**** BEGIN LOGGING AT Mon Jun 07 02:59:56 2010 Jun 07 04:56:52 Android has unique developer signatures on applications .... does openmoko based software stacks or angstrom derivatives or even debian use a similar mechanism? Jun 07 04:57:35 kergoth can u help here Jun 07 05:02:13 mrmoku: can u help here Jun 07 05:13:14 shazkhan: do you want to try the whole list of channel members? Jun 07 05:15:27 yes Jun 07 05:15:52 how can I do that zecke? Jun 07 05:16:08 shazkhan: why do you believe that is appropriate? Jun 07 05:16:49 Maybe someone knows the answer ... and then all members can know about it too. Jun 07 05:17:29 I have to agree with him, very often IRC is more quiet than nntp or email.. Jun 07 05:17:36 still, that's not the right thing to do Jun 07 05:17:57 mailing list I guess ... but was in a hurry :( Jun 07 05:18:07 shazkhan: and what is the difference between "kergoth can u help here", "mrmoku: can u help here"? now, we have seen your request three times Jun 07 05:18:53 shazkhan: you come here, and insist people to help you. This is not how a free community is working... you can not insist on people helping you. You ask a question, it might get answered, it might not get answered. Jun 07 05:19:21 shazkhan: and regarding your question. So with Android the apk is signed? Or what is signed? Jun 07 05:20:15 No response should mean that nobody is watching the screen else someone will answer because I do when I know something. If the person is too rookie then I email him/her in details ... Jun 07 05:20:27 Anyways let me answer your question now ... Jun 07 05:23:10 zecke: the apk hash is signed by the developers private key and this digital signature is distributed along with the apk Jun 07 05:24:27 shazkhan: now the answer you have not been looking for... opkg and dpkg/apt support signing of packets... Jun 07 05:24:40 packets? Jun 07 05:24:57 shazkhan: .deb, .opk, .ipk... Jun 07 05:25:21 shazkhan: but there is only one pool of trusted keys... so you will not get the same protection as with android. Jun 07 05:25:22 u mean packages ... as far as I know the hash or package is considered for its integrity Jun 07 05:25:37 ah right.. packages is a better name. Jun 07 05:26:38 but that does not consider the signature of the developer ... issuer for trust purposes Jun 07 05:27:49 shazkhan: well, the distribution method for debian and angstrom is different. There is a repository with all the packages... and the dev sign the index/packages. Jun 07 05:28:38 u mean the index of packages by the developer ... thats as good as or I might say better than what Android scheme does Jun 07 05:29:14 the index is the repository of packages by a developer Jun 07 05:29:27 or the list of files in the package Jun 07 05:30:41 with debian/apt/dpkg one has a manifest like this Jun 07 05:30:42 http://ftp.debian.org/debian/dists/lenny/Release Jun 07 05:31:10 IIRC it is signed with one of the developer keys (from a pool of trusted debian developers) Jun 07 05:31:18 http://ftp.debian.org/debian/dists/lenny/Release.gpg Jun 07 05:33:04 hmm... is the git server down for everyone, or just me? Jun 07 05:33:17 got it thanx zecke ... my apologies for the inconvenience Jun 07 05:48:09 shazkhan: and the other part of the question. opkg supports the same kind of signature (as debian), so with angstrom you can sign a feed as well Jun 07 05:48:29 johncylee: ni hao ma? Jun 07 05:51:23 zecke: wo han hao Jun 07 05:52:57 johncylee: how was computex? Jun 07 05:53:26 hi,all, after I get pull my oe, it can't work , the error is "ERROR: Error in executing: /media/disk-1/stuff/openembedded/recipes/coreutils/coreutils-native_8.5.bb ERROR: Exception: Message:[Errno 2] No such file or directory" Jun 07 05:53:55 can anyone help me?tks a lot! Jun 07 05:53:56 william_z: we were there already. Have you made yourself more familiar with strace by now? Jun 07 05:54:30 zecke: show girls are not as good as I expected Jun 07 05:55:12 johncylee: yeah, I think they were imported from mainland... Jun 07 05:55:56 zecke, can you give some strace information or some webs about it? Jun 07 05:57:10 ok , I get it, tks zecke! Jun 07 05:58:28 william_z: well, I agree that the above error message is crap... so at least with spawning the process with strace... you will be able to findout which file was not found. Jun 07 05:59:42 yeah, I try to use strace now :) Jun 07 06:02:47 hm, is git down? trying to browse http://cgit.openembedded.org/ but I keep getting the waiting for message .... Jun 07 06:06:31 eFfeM_work: looks like this. Jun 07 07:02:32 good morning Jun 07 07:05:43 moin Jun 07 07:09:17 morning Jun 07 07:11:48 moin Jun 07 07:23:49 morning Jun 07 07:42:16 hi, problems on git.openembedded.org? Jun 07 07:43:58 yes, seems like the server is down Jun 07 08:00:39 morning all Jun 07 08:00:58 zecke: Sorry about the minutes, I've been away all weekend Jun 07 08:00:59 RP: hi Jun 07 08:01:15 RP: that is good! Work or a bike trip? Jun 07 08:01:28 zecke: Three days of trail riding :) Jun 07 08:01:43 hi rp Jun 07 08:02:47 RP: nice! Jun 07 08:02:59 need to search for some food. Jun 07 08:03:26 zecke: ack sent Jun 07 08:08:25 i'm getting the following note: Couldn't find shared library provider Jun 07 08:08:45 RP: committing mud theft again :-) Jun 07 08:09:02 when i look at the final image, that library is missing but executables that depend on that library are installed Jun 07 08:09:52 which library? Jun 07 08:10:47 XorA: Given the lack of rain, attempting that was very hard... Jun 07 08:11:01 RP: wow, it was really heavy rain here, sat/sun Jun 07 08:11:05 * RP has never seen the ground that hard Jun 07 08:11:23 XorA: It rained late yesterday