**** BEGIN LOGGING AT Sun Jan 08 02:59:57 2012 Jan 08 03:28:15 anyone getting this too: oe-core/meta/recipes-support/libiconv/libiconv_1.14.bb' failed Jan 08 03:32:38 it says 'progreloc.c:297: undefined reference to `compute_curr_prefix'' but when I look there I can clearly see the function is included with #include "relocatable.h" so wth? Jan 08 03:36:42 switching to eglibc to see if it resolves that Jan 08 03:43:38 cs_nbp: eglibc does not use libiconv Jan 08 03:43:42 so it will seem to work Jan 08 03:43:55 but if there is a problem with libiconv that remains Jan 08 03:44:14 it seems to work, just not being able to dload all packages as usual ;) Jan 08 03:44:43 so i get them manually from angstrom unstable sources Jan 08 03:44:54 thanks btw Jan 08 03:45:05 but this libiconv thing is weird Jan 08 03:45:11 doesnt make any sense Jan 08 03:45:19 cd .. Jan 08 03:47:28 spent two hours looking at code that had not the error claimed :D Jan 08 03:52:23 are you using angstrom Jan 08 03:52:51 and if yes then is it angstrom-bleeding Jan 08 03:56:40 nope not really but the versions fit and the compilation goes on Jan 08 03:56:56 but jlime is debian-based too i think Jan 08 03:57:47 or is there a better way? Jan 08 04:03:11 hmm Jan 08 04:10:08 now it just told me it cant download a file from sourceforge that just is at the path it says it looks Jan 08 04:10:38 http://downloads.sourceforge.net/project/enlightenment/imlib2-src/1.4.4/imlib2-1.4.4.tar.gz Jan 08 04:11:35 ah i see its a checksum mismatch Jan 08 04:16:25 is there a way to ignore a checksum mismatch.,.i guess not Jan 08 04:16:45 there is Jan 08 04:17:24 how? Jan 08 04:20:19 change the checksum in the imlib2.bb ? Jan 08 04:22:57 well yes that helped but nor there is a checksum eror with bluez-.- Jan 08 04:34:30 k gn8 all Jan 08 15:37:12 hi Jan 08 15:49:24 is there another source than kernel.org to get eggdbus-0.6 ? build always stops there because the kernel.org package doesn't exist Jan 08 15:49:45 or is it safe to leave it out? sounds like its important for dbus-.- Jan 08 15:51:45 perhaps someone here dloaded the archive bfore kernel.org was hacked? Jan 08 15:53:43 cs_nbp: I usually search on google, often I find the packages on some fedora mirrors Jan 08 15:54:15 well if i ask here you can safely assume google didnt work for thia ;) Jan 08 15:54:35 there is packages but its the wrong ones, they dont compile at all Jan 08 15:55:37 hmm, is it some older stuff? Jan 08 15:55:53 what's the exact package name? Jan 08 15:58:57 its eggdbus-0.6.tar.bz2 i think, i need to repeat the error atm, the package name scrolled out of reach .,. Jan 08 16:00:11 yes its eggdbus-0.6.tar.bz2 Jan 08 16:03:12 http://dl.t2-project.org/mirror/9.0/e/eggdbus-0.6.tar.bz2 Jan 08 16:03:49 hey thanx ill try that one Jan 08 16:04:28 I found it on google btw, you can check the md5sums to see if it matches with whatever the OE recipe is expecting Jan 08 16:04:46 thats the main problem Jan 08 16:04:59 the one from sourceforge.jp eg is completely different Jan 08 16:05:11 now I see that I also have one here, downloaded a while ago, so I could send you my copy if you want Jan 08 16:05:39 wait ill test this one first Jan 08 16:09:31 I can't believe it while we talked. the main source http://cgit.freedesktop.org/~david/eggdbus came back online Jan 08 16:18:19 yep it compiled, whatever vodoo you have done, Jin^eLD, it worked like a charm Jan 08 16:25:28 for missing sources one might also check the angstrom mirror Jan 08 16:26:09 this one? http://www.angstrom-distribution.org/unstable/sources/ Jan 08 16:40:25 i cant seem to get through libtool-cross_2.4.2.bb, do_configure) failed with exit code '1', the patches don't seem to apply Jan 08 16:40:53 do i need a newer