**** BEGIN LOGGING AT Fri Jul 08 02:59:58 2011 Jul 08 03:03:22 this googling is not helping. Any input? Jul 08 03:03:29 kergoth, ^ Jul 08 03:07:11 it looks very much like this http://pastebin.com/ZbJi0Kjv Jul 08 03:21:41 apparently there was an typo when I ran your command before... it actually outputs: Jul 08 03:21:42 # BBPATH=/home/m/overo/a_bitbake/openembedded:/home/m/overo/a_bitbake/build Jul 08 03:21:43 export BBPATH="/home/m/overo/a_bitbake/openembedded:/home/m/overo/a_bitbake/build" Jul 08 03:21:51 kergoth, ^ Jul 08 03:25:20 which is in fact where everything should be Jul 08 03:27:54 looks like I figured it out... the DISTRO wasn't recognized Jul 08 03:28:06 not to figure out how to set that up correctly for my env Jul 08 03:30:04 still getting some errors though that are preventing things... it fails on parsing a few bbfiles, including preboot/initramfs-preboot-image.bb Jul 08 03:37:01 what errors ? Jul 08 03:53:15 http://paste.pocoo.org/show/432946/ Jul 08 03:59:52 khem, how easy is it to extract the toolchain for slugos-le? Jul 08 04:00:40 khem, ^ it also outputs a bunch of NOTE:'s with ParameterErrors Jul 08 05:00:46 ka6sox: easy enough. MACHINE=nslu2le bitbake meta-toolchain-slugos Jul 08 05:01:01 khem, thanks Jul 08 05:01:12 we need it for Optware. Jul 08 05:01:37 ok Jul 08 05:01:52 use 2011.03 for best results Jul 08 05:02:09 if daring you can also use new oe-core setup Jul 08 05:03:03 foobraz: paste the complete output Jul 08 05:05:07 2011.3 is fine Jul 08 06:00:52 hello Jul 08 06:03:09 hi Jul 08 06:08:17 where in oe does libudev get built? I see a dependency inside udev.bb - i also need libudev for my project, is it sufficient to add the dependency in my recipe? Jul 08 06:41:55 good morning Jul 08 09:30:36 morning all Jul 08 09:31:44 hi bluelightning Jul 08 09:36:11 hi bluelightning, woglinde Jul 08 09:36:32 hi mckoan Jul 08 10:21:56 jo obi Jul 08 10:22:07 hi woglinde Jul 08 10:24:34 stefan_schmidt: time for a short phonecall? Jul 08 10:26:51 03Koen Kooi  07org.openembedded.dev * r0d424b1eb8 10openembedded.git/recipes/mozilla/firefox_3.6.8.bb: Jul 08 10:26:51 firefox 3.6.8: bump PR to force rebuild against updated libs Jul 08 10:26:51 this fixes the 'missing text everywhere' problem Jul 08 10:26:51 Signed-off-by: Koen Kooi Jul 08 10:52:34 mickey|office: hmm, difficult right now. Jul 08 10:52:50 mickey|office: Possible to do it later today? Around 21:30? Jul 08 10:53:09 mickey|office: if not let me see what I can do here. (At the university lab right now) Jul 08 10:59:00 stefan_schmidt: no problem, lets do it then. i'll set a reminder for me Jul 08 11:01:00 mickey|office: cool Jul 08 11:01:25 * stefan_schmidt hopes he will be fine after wakeboarding and handball at the afternoon and evening Jul 08 11:01:40 Doing a bit more sports again these days Jul 08 11:04:19 morning Jul 08 11:04:24 mickey|office: how is the family stress level? :) Jul 08 11:04:26 hi XorA Jul 08 11:04:31 hey stefan_schmidt Jul 08 11:04:37 long time no see Jul 08 11:07:53 stefan_schmidt: sometimes more, sometimes less. "we"'re still plagued by flatuences, but they might be gone after just 8 more weeks :) Jul 08 11:08:26 other than that, the little one grows fine. Sabine already works again, with a 8h per week job Jul 08 11:10:40 XorA: yup Jul 08 11:10:48 mickey|office: hope that will get better Jul 08 11:10:57 mickey|office: Sabine is working from home? Jul 08 11:11:14 