afternoon just as I was arriving back at the campsite Jun 07 08:11:31 we managed to get our walking in sat morning before the storm hit luckilly and found the cave we searched for Jun 07 08:11:49 XorA: i've created a library for my own applications. library is built using a simple makefile, build objects, build shared object Jun 07 08:12:11 janp: make sure your library has a so name Jun 07 08:13:07 XorA: Sounds like fun and good timing :) Jun 07 08:13:25 RP: it was so hot we ended up just walking up the river back to car Jun 07 08:13:36 RP: then by 3pm it was insane rain Jun 07 08:13:47 RP: luckilly by then we were safely back here and on the rum :-D Jun 07 08:14:40 janp: -Wl,-soname -Wl,libintl.so for linking Jun 07 08:14:51 janp: with correct name for your library Jun 07 08:16:20 XorA: walking in a river plus heavy rain could have been interesting :) Jun 07 08:16:38 * RP has never seen the river crossings he took so low Jun 07 08:16:58 XorA: well, as far a I understand the linker, when DT_SONAME and the name of the library are the same, there shouldn't be a problem Jun 07 08:17:03 RP: yeah that really was flash flood causing rain Jun 07 08:17:19 janp: use objdump to check Jun 07 08:23:28 XorA: checked with objdump, name is indeed that same Jun 07 08:30:19 XorA: added soname to linker anyway. it at least makes the note about not finding a library provider go away Jun 07 08:30:45 janp: you will now need to rebuild everything that depends on it Jun 07 08:30:52 janp: then dependencies should be fixed Jun 07 08:48:32 XorA: in a makefile you always have to add the soname option for the linker? e.g $(LD) -shared -o $@ $^ $(LDFLAGS) -Wl,-soname -Wl,$@ Jun 07 08:49:27 XorA: but for the output file $@ is specified and for the soname option, the same $@ is used. So the name and DT_SONAME should be the same. Jun 07 08:49:49 XorA: does bitbake recognized this soname option somewhere? Jun 07 08:51:39 janp: no, OE uses objdump to retrieve the info and store it in a file which links which package to which so provider Jun 07 08:53:54 morning Jun 07 08:56:58 could somone pls. try building xfwm4-4.6.1? I run into some issues in do_install, created a patch that solves the issue for me, but I'd like to know if it's my local issue Jun 07 09:05:54 in fact, ehen when I get over this, I see this error further: Jun 07 09:06:07 > /mnt/data/zoom2/OE/tmp/sysroots/x86_64-linux/usr/bin/install: will not overwrite just-created `/mnt/data/zoom2/OE/tmp/work/armv7a-angstrom-linux-gnueabi/xfce-utils-4.6.1-r3/image/usr/bin/xfterm4' with `xfterm4' Jun 07 09:06:31 XorA: could it be that if -Wl,-soname is not given, the SONAME in the dynamic section doesn't exist and like you said, bitbake can't link the package with the so provider Jun 07 09:06:46 the cause being: sysroots/x86_64-linux/usr/bin/install is called with the same arg. multiple times (in this case xfterm4) Jun 07 09:07:51 where does sysroots/x86_64-linux/usr/bin/install come from? could it be the version was changed and this triggered the issue? perhaps it was silently ignored before? Jun 07 09:09:13 * XorA would love to know where his mouse pointer just disappeared to Jun 07 09:11:23 Xorg --nocursor *g* Jun 07 09:13:25 maybe the behaviour changed in coreutils 8.5 Jun 07 09:22:35 what's the recommended way to quote arguments given to os.system()? Putting them into '' or using pipes.quote()? Jun 07 09:30:44 hm, install X X dest fails even with coreutils 7.4, then I wonder how come xfce built before Jun 07 09:31:02 maybe another extra-odd-local-issue Jun 07 09:44:19 is git down? Jun 07 09:47:45 09:42:13 < ao2> hi, problems on git.openembedded.org? Jun 07 09:53:29 morning Jun 07 09:54:35 there Jun 07 10:25:46 hi jay7 Jun 07 10:25:50 hi jama Jun 07 10:27:09 XorA, thanks Jun 07 10:30:37 hi woglinde Jun 07 10:31:20 * XorA has no idea who admins the actual server git is on Jun 07 10:32:21 cbrake? Jun 07 10:45:07 http://wiki.openembedded.net/index.php/Infrastructure says melo Jun 07 10:46:02 admins: Mickey, Zecke, RP, cbrake, Rolf, rwhitby, OSUOSL staff Jun 07 10:49:13 yep. melo seems to be down Jun 07 10:49:43 anyone knows if that armdesigner.com's omap3530 kit is compatible/equivalent to the beagleboard? Jun 07 10:50:16 linf ask on #beagle Jun 07 10:50:26 tks Jun 07 10:51:25 XorA: BUG: soft lockup - CPU#0 stuck for 11s! [mysqldump:8039] Jun 07 10:53:57 XorA: try now please, I gave it a hard kick Jun 07 11:00:31 03Jason Kridner  07org.openembedded.dev * r6424ce6c7a 10openembedded.git/recipes/esc/esc-node-demo_git.bb: Jun 07 11:00:32 esc-node-demo: added simple Node.js demo for ESC Jun 07 11:00:32 This is for the Embedded Systems Conference training Jun 07 11:00:33 03Jason Kridner  07org.openembedded.dev * r1df846cc5f 10openembedded.git/recipes/esc-gst/ (6 files in 2 dirs): esc-gst: Added fixup for missing files Jun 07 11:00:35 03Koen Kooi  07org.openembedded.dev * r750b4c97ce 10openembedded.git/recipes/ (7 files in 3 dirs): ESC/2010: move all files into esc/ dir Jun 07 11:00:36 03Koen Kooi  07org.openembedded.dev * r39b527f3a5 10openembedded.git/recipes/esc-gst/esc-gst.bb: esc-gst: fix versioning Jun 07 11:00:37 03Koen Kooi  07org.openembedded.dev * r2ef4f1e00a 10openembedded.git/recipes/qt4/ (qt-4.6.2+4.7.0-beta1.inc qt4-x11-free_4.7.0-beta1.bb): qt 4.7.0-beta1: drop patches for affine and browser Jun 07 11:00:38 03Koen Kooi  07org.openembedded.dev * r144015f302 10openembedded.git/recipes/linux/linux-omap-psp-2.6.32/beagleboard/configs/cpuidle: linux-omap-psp 2.6.32: add cpuidle multi-config for beagleboard Jun 07 11:00:43 03Jason Kridner  07org.openembedded.dev * rd3525f8547 10openembedded.git/recipes/angstrom/ (2 files in 2 dirs): Jun 07 11:00:43 