version of this? Jan 08 16:42:13 ERROR: Function 'do_qa_configure' failed Jan 08 16:43:59 i've no idea what you are building (distro, target, etc) so can't really say anything abou tit Jan 08 16:44:40 im building jlime-lxde-image for machine netbookpro, oe-overlay from tworaz/github Jan 08 16:46:03 oe core or oe classic? what distro ? Jan 08 16:46:18 oe core and the distro is 'Jlime' Jan 08 16:46:43 its defined in the meta-jlime overlay tworaz made Jan 08 16:46:44 cs_nbp: what does the actual error say? if it's do_configure it won't be patches not applying at least Jan 08 16:47:01 I'm guessing it'll be about LIC_FILES_CHKSUM Jan 08 16:47:32 cs_nbp: I'm not too much into oe-core, sry Jan 08 16:48:12 maybe ask the guy who made the overlay Jan 08 16:49:55 cs_nbp: is libtool-cross actually being modified by either of those layers? Jan 08 16:50:09 (I would sure hope not...) Jan 08 16:52:20 cs_nbp: btw http://www.angstrom-distribution.org/unstable/sources/ gave me some textual linux foundation page Jan 08 17:06:35 bluelightning: that wasn't entirely right, when I grep for only 'libtool' there are a lot of matches in meta-jlime Jan 08 17:06:44 yo Jan 08 17:07:11 bluelightning: eg a suspicious configure-libtool.patch Jan 08 17:07:37 libiconv has been supposedly fixed, wheee. gonna go run a build real quick Jan 08 17:07:58 sup i spent 2 hours this night finding a working one -.- Jan 08 17:09:10 brb Jan 08 17:10:06 bluelightning: guess i'll contact tworaz about that Jan 08 17:10:43 cs_nbp: yeah, sounds a bit suspicious Jan 08 17:10:55 anarsoul: hi, are you around? Jan 08 17:14:59 is there another distro for kernel 3.1? Jan 08 17:15:28 angstrom has 'hardcoded';) 2.6.* kernel Jan 08 17:16:33 or can I just change that in the distro definition of angstrom? Jan 08 17:16:42 cs_nbp: the current version of angstrom doesn't hardcode 2.6... Jan 08 17:17:02 at least I don't think it does Jan 08 17:18:33 bluelightning: so i guess there is meta-angstrom? ill take a look Jan 08 17:18:43 cs_nbp: yes Jan 08 17:19:10 I don't know how far you will get with it with an "unsupported" machine such as the netbookpro however Jan 08 17:19:18 also, it mandates systemd now Jan 08 17:19:26 which may or may not be an issue for you Jan 08 17:19:48 abloody cell is useless here Jan 08 17:20:15 well someone has to try it :D Jan 08 17:20:39 gonna have to give this a go from the office tomorrow, can't do crap over this connection. Jan 08 17:20:51 btw what's with the 'include' dir in distro, eg meta-efl provides only that Jan 08 17:24:17 e would rock on the nbp, i think Jan 08 17:24:31 yeah that's the final goal^^ Jan 08 17:25:12 i'd be cool with fluxbox or something Jan 08 17:25:42 at the moment i'd be cool with anything that runs a graphical interface on 3.1 kernel Jan 08 17:25:52 haha, yeah Jan 08 17:26:32 sure I can start one manually after doing pivot_root but that's not entirely the same (plus the rootfs doesn't really fit to the kernel) Jan 08 17:27:09 the e17 mod illume from gta02 would be nice too Jan 08 17:27:11 really only gonna use it for ssh+remote X terminal here. maybe run an mpd client. Jan 08 17:27:19 illume is spiffy. Jan 08 17:27:47 and lean Jan 08 17:27:49 pretty good base for tablet, imo Jan 08 17:28:23 I doubt one of the illume guys (rafael?) has a netbook pro Jan 08 17:28:37 i played with it a bit on an old x86 tablet. some slow-ass P3 slate Jan 08 17:28:52 cs_nbp: the conf/distro/include dir is intended for files that would be included by distro configs Jan 08 17:29:06 cs_nbp: for meta-efl I guess it contains efl-specific distro-policy type stuff Jan 08 17:29:22 yeah, but if you can get angstrom up, illume shouldnt be too hard. Jan 08 17:29:31 bluelightning: so unless i do a distro file first