**** BEGIN LOGGING AT Wed Jan 11 02:59:57 2012 **** BEGIN LOGGING AT Wed Jan 11 05:01:37 2012 Jan 11 06:09:31 GNUtoo: pong Jan 11 06:09:36 and moin :) Jan 11 07:38:33 moin JaMa Jan 11 07:39:03 JaMa: now edje-native is failing for me Jan 11 07:41:56 | src/examples/Makefile.am:67: `pkglibdir' is not a legitimate directory for `PROGRAMS' Jan 11 07:43:23 moin Jan 11 07:44:25 weird, it's probably cause by new automake, but I've built edje-native yesterday without problem and automake wasn't upgraded in shr-chroot yet Jan 11 07:44:34 I'll probably upgrade it today and try Jan 11 07:46:15 btw: there is gcc and eglibc upgrade in queue Jan 11 07:46:16 hmm, ok Jan 11 07:46:22 fun :P Jan 11 07:46:42 with some linaro fixing... I sa Jan 11 07:46:43 w Jan 11 07:47:32 SHR: 03Martin.Jansa 07shr-chroot * r5bc2a9cae6e7 10/ (357 files in 36 dirs): system upgrade Jan 11 07:50:24 SHR: 03Martin.Jansa 07shr-chroot * re7af99c5bce1 10/ (136 files in 15 dirs): automake upgrade Jan 11 07:52:26 * mrmoku retries Jan 11 07:56:43 JaMa: still same failure Jan 11 08:10:32 mrmoku: jenkins has the same but in eina and eina-native, weird Jan 11 08:10:52 heh Jan 11 08:13:47 mrmoku: it's fixed in r66599 Jan 11 08:13:55 that's why I haven't seen it Jan 11 08:14:50 and now the difficult part :), do we want this commit split to multiple patches applied to almost every efl recipe, or do we want EFL bump? Jan 11 08:15:08 hehe Jan 11 08:15:13 well Jan 11 08:15:34 build failure is IMHO a legitime reason for an EFL bump now Jan 11 08:15:47 i can split it and add to meta-efl, but cannot test it as I'm on newer EFL already Jan 11 08:16:12 test it locally I mean, so I would be able to test it on buildhost/jenkins Jan 11 08:19:10 I'll push toolchain upgrade first to keep you and buildhost busy Jan 11 08:19:50 heh, ok :P Jan 11 08:20:12 SHR: 03Martin.Jansa 07meta-smartphone * rd4880c5c5788 10/meta-shr/conf/distro/include/preferred-shr-versions.inc: meta-shr: prefer eglibc-2.15, 2.14 is gone Jan 11 08:20:25 and then efl bump if jenkins build from scratch finishes fine with it Jan 11 08:22:18 JaMa: ok, I will do the upgrade-build test then Jan 11 09:10:39 JaMa: ERROR: '/OE/shr-core/openembedded-core/meta/recipes-core/eglibc/eglibc-initial_2.15.bb' failed Jan 11 09:10:42 ERROR: '/OE/shr-core/openembedded-core/meta/recipes-core/eglibc/eglibc_2.15.bb' failed Jan 11 09:10:52 will try a clean_toolchain_core now Jan 11 09:11:16 how it failed? Jan 11 09:11:27 ahh, moment Jan 11 09:11:30 incremental build worked here fine, gcc yesterday and eglibc today Jan 11 09:12:03 and buildhost is also already in eglibc-initial do_install Jan 11 09:12:19 heh Jan 11 09:12:20 | ERROR: Function 'Fetcher failure for URL: 'svn://www.eglibc.org/svn/branches/;module=eglibc-2_15;proto=http'. Unable to fetch URL svn://www.eglibc.org/svn/branches/;module=eglibc-2_15;proto=http from a Jan 11 09:12:24 | svn: error while loading shared libraries: libcap.so.2: cannot open shared object file: No such file or directory Jan 11 09:12:33 that libcap thing I hadd before already Jan 11 09:12:54 so unrelated to the toolchain update Jan 11 09:13:31 sysroots/x86_64-linux/lib/libcap.so.2 here Jan 11 09:13:59 maybe it was moved with those base_libdir changes and svn needs