**** BEGIN LOGGING AT Wed Jan 23 02:59:59 2013 Jan 23 06:43:45 hello. are there any fuse packages in oe-core that provide archive filesystems? something like fuse-zip, for instance Jan 23 06:45:13 well, oe-core, meta-oe or suh Jan 23 06:45:15 such Jan 23 08:34:22 good morning Jan 23 08:37:49 hi mckoan Jan 23 08:37:54 stupid icu Jan 23 08:38:04 needs std=gnu99 instead of std=c99 Jan 23 08:38:10 c99 does not define timezone Jan 23 08:46:53 morning Jan 23 08:47:23 is it possible to overload the oe-core/meta/files directory with a layer ? i'd like to overload fs-perms.txt Jan 23 08:49:08 hm hm Jan 23 08:51:14 i tweaked the BBPATH, we'll see :) Jan 23 08:54:01 yay it works Jan 23 08:56:26 how did you tweak it? Jan 23 08:57:16 BBPATH =. "${LAYERDIR}:" instead of BBPATH .= ":${LAYERDIR}" Jan 23 08:58:00 so bitbakes looks in my layer first, then oe-core's Jan 23 08:58:46 i guess i can add custom licenses now Jan 23 09:07:57 ah Jan 23 09:08:01 that I need too Jan 23 09:08:57 hm custom liceses should work anyway Jan 23 09:11:30 hi stefn schmidt Jan 23 09:23:03 morning all Jan 23 09:24:35 gm bluelightning, all Jan 23 09:25:28 morning bluelightning Jan 23 09:25:38 hi afournier, mckoan Jan 23 09:47:01 rburton: morning Ross, I tried to bisect the image with no success at all, a few revisions didn't compile and a couple more did odd things as they were clearly meant to be part of a patchset Jan 23 09:47:38 I'm not convinced it is the systemd changes now, but I think I'm going to wipe tmp, update to master again today and try again Jan 23 10:09:30 hi bl Jan 23 10:29:18 okay Jan 23 10:29:26 uclibc does not define __timezone Jan 23 10:29:39 will send a patch later for icu Jan 23 10:37:51 hm and stupid util-linux dev forgot to include rpmatch header a couple of times Jan 23 10:38:34 hi Jan 23 10:38:35 indeed, quite silly Jan 23 10:38:55 does someone know why my fresh image lacks /etc/rc* dirs? Jan 23 10:39:04 (I had some trouble with it trying to port a few things to darwin from util-linux) Jan 23 10:40:24 also lack of /etc/inittab Jan 23 10:45:40 ok. looked at fresh rootfs. /etc/rc*.d/ are present Jan 23 10:46:46 but when I compare working and non-working image then lot of things is missing Jan 23 10:50:40 ok. initscripts are missing Jan 23 10:51:53 with systemd in DISTRO_FEATURES or without? Jan 23 10:54:28 withoutconf/site.conf:DISTRO_FEATURES = "x11 alsa argp ext2 largefile usbgadget usbhost xattr nfs zeroconf ${DISTRO_FEATURES_LIBC}" Jan 23 10:56:59 I do not have good experience with systemd so far so prefer good old sysvinit Jan 23 10:57:56 let me try without own D_E set Jan 23 11:02:48 D_E? meta/recipes-core/packagegroups/packagegroup-core-boot.bb:VIRTUAL-RUNTIME_initscripts ?= "${@base_contains('DISTRO_FEATURES', 'sysvinit', 'initscripts', '',d) Jan 23 11:03:27 hrw: so if you have DISTRO_FEATURES overriden in your distro config you have to include DISTRO_FEATURES_INITMAN Jan 23 11:03:46 to get sysvinit in D_F and initscripts in ackagegroup-core-boot Jan 23 11:03:52 thanks Jan 23 11:08:41 that sounds like a bug to me... surely it ought to have been backfilled Jan 23 11:10:07 bluelightning: by default DISTRO_FEATURES_INITMAN is added to DISTRO_FEATURES. But we define own set to limit amount of builds and services Jan 23 11:10:27 no bluetooth, no wifi, no anything that can be safely removed Jan 23 11:10:55 fastmodel is slow enough without them. Jan 23 11:14:16 hrw: sure, but the idea with backfilling is that features that used to not be optional