**** BEGIN LOGGING AT Sat Jan 07 02:59:57 2012 Jan 07 05:56:57 hello Jan 07 05:58:48 I m building x11-image with uClibc ..... I m getting at last stage of building rootfs...... I m getting http://pastebin.mozilla.org/1437822 Jan 07 05:58:55 can anyone help me' Jan 07 08:45:47 gm Jan 07 11:08:56 :-) Jan 07 12:31:30 Does Oe support dietlibc klibc ? Jan 07 12:34:52 there's klibc support, I think dietlibc is unmaintained Jan 07 12:37:35 bluelightning, hav u tried it ? Jan 07 12:38:05 erwt: I've built/used kexecboot, that makes use of klibc Jan 07 12:38:12 other than that, no Jan 07 12:38:54 erwt: you could ask ant, he's not around now though Jan 07 12:39:09 bluelightning, thanks Jan 07 12:39:47 bluelightning, i have build a basic x11-image using uclibc today , Jan 07 12:40:25 but the size of usr/lib is itself 27 Mb stripped Jan 07 12:42:05 bluelightning, i thought using uclibc it will lessen the size ,but i think is fruitless Jan 07 12:43:49 Is there any technique to reduce the file size completely Jan 07 12:49:31 depends how much u need Jan 07 12:50:36 and if the file can be in an archive Jan 07 12:50:46 (missed most of the discussion though) Jan 07 12:50:58 cs_nbp, how Jan 07 15:13:06 erwt: uclibc will save you 1-1.5 MB, other way to reduce is throw packages out of your image (or build something small like minimal-console-image); and if you are flash constrained use e.g. squashfs Jan 07 15:16:37 eFfeM, thanks a lot i am trying console-base-image Jan 07 15:17:46 erwt: the image for my product is about 2.5 MB, but it contains not too much: busybox, mtd-utils, i2c-tools, syslog and a few other small things and of course the product app Jan 07 15:18:30 eFfeM, i want your package list plz ......................... Jan 07 15:18:59 erwt: give me a minute, it is on another system and I need to finish something else first Jan 07 15:22:02 eFfeM, ya definitely Jan 07 15:23:49 erwt: here is the list of non proprietary packages in my image: http://pastebin.com/twYL8JRX Jan 07 15:24:56 you might even leave out some things, guess the first 4 and maybe dropbear are close to the bare minimum Jan 07 15:25:17 btw and 2.5 MB was measured on power pc, if you have another architecture you might get different sizes. Jan 07 15:25:30 eFfeM, does it have X Jan 07 15:25:37 no Jan 07 15:26:27 X will add quite a lot wrt libs (but if you want to have it display on a remote system you do not need x server, only client libs afaik) Jan 07 15:26:51 eFfeM, did u create a .bb file for your self Jan 07 15:27:41 yes, I always create my own image recipes, use one of the recipes like console-base-image as a starter and add whatever you need to the list of pacakges Jan 07 15:27:54 i am doing that Jan 07 15:28:10 to IMAGE_INSTALL Jan 07 15:28:26 but the file size of console-base-image is 27.3 Mb Jan 07 15:28:31 basic Jan 07 15:28:35 ouch Jan 07 15:28:42 thats my cause Jan 07 15:30:09 i would expect it to be smaller, but there is an option in local.conf that you can set to keep the root fs (forgot the exact name) then after rebuidling the image you can see in deploy/rootfs or so what is installed Jan 07 15:30:15 eFfeM, i used angstrom distribution Jan 07 15:30:28 for what hw? Jan 07 15:30:40 mini2440 Jan 07 15:30:47 it'll probably drag in all kinds of crap that you do not want/need Jan 07 15:31:26 e.g. I have no audio stuff, no usb stuff stuff, no wifi etc etc Jan 07 15:32:38 smallest image is probably micro-base-image.bb, not sure if that builds for angstrom though Jan 07 15:33:04 but if you want full control that could be a starting point (and only add whatever you need) Jan 07 15:33:23 btw angstrom might also give you quite a lot of kernel modules, not sure if you want those Jan 07 15:33:44 I use a hand configured kernel, no modules, only configure what I need Jan 07 15:33:48 i dont want them at all Jan 07 15:34:28 but then again I do not need to deal with people