**** BEGIN LOGGING AT Tue Nov 12 02:59:58 2013 Nov 12 10:21:18 morning all Nov 12 10:36:59 hi all Nov 12 12:59:15 Are there any llvm+clang bitbake recipes? I found some in meta-openembedded for llvm but none for clang. Nov 12 15:12:48 hi, getting the following error when running bitbake. followed the tutorian in oe handboot. nay hints? ERROR: Unable to parse conf/bitbake.conf: ParseError at conf/bitbake.conf:377: unparsed line: 'IMAGE_EXTRA_SPACE = 10240' Nov 12 15:13:01 *tutorial, *handbook Nov 12 15:13:18 sounds like you're trying to use old metadata with a new bitbake, won't work Nov 12 15:13:41 i thought about that. Nov 12 15:14:05 which git branch in openembedded repo is the right one? i am currently on master Nov 12 15:15:05 no openembedded repo is called oe-classic Nov 12 15:15:15 openembedded classic, just 'openembedded', is getting pretty ancient now, it's been replaced by oe-core. read http://www.openembedded.org/wiki/OpenEmbedded-Core Nov 12 15:15:19 JaMa: :) Nov 12 15:15:48 so the official openembedded pdf handbook is horribly outdated? Nov 12 15:18:32 I guess you could say that :] Nov 12 15:20:08 :)) Nov 12 15:20:11 thanx guys Nov 12 15:29:40 is there a preferred location for custom metadata? Nov 12 15:30:13 or just clone the git repos in the oe-core dir? Nov 12 15:35:57 ccube: create a new layer with your custom recipes in Nov 12 15:39:49 yeah Nov 12 15:39:58 so the meta-* folders are the overlays? Nov 12 15:40:33 in a sense Nov 12 15:40:42 :) Nov 12 16:02:23 bb-files are similar to portage ebuilds Nov 12 16:02:31 are they vcompatible in some way? Nov 12 16:03:22 bitbake was based in part on portgage, so that's unsurprising. the format is different in a number of ways, though, and unlike ebuild, is directly parsed, not shell Nov 12 16:03:39 this allows use of python tasks, inline python, and other features Nov 12 16:04:19 kk Nov 12 16:04:55 then, at least, i can feel home a little bit! :D Nov 12 16:06:08 does this mean, that all receiped found in openembedded folder will be build?? BBFILES = "${TOPDIR}/openembedded/recipes/*/*.bb" Nov 12 16:06:18 nope, it just makes them available Nov 12 16:06:34 aight Nov 12 16:06:39 in the oe-core universe, configuration is different, and each layer provides its own additions to bbpath/bbfiles Nov 12 16:06:42 so then i didnt get how to select them Nov 12 16:06:58 bitbake will build that recipe. everything else comes in through build and run time dependencies Nov 12 16:07:28 see the "image" recipes which pull in others and produce root filesystems Nov 12 16:07:36 aaah, so i have to find the core-image-minimal receipe, copy it and add packages to it? Nov 12 16:07:55 there are variables you can set in local.conf to affect what recipes do Nov 12 16:08:13 see the sample local.conf provided by oe-core Nov 12 16:08:29 but yes, that's a viable option too Nov 12 16:41:15 jackmitchell, members list updates done, so email away with website ideas Nov 12 16:45:01 Crofton|work: aye, thanks. Just finding the time to write a sensible proposal now, hopefully sometime this week/weekend Nov 12 16:45:15 keep the umbrella! Nov 12 16:45:48 koen: it shall be present, the image will be the same; I just intend put some scope in on what I want to achieve Nov 12 16:46:01 I'm trying to understand in the 'gcc' configuration what actually enables / disabled fortran support Nov 12 16:46:26 fray: LANGUAGES iirc Nov 12 16:46:42 koen: people commonly look at the design and don't interpret the content (or maybe I've been conversing with sales too much) Nov 12 16:46:51 thats what I thought.. but it appears on first glance that fortran is enabled by default -- yet it's not producing the fortran binaries.. Nov 12 16:47:03 (and no I don't really want fortran, I'm just trying to understand it) Nov 12 16:47:36 # gcc 3.x expects 'f77', 4.0 expects 'f95', 4.1 and 4.2 expect 'fortran' Nov 12 16:47:36 FORTRAN ?= ",f77" Nov 12 16:47:37 LANGUAGES ?= "c,c++${OBJC}${FORTRAN}${JAVA}" Nov 12 16:47:46 maybe newer gcc renamed it again Nov 12 16:47:52 (that's from oe-classic gcc-configure.inc) Nov 12 16:48:16 ahh I may have found it Nov 12 16:48:33 the FORTrAN ?= get evaluated -after- the gcc-4.8.inc which appears to set it to "" Nov 12 16:48:49 fray: but beware that RP hates fortran, so there might be some logic bombs in place ;) Nov 12 16:49:00 bingo Nov 12 16:49:03 # Language Overrides Nov 12 16:49:03 FORTRAN = "" Nov 12 16:49:03 JAVA = "" Nov 12 16:49:10 that is the part I was missing.. Nov 12 16:49:28 (I've looked at this off and on for a couple of days and hadn't been able to figure it out.. now I get it) Nov 12 16:49:58 afair there is a broken symlink in sysroot for fortran Nov 12 16:50:08 wouldn't surprise me.. :) Nov 12 16:50:18 I'm working on a patch to switch the symlinks to alternatives.. Nov 12 16:50:22 ant_work: I think we might have fixed that Nov 12 16:50:37 fray: not in the -cross packages you aren't ;-) Nov 12 16:50:43 I saw that not long ago trying to find a place for klcc-cross Nov 12 16:50:49 no not -cross Nov 12 16:50:51 target Nov 12 16:50:56 as I have more then one compiler on the target Nov 12 16:52:18 koen: I actually fixed fortran issues recently Nov 12 16:52:23 koen: not sure what came over me Nov 12 16:52:27 I'm sorry Nov 12 16:52:50 if anyone can fix libquadmath better than my horrible hack I'd love someone to Nov 12 16:53:24 * mwester2 notes that FORTRAN's legacy (column-sensitivity) lives on --- in Python! Nov 12 16:54:08 http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/wipqueue3&id=ed927d786fed3d207f14f12a3bef7e4f5d8398d5 Nov 12 16:55:10 mwester2: just when I was starting to think python wasn't so bad... ;-) Nov 12 16:58:27 ha Nov 12 17:29:41 moin Nov 12 17:30:17 tach Nov 12 18:34:06 what is the difference between these assignments: = / ?= / ??= Nov 12 18:34:23 both the bitbake manual and the poky/yocto documentation cover that Nov 12 18:35:00 should read that :)) Nov 12 18:35:36 yocto has actual technical writers, so the docs have been coming along nicely Nov 12 18:36:02 to sum up, ?= is an immediate "set if this variable isn't already set", this is based on gnu make's operator Nov 12 18:36:18 thx Nov 12 18:36:19 ??= is a more lazy form of the same, where the last ??= to be defined wins, rather than the first as is the case with ?= Nov 12 18:36:26 both are commonly used for default values **** ENDING LOGGING AT Wed Nov 13 02:59:58 2013