**** BEGIN LOGGING AT Fri Dec 16 02:59:56 2011 Dec 16 07:25:42 hi, all! In what package can I find arecord? Dec 16 07:29:42 I want to test mic on my panda Dec 16 07:29:45 *pandaboard Dec 16 07:42:14 gm Dec 16 08:25:43 good morning Dec 16 08:46:18 morning all Dec 16 08:49:47 Just tried a first build of oe-core. (bitbake core-image-minimal) fails on guilt-native-0.33-r1. "Unable to fetch URL" Dec 16 09:07:38 morning all Dec 16 09:19:56 Hi everybody Dec 16 09:20:56 I have a question about staging directory Dec 16 09:21:06 I have three directories in /tmp/staging: armv6-oe-linux-gnueabi, i686-linux and pc7302-oe-linux-gnueabi Dec 16 09:21:33 first one is assigned to HOST_SYS and second one is BUILD_SYS; what about third one? Dec 16 09:21:57 my machine name is pc7302. Is there any special environment specify pc7302-oe-linux-gnueabi directory Dec 16 09:21:59 ? Dec 16 09:22:09 as /tmp/staging/pc7302-oe-linux-gnueabi Dec 16 09:22:24 I want to use linux headers and I need to define kernel include directory path Dec 16 09:22:36 path is tmp/staging/pc7302-oe-linux-gnueabi/kernel/include/ Dec 16 09:22:51 I tried to examine from environments "bitbake -e" but as usual I could not :) Dec 16 10:25:45 hi all Dec 16 10:25:50 snowing here this morning :-} Dec 16 10:26:25 pb_: remind me, where are you in the UK? Dec 16 10:28:38 cambridge Dec 16 10:29:02 presumably RP has already had about ten feet of snow so far this year, but it's a bit of a novelty down here Dec 16 10:29:21 yes, I'm wondering if we'll see any further south this year Dec 16 10:30:36 or relative to recent years, how much Dec 16 10:30:51 whereabouts are you? Dec 16 10:43:42 pb_: SE london Dec 16 12:06:03 hi Dec 16 12:49:07 does anyone rely on the current values of SECTION for anything? Dec 16 12:49:30 obviously they go into the packaging system Dec 16 12:49:57 iirc some recipes used to define their SRCREV out of the section Dec 16 12:50:05 i don't know whether this has changed in the last 6 months :) Dec 16 12:50:10 err Dec 16 12:50:13 not SRCREV, of course Dec 16 12:50:14 S Dec 16 12:50:27 hmm, that's unusual... Dec 16 12:50:46 opie.bbclass used to use it as well Dec 16 12:50:56 or palmtop, i don't know Dec 16 12:52:15 hmm, a git grep of OE classic for "\${SECTION}" isn't showing up many results Dec 16 12:53:54 ah yes, found the code in opie.bbclass Dec 16 12:57:19 it's even documented ;) Dec 16 12:58:04 indeed :) Dec 16 13:04:48 A oe-core newbie question: How shall oe-core be built? I have made exactly as said on ps://wiki.yoctoproject.org/wiki/OpenEmbedded-Core. But the build crashes. Dec 16 13:05:49 Billy_, you could start by using pastebin to post the crash message where we could see it Dec 16 13:06:05 our crystal balls are off for the holidays Dec 16 13:06:56 The build does not succeed in fetching a package: NOTE: package guilt-native-0.33-r1: task do_fetch: Failed Dec 16 13:13:53 This is the pastebin with some more details: http://pastebin.com/jMZbEpi0 Dec 16 13:14:16 stupid kernel.org Dec 16 13:14:55 I havent used OE for some time, is this a usual problem nowadays? Dec 16 13:39:41 Hi, how can I choose glibc instead of eglibc when I building minimal-image? Dec 16 13:41:06 Am I the only one having problems building a new cloed git version of oe-core and getting it to crash at the fetch of guilt-native-0.33? Dec 16 13:47:48 Billy_, kenel;l.org dies and a lot of the files are not back Dec 16 13:48:25 I suspect a lot of peopl eusing oe core follow the instructions from the aAngstrom site Dec 16 13:48:36 and that will use a source mirror to work around the isssue Dec 16 13:48:38 I think Dec 16 13:49:01 my local sources dir already has it, so I won't see the problem :) Dec 16 13:52:22 Ok, so there are instructions on the Angrstrom site about this. I'll check that out. Thanx! Dec 16 13:59:28 Are 'ubifs' a valid IMAGE_FSTYPE? It isn't mentioned in local.conf Dec 16 13:59:44 s/are 'ubifs'/Is 'ubifs'/ Dec 16 14:00:07 risca: I'm using that here : IMAGE_FSTYPES += "ubi" Dec 16 14:00:17 ericben, roger that Dec 16 14:00:32 It is simply 'ubi', then Dec 16 14:00:46 risca: with MKUBIFS_ARGS = "-m 2048 -e 129024 -c 2030" & UBINIZE_ARGS = "-m 2048 -p 128KiB -s 512" Dec 16 14:00:54 and it builds ubifs then ubi Dec 16 14:01:02 Okay, gotcha Dec 16 14:01:15 Is there any documentation on this? Dec 16 14:01:19 so that you have a .ubi you can directly flash on your board Dec 16 14:01:24 risca: the code ;-) Dec 16 14:01:35 ericben, ^^ Dec 16 14:03:36 package dbus-native-1.0.1: failed Dec 16 14:03:43 can ne one help Dec 16 14:06:16 http://pastebin.com/VK8GSDUf Dec 16 14:16:38 ne1 http://pastebin.com/VK8GSDUf Dec 16 14:34:27 any one plz http://pastebin.com/VK8GSDUf Dec 16 14:36:49 kashyap: seems you are using an old build system with old sources on a recent distribution and that the gcc of this recent distribution doesn't like the old code ... Dec 16 14:37:17 im using opensuse 10.2 Dec 16 14:38:57 can you show me a way out>? Dec 16 14:39:47 kashyap: install a gcc version which can compile this code. Dec 16 14:40:00 kashyap: or upgrade to a recent OE Dec 16 14:40:23 gcc version newer or older> Dec 16 14:40:46 i have 4.1.2 right now Dec 16 14:41:39 kashyap: I would say older Dec 16 14:41:54 your code base seems to be dated in 2007 Dec 16 14:42:19 how can you guess that out? Dec 16 14:42:42 kashyap: you provided a pastebin Dec 16 14:43:13 ya i know that... but how did you guess that from the paste Dec 16 14:43:16 your distro is 2007 Dec 16 14:43:29 and dbus-1.02 was released in 2006 Dec 16 14:44:06 im installing gcc v 4.0 Dec 16 14:45:23 OE/Angstrom package nodejs-native-0.4.12-r0: task do_populate_sysroot: Failed -- http://pastebin.com/fRJgx85C any ideas? Dec 16 14:46:40 smithw101: I had the same problem, I could not really track it down Dec 16 14:46:48 something goes terribly wrong there Dec 16 14:46:58 nodejs has a pretty fsck'd up build script anyway Dec 16 14:47:15 it pretends to be autoconf while its not Dec 16 14:48:20 smithw101: you could try if that does the trick for you: https://gitorious.digitalstrom.org/dss-oe/dss-oe/blobs/master/dS/meta-dss11-testing/recipes-devtools/nodejs/nodejs_0.4.12.bbappend Dec 16 14:50:54 Jin^eLD: It's just adding --destdir=${D} to the ./configure ? Also, it adds PRINC="1" Dec 16 14:52:04 smithw101: well, don't forget the two sed lines Dec 16 14:52:21 The recipe already has those Dec 16 14:52:33 ah, ok Dec 16 14:52:44 I did not update for a couple of weeks Dec 16 14:53:05 Jin^eLD: I reset everything last night and started from scratch Dec 16 14:58:43 Jin^eLD: I just added the --prefix to the recipe and restarted … I don't know if it'll reconfigure it or not. I don't think it'll fix it. Seems that the issue is the path ends up with /// in the middle Dec 16 14:59:46 smithw101: it does because some var turns out empty, I know I fiddled around with it for quite a while and then it worked Dec 16 15:00:01 Jin^eLD: I figured it was something like that Dec 16 15:01:12 Jin^eLD: I'm moving from buildroot to OE/Angstrom, this whole system is new to me. Do you know what variables are empty and where that sysroot-destdir string is built? Dec 16 15:03:02 you can display the whole environment if you run bitbake -g yourpackage Dec 16 15:03:22 as for nodejs, I was not able to figure out where the path gets lost Dec 16 15:03:32 so that means - my fix did not work for you? Dec 16 15:05:28 Jin^eLD: It's still running … I don't know if it'll run do_configure() again since that part previously succeeded Dec 16 15:05:51 you could try bitbake -c cleanall nodejs-native and then build it again Dec 16 15:06:34 Jin^eLD: Ah, trying that Dec 16 15:10:27 help plz http://pastebin.com/VK8GSDUf Dec 16 15:21:56 http://pastebin.com/VK8GSDUf Dec 16 15:34:28 Jin^eLD: The cleanall worked, I retried the "bitbake -g cloud9-image", but it ended without doing anything! I'm