coming with webcam X or wifi dongle Y and expect it to work Jan 07 15:34:37 angstrom might well add lots of those. Jan 07 15:34:58 but for angstrom specifics, peek at your rootfs dir and/or maybe ask in #angstrom Jan 07 15:38:55 eFfeM, i will tell u a list of packages monday .............with which i want to build , Jan 07 15:39:13 just guide me regarding that Jan 07 15:42:17 ericben: hi, btw is all OK on your side with the recent qt patches? Jan 07 15:43:36 eFfeM, bluelightning cs_nbp ,good bye good nite Jan 07 15:43:48 erwt: goodnight Jan 07 15:43:48 nite erwt Jan 07 16:01:12 Is there a way to make bitbake to fetch a file but save it under another name at DL_DIR? Jan 07 16:02:59 otavio: not sure if you do it when fetching but you can always append to do_fetch or prepend to do_configure or so and rename the file Jan 07 16:03:35 eFfeM: the problem is that the file, at the remote site, is not versioned Jan 07 16:03:51 eFfeM: so I'd like to put the version prepended on the file name to avoid clashes Jan 07 16:04:26 otavio: that brings the risk that the remote file may change at some point in time Jan 07 16:04:42 guess I would be inclined to put it with a versioned name on a mirror Jan 07 16:05:02 eFfeM: the version is at file remote path Jan 07 16:05:16 eFfeM: it is the case of firefox l10n modules for example Jan 07 16:05:17 ah ok Jan 07 16:05:30 I think this can be done not sure though Jan 07 16:05:32 how Jan 07 16:44:05 Hmm, anyone know if cbrake has his script for sending out the weekly commit info somewhere public? Jan 07 16:49:54 kergoth: Can you read the backlog and look at what eFfeM and I were talking some minutes ago? Jan 07 16:50:03 kergoth: do you know if this is possible? Jan 07 16:50:29 I don't know of a way to specify a destination filename for a fetch. Jan 07 17:37:54 where does a layer.conf go? next to the local.conf or somewhere else? Jan 07 18:29:59 hello, i have a question regarding LICENSE files, how do i create a recipe which does not provide a license file ? Jan 07 18:40:00 nschle85|2: the first question is why is there no license in the upstream source, if the recipe builds something that is? Jan 07 18:42:38 bluelightning: i wanted to copy termcap from old oe dev into new dev branch and build,it recipes has configured gplv2 but the license is not contained in the upstrieam sources Jan 07 18:44:21 nschle85|2: I think the recommended method now is to take a copy of the appropriate license file from the meta/files/common-licenses/ subdir Jan 07 18:44:45 into the recipe's files/ subdir and then refer to that in LIC_FILES_CHKSUM Jan 07 18:45:12 bluelightning: ok Jan 07 18:48:37 bluelightning: and how do i disable the qa check for licencse if i wat quickly build a recipe without clear license ? Jan 07 18:48:50 s/wat/want/ Jan 07 18:49:38 nschle85|2: there is no way without modifying insane.bbclass Jan 07 18:49:53 it's not something that's good to encourage Jan 07 18:49:55 (IMHO) Jan 07 18:51:15 bluelightning: useing the wrong license instead of no license is also bad :-) Jan 07 18:51:51 bluelightning: but anyway thank you :-) Jan 07 18:52:22 nschle85|2: it's not the wrong license, you're just getting no protection from the checksum Jan 07 18:52:29 but that's the fault of the upstream author Jan 07 18:52:57 bluelightning: i have another question Jan 07 18:53:05 doubly negligent considering AFAICT termcap is a gnu project? Jan 07 18:53:17 sure Jan 07 18:54:35 whet i say: bitbake termcap, the termap 1.3.1 is started to build and not 2.0.8, this is the directory i copied: http://cgit.openembedded.org/openembedded/tree/recipes/termcap Jan 07 18:54:47 s/whet/when/ Jan 07 18:55:41 why is not 2.0.8 choosen ? Jan 07 18:57:13 bluelightning: but my intention is not to have termcap as recipe , my intention is to solve following problem: http://norman-schleicher.de/jenkins/job/shr-core-nokia900-shr-image/36/console Jan 07 19:22:38 nschle85|2: er, we do have ncurses... Jan 07 19:24:23 does someone have a distro recipe for 3.1 kernel? I got the kernel 3.1 recipe but I cant even make a config right now-.