**** BEGIN LOGGING AT Tue Sep 24 02:59:59 2013 Sep 24 07:06:40 good morning Sep 24 07:56:26 morning all Sep 24 07:56:36 Are there any good books, blogs, tutorials etc. on oe and or distributions based on it? I'm using oe for few weeks now, but still don't feel comfortable with it and even trivial tasks sometimes take hours.. Sep 24 07:57:29 hi bluelightning Sep 24 07:58:31 kbart: i learned a *lot* from reading carefully the dev and reference manuals. Sep 24 08:01:21 ndec, of course I have read manuals, guess I wouldn't even be able to start without that.. but learning and development process would be MUCH faster and even with some real world, step-by-step examples Sep 24 08:01:44 even=easier Sep 24 08:01:52 yeah... i agree, there is a steep learning curve... Sep 24 08:03:08 gm bluelightning, all Sep 24 08:04:33 i.e. I want to make a simple "hello world" qte application built-in my image, but still can't do that as other qt application recipes are based on x11 or matrix launcher and my initial attempt failed. Sep 24 08:05:37 kbart: well, i have used qt4e too, and i agree that it is something not well documented. Sep 24 08:06:47 kbart: the first tricky thing is to set DISTRO_FEATURES properly for qt4e. i eventually found the solution in meta/conf/local.conf.sample.extended file. Sep 24 08:06:47 # If you want to get an image based on gtk+directfb without x11, Please copy this variable to build/conf/local.conf Sep 24 08:06:47 #DISTRO_FEATURES = "alsa argp bluetooth ext2 irda largefile pcmcia usbgadget usbhost wifi xattr nfs zeroconf pci 3g directfb ${DISTRO_FEATURES_LIBC}" Sep 24 08:07:51 ndec, I have succeeded in building/deploying minimal q4te, but I still don't have an idea how to develop/deploy applications for it with oe Sep 24 08:08:24 you mean building with a recipe, or with SDK? Sep 24 08:09:35 the ideal solution would be SDK for development (so I don't have to rebuild image after every small change) and then, when application is ready for production, use .bb recipe Sep 24 08:10:11 application itself should be on local git repository so sources are the same for both ways Sep 24 08:11:07 kbart: i haven't used the SDK, but i know some people have done it. what kind of problems did you have with SDK? Sep 24 08:12:44 first of all, I don't know how to build it so that it would use the same toolchain/libraries as oe and my image Sep 24 08:14:11 I have crocc-compiled it manually before, but it would be a mess to set up all environment variables and dependencies by hand.. I bet there should be some easier - oe way to achieve that Sep 24 08:15:00 cross-compiled application using SDK I mean Sep 24 08:15:56 kbart: ok, you can create any SDK with 'bitbake -c populate_sdk ' Sep 24 08:16:41 it would create a tarball that contains the cross compiler and tools required to build , including all -dev packages for packages included in Sep 24 08:17:05 that part is described in the ADT user guide. Sep 24 08:18:05 the generated 'tarball' can be installed anywhere on your PC, and comes with an 'init-env' script that needs to be sourced. once sourced, you can call $CC, qmake, ... and all tools are the one you built from OE. Sep 24 08:20:36 ndec, thanks, I'm trying that now, but it will take few hours to build (if not errors, of course) I guess. Sep 24 08:21:04 well, if you have built the image already, it shouldn't be that long. Sep 24 08:21:11 it depends on your PC too ;-) Sep 24 08:21:33 and by the way, could you give me a link to "ADT user guide"? BitBake and OpenEmbedded manuals I use do not mention "populate_sdk". Sep 24 08:22:53 kbart: FYI, the OE manuals are very out-of-date, I'd recommend using the Yocto Project manuals instead Sep 24 08:23:32 they are linked from the OE documentation page: http://openembedded.org/wiki/Documentation Sep 24 08:24:16 bluelightning, do you mean this one: http://www.yoctoproject.org/docs/current/mega-manual/mega-manual.html? Sep 24 08:24:49 kbart: that's the combined