**** BEGIN LOGGING AT Tue Mar 05 02:59:58 2013 Mar 05 06:31:13 any suggestions on forcing a mozilla build to find the staging include dir? Mar 05 08:24:09 good morning Mar 05 08:55:31 morning all Mar 05 08:55:34 good morning Mar 05 10:01:25 I am looking at the run.docompile script Mar 05 10:01:30 it sets a bunch of exports Mar 05 10:01:38 then in my makefile I do something like Mar 05 10:02:22 ... Mar 05 10:02:55 CXXFLAGS+=-I ../src Mar 05 10:03:10 but somehow the script's export of CXXFLAGS wins Mar 05 10:03:21 and my additions is not added Mar 05 10:03:56 I was wondering how this could happen if clearly in the run.do_compile script the exports are set first and only then make is called Mar 05 10:27:59 pompomJuice: the '+=' takes effect only when the makefile is the origin of the variable; taking it from env ('-e') or as make argument overrides every internal assignment Mar 05 10:29:07 pompomJuice: I suggest to make it like automake: split flags into mandatory parts (AM_CXXFLAGS) and optional (CXXFLAGS) ones where latter can be overridden by env Mar 05 10:29:57 btw, CXXFLAGS is a bad place for -I; (AM_)CPPFLAGS would be better Mar 05 11:24:05 thanks ensc|w Mar 05 11:24:14 I will do that Mar 05 11:42:17 morning all Mar 05 13:34:04 hi pb_ Mar 05 13:43:50 hi bluelightning Mar 05 13:57:46 pb_, can you work with likewise to make sure he is on the board alias? Mar 05 16:58:17 Hi. Anybody getting an error like http://paste.debian.net/239956/ when building nautilus on danny? I do have bcd942a3129c6621486973d3cd92cc1489a57547 on my tree. It seems that gtk headers get installed under target's sysroot /usr/include/gtk-2.0/gtk/gtk.h, but there's no -I/usr/include/gtk-2.0 anywhere in nautilus' build logs. Mar 05 20:19:05 raspberry console-image doesn't power on hmdi port? Mar 05 20:23:06 well, hopefully i'll have a new nss recipe shortly Mar 05 20:23:21 still not sure why it's not in the tree anymore... Mar 05 20:26:24 mr_science: a current nss recipe is in meta-browser I believe Mar 05 20:27:04 hm I think nss should be in meta-oe Mar 05 20:36:31 maybe, if other things use it... right now I guess nothing other than firefox does Mar 05 21:00:21 not true, at least in my case Mar 05 21:00:25 hello. what does the "$@" bit do here? http://fastr.github.com/articles/WAF-on-OpenEmbedded.html Mar 05 21:00:40 so it's an internal build dep for us now Mar 05 21:09:22 dv_: that's a shell parameter meaning all parameters (i.e. it's equivalent to "$1 $2 $3 $4..." Mar 05 21:09:25 ) Mar 05 21:09:54 standard shell stuff, not specific to OE Mar 05 21:09:56 why would one pass this on here? for parallel-make (-jX) for example? Mar 05 21:10:11 I thought ${PARALLEL_MAKE} exists for that Mar 05 21:11:21 I specifically ask because just using "oe_runmake" does not build in parallel Mar 05 21:13:54 dv_: at least with OE-Core (probably OE-Classic as well), PARALLEL_MAKE gets into the make command line via EXTRA_OEMAKE Mar 05 21:14:12 and only for the compile task Mar 05 21:14:23 and oe_runmake alone does not use EXTRA_OEMAKE? Mar 05 21:14:30 it does yes Mar 05 21:14:46 see conf/bitbake.conf and classes/base.bbclass for details Mar 05 21:17:47 dv_: i assume you have it set in your local.conf? Mar 05 21:18:07 oh, it is actually set ... yet the package isnt building in parallel .. wtf Mar 05 21:18:15 either way, its not an OE issue it seems Mar 05 21:18:17 PARALLEL_MAKE = "jN" Mar 05 21:18:29 er, -jN Mar 05 21:18:44 make -j 2 and make -j2 should not make any difference, should it? Mar 05 21:20:32 don't think so Mar 05 21:21:12 you can see what's actually being run by looking at the run.do_compile script if you haven't already looked in there Mar 05 21:24:05 yeah, its using -j 2 . okay, something is wrong with the makefiles themselves. Mar 05 21:42:21 bluelightning: thanks, i hadn't found that one Mar 05 21:42:36 it least 3.13 is newer than 3.12... Mar 05 21:43:18 but this just might work, especially the part about the skip_signing.patch Mar 05 22:07:23 i owe you a beer for that one... Mar 05 22:07:40 in fact i'll give you kergoth's beer Mar 05 22:12:38 heh, i managed to crash gnome-shell and i didn't even notice... Mar 05 22:12:49 hah Mar 05 22:13:10 * kergoth doesn't touch linux desktops anymore, got tired of the frustration :\ Mar 05 22:14:10 do you prefer fruit now? Mar 05 22:18:10 kergoth uh Mar 05 22:27:26 I find KDE to be pretty good these days myself Mar 05 22:28:09 i spent a bit of time getting 3D/gnome3/etc working on ppc, so i'm going to use, by damn... Mar 05 22:30:52 terminal looks almost the same in all DEs I know Mar 05 22:58:17 kde is pretty fine these days, yes Mar 05 23:10:30 JaMa: kexec/arch/i386/x86-linux-setup.c:189: undefined reference to `vfscanf' Mar 05 23:10:35 and x86-linux-setup.c:283: undefined reference to `basename' Mar 05 23:10:49 klibc lacks those Mar 05 23:11:35 ..and both are only used in x86-linux-setup.c Mar 05 23:12:21 JaMa: I'm tempted to restrict the recipe to arm Mar 05 23:16:22 ant_home: ok with me Mar 05 23:16:50 JaMa: ubi-utils-klibc compiled just fine ... qemux86 Mar 06 00:15:05 we need an OE shirt for the next Linaro Connect Mar 06 00:17:56 are the OE lists ok, or has everyone gone quiet? Mar 06 00:36:00 Crofton|work: I've sent poppler upgrade patch few hours ago, do you see it? Mar 06 00:37:01 it's in pipermail, so looks OK http://lists.linuxtogo.org/pipermail/openembedded-devel/2013-March/044230.html Mar 06 00:44:01 yeah< ithink I saw it Mar 06 00:44:35 I'm just over sensitive to lulls in list traffic atm Mar 06 01:23:45 hello Mar 06 01:23:55 hi hrw Mar 06 01:24:23 said suihkulokki while looking into my direction ;D Mar 06 01:27:19 did someone had situation when SRC_URI = "git://" and build tried to fetch from mirrors and then instead of fetching from git it tried to unpack not existing file. ".done" file was for some reason created... **** ENDING LOGGING AT Wed Mar 06 02:59:58 2013