**** BEGIN LOGGING AT Fri Feb 15 02:59:58 2013 Feb 15 08:21:20 good morning **** ENDING LOGGING AT Fri Feb 15 08:38:45 2013 **** BEGIN LOGGING AT Fri Feb 15 08:40:19 2013 Feb 15 09:48:13 is there a working qt5-recipe somewhere? I see there's a meta-qt5 layer out there Feb 15 10:03:33 tasslehoff: https://github.com/meta-qt5 Feb 15 10:09:40 mckoan: thanks. Feb 15 10:10:13 that's the one I chose. just thought I'd say it aloud in case someone would shout "qt5 no work yet!!" Feb 15 10:22:21 he tasslehoff where have you been? Feb 15 10:39:41 I've updated the postfix and the cyrus-sasl recipes from classic Feb 15 10:39:55 in which layer should that go ? Feb 15 10:55:53 ouch, cyrus-sasl :/ Feb 15 10:56:23 abelloni hm meta-openembedded Feb 15 11:08:25 woglinde: hey. I've mostly been here, but occupied with !OE-stuff and not so verbose :) Feb 15 11:25:01 woglinde: hi, may I ask a question about cross-compiling openjdk7? is there still an intermediate step necessary or can be icedtea7 cross-compiled in to steps? Feb 15 11:25:15 s/to/two/ Feb 15 11:32:14 abelloni: you also updated cyrus-sasl? Feb 15 11:32:27 abelloni: can you compare with version I sent to meta-oe?\ Feb 15 11:43:14 wbx hm no just follow the readme and you can compile it Feb 15 11:43:24 meza-java layer Feb 15 11:43:29 ups meta-java Feb 15 11:44:45 woglinde: i am trying to understand the magic behind the scenes ;) Feb 15 11:48:18 oh okay Feb 15 11:49:20 wbx icedtea7 can be compiled in one step with an exisiting jdk Feb 15 11:49:22 i am thinking about this way: gcj - icedtea7(bootstrap) - icedtea7(cross) . ready Feb 15 11:49:44 woglinde: ah, so a new feature Feb 15 11:49:49 hm gcj will not work I think Feb 15 11:49:58 Is it possible to add qt5 to my embedded system now, using the meta-qt5 layer? Feb 15 11:50:11 tasslehoff try it Feb 15 11:50:17 woglinde: I have Feb 15 11:50:21 therefore I ask :) Feb 15 11:50:21 and? Feb 15 11:50:49 I saw last commits from jama in january Feb 15 11:50:59 so it might be not in shape for production Feb 15 11:51:05 woglinde: it works fine. I am at the last step, having some issues with uClibc and MB_CUR_MAX Feb 15 11:51:17 compiled qt5-native and qtbase, but nothing that will give me ipk's I can install on target Feb 15 11:51:34 woglinde: perhaps. will examine the recipes a bit more Feb 15 12:07:15 wbx look at my uclibc patches Feb 15 12:08:56 woglinde: where I can find them? https://github.com/woglinde/meta-java/tree/master/recipes-core/openjdk/openjdk-7-03b21 ? Feb 15 12:09:54 yes Feb 15 12:10:07 look in the .bb file which I apply for uclibc Feb 15 12:12:59 i cannot see anything related to MB_CUR_MAX in the openjdk patches dir Feb 15 12:14:09 hm I did not catch this Feb 15 12:14:15 whcih version do you compile? Feb 15 12:14:27 icedtea 2.3.4 Feb 15 12:14:56 hm maybee you did not compile uclibc with locale support? Feb 15 12:15:29 that might be the problem Feb 15 12:15:37 and I only bulding 2.1 so far Feb 15 12:19:32 | ./aarch64-oe-linux-libtool: line 978: /home/hrw/HDD/devel/canonical/aarch64/openembedded/build/tmp-eglibc/sysroots/x86_64-linux//bin/sed: No such file or directory Feb 15 12:19:37 lovely ;d Feb 15 12:19:45 build completed with a HACK. juhu Feb 15 12:20:00 woglinde: what you do about $ORIGIN and uClibc? Feb 15 12:21:53 uClibc got patches so origin