**** BEGIN LOGGING AT Tue Jan 02 03:00:02 2018 Jan 02 06:56:28 how can we define the sequence of patch applying, if we multiple patches and we want to apply them in specific order... Jan 02 07:01:36 xtron: patches are applied in order they are found. You can specify the sequence in SRC_URI. Jan 02 07:06:34 anujm: ok, I give it another chance, with numeric order 0001, 0002 ... **** BEGIN LOGGING AT Tue Jan 02 08:55:01 2018 Jan 02 10:33:16 Umh, with Yocto 2.4 the ubifs image egneration is failing due to 'mkfs.ubifs: not found'. Do I need to introduce a dependency for mkfsubifs-native? Jan 02 10:37:57 ttllkk_: probably Jan 02 10:42:15 Umh, I already have 'mtd-utils-native:do_populate_sysroot' as dependency.... I think something changed in the image generation rpocess Jan 02 10:42:18 I will keep checking **** BEGIN LOGGING AT Tue Jan 02 12:23:05 2018 Jan 02 13:10:31 I'm having problems building a wic recipe on rocko (that works fine on pyro). It fails in do_image_wic with a py traceback expecting a file that does not exist. My disk image is straight: A ext fs on partition 1. No special boot loaders. Isn't there another image target than wic for this? Jan 02 13:20:57 New news from stackoverflow: Unable to run 64bit app in multilib enabled yocto || downgrading QT-5.7 to QT-5.4 in Yocto( morty 2.2) for nitrogen6sx Jan 02 14:35:42 <_lucko> Hi, I have a question about Yocto Eclipse plugin. I am using it with cmake. I think that there is an issue with generated file toolchain.cmake. For the variable CMAKE_FIND_ROOT_PATH, there is the first path to meta-ide-support/1.0-r3/recipe-sysroot-native/ and the second path is to tmp/sysroots/ (this is used as the sysroot in plugin settings). I think that the first path should not be there at all.) Jan 02 14:51:55 <_lucko> It is related to source code - plugins/org.yocto.cmake.managedbuilder/src/org/yocto/cmake/managedbuilder/YoctoCMakeMakefileGenerator.java:239 Jan 02 14:53:31 <_lucko> The question is if "OECORE_NATIVE_SYSROOT" should be in the list findRootPathValues. And, if it has to be there, is the first position the correct one? Would not it be better to have it as the second one after "OECORE_TARGET_SYSROOT". Jan 02 14:56:15 _lucko: there are two sysroots, one contains native binaries which run on the build host, the other are target binaries Jan 02 14:56:35 _lucko: I don't know enough about cmake to know which ones should be used by which functions though Jan 02 14:58:34 <_lucko> RP: yeah, exactly. And the variable CMAKE_FIND_ROOT_PATH describes where the libraries (by find_library) should be looking for. So, imo it should look for them only in the path with target binaries Jan 02 14:59:14 _lucko: that sounds right... Jan 02 15:06:52 <_lucko> moto-timo: Hi, I see that you have the latest commits in eclipse plugin. Can you please take a look on my question a few lines above? Thank you Jan 02 15:56:28 YPTM: - dial into: 1-800-262-0778 - enter the number: 88748961# Jan 02 15:56:41 YPTM: Stephen Joined Jan 02 15:57:23 YPTM: Armin is on Jan 02 15:57:51 YPTM: Joshua Watt here Jan 02 15:58:05 YPTM: Stephano joined Jan 02 15:58:30 YPTM: Joshua joined Jan 02 15:58:36 YPTM: David Reyna here Jan 02 15:59:07 YPTM: Richard joined Jan 02 16:00:51 Aaron Joined Jan 02 16:00:51 YPTM: Mark is her Jan 02 16:00:54 here Jan 02 16:01:21 YPTM Michael joined Jan 02 16:04:14 halstead, the new AB ran out of disk space over the weekend. Will it auto correct? Jan 02 16:04:35 * armpit nm Jan 02 16:04:51 armpit no. Alright. Jan 02 16:05:19 YPTM: Randy joined. Jan 02 16:07:54 YPTM: wmills (TI) joined Jan 02 16:09:50 Would those changes make it easier to reproduce the same environment/tests the AB does? Jan 02 16:10:19 that's part of the motivation Jan 02 16:16:39 or you can choose to archive the whole machine env Jan 02 16:18:19 YPTM: is over. Jan 02 16:18:21 YPTM: left the room Jan 02 16:18:28 dreyna: I saw that toaster fix, I assume I should just pull that ASAP? Jan 02 16:18:36 * fray goes back to vacation Jan 02 16:18:37 RP; yes please Jan 02 16:19:43 dreyna: done. Not sure how that patch was corrupted like that :/ Jan 02 16:21:28 RP: I think that when I created the "V2 " patch email the paste key got excited, and that you pulled from the email rather than the branch pushed to "poky-contrib" Jan 02 16:21:58 dreyna: I do pull from the emails rather than repos... Jan 02 16:22:26 RP: now I know to double read my emails :-) Jan 02 16:27:42 dreyna: I tend to use the commandline and not edit them :) Jan 02 18:01:27 Hey all. Building Qt app recipes using Rocko. I am copying the format of 'cinematicexperience' recipe to get off the ground. However, it isn't properly populating recipe-sysroot/ on the custom recipe, so do_configure fails. My understanding is do_configure is handled by recipes-qt/qt5/qt5.inc which is also included by my custom recipe so I'm not sure what may be happening. Jan 02 18:03:16 build failure: https://pastebin.com/wBjrcapf Jan 02 18:07:34 https://pastebin.com/39GyPyup Here you can see the folder linux-oe-g++ exists for both recipes, but the working recipe [cinematicexperience] has it in recipe-sysroot-native as well as recipe-sysroot. Jan 02 18:07:53 All instances of that folder have a qmake.conf Jan 02 18:08:04 Any questions or suggestions appreciated. Jan 02 19:44:52 if I see (skipped) next to a recipe in bitbake-layers show-recipes, what is a good way to deetermine why its being skipped? Jan 02 19:46:45 nm tracked it down with bitbake -e, but the strange thing is its selecting a kernel from a layer with a lower priority... Jan 02 19:47:39 theres a set from a machine .conf file to linux-tn-imx for PREFERRED_PROVIDER_virtual/kernel_mx6, but then bitbake -e shows a 2nd operation setting it from an include file, to a different recipe Jan 02 19:47:52 but the machine conf is in the bsp layer which has the highest priority Jan 02 20:28:15 The most frustrating part of this build failures is they built without issue in Dizzy and Morty releases. :( Jan 02 20:28:22 Can't find the change. Jan 02 20:34:12 majuk: we build qt apps. On rocko too, but currently struggeling with image probalems, but the qt apps build fine. What is your problem? Jan 02 20:36:40 majuk after morty, sysroots became recipe specific Jan 02 20:37:28 so if you do not have the recipe thta populate the file it cannot find in the sysroot listed as a DEPENDS, the sysroot will not have the fil Jan 02 20:39:04 sveinse: I'm trying to build the example WebEngine browser. I used the cinemaexperience recipe as a template [and it builds without issue] build failure: https://pastebin.com/wBjrcapf Jan 02 20:40:05 Crofton|work: Interesting. I'm trying to parse your statement. I was leaning on the idea that cinemaexperience recipe builds fine, the voodoo behind the setup of the build environment should be handled similarly and result in success. Jan 02 20:41:07 majuk: I'd like to see the cinemaexperience recipe? Do you have an URL to it? Jan 02 20:41:08 Clearly that's not the case but I'm not experienced enough with Yocto to go much deeper than that. Jan 02 20:41:20 sveinse: Yes, one moment Jan 02 20:42:33 sveinse: https://github.com/meta-qt5/meta-qt5/blob/master/recipes-qt/examples/cinematicexperience_1.0.bb Jan 02 20:43:03 sveinse: https://github.com/meta-qt5/meta-qt5/blob/rocko/recipes-qt/examples/cinematicexperience_1.0.bb there's Rocko, previous link was master Jan 02 20:45:00 If I have understood this right, the "require recipes-qt/qt5/qt5.inc" is for qt internal use. An external recipe should simply "inherit qmake5" and set up the proper dependencies Jan 02 20:45:23 is is qmake or is it cmake based? (I