get added for everyone so their configurations don't break Jan 23 11:14:46 http://www.yoctoproject.org/docs/latest/poky-ref-manual/poky-ref-manual.html#ref-features-backfill Jan 23 11:16:53 I guess it could be that we weren't able to determine which feature value should be backfilled as people might have been using either Jan 23 11:17:16 bluelightning: hi, regarding yesterday's issue with an ancient bitbake on angstrom, updating bitbake on sources/layers.conf hits tons of tab-related errors in almost every recipe :< Jan 23 11:18:04 mnemoc update your layer Jan 23 11:19:04 woglinde: tried with master,HEAD on all of them :< Jan 23 11:19:37 ???? Jan 23 11:19:42 woglinde: in some cases the tabs were in variables I couldn't even find Jan 23 11:19:45 bluelightning: so I should just use DISTRO_FEATURES_BACKFILL_CONSIDERED and list there all things I do not want to have? Jan 23 11:19:56 that's how I understood doc after first read Jan 23 11:20:03 mnemoc oe-core meta-oe are working Jan 23 11:21:00 mnemoc: you can checkout the 1.15.2 tag, that appears to be the last release that didn't warn about that issue Jan 23 11:21:17 hrw: only for those that are in DISTRO_FEATURES_BACKFILL Jan 23 11:21:22 bluelightning: includes the fix for my `git remote add` issue? Jan 23 11:21:25 ok Jan 23 11:21:34 mnemoc: one second Jan 23 11:21:55 bl does intel care about uclibc is it possible to set up a jenkis job for it or should I asked jama? Jan 23 11:22:03 mnemoc: actually no, it doesn't Jan 23 11:22:10 bluelightning: but only pulseaudio is there ;( Jan 23 11:22:47 hrw: right... anything else can just be removed by not including it in the value of DISTRO_FEATURES that you are setting as previously Jan 23 11:23:13 bluelightning: sure, but then I have to track all changes to not get broken images like now I had Jan 23 11:23:30 DISTRO_FEATURES_INITMAN should not be in DISTRO_FEATURES_BACKFILL? Jan 23 11:23:35 woglinde: wait for public jenkins and we'll need more disk space (right now I'm always running out of space with 1 world build) Jan 23 11:24:05 woglinde: it's not really up to intel Jan 23 11:24:49 woglinde: I think though within the yocto project we don't actively test uclibc although it is supported; so if you are able to regularly test it that would be great Jan 23 11:24:50 jama public jenkis, when is it expected? Jan 23 11:25:27 don't know, depends on those disks Jan 23 11:25:37 hrw: I don't know, but the idea of backfilling was to avoid this exact type of situation Jan 23 11:26:05 JaMa: maybe OE should sponsor some disks for you :) Jan 23 11:26:07 well jenkins is public already, but we cannot allow people to add more jobs to it because of disk space Jan 23 11:26:49 bluelightning: OE servers are used as jenkins slaves, so OE should sponsor more disks for OE :) Jan 23 11:27:35 JaMa: I see... well, yes we should :) Jan 23 11:29:40 bluelightning: it work Jan 23 11:30:02 http://pastebin.com/tC2EDkae Jan 23 11:30:25 hrw: right, of course it would... my question is should we not be doing that in OE-Core already Jan 23 11:30:52 I have up to date oe-core and it is not done Jan 23 11:32:13 no, I know that Jan 23 11:37:12 mnemoc: there isn't a release that has the one fix and not the other unfortunately Jan 23 11:38:33 haha another glitch in windows rt Jan 23 11:38:51 sent my fix to ML Jan 23 11:39:15 bluelightning: so switchihg everything to master's HEAD is the way for me? Jan 23 11:40:10 mnemoc: you can, but less drastically you could just use 1.15.2 with