**** BEGIN LOGGING AT Mon Sep 21 02:59:56 2009 Sep 21 05:15:46 openmoko: 03rehar * r5655 10/trunk/gta02-core/modules/ (INFO STATUS fh23-39s-0.3shw.fpd mkloe): added LCM connector FH23-39S-0.3SHW Sep 21 07:30:14 Hi Sep 21 07:30:40 I'm running Debian on my OpenMoko, but I seem to have upgraded it to a point where X does not start any longer :-) Sep 21 07:31:06 X will actually run fine if I just type in 'X', but I think I'm missing some package that would normally do that. Sep 21 07:31:33 On my laptop I'd run GDM, of course, but I'm not sure what people run for that sort of thing on the OpenMoko. Sep 21 07:31:36 Any recommendations? Sep 21 07:31:41 karora: so what happens when you start nodm? Sep 21 07:31:48 Aha. Sep 21 07:31:59 nodm, is that what I should be looking for? Sep 21 07:32:07 karora: check the wiki Sep 21 07:32:57 OK. Sep 21 07:33:12 /etc/init.d/nodm does nothing, but just running nodm appears to start X. Sep 21 07:45:12 karora: I use xdm Sep 21 07:46:06 lindi-: How do you type the login+pass? Sep 21 07:46:23 ...uhm you had a bt keyboard, right? Sep 21 07:46:45 Defiant: with xvkbd Sep 21 07:47:10 Defiant: I think everyone on openmoko uses on-screen keyboards anyway, right? Sep 21 07:48:21 lindi-: ah, so you start xvkbd with nodm, ok. Sep 21 07:48:58 Defiant: no, xvkbd with xdm Sep 21 07:49:07 Defiant: that's documented in xvkbd documentation Sep 21 07:49:13 sorry mean that Sep 21 07:51:16 lindi-: Really? OK, that's curious. Sep 21 07:51:40 I suppose I should log into my phone before I use it, but it seems so wrong, somehow. Sep 21 07:52:00 karora: it's reliability measure Sep 21 07:52:20 karora: if I break something as my normal user I can always login as a fallback user instead Sep 21 07:52:24 Ah, OK, so when everything goes AWOL it goes back to XDM login, and you start again. Sep 21 07:52:35 Oh, right. Sep 21 07:52:49 phone calls of course work without X at all here Sep 21 07:52:57 yes, true. Sep 21 07:53:01 As they should, I guess. Sep 21 07:53:08 I just use the AUX button to answer Sep 21 07:53:21 Yeah, I remember you telling me at Debconf9 Sep 21 07:54:23 * lindi- looks up group photo (I have terrible problem remembering people's faces) Sep 21 07:56:26 karora: btw, you can configure pam to not require you to type a password with xdm Sep 21 07:56:55 karora: that way you can login as your normal user by only typing username Sep 21 07:57:57 True Sep 21 07:58:28 I'm the one in the funny hat Sep 21 07:59:09 any of you ever use maltego? Sep 21 07:59:30 ndnihil: funny hat? Sep 21 07:59:42 Defiant: Curiously, a search on the wiki for 'nodm' returns 'no results' Sep 21 07:59:45 it was a joke Sep 21 08:00:11 I can connect my FR to my Windows machine again. But with 40% package loss and 10822 avg round-trip time. Sep 21 08:00:40 karora: http://wiki.openmoko.org/wiki/Debian - also follow the link to the debian wiki Sep 21 08:00:41 sounds right on par for windows Sep 21 08:01:22 ..but there is not much information there, utsl :) Sep 21 08:01:33 ndnihil: yep :) Sep 21 08:29:56 So if I just *run* nodm, I get a panel and stuff, but if I /etc/init.d/nodm nothing useful happens. Odd? Sep 21 08:30:49 karora: no output? Also check the source Sep 21 08:31:26 karora: a bug Sep 21 08:31:35 karora: is nodm any good yet? Sep 21 08:33:04 lindi-: Well, it would appear not for me, in any case. Sep 21 08:33:48 nodm works fine for my use cases Sep 21 08:34:01 karora: i'd file a bug report anyway Sep 21 08:34:11 Defiant: last time i tried it consumed even more memory than xdm, has that been fixed? Sep 21 08:35:07 lindi-: Don't have any numbers right now, but might check this evenning Sep 21 08:35:19 2963 ? S 0:00 /usr/sbin/nodm Sep 21 08:35:20 2965 ? S 0:00 /usr/bin/xinit /usr/sbin/nodm -- vt4 -nolisten tcp Sep 21 08:35:20 2967 ? Z< 0:00 [Xorg] Sep 21 08:35:37 I wonder what it is doing when it's started via init.d/nodm that means Xorg just dies. Sep 21 08:35:55 karora: strace? Sep 21 08:36:05 Yeah, might need to, I guess. Sep 21 08:36:51 How do you strace so it straces what gets started though? I'm not really a guru with strace, or even an initiate, to be honest. Sep 21 08:37:24 karora: strace -tt -o s -s256 -f /etc/init.d/nodm start Sep 21 08:37:32 karora: saves to file named "s" Sep 21 08:42:00 lindi-: there was that funny update when SHR became available on Debian that broke starting X. Sep 21 08:42:22 lindi-: i hard a hard times figuring the reason, it looks like nodm is not quite friendly if anything goes not as expected. Sep 21 08:45:42 From the strace I see that "Unrecognized option: -nolisten tcp" Sep 21 08:45:42 PaulFertser: isn't shr a distro? how can it become vailable on debian? Sep 21 08:45:44 :-) Sep 21 08:45:50 karora: ha :) Sep 21 08:46:17 lindi-: shr apps i meant Sep 21 08:46:23 So where is that option entering the equation, I wonder. Sep 21 08:46:49 karora: are you using kdrive==xglamo? If yes, i suggest you to stop. Sep 21 08:47:20 karora: because xorg driver for glamo is actively supported and works good and reliable. Sep 21 08:48:32 I thought I was using glamo, but maybe I am not. Sep 21 08:48:38 How can I tell? Sep 21 08:49:00 karora: if you don't know, install xorg and glamo driver for it. Sep 21 08:49:07 karora: are you debian user? Sep 21 08:49:12 Sure. Sep 21 08:49:14 :-) Sep 21 08:49:25 whne someone tell me "can be found iirc" what are they meaning Sep 21 08:49:41 karora: then just do it! Sep 21 08:49:44 if I recall correctly. Sep 21 08:49:48 hi Sep 21 08:49:55 ah Sep 21 08:50:06 PaulFertser: So I did apt-get install xserver-xorg-video-glamo or whatever. Sep 21 08:50:57 /var/log/Xorg.0.log definitely smells like Glamo to me.. :-) Sep 21 08:51:04 (II) Glamo: driver for glamo: Glamo Sep 21 08:51:04 (WW) Falling back to old probe method for Glamo Sep 21 08:51:04 ( Sep 21 08:51:08 E.g. Sep 21 08:51:33 karora: the question is if you're using xorg or not... Sep 21 08:51:55 karora: if you're using xorg, it should support -nolisten tcp, don't you think so? Sep 21 08:52:12 Ah, I see. Sep 21 08:52:27 Yep, I hadn't thought of thinking I was running a non-standard X server. Sep 21 08:53:20 I donxserver-xglamo Sep 21 08:53:31 't seem to have xserver-xglamo installed. Sep 21 08:54:48 I kind of like the idea of an X that doesn't bother implementing enough that it needs a -nolisten tcp :-) Sep 21 08:56:13 karora: kdrive was deprecated for various reasons years ago, parts of its functionality was ported to xorg. Modern modular X.org is good enough even for relatively constrained devices. Sep 21 09:00:11 * SpeedEvil tries to work out hte smallest hardware he's run X on. Sep 21 09:00:30 I think 486/25/8M Sep 21 09:01:04 hah, same here Sep 21 09:01:09 I still have the drive it was installed on Sep 21 09:01:16 controller fried Sep 21 09:01:32 I havent had a workign 486 to tr it on.. mine was p75 32 MB ram Sep 21 09:01:35 1.2.13 I think it was Sep 21 09:01:57 a whopping 240M drive Sep 21 09:02:16 I've now got mp3's larger than that Sep 21 09:04:41 I could load 16 copies of that entire drive into ram on this workstation, and still have enough left over to play enemy territory Sep 21 09:05:22 I have 5 1.3" 20M drives Sep 21 09:05:37 I need to plug them up to a USB converter sometime to see if they still work :) Sep 21 09:05:56 this past summer I threw out a 3b1 Sep 21 09:06:08 had the 40M HD upgrade Sep 21 09:06:30 drive was about the size of a lunchbox Sep 21 09:06:33 I threw away anything under 5 GBs last lear Sep 21 09:07:13 it's scary, I've got about 