**** BEGIN LOGGING AT Fri Aug 01 03:00:00 2014 Aug 01 05:36:04 i am compiling poky with my custom bb file. if i compile this i am getting error please check paste bin. please check this error log http://pastebin.com/9VYNHDZa Aug 01 05:37:17 also check my .bb file http://pastebin.com/NhEXHViS Aug 01 05:42:23 embed: well isn't it obivous: package clancor-qt-sdk71xx-init wants to install the file etc/ppp/peers/provider, which is also provided by the package ppp Aug 01 05:42:37 embed: thats a conflict, so it does not build. Aug 01 05:44:49 so my advice would be to fix package clancor-qt-sdk71xx-init Aug 01 05:50:24 how to fix this.. Aug 01 05:51:27 well clancor-qt-sdk71xx-init seems to be either your package, or something you got from a supplier.. its nothing public. so you have to take care yourself. Aug 01 05:52:46 the trivial solution would be to only include either ppp OR clancor-qt-sdk71xx-init in your dependencies. so no clash. if you need both, you actually *have* *to* look at the clancor-qt-sdk71xx-init recipe, sources and install scripts to find out what they to, and then make them compliant to be installed *on* *top* of ppp Aug 01 06:18:05 kroon: good morning Aug 01 06:18:15 thanks for answeing.. Aug 01 06:19:08 yes i'm using MACHINE="wandboard-solo" Aug 01 07:29:04 good morning Aug 01 08:03:49 good morning if I create my own metalayer how can I access files located in the /files directory ? What is the variable for the path to that location ? Aug 01 08:48:17 good morning if I create my own metalayer how can I access files located in the /files directory ? What is the variable for the path to that location ? Aug 01 09:10:47 wfailla1: what's the context where you need to specify the path to the file? Aug 01 09:10:51 (morning all) Aug 01 09:32:10 bluelightning: I used ${LAYERDIR}/files Aug 01 09:33:12 LAYERDIR only works within the layer.conf file, but maybe that's where you're specifying this ... ? Aug 01 09:33:46 in a lot of contexts you shouldn't need to add any prefix since the root of the layer gets added to BBPATH Aug 01 09:36:10 bluelightning: yes sry I forgot to say I created a variable XXXdir in the layer.conf Aug 01 10:56:57 hi, have anyone had problems booting image on wandboard-solo (imx6 solo) using tftp and nfs, but not when using the sdcard. I am using the same tftp and nfs setup with dora, and everything works well? Aug 01 10:58:52 when booting the zImage from daisy via tftp/nfs it hangs on "Starting kernel..." Aug 01 10:59:27 but when I use the sdcard image from the same build, everythings works fine Aug 01 10:59:47 any ideas? Aug 01 11:16:00 have you tried comparing what is loaded from tftp/sd and seeing if they match up ? Aug 01 12:29:41 please, can you show me some commands for 1) cleaning 2) recompiling kernel only (new config) 3)creating image ? Aug 01 12:30:39 documentations is quite mixed up. i could only find bitbake linux-yocto -c menuconfig Aug 01 12:31:09 but how do i build the kernel after configuring and creating fragments? Aug 01 12:39:58 microd: to compile the kernel, just bitbake virtual/kernel (or linux-yocto; virtual/kernel is just a bit more generic) Aug 01 12:44:38 bluelightning: glad to hear from you;) virtual/kernel throws me an error that there is nothing like virtual/kernel Aug 01 12:44:59 erm, that is strange Aug 01 12:45:12 it only accepts linux-yocto Aug 01 12:45:15 you're not doing bitbake -c virtual/kernel by any chance ? Aug 01 12:45:41 can you pastebin the output ? Aug 01 12:45:49 i do bitbake -c virtual/kernel Aug 01 12:45:57 right, that can't work Aug 01 12:46:09 yap, will take me some minutes Aug 01 12:46:16 no worries, I can see what's wrong Aug 01 12:46:24 the thing that follows -c is meant to be a command, not a target Aug 01 12:46:32 virtual/kernel and linux-yocto are targets Aug 01 12:46:47 so it is bitbake virtual/kernel -c ? Aug 01 12:46:59 just bitbake virtual/kernel, no need for -c Aug 01 12:47:06 ok let me check Aug 01 12:47:46 the strange thing was, the timestamp of my kernel uImage is always the same Aug 01 12:48:10 this is why i'm asking for how to force a rebuild Aug 01 12:50:45 that's just it though, -c menuconfig is supposed to do that already Aug 01 12:50:54 provided you save the edited config that is Aug 01 12:51:30 bitbake virtual/kernel works Aug 01 12:52:43 how to