**** BEGIN LOGGING AT Tue May 20 02:59:59 2014 May 20 05:56:10 Hi all...Gd mg May 20 06:49:44 How do you guys manage your layers? repo? May 20 06:50:17 getting started with the wandboard, and want to build this up properly from the start May 20 06:59:24 meta-fantastic May 20 06:59:28 meta-great May 20 06:59:37 meta-extremely-cool May 20 06:59:42 such or similar :) May 20 07:01:58 LetoThe2nd, hehe. already on board with that. I'm trying to find a way to manage them. May 20 07:02:05 repo seems promising May 20 07:03:04 tasslehoff: i know what you mean. i've even started my own little solution for doing that (i'm no repo man) May 20 07:03:07 until now I've used angstrom and the setup-scripts. there a layers.txt file contains all the layers and info about which branch and revision should be used May 20 07:03:58 but probably we'll just skip onto the autopbuilder train and keep our config magic in its configurations. May 20 07:04:05 I didn't do this properly from the start and now it's a pain if I have to go back and build an earlier version of my rootfs May 20 07:05:18 LetoThe2nd, ah. that is a train I haven't taken before :p May 20 07:06:30 we realized that spending too much time in automating things for a single developer isn't that necessary, because in the end we want a reproductible, automated build anyways, including logs etc. May 20 07:06:53 and most of the developers only need the sdk and rootfs May 20 07:07:10 (which falls out of the autobuilder and they can just grab it) May 20 07:13:23 I guess that *almost*applies to us as well. May 20 07:13:52 Need a bit of control over which packages are added to the image, and have our own layer with some recipes. May 20 07:14:42 yes. but in our case the developers probably wont even have their own layer so far. May 20 07:15:03 but thats all in discussion still. we'll see how it turns out. May 20 09:20:41 morning all May 20 09:48:25 gm May 20 09:48:35 morning Crofton|work May 20 10:15:48 hi May 20 10:16:06 droy: hello May 20 10:16:28 actually i wanted to make an operating system for my MSP-exp 430f5529 board May 20 10:16:37 could any one help me out ?? May 20 10:17:28 well, I don't have any specific experience with that board, but basically you'll need to know something about the machine (specific architecture) and you'll need a kernel configuration May 20 10:18:01 ok May 20 10:18:09 you may find our BSP developer guide useful: http://www.yoctoproject.org/docs/current/bsp-guide/bsp-guide.html May 20 10:19:08 we also have a kernel development manual: http://www.yoctoproject.org/docs/current/kernel-dev/kernel-dev.html May 20 10:23:11 ok May 20 10:25:10 see http://www.ti.com/tool/msp-exp430f5529 this is my board how do i find about its architechture ? May 20 10:26:10 well, you'll probably need to dig for some info about building Linux for this board or perhaps the chip on it May 20 10:26:36 that's really outside the scope of our project May 20 10:27:03 (I mean, unless someone in our community has already done that work) May 20 10:30:05 ok May 20 10:30:57 heres a list of os http://processors.wiki.ti.com/index.php/MSP430_Real_Time_Operating_Systems_Overview thats been used for that board May 20 10:33:10 droy: msp430 are typicall 16-bit low power custom MCU. you wont' get any linux running on it. May 20 10:33:49 ok May 20 10:34:01 yes actually looking at it I would have to say the same thing... May 20 10:34:41 now if the question is whether or not OE can be used to build custom RTOS system , that's a different question ;-) May 20 10:35:26 ndec: I'm sure it could be done... May 20 10:35:34 with enough time and effort, of course May 20 10:35:54 it's all s/w afterall... May 20 10:35:54 so that means yocto cant be used for making os for that May 20 