bitbake revision 2ba23df5fad4b94d38a6aed97f7822226d72eb89 cherry-picked on top Jan 23 11:40:29 let me try Jan 23 11:43:26 ok, restarted fresh build with my fix Jan 23 11:58:39 If I manually put sysvinit in DISTRO_FEATURES in my custom distro will that take precedence over DISTRO_FEATURES_INITMAN or what ever the variable is Jan 23 11:59:22 for example, if I were to then put systemd in my DISTRO_FEATURES would "sysvinit" get backfilled and appened to DISTRO_FEATURES as DESITRO_FEATURES_INITMAN is defaulted to "sysvinit" Jan 23 12:00:11 and would the fix be to set DISTRO_FEATURES_INITMAN in my distro conf? Jan 23 12:04:58 jackmitchell: you need to set DESITRO_FEATURES_INITMAN too because it's used e.g. to populate VIRTUAL-RUNTIME_init_manager in packagegroup-core-boot.bb Jan 23 12:05:51 khem: if libsystemd-core needs symbols from libsystemd-daemon, why not link those to instead of linking test_unit_file with it? Jan 23 12:06:10 JaMa: ok, and is it fine to set it in your distro conf file rather than bitbakes local.conf? Jan 23 12:06:16 hrw: what's D_E? it's also used in commit message you just send Jan 23 12:06:30 jackmitchell: yes distro.conf is fine Jan 23 12:07:36 JaMa: short of DISTRO_FEATURES? ops. should be D_F Jan 23 12:09:05 jackmitchell: something like this: Jan 23 12:09:06 DISTRO_FEATURES_INITMAN = "systemd" Jan 23 12:09:06 DISTRO_FEATURES += "${DISTRO_FEATURES_INITMAN}" Jan 23 12:09:07 # in case http://patchwork.openembedded.org/patch/43209/ gets applied Jan 23 12:09:07 DISTRO_FEATURES_BACKFILL_CONSIDERED += "sysvinit" Jan 23 12:09:17 khem: ok, looks like you only followed the previous style... but still... Jan 23 12:09:57 jackmitchell: if your distro is using own DISTRO_FEATURES definition (not weak assignment from default-setup.inc) Jan 23 12:10:12 JaMa: yep thats what I now have, thanks! Jan 23 12:11:38 fatal the question should be why the corelib needs symbol from some other lib Jan 23 12:13:49 heh I've tried to trim down oe-core to be more minimal (to test world-image) and some core apps/libs are greedy as hell Jan 23 12:14:37 jama yes Jan 23 12:15:14 hm how to fix relocations Jan 23 12:15:23 JaMa: I'm still not totally convinced that this _FEATURES_ chaos is the best way to implement UE flags :/ Jan 23 12:15:27 QA Issue: ELF binary '/devel/oe/setup-scripts/build/tmp-angstrom_2010_x-uclibc/work/armv5te-angstrom-linux-uclibceabi/openjdk-7-jre/03b21-2.1.3-r3.0/packages-split/openjdk-7-vm-zero/usr/lib/jvm/java-7-openjdk/jre/lib/arm/server/libjvm.so' has relocations in .text Jan 23 12:15:29 *USE :) Jan 23 12:16:05 ech. fix was not enough Jan 23 12:16:10 ant_work: PACKAGECONFIG is to implement USE flags, FEATURES are only to select right PACKAGECONFIG Jan 23 12:16:53 ant_work: and good thing is, that when default selection of PACKAGECONFIG options is not optimal for some distro, it's easy to redefine them in .bbappend Jan 23 12:17:26 I'm trying to whittle down my DISTRO_FEATURES to get a very small build for my headless distro and I can't seem to find out what largefile is supposed to do? Is it support for files > 4Gb or something along those lines? Jan 23 12:18:13 it would be really useful if all DISTRO_FEATURES were documented in the yocto manual Jan 23 12:19:23 jackmitchell: +1 Jan 23 12:19:27 jackmitchell you defined you a new task Jan 23 12:19:42 JaMa: afais, the pkgconfig stuff corresponds to the local USE flags while the features should be the global USE flags Jan 23 12:20:35 DISTRO_FEATURES="x11 alsa argp ext2 largefile