**** BEGIN LOGGING AT Tue Dec 02 02:59:58 2014 Dec 02 08:24:37 morning all Dec 02 08:25:02 morning Dec 02 08:25:20 Is it possible to get a native gcc toolchain in the OE generated sdk ? Dec 02 08:25:34 Besides the target compiler that is Dec 02 08:31:32 kroon: why would you want that? Dec 02 08:39:48 mago_, the people using the sdk wants to run test programs on the host as well Dec 02 08:40:33 and the host compiler is too different from the target compiler? Dec 02 08:41:58 yeah, the one shipped in the distro is gcc 4.8.2, and yocto ships 4.9.1. Dec 02 08:42:24 so nativesdk c++ libraries wont play well with distro's gcc Dec 02 08:42:37 s/gcc/g++ Dec 02 08:44:48 kroon: s/yocto/oe-core/ Dec 02 08:45:53 koen, right :) Dec 02 08:46:12 JaMa: Where i can find description which QT version is in each branch of meta-qt5? Dec 02 08:46:26 and morning everyone Dec 02 08:50:31 I guess I could generate the same sdk, but with MACHINE=qemux86_64, instead of the target Dec 02 09:39:08 good morning Dec 02 10:34:09 spaszkoPL: just check the branches Dec 02 10:42:19 hello Dec 02 10:42:30 any openssh expert here Dec 02 10:56:41 need a help on resolving openssh issue Dec 02 10:59:30 * tbr takes 100€/h for "expert openssh consulting", alternatively "your eternal soul" Dec 02 11:03:15 noor, just say what the problem is Dec 02 11:04:08 Crofton|work: we have created rootfs using dizzy branch Dec 02 11:05:05 we are trying to excute command on target using eclipse RSE but connection to the target hangs due to SSH problems Dec 02 11:05:38 we looked into SSH code but it looks like SIGCHLD is not being received Dec 02 11:06:22 are you able to ssh in from the command line? Dec 02 11:07:33 tbr: yes Dec 02 11:07:54 tbr: the problem i described is randon Dec 02 11:08:05 it fails on 3rd or 4th attempt Dec 02 11:09:10 the target board is still alive? have you checked, e.g. over serial? Dec 02 11:09:40 tbr: yes it still alive Dec 02 11:12:32 do you have an actual ssh daemon listening? which one? or is it systemd? Dec 02 11:15:27 tbr: here i sthe status of the ssh service file when it is successfull http://paste.ubuntu.com/9343450/ Dec 02 11:16:02 when it fails we see another child with name sh Dec 02 11:16:25 in that case when we run eclipse again it fails Dec 02 11:16:32 so you have systemd and openssh Dec 02 11:17:21 I believe, whatever your problem is, it's not directly openssh related. Dec 02 11:18:26 tbr: see this http://paste.ubuntu.com/9343470/ Dec 02 11:18:49 it now has sh as a child and I think this was not being terminated Dec 02 11:19:19 look at the journal Dec 02 11:44:10 tbr: there was nothing much in journal Dec 02 11:44:40 is it possbile that kernel is not able to kill sh process started by ssh Dec 02 11:44:56 I mean to say is it possible that its kernel issue Dec 02 11:45:48 that is even less likely Dec 02 12:05:44 Where is description of PACKAGECONFIG variable? I don't see it in http://www.yoctoproject.org/docs/1.6/bitbake-user-manual/bitbake-user-manual.html . Dec 02 12:06:39 spaszkoPL: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#var-PACKAGECONFIG Dec 02 12:07:08 spaszkoPL: the bitbake manual only documents variables understood by bitbake itself, rather than the wider set implemented in OE Dec 02 12:07:32 ok, big thx :) Dec 02 12:07:49 There is a lot of things to learn :) Dec 02 12:08:16 we should probably link to the YP ref manual at the top of the bitbake manual variable reference; at least there is a slightly vague note there Dec 02 12:12:26 bluelightning: hi Dec 02 12:13:12 hi ant_work Dec 02 12:13:19 about this, I was asked yesterday how to dynamically set DEPENDS. I hinted at PACKAGECONFIG ofc Dec 02 12:13:39 but seems that with cmake projects you have to do it in other ways... Dec 02 12:14:03 I finally told he can use base_contains in DEPENDS Dec 02 12:14:38 but maybe there is documentation about this... Dec 02 12:15:17 ant_work: can you be a bit more specific about the situation? Dec 02 12:15:54 he showed me that Dec 02 12:15:55 http://pastebin.com/mXnkt4TU Dec 02 12:16:05 wanting to have conditional DEP on avahi Dec 02 12:17:23 among the many possible examples I pointed to this one Dec 02 12:17:31 commit: http://cgit.openembedded.org/openembedded-core/commit/?id=e668c79de927eff635f29fb5ff001f6b106ccc81 Dec 02 12:18:02 well, it would definitely be PACKAGECONFIG for that sort of thing, which does also work for cmake recipes Dec 02 12:18:25 and elsewhere in the recipe you would use base_contains on the PACKAGECONFIG value rather than DEPENDS Dec 02 12:18:30 (if needed) Dec 02 12:20:31 well, seems you get unknown configure options doing i.e. Dec 02 12:20:32 +PACKAGECONFIG ??