- Jan 07 19:24:26 bluelightning: what does it mean ? Jan 07 19:25:29 nschle85|2: since this is a native recipe, it looks like it's looking for ncurses-devel on your system and not finding it Jan 07 19:26:19 bluelightning: how did you find it out ? ? Jan 07 19:26:54 nschle85|2: by reading the error message Jan 07 19:27:14 nschle85|2: it could be that our until-linux recipe needs to disable ncurses support for -native Jan 07 19:30:11 bluelightning: as i can see readline needs termcap, where is logical connection to ncurses and util-linux ? Jan 07 19:30:21 nschle85|2: further down Jan 07 19:30:34 there are multiple failures in this log :( Jan 07 19:31:07 for the util-linux one, add this to meta/recipes-core/util-linux/util-linux.inc: Jan 07 19:31:13 EXTRA_OECONF_append_virtclass-native = " --without-ncurses" Jan 07 19:31:27 (note the leading space) Jan 07 19:34:42 hmm this doesn't look right at all... the readline recipe already depends on ncurses Jan 07 19:34:59 I don't understand how you could be getting these failures Jan 07 19:35:04 bluelightning: i installed ncurses-dev on buildhost and all is fine now, hmm but what is the strategy ? fix the errors or depend on inmstalled ncurses-dev ? Jan 07 19:35:18 no, fix the errors if you can reproduce them Jan 07 19:35:40 we do provide ncurses-native Jan 07 19:36:25 ah ok i should remove dev dev package and rebuild from scratch Jan 07 19:36:27 it could be that it's not often tested on systems without ncurses-dev installed, but that doesn't mean it shouldn't work without it Jan 07 19:37:03 nschle85|2: shouldn't need to rebuild from scratch, try removing ncurses-dev then bitbake -c cleansstate ncurses-native readline-native Jan 07 19:37:15 then try again, see what happens Jan 07 19:37:36 of course that still leaves the python-cython-native error Jan 07 19:37:39 bluelightning: but broken packages are still build Jan 07 19:37:52 nschle85|2: what do you mean? Jan 07 19:38:58 readline native is still build Jan 07 19:40:00 that's why I said run "bitbake -c cleansstate ncurses-native readline-native" Jan 07 19:40:16 bluelightning: ups, you are right Jan 07 19:45:48 bluelightning: ok ncurses-dev removed, bitbale -c cleansstate .... done Jan 07 19:45:58 now starting build again Jan 07 19:46:12 nschle85|2: did you do the util-linux fix as well? Jan 07 19:46:38 bluelightning: no, i wanted to reproduce the problem Jan 07 19:46:46 well that's a separate issue Jan 07 19:47:15 ok then it needs to reproduced too :-) Jan 07 19:47:29 http://norman-schleicher.de/jenkins/job/shr-core-nokia900-shr-image/38/console Jan 07 19:47:31 well it is definitely an issue, the recipe shouldn't be explicitly excluding ncurses from the deps and not also configuring it out Jan 07 19:48:30 interesting, have never used jenkins before... cool how you can see live output :) Jan 07 19:48:49 ah, I see you are using rm_work Jan 07 19:50:22 bluelightning: you cannot see all live things.. the steps before builkd i had to make on the unix console, so you cannot follow all id do Jan 07 19:50:28 all id do Jan 07 19:50:31 all i do Jan 07 19:51:48 bluelightning: we have a problem, the build does not stop :-) Jan 07 19:52:53 nschle85|2: yes, because the -k option has been specified Jan 07 19:53:10 yes, ill reconfigure it Jan 07 19:54:59 http://norman-schleicher.de/jenkins/job/shr-core-nokia900-shr-image/39/console Jan 07 19:56:20 ok build has stopped Jan 07 19:56:39 hmm, still can't find the header... odd Jan 07 19:56:58 will run a build here Jan 07 19:57:21 bluelightning: i did not change any recipe Jan 07 19:57:27 * bluelightning notes he had libncurses5-dev installed previously Jan 07 19:57:31 nschle85|2: ok Jan 07 19:58:06 but i wanted to reproduce