**** BEGIN LOGGING AT Tue Aug 24 02:59:58 2010 Aug 24 08:29:32 JaMa|GoNe: i still have error with suspend Aug 24 08:30:03 then report it if you didn't Aug 24 08:30:08 actualy sometimes aafter suspend/resume cycle i get not clean reiserfs on card Aug 24 08:30:40 http://dpaste.com/233155/ Aug 24 08:30:45 somthing like this Aug 24 08:30:49 there is similar report for this too Aug 24 08:31:29 also seems it is fs independant Aug 24 08:32:45 http://dpaste.com/233157/ Aug 24 08:32:51 this after reboot Aug 24 08:33:43 seems it because something broken with mmchci Aug 24 08:34:35 ask gena2x or some other skilled kernel hacker to look into it (/me is not kernel hacker) Aug 24 08:35:18 ok =) Aug 24 08:35:37 gena2x: is not mmc expert, ask ThibG :) Aug 24 08:35:45 and try 2.6.34 (here it worked a bit better wrt uSD) Aug 24 08:37:06 hmm Aug 24 08:37:17 will i have working gsm and gps? Aug 24 08:37:22 yes Aug 24 08:37:25 maybe Aug 24 08:37:28 * alexxy uses neo as phone Aug 24 08:37:29 on my non-kms kernel, i never had problems with usd Aug 24 08:37:40 i mean .34 Aug 24 08:37:49 gena2x: its stock shr-u kernel Aug 24 08:37:49 but i think .32 also Aug 24 08:38:07 * JaMa|GoNe had few issues, but nothing reproducible Aug 24 08:38:16 alexxy: you may use other kernels with shr Aug 24 08:38:23 i know =) Aug 24 08:38:27 ok Aug 24 08:38:36 just fs got borked few times (before I moved ~/.efreet to volatile and started to use fsck -y on every boot) Aug 24 08:38:40 where can i get 2.6.34 kernel? Aug 24 08:38:48 shr-kms directory Aug 24 08:38:56 ok Aug 24 08:39:02 I'll upgrade it to 2.6.34.5 today Aug 24 08:39:03 i had reproducible problems with poweroff-sd-not-flushed with shr kernel i rebuild. Aug 24 08:39:07 * alexxy too lazy to compile new one Aug 24 08:39:24 gena2x: seems its related Aug 24 08:39:44 but its on suspend/resume cycle Aug 24 08:40:10 looks like not all fs changes were flushed back to sd Aug 24 08:40:17 you may also add sync, sleep and remount -o ro Aug 24 08:40:25 at least try. Aug 24 08:40:30 to workaround problem Aug 24 08:41:18 JaMa|GoNe: do you realy has wrking gsm? Aug 24 08:41:20 http://build.shr-project.org/shr-kms/ Aug 24 08:41:34 gsm is working on .34 Aug 24 08:41:38 or you simply forgot to adjust info on this page? Aug 24 08:41:42 alexxy: gsm yes with latest patch from gena2x Aug 24 08:41:47 alexxy: forgot Aug 24 08:41:54 =) Aug 24 08:42:17 alexxy: but you still need to adjust alsa scenarios a bit to get working sound in calls Aug 24 08:42:27 hmm Aug 24 08:42:29 dos should know how.. but he didn't push it yet Aug 24 08:42:31 how? =) Aug 24 08:42:38 ahh Aug 24 08:45:08 when he gets here, ask him to update README in shr-kms with that info Aug 24 08:59:24 alexxy: I've seen what I suspect to be 2.6.32 uSD oddities too Aug 24 08:59:33 yep =) Aug 24 08:59:34 but again nothing reproducible/diagnosable Aug 24 08:59:43 its reproducable Aug 24 08:59:43 =) Aug 24 09:00:07 and seems similar to poweroff-sd-not-flushed Aug 24 09:00:42 Installing kernel-2.6.34.5 (2.6.34.5-oe4.6+gitr6+3f6725d43021d2d7597027e36020df5b44d0667e-r0.5) to root... Aug 24 09:02:12 hmmm Aug 24 09:02:21 JaMa|GoNe: do you have feed for it? Aug 24 09:03:44 alexxy: http://jama.homelinux.org/org.openembedded.shr.images/ (use uImage+modules only) Aug 24 09:03:59 alexxy: but it's also building in shr-kms now Aug 24 09:04:06 NOTE: Running task 188 of 1195 Aug 24 09:04:11 ahh Aug 24 09:04:59 feed for kernel is dangerous, because it would remove your old kernel-module*2.6.32*.. while upgrading to 2.6.34 Aug 24 09:06:34 actualy its not a problem Aug 24 09:10:45 Hm, slightly OT, but I gues someone here might know. What's the relation of kernel tags (e.g. v2.6.35) to released tarballs (e.g. 2.6.35.2)? Aug 24 09:10:55 the