i won't be using it soon.,. Jan 08 17:29:43 ill try some minimal angstrom Jan 08 17:29:57 if I can figure out where the distro definitions are hiding Jan 08 17:30:01 cs_nbp: I know nothing about using meta-efl... but I guess you may find you can just include whatever file is in there from local.conf if you need to Jan 08 17:30:51 bluelightning: but still i would have no eg bitbake e17-image until i define it first, or do i got it wrong? Jan 08 17:32:22 aight, afk till tomorrow, catch you folks later. Jan 08 17:33:05 cs_nbp: doesn't look like e17-image is provided in meta-efl Jan 08 17:33:19 cs_nbp: so I guess it's meant to be provided by a distro Jan 08 17:33:38 bluelightning: angstrom has a meta-efl subfolder Jan 08 17:33:59 bluelightning: but i'm having a hard time finding the distro defs even in angstrom -.- Jan 08 17:34:37 cs_nbp: meta-angstrom doesn't have that subfolder... maybe you've used the setup scripts? Jan 08 17:35:06 (which would I guess be the recommended method for building angstrom) Jan 08 17:35:44 bluelightning: what setup scripts? my meta-angstrom has meta-efl meta-gpe meta-gnome meta-oe and meta-xfce Jan 08 17:36:16 cs_nbp: er, that sounds a lot like meta-oe and not meta-angstrom Jan 08 17:36:22 meta openembedded has meta-efl as well Jan 08 17:36:25 yeah Jan 08 17:36:49 thats what i have in meta-oe Jan 08 17:36:53 bluelightning: well i checked out meta-angstrom into the oe-core dir Jan 08 17:37:20 that's in my meta-oe,too Jan 08 17:37:30 cs_nbp: I guess this is why I suggested keeping things at the same level, it avoids layers being confused Jan 08 17:37:51 cs_nbp: what you were describing is not meta-angstrom anyway Jan 08 17:38:13 bluelightning: its what i can see in my meta-angstrom dir Jan 08 17:38:16 meta-angstrom is mostly recipes- folders Jan 08 17:38:53 cs_nbp: I don't know, but meta-angstrom is definitely this: http://git.angstrom-distribution.org/cgi-bin/cgit.cgi/meta-angstrom/ Jan 08 17:39:03 anything else is not it :) Jan 08 17:39:13 i have them all in seperate folders. it seems best to keep them seperate, and then add them in the bblayers.conf Jan 08 17:39:25 avoids confusion Jan 08 17:39:27 muriani: yes Jan 08 17:40:22 damn i wish i had workable desktop internet. tethering flat isnt working for some reason. Jan 08 17:40:28 bluelightning: ah ok, i read over the urls and checked out meta-openembedded meta-angstrom, which indeed is not that Jan 08 17:41:12 right, no wonder you couldn't find the distro configs... Jan 08 17:41:13 ;) Jan 08 17:41:28 *shame* Jan 08 17:42:04 and i got to stop type terminal commands into irc -.- Jan 08 17:42:30 cs_nbp: for example, my checkouts folders are oe-core, oe-tworaz, meta-openembedded, meta-angstrom, meta-ti Jan 08 17:42:44 build direcories i have in the same level as well Jan 08 17:43:13 muriani: so you had any more luck than me with this yet? Jan 08 17:43:38 muriani: i mean nbp Jan 08 17:43:54 maybe, but i was dying at libiconv Jan 08 17:44:09 i got further than you i think, but no success. Jan 08 17:44:09 muriani: yeah that was a hard nut Jan 08 17:44:19 i cant do any more until tomorrow Jan 08 17:44:37 muriani: well yesterday i got to an attempt to build the nBkProOs.img, but it failed Jan 08 17:44:52 oh nice Jan 08 17:45:00 then i realised i probably shouldnt have switched to eglibc in mid-build Jan 08 17:45:09 that's how i got around the libiconv Jan 08 17:45:13 really we just need the rootfs Jan 08 17:45:14 ah Jan 08 17:45:25 and we just need one Jan 08 17:45:39 see, even with eglibc i got libiconv fail Jan 08 17:46:04 i thought it doesnt use that? hmm Jan 08 17:47:02 EGLIBC provides an iconv() implementation and can convert from/to Unicode, therefore libiconv is not required on an CLFS system Jan 08 17:47:37 from cross-lfs-org Jan 08 17:49:48 machine kernel version for 3.1 is still 2.6, right? Jan 08 17:50:01 or is it just linus? Jan 08 17:59:40 is 'DISTRO' name for angstrom not 2010.X ? it's not found even though the layer meta-angstrom is added Jan 08 17:59:59 and i can see it w bitbake-layers show_layers Jan 08 18:03:14 misread-.