the readline-native build problem now Jan 07 20:00:06 si ll do bitbake readline-native in unix console now ok ? Jan 07 20:00:21 nschle85|2: yep that should be fine Jan 07 20:00:47 now that I have a bblayers.conf in my build/conf and a layers.conf in my overlay/conf, why does it still refuse to find the overlayed recipes? Jan 07 20:01:09 the layers.conf was shipped Jan 07 20:01:27 cs_nbp: the file in your layer/conf should be called "layer.conf" not "layers.conf" Jan 07 20:01:31 the bblayers.conf is the one you find everywhere on the net Jan 07 20:02:10 it actually is i typoed here Jan 07 20:02:38 structure is in reality overlay/meta-machine/conf/layer.conf Jan 07 20:02:50 cs_nbp: ok, that's correct Jan 07 20:03:00 cs_nbp: bitbake-layers gives you a few tools to view layer information as bitbake sees it Jan 07 20:03:16 ill give that a run Jan 07 20:03:16 does bitbake-layers show_layers list your layer? Jan 07 20:03:27 bluelightning: util-linux stops before building readline-native http://pastebin.com/ZSriLBKm Jan 07 20:04:03 nschle85|2: ok, so bitbake -c cleansstate util-linux-native and then apply the fix I mentioned earlier Jan 07 20:04:52 a sry i ran bitbake layers without - Jan 07 20:04:57 just a second Jan 07 20:05:30 bitbake-layers: command not found Jan 07 20:05:37 bluelightning: ok ill do Jan 07 20:05:48 cs_nbp: oh, which version of bitbake are you using? Jan 07 20:05:57 1.10.2 Jan 07 20:06:07 cs_nbp: oh right... does that version even support layers? Jan 07 20:06:12 I think it might be too old Jan 07 20:06:22 i could change to 1.12 in a sec Jan 07 20:06:34 cs_nbp: I think if you want to make use of layers you may have to Jan 07 20:08:21 well it still doesnt know about bitbake-layers Jan 07 20:08:41 should i get the git one? Jan 07 20:09:32 bluelightning: is the any code convention where to put --disable-** statements ? Jan 07 20:09:55 nschle85|2: well, in this case it has to be just for native Jan 07 20:10:07 EXTRA_OECONF_append_virtclass-native = " --without-ncurses" Jan 07 20:10:18 that's more or less how it would have to be done in this case Jan 07 20:11:54 bluelightning: now i have to reconfigure jenkins, not to overwrite my changes :-) Jan 07 20:12:00 i got the one from berlios linked in the 'getting started' section Jan 07 20:12:24 cs_nbp: hmm, that documentation is probably out of date... Jan 07 20:12:43 so ill git the git Jan 07 20:13:05 cs_nbp: I presume you are still using OE-Classic ? if so you'll need to stick to 1.12.x Jan 07 20:13:14 bluelightning: http://norman-schleicher.de/jenkins/job/shr-core-nokia900-shr-image/40/console Jan 07 20:13:21 (which you can always get from the appropriate git tag of course) Jan 07 20:13:51 err classic? i just clone the 'openembedded' from github Jan 07 20:13:57 dunnoh if thats classic Jan 07 20:14:44 cs_nbp: openembedded from github? do you have the link? Jan 07 20:14:44 perhaps thats why i got problems building for kernel 3.1, too? Jan 07 20:15:11 git clone git://github.com/openembedded/openembedded.git Jan 07 20:15:22 oh right, that's a clone of OE-Classic yes Jan 07 20:15:40 which should still be fine but you need to stick to bitbake 1.12.x Jan 07 20:15:47 do I need to move away to build for 3.1? Jan 07 20:15:55 ah ok Jan 07 20:17:05 cs_nbp: there's no linux 3.1 recipe in oe-classic no... if you're starting something "new" then I'd recommend making the move to OE-Core now Jan 07 20:17:56 well i havent started it, but Im testing a project thats fairly new and thus only exists as oe-overlay Jan 07 20:18:30 tworaz netbookpro kernel3.1 + bootloader embed code Jan 07 20:20:24 cs_nbp: oh neat, didn't know about those layers Jan 07 20:20:39 cs_nbp: looking at them I'd guess they are meant to be used with OE-Core already Jan 07 20:20:45 even a sort of jlime Jan 07 20:20:50 yes indeed Jan 07 20:21:51 maybe I missed it but I did not hear about anyone working on these Jan 07 