tarballs have one more version number... Aug 24 09:11:20 (talking of tags in the official (linus) git tree) Aug 24 09:12:19 ? then you'll find 4 numbered versions in "stable trees" like linux-2.6.35.y http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.35.y.git;a=summary Aug 24 09:12:31 there are those tags for stable updates Aug 24 09:13:27 or I didn't understand the question Aug 24 09:25:35 JaMa|GoNe: no, you got it, thanks Aug 24 09:25:46 I didn't know I should for other trees Aug 24 09:52:55 hi mickey|office Aug 24 09:53:05 do you want to ear my roadmap? Aug 24 09:53:13 or are you busy? Aug 24 10:08:56 freesmartphone.org: 03GNUtoo 07utilities * r5d60b4f315f9 10/android/rpc/librpc/debug.h: (log message trimmed) Aug 24 10:08:56 freesmartphone.org: android: rpc: librpc: remove debuging prints Aug 24 10:08:56 freesmartphone.org: Debuging is useless now, Aug 24 10:08:56 freesmartphone.org: it only printed a lot of incomprensible(for us) values on the screen Aug 24 10:08:56 freesmartphone.org: Now that the GPS activator works, debugging is not necessary anymore Aug 24 10:27:47 btw http://laforge.gnumonks.org/weblog/2010/08/21/#20100821-playing_with_galaxy_s Aug 24 10:29:10 i wonder what are chances to make it succesfull, but that phone looks quite nice to me Aug 24 10:29:27 GNUtoo|laptop: i'm busy, but want to hear it anyways :) Aug 24 10:29:46 1) keyboard Aug 24 10:30:05 I already found a fix for gtk mapping issue Aug 24 10:30:11 but I've to integrate it Aug 24 10:30:13 2)GUI Aug 24 10:30:16 3) core things Aug 24 10:30:59 radekp: yup, read yesterday, very interesting indeed :) Aug 24 10:33:02 ok, sounds good Aug 24 10:33:42 since it's just the two of us, i don't see anyone else working on Bluetooth, Camera, and the sensors, right? Aug 24 10:38:26 radekp: 512Mb ram sounds exactly most important thing missing on FR. Aug 24 10:38:44 :) Aug 24 10:40:14 gena2x: there is much more missing :) Aug 24 10:40:22 mickey|office, that's for later Aug 24 10:40:37 first sensors Aug 24 10:40:39 then bluetooth Aug 24 10:41:01 and last camera,it would be the last thing Aug 24 10:41:07 because it's complicated Aug 24 10:41:23 camera requires an activator Aug 24 10:41:29 or to modify the kernel driver Aug 24 10:41:37 it's not using shared memory tough Aug 24 10:41:44 it's another thing Aug 24 10:42:30 btw: the plan is just a plan Aug 24 10:42:34 because of waiting times Aug 24 10:42:41 I may parallelize some tasks Aug 24 10:43:01 or rather pipeline some tasks Aug 24 10:43:07 radekp: i think, from raw and i/o performance point of view, memory speed is most limiting factor of fr. second very limiting thing is meory size. if fr has 512mb, i/o speed would be not a problem at all, as you can even just push small distos directly to memory, and everything will be lightning fast. Aug 24 10:45:02 gena2x: i wonder what has to be done for such port - only modem? rest should be in their android kernel? Aug 24 10:46:17 <[Rui]> moin Aug 24 10:46:20 radekp: please, do not go this journey right now :) Aug 24 10:47:25 gena2x: no worries - i am just interested :) Aug 24 10:48:38 radekp: i don't really know, but guess you should get huge bunch of absolutely unexpected troubles. out of article i am clearly see disk i/o toubles. Aug 24 10:49:10 gena2x: i'd boot from SD card :) Aug 24 10:50:17 radekp: would be not bad at all to have qtmoko ported on other device :) Aug 24 10:51:04 radekp: btw, how about v26? :) Aug 24 10:51:19 gena2x: i need a few hours to do it :) Aug 24 10:51:51 gena2x: but still busy with money work and with wife&children when i get home Aug 24 10:52:13 radekp: so, shindlier's list won and produced deep impression? :) Aug 24 10:53:22 gena2x: yup :) Aug 24 10:53:53 gena2x, palm pre