**** BEGIN LOGGING AT Tue Mar 12 02:59:58 2013 Mar 12 06:06:44 tasslehoff: pong Mar 12 06:09:26 Hi JaMa. Could you tell me a bit about the state of the meta-qt5 layer? I want to try to make a rootfs with qt5 support and perhaps an SDK to go with it, so that the Qt-guys at work can check what they can do with Qt5. Mar 12 06:10:20 Should compiling qtbase and qtjsbackend give me the packages needed for running qt5-based applications on target? Mar 12 06:11:15 it should but there are still some issues with it (I haven't fixed those last week, but maybe this one) Mar 12 06:19:05 ok. Mar 12 06:19:55 I'll report/submit/cry about my findings :) Mar 12 06:45:52 JaMa: bad start. cc1plus out of memory allocating 329136 bytes after a total of 900444160 bytes. Mar 12 06:46:08 when compiling qtbase Mar 12 06:46:22 ouch Mar 12 06:52:49 yeah. what does this signify? I'm guessing it won't go away with less threads/parallell makes. Mar 12 07:01:44 tasslehoff: is 900MB really too much for your builder to allocate? Mar 12 07:01:47 maybe a more minimal environment, free up as much ram as you can Mar 12 07:02:22 I have 8GB of ram, so it shouldn't be Mar 12 07:02:53 raise your hand if you sometimes have to compile on machines with less then 512 MB of ram... Mar 12 07:03:46 monitored the build with htop now, and I have 6.5 GB available when it fails with that message. Mar 12 07:04:19 * tasslehoff doesn't raise hand, but sympathises Mar 12 07:04:28 hm, is that with a z? Mar 12 07:06:06 yup Mar 12 07:13:37 nerdboy, does compiling native on the target count? Mar 12 07:27:43 JaMa: have you looked at https://github.com/martiert/meta-qt5? seems he has some sdk recipes that I thought maybe I'd try as a starting point Mar 12 07:34:38 the "out of memory" does not come because I run out of memory, so I wonder why Mar 12 07:35:33 strace maybe? Mar 12 07:35:43 never tried that on bitbake... Mar 12 07:42:39 tasslehoff: sounds like a 32-bit kernel with the wrong memory split? Mar 12 07:43:11 nerdboy: I can confirm that I have a 32-bit kernel. the last part went over my head Mar 12 07:43:39 32-bit kernels have a memory split option Mar 12 07:44:42 CONFIG_VMSPLIT Mar 12 07:48:19 i woul dmake sure you have the right high mem setting and split option that you want Mar 12 07:49:04 if you have the 1G/3G user/kernel split option set, that would explain it Mar 12 07:49:20 why you ran out so fast... Mar 12 07:49:51 is ARM9 v4T? Mar 12 07:50:38 nerdboy: ok. thanks for the input. Mar 12 07:51:38 just a wag, but you never know... Mar 12 07:52:15 I tried to revert the "qt5: upgrade to 5.0.1" since I managed to compile before, and now it seems much happier Mar 12 07:52:36 don't know why... Mar 12 08:02:09 nope, just took a bit longer. Mar 12 08:08:45 nerdboy: I have the 3G split set Mar 12 08:09:54 3/1 or 1/3 ? Mar 12 08:10:57 nerdboy: eh, how do I know? CONFIG_VMSPLIT_3G=y Mar 12 08:11:27 3/1 i think Mar 12 08:11:39 which high mem option? Mar 12 08:12:11 CONFIG_HIGHMEM64G=y Mar 12 08:12:34 should have access to all your memory then... Mar 12 08:13:06 i setup my build server as 32-bit and don't see memory issues Mar 12 08:13:24 need to update the raid firmware tho... Mar 12 08:14:09 full 16G seems to work fine Mar 12 08:14:38 could perhaps be something wrong in a layer somewhere then. Mar 12 08:15:00 I could try denzil instead of danny, I guess Mar 12 08:15:59 sleep... Mar 12 08:16:02 night Mar 12 08:16:26 night. thanks for sparring Mar 12 09:11:07 good morning all Mar 12 09:54:57 morning all Mar 12 09:56:47 good morning Mar 12 09:58:41 Morning Mar 12 10:22:17 hello Mar 