**** BEGIN LOGGING AT Thu Aug 02 02:59:59 2012 Aug 02 07:18:43 SHR: 03Martin.Jansa 07shr-themes * r248c97f406ba 10/e-wm/e-wm-config-illume2-shr/e.src: e-wm-config-illume2-shr: downgrade config_version Aug 02 08:17:00 JaMa: hi, what package provides eet? Aug 02 08:17:22 libeet-tests Aug 02 08:18:20 tnx, missed that one in opkg search.. Aug 02 08:22:21 yup I should have write it in trac comment Aug 02 08:26:01 attached the tar.gz to the ticket Aug 02 08:26:25 fast, thanks Aug 02 08:26:54 yw Aug 02 08:28:47 jake42: sorry my fault.. that tar command was wrong.. I need .src files not .cfg, but I'll extract them here, so fine Aug 02 08:34:56 jake42: are you sure you have removed your .e folder before? Aug 02 08:38:14 jake42: just asking because there are 5 generations of e.cfg Aug 02 08:38:27 but other than that looks goo Aug 02 08:38:28 d Aug 02 08:51:14 JaMa: yes i removed the .e folder Aug 02 08:51:41 and i flashed the image just before Aug 02 08:52:56 SHR: 03Martin.Jansa 07shr-themes * r4916b7248027 10/e-wm/e-wm-config-illume2-shr/ (Makefile.am e.src e_remember_restart.src module.comp.src): Revert "e-wm-config-illume2-shr: downgrade E17 config for version which is in SHR feed" Aug 02 08:52:57 jake42: ok.. then it's fine ^ Aug 02 08:52:59 SHR: 03Martin.Jansa 07shr-themes * r1ba29f122f7b 10/e-wm/e-wm-config-illume2-shr/ (5 files): e-wm-config-illume2-shr: config update based on jake42 config Aug 02 08:53:02 thanks! Aug 02 08:58:24 thanks too :) Aug 02 08:59:49 SHR: 03Martin.Jansa 07meta-smartphone * rdc6c28ef365e 10/meta-shr/recipes-shr/tasks/task-shr-feed.bb: task-shr-feed: add anki Aug 02 08:59:59 SHR: 03Martin.Jansa 07meta-smartphone * r6e23d1000d86 10/meta-shr/recipes-shr/shr/e-wm-config-illume2-shr_git.bb: e-wm-config-illume2-shr: bump SRCREV Aug 02 11:36:26 SHR: 03shr-devel 07buildhistory * r22f866744d0e 10/packages/armv4t-oe-linux-gnueabi/ (42 files in 42 dirs): packages: Build 201208020926 of shr 20120802 for machine om-gta02 on opmbuild Aug 02 12:13:26 SHR: 03Martin.Jansa 07meta-smartphone * r4cdc62d87ce8 10/meta-nokia/recipes-bsp/u-boot/ (u-boot/0001-config-Always-use-GNU-ld.patch u-boot_git.bb): meta-nokia: u-boot: import fixed 0001-config-Always-use-GNU-ld.patch from meta-ti Aug 02 12:14:20 Are there any recent videos of elfe, or documentation? I tried it a while ago and it was not intuitive. Aug 02 12:16:55 jeepingben: http://www.youtube.com/watch?v=mKlDgl-Tquw not sure if it's right video (cannot check now) Aug 02 12:17:06 but there was video about it Aug 02 12:18:05 Thanks Jama, that is of 0.0.1, I know there are also videos of 0.0.2. I was wondering about more recent stuff. I guess that will happen after it is part of SHR and more people are running it. Aug 02 12:18:43 SHR: 03Martin.Jansa 07meta-smartphone * racc8975be920 10/meta-openmoko/recipes-bsp/u-boot/u-boot/0001-config-Always-use-GNU-ld.patch: meta-openmoko: u-boot: import fixed 0001-config-Always-use-GNU-ld.patch from meta-ti Aug 02 12:19:07 agreed, please add it to wiki (even stagingTest page) if you find something usefull about it Aug 02 12:21:08 maybe it was this one from Lukas http://slyon.de/shr/elfe.ogv Aug 02 12:21:34 but you probably know already about this one :) Aug 02 12:23:30 hmm, in the videos it looks fairly intuitive. I wonder if the lack of animation on GTA02 was part of the problem. Thanks, off to work... Aug 02 13:04:52 SHR: 03morphis 07meta-smartphone * r83e5969769c8 10/meta-samsung/conf/machine/tuna.conf: meta-samsung: machine: add machine configuration for the samsung tuna device Aug 02 13:04:57 SHR: 03morphis 07meta-smartphone * rf0b7613b4cc7 10/meta-samsung/recipes-kernel/linux/ (linux-samsung-tuna/defconfig linux-samsung-tuna_git.bb): meta-samsung: linux: add recipe for the linux kernel of the