rebuilding Jan 11 09:14:04 hmm... have it there too Jan 11 09:14:10 rebuilding svn then Jan 11 09:15:46 'll push http://patchwork.openembedded.org/patch/18967/ today as it's fixing rebuild from scratch (stuff like readline-native) Jan 11 09:15:56 but probably not related to your issue Jan 11 09:17:03 ok Jan 11 09:18:27 hmm... it's not svn-native? just svn? Jan 11 09:19:03 subversion-native even Jan 11 09:19:09 :P Jan 11 09:19:13 ok Jan 11 09:20:15 seems like chroot's svn is not even linked to libcap Jan 11 09:21:35 ah neither is the one in sysroot, so probably something indirect Jan 11 09:41:22 JaMa: hmm... now I'm somehow doomed I guess Jan 11 09:41:23 | /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.2/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -lcap Jan 11 10:00:17 mrmoku: while building subversion right? Jan 11 10:00:39 isn't libcap somehow broken on fs? Jan 11 10:04:11 hmm Jan 11 10:04:18 yeah, while building svn Jan 11 10:04:42 OE om-gta04@shr ~/shr-core $ l tmp-eglibc/sysroots/x86_64-linux/lib/libcap* Jan 11 10:04:45 -rw-r--r-- 1 bitbake bitbake 86336 Jan 8 13:33 tmp-eglibc/sysroots/x86_64-linux/lib/libcap.a Jan 11 10:04:48 lrwxrwxrwx 1 bitbake bitbake 11 Jan 8 13:33 tmp-eglibc/sysroots/x86_64-linux/lib/libcap.so -> libcap.so.2 Jan 11 10:04:51 lrwxrwxrwx 1 bitbake bitbake 14 Jan 8 13:33 tmp-eglibc/sysroots/x86_64-linux/lib/libcap.so.2 -> libcap.so.2.22 Jan 11 10:04:54 -rw-r--r-- 1 bitbake bitbake 56157 Jan 8 13:33 tmp-eglibc/sysroots/x86_64-linux/lib/libcap.so.2.22 Jan 11 10:05:18 same here, including filesize Jan 11 10:08:02 export LDFLAGS += " -L${STAGING_LIBDIR} " Jan 11 10:08:14 from subversion recipe Jan 11 10:15:36 * JaMa hates those libdir moves.. :/ Jan 11 10:16:01 * mrmoku too Jan 11 10:16:55 moving other way then fedora/systemd is Jan 11 10:18:17 what if you add -L${STAGING_DIR_HOST}${base_libdir} there? Jan 11 10:18:52 JaMa: NOTE: package subversion-native-1.7.2-r0: task do_compile: Started Jan 11 10:18:55 with exactly that Jan 11 10:22:19 | /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.2/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -lcap Jan 11 10:23:48 * JaMa wondering what exactly are those yocto buildhosts testing.. Jan 11 10:34:35 WARNING: QA Issue: /lib/libgudev-1.0.so.0.1.0 links to something under exec_prefix .. I hope they won't move whole libffi + glib + stuff to /lib too Jan 11 10:37:27 :P Jan 11 10:45:23 NOTE: Running task 3011 of 7618 Jan 11 10:45:33 build from scratch going interestingly good Jan 11 11:02:05 JaMa: will trigger a build from scratch on my desktop too then Jan 11 11:03:05 mrmoku: wait for efl bump Jan 11 11:03:15 ok Jan 11 11:03:46 do you want me to push it now? or should I wait for test build to finish? Jan 11 11:05:00 or you can build from meta-oe-contrib/jansa/test + meta-smartphone/jansa/efl Jan 11 11:05:32 will build from the test branches then Jan 11 11:05:48 I would like to start my build now as I have to go to meetings for the rest of the day Jan 11 11:05:58 so hopefully I have something when I come back this evening :-) Jan 11 11:05:59 I see Jan 11 11:06:23 and this evening I'll push it to shr branches I guess Jan 11 11:06:51 ok Jan 11 11:13:15 bbl Jan 11 12:20:50 SHR: 03Martin.Jansa 