manual yes; I generally look at the separate ones but that has the same content Sep 24 08:25:29 heh, I hope that's the missing piece, thank you. Sep 24 08:26:11 no idea how I have managed to missed it up to now.. I did googling for ages. Sep 24 08:28:01 unfortunately google indexes a lot of old documentation for OE Sep 24 08:31:25 morning al Sep 24 10:38:12 bluelightning: ping Sep 24 10:38:20 hi Noor Sep 24 10:38:27 bluelightning: hello Sep 24 10:38:36 bluelightning: how are you doing? Sep 24 10:38:43 Noor: great thanks and you? Sep 24 10:38:53 bluelightning: doing good as well Sep 24 10:39:05 bluelightning: need minor help from you Sep 24 10:39:17 Noor: sure what's up? Sep 24 10:40:31 bluelightning: I think in current oe-core when a kernel module is built a dependency for kernel is autocratically added in that module Sep 24 10:40:53 so how it is fixed when building initramfs Sep 24 10:41:26 coz in that case if a module is added in initramfs it need kernel but initramfs is needed to build the kernel Sep 24 10:41:42 so its circular dependency there Sep 24 10:43:16 I am getting this error Sep 24 10:43:17 Collected errors: Sep 24 10:43:17 | * satisfy_dependencies_for: Cannot satisfy the following dependencies for cryptodrv: Sep 24 10:43:17 | * kernel-2.6.32.9 * Sep 24 10:43:17 https://bugzilla.yoctoproject.org/show_bug.cgi?id=4072 Sep 24 10:43:23 Noor: ^^ Sep 24 10:43:52 Noor:note that the commit did slip in and broke the 'normal case' initramfs Sep 24 10:44:02 w/out resolving the issue apparently :/ Sep 24 10:44:18 thanks ant_work Sep 24 10:44:23 I'll look into this Sep 24 10:44:46 goosh how people remember all bugs and provide the link in a min :) Sep 24 10:45:14 well, you know, our layer is broken after that commit..I'm in touch with Bruce Sep 24 10:48:00 ant_work: you know I was looking at your layer for reference .... instead if build it I thought lets as on IRC first and they point me on particular recipe I'll go there Sep 24 11:02:08 ant_work: so you are saying that meta-intramfs is broken after Implement kernel INITRAMFS dependency and bundling or the earlier commit I talked about ..... sorry for asking clarification Sep 24 11:20:48 yes, the situation is that Sep 24 11:22:16 Noor: temprarly I reinstate this code: http://paste.debian.net/45585/ Sep 24 11:23:47 Noor: for some reasons the do_bundle_initramfs tasks is not run Sep 24 11:27:45 Noor: I guess that the author did put INITRAMFS_IMAGE and INITRAMFS_IMAGE_BUNDLE in local.conf so that both image and kernel are aware of Sep 24 11:29:10 Noor: the second limitation of this approach is that it is expected to build the same virtual/kernel twice...no space for custom kernel recipes with embedded initramfs Sep 24 11:31:08 I hope this will be fixed soon or I'll ask for revert before 1.5 release Sep 24 12:04:06 ant_work: thanks Sep 24 12:43:35 kbart: on the Yocto mailing list Rudolf Streif mentioned he is working on a book about OE/Yocto. the expected release date is early 2014 Sep 24 12:45:39 tlwoerner, good to know, but it's too late for me - my device should be on production by then..;) Sep 24 12:59:52 hmm, hopefully said book does not refer to the POoky build system Sep 24 13:21:14 Crofton|work: https://lists.yoctoproject.org/pipermail/yocto/2013-June/016837.html Sep 24 13:21:53 Crofton|work: what? poky is not a build system? Sep 24 13:21:59 :) Sep 24 13:48:32 Is there an automated way of installing libraries in a generated (meta-toolchain) sdk? Sep 24 13:55:43 RagBal: you can define your own recipe that contains all package you need. then you can run bitbake -c populate_sdk . it would create a custom SDK that contains all -dev packages included in Sep 24 13:56:19 ndec, my hero, thanks =) Sep 24 13:56:42 well, i am just repeating what's in the doc ;-) Sep 24 13:57:43 I didn't know there was a populate_sdk task =) Sep 24 13:58:47 RagBal: "A final method of building the toolchain installer exists that has significant advantages