- Jan 08 18:03:51 cs_nbp: there must be a conf/distro/distroname.conf file for whatever distroname you set Jan 08 18:04:21 bluelightning: yes i set the version and its the filename Jan 08 18:05:06 cs_nbp: so you have exactly this: DISTRO = "angstrom-2010.x" ? Jan 08 18:05:18 bluelightning: now yes Jan 08 18:05:28 that should work Jan 08 18:05:33 bluelightning: and i'm already at the next error Jan 08 18:05:58 which would be? Jan 08 18:06:52 bluelightning: it says no provider for e-wm-config-default Jan 08 18:07:13 i chose efl-nodm-image, perhaps not the wisest choice Jan 08 18:07:36 no idea sorry Jan 08 18:14:52 little thing at the side (while updatedb is running): i changed some checksums in some recipes, now git pull --rebase doesn't work, what do i say to update the repo anyway? Jan 08 18:18:54 locate e-wm Jan 08 18:22:46 no big difference with systemd-gnome-image NoProvider: systemd-gnome-image Jan 08 18:23:30 ERROR: Nothing RPROVIDES 'gnome-themes' Jan 08 18:30:09 if I remove gnome-themes (as I don't need them anyways), the next two things without a provider are epiphany and gedit Jan 08 18:30:51 is there a way out of this? i mean apart not using systemd-gnome-image obviously Jan 08 18:34:32 is this known broken or am i doing something wrong? Jan 08 18:41:20 ka6sox: around Jan 08 18:42:16 cs_nbp: how did you checkout your souces Jan 08 18:45:17 khem: i was in my topdir that contains build/ oe-core/ etc and did git clone meta-angstrom Jan 08 18:45:47 khem: then added it to bblayers.conf and set distro to 2010.x in local.conf Jan 08 18:46:04 khem: finally did bitbake systemd-gnome-image Jan 08 18:46:33 khem: and this didn't even start fetching sources Jan 08 18:47:43 did you follow http://www.angstrom-distribution.org/building-angstrom Jan 08 18:47:55 its relatively simple Jan 08 18:48:13 khem: wasn't aware of that Jan 08 18:48:24 http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/ Jan 08 18:48:27 khem: so i've got to run some setup scripts Jan 08 18:48:36 here I discuss about oe-core based angstrom Jan 08 18:49:02 for best results yes follow those instructions Jan 08 18:49:20 ok Jan 08 18:49:55 just got to wait til the systemd-image (without gnome) stops building for whatever reason Jan 08 18:51:32 like eg eggdbus checksum error-.- Jan 08 18:51:33 kill bitbake Jan 08 18:51:46 its killing itself Jan 08 18:51:55 actually you should start from scratch it seems you have pretty messed up tree Jan 08 18:52:13 you mean clean tmp and downloads? Jan 08 18:52:39 I means delete the whole stuff Jan 08 18:52:40 tor re-checkout everything Jan 08 18:52:52 setup scripts will get you started groundup Jan 08 18:52:53 ah ok, well it really doesn't take long Jan 08 18:53:13 you dont have to checkout anything Jan 08 18:53:21 setup scripts will do that Jan 08 18:54:03 you mean it checks out oe-core too? Jan 08 18:54:47 Good morning Jan 08 18:55:06 o_O Jan 08 18:55:10 hi Jan 08 18:56:12 still trying to get a rootfs for nbp Jan 08 18:57:46 hi cs_nbp Jan 08 18:58:50 florian: tworaz jlime stuff fails either at libiconv (with uclibc) or libtools-2.4.2 (with eglibc) Jan 08 19:00:08 cs_nbp: okay... i guess I can give it a try unless I fall asleep Jan 08 19:00:49 florian: afk shaving smone Jan 08 19:29:33 cs_nbp: I provided a fix for libiconv/uclibc here http://patchwork.openembedded.org/patch/18767/ Jan 08 19:36:17 | ERROR: Function 'Fetcher failure