07libphone-ui-shr * r5b77d7d7e0e5 10/src/view/ (4 files): adapt to elm_hover API changes from r66989 Jan 11 12:20:55 SHR: 03Martin.Jansa 07libphone-ui-shr * r6859416ba792 10/src/view/ (phone-log-view.c quick-settings-view.c): adapt to elm_toolbar API changes from r66927 Jan 11 12:20:55 SHR: 03Martin.Jansa 07libphone-ui-shr * r5ce1bb482971 10/src/view/quick-settings-view.c: adapt to elm_hoversel API changes from r66343 Jan 11 12:21:05 SHR: 03Martin.Jansa 07libphone-ui-shr * rf149d86442ea 10/src/util/ (ui-utils-contacts.c ui-utils.c): adapt to elm_list API changes from r66796 Jan 11 12:26:09 mrmoku: bumped EFL again so we don't need to ship that enjoy patch from me Jan 11 13:13:02 good morning guys Jan 11 13:29:13 SHR: 03Martin.Jansa 07meta-smartphone * r373ed8efdce7 10/meta-shr/recipes-shr/shr/libphone-ui-shr_git.bb: libphone-ui-shr: bump SRCREV to adapt to newer EFL api Jan 11 15:51:56 JaMa: ok, don't have to build from test branches anymore, right? Jan 11 15:52:06 I ran out of time before and so did not yet start building Jan 11 15:53:17 GNUtoo: ping Jan 11 15:54:46 mrmoku: yes, it's all pushed now to shr branches Jan 11 15:55:30 good :) Jan 11 15:56:03 build started Jan 11 15:56:08 * mrmoku off to the next meeting :/ Jan 11 16:05:55 anyone knowing GPRS very well? Jan 11 16:07:30 my question is: is it possible to connect to multiple APNs at the same time? Jan 11 16:07:36 like one for data and one for mms Jan 11 16:11:09 paulk-desktop: should be Jan 11 16:11:14 nice Jan 11 16:11:54 (I'm working on Nexus S modem and it helps me figuring out a data structure) Jan 11 16:14:49 paulk-desktop: Yes, but it depends on if the modem supports it and if the network does as well. You can also connect the same APN multiple times. Jan 11 16:15:02 ok Jan 11 16:15:07 apparently here it's possible Jan 11 16:21:13 Alex[sp3dev], pong Jan 11 16:45:09 GNUtoo: so, are you using s3c-fb on s5pc100? Jan 11 16:55:16 yes Jan 11 16:55:31 but it's modified Jan 11 16:55:44 and it's s5pc110 Jan 11 16:56:06 it's from the samsung kernel Jan 11 17:08:23 GNUtoo: i tried running vanilla and linaro on my exynos (s5pv310). The kernels with android fb seem to boot and switch to fb, while vanilla hangs. I have no clue how to debug since UART seems to be disabled or not routed (at least, it is said so on xda and I have not yet checked, gonna buy a connector and try tomorrow) Jan 11 17:37:17 Alex[sp3dev], ok Jan 11 17:37:42 Alex[sp3dev], I've UART on my nexus S Jan 11 17:37:58 Alex[sp3dev], do you want me to try something? Jan 11 17:38:35 GNUtoo: no, I think not. my cpu is different, so there may be something different Jan 11 17:39:04 worst thing is that samsung used a single kernel for recovery/boot so i cannot simply reboot and dump ramconsole Jan 11 17:39:27 my arch is s5pv210 Jan 11 17:39:59 ok there is the same on the galaxy tab Jan 11 17:40:06 (single kernel) Jan 11 17:40:18 kexec works? Jan 11 17:40:20 is there uboot for s5p with fb support? Jan 11 17:40:27 possible Jan 11 17:40:36 not sure if it has fb support Jan 11 17:40:42 but there is indeed uboot for it Jan 11 17:40:47 kexec doesn't work (it is generally broken on armv7 smp and i spent some day trying to apply all fixes, but didn't get it working) Jan 11 17:40:47 