plus has 512M Aug 24 10:58:48 but be sure to know the state Aug 24 10:58:50 GNUtoo|laptop: hm, nice to know, but atm, i am not searching for new device, as freerunner seem meets my needs, not without problems, but i guess any 'new' 'anti-vendor' port will have even more troubles. Aug 24 10:58:57 and also the differences between pre/pre-plus Aug 24 10:59:03 <[Rui]> let's hope after this upgrade things go well, I've been unable to take calls since yesterday's upgrade :) Aug 24 10:59:06 ok Aug 24 10:59:11 bbiab Aug 24 10:59:51 <[Rui]> gena2x: do you use the freerunner as a phone as well? :) Aug 24 11:01:11 [Rui]: from time to time - yes. and it work well. not without problems at all, but now it is well. Aug 24 11:01:36 <[Rui]> gena2x: great! :) now there's two of us ;) Aug 24 11:02:52 and most interesting, i am 'discovering' more and more nice features in it. Aug 24 11:03:37 last thing i enjoy is handwriting support in qtmoko Aug 24 11:04:13 <[Rui]> weird, the kernel upgrade seems to have some problem (missing drivers) Aug 24 11:04:25 just never tried it before, and once saw how it works, and wondered how well it is working. Aug 24 11:04:37 now i forgot about that damn keyboard. Aug 24 11:04:50 everywhere except terminals. Aug 24 11:05:01 <[Rui]> I think I'm going to backup the opimd dbs and reflash, this doesn't look good at all Aug 24 11:05:31 [Rui]: what do you mean by missing drivers? Aug 24 11:05:46 [Rui]: kernel config is the same as with 2.6.32.17 Aug 24 11:05:51 <[Rui]> root@om-gta02 ~ # ls -laF /lib/modules/2.6.32.17/kernel/sound/core/snd.ko Aug 24 11:05:51 <[Rui]> ls: /lib/modules/2.6.32.17/kernel/sound/core/snd.ko: No such file or directory Aug 24 11:05:57 so all modules which were installe before are also installed after Aug 24 11:06:05 [Rui]: yes you have to reboot to 2.6.32.20 Aug 24 11:06:13 <[Rui]> it complained about many modules when I did an opkg upgrade Aug 24 11:06:40 that's normal.. because it tries to modprobe them while you're still running 2.6.32.17 Aug 24 11:06:54 but loaded modules are kept loaded Aug 24 11:07:05 so if you reboot to 2.6.32.20 modprobe will find those Aug 24 11:07:12 in /lib/modules/2.6.32.20 Aug 24 11:07:26 <[Rui]> oh it upgraded to a new kernel version, not just a new patch... ok Aug 24 11:10:00 <[Rui]> ok, powering off after solving the terminal-keyboard/e-wm issue Aug 24 11:12:37 <[Rui]> shit, libphoneui-shr windows are still borked. I guess I'll definitely try a new base image, maybe something got too broken with some hard reboot, if this isn't happening to anyone else. Aug 24 11:12:59 * JaMa|GoNe reflashed yesterday and shr apps look normal Aug 24 11:13:47 [Rui]: phone ui apps gui broken? me thought i borked through installing another theme last night/this morning... Aug 24 11:15:58 JaMa|GoNe: 32.20 kernel is in shr feeds or in your testing repo? Aug 24 11:18:16 both Aug 24 11:18:27 vanous: 34.5 is building in shr-kms Aug 24 11:18:42 NOTE: Running task 597 of 1195 Aug 24 11:19:03 34 has this bad jitter on screen that kbd is unusable :( Aug 24 11:19:16 yes.. Aug 24 11:19:28 as mrmoku was talking about the filter i look at it but it goes over my head :) Aug 24 11:19:29 but I didn't notice it much in kbd.. rather in settings Aug 24 11:19:37 evfilter? Aug 24 11:19:37 i see Aug 24 11:19:49 yup Aug 24 11:20:05 userspace filter Aug 24 11:20:40 yup I wanted to try evfilter from metan (Hrubis) Aug 24 11:21:03 there are several modes Aug 24 11:21:51 but perhaps WeightedAverageAbs would be it? Aug 24 11:22:32 maybe.. but I fear "TODO: This filter wasn't tested yet." :) Aug 24 11:22:37 yup Aug 24 11:24:59 JaMa|GoNe: "Second goal is to hack X-server input driver to use evfilter library so we can do much more magic with input devices in