**** BEGIN LOGGING AT Thu Jun 27 02:59:58 2013 Jun 27 03:39:19 trollixx: that path should *not* be hard-coded in a shared lib Jun 27 03:39:47 meaning it shouldn't be your job to fix it either... Jun 27 03:41:05 which would most likely be finding the right linker option and correcting it Jun 27 03:42:28 you might be able to get away with setting LD to the cross g++ instead of of ld... Jun 27 07:35:39 Hello, i'm trying to buid a image to my gumstix Overo, i followed the how-to and documentation, but now, then i run the bitbake a have this error: Pseudo is not present but is required, building this first before the main build - No such file or directory: '/home/Overo/poky-dylan-9.0.0/bitbake/conf/layer.conf' Jun 27 08:17:15 hi guys, im getting stuck on systemd beeing need from meta-oe package, can anyone help me plz ? Jun 27 08:17:47 I get a "Could not inherit file classes/systemd.bbclass" Jun 27 08:18:43 but the systemd.bbclass isn't include in meta-oe bsps Jun 27 08:18:56 how to fix it ? Jun 27 08:20:17 mks: it is in oe-core now (AFAIK) Jun 27 08:20:44 mks: if your not using dylan/master you might want to use a different branch of meta-oe Jun 27 08:21:05 zecke: isn't oe-core include in poky ? Jun 27 08:21:37 mks: yes it is. but systemd.bbclass is only there starting from dylan Jun 27 08:21:55 zecke: I am on danny branch for all meta's Jun 27 08:22:46 zecke: so I should work on dylan on the meta-oe and let poky on danny ? Jun 27 08:24:53 mks: in meta-oe/meta-systemd/classes/ you should have a systemd.bbclass? so if bitbake doesn't find systemd.bbclass Jun 27 08:25:00 your config might not be right. :) Jun 27 08:26:53 zecke: no I don't have it, even after a git checkout master and git pull Jun 27 08:27:30 zecke: I am not sure I understood why my config isn't right Jun 27 08:29:45 hi Jun 27 08:30:48 i'm having trouble with A centos 6.4 build, its complainaing about linux/limits.h. I'm sure i came across this before, but i cannot find any docs on it, or at least my googlefoo has not been up to it. can anyone help? Jun 27 08:42:09 mks: poky/meta/classes/systemd.bbclass exists here in the dylan branch Jun 27 08:42:21 it moved from meta-oe to poky recently Jun 27 08:42:27 (as in, a few months ago) Jun 27 08:42:55 I have an old package laying around in my image.. I have renamed the recipe for this package, and the files of the "old package" now conflicts with the files of the "new package", how can I make yocto forget about the old package? It still attempts to install its RPM. Jun 27 08:43:15 Sput: yeah that's what I just figured out Jun 27 08:44:40 Sput: and now I see the dylan branch need python 2.7 to make bitbake running ... I fear this will make me upgrade all my distro. ? Jun 27 08:45:22 mks: dylan shouldn't need 2.7, just master Jun 27 08:47:23 rburton: "BitBake requires Python 2.7.3 or later" that's what I have when running "source poky/oe-init-build-env build/" Jun 27 08:47:46 mks, about python 2.7, I am successfully using a local python 2.7 install --prefixed to somewhere in my ~. Might be easier than a dist upgrade. Jun 27 08:50:00 mks: that is not dylan then... Jun 27 08:50:15 mks: dylan does not need python 2.7 Jun 27 08:51:07 poky$ git branch danny * dylan master Jun 27 08:52:36 hello again. Has anyone a clue as to what's happening when eglibc-internal complains about linux/limit.h on centos? Jun 27 08:54:33 mks: have you another version of bitbake in your PATH? Jun 27 08:55:06 mks: something is not right there since the 2.7 changes did go into master only Jun 27 08:55:33 RP: no only the one in poky/bitbake Jun 27 08:55:36 