for URL: 'file://0001-Inital-pcf50606-support.patch'. Unable to fetch URL file://0001-Inital-pcf50606-support.patch from any source.' failed Jan 08 19:36:18 how can I find out which directories bitbake is looking in for files? Jan 08 19:36:41 alternatively, how can I find the value of a variable used in a recipe? Jan 08 19:37:44 bitbake -e Jan 08 19:37:48 and FILESPATH Jan 08 19:41:17 yes, got there myself in the end Jan 08 19:41:20 thanks though ;-) Jan 08 19:56:52 khem: sorry i was afk, that's awesome, i'll give it a shot Jan 08 19:59:08 khem: is there any oe patch facility i'm supposed to use or do i use 'patch' like usual? Jan 08 20:01:38 cs_nbp: "git am" should do the trick Jan 08 20:01:48 * florian shoots oe-core Jan 08 20:01:50 florian: back, sometimes i have to prevent a shaving noob from cutting his face off Jan 08 20:02:00 florian: txh Jan 08 20:02:58 it take about 10s to make oe-core fail due to compile problem Jan 08 20:04:24 florian: building libiconv? or the jlime-lxde-image? Jan 08 20:04:52 building pseudo :-( Jan 08 20:05:03 florian: whos that? Jan 08 20:05:07 | /usr/include/gnu/stubs.h:9:27: error: gnu/stubs-64.h: No such file or directory Jan 08 20:06:02 florian: btw i need patch format, not mbox format, right? just to be sure Jan 08 20:07:20 cs_nbp: mbox is fine for am Jan 08 20:09:53 florian: now rebuilding oe-core tree according to khems recipe Jan 08 20:14:19 florian: are you building for 64 bit? I had a similar error last year and had to edit a mekefile and change all 64's to 32's Jan 08 20:14:37 florian: because i compiled for 32 bit Jan 08 20:14:47 florian maybe unrelated Jan 08 20:21:34 cs_nbp: nope 32... but it might be a bug in pseudo build stuff. Jan 08 20:22:35 i guess it reads 32 vs. 64 bit from the kernel (uname or such) wich is simply wrong Jan 08 20:24:40 oh Jan 08 20:24:41 no Jan 08 20:24:46 its even worse Jan 08 20:25:02 bitbake gets it wrong :-( Jan 08 20:26:19 florian: what distro are you running? Jan 08 20:26:37 FWIW, I build regularly on a mix of 32 and 64 bit machines and have not had any issues like this Jan 08 20:26:49 bluelightning: Debian Jan 08 20:27:11 bluelightning: pure 32bit, only a 64bit kernel Jan 08 20:28:15 florian: how does that work? Jan 08 20:28:51 bluelightning: running a linux x86 32bit userland and a 64 bit kernel ist fine Jan 08 20:32:07 florian: ok, there's some code in recipes-devtools/pseudo/pseudo.inc that's supposed to handle this, can you tell which part of the check in do_compile_prepend_virtclass-native is incorrect? Jan 08 20:32:39 s/incorrect/failing/ Jan 08 21:03:41 bluelightning: this basically looks good... let me try to find out why this happens Jan 08 21:13:03 bluelightning: you do not happen to know how SITEINFO_BITS gets set for building native? Jan 08 21:29:53 khem, am now Jan 08 21:30:49 hi ka6sox Jan 08 21:34:03 hiya woglinde Jan 08 21:49:15 bluelightning: it does not seem to be easy to find out but overriding BUILD_ARCH in local.conf works Jan 08 21:50:38 florian: you build for debian means there is a distro conf for debian somewhere?^ Jan 08 21:51:03 cs_nbp: no, i'm building on a debian host Jan 08 21:51:14 florian: ah dam Jan 08 21:51:26 cs_nbp: he would have asked for DISTRO otherwise :) Jan 08 21:51:54 florian: yeh Jan 08 21:57:36 * florian just noticed that deleting old recipes creates a maintenance hell for other layers Jan 08 22:05:45 :) Jan 08 22:05:57 on one hand the layers seem handy, but on the other they can be pure evil Jan 08 22:07:42 yes indeed Jan 08 22:09:10 I am still struggling with situations where I have to undo whatever a layer that comes before my layer appended to a package Jan 08 22:09:33 since there is no way to prevent another layer from appending to a particular