mickey|office: Or are you the full daddy that day? :) Jul 08 11:14:15 stefan_schmidt: ya, she's working from home. we have so much work that i can't take any time off atm. Jul 08 11:15:02 stefan_schmidt: hows life? Jul 08 11:16:19 mickey|office: :/, hope you have still some time for the family life Jul 08 11:16:43 XorA: quite good. A bit stressful due to diploma thesis and freelance work overlap, but okay Jul 08 11:16:53 Will be better from november Jul 08 11:17:00 stefan_schmidt: good to hear Jul 08 11:35:51 Hi. I'm using 2.6.34 with angstrom 2010. i'm trying to figure out why udev doesn't create /dev/mmcblk0p*, if i create it manually, it works fine. any suggestions for debug? Jul 08 11:36:41 the mmcblk0p* is for SD card which is plugged in at bootup, rootfs is on it and boots fine, just no /dev entries for any of the partitions Jul 08 11:53:28 mount.blacklist ? Jul 08 12:04:59 nope, no mmc in mount.blacklist, interesting. if i do udevadm test /sys/block/mmcblk0 , it creates mmcblk0 Jul 08 12:06:38 udevd is running, when i did udevadm test, seems like it is parsing rules correctly, udev_rules_apply_to_event: IMPORT 'path_id /block/mmcblk0' /etc/udev/rules.d/60-persistent-storage.rules:47 Jul 08 12:36:45 Is there an easy way to over ride a PREFFERED package version for the OE Angstrom distro? I tried putting it in build/conf/local.conf but it did not work. I know I can modify angstrom inc but I wanted to keep my oe folder clean and only make changes in my overlay folder Jul 08 12:37:32 I know I can just copy Angstrom.files over to my overlay folder but that seems a bit over kill if I only need to change one PREFERRED version. Jul 08 12:38:22 the package is "PREFERRED_VERSION_tslib" Jul 08 12:46:17 mickeyl what is sabine working on from home? Jul 08 12:46:27 svolpe: IIRC you can do something like PREFERRED_VERSION_pn-tslib = "blah" Jul 08 12:46:27 mickeyl giving e-learning session? Jul 08 12:46:35 svolpe: it's a bit hacky though Jul 08 12:57:28 svolpe: FWIW, there's a big discussion on the issue here: http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg08916.html Jul 08 13:13:38 bluelightning: Thanks, I will read the discussion. Jul 08 14:12:05 morning kergoth__ Jul 08 14:13:49 gm pb Jul 08 14:13:55 hi woglinde Jul 08 14:15:03 greetings. Jul 08 14:17:05 hi pb_, mwester-laptop Jul 08 14:18:07 hi mwester-lpatop, bluelightning, all Jul 08 14:18:38 * mwester-laptop wonders if a reference to the TV show "The Waltons" would make sense to anyone... :D Jul 08 14:33:04 mwester-laptop: goodnight john boy Jul 08 14:36:33 sigh, eglibc packaging busted again Jul 08 14:37:00 oe-classic? It built ok for me a couple of days ago. Jul 08 14:37:10 no, oe-core Jul 08 14:37:24 some kind of localedef related lossage Jul 08 14:37:27 Ah. I can't keep up with oe-core... Jul 08 14:39:04 I see patch sets with dozens of patches go by -- it has to be one's full-time job if one wishes to review and test all of those changes, I expect. Jul 08 14:41:53 hm, now my emacses have all seized up Jul 08 14:42:01 * pb_ throws his computer out of the window again Jul 08 14:42:13 uhm Jul 08 14:42:18 okay Jul 08 14:42:21 I am going Jul 08 14:47:50 pb_: wait I need more RAM before you through it out of window :) Jul 08 14:48:01 good morning all Jul 08 14:48:01 heh Jul 08 14:48:11 mwester-laptop: I used to think the same about oe-dev, FWIW :) Jul 08 14:48:31 yeah, at least with oe-core you can