**** BEGIN LOGGING AT Sun Nov 06 02:59:57 2011 Nov 06 04:46:12 lindi-: i'm sorry but i really had no time/energy to do anything this weekend, and now i'm leaving for a long ride, i'll try to take a look this evening. Nov 06 10:49:49 JaMa|Wrk: just update-env is not enough Nov 06 10:52:19 JaMa|Wrk: hmm... even update-env + rebuilding pseudo-native did not do the trick Nov 06 11:34:28 mrmoku: env-update && source /etc/profile ? Nov 06 11:34:43 but I had seen it too on Friday :/ Nov 06 11:35:18 mrmoku: btw if you're still using jama branch of meta-smartphone, be awere of TCAPPEND removal Nov 06 11:56:15 JaMa|Off: hmm... did not source profile... and yes still using that branch. TCAPPEND means? Nov 06 12:25:21 mrmoku: we're using tmp dir instead of default tmp-eglibc Nov 06 12:25:34 mrmoku: but sometimes it's causing weird issues Nov 06 12:26:00 mrmoku: so locally I'm trying to build without TCAPPEND = "" in shr.conf Nov 06 12:26:32 mrmoku: but if you don't want to rebuild from sstate, then you can put that in local.conf Nov 06 12:26:49 and remove it later when you want to rebuild from scratch Nov 06 12:53:02 ahh, ok Nov 06 12:53:04 thx Nov 06 17:19:18 heyho Nov 06 17:19:20 freesmartphone.org: 03morphis 07aurora * rd9eea983e0e6 10/ (100 files in 13 dirs): Nov 06 17:19:20 freesmartphone.org: Remove old aurora-applications and aurora-components; both are now part of aurora-daemon Nov 06 17:19:20 freesmartphone.org: Signed-off-by: Simon Busch Nov 06 17:19:21 freesmartphone.org: 03morphis 07aurora * ree3a4d23c8f8 10/aurora-daemon/ (71 files in 19 dirs): Nov 06 17:19:22 freesmartphone.org: aurora-daemon: move applications sources below src directory Nov 06 17:19:22 freesmartphone.org: Signed-off-by: Simon Busch Nov 06 18:04:44 morphis: heyho Nov 06 18:07:58 JaMa|Off: Package edbus-enotify version 2:1.0.999+svnr64506-r0.0 has no valid architecture, ignoring. Nov 06 18:09:14 JaMa|Off: and somehow we get both gpsd pulled into images Nov 06 18:10:26 I know about gpsd, I had even .bbappend to drop gpsd-utils (which pulls even udevd) Nov 06 18:10:51 ok Nov 06 18:11:41 | configure: error: Package requirements (openssl >= OPENSSL_REQUIRED) were not met: Nov 06 18:11:47 for libsamsung-ipc Nov 06 18:13:22 hmm... strange Nov 06 18:13:35 did you drop TCAPPEND already? Nov 06 18:13:54 I put TCAPPEND = "" in local.conf Nov 06 18:14:00 but it is building in tmp-eglibc Nov 06 18:14:13 let me check the variable Nov 06 18:14:31 ah :/ Nov 06 18:14:32 # we don't need to build multiple LIBC Nov 06 18:14:33 TCLIBCAPPEND = "" Nov 06 18:14:49 ok :) Nov 06 18:14:55 sorry Nov 06 18:14:57 but should that explain the openssl error? Nov 06 18:15:12 sort of.. Nov 06 18:15:18 NOTE: Running task 2428 of 2428 (ID: 11, /OE/shr-core/meta-openembedded/meta-oe/recipes-connectivity/openssl/openssl_1.0.0e.bb, do_rm_work_all) Nov 06 18:15:21 NOTE: package openssl-1.0.0e-r14.3: task do_rm_work_all: Started Nov 06 18:15:24 NOTE: package openssl-1.0.0e-r14.3: task do_rm_work_all: Succeeded Nov 06 18:15:26 NOTE: Tasks Summary: Attempted 2428 tasks of which 2427 didn't need to be rerun and 0 failed. Nov 06 18:15:54 I guess that even with sysroot trying to be relocateable there will be few issues when you rebuild it from sysroot Nov 06 18:16:17 so cleaning/rebuilding openssl might help? Nov 06 18:16:17 like ie python has special Makefile with hardcoded paths to tmp/sysroot etc Nov 06 18:16:23 yup, should help Nov 06 18:16:28 * mrmoku tries Nov 06 18:16:50 I want to finally have a qemu image to play with :) Nov 06 18:16:57 but on buildhost I wouldn't risk dropping TCLIBCAPPEND without rebuild from scratch (without sstate) Nov 06 18:17:30 * JaMa|Off had one before trying to debug multimachine builds and testing basichash signature algoritm :/ Nov 06 18:27:51 JaMa|Off: hmm... did not help Nov 06 18:27:59 maybe that's something completely different Nov 06 18:31:44 indeed Nov 06 18:31:47 morphis: ping Nov 06 18:32:12 morphis: could you please add the missing $ ;) Nov 06 18:32:14 PKG_CHECK_MODULES(OPENSSL, openssl >= OPENSSL_REQUIRED) Nov 06 18:33:30 paulk_: ping or you as you seem to have access :-) ^^^ Nov 06 18:33:57 yep Nov 06 18:34:00 ah :) Nov 06 18:34:37 hmm Nov 06 18:34:37 what's the problem exactly? Nov 06 18:34:52 I built it with autotools like this afternoon and it was ok Nov 06 18:35:05 configure: error: Package requirements (openssl >= OPENSSL_REQUIRED) were not met: Nov 06 18:35:26 I'm not very deep into autotools Nov 06 18:35:42 but looking at other stuff that should be $OPENSSL_REQUIRED ? Nov 06 18:35:57 JaMa|Off: does not work neither though Nov 06 18:36:01 configure: error: Package requirements (openssl >= 0.0.0) were not met: Nov 06 18:36:13 yes possible Nov 06 18:36:50 hmm Nov 06 18:36:55 it did work for me this afternoon Nov 06 18:37:07 mrmoku: and if you call pkg-config manually? Nov 06 18:37:27 mrmoku: isn't PKG_CONFIG_DIR (or whatever is variable name) set to old TMPDIR? Nov 06 18:38:36 JaMa|Off: how to call it manually with correct env? Nov 06 18:39:01 paulk_: maybe that depends on autofoo versions? Nov 06 18:39:21 I dunno I'm not an autofoo expert either :p Nov 06 18:39:46 JaMa|Off: looking at the run.do_configure script: Nov 06 18:39:47 export PKG_CONFIG_DIR="/OE/shr-core/tmp-eglibc/sysroots/qemux86-64//usr/lib/pkgconfig" Nov 06 18:41:29 gah... so close... and yet so far :-P Nov 06 18:41:45 paulk_: is that openssl new in there? Nov 06 18:42:08 I don't know… sorry Nov 06 18:42:11 ok Nov 06 18:42:25 basically I'm working on the android side of libsamsung-ipc Nov 06 18:42:34 mrmoku: battery OK? Nov 06 18:42:39 DocScrutinizer: yup :-) Nov 06 18:42:44 fine Nov 06 18:43:14 but a few days ago GNUtoo asked me to fix some autofoo-related bug and I also used autotools to build it directly on my laptop (when I didn't have a cross compiler) Nov 06 18:43:58 paulk_: as it does not work for me with the $ ... nvm for now Nov 06 18:44:04 guess I will do a rebuild from scratch Nov 06 18:44:44 JaMa|Off: hmm... maybe I have to come back to your qemu image :) Nov 06 18:45:22 JaMa|Off: which one should I take? there are so many... Nov 06 18:48:44 is there someone working on linux >= 3 for freerunner ? Nov 06 18:48:58 I am interested in doing so Nov 06 18:50:20 also which closest to 3 git tree is known to work ? I tried hackable1 images but they all segfault for me Nov 06 18:50:56 except their 2.6.39 without pers patches (but it lacks support for sound/bluetooth/wifi/etc) Nov 06 18:53:39 prahal: I'm very happy with 3.6.39 Nov 06 18:54:00 prahal: it's amazingly functional. no crashes since I upgraded two weeks ago or so Nov 06 18:55:47 mrmoku: ie latest qemuarm core-image-core Nov 06 18:58:15 lindi-: vanilla ? Nov 06 18:58:20 JaMa|Off: ok Nov 06 19:00:08 ie I would like to target 3 just in case a new udev or such requires the upgrade . Otherwise no hurry Nov 06 19:01:44 mind I am on debian so 2.6.39 is not yet available Nov 06 19:03:19 prahal: I'm preparing this for debian Nov 06 19:04:50 prahal: http://lists.alioth.debian.org/pipermail/pkg-fso-maint/2011-October/004543.html Nov 06 19:07:18 prahal: so it's "vanilla - some non-free stuff + patches to support openmoko" Nov 06 19:10:05 lindi-: I had to rebuild libca-ng0 and to upgrade tslib (git for multitouch support as 2.6.39 use it) and rebuild xserver-xorg-input-tslib . Nov 06 19:11:14 I already reported the issue to libcap-ng to the bts but not yet about tslib (as I wonder if a bug about a not yet in debian kernel is appropriate Nov 06 19:12:55 prahal: I use -evdev now Nov 06 19:13:09 prahal: since xserver-xorg-input-tslib does not seem to be loved very much... Nov 06 19:13:29 prahal: install.sh also uses -evdev nowadays Nov 06 19:13:58 prahal: can you migrate to -evdev or would you like to start working to make -tslib a viable option again? Nov 06 19:14:07 in fact I just