samsung tuna device Aug 02 13:23:50 SHR: 03shr-devel 07buildhistory * rbc67ebcf0a47 10/packages/armv4t-oe-linux-gnueabi/ (68 files in 68 dirs): packages: Build 201208021459 of shr 20120802 for machine om-gta02 on opmbuild Aug 02 13:42:26 SHR: 03shr-devel 07buildhistory * rcbfef1d24955 10/ (16 files in 14 dirs): packages: Build 201208021533 of shr 20120802 for machine om-gta02 on opmbuild Aug 02 14:26:28 morphis, hi Aug 02 14:26:30 for GTA02 should we try that: Aug 02 14:26:34 _GSM line discipline (linux/drivers/tty/n_gsm.c) Aug 02 14:26:40 * N_GSM Aug 02 14:26:46 it seem to be a kernel muxer Aug 02 14:26:58 that might be more efficent Aug 02 14:27:02 and save more CPU cycles Aug 02 14:39:20 GNUtoo: fsogsmd already supports it Aug 02 14:39:27 wow ok Aug 02 14:39:37 in what kernel version was it introduced? Aug 02 14:39:44 is it in use for gta02? Aug 02 14:39:57 don't know Aug 02 14:40:04 ok Aug 02 14:40:14 is it there in 2.6.39? Aug 02 14:40:46 I think so Aug 02 14:40:52 ok Aug 02 14:40:56 is it used in gta02? Aug 02 14:41:16 btw I have a way to get UBX with gpsd Aug 02 14:41:25 GNUtoo: see PM Aug 02 14:42:15 GNUtoo: yes I've enabled it in defconfig even before .39 Aug 02 14:42:20 ok Aug 02 14:42:26 so it's used.... Aug 02 14:42:41 does it have to be enabled in fsogsmd.conf? Aug 02 14:42:45 yes Aug 02 14:42:51 you have to change the pdp provide I think Aug 02 14:47:13 ubxgen.py 06 00 14 00 01 00 00 00 d0 08 08 00 80 25 00 00 01 00 01 00 00 00 00 00 > /dev/ttySAC1 Aug 02 14:47:16 for GPS Aug 02 14:47:20 running that makes it: Aug 02 14:47:25 * switch to UBX Aug 02 14:47:40 * since it send some ubx commands gpsd recognize it and switches to ubx Aug 02 14:47:59 s/send/receive some UBX result from the command/ Aug 02 14:47:59 GNUtoo meant: * since it receive some UBX result from the command some ubx commands gpsd recognize it and switches to ubx Aug 02 14:56:28 GNUtoo: just sending ubx commands blindly to the chip is not a good idea, you are supposed to wait for the ACK Aug 02 14:56:59 lindi-, yes 0x05 0x00 Aug 02 14:57:05 if I remember well Aug 02 14:57:17 and 0x05 0x01 is NACK Aug 02 14:57:18 GNUtoo: also opening and closing the serial port between commands is a really bad idea Aug 02 14:57:23 ah ok Aug 02 14:57:37 you want to have an event loop Aug 02 14:57:41 ok Aug 02 14:58:02 http://lindi.iki.fi/lindi/git/ubx.git/ has some stuff in python Aug 02 14:58:05 I guess I'll try to inject them trough gpsd Aug 02 14:58:06 ok Aug 02 14:58:10 with it you don't need to write raw hex Aug 02 14:58:21 and it has at least a rudimentary event loop using glib Aug 02 14:58:27 yes but I've a big problem: I'm often offline currently Aug 02 14:58:31 and I had your presentation Aug 02 14:58:39 but I count't clone the git Aug 02 14:58:45 GNUtoo: how did it fail? Aug 02 14:58:46 because I was offline Aug 02 14:59:02 ah Aug 02 14:59:07 it didn't fail, I had just no internet to do it Aug 02 14:59:12 I guess I should have embeded it to the PDF Aug 02 14:59:18 ok lol Aug 02 15:10:05 really, where is 2.6.39 kernel git for gta02? git.openmoko.org seem resolved into http://docs.openmoko.org/trac/... Aug 02 15:10:36 gena2x, on gitorious Aug 02 15:10:45 gena2x, project name is SHR Aug 02 15:11:10 we generate patches from it with a script Aug 02 15:11:26 GNUtoo, ok, thanks Aug 02 15:11:44 and then the patches are included in meta-openmoko in meta-smartphone Aug 02 15:12:10 lindi-, are you reporting bugs against this SHR kernel? Aug 02 15:12:36 yes but I think the bugs are reported in the openmoko trac Aug 02 15:12:46 s/yes/I guess so/ Aug 02 15:12:47 GNUtoo meant: I guess so but I think the bugs are reported in the openmoko trac Aug 02 15:18:34 morphis, should I try the basic or the advanced ngsm? Aug 02 15:18:51 puuh don't