actually see (most of) the patches before they get committed Jul 08 14:48:36 there is some sanity to oe-core patches although sometimes we bust Jul 08 14:49:17 pb_: w.r.t. systemd I think fcntl approach seems good. I will try it out Jul 08 14:49:31 ok, very good Jul 08 14:49:35 but so far with my limited testing I have not seen any problems Jul 08 14:50:19 sure, but the kind of problems that result from losing things like O_CLOEXEC are not necessarily going to be apparent in casual testing Jul 08 14:50:31 yeah thats what I thought Jul 08 14:51:33 that's why I keep saying that it isn't good enough to just go "oh, it seems to work, let's ship it". if you're making a change like that then you need to be able to mount a rational argument for why it is a safe thing to do. Jul 08 14:52:02 but thats the corporate mantra isnt it :) Jul 08 14:52:10 after all, it's possible that "limited tsting" might not even be exercising that line of code at all Jul 08 14:53:10 I am trying to make a llvm layer Jul 08 14:53:23 so folks can have an alternative to gcc Jul 08 14:53:39 starting with llvm-gcc Jul 08 14:54:04 atleast on arm/x86 llvm is mature enough to be usable Jul 08 14:55:17 for GPLv3'less world Jul 08 14:55:28 that sounds like a cool project Jul 08 14:56:01 yes at least buildin the compiler is so much better than gcc's funkiness Jul 08 15:10:49 ah, there we go, eglibc working again now Jul 08 15:50:39 re Jul 08 15:50:46 pb what was it? Jul 08 15:53:46 RP, RP__: I'm wondering how long you intended to wait to actually give feedback about the removal of the env bits in the devshell patch. That's, what, the third resend without even telling me there was an issue with it? If there's something wrong with something I try to get merged, I'd really appreciate getting actual feedback so I can fix the problem, in the future. Jul 08 15:54:47 kergoth__: I don't know how it slipped off my radar, sorry :( Jul 08 15:55:21 woglinde: missing "locale-code" in DISTRO_FEATURES turns out to make eglibc unbuildable Jul 08 15:55:24 not a big deal, it happens, there's a lot of email flying around Jul 08 15:57:55 args Jul 08 15:58:05 why get saul's stuff with rdepends commited? Jul 08 15:58:13 woglinde: ? Jul 08 15:58:21 there was enough opinios to fix the bashism Jul 08 15:58:30 woglinde: we can still fix them Jul 08 15:58:36 woglinde: someone just send the patch Jul 08 15:58:47 woglinde: I also did not take them all... Jul 08 15:59:35 rp noboy at intel has the time to look up in debian if there are patches to fix the bashism? Jul 08 15:59:39 ups nobody Jul 08 15:59:59 I'm not quite sure why that needs to be intel's problem. Jul 08 16:00:13 hm okay Jul 08 16:00:17 as things stand, the packages do use bash and the rdepends should reflect that. Jul 08 16:00:42 I don't think you can really argue that adding the RDEPENDS is making things worse. Jul 08 16:01:53 hm it drags in 3 mb of bash Jul 08 16:02:01 or so Jul 08 16:02:11 well, sure, but if the packages didn't previously work without that 3MB then that's no better. Jul 08 16:02:14 The fixes are correct in their own right, adding the dependency. I'm more than happy to take a patch from anyone which removes that though Jul 08 16:02:56 I don't think it reasonable to demand that saul does that in order to fix what is a clear bug though Jul 08 16:04:09 yeah, agreed Jul 08 16:04:48 hm maybee I should have more complained about which stuff out of the