angstrom-uboot-scripts: Added an EDID read script Jun 07 11:00:43 This is just for example purposes so that it is documented how to read Jun 07 11:00:43 the EDID from within u-boot. Jun 07 11:00:44 03Koen Kooi  07org.openembedded.dev * rbdfaa71fc6 10openembedded.git/recipes/esc-media/esc-media.bb: esc-media: fix versioning Jun 07 11:00:46 03Koen Kooi  07org.openembedded.dev * r87ec7b6c9a 10openembedded.git/ (50 files in 3 dirs): linux-omap-psp 2.6.32: rebase patches onto new baseline Jun 07 11:00:50 03Jason Kridner  07org.openembedded.dev * r9cd504dd20 10openembedded.git/recipes/angstrom/ (6 files in 2 dirs): Jun 07 11:00:50 angstrom-uboot-scripts: Added various resolutions Jun 07 11:00:50 Also added a ramdisk configuration with mem= statements. Jun 07 11:14:23 is there any recommendation w.r.t. LC_ALL/LANG? Jun 07 11:15:08 for some reason gcc-cross-sdk-4.3.3-r12.2 fails for me in do_compile unless I run it it LC_ALL=C Jun 07 11:15:17 even en_US.UTF-8 breaks it Jun 07 11:15:54 Interestingly, gcc-cross-4.3.3 is not affected :-/ Jun 07 11:21:58 rwhitby: cheers Jun 07 11:44:05 Jay7: hello Jun 07 11:45:46 zub: Where did you get the sdk? Jun 07 11:46:47 zub: If you are interested in any sdk then try code-sorcery toolchains. Jun 07 11:46:56 cbrake: your nick was answer to question before (about git server admin) :) Jun 07 11:47:42 zub; u can also try http://www.angstrom-distribution.org/narcissus/ Jun 07 11:49:09 zub: If you are using one from oe builds than why not try the latest? Jun 07 11:52:07 well, I'd like the toolchain to contain some extra libs, so having an OE-built toolchain this could be easily done Jun 07 11:52:20 as for the version... I just kept it at whatever angstrom choses Jun 07 11:52:28 at least that's what I think :) Jun 07 11:52:47 having distro=angstrom-2008.1 in my local.conf Jun 07 11:53:01 looks like 2 yrs old, but I always assumed this is the current angstrom, is it not? Jun 07 11:57:28 * hrw -> lunch Jun 07 12:15:01 Jay7: oe git seems to be working fine here Jun 07 12:15:16 * cbrake wonders if someone just fixed it or if we have a networking problem Jun 07 12:16:00 ahh I see rwhitby's "I gave it a hard kick" -- that must have done it Jun 07 12:21:29 It seems that many thing in life respond well to a good solid kick. Jun 07 12:33:52 * MWelchUK_work feels like giving a computer a good kicking on a regular basis. Jun 07 12:34:30 usb powered auto-kicker should be good gadget :) Jun 07 12:35:54 I was thinking along the lines of giving them the type of kicking that makes them a lot quieter and cooler, though trying to calculate performance per watt results in a divide by zero error. Jun 07 12:38:22 MWelchUK_work, if you consider the as a limit, after that kind of kick you get infinite performance/watt ratio... even a lot more eco-friendly :) Jun 07 12:38:35 s/the/the 0/ Jun 07 12:39:18 ao2, Don't know, may still be pulling a bit of power... Jun 07 12:40:02 MWelchUK_work, right, forgive my naivety :) Jun 07 13:18:52 khem: http://paste.pocoo.org/show/222823/ Jun 07 13:24:31 hey JaMa|Wrk Jun 07 13:24:39 hey ant Jun 07 13:25:06 I've heard on irc gcc-4.5 has on real life some dryst. performance as gcc.4.3 Jun 07 13:25:19 this on omap Jun 07 13:25:46 ant_work: have you seen my benchmarks? Jun 07 13:25:57 yes, armv4t Jun 07 13:26:18 but no gcc-4.3 iirc Jun 07 13:26:50 JaMa|Wrk: I even linked your benches ;) Jun 07 13:27:03 yes I don't have old image to test it (and not in a mood to start another build with gcc-4.3 just for benchmark sake..) Jun 07 13:27:12 he he Jun 07 13:27:30 I'll do one soon, when khem says work is done Jun 07 13:27:39 ant_work: still fighting (again) with armv5te+gcc-4.5 build as in link just sent to khem Jun 07 13:27:49 yea :/ Jun 07 13:28:55 optspace was problem on armv4t, while on armv5te it worked fine.. but I started rebuild from scratch for spitz (just to get disalbed optspace too for spitz) and now it was failing.. Jun 07 14:05:02 having troubles building the latest org.openembedded.dev tree with grep-native, error message indicates that xz is missing but the new grep bb file from khem depends on xz-native so this should be installed Jun 07 14:06:22 janp: had the same issue (on 2nd host without xz already installed), but after adding xz-native and updated DEPENDS it worked as expected Jun 07 14:08:48 JaMa|Wrk: I believe the updated grep bb files is already there, as there is a DEPENDS += "xz-native" line in it, but somehow bitbake ignores it Jun 07 14:24:45 JaMa|Wrk: the problem you are seeing is bit strange Jun 07 14:24:58 JaMa|Wrk: I dont see it here on ubuntu host Jun 07 14:25:35 khem: but not seems like binutils version is not important only optspace triggers it Jun 07 14:25:38 JaMa|Wrk: problem seems to be that this symbol is far away and linker emits a veneer which is fine but messes the symbol attributes which is not fine Jun 07 14:26:26 JaMa|Wrk: optspace could be forcing linker to do it Jun 07 14:26:59 ah OK, now rebuilding again.. because libgcc1 is now just libgcc and libstdc++6 is just libstdc++ (not sure why...) Jun 07 14:27:20 JaMa|Wrk: debian shared library naming ? Jun 07 14:29:22 khem: wasn't changed.. those 2 builds differ only in OPTSPACE variable.. Jun 07 14:30:15 janp_away: you need to pull in latest master xz-native was missing which was added later Jun 07 14:31:18 khem: and yes.. using debian naming, distro is angstrom Jun 07 14:32:14 JaMa|Wrk: hmmm angstrom I used minimal here Jun 07 14:49:23 morning kergoth Jun 07 14:49:31 hey Jun 07 14:50:39 JaMa|Wrk: if you're getting libgcc rather than libgcc1 then either your distro has disabled debian naming, or something has happened to the package to cause the