mks: what's your output from 'git describe'? Jun 27 08:56:02 zibri: 1.5_M1.final-266-g353212a Jun 27 08:56:22 mks: that is clearly not 1.4, 1.5 is master Jun 27 08:56:23 ah, so your dylan branch is for some reason on master Jun 27 09:00:13 zibri: Hum yes, im confused on this one Jun 27 09:47:04 RP, zibri, ok fixed systemd.bbclass by pointing poky on branch dylan Jun 27 09:47:13 thanks guys Jun 27 09:51:34 RP: I read your bitbake commit messages.. it is funny that the cookie monster is still around. :} Jun 27 10:03:02 zecke: I smile when I see those comments in the code so I've left them around :) Jun 27 10:04:54 RP: hi. based on your suggestion, i looked at image-live , but either i don't understand it, or it's not doing what i wanted... what i would need to be able to do is to build a 'regular' bootable image which includes 'chroot' environments let's say in /var/chroot, and have such 'chroot' created by OE, as part of the regular image build. Jun 27 10:06:01 ndec: make your image depend on another image, and then you can take the first image from deploy and put it into your second image Jun 27 10:06:27 oh. that's interesting. i didn't know we could 'chain' images that way. Jun 27 10:06:48 ndec: image-live chains two images that way Jun 27 10:07:06 hmm. as i said, it's possible that i didn't understand it properly... Jun 27 10:07:07 ndec: no, it doesn't extract one into the other, it creates an ext3 file but the same principle applies Jun 27 10:07:23 ndec: nothing special about an image, it's something you can depend on. Jun 27 10:07:25 ndec: just swap the cp for an untar Jun 27 10:07:47 is there any example of -live being used in a layer? Jun 27 10:08:28 ndec: rather than create a class, just write an image recipe which depends on another image recipe's do_rootfs task Jun 27 10:08:53 ndec: I mainly suggested image-live as its proof this does work Jun 27 10:09:24 ok. i can try that. Jun 27 10:10:46 i guess i won't even need package-core-boot for the 'chroot' image, but something simpler than that, right? Jun 27 10:12:43 ndec: whatever makes sense for your use case. Your biggest risk is circular dependencies on the kernel/kernel modules Jun 27 10:13:49 RP: rburton: many thanks. i will try these things, and see if i can get what i need! Jun 27 11:38:10 it's surprising how many recipes are depending (transitively on x11 DISTRO_FEATURE), would anyone object if I send patch adding "raise bb.parse.SkipPackage("'x11' not in DISTRO_FEATURES")" into all of them? Jun 27 11:38:45 JaMa: can you give an example? Jun 27 11:38:51 it would fix "bitbake world" to stop showing long list of ERRORs at the beginning Jun 27 11:39:07 gtk+3, xterm, qt4-x11-free, ... Jun 27 11:39:43 well, not gtk+3, as that builds without x Jun 27 11:39:45 I have list of maybe 100 recipes Jun 27 11:40:03 presumably the situation is eg xterm -> libx11, but x11 is skipping, so bad dependency? Jun 27 11:40:27 maybe there should be a way for the world build to handle deps that are being skipped, and bubble that up the tree without errors? Jun 27 11:40:46 PNBLACKLIST[gtk+3] = "depends on blacklisted libxrandr, libxext, libxcursor, libxdamage, libxrender, libxcomposite, libxi" Jun 27 11:40:46 ^ even for distro without x11 in DISTRO_FEATURES Jun 27 11:41:03 yes that would be nicer than maintaining such dependency tree manually Jun 27 11:41:52 PACKAGECONFIG ??