over the previous two methods. This method results in a toolchain installer that contains the sysroot that matches your target root filesystem. To build this installer, use the bitbake image -c populate_sdk command." Sep 24 13:58:48 http://www.yoctoproject.org/docs/current/adt-manual/adt-manual.html Sep 24 14:01:00 Yea it is exactly what I need Sep 24 14:01:18 Found it strange that only libc libraries were included in the meta-toolchain Sep 24 14:02:16 RagBal: meta-toolchain is pretty slim on purpose Sep 24 14:20:09 rburton, it fails because it cannot satisfy dependencies for read-edid-dev, I don't have this when I generate the image itself Sep 24 14:20:22 It also complains about dropbear conflicting with openssh Sep 24 14:20:28 Any suggestions? Sep 24 14:20:50 i can't recall the fix for dropbear and openssh, but they do conflict. Sep 24 14:44:46 hi. i would like to be able to split downloads folder (and probably sstate) by license type. is that something do-able today, if not, do you think it would be a wishful feature for more people? Sep 24 14:45:10 one main use case for that would be to split out 'CLOSED' stuff for example. Sep 24 14:54:47 ndec: if this is for archiving source for license compliance then you may find classes/archive* useful Sep 24 14:54:58 mckoan|away, we (the board) talked a little about usage yesterday. I just got to doing email today Sep 24 14:55:18 bluelightning: ok. will look into that right now Sep 24 14:55:22 ndec: http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#maintaining-open-source-license-compliance-during-your-products-lifecycle Sep 24 15:00:46 bluelightning: thanks. it is useful. Sep 24 15:45:58 rburton, I'm just using the rootfs from the OE build dir now, not generating a sdk. Works like a charm, thnx again Sep 24 17:08:41 Hello everybody. I wonder if it possible to generate an image with a given size (e.g 100 Mb). But when I boot the image, the operating system would see it 2Gb instead. Any help ? Sep 24 17:10:15 vadmeste: hmm. what should the 'real' size be eventually? Sep 24 17:13:44 ndec: 2Gb :D Sep 24 17:13:58 but as I see, parted will not allow that and it's not possible to force it Sep 24 17:14:18 one option would be to resize the root fs upon first boot Sep 24 17:14:49 that could be done from initramfs Sep 24 17:18:13 ndec: when I flash my 100Mb image to the sdcard, will Linux be able to see the real size of my sdcard which is 2Gb ? Sep 24 17:18:21 yes Sep 24 17:18:26 indeed. Sep 24 17:18:43 in fact, that's how it was done for ubuntu installer for pandaboard. Sep 24 17:18:46 AFAIK, dynamically expanding an ext2/3/4 partition is allowed while it's mounted, so it could just be done on boot Sep 24 17:19:07 the image size was 1G to fit 'any' SD cards, and it was resized to the actual SD card size during first boot. Sep 24 17:19:08 i.e. shouldn't need to be done outside in the initramfs Sep 24 17:19:50 ndec, bluelightning: thank you very much Sep 24 17:19:58 bluelightning: yes, that's right. it can be expanded live.. Sep 24 17:20:25 we did it in the initramfs because of some other constrainst, but if it's just extending, it should be fine. Sep 24 17:26:35 is 'dora' the new release name? Sep 24 17:27:27 ensc|w: it is, yes Sep 24 17:28:47 how are these names generated? they begin all with 'd', but I do not see a pattern in the rest Sep 24 17:30:34 ensc|w: it was supposed to be a bit of fun on RP's part, only he knew the pattern originally... but if you google it you can figure out where they come from Sep 24 17:36:29 bluelightning: eheh... didn't know about the name story ;-) Sep 24 17:36:41 but, yes, easy to find with google! Sep 24 17:40:56 I liked the old scheme myself, but we ran out of names with that one Sep 24 17:46:58 i suppose you could use the current list of storm names... Sep 24 22:49:59 oe-core branch names come from where exactly? http://en.wikipedia.org/wiki/Dizzy_(series) ? **** ENDING LOGGING AT Wed Sep 25 02:59:59 2013