10:36:59 droy: not out of the box, no - we are primarily geared towards building OSes based on Linux May 20 10:39:28 ok actually its my summer project which my proffessor gave me i think i have to say no to him May 20 11:35:11 droy: using yocto/oe would really not fit the task May 20 11:35:23 have a look at eg. http://processors.wiki.ti.com/index.php/MSP430_Real_Time_Operating_Systems_Overview May 20 11:35:46 some open source rtos (real time operating system) are mentioned there May 20 11:35:54 may be a start for research May 20 11:36:19 you may be able to use bitbake, to ease compilation May 20 11:37:20 but i would advise you to start simple, with shellscripts, later Makefiles, then, if you reached a state where the system works, you may want to investigate using an embedded build system like bitbake May 20 11:37:41 good luck May 20 11:57:03 thanks simmel80 May 20 12:14:17 droy: you are welcome, and sorry for posting the link you already found again, i must have been sleeping ;-) May 20 12:16:07 thanks for help May 20 12:43:35 Hi, I'm only two days into Yocto, customising an image for an imx6 board using Dora. Basics are all good, and I've been able to add openjdk to an image and confirm that it is runs on our target hardware. Now I'm trying to create a new recipe that installs our java application and adds an init script, but it is erroring out when I attempt to build the image. May 20 12:44:03 Can anyone here point me in the right direction if I give little more info? May 20 12:50:48 CaCtus491: sure, if you pastebin the error we can try May 20 12:53:10 Thanks. I'm sure it is something pretty simple. All I'm trying to do is create a recipe that puts the contents of a tar.gz file into a specific location in the final image. May 20 12:54:10 http://pastebin.com/1C2rXHwE for the recipe May 20 12:55:01 General pointers welcome too :) May 20 12:55:12 The error occurs when trying to build the final image: May 20 12:55:14 | Error: etrac-client not found in the base feeds (varsommx6q cortexa9hf-vfp-neon-mx6 cortexa9hf-vfp-neon cortexa9hf-vfp armv7ahf-vfp-neon armv7ahf-vfp armv6hf-vfp armv5ehf-vfp armv5hf-vfp noarch any all). May 20 12:56:24 what's the recipe file name? is it parsed by bitbake? May 20 12:57:34 Recipe appears to be parsed by bitbake because if I add an intentional typo to the SRC_URI varilable, it complains about not being able to find that source file May 20 12:58:25 path to the recipe file is ./meta-etrac/recipes-etrac/etrac-client/etrac-client.bb May 20 12:58:39 (part of a new layer I'm creating for our project) May 20 13:01:39 I've just been poking around in the build area, and it seems like it has extracted the contents of the .tar.gz file to tmp/work/cortexa9hf-vfp-neon-poky-linux-gnueabi/etrac-client/1.0-r0, which is mushed up with other directories such as sysroot-destdir, pkgdata, etc May 20 13:02:26 yes, that's where the recipe is 'built'. May 20 13:02:30 I suspect it should have been extracted into a sub directory in there? Not too sure, a bit too new May 20 13:03:18 CaCtus491: that depends on the structure within the tar.gz file May 20 13:03:39 there is an option to put it explicitly into a subdirectory if none has been used in the tar.gz, one sec May 20 13:04:40 the tar.gz has no subdirectory defined. I'm looking to put the contents of the tar.gz directly into /usr/local/etrac May 20 13:04:45 just append ;subdir=yourdesiredsubdirname to the entry in SRC_URI for the tar.gz May 20 13:05:14 ah, ok. I'll give that a go. May 20 13:05:16 right.. you probably generated 'empty' packages then. May 20 13:05:21 hence they don't exist. May 20 13:05:33 then you can copy the contents from ${WORKDIR}/yourdesiredsubdirname to /usr/local/etrac within do_install (or a do_install_append