20:22:00 jlime is based on classic OE Jan 07 20:23:01 no its easy to miss theres just one thread on jlime http://jlime.com/forum/viewtopic.php?f=6&t=2775&start=150 Jan 07 20:23:23 and the stuff on github google finds when you put nBkProOs.img ^^ Jan 07 20:24:23 Jay7: this could be an unofficial layer, it's hard to tell Jan 07 20:24:55 I'm on #jlime and never seen there any talks about migration to OE-core :) Jan 07 20:25:04 I would be reluctant to list the jlime layer until it's clear that it's official Jan 07 20:25:16 even more - I was first one who said about OE-core there :) Jan 07 20:25:35 so seems this is unofficial layer really Jan 07 20:25:58 Jay7: kristoffer is at least aware of OE-Core, I guess he lacks time Jan 07 20:26:14 iirc, jlime was fork of angstrom Jan 07 20:26:20 i'm only getting bitbake-1.5 it seems, and am relatively new to git (used svn so far) Jan 07 20:26:36 so.. at least one may fork angstrom again :) Jan 07 20:26:39 is there a way to check out a specific tree? google gives very different responses Jan 07 20:27:01 cs_nbp: if you're using OE-Core (and it looks like you should be for this) then you can use git master of bitbake Jan 07 20:29:27 you mean the openembedded/bitbake one? Jan 07 20:32:07 cs_nbp: I mean the very latest master branch of bitbake Jan 07 20:32:23 cs_nbp: for OE-Core see http://www.openembedded.org/wiki/OpenEmbedded-Core Jan 07 20:32:47 nschle85|2: btw readline-native succeeded here... Jan 07 20:33:12 ah now I get it I was still at oe-classic and bitbake 1.12 mentally .,. Jan 07 20:33:38 that was the 1.15 I got above I think Jan 07 20:33:42 master Jan 07 20:33:47 yep Jan 07 20:35:01 bluelightning: but util-linux did not Jan 07 20:35:34 bluelightning: hmm i am a little bit confused, because this is what i did not expect Jan 07 20:36:27 oe-core explains everything thats why I got a distroless kernel^^ Jan 07 20:37:16 i guess it boots to initfs? Jan 07 20:37:23 and spawns a rootshell? Jan 07 20:37:45 k back to work -.- Jan 07 20:39:06 cs_nbp: I've emailed tworaz to see if I should list those layers Jan 07 20:49:32 Hello folks. I just created a meta-mozilla layer and it would be nice if people could look at it and help. https://github.com/OSSystems/meta-mozilla Jan 07 21:36:43 0 Jan 07 21:38:27 1 Jan 07 21:38:31 ;) Jan 07 21:39:52 heh Jan 07 21:40:25 Ok; enough for me ... Jan 07 21:40:32 I will be back on Monday Jan 07 21:40:33 see ya Jan 07 21:51:02 otavio: bye Jan 07 22:03:30 bluelightning: sorry i was inwolved in changing shr wiki, what about fixing util-linux-native ? Jan 07 22:03:41 i have no idea Jan 07 22:16:30 nschle85|2: it would be worth confirming if that --disable-ncurses switch ever made it to the configure command line Jan 07 22:16:59 nschle85|2: check the tmp/work/*/util-linux-native-*/temp/log.do_configure Jan 07 22:18:04 ok ill do Jan 07 22:20:42 bluelightning: http://norman-schleicher.de/jenkins/job/shr-core-nokia900-shr-image/ws/shr-core/tmp-eglibc/work/x86_64-linux/util-linux-native/util-linux-native-2.20.1-r1/temp/ i have several do configure logs Jan 07 22:21:44 bluelightning: i guess latest one is linked to the highest number Jan 07 22:25:43 bluelightning: checking ncurses.h presence... yes checking for ncurses.h... yes checking for initscr in -lncursesw... no checking for initscr in -lncurses... yes checking ncurses/ncurses.h usability... no checking ncurses/ncurses.h presence... no checking for ncurses/ncurses.h... no Jan 07 22:30:36 did the disable-ncureses flag apply ? Jan 07 22:41:55 bluelightning: http://norman-schleicher.de/jenkins/job/shr-core-nokia900-shr-image/ws/openembedded-core/meta/recipes-core/util-linux/ this is my directory Jan 07 23:09:06 good night bye Jan 07 23:09:18 nschle85|2: it shouldn't even be trying to use it Jan 07 23:09:22 nschle85|2: cya Jan 07 23:10:18 bluelightning: i too lazy Jan 07 23:10:49 iam too lazy , lets talk tomorrow again :-) Jan 07 23:10:53 nschle85|2: ok Jan 07 23:10:53 bye Jan 07 23:11:45 bye Jan 07 23:11:53 cya Jan 07 23:21:41 hmmm Jan 07 23:21:42 ./alsaloop -P plug:dmix:0 -C plug:dsnoop:1 -S 0 Jan 07 23:21:49 produces 100% CPU usage too Jan 07 23:21:57 I worked arround in my code Jan 07 23:25:01 could it be a bug in the kernel? Jan 07 23:28:36 strange.... Jan 07 23:29:00 ./alsaloop -P plughw:0 -C hw:1 -S 0 -f S16_LE -r 8000 -c 1 -t 1000000 works Jan 07 23:29:08 but the CPU usage is still very high Jan 07 23:29:40 and it's way higher when not in call Jan 07 23:31:22 note the ./ Jan 07 23:31:33 the ./ is the version recompiled with libsamplerate support Jan 07 23:31:41 libsamplerate is autodetected Jan 07 23:39:07 mickeyl, hi Jan 07 23:39:15 hangup blocks Jan 07 23:39:25 I'll tar the logs Jan 07 23:39:58 GNUtoo|laptop: does this use some neon code Jan 07 23:41:22 oops wrong channel Jan 07 23:41:28 khem, hi Jan 07 23:41:44 khem, let me look Jan 07 23:43:00 khem, you're talking about the gcc problem I had? Jan 07 23:43:20 in angstrom I can't compile gcc-cross Jan 07 23:43:29 for my bug2.0 Jan 07 23:44:49 /home/gnutoo/embedded/oe/oe-core/angstrom/oetmps/angstrom/work-shared/gcc-4.5-r45+svnr181733/gcc-4_5-branch/gcc/cp/cfns.gperf:101:1: error: 'gnu_inline' attribute present on 'libc_name_p' Jan 07 23:44:52 I'll look Jan 07 23:45:33 hmm Jan 07 23:45:45 the .gperf file seem very strange Jan 07 23:45:50 I'm not used to its format Jan 07 23:45:54 it's not code Jan 07 23:46:38 isalpha is what is at line 101 Jan 07 23:46:53 I've the issue since a very long time ( I can't remember) Jan 07 23:47:10 so because of that issue I abandoned working on the bug 2.0 Jan 07 23:47:31 somehow your build is deciding to rebuild cfns.h Jan 07 23:47:41 ah? Jan 07 23:47:59 do u have gperf installed ? Jan 07 23:48:03 where? Jan 07 23:48:07 on the build system Jan 07 23:48:09 let me look Jan 07 23:48:15 yes I do Jan 07 23:48:21 I'll remove it Jan 07 23:48:37 under SHR the toolchain work btw Jan 07 23:48:47 I can rebuild at will, with sstate disabled Jan 07 23:49:40 I aptitude remove gperf.... Jan 07 23:49:46 and retry now Jan 07 23:50:13 no you dont have to remove it Jan 07 23:50:22 ok what should I do then? Jan 07 23:50:46 it happens everytime ? Jan 07 23:50:47 anyway there is a gperf-native.... Jan 07 23:50:49 yes Jan 07 23:50:54 I tried everyhting Jan 07 23:50:58 rebuild from scratch Jan 07 23:51:08 did you try to delete the tar file Jan 07 23:51:11 going trough 1 BB_THREAD Jan 07 23:51:20 what tar file? Jan 07 23:51:23 in DL_DIR Jan 07 23:51:23 the sstate file? Jan 07 23:51:25 no Jan 07 23:51:26 I didn't Jan 07 23:51:39 I'll move them Jan 07 23:51:44 ok my guess is somehow the gcc sources are corrupted Jan 07 23:52:00 and since there are no checksums for scm based recipes Jan 07 23:52:03 they dont get checked Jan 07 23:52:19 it could be Jan 07 23:52:21 so remove the tar file Jan 07 23:52:23 it's very probable Jan 07 23:52:24 ok Jan 07 23:52:33 as well as the gcc checkout under dL_DIR/svn Jan 07 23:52:38 ok Jan 07 23:52:41 thanks a lot!!!! Jan 07 23:52:45 then retry Jan 08 00:05:10 when i removed a broken package in downloads/ , what is the correct procedure to tell oe-core that it's gone and instead of trying to unpack it it should refetch it? Jan 08 00:05:48 #nick cs_nbp Jan 08 00:11:04 cs_nbp: bitbake -ccleanall Jan 08 00:11:31 infact that will delete it as well u dont have to do that manually Jan 08 00:12:00 ah ok thanx Jan 08 00:12:31 so u cant forget that u cded into download/ and cant do bitbake from there :D Jan 08 00:40:15 if the crosscompiler requires a header (eg libelf.h) he does mean my i686 machine, right? juast a bit confused-.- Jan 08 01:13:04 has someone seen this? build/tmp/work/armv5te-oe-linux-uclibceabi/libiconv-1.14-r0/libiconv-1.14/srclib/progreloc.c:297: undefined reference to `compute_curr_prefix' **** ENDING LOGGING AT Sun Jan 08 02:59:57 2012