retrying without -g Dec 16 15:35:03 smithw101: oops, my bad, -g does a dependency graph, I meant -e for environment, sorry Dec 16 15:36:53 smithw101: but I'd actually try rebuilding node-js to see if the change I showed you fixes the problem for you Dec 16 15:37:37 Jin^eLD: I'm running that now … I assume I probably could have done just: bitbake nodejs-native rather than the whole cloud9-image ... Dec 16 15:38:38 yep Dec 16 15:46:48 Jin^eLD: That did not fix it Dec 16 15:47:38 smithw101: interesting... I also believe that it has something to do with the build host as I had different behaviour on different systems Dec 16 15:47:57 Jin^eLD: I'm running this on Ubuntu 11.10, 64-bit Dec 16 15:49:03 I got it working on debian stable Dec 16 15:50:06 help me http://pastebin.com/VK8GSDUf Dec 16 16:15:52 Jin^eLD: Ok, lots of debug output, where do I start looking for these missing variables? Dec 16 16:17:04 smithw101: I don't know, I was not able to figure it out for real; tried various things in workdir and the solution I pasted worked for me in the end, but I was not really able to figure out what this thing happened that way in the first place Dec 16 16:17:19 smithw101: I think the build system of nodejs installs it to some weirdo location Dec 16 16:17:25 and OE can't find it then Dec 16 16:18:13 smithw101: I'd do a -c install, followed by a -c devshell and then play around with ../temp/run.do_compile/configure/etc and see where it gets installed on ../image/ Dec 16 16:18:19 03Philip Balister  07master * ra0d5f8af69 10openembedded.git/recipes/uhd/uhd-firmware.inc: Dec 16 16:18:19 uhd-firmware : File name no longer includes git hash. Dec 16 16:18:19 Signed-off-by: Philip Balister Dec 16 16:18:20 at least that's what I did Dec 16 16:18:20 03Philip Balister  07master * rf6df103372 10openembedded.git/recipes/orc/orc_0.4.16.bb: Dec 16 16:18:20 orc : Update to 0.4.16 per customer request. Dec 16 16:18:20 uhd needs at least 0.4.12. oe-core already has 0.4.14, so hopefully this Dec 16 16:18:20 does not have the segfault issues reported with other versions of orc. Dec 16 16:18:20 Signed-off-by: Philip Balister Dec 16 16:20:29 Jin^eLD: To double check on your solution: that paste was just a snippet, not a diff. Did you just replace the do_configure() as given and add the PRINC="1" ? Dec 16 16:24:01 smithw101: nono, this paste was a .bbappend file Dec 16 16:24:24 so, a recipe that gets appended to the existing recipe by bitbake on the fly Dec 16 16:25:20 Jin^eLD: Ahh … so I put that in the same directory as the nodejs recipe? Dec 16 16:25:37 for example, yes, I use it in my layer Dec 16 16:26:34 Jin^eLD: So I drop it into here: /home/wsmith/Work/beaglebone/setup-scripts/sources/meta-openembedded/meta-oe/recipes-devtools/nodejs/ Dec 16 16:26:42 And the system will pick it up? Dec 16 16:26:44 yes Dec 16 16:26:57 hence the PRINC - it is added to the .bb's PR Dec 16 16:27:03 and thus bumps the PR of the package Dec 16 16:29:59 it's really too bad bitbake doesn't have more types than just string internally Dec 16 16:30:13 PR = 1; PR += 1; and have PR be 2 would be nice Dec 16 16:30:20 heh Dec 16 16:30:27 :) Dec 16 16:30:49 oe-lite is starting to look in that direction, certain types are handled at parse time Dec 16 16:31:58 huh? oe-lite? we have classic, core, now lite? :) what is that? Dec 16 16:32:25 it's an earlier fork Dec 16 16:32:39 with some rather more adventurous changes to bitbake in particular Dec 16 16:32:46 it's not just a variant, it's deviated quite a lot, their bitbake code is almost entirely different Dec 16 16:32:58 the metadata is held in an in-ram sql database, etc Dec 16 16:33:01 file format changes Dec 16 16:33:12 (their event handles have dependencies like tasks, which is rather cute) Dec 16 16:33:37 * kergoth keeps an eye on it, but has never actually *used* it to do something useful Dec 16 16:35:11 richard and I talked to esben about trying