check if the defconfig in my layer is applied to kernel config ? Aug 01 12:56:43 Hi, dose any one other than pidge has access to the yocto-autobuilder documentation? Aug 01 12:58:52 wfailla1: I'm not sure there is a huge amount of documentation atm, I believe that's something that lynn is helping with at the moment Aug 01 12:59:53 wfailla1: pidge is definitely the autobuilder expert though Aug 01 13:00:36 bluelightning: yes she told me two days ago she whanted to sync with the intern who is working on the documentation and then email me a copy. But I can't reach her here. Aug 01 13:01:12 wfailla1: FWIW she's based in the US so she'd only be around starting in a few hours Aug 01 13:01:46 wfailla1: based on a bug report I'm subscribed to I don't think there's been a lot of progress on the docs yet Aug 01 13:03:06 microd: you could have a look at the .config file in the build directory for the kernel recipe... if you're not sure where that is you can find it by running: bitbake -e virtual/kernel | grep ^B= Aug 01 13:03:47 bluelightning: ok what time PST is she usually arround ? Aug 01 13:05:17 wfailla1: I think 9 or 10 Aug 01 13:05:30 bluelightning: AM? Aug 01 13:05:35 yeah Aug 01 13:06:02 you could try emailing if you haven't already Aug 01 13:07:10 bluelightning: can i get you a beer? ;-) Aug 01 13:08:01 microd: sure :) Aug 01 13:09:04 checked that config is applied! u-boot still tells me image creation time is 12:09 UTC.... there's sth wrong. maybe delete tmp directory? Aug 01 13:37:41 hi, is it reasonable to only update meta, bitbake and oe-init-build-env from Yocto if we do not seemingly use anything else from Yocto, or we should always grab everything together? Aug 01 13:40:39 lpapp_: well you won't get too far without the scripts directory, since some of that is used by the build system Aug 01 13:41:00 ah, yes, true, + scripts. Aug 01 13:41:16 so basically getting rid of: Aug 01 13:41:18 * meta-hob Aug 01 13:41:26 * meta-skeleton Aug 01 13:41:31 * meta-yocto-bsp Aug 01 13:41:37 * meta-yocto Aug 01 13:42:36 I'm not sure you'll really gain much by deleting those things Aug 01 13:43:02 if you don't have them in bblayers.conf they won't have any influence over your builds Aug 01 13:48:06 bluelightning: well, the gain is clear repository, not cluttered with irrelevant directories in the root. Aug 01 13:48:37 and less directories to sync up or replace when updating our software stack on top of Yocto. Aug 01 13:48:43 s/less/fewer/ Aug 01 13:49:18 basically the "Yocto core". Aug 01 13:49:41 you can certainly use bitbake + OE-core rather than poky, lots of folks do that Aug 01 13:49:54 we have been already doing that; that is a no issue. Aug 01 13:50:29 as far as I understand Yocto is nothing more than just verification of bitbake + oe-core together with dedicated versions. Aug 01 13:50:37 leave Poky alone. Aug 01 13:51:05 in other words, Yocto seems to be just a Q/A process for us. Aug 01 13:51:32 that's a pretty huge simplification (and minimises a lot of the work that we do) Aug 01 13:52:06 so what is Yocto then if not only a Q/A process for our use case? Aug 01 13:52:38 Q/A process for a coherent release, that is. Aug 01 13:54:29 I mean, bitbake and oe-core are developed in dependently from Poky and all that. Aug 01 13:54:40 and it had been developed before Yocto, too, etc. Aug 01 13:55:03 Yocto seems to have added a release process for a coherent release. That is how I see Yocto's rule. Aug 01 13:56:43 (I am not saying Intel, Linux Foundation, et all are not developing bitbake, meta, and some patches to upstream, but the addition with the Yocto seems to be the release process for those under an umbrella, isn't it) Aug 01 13:56:47 et al* Aug 01 14:07:17 I have a device that needs to be able to show a text console if a monitor is hooked to it. Right now it is working for 1 out of 3 monitors I have tested. All the monitors work with the old image we are trying to replace. Any suggestions how to troubleshoot the video issue? Aug 01 14:07:43 (the old image is not yocto based) Aug 01 14:51:24 blloyd: I guess the question is what the differences are and the kind of problems you're seeing. Same kernel or not? Can you try the other kernel? or is the problem userspace? Aug 01 15:24:00 RP: new kernel (who wants to deal with a 2.4 kernel these days). ssh