and it's mergerd in master Jan 11 17:40:54 *upstream Jan 11 17:41:01 ok Jan 11 17:41:17 what's the device btw? Jan 11 17:41:25 you see, there's no way to debug except fb since uart is not there.. do you think simply memcpy'ing to fb address will work? Jan 11 17:41:35 but fb was merged upstream right? Jan 11 17:41:39 it's galaxy s2, of course Jan 11 17:41:42 ok Jan 11 17:44:32 http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=1c248b7d2960faec3e1b8f3f9c5d9d0df28e0a3c Jan 11 17:45:44 Alex[sp3dev], ^^^ Jan 11 17:46:04 well, origen and other boards just use s3c-fb. if nothing works, i could just stick to the 3.1 kernel with android fb. for now, i just want to dump memory to figure out which atags are passed by default by sgs bootloader Jan 11 17:46:18 ok Jan 11 17:49:28 GNUtoo: which toolchain are you using? Jan 11 17:50:44 ah right Jan 11 17:50:55 I'm using the linaro one, gcc 4.6 Jan 11 17:51:04 link to binaries? Jan 11 17:51:15 it's from oe Jan 11 17:51:22 I build shr-core Jan 11 17:51:29 and then I export the path to the toolchain Jan 11 17:51:35 there are no official binaries for it Jan 11 17:51:42 but it unalign stuff Jan 11 17:51:51 so older kernels have the alignement problem Jan 11 17:51:54 could you tar it? my host is i386 debian, dunno if your binaries will work Jan 11 17:52:02 ah ouch Jan 11 17:52:09 I've x86_64 Jan 11 17:52:28 if you're looking for a toolchain I know where to find one tough Jan 11 17:52:41 but it may be for armv4 Jan 11 17:53:15 i've been building kernels with 2010q1 4.4 iirc. will that work for 3.2 or does it break? Jan 11 17:54:06 I guess it would work Jan 11 17:54:10 do you want me to try? Jan 11 17:55:52 well, if you have time, you could try. Honestly i think the problem is elsewhere. If i launch zImage, the screen is filled with noise. if i launch Image, it just hangs. if i launch zImage of 3.1 for origen, it fills with noise and then switches and fills with some other color, so i guess fb registers. I don't feel like bisecting from 3.1 to 3.2 since there were a lot of renames and fb driver was replaced Jan 11 17:58:26 ok Jan 11 17:58:38 do you have a link to your toolchain? Jan 11 17:59:15 2010q1-188 from sourcery Jan 11 17:59:32 https://sourcery.mentor.com/sgpp/lite/arm/portal/release1294 Jan 11 18:18:34 larsc, hi Jan 11 18:18:49 larsc, here's the thread: http://mailman.alsa-project.org/pipermail/alsa-devel/2011-December/046853.html Jan 11 18:24:26 Alex[sp3dev], btw does your device have the address changed? Jan 11 18:24:37 GNUtoo: which address? Jan 11 18:24:42 like that for instance: Jan 11 18:24:44 -#define PLAT_PHYS_OFFSET UL(0x20000000) Jan 11 18:24:44 +#define PLAT_PHYS_OFFSET UL(0x30000000) Jan 11 18:25:09 let me check. should not be Jan 11 18:25:16 mine does Jan 11 18:25:49 no, 0x40000000, as other exynos boards Jan 11 18:25:50 btw there is a 3.0 for crespo if I remember correctly the version Jan 11 18:25:51 ok Jan 11 18:26:34 i think i can make 3.1 android tree run, but i want vanilla because ultimately everyone will be using that, even android Jan 11 18:27:02 if nothing works, i could start with backporting vanilla fb to a working kernel, actually Jan 11 18:28:56 ok I understand and I know for android and vanilla Jan 11 18:32:25 