found time to look back at my freerunner . A few years ago I submitted patches to neon viewer upstream (I believe upstream was dead and never got a reply ) and found out that gstreamer was as fast if not faster than mplayer if one removed the floating point plugins from gstreamer (and so did an e17 emotion version based on gstreamer that I never had time to clean up) for intone Nov 06 19:14:56 lindi-: tslib works very well with latest git of tslib + rebuild of the xorg extension Nov 06 19:15:21 is the debian maintainer mia for tslib ? Nov 06 19:16:24 also one has to add MatchDevicePath "/dev/input/event*" to the 50-tslib.conf because when applied to mouse0 the extension crash Nov 06 19:16:43 no need for any config in xorg.conf Nov 06 19:17:33 I could use evdev . Exxxxxcept it is working badly here . At least with hackable1 2.6.39 uImage Nov 06 19:18:04 I will give it a second shot with your build Nov 06 19:18:05 > It seems that whatever InputDevice is defined in xorg.conf for the Nov 06 19:18:05 > touchscreen, X.org defines another one from udev inputs if any, leading Nov 06 19:18:05 > to two X InputDevices competing to handle the touchscreen events. I Nov 06 19:18:05 > don't know X enough to characterize this as a bug or not. Nov 06 19:18:10 prahal: http://www.mail-archive.com/smartphones-userland@linuxtogo.org/msg02087.html Nov 06 19:18:19 prahal: that is at least one problem that caused people to remove -tslib Nov 06 19:18:57 prahal: also, #openmoko-debian Nov 06 19:19:42 mrmoku: did you finish downloading? Nov 06 19:19:48 it is only that the /usr/share/X11/org.conf.d/50-tslib.conf is broken ie it is lacking the MatchDevicePath to only match /dev/input/event* items Nov 06 19:19:50 I'm going to shutdown for some time Nov 06 19:19:51 JaMa: yup Nov 06 19:20:00 ok, bbl Nov 06 19:20:02 prahal: http://lists.linuxtogo.org/pipermail/smartphones-userland/2011-August/003037.html Nov 06 19:20:08 JaMa: still trying to figure out how to make it run Nov 06 19:20:13 but everything downloaded Nov 06 19:21:16 other than that tslib refuse to work on 2.6.39 due to the version in debian lacking the october patches from git for MT support and a too strong version check Nov 06 19:22:25 thanks with those calibration settings I guess evdev will behave better Nov 06 19:23:52 prahal: anyways, now install.sh uses -evdev Nov 06 19:24:16 prahal: so I'm using that to test what others use :) I can't test all combinations... Nov 06 19:31:47 also might be of interest . I have partial success i just letting the phone autoconfigure to network-manager via dhcp (using network-manager shared computer option) as of 0.9.1.95-1 in debian . ie the only drawback currently is that the freerunner is bound to 10.42.43.68 but one have to look in syslog to find out this ip (and as network-manager disable all host dns-masq configs - be it /etc/hosts or dns - it is not user manageable) Nov 06 19:32:39 this avoid one to require tweaks his freerunner if the defaults freerunner subnet conflict with a local subnet , if one want to plug mutliple freerunner at the same time , etc Nov 06 19:33:49 fair enough about evdev Nov 06 19:41:35 prahal: I just use static addresses now, I have three freerunners :) Nov 06 19:43:18 prahal: but any testing of linux-2.6 2.6.39-1-gta021 is greatly appreaciated. I'm going to upload it to pkg-fso repository unless major issues are found Nov 06 19:45:08 testing asap (installing currently) Nov 06 20:13:54 lindi-: there only is missing the linux-kbuild-2.6.39 (it is only in squeeze-backports) Nov 06 20:14:02 for the -headers-.deb Nov 06 20:14:25 prahal: you can get it from snapshot.debian.org using debsnap command Nov 06 20:16:18 agreed . Just the only glitch I found in the install process . Now I will boot into it Nov 06 21:35:46 SHR: 03Martin.Jansa 07shr-chroot * r9bf773ea349b 10/ (135 files in 19 dirs): system upgrade Nov 06 21:37:30 lindi-: the gta02 snd soc