**** BEGIN LOGGING AT Thu Jul 10 02:59:58 2014 Jul 10 07:10:32 Hi all.. Jul 10 07:53:25 Hi yocto..I am working in beaglebone development board..in that i want to run qt programs so i enabled QT support in conf file and i compiled my own .bb files for my custom project purpose. Jul 10 07:53:47 and i generated sdk using populate_sdk Jul 10 07:55:17 if i install that SDK there is no qmake binary in sysroots/x86_64-linux/usr/bin directory. Jul 10 07:59:10 please check my conf file and my own bb file as mentioned in link http://pastebin.com/KaFWLtKH. Jul 10 08:00:17 what went wrong ?. and what i want to do to get qmake in sysroots/x86_64-linux/usr/bin disrctory ? Jul 10 08:00:36 please any one help me to solve this issue. Jul 10 08:13:50 wasn't there some specific target for a qt-enabled sdk? Jul 10 08:28:27 elinuxer, when I've delivered SDK's including Qt, I've always made custom toolchain-recipes based on e.g. meta-toolchain-qte.bb (or something similar, possible from a vendor layer etc). Jul 10 08:28:47 These are built by `bitbake meta-toolchain-qte` Jul 10 08:31:33 hi guys Jul 10 08:32:03 AndersD: yeah that was what i couln't remember... meta-toolchain-qte Jul 10 08:33:44 I have a question regarding toaster. I am wondering why I can't see in the report among recipes nor among packages that u-boot was built, even though I know it was? Jul 10 08:36:10 forget that I asked...I found why :) Jul 10 08:53:28 AndersD:my custom bb files named as bbb-74xx.bb and please check this link http://pastebin.com/cKvqy0TT after compilation i generated SDK using "bitbake bbb-74xx -c populate_sdk " Jul 10 08:54:12 IS this right? Jul 10 08:55:28 Yes, but if I haven't missed anything lately, you need a toolchain-recipe to get Qt into your sdk... Jul 10 08:57:48 Please tell me how to add toolchain-recipe in my sdk.. Jul 10 09:01:02 I've created a recipe meta-toolchain-customer.bb that inherits e.g. meta-toolchain-qte.bb. Then I just need to make sure that meta-toolchain-customer.bb adds all the libraries etc I need to my image. Jul 10 09:01:53 It would be preferably to be able to run -c populate_sdk on your standard image, but I don't thin it works yet for create a qt-based SDK. Jul 10 09:02:51 Even If i check my yocto build system path entry in sysroots/usr/bin directory,there is no qmake instead of that qmake2 is located. Jul 10 09:03:58 Are you using qt4 or qt5= Jul 10 09:05:25 please see my build directory.... build-74xx/tmp/work/x86_64-linux/qt4-native/4.8.4-r19.0/ Jul 10 09:07:37 Well, you're using qt4 => qmake2. Don't you have a qmake2 in tmp/sysroots/x86_64-linux/usr/bin/qmake2? Jul 10 09:14:26 AndersD: qmake2 is there in tmp/sysroots/x86_64-linux/usr/bin/qmake2 and please check this link http://pastebin.com/msh3TPri Jul 10 09:15:53 Then what was the issue with qmake/qmake2? Jul 10 09:21:03 In my previous project with same qt4 i was getting qmake and qmake2 but now only qmake2 is located, thatswhy i am getting confuse is there any mistakes on configuration. Jul 10 09:28:45 Well, I've only got qmake2 in my old qt4-based branch of one of my projects. Unless you're having any issues, I wouldn't worry. Jul 10 09:41:27 AndersD: ok...I have one more doubt, I want to compile QT sample application called Imageprinttest and i trying to compile this its error please check this link http://pastebin.com/cc4WZu6g Jul 10 09:49:33 I have a init_ifupdown_1.0.bbappend to append a new interfaces file. In tmp/work/overo.../init_ifupdown/image my interfaces file is found. In tmp/work/arm....../init_ifupdown/image it is another bbappend file present. I believe my image come from arm... why this could append ? Jul 10 09:49:45 elinuxer, Are you trying to build it outside of oe/bitbake? Jul 10 09:50:47 It looks like that. In