packages needs bash in the commit-msg'es Jul 08 16:06:42 ugh what in usbutils needs bash Jul 08 16:08:05 koen sent some mail about that, I think Jul 08 16:12:18 rather than spending your time complaining, maybe a more constructive thing to would be to write a QA test to detect any more as-yet-undiscovered bashisms that might be lurking. :-} Jul 08 16:15:51 I did wonder if someone couldn't have written patches to remove some of these in the time the discussions on the mailing list took :) Jul 08 16:37:24 03Joshua Lock  07master * r737d1bc819 10bitbake.git/lib/bb/ui/hob.py: (log message trimmed) Jul 08 16:37:25 lib/bb/hob: fix changing base image Jul 08 16:37:25 The path is not guaranteed to always point to the same value so do not rely Jul 08 16:37:25 on it to change the image contents. Further, when changing the base image Jul 08 16:37:25 we should maintain user selections. Jul 08 16:37:25 Addresses [YOCTO #1225] and fixes [YOCTO #1226] Jul 08 16:37:25 Signed-off-by: Joshua Lock Jul 08 16:37:34 03Richard Purdie  07master * r3725602ec5 10bitbake.git/lib/bb/fetch2/git.py: (log message trimmed) Jul 08 16:37:34 fetch2/git: Tweak git fetcher to handling repo updates correctly Jul 08 16:37:34 Currently the git fetcher can malfunction when branches change in remote Jul 08 16:37:34 repositories since whilst the update code updates the "origin" remote, Jul 08 16:37:34 this isn't linked to the local heads. Jul 08 16:37:36 By passing the --mirror option to 'git clone' and 'git remote add', Jul 08 16:37:36 linkage between the local heads and remote heads is created with a 1:1 Jul 08 16:37:36 03Joshua Lock  07master * rf3fbc97471 10bitbake.git/lib/bb/ui/crumbs/tasklistmodel.py: (log message trimmed) Jul 08 16:37:36 ui/crumbs/tasklistmodel: fix reset method Jul 08 16:37:37 The reset() method only touched the contents sub-model, which does not Jul 08 16:37:37 include the selected image(s). This patch ensures that reset correctly unsets Jul 08 16:37:38 any image selection when called. Jul 08 16:37:38 Further we re-initialise the COL_IMG column when resetting packages. Jul 08 16:37:39 Signed-off-by: Joshua Lock Jul 08 16:38:34 regardless. Jul 08 16:38:35 Fixes [YOCTO #1220] Jul 08 16:38:35 Signed-off-by: Joshua Lock Jul 08 16:38:35 Signed-off-by: Richard Purdie Jul 08 16:43:26 03Lianhao Lu  07master * r521909d135 10bitbake.git/lib/bb/ (5 files in 2 dirs): Jul 08 16:43:27 bitbake cooker/ui: handle cmd line parsing result by individual UI. Jul 08 16:43:27 Changed the return result of "getCmdLineAction" to a dictionary Jul 08 16:43:27 {'action', 'msg'} to allow the individual UI decide how to handle the Jul 08 16:43:27 cmd line parsing result. Jul 08 16:43:27 Signed-off-by: Lianhao Lu Jul 08 16:43:27 Signed-off-by: Richard Purdie Jul 08 16:47:10 hm, spoke too soon, eglibc still not quite working right. Jul 08 16:47:22 I seem to get a different failure every time I try to build it which is rather irritating. Jul 08 17:24:54 try luck here... Jul 08 17:25:18 try luck here Jul 08 17:25:39 i have dhcp3 and dhcp4 recipes under the same dir, the default is 4, i need bitbake dhcp3, how to choose that? Jul 08 17:37:00 xxia pin Jul 08 17:37:33 pong Jul 08 17:37:41 still could not figure out how to bitbake dhcp3 Jul 08 17:37:54 ? Jul 08 17:37:58 grep failed to find out where it's set to default Jul 08 17:38:02 xxiao no I meant pin the revision Jul 08 17:38:16 bitbake dhcp3? Jul 08 17:38:18 