= "${@base_contains("DISTRO_FEATURES", "x11", "x11", "", d)} Jun 27 11:41:54 PACKAGECONFIG[x11] = "--enable-x11-backend,--disable-x11-backend,at-spi2-atk fontconfig libx11 libxext libxcursor libxi libxdamage libxrandr libxrender libxcomposite libxfixes" Jun 27 11:41:56 says gtk3 Jun 27 11:42:05 so you can remove that Jun 27 11:42:20 or showing only WARNING instead of: Jun 27 11:42:21 ERROR: Nothing PROVIDES 'libfontenc-native' (but virtual:native:/home/mjansa/build-starfish/BUILD-goldfinger/.././oe-core/meta/recipes-graphics/xorg-app/mkfontscale_1.1.0.bb DEPENDS on or otherwise requires it) Jun 27 11:42:26 ERROR: libfontenc-native was skipped: 'x11' not in DISTRO_FEATURES Jun 27 11:42:36 yeah, that as a warning makes sense to me Jun 27 11:42:47 so you're aware stuff was skipped, but that's because it was skipped. :) Jun 27 11:42:59 * rburton -> left-over pizza Jun 27 11:43:20 rburton: then something with gtk+3 PACKAGECONFIG doesn't work, because the same ERROR is shown.. Jun 27 11:43:38 I'll test with bitbake -g Jun 27 11:44:09 JaMa: i've done builds of gtk3 without x so it should be working fine Jun 27 11:46:43 and still it fails :) http://pastebin.com/13UHgDvT Jun 27 11:47:30 rburton: ah this is dylan, not master Jun 27 11:48:02 it's fixed in master, but not in dylan.. Jun 27 11:50:35 JaMa: are you still seeing those hangs after the 2.7.3 changes? Jun 27 11:52:01 RP: yes :/ Jun 27 11:52:23 I've seen it even without multiple bitbake processes running Jun 27 11:52:42 but also under heavy load (from portage building libreoffice or something like that) Jun 27 11:54:28 RP: I'll send more detailed info to that ticket when I get it hanging in better time, but IIRC most significant change from last time, is that only one process is hanging (workers are not created) Jun 27 11:54:40 JaMa: ok, thanks for the extra data points :( Jun 27 11:55:10 JaMa: haven't forgotten it but really don't know what to do about it Jun 27 11:56:32 no problem, I understand that it isn't possible to fix something we cannot reproduce reliably Jun 27 11:57:30 I'm glad to see latest bitbake changes going in right direction and still hoping that one of those will magically reveal where it's hanging for me :) Jun 27 11:58:08 have you seen my question about rm_work and setscene tasks? Jun 27 12:03:06 JaMa: I have now. I think this is an artefact of the way the system works. Its not a bug as such :/ Jun 27 12:05:59 OK, I was just interested if you can see some easy way to change this behavior Jun 27 12:06:46 because now in clean build it leaves about 50-100MB in tmp-eglibc/work, but then just the rebuild from sstate-cache leaves couple of GB Jun 27 12:08:57 JaMa: I have to admit I'm a little unclear on the details here. So you're saying an image from scratch with rm_work enabled doesn't remove everything? Jun 27 12:09:49 no other way around Jun 27 12:11:06 image from scratch leaves only temp and sometimes some pseudo files Jun 27 12:11:06 but then when you rebuild the same image from sstate-cache it leaves all those package/packages-split directories from setscene tasks Jun 27 12:11:11 JaMa: oh, I see. That is a bug :/ Jun 27 12:11:20 JaMa: something is causing it to install dependencies it doesn't need :/ Jun 27 12:30:36 RP: ah then it will also resolve package-setscene tasks being reexecuted too often I guess Jun 27 12:31:05 RP: right now when you do "bitbake libx11" with rm_work and libx11 package-setscene stamp is removed Jun 27 12:32:05 RP: and then e.g. "bitbake xserver-xorg" executed libx11.package-setscene again (which from my understanding of some of your comments shouldn't be needed as we shouldn't depend on many package-setscene task nowadays) Jun 27 12:32:22 rp, I haven't forgot to reply to your email :) Jun 27 13:30:27 JaMa: so the question is whether it needs to