X-server. " Aug 24 11:26:44 yes.. it's a lot of work.. you should also hack xinput to allow setup evfilter lib in runtime etc.. Aug 24 11:27:35 or even xinput-calibrator for testing evfilter settings before applying those Aug 24 11:28:07 hmm Aug 24 11:29:36 <[Rui]> JaMa|GoNe: well, now I don't have the cursor either :) Aug 24 11:30:37 [Rui]: and shr apps ok? maybe your theme wasn't upgraded during last upgrade and that's why you had broken shr apps and working cursor Aug 24 11:31:15 <[Rui]> JaMa|GoNe: not so ok Aug 24 11:31:28 <[Rui]> JaMa|GoNe: after the wizzard most UI is in a frozen state Aug 24 11:31:44 <[Rui]> config button reacts but no config shows up Aug 24 11:32:01 config as wrench icon? Aug 24 11:32:29 <[Rui]> yes, wrench icon. battery meter shows percentage when clicked, and tapping the tiny-wee free space in the top bar doesn't popup the window actions bar Aug 24 11:32:38 <[Rui]> I'm going to restart X from ssh Aug 24 11:32:54 check /tmp/x.log for errors Aug 24 11:33:23 <[Rui]> now it looks ok. Aug 24 11:33:28 <[Rui]> but still doesn't ask for pin Aug 24 11:34:06 <[Rui]> libphoneui-shr windows now look normal Aug 24 11:34:50 good Aug 24 11:35:13 and just for confirmation, no cursor, right? Aug 24 11:35:31 <[Rui]> JaMa|GoNe: no cursor Aug 24 11:58:08 am I the only one who can't pull ? Aug 24 12:00:18 <[Rui]> pull what? Aug 24 12:01:06 oe Aug 24 12:01:41 git://git.openembedded.org/openembedded Aug 24 12:01:45 fatal remote hung up Aug 24 12:02:07 <[Rui]> oh, heven't pulled in a couple of weeks, so I couldn't tell... too afraid to do it now and have to wait a few days before I'm able to compile again :) Aug 24 12:02:28 <[Rui]> JaMa|GoNe: remember my "freeze"? I think it's e that's freezing Aug 24 12:02:40 note your current commit and do git reset --hard to recover Aug 24 12:03:31 <[Rui]> enlightenment_remote -restart hangs as well Aug 24 12:08:33 [Rui]: any special way to reproduce it? Aug 24 12:10:43 <[Rui]> JaMa|GoNe: I was just loading the e config (wrench) and viewing options looking for the fps section when it last happened, but it happened on first boot as well, just after the shr wizzard finished. Aug 24 12:10:51 <[Rui]> JaMa|GoNe: might be some e instability Aug 24 12:11:02 <[Rui]> or illume Aug 24 12:11:08 and something in x.log? Aug 24 12:12:23 <[Rui]> brb baby stuff Aug 24 12:15:15 <[Rui]> JaMa|GoNe: let's see if it happens again and if sth appears there Aug 24 12:15:26 <[Rui]> it seems like that file is overwritten, not appended Aug 24 12:17:46 yes, because it gets long, you can change that in /etc/init.d/xserver-nodm Aug 24 12:22:38 <[Rui]> ca-certificates should be installed by default, btw Aug 24 12:22:53 <[Rui]> JaMa|GoNe: yeah... I can imagine so :) Aug 24 12:34:15 dos1: ping Aug 24 12:34:37 dos1: http://build.shr-project.org/shr-kms/ says Sound won't probably work in phone calls, ask dos1 on #openmoko-cdevel to publish diff for alsa scenarios. Aug 24 12:35:11 i have only speakerout :( Aug 24 12:35:57 mickey|office, also for the shr-themes that is in fso repo....JaMa said that I should patch in oe for making the htcdream theme, is it ok? Aug 24 12:36:21 GNUtoo|laptop: shr-themes are in shr repo, why to you still talk about fso? Aug 24 12:36:29 ah? Aug 24 12:36:30 s/to/do Aug 24 12:36:51 ah indeed Aug 24 12:36:53 url = http://shr.bearstech.com/repo/shr-themes.git/ Aug 24 12:37:02 it was in my fso dir Aug 24 12:37:05 or something like that Aug 24 12:37:13 that's why I mis-thought that Aug 24 12:37:22 sorry then Aug 24 12:37:30 np Aug 24 12:39:07 dos1: publish it anyways.. higher chance that someone will fix other scenarios Aug 24 12:39:21 dos1: please.. Aug 24 12:42:34 JaMa|GoNe: ok Aug 24 12:48:14 gena2x: I built mplayer from svn. it is a lot