9T of total storage here Sep 21 09:07:14 I still have a couple of double height SCSI drives that I'm using as bookends. Sep 21 09:09:14 http://en.wikipedia.org/wiki/3B1 Sep 21 09:09:18 I kinda wish I had kept it Sep 21 09:09:30 but I was in one of those "shit has to go" kind of cleaning moods Sep 21 09:09:45 tossed out a ][c clone along with it Sep 21 09:10:05 80 column card applecat 300baud Sep 21 09:10:08 it was pimped out Sep 21 09:10:30 (I never found that copy of the original oregon trail I was hoping to find) Sep 21 09:12:35 SpeedEvil: yeah, with cyrrusLogic graca. Was a PITA to set up the X modelines Sep 21 09:13:25 Oh, i remember X modelines. I understood everything and could easily prepare a modeline based on information from my monitor's user manual. Also to convert between X modelines and fbset :) Sep 21 09:14:00 I'm at the point of wondering how to change laptop modelines on an LCD swap Sep 21 09:14:03 PaulFertser: try that with the fsckng cyruslogic chip! Sep 21 09:14:40 DocScrutinizer: i remember i had et6000 on my first X-capable PC. Sep 21 09:14:59 So it wasn't CL luckily. Sep 21 09:15:12 SpeedEvil: same here. now sis graca, and 768*1280, instead of the 800*1280 the display needs :-( Sep 21 09:15:38 moin btw Sep 21 09:17:04 PaulFertser: Pretty sure there's no kdrive running on here. I installed Debian on it some months ago (well, mid-July :-) and then upgraded and now nodm gives me this strange error. Or lack of error, really. Sep 21 09:17:21 karora: xglamo is kdrive Sep 21 09:17:46 OK, so that package (xserver-xglamo) is not installed. Sep 21 09:39:17 So is the package xserver-xorg-video-glamo also kdrive, in some way? Sep 21 09:40:14 karora: no Sep 21 09:40:18 karora: it's normal Xorg driver. Sep 21 09:40:32 Strange. Sep 21 09:41:14 If I create an xorg.conf which says Driver "glamo" then it seems to use that driver, but doesn't accept the -nolisten tcp option Sep 21 09:43:12 karora: i'm using this xorg.conf: http://www.bitwiz.org.uk/openmoko/xorg.conf Sep 21 09:43:35 but haven't tried -nolisten tcp Sep 21 09:44:22 It's just that /etc/default/nodm has -nolisten tcp in the X options that it sets. Sep 21 09:44:36 So X doesn't start when the glamo driver is being used. Sep 21 09:46:30 karora: does it start with xserver-xorg-video-fbdev driver? Sep 21 09:47:30 lindi-: Yes, it did. -nolisten tcp worked OK with that one. Sep 21 09:49:44 karora: very odd Sep 21 09:51:05 karora: ask Weiss about that, he can probaly fix that Sep 21 09:51:09 karora: in any case, file a bug, i'm sure Mirv is interested in this (he maintains the driver package in debian) Sep 21 09:51:30 OK. Sep 21 12:52:14 are there settings in alsamixer that i should adjust when using my freerunner for listening to music via a media player like intone? Sep 21 12:52:58 i've adjusted bass filter to 100Hz @ 8kHz Sep 21 12:53:12 and cranked up bass to 15 Sep 21 12:53:28 as recommended in the wiki Sep 21 13:00:28 djp: music? Sep 21 13:00:39 djp: without bass fix you won't gain quality with any settings. Sep 21 13:02:33 PaulFertser: is a bass fix ever liklely? Sep 21 13:02:45 djp: if you do it yourself yes Sep 21 13:03:10 Or if you use external phones Sep 21 13:03:18 Bluetooth phones Sep 21 13:03:25 Or that Sep 21 13:03:30 I meant that. Sep 21 13:03:32 Defiant: hi. i see. so it's a hardware fix? Sep 21 13:03:38 Or amplified phones - which are rare Sep 21 13:03:41 djp: yes Sep 21 13:03:41 djp: GTA02_bass_fix on the wiki Sep 21 13:03:48 djp: it's a hardware bug. Sep 21 13:03:55 thanks PaulFertser Sep 21 13:03:57 djp: not fixable really in software Sep 21 13:04:06 won't BT drain way more power? Sep 21 13:04:11 SpeedEvil: no, even with an amplifier of infinite resistance the cutoff is still too high due to