that case generate a toolchain (bitbake meta-toolchain-qte or bitbake meta-toolchain-qt, depending on what you're using). Jul 10 10:50:08 RP: last package.bbclass changes are great, thanks (I'll include them all in the next world build) Jul 10 12:11:28 Is there a way to disable recipes or set of recipes in Yocto? e.g. I made a new meta layer which contains one recipe busybox. Since poky/meta contains busybox recipe and my layer contains busybox, Yocto will not know which recipe to build. Therefore, I would like to disable recipe without disabling the whole layer (since I need other recipes in that layer). Do you know if there is a way to do that? Jul 10 12:18:03 ivanstojanovic: just increment your layers' priority Jul 10 12:20:37 ivanstojanovic: or you can use BBMASK, see the ref manual for details Jul 10 12:27:40 @Letothe2nd: I will try that. I know that this works for the two packages with different version, but let's see if it works when the version is same. Jul 10 12:27:52 @ndec: I will have a look at that as well. Jul 10 12:27:55 Thanks guys! Jul 10 12:43:59 Hello all, I'm writing some "complexe" recipe and I have a question... How is the best way to use a file generate by a recipe A in a recipe B ? Jul 10 12:44:07 JaMa: Just a question of how quickly to merge them now :) Jul 10 12:46:31 detail: I would create a xilinx boot image, it contain uboot binary, a kernel with an initrd and a FPGA design, in a recipe, I have the kernel, in another I have boot image, I have also the bitstream in a SVN server... I want to get all them in the "fetch" of the recipe... Jul 10 12:56:21 RP: :) I'll have world status after weekend.. fwiw Jul 10 12:57:12 rburton: fyi: I'll send meta-demos upgrade together with 2 more PACKAGECONFIGs for glew and glu (we talked about it few months ago..) Jul 10 12:57:37 it will be build-able without x11 in DISTRO_FEATURES Jul 10 12:58:03 JaMa: I'm tempted to get these ones in and then we can start fixing the warnings. Where are we at on the autotools one? Jul 10 12:59:12 RP: I've sent patches for 1/2 of test-dependencies run Jul 10 12:59:24 but it's still running (takes 14 days to finish, 2-3 days remaining) Jul 10 13:00:00 and I wonder if it caused few "silent" do_configure failures Jul 10 13:00:15 because few recipes were now failing in do_install with "nothing to do" Jul 10 13:00:29 JaMa: those are more likely B!=S issues? Jul 10 13:01:17 that's possible, but most B!=S were blacklisted/resolved already Jul 10 13:01:24 so I'm not sure if it's new or not Jul 10 13:01:49 it could be newly revealed B!=S issue (which started to show now when its dependencies were fixed) Jul 10 13:02:45 JaMa: that is probably more likely, I've not seen the autotools one give silent failures Jul 10 13:03:16 please wait 1 week, for me to send reports and fix trivial issues like pkgconfig/intltool-native/gtk-doc Jul 10 13:04:02 JaMa: ok Jul 10 13:04:38 JaMa: part of the problem is the master-next and master divergence is blowing up the size of the sstate cache and the build times :/ Jul 10 13:29:53 RP: how long do you keep sstate from builds? Jul 10 13:30:30 we're generating about 2TB per month and we're not tracking master (still based on dylan) Jul 10 13:30:38 I have a SRC_URI consisting of two GIT repo's, each has the option name=... attached. How can I specify the checkout directory for each? Jul 10 13:31:29 (GITDIR_ seems not to work as I expect) Jul 10 13:32:00 Ah GITDIR is the DL dir, not the source dir. Jul 10 13:32:12 likewise: destsuffix= Jul 10 13:32:59 JaMa: tnx Jul 10 13:33:15 I had destdir in mind, but didn't grep much Jul 10 14:16:36 Hello all, had a quick question about the PR server. My team is looking to start using it for our project and I'm wondering the best way