is not a problem anymore Feb 15 12:22:19 hrw checkout the sed discussion Feb 15 12:22:23 woglinde: are they in upstream or just in OE world? Feb 15 12:22:27 woglinde: thanks Feb 15 12:22:40 wbx no upstream Feb 15 12:22:48 since 0.9.33 i THINK Feb 15 12:25:02 oh, I missed this... will see if I run the java binary if 0.9.33.2 is fine Feb 15 12:27:42 java and uclibc? Feb 15 12:28:31 HRW YES Feb 15 12:28:35 i have it running Feb 15 12:28:41 since 4 years Feb 15 12:29:22 I do not say that it impossible - just a bit strange Feb 15 12:29:27 woglinde: i think $ORIGIN support is a patch in OE Feb 15 12:29:33 wbx you maybee need to link the zero libjvm.so from arm/lib/server to arm/lib Feb 15 12:29:33 woglinde: btw - nice weather in Berlin today Feb 15 12:30:00 hrw hu? Feb 15 12:30:05 its grey Feb 15 12:30:13 the whole day Feb 15 12:30:55 woglinde: sun not hit my eyes and it is not raining - nice weather ;d Feb 15 12:32:09 it's cracking weather in my part of the UK today, first sun I've seen in ages! Feb 15 12:32:10 hrw waiting for flight? Feb 15 12:32:38 http://cgit.openembedded.org/openembedded-core/tree/meta/recipes-core/uclibc/uclibc-0.9.33/orign_path.patch Feb 15 12:32:51 wbx ah hm Feb 15 12:32:58 I thought khem pushed it upstream Feb 15 12:33:02 than its in master Feb 15 12:35:02 woglinde: no, visiting friend Feb 15 12:35:23 oh ,) Feb 15 12:55:41 morning all Feb 15 13:01:08 hi pb Feb 15 13:01:38 hi woglinde Feb 15 13:35:01 date Feb 15 13:35:06 oops Feb 15 13:35:44 piÄ…, 15 lut 2013, 14:34:59 CET Feb 15 13:50:49 Friday! Feb 15 17:34:56 argh... busybox-syslog depends now on systemd :( Feb 15 17:35:10 merging -systemd packages into main packages was a horrible idea :( Feb 15 17:47:57 moin Feb 15 17:50:32 ensc|w: its not as horrible, if you think Feb 15 17:50:46 initscripts were always part of main paxkage Feb 15 17:53:58 khem: its is more horrible I guess :( Feb 15 17:54:25 khem: I could live with the additional sysvinit dep because it was very minimal and easy to remove at image creation time Feb 15 17:55:00 but systemd brings in a bunch of extra package which are impossible to handle in IMAGE_PREPROCESS_COMMAND or so Feb 15 17:55:31 e.g. formerly I created two kinds of images: the main, full featured one and a very minimal rescue image with busybox and some filesystem utils Feb 15 17:55:55 now, due the recent packaging the rescue image requires all the systemd stuff Feb 15 17:57:11 dedicated -systemd and (ideally) -sysv would be a perfect way to create custom images Feb 15 18:09:41 ensc|w: hmm if you see ways to use PACKAGECONFIG and reduce deps for systemd that would be nice improvement Feb 15 18:10:17 ensc|w: I think controlling it with DISTRO_FEATURE is right way to go IMO Feb 15 18:10:31 systemd may seem like IMAGE_FEATURE but it really is not Feb 15 18:10:50 it defines lot of distro policy bits Feb 15 18:10:54 khem: that has nothing to do with package config; e.g. a hard util-linux -> systemd dependency exists and might be wanted for full featured systems. Feb 15 18:11:33 I really do not want to define an extra distribution for a rescue system with 3-4 relevant packages Feb 15 18:11:58 ensc|w: so you can use may be busybox and mdev ? Feb 15 18:12:13 khem: busybox brings in systemd too Feb 15 18:13:26 ensc|w: hmm Feb 15 18:13:34 is it because of