slower :) Aug 24 12:48:47 gena2x: and there's some offset in the fbdev output Aug 24 13:42:49 can anybody toss me a pointer please where to learn more about kobject.uevent, /sys/*/uevent and generally how a kernel device driver would signal an event to userland without userland processes polling /sys or whatever? Aug 24 14:06:52 mickey|office, any news about the activator integration? Aug 24 14:07:03 btw I removed debugging in librpc Aug 24 14:08:31 * GNUtoo|laptop tried openBmap but had that issue: Aug 24 14:08:56 dbus.exceptions.DBusException: org.freesmartphone.Resource.UnableToEnable: Can't open the device. Aug 24 14:09:07 so I really want the activator integration Aug 24 14:15:38 lindi-: exactly. i noticed that slowness too :) Aug 24 14:15:57 lindi-: and have no idea how to fix it :) Aug 24 14:16:56 last time I tried to send a patch to mplayer they told me that: Aug 24 14:17:02 *they don't care a lot about armv4 Aug 24 14:17:18 ah, sounds bad. Aug 24 14:17:30 (*they refused the patch because it should have gone into mpeg2 lib and not mplayer) Aug 24 14:17:45 ah, sounds ok :) Aug 24 14:17:54 ???? Aug 24 14:18:04 the * were like 1) 2) Aug 24 14:18:14 if it should go to mpeg2 lib, you should submit it to mpeg2 lib, no? Aug 24 14:18:21 not like * and things written in the smallest possible font possible Aug 24 14:18:29 I tried Aug 24 14:18:31 no one responded Aug 24 14:18:36 but it was long time ago Aug 24 14:18:52 interestion, what did that patch fixed? Aug 24 14:18:58 *interesting Aug 24 14:19:12 basically there were some > armv4 assembly Aug 24 14:19:20 it didn't compile with armv4 Aug 24 14:19:32 .pld or something like that Aug 24 14:20:07 the mpeg2 mailing list is not very active.... Aug 24 14:20:49 'don't care a lot' sounds ok in fact, as this is not really major platform. Aug 24 14:21:02 ineded Aug 24 14:21:23 I had to said that armv4 was used in freerunner, and that a lot of people bought it Aug 24 14:23:24 may be something were wrong with your patch? Aug 24 14:23:41 why it were not accepted? did it introduce regressions for armv5? Aug 24 14:24:15 imo, it may be worth also to tell 'debian is armv4' Aug 24 14:24:41 no it was right Aug 24 14:24:49 any link? Aug 24 14:24:51 it's only that mpeg2lib was an external lib Aug 24 14:24:55 just interesting to see Aug 24 14:25:13 and patching an imported external lib is not good Aug 24 14:25:22 it is right Aug 24 14:25:27 you have to reduce the diff as minimum as possible Aug 24 14:25:30 for external libs Aug 24 14:25:32 for instance Aug 24 14:25:41 if there is a security issue in let's say pnglib Aug 24 14:25:51 if everyone import and modify pnglib Aug 24 14:26:00 the packagers have to fix every packages Aug 24 14:26:06 and worse... Aug 24 14:26:13 the proprietary packages don't get fixed Aug 24 14:26:32 so better not using proprietary things Aug 24 14:26:37 so problem were actually with fact that external lib absence of maintainer, not with mplayer. Aug 24 14:26:57 yes Aug 24 14:27:05 basically mplayer told: Aug 24 14:27:12 send patch to lib maintainers Aug 24 14:27:21 lib maintainers mailing list was nearly dead Aug 24 14:27:27 so no way to get the patch upstream Aug 24 14:27:46 why do need this lib at all? only way to get mpeg2 for mplayer? Aug 24 14:28:45 any test show it is better than other implementation? Aug 24 14:28:46 no Aug 24 14:28:51 there is ffmpeg too Aug 24 14:28:54 it depends Aug 24 14:29:05 the only way to know which lib is better is to test Aug 24 14:31:25 you had patch :) Aug 24 14:31:55 I'll try to find again the mails Aug 24 14:32:18 i think only real difference in mpeg2 will be fourier floating point/fixed point type Aug 24 14:33:24 GNUtoo|laptop: i can't help you anyway, but it might be