but I was telling that in another perspective Jan 11 18:32:39 the perspective was to import the fb driver from 3.0 Jan 11 19:30:10 heyho Jan 11 19:30:59 hey morphis! Jan 11 19:31:34 angelox|laptop: I hope you will not kill me; I rewrote most aurora stuff again :) Jan 11 19:31:54 but I will integrate your stuff before pushing Jan 11 19:32:10 most important thing is we got rid of the symbian/qt components Jan 11 19:32:14 * angelox|laptop searches for his colt :P Jan 11 19:32:16 as they does not scale Jan 11 19:32:28 for various display resolution without much adjustments Jan 11 19:32:34 that is good Jan 11 19:32:43 but I will tell you more when it's ready Jan 11 19:32:58 currently I don't have much time to be available online due to day work Jan 11 19:33:20 ok Jan 11 19:34:05 do you know what can i do now? Jan 11 19:34:12 maybe documentation ;) Jan 11 19:36:06 yes but please wait until it's ready Jan 11 19:36:11 maybe I will push this weekend Jan 11 19:37:46 ok Jan 11 19:49:06 mickeyl, ping Jan 11 19:49:08 mrmoku, ping Jan 11 19:49:15 I think alsaloop won't work Jan 11 19:49:21 the usage is near 20% Jan 11 19:49:26 at max Jan 11 19:49:36 it varies between 10 and 20% Jan 11 19:49:47 only for modem->CODEC Jan 11 19:52:55 ping mrmoku mickeyl Jan 11 20:12:05 I took some advises in #alsa Jan 11 20:12:10 and I'll use the kernel module Jan 11 20:26:54 GNUtoo: pong Jan 11 20:28:23 mrmoku, I'm trying to lower the CPU usage of alsaloop Jan 11 20:28:26 GNUtoo: snd-aloop? Jan 11 20:28:27 in #alsa Jan 11 20:28:32 not a good idea Jan 11 20:28:35 it's not what we want Jan 11 20:28:38 it's a virtual card Jan 11 20:28:42 where you write in it Jan 11 20:28:46 and then read in it Jan 11 20:28:51 it doesn't forward Jan 11 20:28:58 ok Jan 11 20:29:55 GNUtoo: how's the cpu usage of your forwarder? Jan 11 20:31:08 between 10 and 20% ony for modem->CODEC Jan 11 20:31:28 I guess it's way too high Jan 11 20:32:05 hmm, so what is the solution? Jan 11 20:32:10 no idea Jan 11 20:32:16 what's perf? Jan 11 20:33:00 community/perf 3.2-1 Linux kernel performance auditing tool Jan 11 20:33:50 yes http://www.pixelbeat.org/programming/profiling/ Jan 11 20:33:51 hmmm Jan 11 20:33:56 s/yes// Jan 11 20:40:47 ok I now know how to use gprof again(I forgetten how to use it) Jan 11 20:41:11 I need to cross compile gprof now that I've the dump I guess Jan 11 20:55:23 GNUtoo: do you have any experiences with pandaboard (ES) ? Jan 11 20:59:01 no but what's the question? Jan 11 21:01:40 GNUtoo: i need to setup a productive streaming client in my living room. so the option is to use xbmc or a kde program. Jan 11 21:02:18 GNUtoo: which distribution should i take ? Jan 11 21:03:03 GNUtoo: and it may be necessary to compile some stuff because my infrared receiver is not supported yet in LIRC Jan 11 21:03:53 so ithought somethink oe core based would be good Jan 11 21:04:46 the board itself is not supported in shr so i would like to use somethink which actively supports the board Jan 11 21:06:44 yes like angstrom Jan 11 21:07:04 it's for music only? Jan 11 21:07:09 or also for videos? Jan 11 21:07:15 do you know mpd? Jan 11 21:10:06 GNUtoo: its also for video Jan 11 21:10:23 GNUtoo: i would like to stream kartina.tv Jan 11 