package Jan 08 22:09:50 yeah I'm stuck at pseudo too, but mine won't even fetch Jan 08 22:10:01 hmm yes, that's evil as well... Jan 08 22:10:11 WARNING: Fetcher failure for URL: 'None' Jan 08 22:10:20 uh?! Jan 08 22:10:23 ERROR: Function 'Fetcher failure for URL: 'http://downloads.yoctoproject.org/releases/gnu-config/gnu-config-yocto-20111111.tgz' Jan 08 22:10:36 first the one than the other, perhaps 1st is normal Jan 08 22:10:38 There is an issue I di not ran into so far? ;) Jan 08 22:11:16 if you give me root access I can make more :D Jan 08 22:12:10 :-) Jan 08 22:13:50 cd .. Jan 08 22:13:56 onoz not again-.- Jan 08 22:14:45 k i thieved the one from angstrom unstable, checksum seems to work out .,. Jan 08 22:14:56 hmm, it seems that bitbake will sometimes return a retcode of 1 even if no errors were printed Jan 08 22:15:04 i.e. all seems fine, but it exits with 1 Jan 08 22:15:14 yeah make me courage Jan 08 22:16:06 Jin^eLD: The worst problem I currently see is fragmentation... everyone now creates own layers and does not care about contributing back. Jan 08 22:16:29 its like walking around after a plane crash Jan 08 22:17:16 ... because this additionally became harder - more restrictions and more complicated to decide where to contribute) Jan 08 22:17:24 florian: yes, that's true as well Jan 08 22:17:36 and some layers still have duplicates I think Jan 08 22:19:59 omg building this pseudo takes longer than the main build ;) Jan 08 22:20:32 Another thing I do not understand is why everyone cares about tiny recipe QA details but not about checking if download locations are still valid. Jan 08 22:20:47 This is what breaks OE for many users Jan 08 22:21:18 plus download locations temporarily being down, like for eggdbus on kernel.org Jan 08 22:21:50 most kernel.org downloads are down for months now :-( Jan 08 22:22:03 well it came back up today Jan 08 22:22:09 at least the one i needed Jan 08 22:23:28 the kernel.org situation has been quite ridiculous... although I'm not blaming OE for that Jan 08 22:23:42 true Jan 08 22:32:57 had to download pseude from somewhere else, too Jan 08 22:34:54 eg module-init-tools-3.16.tar.bz2 is not on kernel.org either Jan 08 22:36:40 usbutils, PAM, usbutils... Jan 08 22:36:48 libcap Jan 08 22:37:19 florian: I would agree with the layers leading to fragmentation but at the same time a huge jump in functionality. Upstreaming seems to not always be working as well as it should ;). Jan 08 22:38:08 DJW|Home: It worked really good for oe... much better compared to other projects. Jan 08 22:38:56 florian: yep, not perfect but I can't disagree with that. Jan 08 22:43:09 so you say I should start downloading them before bitbake asks? :) Jan 08 22:47:18 we should write a fetcher tat googles for an alternate download location ;) Jan 08 22:47:52 or at least asks Jan 08 22:48:00 without stopping the whole process Jan 08 22:48:33 this would be pretty cool Jan 08 22:48:42 * florian needs more time... Jan 08 22:48:47 should provide filename and checksum so i can search on google myself Jan 08 22:48:59 :D Jan 08 22:54:11 brb changing computer Jan 08 23:16:19 Jin^eLD: if there's an exit code of 1 there was an error Jan 08 23:17:05 bluelightning: but I do not see any errors Jan 08 23:17:17 Jin^eLD: not a single ERROR: line? Jan 08 23:17:21 notespace: Tasks Summary: Attempted 77 tasks of which 70 didn't need to be rerun and 0 failed. Jan 08 23:17:28 that doesn't mean there were no errors Jan 08 23:17:52 there is an open bug to improve the summary however in the case where errors did occur Jan 08 23:17:53 aah, ok ok wait, I think I know then Jan 08 23:18:05 what that summary does mean is that no tasks failed, and that is true Jan 08 23:18:06 it as before it