to use it. We have a main gito repo server on our network, and multiple users pushing/pulling on a daily basis. I'm thinking to run the PR server there and have everyone set their local.conf on the local machine to point at the remote git server where it runs for everyone. Is that the normal way to Jul 10 14:42:10 mrschmitz: you'd only need to do that if they were expecting to share packages built on multiple machines and the versions had to be in sync Jul 10 14:44:00 rburton: Thanks. We want the versions to be in sync. I think. Our main git repo is also our main build server, so it would probably be best for everyone to stay in sync. Unless I'm thinking wrong. Jul 10 14:45:22 rburton: I suppose the main build server is the real build for the project, and not just local minor builds though. Perhaps that makes it not necessary. Jul 10 14:47:13 mrschmitz: yes, if you're not mixing developer's builds with master builds then you don't really need it. Jul 10 14:49:12 rburton: OK. I removed the "PRINC" references in the packages for our custom layers, do I just replace that with PR in the recipes? Jul 10 14:49:33 rburton: we definitely won't be mixing developer builds and master builds. Jul 10 14:52:25 RP: this kind of exceptions from package_qa isn't expected, right? Jul 10 14:52:26 Exception: OSError: [Errno 2] No such file or directory: '/OE/build/shr-core/tmp-eglibc/sysroots/qemux86-64/pkgdata/runtime-rprovides/xorg-abi-video-15/' Jul 10 14:53:14 I was in build still using OEBasic (without Hash) so it's very likely that xserver just wasn't rebuilt automatically, but maybe this exception should be error that there is no RPROVIDER for xorg-abi-video-15? Jul 10 14:58:08 JaMa: right, its not expected to fail that way Jul 10 15:41:03 Hello. I have been dealing with a bug for some time now. Basically, running our software on our embedded device (OMAP 4) our program occasionally (very rarely actually) gets aborted and "Bus Error" is printed to stderr. I have been operating under the assumption that this was a bug in our code but have not been able to find it. I recently observed the same "Bus Error" while GCC was compiling code (not while running our software). I ran the compiler again ex Jul 10 15:41:03 actly the same way and it completed successfully (no Bus Error). I have seen this twice now while compiling code. It seems to me that the issue may not be in our software after all - it may just happen occasionally when under heavy load. Now I am a bit stumped. I understand that on ARM, a Bus Error typically means a non-aligned memory access. When the problem happens, the computer does not crash - it just aborts the process it was in. To me this suggests (b Jul 10 15:41:04 ut I am no expert) that the bad memory access is coming from user space and not kernel space (I would expect a less graceful failure). Does anybody have any idea where I might look? Jul 10 15:58:20 Has anybody dealt with Bus Errors before? Jul 10 15:58:55 is it possible to run the s/w under a debugger like gdb? Jul 10 16:00:10 I could probably run our software in GDB in the lab. But the issue also occurs when compiling code with GCC (randomly, it seams). Not sure if I can run GCC in GDB. Jul 10 16:00:27 sure you can run gcc in gdb Jul 10 16:00:32 K Jul 10 16:00:51 that would probably be the best way to stimulate an error Jul 10 16:01:01 what would I be looking for Jul 10 16:01:03 ? Jul 10 16:02:57 gdb should show you what line of code triggered the bus error Jul 10 16:03:30 another q, are you running a custom kernel on your device? Jul 10 16:11:11 I built everything with Yocto. We are using the Gumstix DuOvero module and I started with their recipes. I think it is mostly stock, but with some patches. Jul 10 16:13:41 I ran