usbgadget usbhost xattr nfs zeroconf ipv4 ipv6 libc-backtrace libc-big-macros libc-bsd libc-cxx-tests libc-catgets libc-charsets libc-crypt libc-crypt-ufc libc-db-aliases libc-envz libc-fcvt libc-fmtmsg libc-fstab libc-ftraverse libc-getlogin libc-idn libc-inet-anl libc-libm libc-libm-big libc-locales libc-locale-code Jan 23 12:20:48 notice sysvinit at the end... Jan 23 12:20:56 build lacks initscripts Jan 23 12:22:37 JaMa: what fundamentally differs is, you can hardly synthetize things like USE="-kde -qt4 ldap" Jan 23 12:28:26 hm what I have to find out again what the qemustart command was Jan 23 12:29:33 ah runqemu Jan 23 12:30:16 rpm's are borked for uhd Jan 23 12:42:58 do_rootfs takes such a long time - no hdd activity and minimal cpu activity I have no idea what it's doing - it's certainly on my list to investigate and possibly try to slim down a bit Jan 23 12:44:27 slaughtering is already planned Jan 23 12:51:28 ant_work: you still do some zaurus c7x0 work? Jan 23 12:51:40 yes Jan 23 12:51:48 ant_work: will you be at fosdem? Jan 23 12:52:08 no, probably not :/ Jan 23 12:52:21 ant_work: I still have spare leather case for c7x0 Jan 23 12:52:24 btw, I had the bad idea to get a collie for Xmas :) Jan 23 12:52:30 (10 Eur ebay) Jan 23 12:52:36 expensive ;D Jan 23 12:52:57 is 10yrs old by now :) Jan 23 12:53:13 is booting 3.8 :) Jan 23 12:53:29 jackmitchell: try with iotop Jan 23 12:53:29 but I've lost trace of thesing, the developer Jan 23 12:54:10 ant_work: probably moved to normal devices Jan 23 12:54:13 hrw: I have to send one Z in Poland soon Jan 23 12:54:42 maybe I've found a kernel developer interested Jan 23 12:54:54 o! Jan 23 12:55:17 I still dream about a kernel driver for reading proper bootyargs / mtdparts from nand Jan 23 12:55:28 he did one for bcm4x machines Jan 23 12:56:50 hrw: well, soon I'll have read enough to write one myself ;) Jan 23 12:57:02 hrw: any mips device there? Jan 23 12:57:31 ant_work: I only had wrt54gs when it comes to mips. and I donated it to OE few years ago Jan 23 12:58:09 I see Jan 23 12:58:34 reading teh ML, seems nobody is complaining about mips Jan 23 12:58:57 or there is just little testing... Jan 23 13:03:13 khem cares for mips Jan 23 13:06:34 I was thinking about a ben nanonote or something like that Jan 23 13:07:37 something that doesn't take too much space in the drawer Jan 23 13:15:21 afournier: that's how I came to the conclusion it wasn't doing anything! :D Jan 23 13:15:50 maybe tiny reads/writes of ~500kb every few seconds, but apart from that Jan 23 13:16:14 I'm upgrading to an SSD within the next week and I'm hoping that will improve the situation Jan 23 13:17:35 i checked this yesterday and it was spending a lot of time in tar / bzip, but i did not check for ext3 Jan 23 13:19:17 I sent a list of devices I can bring to fosdem and give to people. Check oe-devel ML for details. Jan 23 13:33:08 hm yes khem systemd patches dont work Jan 23 13:49:27 woglinde: for textrel I think you need to enable PIC Jan 23 13:50:16 bluelightning? Jan 23 13:50:31 ah its the other round khem patches don hit oe yet Jan 23 13:51:59 woglinde: to fix text relocation issues you enable to enable PIC when compiling (based on other fixes for such warnings) Jan 23 13:52:52 ah Jan 23 13:52:55 sorry Jan 23 13:53:12 its quite some time I asked it Jan 23 13:53:28 hm I dont know if libjvm works with pic Jan 23 13:54:35 hm am I stupid didnt send khem a bunch of systemd patches Jan 23 14:01:52 does anybody know