**** BEGIN LOGGING AT Tue Mar 31 02:59:59 2015 Mar 31 03:02:38 I've had issues before trying to boot without genericx86-64, unfortunately its been awhile and I've forgotten the solution Mar 31 03:03:33 is this your first build? Mar 31 03:06:06 what you mean by first build? Mar 31 03:06:29 i have tried bitbaking several times...rebooted several times but no luck Mar 31 03:11:00 have you build a successful image yet? Mar 31 03:13:43 only with core-image-minimal Mar 31 03:14:05 core-image-base and core-image-sato-sdk fail Mar 31 03:14:45 I don't think core-image-base will boot, I think to boot you need atleast minimal, but I've never tried the sato images Mar 31 03:16:59 I thought core-image-minimal is smaller than core-image-base? Mar 31 03:19:31 ah, you're right Mar 31 03:20:36 do you know how to configure serial console to get the kernel boot log? Mar 31 03:23:03 redengin: afaik u will need to start up something like putty or some console that can read from serial console Mar 31 03:23:27 then u will need to configure the bauds and so on.... Mar 31 03:23:43 virtualbox will allow you to configure serial to dump to a file Mar 31 04:43:49 redengin: ok..will try that Mar 31 04:43:50 thanks Mar 31 11:34:17 hello Mar 31 11:37:50 I have a recipe, let's call it abc_1.0.bb. I want to add 2 different abc_%.bbappend files in two different machine specific layers (for two different machines) and I want each one of those to be used only for one machine. How can I do that? Mar 31 11:38:23 At the moment, I have both machine specific layers in the bblayers.conf and I want it to stay like that. Mar 31 11:38:33 Can anyone help me? Mar 31 11:40:17 wullewaisch, well, in that case both abc_%.bbappend will be parsed and processed. You can have the respective bbappen only modify things for it's MACHINE, though. Mar 31 11:42:59 AndersD: it would be also ok if I could configure bblayers.conf.sample to be machine specific, but I am not sure if that's possible...if that could work, I could than add each machine specific layer to its machine Mar 31 11:43:54 AndersD: do you have any other idea how to solve this problem? Mar 31 11:47:15 AndersD: I tried using COMPATIBLE_MACHINE, but that doesn't work for some reason. Mar 31 11:47:17 What are you doing in those two bbappends Mar 31 11:47:39 nothing much...only applying different patches for two different machines Mar 31 11:48:21 Well in that case use something like: SRC_URI_append_machine = "file://...." Mar 31 11:48:45 That will only append the listed patches when building for "machine" Mar 31 11:49:07 Doing it like that allows you to keep both bbappends enabled at the same time Mar 31 11:50:10 ok, in that case I don't need those bbappend files at all Mar 31 11:50:52 Not if the recipe in question is in your own layer Mar 31 11:51:26 it is in my own layer Mar 31 11:51:36 ok, great, you helped me a lot, thanks!!! Mar 31 11:52:21 I will let you know if it works :) Mar 31 11:54:23 hi, what is the difference on some pakages if the -dev package occurs in the sdk with populate_sdk and some pakages the -dev package is not included in the sdk, how is this difference done ? Mar 31 11:57:41 grma, if the package is included in the image being built, then it's -dev package should be included in the SDK. Mar 31 11:58:16 Ok, than i do something wrong Mar 31 11:58:38 any particular package that's not being included? Mar 31 12:04:40 AndersD: i saw it eg at libtinyxml2, does it make a differnece if i install something with IMAGE_INSTALL or CORE_IMAGE_EXTRA_INSTALL ? Mar 31 12:07:25 No, that shouldn't really make any difference. Mar 31 12:10:25 and IMAGE_INSTALL += "blah" is the same like IMAGE_INSTALL_append = "blah" ? Mar 31 12:12:04 In this regard yes. Mar 31 12:17:36 grma: with the _append you will want to have a space before 'blah' Mar 31 12:17:45 since that's appending without adding a space, IIRC Mar 31 12:17:52 _append = " blah" Mar 31 12:18:30 ahh ok, thx Mar 31 13:00:22 hi all Mar 31 13:03:10 why Im not able to apply a kernel config fragment without having issues with do_kernel_configcheck (invalid.cfg/required_redefinition.txt/missing_required.cfg) ?? help appreciated Mar 31 15:29:46 * kergoth hopes the pending os-release fixup and RP's appends determinism fix, and possibly also the sstate fetch fix, improve his sstate reuse problems Mar 31 15:54:18 maybe someone can help me... I've been bashing my head against the wall for a few days on something that seems to simple Mar 31 15:54:38 I've got a custom image, and I'm building an SDK using bitbake -c populate_sdk my-image Mar 31 15:55:03 is there a way to include x86 tools in the populated sdk? Mar 31 15:55:09 like... my own tools Mar 31 15:55:19 i've got -native packages for them that work