**** BEGIN LOGGING AT Mon Dec 15 02:59:59 2014 Dec 15 09:30:42 morning all Dec 15 09:41:27 hi bluelightning Dec 15 09:42:18 hi woglinde Dec 15 10:29:04 morning everyone. Dec 15 10:40:19 moaning Dec 15 12:17:17 morning Dec 15 12:48:39 JaMa, I nuked tmp and sstate, set MACHINE=qemuarm and ran bitbake python-pyqt Dec 15 12:49:13 and doubel checked that sip.h is not installed on the build machine Dec 15 13:02:03 bluelightning, any word on FOSDEM stand? Dec 15 13:09:16 Crofton|work: ok, I'll try to restore the change in master-next before running next test-dependencies build and lets see Dec 15 13:09:44 be sure you have the other patches from the seies Dec 15 13:10:14 and thanks for your attention to detail Dec 15 13:20:37 Crofton|work: I got an acknowledgement, no word on approved / not approved yet (should come today or tomorrow I guess) Dec 15 13:20:54 ok Dec 15 13:20:57 just chcking :) Dec 15 13:21:03 I checked website first Dec 15 13:21:47 right, it's supposed to be today according to the site Dec 15 13:27:50 you guys all going to FOSDEM? Dec 15 13:28:37 I will be Dec 15 13:28:51 I think Crofton|work will be as well since he's running a devroom Dec 15 13:29:08 yep Dec 15 13:29:30 cool, a OE devroom? Dec 15 13:29:35 no SDR Dec 15 13:30:42 i see. are you related to Ettus somehow? Dec 15 13:30:52 I am a contractor for them :) Dec 15 13:31:06 :) Dec 15 13:31:31 I just rememberd you posted example links to meta-ettus, and now you do a devroom on SDR. Dec 15 13:31:40 ah Dec 15 13:31:58 (not spying on you, or anything :D) Dec 15 13:32:32 rofl, someone just asked about stands in #fosdem Dec 15 16:09:41 JaMa : Why in llvm.inc is EXTRA_OECONF += "(...) --enable-targets=host-only ". Does it means that it builds only on my host, not on embedded device? Dec 15 16:10:33 I am asking because after i run eglfbdev i have an error "LLVM ERROR: Program used external function '__aeabi_i2f' which could not be resolved!" Dec 15 16:13:55 spaszkoPL: I don't remember what exactly enable-targets does Dec 15 16:14:10 spaszkoPL: but llvm on target is good enough for gallium-llvm Dec 15 16:21:52 ok Dec 15 16:21:58 thx for reply Dec 15 16:52:01 If an allarch build uses pkgconfig, it uses the machine sysroot, not the native sysroot. That's a bug, right? Dec 15 16:52:40 rtollert: should an allarch recipe even be using pkg-config ? Dec 15 16:53:19 bluelightning: maybe? this came up because I was trying to make the xorg font packages allarch -- they're arch-specific right now, which doesn't make sense Dec 15 16:54:01 bluelightning: and one of the issues I ran into with that was that PKG_CONFIG_PATH etal isn't pointing to where font-util-native is installed to Dec 15 16:54:31 ok, I'm not really sure how you would best solve that to be honest Dec 15 16:54:34 bluelightning: I agree that an allarch sysroot doesn't make a whole lot of sense, but there's nothing preventing an allarch build from using native packages as part of its build process Dec 15 16:55:09 so pointing it at the native sysroot makes a whole lot of sense to me, in and of itself; it certainly makes a whole lot more sense than pointing at the machine sysroot Dec 15 16:58:45 bluelightning: ... oh, I think I see what you're getting at. if you're installing an allarch package on an actual target, you *want* pkg-config to be pointing at the actual (machine) sysroot, because that reflects the reality of the install Dec 15 16:59:23 rtollert: that would be one issue yes Dec 15 16:59:34 what are the others? Dec 15 17:05:52 I haven't thought too hard about all of the issues, that's just one I can see right now Dec 15 17:09:40 I'm still leaning towards pointing at the native sysroot, because a) I don't think there's anyway to build a "fake" allarch package, ie build an arch-specific package and just call it allarch (plus that's ugly); b) building allarch allows these packages to be built only once for all archs, which seems like a significant win to me **** ENDING LOGGING AT Tue Dec 16 02:59:58 2014