autonamer to stop thinking it is a library. see debian.bbclass. Jun 07 14:50:45 khem, are you sure that grep_2.6.3.bb DEPENDS is ok? BBCLASSEXTEND="native" automatically adds -native to DEPENDS, i suspect folks will end up with DEPENDS=xz-native-native Jun 07 14:52:16 khem, of course i didn't look nor try it, see someone mentioning xz earlier today :) Jun 07 14:53:52 pb_: the problem is that I built whole image with same distro, then changed binutils version on host and OPTSPACE param in gcc-configure-common.in and now I got lots off missing libgcc1 from do_rootfs Jun 07 14:54:10 pb_: so it's the same gcc recipe and same distro config Jun 07 14:54:39 pb_: so maybe objdump output changed.. Jun 07 14:58:16 * JaMa|Wrk -> home, bbl Jun 07 14:59:06 yeah, maybe. I'm not quite sure why debian.bbclass is using BUILD_PREFIX, that sounds rather like a bug. Jun 07 15:18:40 hm does some knows how to boot into the initramfs on ubuntu? Jun 07 15:28:29 khem, sorry for the late response. The autobuilder picked up your change, and gcc did indeed build correctly for the target. :) So thank you for fixing 4.2.4 for us! Jun 07 15:30:37 khem: To followup on the image testing I was able to do yesterday, the same image (SlugOS, NSLU2 BE and LE, with the same set of PREFERRED_VERSIONS other than gcc) had dissapointing results -- gcc 4.2.4 booted, but both gcc4.4.4 and 4.4.2, for both LE and BE, resulted in kernel panic when starting init. So something is very wrong. I'll work to debug that. Jun 07 15:31:26 Having a working 4.2.4 as a reference will make debugging the 4.4.x images much easier -- again, thank you for your efforts! Jun 07 15:32:26 (BTW, will work on extracting an NSLU2 that we can send you as a test device; all the ones with serial ports are in service so I have to arrange a swap-out for one of the upcoming weekends when I'm home. I'll keep you posted.) Jun 07 15:56:29 03Tom Rini  07org.openembedded.dev * r5e813f8a9e 10openembedded.git/ (3 files in 2 dirs): (log message trimmed) Jun 07 15:56:29 rootfs_ipk/meta-toolchain: Add locking on tasks that add package_update_index_ipk Jun 07 15:56:29 We add a lockfile of do_populate_sdk and do_rootfs (in rootfs_ipk) that is also Jun 07 15:56:29 the lockfile that package_update_index_ipk uses. With enough threads it is Jun 07 15:56:29 possible that one image (or meta-toolchain) is attempting to use the package Jun 07 15:56:29 index while another has only reached the point of generating the index leading Jun 07 15:56:30 to an empty index being seen later on. Jun 07 16:08:29 khem, ping Jun 07 16:08:38 hm anyone knows if ext3 mounts itslef into ro? Jun 07 16:09:02 I have here a strange problem where my rootpartition is always mounted ro Jun 07 16:09:52 woglinde, IIRC it does, yes Jun 07 16:10:15 Er, by default root is ro, yes? Jun 07 16:10:25 unless you pass rw on the cmdline Jun 07 16:10:36 tartarus yes I did this Jun 07 16:10:47 Then it should be rw Jun 07 16:10:48 I even forced the initramfs to mount rw Jun 07 16:10:53 still ro Jun 07 16:10:55 tune2fs show anything? Jun 07 16:11:29 mh which option? Jun 07 16:11:36 -l Jun 07 16:11:45 afterb booting I can remount it with rw all is fine Jun 07 16:11:58 ok, so the fs itself is fine Jun 07 16:12:08 not quite sure then, sorry Jun 07 16:12:12 filesystem state is clean Jun 07 16:12:45 .../coreutils-native_8.5.bb, do_install) failed with 256. Does that sound familiar to somebody or did i goof something up locally? Jun 07 16:13:20 blindvt: zub had the same issue IIRC Jun 07 16:33:53 03Cliff Brake  07org.openembedded.dev * rdd61383cc9 10openembedded.git/classes/packaged-staging.bbclass: (log message trimmed) Jun 07 16:33:53 packaged-staging.bbclass: fix issue with no SRC_URI Jun 07 16:33:53 Currently with srctree, we get errors like the following: Jun 07 16:33:53 | + srcuri= Jun 07 16:33:53 | + '[' '' == '' ']' Jun 07 16:33:54 | + srcuri=OpenEmbedded Jun 07 16:33:54 | + echo 'Source: ' Jun 07 16:47:14 mwester: nice to know. Jun 07 16:47:35 mwester: w.r.t gcc 4.4 issue what binutils version do you use with it Jun 07 16:47:57 hey likewise Jun 07 16:52:17 khem: btw why there is no bitbake-native and gcc-native (for newer gcc versions)? maybe it would make builds more reproducible between different build hosts? Jun 07 16:52:43 khem, you're a coward. If you would have picked the .tar.xz for coreutils somebody would have had to fixup unpacker dep handling ;P Jun 07 16:53:22 blindvt: yeah pandora of bugs Jun 07 16:53:44 JaMa: it would take quite a bit to build host env then :) Jun 07 16:54:10 JaMa: while in pure sense its ok but practically there is no end to having host packages Jun 07 16:54:26 you might end up rebuilding host packages from scratch Jun 07 16:54:39 I was just curious because of this last gentoo-binutils issue.. Jun 07 16:55:03 JaMa: I think you should report this issue to gentoo Jun 07 16:55:15 they should fix it Jun 07 16:55:23 and I'm already trying to minimize build host influence by using minimal chroot Jun 07 16:55:30 regarding the missing symbol I am still perplexed Jun 07 16:55:41 as I am not able to reproduce it here. Jun 07 16:55:43 khem: I'm waiting for at least ant's confirmation Jun 07 16:55:52 khem: armv5te-linux-gnueabi/binutils-cross-2.18.50.0.7-r10.1 Jun 07 16:56:04 JaMa: is that what you are using ? Jun 07 16:56:19 khem, i've a tree somewhere (it's in my virtualize-coreutils tree that uses busybox instead of the coreutils bloat) that handled this properly. I'll get to factoring it out someday Jun 07 16:56:24 because on host I'm using not-yet-supported gcc and maybe my bug should be confirmed with supported 4.4.4 on host first Jun 07 16:56:31 between 2.18 and 2.19 the