why usbsutils is a DEPENDS of systemd? The systemd readme doesn't say anywhere that it needs it Jan 23 14:11:22 maybee it was earlier Jan 23 14:11:30 or maybe it was of ude Jan 23 14:11:31 v Jan 23 14:15:51 possibly udev should get the finger, I'm going to try building without now and see if I run into any issues, if not; patch incoming ;) Jan 23 14:26:44 fray, the spec file only has 718 lines .... Jan 23 14:26:54 do you mind if I mail it direct to you? Jan 23 14:53:46 what package is responsible for creating the /run directory for systemd? Jan 23 15:03:33 bother boost 1.52 has a nasty bug related to threads that annoys gnuradio Jan 23 15:19:57 jackmitchell: volatiles ? Jan 23 15:20:31 afournier: hmm, there has been some noise about the volatiles on the list recently too Jan 23 15:20:34 * jackmitchell goes off to read the thread Jan 23 15:20:56 i am trying to get rid of it Jan 23 15:22:01 hmm seems it is read-only rootfs related anyway Jan 23 15:22:28 and you don't have symbolic link from /var/run to /var/volatile/run ? Jan 23 15:24:07 but wait you are talking about /run not /var/run Jan 23 15:24:12 sorry Jan 23 15:24:18 I am, there is a symlink for /var/run Jan 23 15:29:06 then it's not related to volatiles i think Jan 23 16:01:01 bluelightning: hi again. tried a fresh angstrom tree, switching bitbake to 1.15.2 + the commit removing `git remote prune` .... and it fails too ... strangely, if I run the commands myself it works fine :| Jan 23 16:08:46 mnemoc: the commands you ran aren't the commands in the log though. Jan 23 16:09:06 git fetch -f --prune git://git.yoctoproject.org/prelink-cross.git refs/*:refs/* Jan 23 16:09:12 that was the one that failed Jan 23 16:09:19 bluelightning: it was from lib/bb/fetch2/git.py Jan 23 16:09:41 bluelightning: the 5th '$' Jan 23 16:10:45 oh right, sorry I misread Jan 23 16:10:50 I don't know :( Jan 23 16:11:01 can it be because of the output or the branch switching? Jan 23 17:56:21 does OE classic have a "do_populate_packages" stage? or is it called something else? Jan 23 18:12:38 hms what is the default passwd in oe-core images now? Jan 23 18:13:15 disabled? Jan 23 18:13:46 hms Jan 23 18:14:02 or it is systemd problem on uclibc Jan 23 18:14:05 you need to add zap-root-password IMAGE_FEATURE Jan 23 18:14:11 jama thanks Jan 23 18:14:12 or that Jan 23 18:14:22 so we do not have a default accout with passwd anymore Jan 23 18:14:25 not sure about spelling, git grep before using it Jan 23 18:14:35 I think it was always like that Jan 23 18:15:12 no? Jan 23 18:15:18 we could use root withouth passwd Jan 23 18:15:21 hm Jan 23 18:15:25 or only angstroem images Jan 23 18:15:37 with that IMAGE_FEATURE yes Jan 23 18:16:00 and when dropbear is setup correctly (see last changes from Paul) Jan 23 18:18:08 hm is it enough in local.conf Jan 23 18:18:25 or do I have to play whith this strange other variable Jan 23 18:23:54 latest master? Jan 23 18:24:04 if yes then local.conf is enough Jan 23 18:24:34 er.. I think you mean you need debug-tweaks to be in IMAGE_FEATURES to enable no-password login Jan 23 18:26:37 ah right, sorry Jan 23 18:26:46 that's why git grep is handy :) Jan 23 18:27:12 is there a standard way to set the root (or other user) passwords in oe? Jan 23 18:27:20 with latest master the dropbear evility is also gone :) Jan 23 18:27:29 I thought someone suggested one, but I lcan't find that it was ever accepted Jan 23 18:27:43 fray: I don't think there is... I've often wondered if it might be useful