to get some of the improvements piecewise into bitbake, he said he'd think about it Dec 16 16:35:23 (at ELCE this year) Dec 16 16:36:58 NOTE: Handling BitBake files: | (4957/4958) [99 %]ERROR: opening /home/group2/gu                              mstix/gumstix-oe/user.collection/packages/*/*.bb: [Errno 2] No such file or dire                              ctory: '/home/group2/gumstix/gumstix-oe/user.collection/packages/*/*.bb' Dec 16 16:37:06 what how to correct this Dec 16 16:37:20 you'd be better off talking to whomever created taht user.collection Dec 16 16:41:36 Can you please explain further Dec 16 16:42:08 how to change that path>? Dec 16 16:45:00 you need to fix your BBFILES. beyond that, we have insufficient context to say. talk to the gumstix people Dec 16 16:45:24 has nayone meet this problem : http://lists.linuxtogo.org/pipermail/openembedded-core/2011-December/014480.html ? Dec 16 16:53:25 Jin^eLD: one interesting, slightly controversial thing that oe-lite does is uses binary packages as the main artifact. the sysroots/staging is populated using binary packages, not a separate staging/sstate package, and each recipe gets its own isolated sysroot to build against Dec 16 16:55:21 isn't that a lot of overhead? Dec 16 16:56:33 it can add some, but you can also cache based on dependency combinations. not sure how much of that they do Dec 16 16:56:49 the upside is, you don't have to worry about optional dependencies and build ordering affecting things Dec 16 16:56:59 makes it more deterministic Dec 16 16:57:02 tradeoffs Dec 16 17:01:39 what I worry about is, and I had that with OE-core now, that build result differs depending on the host, i.e. on some hosts (distros) the build works, on others it fails Dec 16 17:01:58 yeah, that's always a concern Dec 16 17:02:09 yocto has a tool for identifying such contamination Dec 16 17:02:22 swabber Dec 16 17:02:56 analyzes an strace of a build to identify use of any files outside the build directory Dec 16 17:14:15 I also had some problems with utils like find, there seem to be differences on debian stable and on fedora/ubuntu Dec 16 17:14:26 I had to patch kernel.bbclass for stable Dec 16 17:50:15 Jin^eLD: Ok, it failed again with nodejs_0.4.12.bbappend dropped in. Same problem! Dec 16 18:03:20 khem, ping? Dec 16 18:08:16 hi ka6sox Dec 16 18:09:21 is there a oe image that can do something like openfiler does? SAN storage(not NAS) Dec 16 18:09:37 hiya woglinde Dec 16 18:09:48 i saw some NAS recipes to make the NAS image, but nothing for SAN(iscsi,fcoe,etc) Dec 16 18:09:59 don't really like openfiler's rpath stuff Dec 16 18:16:10 xxiao: as part of the yocto project we did a nas project called "baryon" a few months ago Dec 16 18:16:31 it was more an exercise for us in testing the tools, but it is functional Dec 16 18:16:51 I know I'd be keen to see someone extend it Dec 16 18:18:02 http://www.yoctoproject.org/blogs/jeff/2011/baryon-nas-software-created-yocto-project Dec 16 18:48:59 I wnat to ask something: Dec 16 18:49:31 how to exlude a cpufrequtils.bb from conssole-image compilation? Dec 16 18:49:54 I'm just new in OE and Angstrom ... Dec 16 18:50:06 look at the graph Dec 16 18:50:11 bitbake -g conssole-image Dec 16 18:50:20 where is the graph? Dec 16 18:50:38 bitbake -g produces it Dec 16 18:51:03 I will try it - but what to do after? Dec 16 18:51:32 first you can search what it drags in Dec 16 18:51:46 or you do the second now Dec 16 18:51:56 git grep cpufrequtils Dec 16 18:54:30 lovejoy68_ do you mean included in console-image or build process? Dec 16 18:58:17 lovejoy68_: bitbake -g -u depexp might be more friendly (brings up a dependency browser UI) Dec 16 18:59:19 I just try to to compile a console-image or x11-image (preferable) with hardfp ABI + neon optimisations. Dec 16 19:00:16 When I try to compile console-image , mostly all of the packages are succe3sfully compiled but cpufrequtils , so I