make in GDB and everything went fine that time. I will repeat several times until I get the problem. Jul 10 16:14:35 Thank you for the tips. Jul 10 16:15:46 I asked about the kernel because an older kernel could have CONFIG_ALIGNMENT_TRAP disabled, which might cause similar symptoms Jul 10 16:16:08 I believe it's not optional on newer kernels Jul 10 16:27:21 What is CONFIG_ALIGNMENT_TRAP? My kernel version is 3.6.0 Jul 10 16:27:51 bryan_: some details on this page http://jsolano.net/2012/09/06/arm-unaligned-data-access-and-floating-point-in-linux/ Jul 10 16:29:00 Thank you! Jul 10 16:32:09 bryan_: np Jul 10 17:24:15 JaMa: that is a general bug in the code, its not siggen related. Will fix, it just needs a try: except OSError Jul 10 17:26:50 JaMa: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t222&id=c9284553e826858e063ef233c620c1342350c37d is the fix, will squash into the main patch in due course Jul 10 17:29:04 General (and basic) question: How can I re-build my distro using a newer kernel version? I am building for the DuOvero module and my kernel version is 3.6.0 Jul 10 17:34:05 RP: thanks Jul 10 17:38:08 RP: in my case I've seen a lot of recipes failing like this in the end, because their dependencies weren't rebuilt to populate rprovides entries Jul 10 17:38:32 RP: so I've started new build - that's how it was related to siggen for me Jul 10 17:40:41 RP: question. i know recently you made it so native changes result in rebuild of target stuff, does that include change of BUILD_ARCH? Jul 10 17:42:31 we're hitting a case where target sstates aren't reused between 32 and 64 bit build hosts, but i'm not sure if its becuase we needed to add -march= to BUILD_CFLAGS to deal with gcc <4.5 hosts, or if it's the way it always behaves Jul 10 17:45:02 hmm Jul 10 18:38:43 hi all Jul 10 18:39:44 can someone tell me how to install a -dev package to the sdk Jul 10 18:40:03 i remember opkg-target Jul 10 18:40:16 but it is not there Jul 10 18:41:46 any help will be realy apresiated Jul 10 18:45:13 jopado1: can you tell me how to install a -dev package to the sdk Jul 10 18:46:03 florian_kc: can you tell me how to install a -dev package to the sdk Jul 10 18:47:33 nobody can help ? Jul 10 19:18:03 jopado1 Jul 10 19:26:44 Not sure if this is the correct forum for this (yocto noob). If not please advise. I am using meta-fsl-arm on a freescale vybrid platform and am trying to create encrypted filesystem partition at runtime. I'm using cryptsetup (LUKS) on a jffs2 partition. I can format and open the LUKS container and see the entry in /dev/mapper. I successfully create the filesystem via mkfs.jffs2. However I can not mount the filesystem - mount fails Jul 10 19:26:57 Does anyone know if this is supported or if there are known issues with this. Thanks Jul 10 19:39:40 Don't know the answer to that one, sorry. Jul 10 19:39:59 (Just a note, the way IRC tends to work, you might get an answer in several hours, so try to stay logged in for a long time.) Jul 10 19:40:14 or better yet, always Jul 10 19:40:15 :) Jul 10 20:03:22 OK. Thanks. I will post to the email subscribers also. Jul 10 22:27:20 RP: it looks like it was already merged without your last fix for OSError Jul 10 22:27:51 ah not all, just package* changes were merged Jul 10 23:22:35 JaMa: right, I've held the final patch for now, just enabled the other pieces Jul 11 02:23:50 It seems the insane change has broken some recipes Jul 11 02:24:42 In meta-fsl-arm, we do some additions to the INSANE_SKIP list in the populate_packages_prepend. Jul 11 02:25:06 It seems the QA tests are run before this completes so they cannot be skip Jul 11 02:27:21 * nerdboy is still not sure what broke x keyboard layout **** ENDING LOGGING AT Fri Jul 11 02:59:58 2014