arm thumb long call interworking was horrible Jun 07 16:56:36 it was under work Jun 07 16:57:09 blindvt: sounds cool. Jun 07 16:58:08 mwester: first think I would suggest it to use newer binutils like 2.20.1 Jun 07 16:58:23 khem: mwester's binutils version? no 2.20.1 (as set from sane-toolchain) Jun 07 16:58:25 ok. I'm trying to find where that version is set. Jun 07 16:58:53 JaMa: ok I mistook the reply (your nick and mwester's nick share same color on my screen :) Jun 07 16:59:11 mwester: in distro.conf Jun 07 16:59:20 Found it. thanks. Jun 07 16:59:22 mwester: your distro is slugos isnt it Jun 07 16:59:26 2.20 best, you think? Jun 07 16:59:33 2.20.1 Jun 07 16:59:40 it has bug fixes on top of 2.20 Jun 07 17:00:03 mwester: and whats your glibc version ? Jun 07 17:00:06 Ok. I can at least build it, but it'll have to wait 10 days or more until I get back here to test the image. (This job is killing me) Jun 07 17:00:12 2.22.1 Jun 07 17:00:19 kergoth, could we potentially get away with not parsing all .bb? If i assume that a specific version of a recipe is requested, could i spare parsing the "not interresting" recipes for that package, perhaps? Jun 07 17:00:20 er, 2.6.1 Jun 07 17:00:58 mwester: hmmm could you use 2.9 although 2.6.1 is not bad but to try it out Jun 07 17:01:03 and switch to eglibc Jun 07 17:01:06 ~tell kergoth, could we potentially get away with not parsing all .bb? If i assume that a specific version of a recipe is requested, could i spare parsing the "not interresting" recipes for that package, perhaps? Jun 07 17:01:20 eglibc? Hmmm.... ok. Jun 07 17:01:24 * mwester takes some notes... Jun 07 17:01:29 mwester: yes thats the best suited Jun 07 17:01:39 and you could choose 2.12 Jun 07 17:01:47 ibot, renitent bot Jun 07 17:01:49 now when you are upgrading then do it properly :) Jun 07 17:01:59 blindvt: that wud be really nice Jun 07 17:02:17 blindvt: it wud speed up parsing Jun 07 17:02:31 khem, i didn't think it through yet, but it would be nice, yea Jun 07 17:02:42 blindvt: whenever I add a new recipe a thought comes oh it will slow parsing by a bit Jun 07 17:03:42 khem, well, bitbake is pretty fast as it is iff you're lucky to have fasts disks and alot of RAM, but doing less work is always good, of course Jun 07 17:03:57 mwester: I am giving you combinations that works on arm for many folks here Jun 07 17:04:00 with gcc 4.4 Jun 07 17:04:18 That would be good. Jun 07 17:04:22 blindvt: yes its fast true I still have old era machines Jun 07 17:04:36 (clearly one cannot just pick and choose...) Jun 07 17:04:40 gcc 4.4.4 + binutils 2.20.1 + eglibc 2.12 Jun 07 17:05:01 well you could choose eglibc 2.10 2.11 as well if you want Jun 07 17:05:28 ok Jun 07 17:05:33 * khem still envies mwesters 12G RAM Jun 07 17:05:52 and under 45 min build Jun 07 17:06:37 khem, i always have old machines that are utterly outdated. That's why i like small and fast ;) Jun 07 17:07:29 khem, mwester: why not just include sane-toolchain.inc to slugos.conf? Jun 07 17:07:30 khem, but they are a PITA to develop on. oh well Jun 07 17:12:46 JaMa: a good idea Jun 07 17:13:17 khem: btw debian naming issue: DEBUG: LIBNAMES: pkg libstdc++ libs 1 bins 0 sonames [] Jun 07 17:13:27 khem: I guess that before there was some soname Jun 07 17:13:42 JaMa: hmm Jun 07 17:14:13 JaMa: I wonder if there is issue in naming class Jun 07 17:14:39 JaMa: can you upload the faulty and good libstdc++.so.6 somewhere ? Jun 07 17:15:25 JaMa: another thing you could do is that in binutils-cross.inc comment out --enable-shared Jun 07 17:15:34 then rebuild binutils-cross Jun 07 17:15:53 and gcc-cross or gcc from whereever libstdc++ is being emitted Jun 07 17:15:57 as package Jun 07 17:16:09 gcc-cross Jun 07 17:16:23 ok then you can just rebuild gcc-cross Jun 07 17:16:31 and binutils-cross Jun 07 17:16:43 I'll, thanks for hints and I can upload only faulty, because I'm trying to build again good one for whole afternoon :) Jun 07 17:16:46 JaMa: for me with or without optspace it works well Jun 07 17:16:52 khem, another thing JaMa could do is make shared support a machine feature where it belongs and derive --enable-shared/static from the machine, no? :) Jun 07 17:17:27 khem, s/machine/distro/ ? Jun 07 17:17:33 blindvt: well actually on some architectures where PIC makes difference we need it --enable-shared Jun 07 17:17:42 othereise binutils we generate are wrong Jun 07 17:17:45 they wont work on target Jun 07 17:18:00 or infact any binary that will link to libiberty.a Jun 07 17:18:03 will be wrong Jun 07 17:18:03 khem: ah faulty as symbol issue.. yeah I can upload those easily, but it will take a while.. Jun 07 17:18:13 JaMa: ok Jun 07 17:18:36 JaMa: in faulty case linker is emitting a veneer for that symbol for some reason **** BEGIN LOGGING AT Mon Jun 07 18:28:14 2010 Jun 07 18:29:14 kergoth: ok thx Jun 07 18:29:17 khem: but maybe target cross dir is not in PATH Jun 07 18:29:33 * kergoth expects not many people are using that class, but it implements per-subpackage -dbg packages Jun 07 18:29:39 blindvt please try git? Jun 07 18:29:51 ka6sox, works ATM Jun 07 18:30:25 thanks Jun 07 18:30:59 kergoth: what do you mean by subpackage is it less than PACKAGES ? Jun 07 18:31:05 ? Jun 07 18:31:21 i mean each binary package listed in PACKAGES gets its own -dbg package, with package_dbg.bbclass Jun 07 18:31:27 rather than just ${PN}-dbg Jun 07 18:31:45 yeah got it now Jun 07 18:32:04 for every ipk there is a corresponding _dbg.ipk Jun 07 18:32:26 forget doc static and dev Jun 07 18:32:29 yup Jun 07 18:35:22 hi I bet netbase is core Jun 07 18:36:34 JaMa: kergoth still it does not land dbg packages into the image hmm I guess