do that and if not, why it is... Jun 27 13:30:43 Crofton|work: I'll be gone shortly ;-) Jun 27 13:34:55 heh Jun 27 13:34:59 all the more fun Jun 27 13:35:02 when are you gone? Jun 27 13:36:44 RP: OK I'll debug it a bit more, thanks for confirmation that it's not (probably) expected to happen Jun 27 13:37:18 a lot of my builds are still using OEBasic, so where it doesn't seem to happen Jun 27 13:48:37 RP: and one last question, do you remember a bit more about this change http://git.openembedded.org/openembedded-core/commit/meta/lib/oe/path.py?id=2f954a9a6932f1e6c564e7e7aacaac628a75eed7 then what's already included in commit message? Jun 27 13:49:24 RP: we're seeing strange issues from sstate.bbclass in Dylan on one very fast builder, so I wonder if it's possible to trigger that race-condition even without the patch which introduced hardlinks for sstate Jun 27 13:50:29 This is the exception we're seeing http://pastebin.com/RZefUEv5 Jun 27 13:53:52 JaMa: I think it would be possible to trigger the race but that patch should fix those kind of errors Jun 27 13:55:04 JaMa: We ended up using the copyhardlinktree() in more places which just made the race much more visible Jun 27 13:57:09 I've tested that backporting just those 2 commits doesn't break Dylan builds Jun 27 13:57:33 but I'm not 100% sure that it fixes the issue we're seeing as it also doesn't show every build Jun 27 14:01:17 hello there. CORE_IMAGE_EXTRA_INSTALL vs IMAGE_INSTALL_append ... how do they differ? if I append some packages in my image with the latter, I noticed that not all relevant packages will be installed in the image. meaning: i appended openssh in my local.conf upon completion of bitbake, ssh server was sure enough the openssh server, BUT, sftp had not been installed as well. turns out the openssh recipe prepared the openssh-sftp-server package as well Jun 27 14:01:17 , but did not install it. Should I expect CORE_IMAGE_EXTRA_INSTALL would have installed the sftp-server package, too? Jun 27 14:19:23 core_image_extra_install is just another way to add to IMAGE_INSTALL. neither is going to magically pull in packages you didn't add to it. add openssh-sftp-server. Jun 27 14:22:04 Greetings. Having a slight problem with a recipe I wrote. I wrote one recipe, and then a number of others that the first one depends upon. This seems to work, but the recipes on which the first depends are not included in the final image unless we add them to the configuration file. However, it seems that other packages don't have to specify every dependency they have, and just adds them to DEPENDS to achieve this effect. Any pointers? Jun 27 14:23:12 Stygia: when recipe A links against a lib in recipe B, then B will end up in the runtime dependencies of the package from A. that's not the case for other sorts of dependency, however. Jun 27 14:24:18 kergoth, Hmm, right. So do I set SECTION to "lib" for the extra dependencies, or is that a terrible solution (or not one at all)? Jun 27 14:24:40 section is irrelevent. it has no meaning. it's just a categorization mechanism for package manager frontends to use on the target Jun 27 14:24:55 kergoth, Ah. Can I somehow make the extra dependencies libs then? Jun 27 14:24:56 if you want to manually add a runtime dependency, you use RDEPENDS Jun 27 14:25:04 no, that wouldn't make sense either Jun 27 14:25:08 just explicitly add the dependencies Jun 27 14:25:10 kergoth, Ah, alright. I will do that. Jun 27 14:25:14 e.g. RDEPENDS_${PN} += "some-other-package" Jun 27 14:25:25 PN for PackageName, yea?