12 10:22:25 did someone used bzr:// in SRC_URI? Mar 12 10:50:39 what is the best way to blacklist an RRECOMMENDS Mar 12 10:50:43 for example: Mar 12 10:50:48 recipes-core/udev/udev.inc:RRECOMMENDS_${PN} += "udev-extraconf usbutils-ids pciutils-ids" Mar 12 10:51:01 now, I don't want pciutils-ids or udev-extraconf Mar 12 10:51:43 can I just append RRECOMMENDS_udev = "usbutils-ids" Mar 12 10:52:04 but then I would possibly wipe out earlier (sane) RRECOMMENDS Mar 12 10:55:16 jackmitchell: setting BAD_RECOMMENDATIONS in the image recipe Mar 12 10:56:23 ensc|w: perfect, thanks Mar 12 11:18:43 I have a question about QA implemented in OE Mar 12 11:19:12 OE-core, there is some document that describe it or the work flow? Mar 12 11:21:39 hi ALl Mar 12 11:22:42 Can any body resolve this error Mar 12 11:22:46 ERROR: QA Issue: llvm-2.9 rdepends on external-simul-toolchain-dev ERROR: QA run found fatal errors. Please consider fixing them. ERROR: Function failed: do_package_qa Mar 12 11:22:56 i'm using external prebuilt tolchain Mar 12 11:22:59 toolchain* Mar 12 11:23:22 everything is working fine but packagin is giving this error... Mar 12 11:27:18 if I have an .bbappend and it has a do_install() will that override the do_install in the origianl .bb or will it append it? Mar 12 11:31:46 jackmitchell, I m not sure, but i think that not ovveride it Mar 12 11:32:25 jackmitchell, Could u use RREPLACE or a new layer? Mar 12 11:36:00 silvio_: RREPLACE looks good Mar 12 11:37:58 if you have any function in original .bb file then it will be given priority and is executed.. the fucntion which is in inherited class won't be executed @jack Mar 12 11:38:44 Guest12557: ah ok, RREPLACES it is ten Mar 12 11:38:46 then* Mar 12 11:39:28 any body knows about adding external toochain??? Mar 12 11:39:32 toolchain* Mar 12 12:04:01 Guest12557, thanks for info, sorry I dont know how help u Mar 12 12:17:03 how can i set a default password for an new created image ? Mar 12 13:06:04 Hello Mar 12 13:06:34 It seems we have some problem loading firmware with kernels < 3.7 in current udev. Do someone knows a solution for it? Mar 12 13:44:45 is it possible to generate a list of the packages (and version) included il the final image? Mar 12 13:45:38 without building image? Mar 12 13:47:17 hrw: after building would be ok as well, like happened with oe-classic Mar 12 13:47:39 do you know about bitbake -g image-name ? Mar 12 13:47:43 hrw: or without Mar 12 13:48:37 jackmitchell: yep, I thought -g was useful for graphviz only Mar 12 13:48:56 mckoan: it does a little bit of both, with a bit of scripting you could match the packages up to versions Mar 12 13:49:09 not ideal though Mar 12 13:49:22 maybe buildhistory? Mar 12 13:49:35 mckoan: cp var/lib/opkg/status from image Mar 12 13:52:10 hrw: that's true only if you have PACKAGE_CLASSES ?= "package_ipk" Mar 12 13:52:52 mckoan: var/lib/dpkg/status for deb image and similar for rpm? Mar 12 13:54:42 I would like tomething like *testlab with oe-classic Mar 12 13:54:51 s/tomething/something Mar 12 13:56:17 mario-goulart: meta/classes/buildhistory.bbclass:# Based in part on testlab.bbclass looks a derivative work Mar 12 13:56:42 s/looks/seems to be Mar 12 15:34:41 mario-goulart: INHERIT += "buildhistory" created something like what I needed, thx Mar 12 16:45:55 mckoan: you're welcome. Mar 12 18:30:27 JaMa: have you compiled qtjsbackend-native_5.0.1.bb? Mar 12 18:38:52 JaMa: seems so, since using jansa/test instead of master rid me of the configure error :) Mar 12 21:43:59 hello **** ENDING LOGGING AT Wed Mar 13 02:59:58 2013