interesting to just see test results, as i've found mpeg2 is best codec (in software) for my freerunner. Aug 24 14:33:54 http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2008-October/058726.html Aug 24 14:34:06 ok Aug 24 14:34:14 maybe it's the same for htcdream Aug 24 14:35:38 hm, strange. my cross-compiled mplayer had no cross-compilation-specific issues. Aug 24 14:36:17 afair Aug 24 14:41:22 gena2x, maybe it has been fixed noe Aug 24 14:41:24 *now Aug 24 14:44:12 does someone know gps subsystem in fso? : Aug 24 14:44:13 I've: Aug 24 14:44:13 *no fsotldt.conf Aug 24 14:44:14 dbus.exceptions.DBusException: org.freesmartphone.Resource.UnableToEnable: Can't open the device. Aug 24 14:44:17 and I've: Aug 24 14:44:41 org.freedesktop.Gypsy available, I don't know from where it comes tough Aug 24 14:44:47 it's the gpsd compatible thing Aug 24 14:45:06 but I remember that stopping fso-gpsd didn't stop that Aug 24 14:45:09 but I'll check Aug 24 15:20:24 GNUtoo|laptop: GPS still comes from FSO1 and has only support for UBlox or NMEA. GPS in FSO2 will likely be part of a completely different location API, if at all Aug 24 15:20:48 ok Aug 24 15:20:52 I just found using the logs Aug 24 15:20:55 fsotldt.conf Aug 24 15:21:02 s/conf/log Aug 24 15:21:10 does that interact with frameworkd? Aug 24 15:22:51 not that i remember Aug 24 15:30:31 ok so what's gps config doing in fsotldt.conf? Aug 24 15:31:55 this is for my tests w/ the NMEA protocol parser in fsotdld Aug 24 15:32:03 which already works as time source Aug 24 15:32:21 ok Aug 24 15:32:24 I understand Aug 24 15:32:44 can frameworkd and fsotldt work at the same time? Aug 24 15:33:11 sure, they already do Aug 24 15:33:34 ok Aug 24 15:33:40 SHR-U is using fsotdld for time synchronization and alarm Aug 24 15:37:38 JaMa|GoNe: wha't the error while on boot: can't registed miscdev on minor 130? Aug 24 15:50:45 mickey|zzZZzz, where could the registration that occur be? Aug 24 15:50:48 for GPS Aug 24 16:55:44 bluez is not detecting bluetooth adapter in shr-u!!! Aug 24 16:56:06 neither bluetoothd is being run when powering up the bluetooth device Aug 24 18:34:39 hi mickey|zzZZzz Aug 24 18:35:00 I wait with impatience the gps integration commit Aug 24 18:35:25 should I write an ogpsd plugin Aug 24 18:35:29 what should I do to speed it up a bit? Aug 24 18:39:26 how does fso-gpsd ogpsd and fsotldt interact with each other? Aug 24 18:39:32 and what about the resource system? Aug 24 19:12:15 does anybody know what happened to this patch? http://marc.info/?l=linux-arm-kernel&m=122468318011905&w=2 Aug 24 19:15:45 gena2x: I think I figured out how to rotate video Aug 24 19:16:11 gena2x: so now the only missing part is to make mplayer byteorder correct Aug 24 19:17:10 lindi-, wow nice Aug 24 19:20:18 GNUtoo|laptop: rotate on PC that is Aug 24 19:29:31 ok Aug 24 19:29:56 mickey|zzZZzz, ping Aug 24 19:30:04 does anyone knows GPS? Aug 24 19:30:13 more specifically: Aug 24 19:30:38 what's the link between fso-gpsd ogpsd fsotldt adn the resource system Aug 24 19:30:49 The goal is to integrate the code I wrote Aug 24 19:30:54 that is to say: Aug 24 19:31:05 I wote a resource managememnt for htcdream's gps Aug 24 19:31:33 and I've Aug 24 19:31:35 ** (process:338): CRITICAL **: file MsmGpsResourceActivator.c: line 1430: unexpected error: Resource GPS already registered ((null), 153600) Aug 24 19:31:44 so should I modify ogpsd? Aug 24 19:32:01 there is a NMEA system, an openmoko one, an eten one.... Aug 24 19:32:26 # possible options are NMEADevice, UBXDevice, GTA02Device, EtenDevice Aug 24 19:32:30 in frameworkd.conf Aug 24 19:35:40 GNUtoo|laptop, I got a similar vala error with dbus the other day too, and it was due to sending wrong data types to