I will try to debug it a bit Jun 07 18:38:55 kergoth: the bitbake clone for committers is different url Jun 07 18:39:16 I'll send a patch Jun 07 18:40:15 khem: this works, but looks awfull: cmd = bb.data.getVar('CROSS_DIR', d, 1) + "/" + bb.data.getVar('bindir_cross', d, 1) + "/" + (bb.data.getVar('TARGET_PREFIX', d, 1) or "") + "objdump -p " + fp Jun 07 18:42:42 mmt food, bbl Jun 07 18:42:57 JaMa: could use TOOLCHAIN_PATH Jun 07 18:45:01 hmmm but then I wonder why cross compilers are not in path Jun 07 18:45:33 kergoth, zecke said yesterday that he added my ssh key to let me write to a personal branch of bitbake to journal back my little patches. What's the clone url i should be using? Jun 07 18:45:45 same as for OE Jun 07 18:45:54 git@git.openembedded.org: Jun 07 18:46:53 if these sanity check builds go fine, i intend to push the patches you've sent to the list, along with rpjay's pending bitbake patches and a couple others, today Jun 07 18:46:54 kergoth, mhm. That denies my publickey ? Jun 07 18:47:02 then it isn't set up right Jun 07 18:47:08 * kergoth goes to take the dog out Jun 07 18:47:45 * blindvt continues to route stuff through the list then *shrug* Jun 07 18:48:03 i think i have admin access to the gitosis setup, i'll take a look when i get back Jun 07 18:48:50 kergoth, TIA. take your time Jun 07 18:56:20 erk, -k world = 70,403 tasks Jun 07 18:57:34 kergoth: you will need all the time in world to bitbake those :) Jun 07 18:57:56 i used to do a world build a couple times a month, but that was many years ago :) Jun 07 18:58:19 kergoth: heh when world was small :) Jun 07 18:58:30 only a couple thousand recipes at that time i think Jun 07 18:58:46 kergoth: world will build every recipe isnt it Jun 07 18:59:08 so say if I have 10 versions of gcc will it try to build and install all 10 ? Jun 07 18:59:12 pretty much. if there's something provided by multiple recipes, it only builds the preferred one, and it only builds the preferred version of everything Jun 07 18:59:12 nope Jun 07 18:59:15 I mean cross gcc Jun 07 18:59:26 oh ok Jun 07 18:59:26 it'll only build one gcc, one cross gcc, etc Jun 07 18:59:43 so one of each Jun 07 19:01:32 kergoth: how about libcs Jun 07 19:01:36 blindvt: okay, pushed most everything that's gone out Jun 07 19:01:41 khem: again, it builds one of each provider Jun 07 19:01:56 and nothing that is EXCLUDE_FROM_WORLD Jun 07 19:02:00 (unless it comes in via deps) Jun 07 19:02:05 so if I choose a DISTRO and a MACHINE then its set Jun 07 19:02:14 should be, yeah Jun 07 19:02:38 course, it takes a while to generate teh runqueue, and there are lots o fthings that get left out due to dependency problems, but those are busted recipes that should be fixed at some point too, so even that is useful information Jun 07 19:02:39 * khem is contemplating a world build on the weakest machine in room Jun 07 19:03:29 masochist ;) Jun 07 19:03:56 * kergoth has to run a world build to try to spot missing var refs for his variable reference tracking code :| Jun 07 19:04:14 khem, did you find somthing WRT baking something asume_provided, btw? Jun 07 19:04:15 blindvt: just pushed most of your stuff that's hit the list to master Jun 07 19:04:18 heh then you have to do the heavylifting of revival Jun 07 19:04:26 kergoth, thanks alot! Jun 07 19:04:31 with most of rpjay's stuff and a modified version of cbrake's exception fix, as well Jun 07 19:04:49 blindvt: no other than bitbake hung Jun 07 19:06:08 blindvt: what's the address you used in your public key? Jun 07 19:06:13 khem, nod. woglinde said that current master is much slower than 1.8.18 if he tries bitbake somethingthatdoesnotexist Jun 07 19:06:38 blindvt hm not anymore Jun 07 19:06:40 oh I did not see that yet Jun 07 19:06:58 hi woglinde Jun 07 19:10:02 kergoth, i sent him my ssh id_dsa.pub, there's no addr assigned to that AFAIR, isn't it? Jun 07 19:10:11 yes, there is Jun 07 19:10:14 cat the .pub Jun 07 19:17:27 kergoth, that addr is optional. md5sum of my pubkey is 47dc0bfe570276fc76df87aa5018dde1, but let me mail it to you Jun 07 19:19:06 your key isn't in the repository Jun 07 19:19:14 so i guess zecke forgot to push or something Jun 07 19:19:25 . Jun 07 19:22:17 khem: TOOLCHAIN_PATH is only CROSS_DIR without /bin, isn't it? Jun 07 19:25:21 sat top boxes have JaMa yes Jun 07 19:25:32 ah man I must restart this KDE Jun 07 19:25:43 :) Jun 07 19:25:58 its paste/klipper has gone insane Jun 07 19:26:00 today Jun 07 19:26:11 upgrade to 4.4.4? Jun 07 19:26:16 JaMa: and why it does not appear in path I thought it must be in path Jun 07 19:26:23 I am on 4.4.4 Jun 07 19:26:37 I wonder if I should try 4.5 beta now Jun 07 19:26:47 and strangle myself Jun 07 19:28:37 blindvt: gitosis doesn't provide branch level access control. its all or nothing, so just don't push master :) Jun 07 19:28:55 there is an alternative to gitosis that provides it, but we aren't using it yet Jun 07 19:28:58 khem: it is in bitbake.conf PATH and I don't see debian.bbclass changing it.. but still without calling it without path it fails and no output is written Jun 07 19:29:33 kergoth: btw: that's name of that gitosis alternative? Jun 07 19:29:54 JaMa: http://github.com/sitaramc/gitolite Jun 07 19:30:16 gitolite? Jun 07 19:30:18 ah right :) Jun 07 19:30:22 :) Jun 07 19:31:00 * JaMa was searching for this blog post.. http://www.bitwiz.org.uk/s/2010/01/gitolite.html Jun 07 19:31:31 blindvt: hmm, do you not have write access to OE? or did oyu switch keys? Jun 07 19:32:36 kergoth, Would be handy but i don't have write-access to OE yet, no Jun 07 19:34:02 i'd give it to you, but i think we're supposed to do that whole proposal on the list