is missing and the debugfs is not hardcoded into the kernel as on debian kernels (ie one has to add the debufs line in /etc/fstab Nov 06 21:38:04 I will do a rebuild of the kernel and report if anything else arise . Tahank you for all the work Nov 06 21:38:10 prahal: snd soc is missing? Nov 06 21:38:20 Tahank/thank Nov 06 21:38:21 prahal: what do you mean? sound works here Nov 06 21:38:54 prahal: you just have to load the modules Nov 06 21:39:09 there is only the snd-soc-neo1973-wm8753 Nov 06 21:39:22 the -gta02- ojne is missing for freerunner Nov 06 21:39:55 could it be that one previous build left the file in your install ? Nov 06 21:40:06 prahal: they might have changed names Nov 06 21:40:27 $ cat /etc/modules|nc paste.dyndns.org 1234 Nov 06 21:40:28 http://paste.debian.net/143347/ Nov 06 21:41:25 snd-soc-neo1973-gta02-wm8753 is not available in the package still you have it loaded Nov 06 21:41:41 prahal: no, I have it listed Nov 06 21:41:47 prahal: since I want to be able to boot older kernels too Nov 06 21:41:58 prahal: it is snd_soc_neo1973_wm8753 nowadays Nov 06 21:42:27 hum could be that I had a shortage of memory at one point in the install and the next run did not applied cleanly . I am checking this now Nov 06 21:42:48 ah ok . Though your pastebin still list the -gta02- one Nov 06 21:43:04 hum sorry I am confused . Nov 06 21:43:15 I get it now . Nov 06 21:43:46 cat /proc/asound/cards . no soundcards (I load the module by hand . lsmod shows it but nothing in dmesg Nov 06 21:43:52 prahal: my pastebin shows whta I have in /etc/modules Nov 06 21:44:04 prahal: you need to load the modules in the right order Nov 06 21:44:55 got it working now . Nov 06 21:45:15 still I need the debugfs entry in fstab otherwise the module fails to load Nov 06 21:45:58 prahal: hmm? Nov 06 21:46:03 this time I loaded it by hand after the boot (I did a few attempts in /etc/modules beforehand without the -gta02- name and it failed . I will try again Nov 06 21:46:12 prahal: I'll test this Nov 06 21:50:08 prahal: sound works here even if debugfs is not mounted Nov 06 21:50:18 prahal: how does it fail for you? Nov 06 21:59:44 I get a line soc-audio soc-audio: ASoC: Failed to create codec debugfs directory Nov 06 22:00:13 prahal: odd Nov 06 22:00:28 prahal: can you put /etc/modules, dmesg and lsmod online somewhere? Nov 06 22:01:30 I bet I got any oldities into my init scripts . This is two years old . Nov 06 22:01:36 sure Nov 06 22:03:07 prahal: ok, I run unstable Nov 06 22:06:47 me too though the install process was not streamlined when I installed this debian in the sd . I will spend some time on it to avoid chasing a local issue Nov 06 22:08:02 http://prahal.homelinux.net/~prahal/freerunner/0/ (mind I loaded the module by hand so I am missing the right line for the sound module in /etc/modules Nov 06 22:12:16 prahal: I think that shows you loaded the modules in the wrong order Nov 06 22:12:57 prahal: http://prahal.homelinux.net/~prahal/freerunner/0/dmesg does not mention debugfs Nov 06 22:19:06 thnk you . I am trying with your /etc/modules Nov 06 22:19:58 as there are big snd soc changes Nov 06 22:20:38 dmesg is froma successful run (ie modprobing the snd soc by hand (the failed run is in messages Nov 06 22:20:55 prahal: the web server does not let me read messages Nov 06 22:21:09 prahal: You don't have permission to access /~prahal/freerunner/0/messages on this server. Nov 06 22:31:44 fixed sorry Nov 06 22:32:02 prahal: no problem, thanks for testing :) Nov 06 22:32:18 prahal: the kernel should definitely automatically detect the modules but afaik that is not so easy Nov 06 23:01:40 got the sound working :) had to remove /etc/ alsa conf , asound state , alsactl init . Oldies were left behind from time they were required Nov 06 23:09:48 prahal: grat Nov 06 23:09:51 prahal: great Nov 07 00:16:43 * [Rui] waves **** ENDING LOGGING AT Mon Nov 07 02:59:57 2011