dbus method Aug 24 19:35:57 ah Aug 24 19:36:03 here it's because it does what it says Aug 24 19:36:05 like sending an uint where it was a bool Aug 24 19:36:14 Resource GPS already registered Aug 24 19:36:20 it's already registered Aug 24 19:36:22 that is to say Aug 24 19:36:30 if I killall Aug 24 19:36:37 and start only some daemons Aug 24 19:36:39 it works Aug 24 19:36:48 but I've to integrate it into somehting usable Aug 24 19:37:00 I'll be back in a bit Aug 24 19:41:46 # mdbus2 -s org.bluez / org.bluez.Manager.DefaultAdapter Aug 24 19:41:47 org.bluez.Error.NoSuchAdapter: No such adapter Aug 24 20:27:41 hi mickey|zzZZzz Aug 24 20:27:45 I hope you're there Aug 24 20:27:47 I've big issues Aug 24 20:27:51 I'm trying to help with GPS Aug 24 20:28:03 but I'm still hitting the walls with my head....again Aug 24 20:28:17 basically I can't find who register the gps resource Aug 24 20:28:23 if I do that: Aug 24 20:28:31 /etc/init.d/dbus-1 stop Aug 24 20:28:37 fsodeviced Aug 24 20:28:45 I can register the resource Aug 24 20:28:47 s/the/the GPS Aug 24 20:28:51 but... Aug 24 20:29:03 else I can't find the gulty program Aug 24 20:29:08 frameworkd => no Aug 24 20:29:13 fso-gpsd => no Aug 24 20:29:16 etc... Aug 24 20:29:25 the procedure is to kill them Aug 24 20:29:30 or to stop them Aug 24 20:29:34 I bet that's not right Aug 24 20:29:41 maybe it doesn't unregister it Aug 24 20:29:47 I'll have to reboot a lot of times then Aug 24 20:30:17 please.... Aug 24 20:30:25 I'm doing that so you don't have to do it Aug 24 20:30:39 so you can concentrate on autotools for the plugin Aug 24 20:30:43 or on something else Aug 24 20:31:17 it's very frustrating to look to the code, and not to understand what it does, and to have no one to refer to Aug 24 20:31:21 and that mail doesn't work Aug 24 20:31:40 (I've issues with mail such as huge mailbox => risky on 3g) Aug 24 20:33:21 GNUtoo|laptop: frameworkd does register the GPS resource. Aug 24 20:33:30 ok Aug 24 20:33:40 GNUtoo|laptop: ogpsd part of it Aug 24 20:33:44 then why doesn't it go away when doing that: Aug 24 20:34:09 [ogpsd] Aug 24 20:34:09 disable = 1 Aug 24 20:34:24 maybe the [ogpsd.factory] Aug 24 20:34:25 ? Aug 24 20:34:31 but ogpsd is already disabled Aug 24 20:34:36 I'll look the logs Aug 24 20:34:39 thanks a lot!!!! Aug 24 20:34:58 GNUtoo|laptop: do you use lsfsor.py to list registered resources? Aug 24 20:35:01 GNUtoo|laptop: maybe fsotdld? Aug 24 20:35:18 PaulF, no, I used a separate vala code Aug 24 20:35:22 dos1: should it register the resource ever? I thought it can use GPS as a timesource if configured. Aug 24 20:35:23 it's separated from fso Aug 24 20:35:28 it doesn't use its libs etc... Aug 24 20:35:31 just dbus interface Aug 24 20:35:39 PaulF: dunno, just guessing Aug 24 20:36:09 mdbus2 -s | grep ogps Aug 24 20:36:09 org.freesmartphone.ogpsd Aug 24 20:36:16 but tdld stands for time&date&*location* daemon AFAIK Aug 24 20:36:17 that's what I have with ogpsd disabled in config Aug 24 20:36:30 GNUtoo|laptop: check again after killing fsotdld Aug 24 20:36:40 hmmm Aug 24 20:36:42 /var/log/fsotdld.log:2010-08-24T20:29:47.855834Z [INFO] DBusServiceResource : Ok. Registered with org.freesmartphone.ousaged Aug 24 20:37:10 dos1: wow, you're better informed :))) Aug 24 20:37:20 thanks a lot!!!!!!!! Aug 24 20:37:23 ~praise dos1 Aug 24 20:37:23 wow Aug 24 20:37:24 All hail dos1! Aug 24 20:37:26 indeed Aug 24 20:37:30 thanks a lot Aug 24 20:37:33 really Aug 24 20:37:47 dos1, basically I'll explain from scratch Aug 24 20:37:58 I've an htcdream which I want SHR on Aug 24 20:38:12 it came with android which had a proprietary gps lib Aug 24 20:38:20 gps was reverse-engineered Aug 24 20:38:24 dos1: i didn't even think about it as i thought