarm/build-eglibc/work/i686-linux/gnome-doc-utils-native-0.20.6-r2/gnome-doc-utils-0.20.6/doc/gnome-doc-make/../../xml2po/xml2po/xml2po: /var/hudson/workspace/sb_oe-core_autobuilder/BBPKGS/world/label/sb-u1004/machi: bad interpreter: No such file or directory Jul 15 22:13:45 blech, yeah that's not good it it :-( Jul 15 22:15:30 Yeah, we'll see if this is one I fixed in oe.dev before or not Jul 15 22:15:43 put out the mercurial-native/help2man-native branch just now Jul 15 22:15:57 So maybe next weekend I can run oe-core+qemu*+stuff in the env here Jul 15 22:55:41 'night alll Jul 16 00:42:27 Greetings - I wonder if I might ask a few questions re: new yocto BSPs and the expected workflow ? Jul 16 00:50:13 deckard026354: you may ask, it's getting late for most folks mind Jul 16 00:50:47 OK I'll be quick Jul 16 00:51:34 Following the setup instructions for a new BSP - doesn't populate a copy of a kernel tree Jul 16 00:52:11 deckard026354: which instructions are you referring to? TBH you may have more luck on the mailing list Jul 16 00:52:16 so - is the expected thing to do to do a git clone of the yocto kernel into some directory somewhere ? Jul 16 00:52:19 OK Jul 16 00:52:21 * incandescant is not au fait with the kernel pieces Jul 16 00:52:42 deckard026354: iirc the kernel manual suggests that? Jul 16 00:52:57 Nope Jul 16 00:53:30 I had expected cloning an existing BSP that it would be automagic Jul 16 00:53:49 no, the BSP is a ready to use machine setup Jul 16 00:54:33 if by BSP you're referring to, for e.g., meta-intel/meta-someintelboard ? Jul 16 00:54:44 spot on Jul 16 00:54:53 you still haven't told us which instructions you're following though :-) Jul 16 00:55:27 https://wiki.yoctoproject.org/wiki/Transcript:_creating_one_generic_Atom_BSP_from_another Jul 16 00:56:00 ah, not a document I've seen before Jul 16 00:56:14 you may need to complement it with http://www.yoctoproject.org/docs/1.1/kernel-manual/kernel-manual.html or http://www.yoctoproject.org/docs/1.1/bsp-guide/bsp-guide.html Jul 16 00:56:58 you can usually get good kernel help on the mailing list though if you want or need Jul 16 00:57:01 You're right Jul 16 00:57:37 I assume the thing to do is develop against the head of git Jul 16 00:57:56 .... but perhaps not ? Jul 16 00:58:03 yeah, there are pieces in place for using the git yocto-kernel recipes but I'm not familiar with them personally Jul 16 00:59:00 it may even be an extra layer to enable you to do so Jul 16 00:59:16 Ok. My impression. Upstream kernel drivers to regular Linux mailing lists. Upstream yocto-bsp stuff to yocto mailing list.... profit Jul 16 01:00:13 sounds about right Jul 16 01:00:35 and our kernel maintainers are usually happy to backport (or accept patches to) from upstream kernel to one of the yocto stable kernels so you don't have to maintain a tonne of patches in your layer Jul 16 01:01:06 I believe there's work on more documentation in this area Jul 16 01:01:15 and indeed, more tools being developed Jul 16 01:01:16 cool Jul 16 01:02:00 One thing is that I don't see defconfigs in the yocto kernel git repo "grep -r fishriver *" gives me nothing Jul 16 01:02:18 are these maintained in yocto seperately ? Jul 16 01:08:42 they are maintained as part of the kernel repo Jul 16 01:09:01 there are meta branches that include a bunch of config fragments that all get pieced together Jul 16 01:09:14 it enables very flexible config files Jul 16 01:09:44 groups of related options are maintained as features which can be easily enabled/disabled Jul 16 01:10:49 I think that's covered in http://www.yoctoproject.org/docs/1.1/kernel-manual/kernel-manual.html Jul 16 01:11:31 ok I'll rtfm a bit more Jul 16 01:11:35 cheers ! Jul 16 01:14:04 np Jul 16 01:14:08 * incandescant -> away **** ENDING LOGGING AT Sat Jul 16 02:59:57 2011