thing ) Jun 07 19:34:03 kergoth, and i will not push to any master for sure Jun 07 19:34:08 k Jun 07 19:35:16 blindvt: okay, should work now Jun 07 19:36:52 kergoth, i'll continue to route eventual OE stuff through khem then. I'm trying to refrain from doing broader stuff ATM anyway until i have my base covered. Tricky at times but i hold off as much as i can for now Jun 07 19:36:58 k Jun 07 19:37:26 khem: would you mind proposing oe write access for him on the list for me, if you have a chance sometime in the near future Jun 07 19:37:29 bit swamped atm Jun 07 19:37:38 at least, my brain feels that way, even if my task list isn't Jun 07 19:37:39 hrmph Jun 07 19:37:42 kergoth, bb access works fine. Thanks! Jun 07 19:37:46 great, np Jun 07 19:38:27 * kergoth thinks about adding a slightly tweaked pylintrc to the bitbake repo Jun 07 19:38:57 kergoth, OE seems to use $nick/$topic for branches, is that ok for you, too? Jun 07 19:39:09 fine with me Jun 07 19:39:13 k Jun 07 19:39:15 personally, i have my personal branches on github Jun 07 19:39:24 just to avoid issues with not having access to rebase or remove them Jun 07 19:39:34 but when i had them her,e it was $nick/$topic Jun 07 19:40:16 khem: more about ____cxa_end_cleanup_veneer.. disabled OPTSPACE, gcc-cross seems fine (at least before stripping) only target gcc seems wrong (rebuilding version with OPTSPACE now), but because target gcc overwrites staged version from gcc-cross, then I have problem with building other recipes only after gcc was built (ie for eet) Jun 07 19:40:47 khem: both builddirs will be available here http://jama.homelinux.org/org.openembedded.dev.images/ now only cross+OPTSPACE and target+no-OPTSPACE Jun 07 19:41:23 khem, so what about OE's conf/bitbake.conf's BUILD_{ARCH,OS} pinning that saves a couple of ten-thousand uname() calls? please install :) Jun 07 19:45:21 khem, my paste already timed out, but fortunately kergoth put it into the queue as http://patchwork.openembedded.org/patch/2111/ and i'd happily ACK this one (and the rest of those 4 are also ok, from the looks) Jun 07 19:50:33 JaMa: hmmm interesting. I have staged gcc here but did not have that problem Jun 07 19:51:43 khem: moving builddir from target gcc with optspace enabled Jun 07 19:54:04 JaMa: blindvt yeah those patches seem ok. I will ack them Jun 07 19:55:10 kergoth: ERROR: QA Issue with kernel-modules: kernel-modules rdepends on kernel-dev-dbg Jun 07 19:55:24 ah, right, i forgot to push the fix for that Jun 07 19:55:33 its because it pulls in everything from PACKAGES minus certain things Jun 07 19:55:36 after I enable package_dbg Jun 07 19:55:39 so when package_dbg adds mre to it.. Jun 07 19:56:02 its easy to fix, just make it exclude ones that end with -dbg Jun 07 19:56:39 in the recipe ? Jun 07 19:56:58 i don't recall exactly where, probably kernel.bbclass or linux.inc Jun 07 19:58:03 must be in kernel.bbclass Jun 07 19:58:41 bb.data.setVar('RDEPENDS_' + pkg, ' '.join(rdepends), d) Jun 07 20:29:40 JaMa: problem happens when you have target gcc compiled with OPTSPACE disabled right Jun 07 20:30:54 khem: yes Jun 07 20:31:25 JaMa: hmmm Jun 07 20:31:35 let me look at the build logs of failing case Jun 07 20:32:58 kergoth: We've found a nice bitbake bug in poky Jun 07 20:33:02 fun Jun 07 20:33:18 kergoth: bb.decodeurl passes the path in the host variable for file:// urls Jun 07 20:33:24 and base.bbclass uses it Jun 07 20:33:28 JaMa: I can not access homelinux.org :( its firewalled Jun 07 20:33:32 khem: ah already removed, but mmt I'll rebuild Jun 07 20:33:40 kergoth: We can fix bitbake but break OE Jun 07 20:34:00 JaMa: logs are important if you want to refer keep them for a while Jun 07 20:34:24 that's pretty much what i'd expect that code to do. Jun 07 20:34:31 we should really think about using urlparse, rather than our own thing Jun 07 20:34:55 khem: because of lack of disk space I moved only builddir and removed rest for another rebuild... Jun 07 20:35:16 kergoth: well, yes. I took a fix for this, then watched the builds explode :/ Jun 07 20:35:18 khem: dns is firewalled or apache doesn't respond to you? Jun 07 20:35:52 JaMa: no idea Jun 07 20:35:53 i wouldn't expect it to be terrible to work around. couldn't base.bbclass just check for host set path unset, path unset host set, and fall back? Jun 07 20:35:57 khem: 78.24.13.161 Jun 07 20:36:48 kergoth: I think we'llhave to Jun 07 20:37:20 that's what i'd think would be best, at least until OE requires a release version with the fix Jun 07 20:37:30 kergoth: agreed Jun 07 20:38:28 khem: and to be sure we understand each other, the problem happens when OPTSPACE (as variable) is empty so when --enable-target-optspace is not used Jun 07 20:40:00 khem: and dirname no-optspace means empty OPTSPACE variable = non-working case Jun 07 20:40:37 what bitbake version is recommended to use with .dev now? Jun 07 20:40:39 1.10? Jun 07 20:41:52 Jay7: master Jun 07 20:42:10 I fear :) Jun 07 20:42:12 JaMa: hmmm .lib dir is missing which has the libsdtc++ binary Jun 07 20:42:33 Jay7: no need to fear its better than previous releases and it stable Jun 07 20:42:48 and more we use it more we will find breakges if it happens early enough Jun 07 20:42:53 khem: arm-angstrom-linux-gnueabi/libstdc++-v3/src/.libs/libstdc++.so.6.0.14 ? Jun 07 20:43:06 JaMa: yes Jun 07 20:43:13 hrr apache hiddes it.. Jun 07 20:43:20 well.. let's try :) Jun 07 20:43:35 JaMa: oh I just dont see it if I type it in url it opens Jun 07 20:43:46 khem: but only from index.. Jun 07 20:43:49 Jay7: you wont be disappointed Jun 07 20:47:02 well.. first issue - chrpath is not installed :) Jun 07 20:47:14 * Jay7 going to add this