\ Jun 27 14:25:41 yes, it's the main binary package from PACKAGES Jun 27 14:28:57 Hello everybody! I just want to know how much space do I need for my virtual machine when I am building some yocto files on it. Everytime I am trying to build something I get an error message saying I am running out of space. Jun 27 14:29:36 I had 20GB and ran out of space in seconds. Jun 27 14:31:20 Lahar: turn on rm_work Jun 27 14:31:57 put INHERIT += "rm_work" into local.conf Jun 27 14:32:22 rburton: allright Thanks! and that does what? Jun 27 14:34:40 Lahar: deletes the build trees of each recipe once it's built, massively saving space. Jun 27 14:34:48 that said, 20G for building an entire OS is still a bit tiht Jun 27 14:34:51 tight Jun 27 14:35:22 I am giving it 50GB now...You think that would be enough? Jun 27 14:37:32 Also, should I avoid doing bitbake core-image-minimal and bitbake core-image-full? Actually I am currently learning how to use yocto so I didn't know which image to use....I tried using minimal first and then moved on to full....Is this taking even more space? Should I just decide which one to build and avoid doing it again and again? Jun 27 14:37:51 yes Jun 27 14:38:10 rburton: Thanks a lot! Jun 27 14:38:12 keep an eye on your deploy/images as every image you build will sit in there Jun 27 14:38:20 and if you're just experimenting, use a small image Jun 27 14:39:21 got u! Jun 27 14:44:51 Sorry I am full of questions today, I am working on this freescale BSP and reading there documentation to build the image. In their documentation they are doing bitbake fsl-image-minimal or fsl-image-full etc. but in yoctoproject.org quick start guide we have bitbake core-image-minmal....I am just trying to understand the difference between two. Jun 27 14:49:21 does fsl-image-minimal really exists anymore ? :) Jun 27 14:50:15 Lahar: try looking at fsl-image-minimal.bb and core-image-minmal.bb to compare Jun 27 14:53:06 I mean I downloaded the source this morning, the denzil branch only gets fsl-image-gui, fsl-image-gui-sdk and fsl-image-test Jun 27 15:04:36 mks: I tried to google both of these files....I see both has variables that tells you where to install image and stuff but not really understand everything Jun 27 15:14:00 kergoth: is there a systematic way to know the exact name of the package I want to add to my image? Because after generating the openssh* packages , I was searching for relevant packages under tmp/ and came across openssh-sftp-server packages. Jun 27 15:14:32 inspect deploy, or read the recipe or use bitbake -e to examine PACKAGES Jun 27 15:17:47 or use bb ;) Jun 27 15:18:00 kergoth: speaking of which, you know bb is totally broken with bitbake master? Jun 27 15:18:17 kergoth: my branch has some fixes, and i just hacked/fixed up bbcmd Jun 27 15:19:57 Lahar: I have recently bitbaked fsl-image-gui-sdk (which is the most HD hungry image), and the meta-toolchain and adt-installer. for all those, I have used approx 120GB of space. but I have not added the "rm_work" option in my local.conf Jun 27 15:20:13 I have a need to mk an extra directory after a packaging step for a recipe. Essentially, the application expects that directory to be available somewhere but the applications existing install scripts don't make it. What's the best way to go in to fix it? Jun 27 15:20:48 mulhern: in a do_install_append Jun 27 15:21:11 mulhern: rburton: OK! Jun 27 15:21:19 rburton: ah, not too surprising what with the change to configuration in cooker and all, i haven't had a chance to even touch it lately Jun 27 15:21:49 synthnassizer: Thank you! I that's a good information to know. Jun 