Jan 23 18:27:59 -yes- it would be.. I'd love to list passwords and accounts to be added.. Jan 23 18:28:23 xuser already does that but with pw hash in metadata Jan 23 18:28:23 list of users/passwords and making /home/xxx with skel files :-D Jan 23 18:29:17 ya.. add users and setup passwords.. Jan 23 18:29:46 could very quickly provision netbooks/OLPC type stuff then Jan 23 18:29:59 bbl Jan 23 18:30:02 I have no real user for this though! Jan 23 18:31:36 I've had a coupel of instances in the past week where it would have helped generate deployable filesystems Jan 23 18:32:17 would be good for "demo booth" type stuff Jan 23 18:32:28 all the demos could have consistant users and files Jan 23 18:33:35 ya.. that is pretty much what I was trying to do Jan 23 18:35:44 * XorA checks one of clones is not called fray Jan 23 18:36:09 :P Jan 23 18:38:23 scale booth, or "generic" booth? Jan 23 18:43:51 mr_science, unfortunately I do not believe with ELC just the week before that a booth at SCaLE is possible :( Jan 23 18:44:20 unless enough people will step forward. Jan 23 18:51:33 fray ping Jan 23 18:52:09 I'm here.. but distracted for a few minutes.. Jan 23 18:53:10 k Jan 23 18:53:26 when you get a chance, I'd like to talk about my rpm issue Jan 23 18:53:52 it was exposed by someone using "yocto" and I need to solve the problem Jan 23 18:54:15 not just an academic exercise, and apparently dropping the rpm bit might be an issue Jan 23 18:54:28 which issue? the '-' thing? Jan 23 18:56:56 yeah Jan 23 18:57:12 I just need to see the files and figure out where the translation failed.. Jan 23 18:57:19 basically, the learn reported is off the end of the spec file Jan 23 18:57:27 the system is supposed to translate dependency versions and releases that have a '-' in them to something else that is allowed.. Jan 23 18:57:31 let me know what to send Jan 23 18:57:39 if it doesn't then packages_rpm.bbclass may be broken Jan 23 18:57:46 if I have two python functions in my recipe, populate_packages_prepend_poky() and populate_packages_prepend() , which one is called if I build with yocto? the former or the latter? or both? Jan 23 18:58:03 do I need to run the entire build with rpm enabled Jan 23 18:58:10 in other words, is a specialized version preferred over the generic one? Jan 23 18:58:14 what is prepend_poky? Jan 23 18:58:48 on yocto, which uses oe-core, the _poky suffix makes sure this function is used only if poky is used as distro Jan 23 18:59:15 and _prepend prepends the contents of this function to populate_packages Jan 23 19:02:19 out of interest does anyone know the replacement sequence for under OE class bitbake -c compile -f ; bitbakc -f deploy Jan 23 19:02:34 would be nice to update my kernel workflow doc as people seem to be rading it Jan 23 19:08:02 ka6sox, what, not OE booth this year? Jan 23 19:08:31 i just signed up for sat/sun, so I can hang for a while Jan 23 19:08:50 XorA: bitbake -C compile; done. -C says invalidate the stamp for copmile, so rerun compile and all the other tasks up through do_build Jan 23 19:08:57 but i think i want to see the advanced git stuff on saturday morning Jan 23 19:09:01 at least, thats probably the intent of the sequence you were using.. Jan 23 19:09:07 <3 -C Jan 23 19:09:57 kergoth, are you going to scale this year? Jan 23 19:10:08 probably not Jan 23 19:11:00 i'll buy you a beer or three... Jan 23 19:17:14 kergoth: thanks I will give that a try and update the guide Jan 23 19:17:39 kergoth: intent is fiddle