if you're relying on some default implementation) May 20 13:09:27 ok, that's better, the extracted files are going into the subdir, but now I get the following error: May 20 13:09:28 ERROR: QA Issue: etrac-client: Files/directories were installed but not shipped May 20 13:09:38 May 20 13:09:48 ERROR: QA run found fatal errors. Please consider fixing them. May 20 13:09:48 ERROR: Function failed: do_package_qa May 20 13:09:48 ERROR: Logfile of failure stored in: /mnt/yocto/build_mx6q/tmp/work/cortexa9hf-vfp-neon-poky-linux-gnueabi/etrac-client/1.0-r1/temp/log.do_package.9174 May 20 13:09:48 ERROR: Task 489 (/mnt/yocto/sources/meta-etrac/recipes-etrac/etrac-client/etrac-client.bb, do_package) failed with exit code '1' May 20 13:13:57 CaCtus491: ok, so that means the files were installed but weren't included in any package May 20 13:14:23 you need to set/add to FILES for a particular package to have them picked up by that package May 20 13:15:11 ok, so I should add a FILES_${PN} += ... to my recipe? May 20 13:15:34 (reading about the FILES variable in the mega manual now) May 20 13:17:37 ..that seems to have gotten past that error, waiting for the rootfs to build now - looking good so far May 20 13:17:51 CaCtus491: yes, assuming it's the main package you want those files in May 20 13:20:14 I'm still stuck on nativesdk-omniorb. Do anyone have any patch for this ? May 20 13:21:29 can you pastebing the current recipe May 20 13:21:36 I'm running some other builds atm May 20 13:21:42 but I can run it in a bit May 20 13:22:10 bluelightning / ndec: Thanks heaps. The files appear in the expected location in the final image. May 20 13:22:17 CaCtus491: np May 20 13:22:59 Now to look at adding the init script, I think I saw an example on this in the manual somewhere - will start with that and see how we go... May 20 13:26:49 base recipe : http://pastebin.com/mENXqJ96 // append : http://pastebin.com/3w3r2PTx // patch file : http://pastebin.com/MW9qW3Nn May 20 13:27:23 for omniorb May 20 13:33:30 I want to create an image which just contains one package. The reason for this is because the embedded system I'm using has multiple flash partitions and I want to create an ext3 image for each. I'm using "inherit image" in the recipe for the image but the final image is getting clutter with libc, base-passwd, busybox etc. May 20 13:33:56 Is there a way to limit the image to just one RPM package May 20 13:42:08 dguthrie: looking at image.bbclass, you probably want to set PACKAGE_INSTALL to avoid the fundamental packages May 20 14:16:15 hi guys, i just want to wish you all the best, this would be my last meetings as an engineer at ENEA, i will still be involved in open-source projects but right now i am going for a short vacation, good luck to all May 20 14:18:35 vali_at_enea: thanks for your contributions, all the best :) May 20 14:23:29 hello guys May 20 14:23:50 I have o problem with the following bug: https://bugzilla.yoctoproject.org/show_bug.cgi?id=6203 May 20 14:23:51 Bug 6203: normal, Medium, 1.7, vaduva.jan.alexandru, REOPENED , perl ptest installed in images without ptest-pkgs in IMAGE_FEATURES May 20 14:24:00 do you have any suggestions May 20 14:26:46 AlexVaduva: try searching the metadata for where perl-ptest gets pulled in, and figure out why it happens May 20 14:26:56 AlexVaduva: (its probably similar to the dbus situation) May 20 14:29:59 rburton: I alreaddy tried this May 20 14:30:24 rburton: but no visible results May 20 14:39:08 rburton: I also tried integrating this patch - http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=a2e612c7763002c2fdc5f24cd015a8fd5b12b1e0 May 20 14:43:05 rburton: and do the changes accordingly, just for creating a more clear image over the 'ptest' enabled