to wiki Jun 07 20:49:13 Jay7: thats OE wanting it not bitbake Jun 07 20:49:21 khem: I know.. Jun 07 20:49:22 Jay7: but still it told you what it wants Jun 07 20:49:40 rathar bombing on you Jun 07 20:49:45 somewhere else Jun 07 20:49:54 hm.. Jun 07 20:50:03 now I can't just start bitbake -i Jun 07 20:50:09 JaMa: interesting part is that there is only one veneer generated Jun 07 20:50:31 it wants package name Jun 07 20:51:27 JaMa: your upload link is slow. I dont know how long it will take to download stuff Jun 07 20:56:13 khem: yeah its only 12/2Mbps..with 1:20 agregation.. Jun 07 20:59:20 JaMa: is it gcc 4.5 ? Jun 07 20:59:30 yes Jun 07 21:00:39 the symbol should be in the same libstdc++.so but it does not find it and hence creates a external call Jun 07 21:00:47 so where does it get when linking Jun 07 21:01:42 should I call nm for every .so and try to grep it somewhere? Jun 07 21:02:16 JaMa: I think its there in libstdc++.so thats installed by gcc-cross Jun 07 21:02:26 and somehow that creeps in to link step Jun 07 21:02:30 thats my guess Jun 07 21:04:56 libstdc++-v3/libsupc++/eh_arm.cc defines it Jun 07 21:09:16 JaMa: hmm your version of eh_arm.o does not have it Jun 07 21:09:24 thats the problem Jun 07 21:13:25 JaMa: if you have the faulty build tree around then preprocess eh_arm.cc Jun 07 21:13:30 and lets see what it produces Jun 07 21:15:00 ok, mmt Jun 07 21:24:10 khem: preprocess http://paste.pocoo.org/show/223024/ Jun 07 21:24:48 JaMa: ok now can you also generate a .s file Jun 07 21:24:52 and paste it somewhere Jun 07 21:24:58 and also the commandline options plz Jun 07 21:25:19 khem: btw logs uploaded to temp subdir Jun 07 21:25:50 khem: here is cmdline http://paste.pocoo.org/show/223027/ Jun 07 21:26:58 ok so it used -O2 Jun 07 21:28:50 JaMa: ok I got it Jun 07 21:30:26 JaMa: it seems to do the right thing hmmm Jun 07 21:30:29 let me try once more Jun 07 21:30:35 khem: http://paste.pocoo.org/show/223029/ Jun 07 21:30:43 khem: with -S it shows few warnings Jun 07 21:32:58 JaMa: here is fix for you :) Jun 07 21:33:04 I think I see the problem Jun 07 21:33:10 and I see the reason also Jun 07 21:37:10 JaMa: try this patch http://pastebin.com/qpnN5vf9 Jun 07 21:37:18 to gcc-4.5.inc Jun 07 21:37:28 and rebuild gcc Jun 07 21:37:41 then rebuild eet or whatever was failing for you Jun 07 21:37:49 and report back here Jun 07 21:41:22 khem: nm output should be the same right? only now it should find that symbol Jun 07 21:55:13 JaMa: yes now nm should show you __cxa_end_cleanup as global Jun 07 21:57:11 hmm pastebin messed white-spaces somehow and it refused to apply, applied manually and restarting build Jun 07 22:01:41 JaMa: hmmm thats after I disabled klipper Jun 07 22:01:54 something is messed up with my pasting Jun 07 22:02:05 anyway I am out for few hours Jun 07 22:02:15 send me a note here how your compile went Jun 07 22:02:20 and i will read it later Jun 07 22:04:35 khem: ok, thanks a lot again, you're great! Jun 07 22:16:03 khem: xinput-calibrator rebuild went fine, __cxa_begin_cleanup is normal symbol now and no ____cxa_begin_cleanup_veneer shown in -nm Jun 07 22:19:35 qemu-native-0.12.4-r1 failed Jun 07 22:20:17 http://tinderbox.openembedded.org/builds/73262/ Jun 07 22:20:52 i586-alt-linux-gcc: Internal error: Killed (program cc1) Jun 07 22:20:54 hehe.. Jun 07 22:22:06 well.. will look tomorrow Jun 07 23:38:25 03Henning Heinold  07org.openembedded.dev * rf6b09dcb60 10openembedded.git/recipes/xorg-proto/ (14 files): xproto: convert to BBCLASSEXTEND Jun 07 23:38:35 03Henning Heinold  07org.openembedded.dev * re536b582cb 10openembedded.git/recipes/xorg-proto/xorg-proto-common.inc: xorg-proto-common.inc: remove do_stage function Jun 07 23:38:36 03Henning Heinold  07org.openembedded.dev * rcda3842850 10openembedded.git/recipes/cacao/cacao-native_hg.bb: cacao-native: add mercurial recipe Jun 08 00:09:48 Jay7: thats your host compiler. Which version is it and what distro do you use Jun 08 00:35:57 khem was my xproto commit okay? Jun 08 00:40:26 good nite Jun 08 00:40:27 exit Jun 08 02:03:22 hi Jun 08 02:09:42 I have a problem when I cross compile the GTK+ with OE,it always display " Error in executing: /home/stuff/openembedded/recipes/coreutils/coreutils-native_8.5.bb" and "coreutils-native_8.5.bb do_populate_sysroot failed",and yesterday the Error of coreutils-native version is 7.2 Jun 08 02:12:41 this have trobled me many days ,somebody help ? Jun 08 02:31:38 ~pastebin Jun 08 02:32:00 no CIA? bah Jun 08 02:32:02 [~pastebin] A "pastebin" is a web-based service where you should paste anything over 3 lines so you don't flood the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://bin.cakephp.org/ , http://asterisk.pastey.net/ , or install pastebinit with yum or aptitude. Jun 08 02:32:09 there we go Jun 08 02:32:20 joe123, use that and show us the contents of the log.do_populate_sysroot Jun 08 02:44:56 thanks ,i have submit the error on http://www.pastebin.com Jun 08 02:56:05 what is the url it posted it to? Jun 08 02:56:24 when you paste something, it puts it on its own page and you tell us that :) Jun 08 02:59:50 03Khem Raj  07org.openembedded.dev * rbf1756a16e 10openembedded.git/recipes/gcc/gcc-package-cross.inc: (log message trimmed) Jun 08 02:59:50 gcc-cross: Don't stage libiberty.a in sysroot. Jun 08 02:59:50 * We already stage libiberty from target binutils Jun 08 02:59:50 problem is that when we stage it from gcc-cros Jun 08 02:59:50 it conflicts with the ones installed in cross dir Jun 08 02:59:51 by binutils-cross which should be used to link Jun 08 02:59:51 binutils for target but it picks up the one **** ENDING LOGGING AT Tue Jun 08 02:59:56 2010