try to exclude it temporary from the compilation chain Dec 16 19:00:32 or more exactly - the bitbake chain. Dec 16 19:01:56 The problem with cpufrequtils is the linking between cpufreq.o nad libcpufrequtils.so.0.0.0 - the first has VFP register arguments , the second dont have them Dec 16 19:19:19 * Russ wonders why bitbake keeps trying to fetch commit "1" from git Dec 16 19:19:35 didn't set SRCREV Dec 16 19:20:22 * Russ notices ${AUTOREV} Dec 16 19:22:40 bluelightning: thanks for showing depexp, this is far easier to use than .dot files ! Dec 16 19:26:39 ericben: no problem :) Dec 16 19:26:56 btw I'm still trying to fix an issue with the qt 4.8 recipes Dec 16 19:27:31 for some reason it's linking its libraries against the .so symlinks rather than .so. Dec 16 19:28:03 which causes do_rootfs to fail with unresolved dependencies :( Dec 16 19:35:57 bluelightning: that's very interesting, though i'm looking for SAN(iscsi) this time Dec 16 19:36:40 scst looks great, but was reject by the kernel, lio/tcm is used instead, and it sucks :( Dec 16 19:37:41 yes depexp is very usefull - thanks also! Dec 16 19:39:12 does somebody knows how from local.conf one can determine the ABI (softfp or hardfp) for all the packages? Dec 16 19:39:21 without exceptions? Dec 16 20:09:21 bluelightning: do you have your 4.8 recipes somewhere online ? Dec 16 20:10:14 ericben: I don't but I can push them now if you like Dec 16 20:11:32 well I can test if that can help Dec 16 20:14:17 ericben: http://cgit.openembedded.org/openembedded-core-contrib/log/?h=paule/qt-4.8.0-test Dec 16 20:14:31 ericben: testing would be welcome :) Dec 16 20:14:43 still stumped on this library version thing though Dec 16 20:14:53 bluelightning: I'll give it a try this weekend Dec 16 20:15:04 and let you know Dec 16 20:15:07 thanks Dec 16 20:15:37 hm newer A5 are build in germany, nice Dec 16 20:15:51 ups Dec 16 20:15:52 A9 Dec 16 20:15:53 sorry Dec 16 20:31:27 woglinde_: A9 from which manufacturer ? Dec 16 20:32:25 globalfoundries of course Dec 16 20:32:31 2.5 ghz Dec 16 21:02:41 ka6sox-away: whats up Dec 16 21:03:10 woglinde_: a9 is that a new audi model :) Dec 16 21:03:24 khem, are you getting status notifications from redmine? Dec 16 21:03:30 re: agate and garnet? Dec 16 21:03:32 ka6sox-away: yes I did thank Dec 16 21:03:39 ka6sox-away: yes Dec 16 21:03:40 old Garnet has to be offline. Dec 16 21:03:50 ka6sox-away: in progress Dec 16 21:03:57 ka6sox-away: ok take it away Dec 16 21:04:08 I have all needed stuff offloaded from it Dec 16 21:05:11 khem cortex-a9 auf course Dec 16 21:05:52 wo Dec 16 21:06:49 I want an r10 Dec 16 21:07:11 ? Dec 16 21:07:18 I want to drive the r18 Dec 16 21:07:18 audi) Dec 16 21:07:59 I have sat in an R10 :D Dec 16 21:08:12 I'd rather have the Lamborghini it's based off of Dec 16 21:09:20 heh I got a ride in one.. sometimes Silicon Valley is a fun place to be.. :) Dec 16 21:10:24 nice Dec 16 21:11:18 you might know the owner kergoth.. ;) Dec 16 21:19:38 hmm don't think so Dec 16 22:11:21 fray: some one at MV :) Dec 16 22:11:38 na, someone who was at MG Dec 16 22:11:53 hmm Dec 16 22:12:00 Dan Malek Dec 16 22:12:13 he borrowed one from a car club last time I was in town Dec 16 22:12:14 heh ironically is ex MV too Dec 16 22:12:16 isnt it Dec 16 22:12:19 yes he is Dec 16 22:12:39 cool Dec 16 22:13:00 they are nice I never had a chance to get into one yet Dec 16 22:13:42 although with age my adrenaline is less and less Dec 16 22:13:54 drink more coffee Dec 16 22:14:55 * Jin^eLD finds cars that one can hardly fix himself - boring and misses the good old times ;P Dec 16 22:18:55 woglinde_: we use to rent porche's in DE and drive like crazy Dec 16 22:19:18 single with some money and whacked brain :) Dec 16 22:19:57 heh Dec 16 22:20:15 we would try to see how fast it can go Dec 16 22:53:46 mtr: ich habe jetzt dss11-websetup upgedated (0.18.1), die version hat die heutigen fixes Dec 