feature, May 20 14:43:58 rburton: I believe that I I need is to disable the perl test install when ptest disabled, but I do not know the way to do this May 20 14:44:20 if ptest is enabled at the distro level you need to *build* ptest packages May 20 14:44:26 if its enabled at the image level you *install* packages May 20 14:48:19 i know the difference May 20 14:48:27 it is enabled in distro May 20 14:48:32 but not into image May 20 14:49:07 so it shouldn`t appear in mi package dirs May 20 14:49:24 but it does May 20 14:56:32 YPTM Bridge: Participant passcode: 42001078 Dial-in number: 1.972.995.7777 US Toll Free number: 1.877.561.6828 May 20 14:56:40 YPTM Stephe Joined May 20 14:57:33 YPTM Saul is on early for a change! May 20 14:58:11 YPTM: Paul Eggleton is on May 20 14:58:41 AlexVaduva: by your package dirs, what do you mean exactly? May 20 14:58:59 YPTM: Bruce Ashfield is on. May 20 14:59:11 YPTM: Tom Z on May 20 14:59:15 YPTM: Beth Flanagan on May 20 15:01:25 YPTM: Matthew Weigel on the call May 20 15:01:30 YPTM: nitin on the bridge May 20 15:01:36 YPTM: armin May 20 15:01:46 YPTM: Scott Rifenbark joined. May 20 15:02:43 YPTM: AlexG on the call May 20 15:02:51 YPTM: ross joining May 20 15:02:59 YPTM: 1.5.2 status. OPW intern starting. May 20 15:03:07 YPTM: Richard has joined May 20 15:04:25 YPTM: Sean Hudson has joined May 20 15:04:40 YPTM: Michael on the call. May 20 15:07:17 pidge: we will eventually want the meta-amd bsp released, but I don't think we're ready for that yet May 20 15:09:35 darknighte: ok, let me know when you're ready. This has guidance: https://wiki.yoctoproject.org/wiki/Third_Party_BSP_Release_Process May 20 15:10:07 pidge: right. thx. May 20 15:11:10 pidge: I didn't quite catch the name. Was it Lin? May 20 15:11:18 lynn May 20 15:11:23 ahh, thx. May 20 15:12:45 out of curiousity, does lynn have a nick on IRC? May 20 15:13:31 * LCyrin aka Lynn shows up in the IRC May 20 15:14:00 * darknighte sheds a tear for dave's departure May 20 15:14:06 waves at LCyrin May 20 15:14:48 welcome LCyrin May 20 15:14:55 welcome LCyrin May 20 15:15:55 halstead: nice :) May 20 15:16:08 * rburton googles amd, and thinks that Bald Eagle isn't a great name for a produt May 20 15:16:43 :) May 20 15:16:52 * LCyrin -bows reverently- May 20 15:17:04 rburton: no dissing on the name now. (although, i didn't name the things) May 20 15:17:24 darknighte: the pcmag article mentions yocto though :) May 20 15:17:33 Welcome LCyrin. May 20 15:17:34 rburton: "The bald eagle is both the national bird and national animal of the United States of America." May 20 15:17:34 YPTM is over May 20 15:17:41 rburton: so some might disagree ;) May 20 15:18:09 rburton, probably keeps thinking of: http://en.wikipedia.org/wiki/Eddie_%22The_Eagle%22_Edwards May 20 15:18:18 bluelightning: so where do you go from there? super bald eagle? May 20 15:19:06 rburton: not sure, I'm not in marketing :) May 20 15:19:31 darknighte: i should probably point out that obviously the Intel(R) Atom(TM) solutions offer better performance per dollar, or something May 20 15:19:35 zecke: hah! nice. May 20 15:19:53 zeddii: eddie is awesome :) May 20 15:19:59 rburton: them's fightin' words!!!! or something. May 20 15:25:33 * darknighte just *now* notices that tab completion got him again May 20 15:25:44 * darknighte quietly replaces zecke with zeddii May 20 15:28:35 darknighte: too bod for me May 20 15:28:58 * zeddii answers to almost anything May 20 15:30:09 just don't call zeddii late for dinner? ;) May 20 15:30:27 darknighte, I was going to make that joke .. but wasn't sure who else would get it :P May 20 15:31:13 scary. we may be in a similar mental space. May 20 15:31:21 * darknighte gives a quick shudder at the thought May 20 15:33:22 