pulldowns on lines. Sep 21 13:04:22 Ah - I forgot those. Sep 21 14:35:12 hi all Sep 21 15:28:16 i'm not getting any response from the gps using the traditional means on shr-u...is there a way to query it to determine if it's alive? Sep 21 15:29:37 I'm trying to pair my moko with shr-unstable with my PC over USB, but it isn't working anymore, have there been destructive updates the last few days ? Sep 21 15:34:15 yay for correct timestamps on sms after upgrade Sep 21 15:41:15 Bert_2: shr-u? Sep 21 15:45:16 smt line for free woohoo Sep 21 15:54:53 Phone GPS the better way. http://www.youtube.com/watch?v=4Fl718QO_xQ Sep 21 15:56:46 cute Sep 21 15:59:45 nice :] Sep 21 15:59:57 is anybody implementing that stuff to F? Sep 21 15:59:59 FR? Sep 21 16:22:06 Say, can I update my moko inside a chroot ? Sep 21 16:22:14 or is that impossible because it's another architecture ? Sep 21 16:22:42 chroot does not do other architectures Sep 21 16:22:56 It simply runs programs on the CPU as normal with a different filesystem / Sep 21 16:23:27 SpeedEvil: damn, cause the eth device doesn't appear anymore (probably did an update on the wrong time) Sep 21 16:24:00 you don't need eth device to flash Sep 21 16:24:47 SpeedEvil: I don't want to flash, I want to opkg update && opkg upgrade Sep 21 16:24:55 (so I need to login over ssh) Sep 21 16:25:53 it might work via wifi Sep 21 16:26:15 opkg update is generally quite a flaky way to upgrade Sep 21 16:26:22 tschaka: yeah, but I'm on the 2nd floor here with special isolation for the winter cold, so I can't pick up anything Sep 21 16:26:23 it often doesn;t work well Sep 21 16:26:53 SpeedEvil: I know, but it's been quite some time since a new image of shr-unstable was released Sep 21 16:26:59 k Sep 21 16:32:24 eth2 is back, did a few reboots, probably had something to do with the moment of USB connecting the device Sep 21 16:35:51 :) Sep 21 16:37:33 quite some time? Sep 21 17:21:15 Hi PaulFertser Sep 21 17:21:37 panicking: hey :D Sep 21 17:21:58 ok, I decided to move Sep 21 17:22:03 to the 2.6.3x kernel Sep 21 17:22:22 Can we have a reposistory on openmoko Sep 21 17:22:26 and rebase on it Sep 21 17:22:28 or merge Sep 21 17:23:48 the openwrt patches are good start point Sep 21 17:23:59 panicking: we will have it very very soon Sep 21 17:24:09 panicking: just wait for larsc, he's on it. Sep 21 17:24:22 Ok, but I will do the same thing Sep 21 17:24:28 I prefer a public repository Sep 21 17:24:37 because there are a lot of thing to discuss Sep 21 17:24:57 The accelerometer for example Sep 21 17:25:01 are still present Sep 21 17:25:06 on the linux-2.6 Sep 21 17:25:14 and the bluetooth interface has changed Sep 21 17:25:21 I just compile on the new kernel Sep 21 17:25:31 I would like to have some good and clean code Sep 21 17:28:13 I think that step by step patches with some discussion is the best thing Sep 21 18:02:35 larsc, Hi Sep 21 18:02:38 are you there? Sep 21 18:12:15 panicking: it will be public OM repository. Sep 21 18:12:27 panicking: hi, yes i am Sep 21 18:12:35 panicking: what's about accelerometer? Aren't you happy with the interface current andy-tracking has? Sep 21 18:12:45 No, it's wrong Sep 21 18:12:51 the spi change Sep 21 18:12:51 panicking: and what about bluetooth? I think it's exactly the same on .29 and .31. Sep 21 18:12:56 no Sep 21 18:13:11 there are a lot of change Sep 21 18:13:13 panicking: on FR bluetooth is USB Sep 21 18:13:13 I apply Sep 21 18:13:26 rfkill Sep 21 18:13:29 interface Sep 21 18:13:41 I use pastbin Sep 21 18:13:45 panicking: i have fixed versions of them Sep 21 18:13:45 to show somenting Sep 21 18:13:52 panicking: our FR bluetooth is generic usb HCI, nothing special. Sep 21 18:14:10 panicking: why do you think you need