know Aug 02 15:18:56 I never played with ngsm Aug 02 15:19:03 mickeyl implemented this long ago Aug 02 15:22:13 yes Aug 02 15:22:14 in 2011 Aug 02 15:22:29 a77e3ee12bb53d83d48af9aa50f2377254f8720a Aug 02 15:22:45 hm, so basically shr has GTA02 patches rebased to latest 3.5 kernel? Aug 02 15:23:07 cool Aug 02 15:23:40 really, GTA is only obsolete phone with bleeding edge kernel support =) Aug 02 15:23:54 gena2x, yes and no Aug 02 15:24:06 gena2x, the kernel must be bisected Aug 02 15:24:09 to fix boot issues Aug 02 15:24:13 on 3.5 Aug 02 15:24:20 + it's not tested well etc... Aug 02 15:24:26 basically it doesn't work Aug 02 15:24:27 GNUtoo: he was testing it on 2.6.37.1 and something older (based on linux-next) http://git.openembedded.org/openembedded/commit/recipes/linux/linux-2.6.37/om-gta02/defconfig?id=17f3253e285b828adb4daa0b8b2eb9cda9e5bdfe I don't remember if there was basic/advanced back then Aug 02 15:24:30 because it must be debugged Aug 02 15:24:35 That's the latest version which boots? Aug 02 15:24:48 it boots Aug 02 15:24:51 but too slowly Aug 02 15:24:52 3.2 works good enough for me Aug 02 15:25:04 like you must wait like more than one hour Aug 02 15:25:08 for the boot to complete Aug 02 15:25:12 3.2 doesn't have that issue Aug 02 15:25:23 cool, i see. Aug 02 15:25:48 Sounds like pretty nice thing to fix Aug 02 15:52:16 SHR: 03shr-devel 07buildhistory * rc0ed84e492d2 10/packages/armv4t-oe-linux-gnueabi/ (62 files in 57 dirs): packages: Build 201208021718 of shr 20120802 for machine om-gta02 on opmbuild Aug 02 15:54:59 hmmm N_GSM doesn't work out of the box with the calypso plugin Aug 02 15:55:13 calypso plugin depend on the 07something mux plugin Aug 02 16:13:38 SHR: 03Martin.Jansa 07shr-makefile * r53182bab6bad 10/conf/shr-core/local.conf: shr-core: local.conf: enable buildhistory and set better BB_CONSOLELOG Aug 02 16:17:30 SHR: 03Martin.Jansa 07shr-makefile * rbc6a8a9ac781 10/conf/shr-core/local.conf: shr-core: local.conf: example how to enable SSTATE_MIRRORS, but it's not populated yet Aug 02 16:17:41 SHR: 03Martin.Jansa 07shr-makefile * r908d2a7c5d7b 10/conf/shr-core/local.conf: shr-core: local.conf: comment about how to disable buildhistory Aug 02 16:35:07 JaMa, hi Aug 02 16:35:08 hmmm Aug 02 16:35:14 gpsd has hook script Aug 02 16:35:19 that seem easy at first Aug 02 16:35:21 but.... Aug 02 16:35:22 but.... Aug 02 16:35:32 gpsd runs as "nobody" Aug 02 16:35:53 any ideas? Aug 02 16:35:59 what's the correct way to handle that? Aug 02 16:36:03 make gpsd run as gpsd Aug 02 16:36:20 and add a gpsd group Aug 02 16:36:27 and put the device nodes as gpsd? Aug 02 16:36:29 or.... Aug 02 16:36:34 chmod +s omhacks Aug 02 16:36:39 or something else? Aug 02 16:36:58 like run gpsd as root(maybe a not-so-good idea because we're not alone in the oe world) Aug 02 16:37:25 gpsd as user seems good Aug 02 16:37:41 ok Aug 02 16:37:53 I wonder where the nobody is set Aug 02 16:37:57 maybe in the source code? Aug 02 16:39:40 maybe it's default for systemd .service, but you can specify different user in [Service] section Aug 02 16:40:10 but that's not the case for xserver-nodm.service Aug 02 16:40:42 hmmm Aug 02 16:40:43 gpsd must start up as root in order to open the NTPD shared-memory segment, open its logfile, and create its local control socket. Before doing any processing of GPS data, it tries to drop root privileges by setting its UID to "nobody" (or another userid as set by configure) and its group ID to the group of the initial GPS passed on the command line — or, if that device doesn't exist, to the group of /dev/ttyS0. Aug 02 16:41:01 we have a serivice file for gpsd? Aug 02 16:42:47 JaMa, morphis do we handle the sysfs trough gpsd hooks or trough the quirks plugin trough gps = true