syslog ? Feb 15 18:13:38 yes Feb 15 18:26:38 Is opkg still maintained? Feb 15 18:26:56 With the dropbear fix of today.. I noticed that a SIGHUP doesn't appear to be handled by opkg Feb 15 18:27:00 (e.g. no state is written out) Feb 15 18:27:55 mailng lists OK? I do not have any yocto emails Feb 15 19:10:06 is there a good way to detect at run-time if a UBI partition exists? Feb 15 19:57:26 is there an image feature that install perf? Feb 15 20:07:55 If I set tools-profiile in an image recipe, how do I know what packages get installed? Feb 15 20:28:11 found it Feb 15 20:28:20 recipes-core/packagegroups Feb 15 20:51:08 is ntp in oe-core, I'mm having troubel finding it Feb 15 20:52:55 hmm, seems to come from busybox Feb 15 22:04:44 Is there a way to tell bitbake to re-run a recipe from a certain point and then keep going? i.e., a shorter version of "bitbake recipe -c compile -f ; bitbake recipe" Feb 15 22:05:38 I'm trying to build a kernel including an initramfs. I keep needing to re-build the kernel starting from the configure step. Feb 15 22:05:53 while I adjust the initramfs image Feb 15 22:09:42 IIRC yes Feb 15 22:10:05 but also look at some blog posts describing 'angstrom kernel workflow'. that might help Feb 15 22:11:36 thanks Feb 15 22:11:58 I just remembered reading that there are stamp files. Feb 15 22:12:19 perhaps if I touch the file before the configure step making them out-of-date. Feb 15 23:56:21 Cannot create link over existing -/tmp-. Feb 15 23:56:24 hmmm **** ENDING LOGGING AT Sat Feb 16 01:22:35 2013 **** BEGIN LOGGING AT Sat Feb 16 03:02:18 2013 Feb 16 06:04:30 hello i m developer for windows and mac. I m making cross platform tool.I can get foreground window on the system on win32 by utilizing GetForegroundWindow , how can i get same ? i want foreground window title name? possible ? Feb 16 06:17:46 JDuke128, this is mostly for embedded systems. Feb 16 10:37:00 hi, quick question: in general, as how severe (if existant at all) would you classify the problem of build dependency cycles when cross compiling a big set of binary packages for a certain target using openembedded? Feb 16 15:29:39 morning Feb 16 16:31:22 evein' guvnor Feb 16 18:50:32 is it possible to get -dbg packages with /usr/src/debug? Feb 16 18:50:41 I really don't need the source of eglibc on the device Feb 16 18:54:26 i was thinking there was a variable to disable it, but i guess there isn't, from a quick review of package.bbclass.. Feb 16 22:08:17 kergoth: thanks, yeah I thought switching to 'OE'-style would remove the /usr/src/debug.. but they don't. :} Feb 16 22:08:50 kergoth: but perf report is soo slow that I will probably just build a dedicated -dbg rootfs as tar.bz2 and use perf there. Feb 16 22:12:01 he zecke Feb 16 22:13:08 woglinde: hi Feb 16 22:13:35 woglinde: and I tried building perf.. but it will install everything on perl.. and I can't make my ubifs big enough to make it fit. :) Feb 16 22:13:48 and journald is taking 80% cpu time. :} Feb 16 22:14:51 switch back to sysvinit Feb 16 22:15:14 woglinde: then I need to start to learn daemontools or such Feb 16 22:15:45 or runit Feb 16 22:16:05 hm can you not disable journald? Feb 16 22:16:24 yes, I can.. but I want to see if it is broken beyond repair first **** ENDING LOGGING AT Sun Feb 17 02:59:57 2013