the default configs have gps disabled as it's not working on FR anyway. Aug 24 20:39:18 the first version was a standalone android activator,which was converted to autotools, GNU etc...., its source code was unreadable Aug 24 20:39:35 then I found out that the author made an android lib for it Aug 24 20:39:44 with readable source code Aug 24 20:39:53 readable means not functions(hex,hex,hex) Aug 24 20:40:13 but function (MEANS_SOMETHING,MEANS_SOMETHING etc...) Aug 24 20:40:23 and still some small amount of hex Aug 24 20:40:29 I wanted to use the second source Aug 24 20:40:39 but it used some google MEANS_SOMETHING Aug 24 20:40:45 so it had to be apache2 Aug 24 20:40:50 fso is lgpl Aug 24 20:40:57 which is not compatible with apache2 Aug 24 20:41:08 so the only way is to use dbus to register it Aug 24 20:41:15 here's the background Aug 24 20:41:17 so... Aug 24 20:41:23 now it say something already registered gps Aug 24 20:41:31 how do I remove this something in fsotldt? Aug 24 20:42:19 and then: how do I make gps work within FSO? I mean you go in settings, you put it to on, you wait for a fix, and you get the data like position,3dfix,altitude etc... Aug 24 20:42:30 I already made the activator resource-aware Aug 24 20:42:38 it measn that If I do that: Aug 24 20:42:38 GNUtoo|laptop: Aug 24 20:42:40 #[fsotdl.provider_gps] Aug 24 20:42:41 #provider_type = nmea Aug 24 20:42:44 GNUtoo|laptop: in fsotdld.conf Aug 24 20:42:47 ok Aug 24 20:42:58 that will remove all fsotdl? Aug 24 20:43:12 or just the resource activaror? Aug 24 20:43:13 GNUtoo|laptop: no, just GPS provider i think Aug 24 20:43:17 ok nice Aug 24 20:43:19 thanks a lot Aug 24 20:43:52 GNUtoo|laptop: about functional GPS - implement point 7 of docs.freesmartphone.org Aug 24 20:43:56 org.freedesktop.Gypsy Aug 24 20:44:40 it's plain NMEA Aug 24 20:44:51 basically with the standalone activator you do that: Aug 24 20:44:51 gps Aug 24 20:44:55 cat /dev/smd27 Aug 24 20:45:08 and cat see plain NMEA on /dev/smd27 Aug 24 20:45:38 so I wonder Aug 24 20:45:42 I clone the nmea class Aug 24 20:45:43 hmmm Aug 24 20:45:45 ? Aug 24 20:45:48 I inherit it? Aug 24 20:45:51 or something like that? Aug 24 20:45:54 ask mickey|zzZZzz then Aug 24 20:45:56 and remove the activation Aug 24 20:45:58 ah ok Aug 24 20:46:04 that's out of my knowledge now :x Aug 24 20:46:13 ok thanks Aug 24 20:46:26 he also told me that fso could launch daemons Aug 24 20:50:28 thanks a lot dos1 Aug 24 20:50:35 I'll read fsotdlt source code then Aug 24 20:53:46 dos1, I'd like to understand the big picture Aug 24 20:53:50 there is fsotldt Aug 24 20:53:57 there is ogsmd in frameworkd Aug 24 20:54:09 there is odeviced Aug 24 20:54:11 etc... Aug 24 20:54:16 how do they work together for GPS Aug 24 20:54:28 s/ogsmd/ogpsd/ Aug 24 20:54:41 basically frameworkd starts ogpsd? Aug 24 20:54:56 if it's not present fsotdlt starts it, and it use it Aug 24 20:54:57 ? Aug 24 20:59:59 hmm Aug 24 21:00:14 fsotdld.source_gps is just copy of source_gsm with s/gsm/gps/ :x Aug 24 21:00:24 GNUtoo|laptop: ogpsd is part of frameworkd Aug 24 21:00:58 GNUtoo|laptop: i think they don't work together well, and that's why GPS part of fsotdld is disabled by default in SHR ;x Aug 24 21:01:27 fsotdld was supposed to replace ogpsd, but mickey|zzZZzz don't want to implement Gypsy api again for some reason Aug 24 21:01:46 so full GPS support in fsotdld is still missing, it can only sync time Aug 24 21:03:20 ah ok Aug 24 21:03:22 I get it Aug 24 21:03:33 so basically there are 2 conflicting implementations Aug 24 21:03:52 which both provide ogpsd Aug 24 21:03:57 dbus interface Aug 24 21:04:15 without both disabled one takes the ownership of the bus **** ENDING LOGGING AT Wed Aug 25 02:59:57 2010