to do some additional rfkill magic? Sep 21 18:14:12 bt-rfkill Sep 21 18:14:13 ? Sep 21 18:14:22 struct rfkill is changed Sep 21 18:14:31 larsc, can you confirm Sep 21 18:14:32 ? Sep 21 18:14:35 yes Sep 21 18:15:00 and i adjusted the driver to use the new one Sep 21 18:15:01 I rewrite the pwm interface Sep 21 18:15:02 too Sep 21 18:15:05 and you the mainlin Sep 21 18:15:23 ok , but where is the git because we have rewrite the same thing Sep 21 18:15:27 I have done today Sep 21 18:15:31 a couple of change Sep 21 18:15:52 I have a ready compiled version Sep 21 18:15:54 panicking: openwrt patches are public and available for several months already. Sep 21 18:16:03 There are wrong Sep 21 18:16:08 for the mainline kernel Sep 21 18:16:13 no they arn't Sep 21 18:16:17 sorry Sep 21 18:16:18 :) Sep 21 18:16:42 Do you change Sep 21 18:16:43 the pwm Sep 21 18:16:46 initialization Sep 21 18:16:49 no Sep 21 18:16:50 in the mach-gta02 Sep 21 18:16:51 ? Sep 21 18:17:03 Do you remove the spi accelerometer Sep 21 18:17:07 _sync hack Sep 21 18:17:08 ? Sep 21 18:17:11 yes Sep 21 18:17:25 the sync hack didn't even work for me Sep 21 18:17:32 Can I clone your repository Sep 21 18:17:37 and apply and test my change Sep 21 18:17:38 ? Sep 21 18:17:51 i don't have a repository Sep 21 18:19:20 Can you send your updated versione Sep 21 18:19:26 larsc: but no kidding, when are you going to push your .31? Sep 21 18:19:27 of pm-bt Sep 21 18:19:35 pm-wlan Sep 21 18:19:36 etc Sep 21 18:19:36 ? Sep 21 18:19:46 panicking: can you please tell me why do you care at all about rfkill??? Sep 21 18:19:51 panicking: what's use of it? Sep 21 18:20:18 PaulFertser: i'm working on it Sep 21 18:20:20 I try to understand all the change from 2.6.30 and 2.6.31 needed Sep 21 18:20:30 and cleanup the patches Sep 21 18:20:30 larsc: i know :) Sep 21 18:20:44 to have less diff from the mainline Sep 21 18:21:10 panicking: every time we talk you mention rfkill. It can't be a simple coincidence :) I just can't understand why this kernel subsystem is needed at all and you care about it. Sep 21 18:22:35 we are several question: Sep 21 18:22:44 - fiq integration and pwm Sep 21 18:22:58 - accelerometer integration Sep 21 18:23:06 - glamo video device Sep 21 18:23:12 - board revision Sep 21 18:23:20 make them as clean as possible Sep 21 18:23:29 so I'm working on it Sep 21 18:23:34 panicking: have you seen https://dev.openwrt.org/browser/trunk/target/linux/s3c24xx/patches-2.6.30 Sep 21 18:23:46 They are wrong and not clean Sep 21 18:23:54 do you have apply on a git and try to compile? Sep 21 18:24:15 on linus-2.6 tree Sep 21 18:24:16 ? Sep 21 18:24:46 We need a repository and not a couple of patches Sep 21 18:24:54 we need a review process Sep 21 18:24:57 panicking: larsc is working on it Sep 21 18:25:07 The community must work on it Sep 21 18:25:17 or is a closed process Sep 21 18:25:18 ? Sep 21 18:25:46 better patches come from a discussion Sep 21 18:25:53 panicking: no, but he's .31 already running. Just wait for him to push it to the OM repo and then join... Or is it too long for a wait? Sep 21 18:25:54 on a mailing list with some review process Sep 21 18:26:25 I think that each change must be discussed. So start step by step can be a good thing Sep 21 18:26:36 for all Sep 21 18:26:41 to have some clean tree Sep 21 18:26:52 that is my idea Sep 21 18:26:56 panicking: aha, that's what you want... Sep 21 18:27:10 that is how the project go ahead Sep 21 18:27:13 linux Sep 21 18:27:14 u-boot Sep 21 18:27:19 panicking: why didn't you say it explicitly earlier, when larsc wrote that email to the kernel ML about our future? Sep 21 18:27:20 openmoko (andy) Sep 21 18:27:47 panicking: i thought you were ok with the way when larsc pushes all