27 15:22:49 I am trying to build hypervisor image any idea where can I get some information on it? Jun 27 16:46:25 Hi! I was successfly able to set up and configure my environment for yocto build but whenever I do bitbake embedded-hv ....it comes back to me with an error saying ERROR:Nothing PROVIDES 'embedded-hv' ..I have confirmed the BSP iso image I am downloading has hypervisor code in it too. Jun 27 16:50:53 I was able to run bitbake fsl-image-minimal just fine..just having issues running bitbake embedded-hv Jun 27 16:53:01 I'd say contact the maintainers of the fsl layer you're using (meta-fsl-arm / meta-fsl-ppc). read hte readme in the layer Jun 27 16:57:44 khem: around? Jun 27 16:59:48 seebs: http://www.ellexus.com/blog.php Jun 27 17:01:30 kergoth: ok so if I don't have .bb file for embedded-hv in one of the fsl-layer that means it won't build embedded-hv correct? Jun 27 17:50:01 I just saw there is a folder called "embedded-hv" inside my yoctobuildpath/meta-fsl-ppc-private/recipes-tools/embedded-hv ....Do I need to give path when I do bitbake embedded-hv? Jun 27 18:20:48 Lahar: sounds like you didn't add meta-fsl-ppc-private to your bblayers.conf Jun 27 18:22:39 kergoth: I just checked the path is actually there Jun 27 18:23:17 kergoth: but I don't find a .bb file inside yoctobuildpath/meta-fsl-ppc-private/recipes-tools/embedded-hv/ dir ....Does this mean I am missing something? Jun 27 18:43:34 Lahar: you could do "bitbake -e" to confirm the layer config Jun 27 18:51:27 nerdboy: Thanks for info. Looks like system linker is used for qtwebkit while I cross-compile for ARM, I'm not sure though. Jun 27 18:52:48 nerdboy: Ok, Thanks! Let me try that. Jun 27 18:55:25 trollixx: did you look in the recipe files to see if LD is already there in EXTRA_OEMAKE or some such? Jun 27 18:58:59 you could try adding LD='${CXX}' there... Jun 27 19:00:14 nerdboy: I found OE_QMAKE_LINK = "${CXX}" Jun 27 19:01:11 nerdboy: https://github.com/meta-qt5/meta-qt5/blob/master/classes/qmake5_base.bbclass Jun 27 19:10:13 you could try setting it to BUILD_CXX instead Jun 27 19:10:36 but that should evaluate correctly as-is Jun 27 19:12:02 mostly working on a lightweight openbox image, so i've avoided qt so far... Jun 27 19:23:46 nerdboy: so, I looked for .bb class for embedded-hv and I found something called hv-test so I ran bitbake hv-test but ended up with an error which says Summary: 1 task failed: /home/lahar/QorIQ-SDK-V1.3.2-20130325-yocto/meta-fsl-ppc-private/recipes-tools/hv-test/hv-test_gi.bb, ...returning a non-zero exit coded. Any idea what that could be? Jun 27 19:24:18 nerdboy: the thing is that everything else is building correctly, so only webkit fails, I'll try BUILD_CXX now Jun 27 19:49:38 trollixx: which branch are you building from? Jun 27 19:49:59 Lahar: not without more info... Jun 27 19:51:16 nerdboy: No problem! I think I found another file called hypervisor_git.bb and currently running bitbake hypervisor....let's see if that helps ...if not will ask you again. Jun 27 19:55:56 trollixx: maybe that last commit is pooching it Jun 27 19:56:53 if you're building master, then you could try a git checkout on the commit just prior to head Jun 27 19:57:27 if that passes qa packaging then i would file a bug Jun 27 20:04:33 damnit, keep getting an ICE trying to build nativesdk-qemu Jun 27 20:05:06 nerdboy: I used several week old master branch first, and only then updated, so the problem is not in the latest commit. Jun 27 20:06:19 ah, then resetting LD in qtwebkit.inc is your best bet Jun 27 20:06:26 nerdboy: I guess