great Mar 31 15:55:23 add nativesdk packages to TOOLCHAIN_HOST_TASK Mar 31 15:55:30 native won't do, nativesdk will. add to your BBCLASSEXTEND Mar 31 15:55:36 aha! Mar 31 15:55:46 will give it a shot Mar 31 15:56:08 see lines 33 and 34 of populate_sdk_base.bbclass Mar 31 15:56:12 and 35 Mar 31 15:56:17 those control what packages go into the two sysroots in the sdk Mar 31 15:57:59 awesome! I was so close :D Mar 31 15:58:13 had a bunch of notes written down about those lines but hadn't quite figured out what they did yet Mar 31 15:58:52 how on earth did you know the line numbers so quickly?! Mar 31 16:02:01 kergoth: thank you so much! It's now complaining about not being able to install the nativesdk- version of my package, which is good enough to start debugging. Much better than just flat out ignoring the package :D Mar 31 16:02:25 that's something I've successfully debugged before Mar 31 16:14:20 tonywin32: as i said, add nativesdk to BBCLASSEXTEND, just as native is Mar 31 16:14:29 that's all you need to do Mar 31 16:43:44 kergoth: turns out the package itself doesn't like the environment :) Mar 31 16:44:18 and that's not a yocto problem... that's a "dumb package build" problem Mar 31 16:44:29 trying to be too clever Mar 31 16:59:44 moin Mar 31 17:17:29 this gphoto build is going to kill me. why do things do strange things?! Mar 31 17:24:21 bmaptools is a huge win for copying disk images to sd cards Mar 31 17:25:18 sadly it appears the output of wic may not be sparse Mar 31 17:25:27 need to study more Mar 31 18:17:31 When I build an image with package-management in the IMAGE_FEATURES, should I not expect that opkg list-installed should show me the packages that are installed by the IMAGE_INSTALL settings so that I can easily upgrade those packages in the future? Mar 31 18:18:23 that's the idea, yes Mar 31 18:20:16 i hate when -S printdiff only shows me a bunch of do_fetch's, with no details Mar 31 18:22:44 then -vDD ? Mar 31 18:22:53 (if possible) Mar 31 18:37:43 RP: when -S printdiff only displays a list of fetch/unpack tasks with no details, what's the usual suspect, missing intermediate task sigdata? Mar 31 18:37:57 hmm Mar 31 18:38:31 * kergoth digs Mar 31 18:59:08 Cardoe: worked last time for me... Mar 31 18:59:28 IMAGE_FEATURES += "splash package-management ssh-server-openssh" <= in local.conf Mar 31 20:35:35 kergoth: A test case for trying to improve this would be really really useful when you do figure it out... Mar 31 20:35:52 kergoth: I don't remember what that means offhand :( Mar 31 21:04:15 k Mar 31 22:40:24 If i want to specifically tell the configure script where a header is.. say libjpeg.h, what $VAR would that be? Mar 31 22:40:37 in a recipe Mar 31 22:58:06 figured it out Mar 31 22:59:04 rburton: i got one step closer, do_configure works, now i'm getting errors on do_compile Mar 31 23:05:43 wait.. so if it the yocto tells the compiler that there is a sysroot, shouldn't any -I/usr/bla be at /usr/bla? Mar 31 23:14:16 and… does anyone know what package would provide predefs.h? it doesn't seem to be in the sysroot include dir Mar 31 23:17:02 looks like stdc-predef.h might be the file.. something keeps telling this program to look in local include locations instead of using the sysroot Mar 31 23:34:36 figured it out as well… i just keep quiet about my last problem and continue to work it out ;) Apr 01 00:24:29 anyone getting weird problem of /sbin/init missing Apr 01 01:26:23 damnit, need to fix sstate for my split recipe setup in meta-sourcery, forgot that EXTERNAL_TOOLCHAIN needs to cause *-external-cross to rebuild, since they add links / wrap binaries in that path, yet recipes depending on those shouldn't rebuild in such a case Apr 01 01:26:26 oops Apr 01 01:52:15 * kergoth grumbles Apr 01 01:53:02 Do we have a tool that will let you feed in a variable name, and a target, and have it give you the recipes/tasks which will have their sigantures changed as a result, and everything that depends on those? I think the closest is to bitbake -S none target; modify the variable; bitbake-whatchanged -v target Apr 01 01:58:16 zeddii: around ? Apr 01 01:58:38 if you have time to kill, then I have a future problem for you Apr 01 01:59:03 linux-yocto 3.19 does not boot on qemuarm when built with gcc5 Apr 01 01:59:26 you can use my contrib master tree Apr 01 01:59:45 http://git.openembedded.org/openembedded-core-contrib/log/?h=kraj/master Apr 01 02:00:01 to get 5.0 gcc recipes Apr 01 02:01:32 * kergoth grumbles some more Apr 01 02:02:09 heh Apr 01 02:02:31 well I have clang too if you have more time Apr 01 02:02:49 and didnt mention musl as well if you have 25th hour Apr 01 02:03:01 need help on all fronts **** ENDING LOGGING AT Wed Apr 01 02:59:58 2015