**** BEGIN LOGGING AT Tue Oct 11 02:59:57 2011 Oct 11 03:33:15 grr, gclient's git support is weak at best Oct 11 06:02:24 good morning all Oct 11 07:05:09 good morning Oct 11 07:57:59 boah angstrom-distribution.org is obviously hosted on a 8086 via 300bd :-( Oct 11 08:03:41 Hi Oct 11 08:04:20 Hi have a problem building a problem with libraries Oct 11 08:04:23 I get this message Oct 11 08:04:26 NOTE: Multiple libraries (libntcip.so.1, libmulticast.so.1, libtcp.so.1, libdcz.so.1, librtsp.so.1, libpelcod.so.1, libserial.so.1, libhdlc.so.1, libvtcooperation.so.1, libcias.so.1, libpmpp.so.1, libmodbus.so.1, libzeroconf.so.1, libmacro.so.1, libcalibration.so.1) found and LEAD_SONAME not defined Oct 11 08:04:45 I tried googling a while, but I did not understand where I have to put LEAD_SONAME Oct 11 08:05:01 a what value I have to assign to it Oct 11 08:05:06 can you give me a hint? Oct 11 08:08:58 hat recipe is that? Oct 11 08:09:00 what Oct 11 08:10:21 it's for debian naming to know which lib is most important.. so if you put LEAD_SONAME = "libtcp.so" then it will be named ie libtcp1_1.0.ipk Oct 11 08:10:42 other option is to package each lib in separate package, then you won't need LEAD_SONAME Oct 11 08:12:06 it's a recipe for a software of mine Oct 11 08:12:20 JaMaOff, not a public software Oct 11 08:12:46 but I'd like to get a single ipk containing both binaries and libraries Oct 11 08:53:56 eFfeM_work: ping Oct 11 10:08:12 hello Oct 11 10:26:55 in oe-core i miss a recipe for gcc native (execute gcc on the target) is there some place ? Oct 11 10:43:27 lautriv: -native are for host system, without prefix are for target.. so gcc is to execute gcc on target Oct 11 10:47:53 JaMaOff, unfortunately it won't do the trick. i got a package from angstrom distro for armv5te but gcc -v says --build=x86_64-linux and a try ro compile a kernel leads to missing cc1, even with an additional entry in $PATH for the /usr/libexec/gcc/arm........... Oct 11 10:50:03 lautriv: install gcc-symlinks Oct 11 10:50:32 did so too, solved nothing. find / -name gcc1 won't tell me either. Oct 11 10:52:44 why gcc1? Oct 11 10:53:08 err cc1 Oct 11 10:53:54 then configure kernel to use right cc Oct 11 10:54:08 this doesn't look like gcc fault Oct 11 10:55:23 i made a symlink in /usr/bin nope. But as far as i can tell on the host-system (x86_64) a make -j8 brings a lot of cc1 processes up. Oct 11 10:59:14 there is also /usr/lib/gc/x86_64-linux-gnu/4.6/cc1 but not on the target in the resp. dir Oct 11 11:03:48 that's in your distribution.. my doesn't have cc1 and my kernel build is not using cc1.. so again.. configure kernel to use available cc not the one you had on some other system Oct 11 11:04:57 that's a chicken-egg problem because i need gcc on the target OR some config on the host (which i miss) Oct 11 11:05:32 but you have gcc on target.. Oct 11 11:05:40 it's just not named cc1 Oct 11 11:07:27 yes and a link comes up with unrecognized functions ( btw. on the x86_64 there is a hughe diference where gcc is approx. 360k and cc1 about 11,2M ) Oct 11 11:08:18 what distribution is that? Oct 11 11:12:01 should be angstrom 2011.x, according to this guide ->http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/ with slightly changed paths. uname -r gives me 3.0.4-yocto-standard+ Oct 11 11:17:15 lautriv: still don't see the problem.. here it works fine http://paste.pocoo.org/show/490954/ Oct 11 11:31:12 here you can see mine -> http://paste.pocoo.org/show/490958/ where obv. something is missing (...is installed and has the following files : ) Oct 11 11:36:56 that's because of sort Oct 11 11:38:06 ok, forgot that Oct 11 11:43:15 lol, yesterday