= "${@base_contains('DISTRO_FEATURES', 'zeroconf', 'avahi', '', d)}" Dec 02 12:20:32 +PACKAGECONFIG[avahi] = "--enable-avahi,--disable-avahi,avahi" Dec 02 12:21:34 so he's using DEPENDS = "libxml2 ${@base_contains('DISTRO_FEATURES', 'zeroconf', 'avahi', '', d)}" Dec 02 12:21:35 well naturally you must use options which the configure (or cmake) script understands... Dec 02 12:21:54 wondering about EXTRA_CMAKE Dec 02 12:22:09 PACKAGECONFIG will feed into EXTRA_OECMAKE Dec 02 12:22:22 err.. EXTRA_OECMAKE right Dec 02 12:22:24 for cmake it would usually be something like -Dxxxx=1 Dec 02 12:22:29 yep Dec 02 12:23:09 and if there is no option, patch it in and send the patch upstream ;) Dec 02 13:13:24 i have an error "configure: error: cannot enable egl_gallium without Gallium" as an output of do_configure of Mesa recipe. I've added to my conf 'PACKAGECONFIG_append_mymachine = "gallium gallium-egl gallium-gbm gallium-llvm"' and i don't understand what's wrong? Dec 02 13:24:25 missing leading space Dec 02 13:25:28 ah yes my mistake :) Dec 02 13:27:20 ok now i see it. :) Dec 02 13:27:26 thx ^_^ Dec 02 15:08:58 if my hw requires special packaging of the bootloader, where would be the proper place to package it? in the u-boot recipe, by inserting a task after deploy, or by adding a separate recipe that depends on u-boot, and then have my image depend on that recipe? Dec 02 15:09:43 i guess in the latter case, i could also make sure it's always a dependency by adding it to CORE_IMAGE_EXTRA_INSTALL in my machine config Dec 02 16:50:02 JaMa : I'm using meta-qt5 branch dizzy and probably in recipe meta-qt5/recipes-qt/examples/qt5everywheredemo_1.0.bb is error because when i try to build it i have "Fetcher failure: Unable to find revision 9a868f96ee63c21ceda890d8dfc9d33f093d1b6d in branch master even from upstream". Dec 02 16:50:26 otavio : up Dec 02 16:53:28 9a868f96ee63c21ceda890d8dfc9d33f093d1b6d doesn't exist anymore in upstream sources :/ Dec 02 16:54:31 try to build it with latest SRCREV 6178748a6ea34df40a8e3c9ce67137e33383bb0e Dec 02 16:57:02 ok:) Dec 02 16:57:11 for me is all for today:) Dec 02 16:57:15 bye Dec 02 17:14:11 JaMa: the good thing is that he does not let us know if it works or not lol Dec 02 17:15:35 otavio: lets see tomorrow :) Dec 02 18:21:14 Has something changed with source forge downloads recently? Dec 02 18:24:57 Crofton|work: you mean the windows virus thing? Dec 02 18:25:12 swig downlaod got a 404 Dec 02 18:25:28 I didn't mean the wondpws virus thing Dec 02 18:25:35 grr typing is hard Dec 02 18:26:16 wgeting the url from bitbake says 404 Dec 02 18:26:50 no idea, sorry Dec 02 18:27:01 bother Dec 02 18:27:47 crofton is it cold as hell at your place too? Dec 02 18:28:05 so you shiver when typing? Dec 02 18:28:18 heh Dec 02 18:28:22 not really Dec 02 18:40:36 ka6sox, naked ping Dec 02 18:40:48 can we put the swig tarbal in the oe source mirror? Dec 02 18:43:34 so you shiver when typing7 Dec 02 18:43:38 args Dec 02 18:45:13 Croftonsure…do you have a good source? or can you send it to me with a md5? Dec 02 18:47:19 awozniak, I get a 403 on that repo Dec 02 18:47:37 awozniak, can you get them for me and email them to me along with the MD5 Sums? Dec 02 18:47:47 Crofton|work, you know my email :) Dec 02 18:47:54 will do Dec 02 18:48:07 ka6sox: sorry, can't too overloaded with tasks right now. sorry Dec 02 18:48:08 I'll put one in my dropbox after this meeting Dec 02 19:18:40 ffs fuse also fails Dec 02 19:20:05 hmm, may have been transient? Dec 02 19:21:42 ka6sox, seems like it may ahve been a transient issue Dec 02 19:42:36 Crofton|work, lets throw it in there in case it goes away Dec 02 19:43:13 I will emailurls for swig and fues in a bit **** ENDING LOGGING AT Wed Dec 03 02:59:58 2014