with source, rebuild and re-deploy Jan 23 19:17:46 * kergoth nods Jan 23 19:17:49 -C is ideal for that case Jan 23 19:17:53 rerun this task and all the takss after it Jan 23 19:18:10 kergoth: awesome, I can update guide tonight and make lots of people happy Jan 23 19:18:55 hehe Jan 23 19:19:50 there are so many lesser known features floating around, both bitbake and metadata Jan 23 19:19:57 should do a talk about them or something Jan 23 19:20:01 heh Jan 23 19:20:29 ELC-E would be a good place :-D Jan 23 19:20:47 hrms what was the config to get sysvinit instead of systemd Jan 23 19:21:24 IMAGE_INIT_MANAGER = Jan 23 19:21:26 I think Jan 23 19:21:31 setting IMAGE_INIT_MANAGER and DISTRO_FEATURES_INITMAN did not help Jan 23 19:21:35 still systemd Jan 23 19:21:38 we just had the same discussion in Angstrom :-) Jan 23 19:21:51 yes is angstrom Jan 23 19:21:58 agstrom master is broken anyway Jan 23 19:22:05 with cworth gone from core Jan 23 19:22:13 VIRTUAL-RUNTIME_somethingorother too Jan 23 19:22:18 stumbled over it yesterday Jan 23 19:22:35 but i'm not sure as to the sysvinit vs systemd status what with the merge in progress to oe-core Jan 23 19:22:38 heh Jan 23 19:22:51 yeah hrw fight this morning too Jan 23 19:26:16 ah there exists IMAGE_INIT_SCRIPTS Jan 23 19:47:44 hm okay looks udev provider must be udev Jan 23 19:52:14 hi Jan 23 19:53:54 hi zecke Jan 23 20:02:05 some idiot forgot to turn on the laptop power power brick switch Jan 23 20:18:41 hello Jan 23 21:11:16 gah, systemd build is failing due to trying to use python-dev bits from the host Jan 23 21:14:39 hu? Jan 23 21:15:16 * kergoth wonders why it wants python at all Jan 23 21:16:04 is not big enough? Jan 23 21:16:08 hehe Jan 23 21:16:14 insufficient bloat, quick, add more features Jan 23 21:16:44 kergoth: systemd from oe-core or systemd from meta-systemd? Jan 23 21:16:57 * kergoth investigates Jan 23 21:20:13 kergoth: meta-systemd has python support enabled, but should use right python Jan 23 21:22:13 * woglinde removed meta-systemd now Jan 23 21:22:40 too soon :) Jan 23 21:22:49 why? Jan 23 21:22:56 missing all the .bbappends Jan 23 21:23:06 no Jan 23 21:23:13 I switched back to sysvinit Jan 23 21:23:20 ah OK Jan 23 21:23:56 but even dbus depends on systemd bitbake rebuilds all stuff up to openjdk Jan 23 21:27:42 jama and I wondered why systemd was still put in the image then I regonized that it must be normal package dependencies Jan 23 23:32:34 JaMa: atleast we can delete meta-systemd/recipes-core Jan 23 23:32:50 JaMa: I am currently testing a systemd upgrade to 197 Jan 23 23:32:54 for oe-core Jan 23 23:33:58 I have been testing that combo extensively here Jan 23 23:46:19 ka6sox, did i miss the answer? Jan 23 23:48:12 mr_science: whats is Q Jan 23 23:49:06 * ka6sox wonders too... Jan 23 23:51:37 your comment about no booth or something... Jan 23 23:53:23 so do you mean no OE booth at scale without more booth victims, er, volunteers? Jan 23 23:58:36 does anyone had any problems trying to link a Qt project with opengles using a QT += opengl on OE? Jan 23 23:59:03 the problem here is that all the makefiles generated with this configuration links -lGL -lGLU also Jan 24 00:00:20 not often i'll be excited by seeing a square wave on teh scope Jan 24 00:04:33 thats bizzare Jan 24 00:04:53 in debian, the /sys/kernel/debug was not mounted Jan 24 00:05:16 so /sys/kernel/debug/omap_mux was not visisble Jan 24 00:34:18 :-) happy hobbit **** ENDING LOGGING AT Thu Jan 24 02:59:58 2013