let me try Jul 08 17:38:25 or do the badboy variant and remove the dhcp4 recipe Jul 08 17:38:35 anyway why you dont udhcp? Jul 08 17:39:47 using busybox udhdpc and it had troubles with dhcp server here Jul 08 17:39:57 is udhcp the same or something totally different Jul 08 17:40:08 i can try udhcp Jul 08 17:42:03 yeah busybox thinggie Jul 08 17:42:15 bitbake dhcp3, dhcp_3, dhcp_3.0.2, none baked Jul 08 17:42:25 how to pin the revision ? Jul 08 17:44:21 look into config dir Jul 08 17:48:33 i assume this udhcp is irrelevant to busybox's udhcp Jul 08 17:48:48 it is the same Jul 08 17:48:55 *sigh* Jul 08 17:49:37 what?, ok Jul 08 17:50:08 dhcp_3.0.1.bb dhcp_3.0.2.bb dhcp_3.1.2p1.bb dhcp3.inc dhcp-4.1.1-P1 dhcp_4.1.1-P1.bb dhcp4.inc files Jul 08 17:50:17 that's everything under recipes/dhcp Jul 08 17:50:24 i could not find a hint from conf/ Jul 08 17:50:31 how to pin to, say, 3.0.2? Jul 08 17:53:02 ok i switched to arm 2010q1 and dhcp4 built Jul 08 17:53:15 somehow it dislikes 2009q1 gcc Jul 08 20:20:22 uhm the ldd bashism fix is just replace bash with /bin/sh Jul 08 20:20:32 why they dont fix it upstream Jul 08 20:30:38 woglinde: that's not a correct fix Jul 08 20:30:54 the script really does use bash-specific constructs, eg $"string" Jul 08 20:31:30 possibly others too, I dunno. someone would need to audit it. Jul 08 20:41:08 pb okay so you saw the patch too Jul 08 20:42:17 no, I was just going on what you said above Jul 08 22:08:42 heh, lamawithonel's connection seems to be abnormally stable today, only about one timeout per hour Jul 08 22:08:53 let's hope this happy state of affairs persists Jul 08 22:09:02 lol Jul 08 22:10:48 pb_: I'm bothering khem to death à propos klibc. Now it's your turn...just general questions, don't worry :) Jul 08 22:11:45 first: I'm a bit dubious about the nature/naming of klibc. We install on buildhost to be able to crosscompile binaries for target. Jul 08 22:12:25 but the lib itself is packaged to be installed on target... Jul 08 22:14:30 now, where would you expect the headers in sysroots? Jul 08 22:15:42 corollary is klcc-cross Jul 08 22:54:07 well, the location in sysroot ought to exactly mirror the location on the target Jul 08 22:54:32 that's the same as we do with eglibc and uclibc Jul 08 22:56:42 so, I guess that'd be sysroots/TARGET/lib/... Jul 08 23:00:11 ah, I almost got it then Jul 08 23:00:12 http://paste.debian.net/122373/ Jul 08 23:02:03 pb_: and klcc-cross can be BASE_PACKAGE_ARCH while klibc is MACHINE_ARCH ? Jul 08 23:02:37 (mimic of gcc-cross) Jul 08 23:06:54 anyway, there is something wrong there because if I package it... Jul 08 23:07:09 I end up always with smthg like http://paste.debian.net/122165/ Jul 09 00:14:52 is there some way that .bbappend files are appended to ${PN}.bb instead of to the literal filename of a recipe (e.g. to emulate amend.inc)? The default behaviour of .bbclass seems to be very naiv and causes lot of maintainance effort :( Jul 09 00:27:45 ensc, you can still use amend.inc iirc Jul 09 00:28:00 Or maybe that bbclass needs moving over to oe-core? Jul 09 00:28:07 no; does not exist anymore Jul 09 00:28:20 and needs some infrastructure which is not available in oe-core Jul 09 00:31:53 bah Jul 09 00:32:01 * Tartarus adds to his Monday pile Jul 09 00:38:58 mmh... 'bitbake -b ... -e' is broken :( Jul 09 00:47:59 sigh Jul 09 00:48:01 bug the ml? **** ENDING LOGGING AT Sat Jul 09 02:59:57 2011