i installed the nouveau-fb and since then i get randomly an acpi-event for "power-button pressed" :-) Oct 11 11:48:27 hello Oct 11 11:51:59 hello Oct 11 11:53:55 hiii Oct 11 12:40:18 Hi everybody Oct 11 12:40:32 I create my rootfs image by OE Oct 11 12:40:52 I dont use ipkg packages and I dont need them Oct 11 12:41:09 is it possible to prevent creating any packages such as ipkg? Oct 11 12:41:29 can I tell oe not to create any packages? Oct 11 12:51:41 ERROR: No recipes available for: Oct 11 12:51:41 /home/balister/src/git/oe-core/sources/meta-smartphone/meta-openmoko/recipes-graphics/mesa/mesa-dri_7.11.bbappend Oct 11 12:51:46 anyone seeing this? Oct 11 12:55:54 mgundes: OE uses the packages to build the rootfs Oct 11 12:56:14 mgundes: however, you can have all the packaging info removed at the end of building the rootfs Oct 11 13:03:01 bluelightning: you mean OE first create ipkgs and then install all ipkgs to rootfs directory and creates image? Oct 11 13:03:13 mgundes: that's correct Oct 11 13:03:30 But what about ${D} destination directories? Oct 11 13:03:46 this directory used for creating ipkg? Oct 11 13:03:49 hm Oct 11 13:04:23 I thought everything for rootfs directory are copied from desctination directories ${D} Oct 11 13:04:59 so that I thought I can prevent ipkg creation since it takes extra times Oct 11 13:05:37 mgundes: the way things work you can't really Oct 11 13:07:45 Is there any document about make build quickening? Oct 11 13:08:16 I want to learn how can I accelerate build issues If I can Oct 11 13:12:00 mgundes: AFAIK there's not such document, you have to read the python code Oct 11 13:13:56 you mean classes code? Oct 11 13:20:48 mgundes: yes Oct 11 13:23:18 ERROR: Nothing RPROVIDES 'util-linux-blkid' (but .../meta-oe/recipes-core/udev/udev_173.bb RDEPENDS on or otherwise requires it) Oct 11 13:24:17 Zoxc: http://patches.openembedded.org/patch/12969/ Oct 11 13:27:49 hmm, it does not appear that -DDDD gets passed through the scripts/bitbake wrapper Oct 11 13:36:07 I see that patch is in master-next. What is that used for and why doesn't that page say it's been applied there? Oct 11 13:38:45 JaMaOff, i guess your gcc-4.6 is from development, will it reach core next time ? Oct 11 13:46:13 Zoxc: master-next is just work queue for patches planed to be pushed to master later Oct 11 13:46:36 Zoxc: and RP who maintains oe-core doesn't use patchwork.. so he doesn't update status there AFAIK Oct 11 13:46:47 lautriv: it's from oe-core/meta-oe Oct 11 13:47:08 so meta-openembedded shouldn't depend upon things in master-next? Oct 11 13:48:49 no it shouldn't, this was accident Oct 11 14:02:04 mickey|office: pong (missed your ping from this morning, buried into work) Oct 11 14:16:20 is this patch correct? http://dl.dropbox.com/u/152908/0001-Add-a-native-hicolor-icon-theme.patch Oct 11 14:51:26 anyone built cmake for the target? Oct 11 14:51:39 I see a build failure Oct 11 14:53:06 suggestions? http://pastebin.com/9ZCkLkst Oct 11 14:53:51 urkx ugly... Oct 11 14:54:00 * florian dislikes cmake Oct 11 14:54:35 well, it has its pluses and minuses Oct 11 14:55:09 this is in oe-core, which is uspposed to work :) Oct 11 14:56:16 linux-libc-headers is in oe core as well... and currently fails Oct 11 14:56:48 * cbrake kind of likes cmake, but I've never cross compiled cmake itself Oct 11 14:56:50 ok, i admiot for different reasons Oct 11 14:57:34 * Jin^eLD really hates cmake :P Oct 11 14:58:15 Crofton|work: did you check these libraries? - maybe use of the wrong compiler or things like this Oct 11 14:58:37 I need to run it on the target Oct 11 15:00:48 we definitely need to improve the testing coverage Oct 