checked the cache Jan 08 23:18:09 so I did not notice it Jan 08 23:18:10 ERROR: Task do_prepare_source from /oe/jin/dss-oe/dS/meta-dss11-staging-fieldtest/recipes-support/yoke/yoke-native_0.1.bb seems to be empty?! Jan 08 23:18:23 actually wanted to ask what that error means Jan 08 23:18:24 however if anything calls bberror / bb.error then that's what happens Jan 08 23:18:48 Jin^eLD: well, is that task empty? Jan 08 23:19:06 I have a SRC_URI Jan 08 23:19:22 but that points to a .js file that is fetched from the web Jan 08 23:19:41 so I guess there is no prepare source task then? Jan 08 23:20:00 I'm not familiar with do_prepare_source, is that a task you've defined yourself? Jan 08 23:20:18 oh damn indeed, what the hell was I thinkin there? Jan 08 23:20:39 sorry for all the confusion :) Jan 08 23:20:46 no problem :) Jan 08 23:20:51 I guess you found it? Jan 08 23:20:56 I looked at that recipe a dozen of times Jan 08 23:21:20 yeah, at the very end of the recipe there was an addtask line where I obviously tried to do some hacking but then removed the task and forgot to remove the addtask thing Jan 08 23:21:33 aha, bingo :) Jan 08 23:22:07 two problems solved ;) Jan 08 23:24:16 cs_nbp: here the nbp kernel fails to build... it misses the bootcode file Jan 08 23:25:18 Jin^eLD: btw with your need to "undo" what other layers are doing... can you be more specific? I know we've talked about this before but I don't remember the details Jan 08 23:25:27 Jin^eLD: is this a public layer or an internal one? Jan 08 23:26:39 bluelightning: there is a recipe, I have to look up what it was, I don't remember by heart now, the angstrom layer appends to that recipe and does some stuff which I do not need and which contradicts with my intentions; so in my "private" layer I am appending again and trying to filter_out and undo all the things that the angstrom layer does to this recipe Jan 08 23:30:33 khem: with your patch libiconv successfully installed Jan 08 23:33:42 Jin^eLD: when you start needing to change distro policy either you need to try to work to change that policy or you need your own distro Jan 08 23:34:19 bluelightning: that is true, but for now Angstrom fits me in all regards except for the systemd move Jan 08 23:34:31 maintaining an own distro is more effort Jan 08 23:34:58 Jin^eLD: well, I would imagine you're going to find it harder and harder to use in that case, if angstrom is moving to systemd only Jan 08 23:35:10 indeed... Jan 08 23:35:47 but so far I only had to make sure the systemd stuff lands in an own package and that the init scripts are still around Jan 08 23:35:56 hrm... imlib2 md5sum is wrong... wtf?! Jan 08 23:35:58 in most cases systemd is alredy put in an own package Jan 08 23:36:07 so it was not too much work yet Jan 08 23:36:41 florian: yes i had that too Jan 08 23:36:53 florian: but whats the bootcode file Jan 08 23:37:00 florian: do we have thatß Jan 08 23:37:38 florian: i thought i got that because i switched from ulibc to eglibc in the middle of a build .,. Jan 08 23:38:39 cs_nbp: I think so... https://github.com/tworaz/boost-bootcode Jan 08 23:39:22 floraian: yeehaw but wth isn't that in the oe. k i guess he has no problem using his stuff Jan 08 23:39:36 florian: do you know where it goes? Jan 08 23:39:51 florian: perhaps I can put it there before my build fails Jan 08 23:40:40 well too late forgot about linux-pam Jan 08 23:40:56 cs_nbp: looks like we need to make a bb for it (or well sure - build it manually and place where it needs to go) Jan 08 23:41:00 tomorrow Jan 08 23:41:07 * florian needs sleep Jan 08 23:41:51 good night Jan 08 23:41:56 florian: well gn flo Jan 08 23:42:00 cyaZ Jan 09 00:52:46 ok gn8 all **** ENDING LOGGING AT Mon Jan 09 02:59:57 2012