dont have any experience with the latter) Jan 02 20:45:49 Let me see if I can strip down my recipe here and share Jan 02 20:46:47 majuk, people with problems post morty likely have subtle recipe specific sysroot issues Jan 02 20:46:57 the obvious problem is missing DEPENDS Jan 02 20:47:11 that only appeared in certain race conditions pre pyron Jan 02 20:48:09 Crofton|work: Yes, I suspect that is my issue, but not sure how to address. The same DEPENDS and RDEPENDS_${PN} strings are in cinematicexperience and my custom recipe... I dunno. Seems like one works, the other should. Jan 02 20:48:23 Yet here I am. Jan 02 20:50:39 poke around in the recipe sysroot and see if th efile is there, then try to figure out what recipe is supposed to popualte it Jan 02 20:50:51 so confirm the problem is it is missing from the sysroot Jan 02 20:52:18 majuk: here is my crudely reduced recipe for qt apps. I agree with Crofton|work, it is often about finding the right dependencies, both for target and native. Jan 02 20:52:27 ^ https://bpaste.net/show/87e406e5b2d1 Jan 02 20:53:35 sveinse: Thanks. I'm adding 'qtbase' to my depends Jan 02 20:53:36 here' Jan 02 20:53:43 Nope, same result. Jan 02 20:54:13 Oh! Ok, here we go, it's different error Jan 02 20:54:28 Project ERROR: Unknown module(s) in QT: webengine Jan 02 20:54:42 ok ok, I'll tinker with this a bit longer. Thanks for y'alls input. Jan 02 20:55:03 you're welcome Jan 02 20:55:08 I think the recipe sysroot is in the work directory, so check there Jan 02 20:55:12 and good luck Jan 02 20:55:24 long term, it make your build stabler Jan 02 20:55:33 not susceptible to floating DEPENDS Jan 02 20:55:42 which led to mysterious build issues Jan 02 20:56:44 Sweeeet, it's building. Yea, needed to add qtbase, that solved with 'no qmake config file' error, then added qtwebengine and we're off to the races. Jan 02 20:58:07 amazing how the end of the year problems are solved after the first of the year Jan 02 20:58:27 speaking of mysterious: Just changed to rocko, and now the wic image generation fails. Is there other ways to generate a partitioned image with extfs on it? No special bootloader needed? Jan 02 20:58:33 majuk, so likely those were added to the sysroot by another recipe Jan 02 20:58:49 that built before yours only by "chance" Jan 02 20:59:18 dependency leakages are always fun! Jan 02 20:59:24 indeed Jan 02 21:00:14 only matched by taskhash issues. Now, that's painful Jan 02 21:03:26 Crofton|work: Yes, you're right, they would have been. That differentiation definitely bit me. Thanks again for the clarification, I definitely would not have figured this out on my own much less understood it. Jan 02 21:04:39 I'm a bit intersted in how cinematicexperience was able to build without at least the qtbase in its depends Jan 02 21:11:29 something in the build chain built first and left enough stuff in the sysroot so it could build Jan 02 21:39:59 wompwomp, qtwebengine recipe failed to build. Its recipe came from meta-qt5 so it shouldn't be my mucking something up at least. https://pastebin.com/apeZvDrd Jan 02 21:40:23 /usr/include/c++/5/string:38:28: fatal error: bits/c++config.h: No such file or directory Jan 02 21:40:38 Definitely have gcc-multilib installed Jan 02 22:23:02 Reinstalled gcc- and g++-multilib packages on host machine and compilation is continuing now. Jan 02 22:49:26 i just ran bitbake-whatchanged *imagename* and killed it with KeyboardInterrupt now bitbake *imagename* runs for about 4 minutes without any output... will it recover? Jan 02 22:50:35 Probably not of interest to anyone, but I made a small crude build-core-image-minimal-poky-for-qemux86 benchmark. I use it to compare build times on various systems. https://github.com/sveinse/yocto-benchmark **** ENDING LOGGING AT Wed Jan 03 03:00:00 2018