**** BEGIN LOGGING AT Wed Mar 02 02:59:59 2016 Mar 02 08:17:34 what is yocto.? Mar 02 09:06:47 good morning Mar 02 09:28:42 Hello Mar 02 09:30:05 When bitbake tells me "Could not inherit file classes/bluetooth.bbclass", is it good practice to just copy that file from oe-core to a declared layer? Mar 02 09:47:45 zeechs: its a good idea to use layers with a version that match Mar 02 09:48:09 sounds like you need to stop using master of one layer with an older release of oe-core Mar 02 09:48:56 Hello everyone! I'm new here and I trying to learn about the project. can anybody recommend me the best reference system or example to begin with to make a demo of an embedded x84 application and how to get started? Mar 02 09:54:58 rburton, you caught me hehe, I just had to checkout the right version of meta-qt5. Now it's building without any error, thanks! Mar 02 09:56:50 N-a-N: the yocto docs and poky are a good start Mar 02 09:57:32 ok, thanks rburton! I'll start there Mar 02 10:05:17 morning Mar 02 10:33:20 FOO_append .= "bar" is redundant, right? Mar 02 10:33:42 and should do the same thing as FOO_append = "bar" Mar 02 10:42:47 howto switch opkg from gzip to pigz in bitbake builds? packaging some binaries seems to take ages when gzip is using a single core only. Mar 02 11:04:36 How to remove a patch from SRC_URI list based on ${MACHINE} value ? Mar 02 11:05:06 Does some thing line SRC_URI_remove should work ? Mar 02 11:19:18 I have a problem with tiny receipe I wrote for a c++ cmake project. It downloads and configures the project just fine, however when compiling it does not provide a -isystem for the c++ library, only -isystem /home/builder/poky/build/tmp/sysroots/qemuarm64/usr/include Mar 02 11:20:30 Have anybody exprience with how and why the cxxflags are not forwarded? (The problem is that it can not find c++ stdlibrary headers) Mar 02 12:19:38 hey folks. I can't find meta layer for AVR32 machines. Is it supported? Any links? Mar 02 12:52:26 rburton: thanks :) Mar 02 13:45:42 hello everybody Mar 02 13:46:06 does anybody know about incompatibility between out-of-tree kernel modules and using rm_work.bbclass ? Mar 02 13:46:33 they seem to need a .config that is removed by do_rm_work Mar 02 13:46:39 (from what I can tell...) Mar 02 15:37:02 Hmm, getting lots of -initial manifest not found warnings from nativesdk recipes Mar 02 18:51:34 okay, this is probably the weirdest thing I've yet seen while debugging pseudo Mar 02 18:51:44 I'm having problems with LD_PRELOAD staying set, and... I don't get it. Mar 02 18:51:57 After unsetenv("LD_PRELOAD"), I note two things: Mar 02 18:52:00 1. getenv("LD_PRELOAD") is NULL. Mar 02 18:52:09 2. environ[75] is "LD_PRELOAD=libpseudo.so" Mar 02 18:52:23 I don't understand how this is possible. Mar 02 18:52:32 on an exec there are two copies of the environment.. Mar 02 18:52:38 This is before the exec. Mar 02 18:52:41 the current RUNNING enviromment, and the one being passed to exec Mar 02 18:52:51 execv() just uses "environ", apparently. Mar 02 18:53:03 Which is the current running environment. Or should be. Except obviously, it's not. Mar 02 18:53:08 also getenv/setenv has pseudo wrappers doesn't it? that way we can remove and add to the virtual enviromment without actually doing it? Mar 02 18:53:37 No, we just fix them up right before an exec. Mar 02 18:53:50 And I'm looking at that fixup code, and there's a pseudo_dropenv() which is trying to remove libpseudo.so from LD_PRELOAD. Mar 02 18:54:09 And it seems like it thinks it's done so, but getenv("LD_PRELOAD") disagrees with the actual contents of environ, and I don't know why/how. Mar 02 18:54:10 ok Mar 02 18:54:31 inside of pseudo I know we keep a local copy of the environment (or well did a long time ago) for parts of it.. Mar 02 18:54:37 Yup. Mar 02 18:54:37 this was done to help preserve some of the pieces of it Mar 02 18:54:41 could that be contributing? Mar 02 18:54:48 And actually this one isn't one of those, we just remake it or delete it as needed. Mar 02 18:55:21 Anyway, the short answer is, I can fix it easily enough, because I can pass in an environ pointer, and then that one should in theory get used. Mar 02 19:03:06 okay, so Mar 02 19:03:14 I think that actually explains the mysterious startup problems. Mar 02 19:03:34 So, if the server ends up being invoked with the pseudo wrappers, it tries to rexec itself without them. That works. Mar 02 19:03:36 But! Mar 02 19:04:19 The exec call itself is wrapped, and the exec call wrapper ends up trying to send a message to the server to notify it about the exec. Which causes it to start a server. And that means that we have a server trying to start another server, and I think that's where the problem was coming from. Mar 02 19:05:32 And now my restart logic is working as expected and not producing extra server processes or taking so long to start up. Yayyyyyy. Mar 02 19:27:43 RP: Hi, is "sstate: Allow late expansion of NATIVELSBSTRING" related to sstate_create_package failures in meta-ide-support ? Mar 02 19:28:48 http://errors.yoctoproject.org/Errors/Details/40844/ **** ENDING LOGGING AT Thu Mar 03 02:59:58 2016