21:10:52 there is a plugin for xbmc and a kde app vlc-record Jan 11 21:11:22 I don't know xbmc sorry Jan 11 21:11:29 in fact I only know mpd and xmms2 Jan 11 21:11:34 I never used media centers Jan 11 21:11:44 there is even one made in efl Jan 11 21:11:49 made by captainigloo Jan 11 21:11:53 enna I think Jan 11 21:13:18 GNUtoo: should i investigate angstrom ? Jan 11 21:13:58 yes they have an IRC channel Jan 11 21:14:07 GNUtoo: i think shr would also be good, but its missing the board support Jan 11 21:14:20 shr or angstrom is nearly the same thing Jan 11 21:14:24 for such device Jan 11 21:14:35 and angstrom would be a bit better.... Jan 11 21:14:46 else there is ubuntu or other distros Jan 11 21:14:51 look at the pandaboard wiki Jan 11 21:15:00 or in wiki that talk about the pandaboard Jan 11 21:18:27 GNUtoo: ok thank you Jan 11 21:22:31 GNUtoo: do you know why jama is not only ? Jan 11 21:22:37 online ? Jan 11 21:23:12 no idea Jan 11 21:48:10 GNUtoo: do you know the irc name of Koen Kooi? Jan 11 21:48:17 yes Jan 11 21:48:19 koen Jan 11 21:48:22 why? Jan 11 21:48:37 It's not a good idea bothering him, he's known to be rude with people Jan 11 21:49:12 GNUtoo: then he should invest more time in documentation :-) Jan 11 21:49:23 why? Jan 11 21:49:25 what's wrong? Jan 11 21:50:07 GNUtoo: how do i setup a build env for angstrom-core / pandaboard ? Jan 11 21:50:18 in shr we have makefile Jan 11 21:50:21 ah ok Jan 11 21:50:25 easy Jan 11 21:50:40 http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/ Jan 11 21:50:49 but there is also some images and feeds Jan 11 21:50:54 for instance narcisius Jan 11 21:51:05 http://narcissus.angstrom-distribution.org/ Jan 11 21:51:34 only that it's slow as hell here Jan 11 21:53:32 http://narcissus.angstrom-distribution.org/ server not fould Jan 11 21:53:35 found Jan 11 21:53:50 yes same for www.angstrom-distribution.org Jan 11 21:53:55 they have a problem currently Jan 11 21:54:21 no, for me it works Jan 11 21:54:39 may be a dns problem Jan 11 21:54:40 ok Jan 11 21:54:55 I hope it's not my DNS Jan 11 21:55:05 GNUtoo: or routing Jan 11 21:55:26 i sometimes have routing problems to openembedded Jan 11 22:00:03 I run my own DNS Jan 11 22:00:27 nschle85: btw, http://sourceforge.net/apps/mediawiki/mbm/index.php?title=MBM http://sourceforge.net/apps/mediawiki/mbm/index.php?title=Using_mbm-gpsd suggests there're some GSM/UMTS modems that have GPS as well, and they provide gpsd with pseudo-devices, just like i suggested to you it seems. Jan 11 22:01:13 PaulFertser: thank you Jan 11 22:15:24 nschle85, did you try that: http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/ Jan 11 22:15:25 ? Jan 11 22:16:42 no, but page opens Jan 11 22:16:49 now ill try Jan 11 22:17:04 it's to build angstrom-core Jan 11 22:17:18 please visit #angstrom channel, i found a brother with same problem :-) Jan 11 22:30:24 GNUtoo: what do you think about a "Desktop Phone" ? Jan 11 22:30:39 what would that mean? Jan 11 22:30:47 I must go to sleep btw Jan 11 22:31:28 FSO+USB internet stick, + SHR applications Jan 11 22:31:52 ok Jan 11 22:31:56 I must go bye Jan 11 22:32:01 bye Jan 11 22:46:32 mickel ping Jan 11 22:46:43 mickeyl: ping **** ENDING LOGGING AT Thu Jan 12 02:59:57 2012