https://github.com/meta-qt5/meta-qt5/blob/master/recipes-qt/qt5/qtwebkit.inc#L18 should address this problem, but not in my case Jun 27 20:06:34 unset LD? Jun 27 20:07:44 unset/set or override it, however you want to look at it... Jun 27 20:08:16 ok, now rebuilding, it takes a lot of time... Jun 27 20:08:44 i would try adding EXTRA_QMAKEVARS += "${LD}=${BUILD_CXX}" Jun 27 20:09:12 er, LD=... Jun 27 20:10:21 ok, will try that if just unset won't succeed... Jun 27 20:13:51 i think i found it Jun 27 20:14:07 gotta clone it to check Jun 27 20:14:28 also in the middle of my rpi image build Jun 27 20:29:51 I was just fixing a permission error on the minnowboard IRC logs and noticed the #yocto logs aren't linked to in the title. Jun 27 20:30:04 Should we add https://www.yoctoproject.org/irc/ ? Jun 27 20:34:22 trollixx: that sed expression doesn't look right Jun 27 20:34:38 i would change it to \${B} Jun 27 20:35:42 can you check the makefiles and see if that line actually worked? Jun 27 20:36:21 there aren't any linker options like that in the older qt makefiles Jun 27 20:37:04 looks like it's mostly qmake project files Jun 27 20:42:40 nerdboy: i'm curious, why doesn't the sed line look correct? the variable should be expanded long before sed sees it. but then again, i'm new to oe/yocto, i may be missing some weird detail :) Jun 27 20:42:51 nerdboy: what should look for? Jun 27 20:42:58 *I look Jun 27 20:51:53 the -rpath,${B}/lib part Jun 27 20:53:24 zibri: i assumed it was literally looking for ${B} Jun 27 20:53:59 ah, if that's the case i think it would have to be double escaped, as to not let sed treat $ like end of line Jun 27 20:54:05 i don't have qt5 source to look at yet Jun 27 20:54:09 nerdboy: find -name "Makefile*" | xargs sed -i "s|-Wl,-rpath,/home/oleg/Projects/willow-build/tmp/work/armv7a-vfp-neon-willow-linux-gnueabi/qtwebkit/5.0.2-r0.0/build/lib||g Jun 27 20:54:35 oops Jun 27 20:55:19 I would think it was meant to remove host specific paths... like that one Jun 27 20:57:42 nerdboy: here's line from makefile: Jun 27 20:57:44 LFLAGS = $(OE_QMAKE_LDFLAGS) -Wl,--gc-sections -Wl,--no-undefined -Wl,--no-undefined -Wl,-O1 -Wl,-rpath-link,/home/oleg/Projects/willow-build/tmp/sysroots/g2/usr/lib -shared -Wl,-Bsymbolic-functions -Wl,-soname,libQt5WebKit.so.5 Jun 27 20:58:58 trollixx: it must be in $(OE_QMAKE_LDFLAGS) Jun 27 20:59:43 zibri: it appears to be trying to remove the whole -rpath argument Jun 27 21:01:51 -rpath-link should be safe as it's only a link time option Jun 27 21:03:15 yes, i guess the current sed line is correct, then Jun 27 21:04:34 he's still getting a host rpath in a library Jun 27 21:04:45 ah Jun 27 21:08:06 still waiting for my image build so i can bitbake qtwebkit -c configure and look for myself Jun 27 21:15:54 only 1800 tasks left to go... Jun 27 21:22:58 nerdboy: OE_QMAKE_LDFLAGS='-Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed' Jun 27 21:33:12 I am not sure if I am asking my question correctly - Is it possible to run a single recipe on QEMU. I actually downloaded BSP of P4080ds board but bitbaked it's hypervisor code only and would like to see it's image running on QEMU. but when I do runqemu p4080ds or runqemu hypervisor nothing works.... Jun 27 23:02:57 * pidge kicks 1.4.1 is out the door! Jun 27 23:03:05 s/is// Jun 28 01:27:51 hmm, looks like nativesdk-git hardcodes the path to the templates dir, breaking buildtools-tarball Jun 28 02:22:40 /me switches wife's chromebook to developer mode Jun 28 02:24:55 don't suppose there's a meta-chromebook layer somewhere... **** ENDING LOGGING AT Fri Jun 28 02:59:58 2013