Aug 02 16:42:54 or handle_gps = true rather Aug 02 16:43:19 *ignore_gps = false Aug 02 16:43:36 GNUtoo: yes it's in meta-systemd now Aug 02 16:44:27 ok Aug 02 16:44:32 imho would be better through gpsd hooks Aug 02 16:44:40 I may just not have it installed Aug 02 16:44:40 ok Aug 02 16:45:12 indeed you're right because it would make all gpsd program work without fsoraw wrapping Aug 02 16:45:46 GNUtoo: but there is also gpsd.socket waiting on gpsd port Aug 02 16:46:56 ok Aug 02 16:47:32 what's the implications exactly? Aug 02 16:50:57 that if you need to change that that it should be consistent between socket activation and manual start Aug 02 16:52:09 hmmm Aug 02 16:52:13 else I've another idea.... Aug 02 16:52:16 we could do that: Aug 02 16:52:25 chown dialout the sysfs nodes Aug 02 16:52:43 normally gpsd should become nobody:dialout Aug 02 16:53:50 like in systemd machine unit scripts Aug 02 17:06:44 JaMa, how to see the group of a program? Aug 02 17:06:51 at runtime Aug 02 17:07:40 group of user running that program? Aug 02 17:08:52 GNUtoo: cd /proc/`pidof foo`/status. or ps aux better Aug 02 17:09:27 I'll try ps aux again(I use that all the time but I didn't notice it had group) Aug 02 17:09:46 no it has no group here Aug 02 17:09:58 I'll look at status Aug 02 17:10:55 GNUtoo: ps -eo pid,group Aug 02 17:11:14 or, even better, "info ps" Aug 02 17:11:24 ok thanks Aug 02 17:11:38 and is it possible to have something else then group of uid running that? Aug 02 17:11:45 no problem. have fun. ps has an incredible amount of options Aug 02 17:11:50 ok Aug 02 17:12:08 JaMa: what do you want to get? Aug 02 17:12:53 I don't want anything.. just always expected that process has uid as owner and doesn't track "group" Aug 02 17:13:05 other than group in which that uid is listed Aug 02 17:14:11 I would assume you just need to add gpsd user to the group owning device nodes, no? Aug 02 20:39:16 gena2x: ? Aug 02 22:18:07 SHR: 03Martin.Jansa 07shr-chroot * rfe5f64aafd40 10/ (160 files in 20 dirs): system upgrade Aug 02 22:19:00 SHR: 03Martin.Jansa 07shr-chroot * r85ede118592f 10/ (298 files in 28 dirs): system upgrade Aug 02 23:45:12 hi it seems I had to trick the charging code (the battery was drained, I let it trickle charge, then it was again depleted before full boot ) . I unplugged , plugged back the usb charging cable while shr was booting . Could it be that the full charge code is only triggered by a plug action ? I had tried like 7 times before without this trick without success Aug 03 01:37:35 prahal2, I have never recovered from a drained battery with SHR, I boot qtmoko when it happens. Aug 03 01:45:08 it takes around 20min with usb cable plugged then it start (I never tried long enough before I saw a post on the ML telling it worked) . this is for gta04 only . my trick is only a addition to this http://lists.goldelico.com/pipermail/gta04-owner/2011-November/000656.html . Only for me it took a little more than 15 min ... and Aug 03 01:46:35 then it start to boot but fail while starting enligtnenment . at this point the battery is drained anew . So I tried unplugging the cable and plugging it back after the shr logo showed up . This seems to trigger full charge faster Aug 03 01:47:00 thanks, I will try that next time Aug 03 01:48:32 so my wondering if I was just lucky this time or if the full charge trigger really happen faster or at all only when unpluging the cable and pluging it back Aug 03 01:50:14 I just put a spare empty battery in my gta04. I will try when it turns on. Aug 03 01:53:00 It looks like it worked. Aug 03 02:00:30 :) I find this especially neat as I lost my hand on the gta02 I used to charge back the drained battery for a few weeks **** ENDING LOGGING AT Fri Aug 03 02:59:58 2012