16 22:53:49 oops wrong window Dec 16 22:54:45 go on we spreak german Dec 16 22:55:06 well, the guy who's supposed to get this is not on this channel :) Dec 16 22:55:07 seit wann? ;) Dec 16 22:55:56 :> Dec 16 22:56:20 florian: if he's testing porshe speed limits in .de he must speak at least basic german to communicate with german polizei ;) Dec 16 22:57:06 Jin^eLD: every german is supposed to speak a little bit english ;) Dec 16 22:57:50 :) Dec 16 22:59:34 ich hab in DE gelebt Dec 16 22:59:41 hrm.... xf86-video-v4l2 is a nice example for a failure due to limited view on things by its developers Dec 16 23:00:58 good to declare the C implementation stuff that exists in assembler as well as "weak"... but trying to compile the armv7a assembler with a mips toolchain is doomed to fail somehow ;) Dec 16 23:01:11 khem: oh... where? Dec 16 23:01:39 florian: in Ulm :) Dec 16 23:01:45 and Munich Dec 16 23:02:01 HD video playback doesnt work for you ? Dec 16 23:02:18 khem: ok, not the worst area at least ;) Dec 16 23:02:39 florian: I still have mt lederhosen Dec 16 23:02:46 lol Dec 16 23:03:16 khem: heh... these strange stuff they wear in bavaria? Dec 16 23:03:22 I know enough to differentiate an austrian, bavarian, swiss german Dec 16 23:03:41 and of course hochduetsch Dec 16 23:04:55 khem: the one in oe-classic failed to fetch and lists an assembler file in its sources which gets compiled in any case. for some reason my mipsel toolchain does not like this ;) Dec 16 23:05:25 hmm heh Dec 16 23:05:32 is this asm for x86 only Dec 16 23:06:22 khem: austrian and bavarian is not that easy... but swiss people you recognize by slowness :) Dec 16 23:06:24 I once knew a CTO who only thought that intel is the sole company in world who makes CPUs Dec 16 23:06:44 heh Dec 16 23:07:32 no... its armv7a only so its likely that one of the ti folks is responsible Dec 16 23:07:33 florian: austrians have a singing tone and swiss have even more of a singling tone Dec 16 23:07:43 florian: ah. Dec 16 23:07:51 yes overrides were made for a reason Dec 16 23:08:08 and gcc has those builtin defines too Dec 16 23:09:00 * khem contemplates llvm as system compiler for OE Dec 16 23:20:03 that'd be cool Dec 16 23:30:48 kergoth: there is a project called elftoolchain which provides all binutils Dec 16 23:30:58 and lldb can be used for debugging elf Dec 16 23:31:12 so avoid all gpl if need be Dec 16 23:31:16 interesting Dec 16 23:31:37 see http://sourceforge.net/apps/trac/elftoolchain/ Dec 16 23:31:40 sounds like it'd be fun to experiment with Dec 16 23:32:08 and llvm has replacement for libgcc and libstdc++ already in place now Dec 16 23:32:33 its a viable solution for arm and x86 at least Dec 16 23:32:58 I think problem they have is no full distro is using these tools to compile the system Dec 16 23:33:27 there was a post where someone successfully compiled linux kernel with llvm for amd64 Dec 16 23:33:51 the stability of these tools will be way immature than what we have Dec 16 23:34:23 but if I have some spare time I would like to put together a meta-llvm with all above packages in it Dec 16 23:34:45 and it can then override gcc and friends Dec 16 23:35:42 actually its the gcc'ism that lot of software use Dec 16 23:35:50 that hinders other compilers getting in the game Dec 16 23:48:54 it'd be quite cool to use oe to exercise it, since often other distros don't rebuild their entire systems, just pieces as necessary Dec 16 23:49:00 * kergoth nods Dec 16 23:58:40 otavio: ping Dec 17 00:48:25 does oe-classic support git in SRC_URI ? Dec 17 00:48:38 yes Dec 17 00:49:18 awozniak: never heard of grep? :P Dec 17 00:49:38 kergoth: I think you have a special fondness for me now. Dec 17 00:49:59 was that supposed to make sense? Dec 17 00:50:44 locate linux_git.bb Dec 17 00:50:50 frack, wrong window Dec 17 02:13:57 /win 10 Dec 17 02:15:59 heh, irssi? **** ENDING LOGGING AT Sat Dec 17 02:59:57 2011