indeed :) beers at a conference to dicuss the horror! May 20 15:47:30 zeddii: you're on. I'll buy the first, if you buy the 2nd. May 20 15:47:49 * darknighte checks conference schedule May 20 15:49:47 elc-e looks next for me. May 20 15:50:17 zeddii: see you there then May 20 15:57:08 zeddii: I am tentative, but will submit a talk to ELCE. If it gets accepted, I will most likely see you there. May 20 15:57:22 sounds like I might see rburton there too. :) May 20 16:21:17 bluelightning: package and packages-split from tmp/work//perl/5.14.3-r1/ May 20 16:21:36 AlexVaduva: which directory under packages-split? May 20 16:23:21 I just gave an example. I know it appears in package May 20 16:23:38 AlexVaduva: yes, but that's the point - it absolutely will appear there May 20 16:23:48 but you asked what type of packages I am reffering to May 20 16:23:54 so I gave you an example May 20 16:24:04 why? May 20 16:24:56 the critical question is which package does it appear in May 20 16:25:06 isn`t the whole purpose of not having the ptest image feature to make sure it doesn`t May 20 16:25:25 yes, but we're in the process right now of determining how it's getting there May 20 16:25:27 and only appear in the image dir May 20 16:26:02 if you are talking about under the workdir, everything in image/ should appear in package/ and packages-split/ May 20 16:26:23 bluelightning: this is where i think WORKDIR/image should be renamed to avoid confusion May 20 16:26:35 rburton: definitely, it's silly naming May 20 16:26:48 WORKDIR/staging seems more obvious? May 20 16:27:08 rburton: staging wouldn't be ideal either because it's not directly related to staging to the sysroot ;) May 20 16:27:21 something like "install" would work May 20 16:27:39 rburton: I believe so, but wont that be a confusion with staging kernel task May 20 16:28:13 AlexVaduva: so do the files appear under a "perl-ptest" directory under packages-split, or some other directory? May 20 16:29:05 ./package/usr/lib/perl/ptest ./packages-split/perl-ptest/usr/lib/perl/ptest ./sysroot-destdir/usr/lib/perl/ptest ./image/usr/lib/perl/ptest May 20 16:29:15 yes it does May 20 16:31:07 should I make this package empty with somethin like: ALLOW_EMPTY_perl-ptest = "1" ? May 20 16:32:36 AlexVaduva: no, ALLOW_EMPTY tells the system to create the package if it is empty, not make it empty May 20 16:33:00 AlexVaduva: fundamentally, if ptest is in DISTRO_FEATURES then it is correct for the package to be created May 20 16:33:58 AlexVaduva: if it appears in the final image without you explicitly adding ptest-pkgs to IMAGE_FEATURES, that is a problem - is that what is happening in your case? May 20 16:34:11 I assumed that, but I wasn`t sure about ALLOW_EMPTY purpose May 20 16:34:27 I do not believe so May 20 16:34:51 it is not in image features May 20 16:35:11 AlexVaduva: but it is appearing in the final image then? May 20 16:35:20 yes May 20 16:35:31 ok, so then the thing to determine is how it is getting there May 20 16:35:56 one way to do that would be to enable buildhistory and look at the dependency graphs produced by it May 20 16:37:15 ok, thanks I will try that May 20 16:37:58 http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#maintaining-build-output-quality May 20 16:41:31 thanks for the link, I already did that and started a build, I will let you know about the progress after the build is finished, it mak take a while May 20 18:00:23 Jefro: hah. I should have read your email about vote status more closely. May 20 18:30:35 * nerdboy launches an udoo May 20 18:33:58 should i assume meta-fsl has a machine for this thing? May 20 21:06:09 bitbake has completely lost its mind on me May 20 21:16:47 what changed? **** ENDING LOGGING AT Wed May 21 02:59:58 2014