11 15:36:53 eFfeM_work: hi. will you be at the GA? Oct 11 15:39:54 how can I use my "meta-toolchain" as my external toolchain? Oct 11 15:44:43 eFfeM_work, did anyone check back on you about the wiki? Likely the wiki people are still asleep though Oct 11 16:11:39 building oe-core bitbake core-image-minimal Oct 11 16:11:49 Unable to fetch URL http://kernel.org/pub/linux/kernel/v2.6/linux-2.6.37.2.tar.bz2 from any source.' failed Oct 11 16:13:00 mckoan: I guess you might want to consider enabling some mirrors Oct 11 16:13:22 the kernel mirrors are not yet up again, right? Oct 11 16:13:31 I had a similar problem recently and found out that kernel.org got hacked Oct 11 16:13:38 but I could not find any kernel tarballs via google :P Oct 11 16:13:40 i use KERNELORG_MIRROR = "http://mirror.nexcess.net/kernel.org" Oct 11 16:13:41 all links were broken Oct 11 16:13:46 KERNELORG_MIRROR = "http://mirrors.arsc.edu/kernel" Oct 11 16:13:48 works too though Oct 11 16:13:51 take your pick Oct 11 16:16:04 or use the angstrom/yocto mirrors Oct 11 16:16:20 indeed Oct 11 16:16:35 mtr: ich weiss nicht ob das nicht ich damals implemented hab? k.a. Oct 11 16:16:37 oops wrong window Oct 11 16:18:19 bluelightning: I am running on a fresh installed Ubuntu 10.04, sorry Oct 11 16:18:36 I am using oe-core as I found it Oct 11 16:18:46 right, which has no mirrors enabled.. Oct 11 16:18:59 I know I have a lot of fine-tuning to do since now Oct 11 16:19:08 I guess kernel.org are still verifying their tarballs Oct 11 16:22:16 BTW looks like is not a matter of mirror Oct 11 16:22:27 linux-2.6.37.2 does not exist Oct 11 16:23:37 looks like http://mirror.nexcess.net/kernel.org works Oct 11 16:23:51 kergoth_: "http://mirrors.arsc.edu/kernel" not Oct 11 16:24:28 * mckoan lets build continue Oct 11 16:24:36 have a nice rest of the day Oct 11 18:17:49 Jefro, no one replied to my projector question Oct 11 18:18:07 I think it would be nice, but not critical Oct 11 18:26:19 strange thing here, got this err ->I/O error : Attempt to load network entity http://docbook.sourceforge.net/release/xsl/current/manpages/docbook.xsl ...........I/O error points to hdd or net, both are fine and i can open that page in a browser ? Oct 11 18:30:20 lautriv: do you have docbook-xsl installed on your host or not ? Oct 11 18:30:24 if not then do install it Oct 11 18:32:37 Crofton|work: A whiteboard would do to Oct 11 18:33:32 khem, installing ... but should it not warn about "xyz missing" instead I/O error ? Oct 11 18:33:53 lautriv: its not in pre requisites Oct 11 19:57:20 ls Oct 11 20:03:33 Permission denied Oct 11 20:03:39 ;) Oct 11 20:30:40 where is the best place to look to see what nativesdk is doing? Oct 11 20:30:48 exit Oct 11 20:30:52 whoops Oct 11 21:01:34 who manages OE mailing lists? Oct 11 21:10:41 msm, what is the issue? Oct 11 21:12:57 Crofton: i was just going to suggest not striping out to: and cc: Oct 11 21:13:18 its so easy to lose track of messages when you are not on the to: Oct 11 21:13:27 although i could set up filters just for OE-core Oct 11 21:13:49 just a suggestion is all Oct 11 21:14:49 thanks Oct 11 22:43:28 is bitbake-runtask depreciated? Oct 11 22:43:53 or do we just use runqueue? Oct 11 22:44:51 hmm nevermind, i think i'm just lost =) Oct 11 23:10:00 msm: that was used when we were forking and execing each task rather than just forking Oct 11 23:27:19 kergoth: so not used then Oct 11 23:27:46 not as far as I'm aware. unless poky is doing something different there, but I don't believe so Oct 11 23:29:36 anyone got an idea how to insert -ltinfo into the linker flags for the cmake recipe **** ENDING LOGGING AT Wed Oct 12 02:59:57 2011