works, but console flashes like it is changing mode then stays blank. Using video=vesafb vga=0x318 one monitor works. The old system used vga=788 with a lilo bootloader and 2.4 linux kernel. Aug 01 15:24:53 and 0x314 doesn't work on any of the monitors (even the one that works at 0x318) Aug 01 15:26:18 blloyd: It does sound primarily a kernel problem. The console parts have changed a lot in more recent kernels with KMS and so on. Anything obvious missing in the defconfig? Aug 01 15:29:08 no, but not sure what I would be looking for. Aug 01 15:32:33 blloyd: The CONFIG_FB_* pieces I guess Aug 01 15:33:10 blloyd: vesafb is a graphics framebuffer console FWIW, not text mode Aug 01 15:54:33 true, but what gets me, I can hand edit the resulting syslinux.cfg file, remove the video=vesafb and vga=X, and no monitors work. I can just change the 0x318 to 0x314 and again nothing works. But for 0x318 one of the displays work. Because one can work, I don't think it is a missing driver or anything (never should work at all in that case, right?), but I don't know what to do from here to further isolate why this is broken. Aug 01 15:58:16 ok, well technically it isn't blank at 0x314. I get green mess in the top 1/4 of the screen. It almost looks like it is overdriving the monitor. It's like it's selected the wrong mode, and so doesn't display a good picture. Aug 01 15:59:33 more specifically, it's like the top area has 4 poor quality thumbnails of the requested display, complete with blinking cursor near the bottom of each. Aug 01 16:05:22 I have a bb recipe that I am trying to create that sets the package version to the latest changelist number in my version control system. I am running into an issue where between the different states that value is being re-evaluated and then breaks the build. How can I automatically set a variable to this value and then have it stay fixed through the whole build process? Aug 01 16:28:31 moin Aug 01 17:24:38 I love you guys. Aug 01 17:24:41 Just so you know. Aug 01 17:38:01 wfailla1: http://git.yoctoproject.org/cgit/cgit.cgi/yocto-autobuilder/log/?h=lcyrin/working Aug 01 17:40:10 wfailla1: they're beginning stages, but it will bring some basic autobuilder docs in line with the rest of the project. for more complex things, the class docs (which I'll be pulling in soon) are the way to go. Aug 01 17:40:37 wfailla1: talking to the person who has been working on them now actually Aug 01 17:41:50 wfailla1: branch with additional docs Aug 01 17:41:51 http://git.yoctoproject.org/cgit.cgi/yocto-autobuilder/log/?h=sullical/documentation Aug 01 17:41:55 pidge: THANKS SO MUTCH!!!! Aug 01 17:42:58 wfailla1: I'm sure there are holes that need patching in them, especially in the area between "what is a buildbot doc thing and what is a yocto-autobuilder doc thing". Buildbots docs *should* be helpful in those places Aug 01 17:44:05 wfailla1: np. sorry this took a bit. I'm in the middle of releasing 1.5.3 Aug 01 17:45:27 pidge: Well I will read all of this this weekend. And then I will give you a feedback. Aug 01 17:45:47 pidge: no problem Aug 01 17:50:54 wfailla1: please do and copy cal and lynn. Aug 01 17:52:42 pidge: Sry for asking away write this moment. But is there a way for builds to be trigured by git events ? I know build bot can do that but does the autobuilder have configure options for this ? Aug 01 17:54:42 pidge: FYI I'm in CEST so I dont know when I get the chance to catch you in the chat again ;) Aug 01 17:55:35 wfailla1: We support.... *looks at code* Aug 01 17:56:24 ok Aug 01 17:56:27 timed, SingleBranchScheduler, trigger and forced. Aug 01 17:56:39 adding support for others is relatively trivial Aug 01 17:56:42 look at: Aug 01 17:56:51 lib/python2.7/site-packages/autobuilder/Autobuilder.py Aug 01 17:59:00 wfailla1: I'd be more than happy to take a patch that supports the others. In fact, I'd like to refactor the scheduler code a bit to make it a little less icky (it's not horrible, it could just be a little tidier) Aug 01 18:01:42 wfailla1: I'm on the west coast of the US for the next few weeks. I'll be closer to you in about 2 months. Aug 01 18:04:18 pidge: Ok yes I will be trying to use the autobuilder to setup a CI for tplino the destrebution we are developing here. **** ENDING LOGGING AT Sat Aug 02 02:59:59 2014