**** BEGIN LOGGING AT Mon Nov 21 02:59:58 2011 Nov 21 11:05:26 morning all Nov 21 15:56:20 RP__: runqemu script is defaulting to vnc server and not spawning a separate xterm as it use to. were there some changes ? Nov 21 15:56:27 this is on ubuntu 11.10 Nov 21 16:46:10 khem: I don't know :/ Nov 21 17:38:20 RP__: somehow on this box sdl support is disabled when qemu-native was built Nov 21 17:38:26 RP__: I think thats the problem Nov 21 19:03:19 khem: ah, that could cause a problem, yes Nov 21 19:21:20 RP__: I am still digging more but it seems libSDL is not able to find some versioned symbols which come from libasound.so Nov 21 19:21:48 and these symbols are there in system provided libasound.so but not in one we build with alsa-libs-native Nov 21 19:22:08 and thats the reason sdl detection fails in qemu Nov 21 19:22:13 configure Nov 21 19:22:39 let me see if something changed in alsa-libs Nov 21 19:27:00 hmmm alsa-lib-native configure has --with-versioned=no Nov 21 19:27:02 I wonder why Nov 21 19:28:49 hmmm here we go Nov 21 19:29:20 commit b354eb957ce08ac7814ce46c13ca3a8449b4063a Nov 21 19:29:20 Author: Daniel Lazzari Nov 21 19:29:20 Date: Thu Sep 22 00:29:27 2011 +0000 Nov 21 19:29:20 alsa-lib: Don't use versioned symbols on uclibc builds as it causes strange hangs Nov 21 19:29:23 alsa-lib: Don't use versioned symbols on uclibc builds as it causes strange hangs. Taken from oe-classic. Nov 21 19:29:27 Signed-off-by: Daniel Lazzari Jr Signed-off-by: Richard Purdie Nov 21 19:29:45 so when we do TCLIBC=uclibc bitbake alsa-lib-native this kicks in Nov 21 19:30:08 this commit should have made it only for target recipe Nov 21 19:30:19 and spared native and nativesdk ones Nov 21 20:08:42 if i do a bitbake-diffsigs Nov 21 20:09:03 against two stamps with obviously different SHA's from the file name Nov 21 20:09:14 and bitbake-diffsigs does not say anything Nov 21 20:09:23 what else could be different between the two sigs/stamps? Nov 21 20:29:20 this makes no sense… these hashes should be the same Nov 21 20:31:28 RP__: any feedback on http://git.openembedded.org/openembedded-core-contrib/commit/?h=kraj/uclibc-updates&id=d06edef859ff936294d76bc07523f50eaaa5baf4 Nov 21 22:59:13 khem: ping Nov 21 23:32:12 msm: yes Nov 21 23:33:35 khem: any thoughts on http://bugzilla.yoctoproject.org/show_bug.cgi?id=1357 Nov 21 23:34:08 the toolchain in our tmp/sysroot can not find libgcc if it's moved - this ends up effecting sstate-cache negatively Nov 21 23:44:00 khem: I was meaning to have a look at that source. If it is appropriate, I can get it upstream Nov 21 23:46:58 RP__: OK I think its needed upstream Nov 21 23:48:51 khem: me too, I just need to check some things Nov 21 23:53:37 RP__: OK np Nov 21 23:53:48 msm: hmm, I have to study this problem Nov 21 23:54:02 msm: geneally toolchains should be relocatable Nov 21 23:54:29 msm: but I have to check this given case. Nov 21 23:54:51 msm: update me the gcc in question is gcc-cross rigght ? Nov 21 23:58:17 msm: as I see. We cross access two sysroots Nov 21 23:58:39 msm: usually gcc wants libgcc relative to its own installation Nov 21 23:58:51 which will be under host sysroot Nov 21 23:59:31 then it will be relative to gcc's installation Nov 22 00:00:01 but we have separated the cross bits from target bits nicely when building cross gcc Nov 22 00:27:30 khem: gcc-cross yes Nov 22 00:27:44 khem: it might not be passing —sysroot though per RP's updated comment Nov 22 00:27:55 i'm going to double check this Nov 22 00:28:00 gotta run now... **** ENDING LOGGING AT Tue Nov 22 02:59:58 2011