his patches and then we use it as a base. Sep 21 18:28:36 larsc, PaulFertser Ok, but is it possible to have a step by step review on the mailing list Sep 21 18:28:44 in every single patch Sep 21 18:28:50 before commit it on the repository Sep 21 18:28:50 ? Sep 21 18:28:52 panicking: i guess you should have requested it a little bit earlier. Sep 21 18:29:06 panicking: do you have any particular comments about openwrt patchset? Sep 21 18:29:12 Yes Sep 21 18:29:15 a lot Sep 21 18:29:43 panicking: feel free to post them to the ML or discuss here. Sep 21 18:30:06 use of the gpilib Sep 21 18:30:20 rename of all gpio to GPG(4) instead of GPG4 Sep 21 18:30:24 etc Sep 21 18:30:29 there a lot of issue Sep 21 18:30:40 Cleanup comment do avoid difference from the mailine Sep 21 18:30:54 move to the pwm interface Sep 21 18:31:17 but without the latest patcheset I can't say a lot Sep 21 18:31:32 panicking: https://dev.openwrt.org/browser/trunk/target/linux/s3c24xx/files-2.6.31 feel free to clean it up Sep 21 18:32:39 ok thanks Sep 21 18:33:10 and wanted to do more cleanup before commting, so it might not be in it's best shaped Sep 21 18:33:29 but at least it should compile and boot Sep 21 18:35:12 it is possible file that are present Sep 21 18:35:16 in the mainline Sep 21 18:35:20 to have a diff file Sep 21 18:35:23 instead a new file Sep 21 18:35:41 mach-gta02 Sep 21 18:35:43 is present Sep 21 18:36:00 nope Sep 21 18:36:05 not yet Sep 21 18:36:06 I notice that with the previus patches Sep 21 18:36:16 ok Sep 21 18:36:22 the previous patches were against 2.6.30 Sep 21 18:36:28 2.6.30 had no mach-gta02.c Sep 21 18:36:47 2.6.31 Sep 21 18:36:48 yes Sep 21 18:36:57 and in the files-2.6.31 Sep 21 18:37:01 i know Sep 21 18:37:02 there is mach-gta02.c Sep 21 18:37:14 thats one of the cleanups i was working on Sep 21 18:37:20 ok Sep 21 18:37:27 I have done a change Sep 21 18:37:31 in the accelerometer Sep 21 18:37:34 for android Sep 21 18:37:39 what change? Sep 21 18:37:50 deactivate accelerometer Sep 21 18:38:01 and activate accelerometer Sep 21 18:38:08 for now is raw hack Sep 21 18:38:22 but it is very usefull Sep 21 18:38:30 because close all descriptor Sep 21 18:38:34 it's deactived unless somebody queries data from it Sep 21 18:38:35 is the only alternative Sep 21 18:38:41 fd Sep 21 18:39:10 In android for example the module Sep 21 18:39:15 abstraction open the fd Sep 21 18:39:23 and then require activation/deactivation Sep 21 18:40:08 http://git.moko.serdar-dere.net/cgi-bin/?p=kernel/openmoko.git;a=commitdiff;h=9e8fb53b4bca8ef61588653dc6db4d010bf7eb22 Sep 21 18:40:15 somenthing like this Sep 21 18:40:50 this is an HACK Sep 21 18:41:01 I don't find a lot of accelerometer Sep 21 18:41:03 in the mainlin Sep 21 18:43:41 panicking: how is accelerometer any different from usual input devices except that you have additional sysfs nodes to change some device-specific parameters? Sep 21 18:45:22 ok I increase the sensitivy Sep 21 18:45:32 and deactivate if is to high Sep 21 18:45:52 I don't change the interface Sep 21 18:47:58 panicking: i'm sorry but i have difficulties understanding you. I'm not sure about larsc. Sep 21 18:50:44 If you close all file descriptor Sep 21 18:50:57 you basically shutdown the accelerometer Sep 21 18:51:04 ok, this is impossible using android Sep 21 18:51:07 because the interface Sep 21 18:51:13 open the file descriptor Sep 21 18:51:23 and use activate/deactivate Sep 21 18:51:31 to shutdown the acceleromente Sep 21 18:51:38 but isn't it androids fault then? Sep 21 18:51:39 I ask if it is possible to change the interface Sep 21 18:51:41 panicking: well, it doesn't prove that the current implementation is wrong Sep 21 18:51:51 It's just a feature Sep 21 18:51:59 deactivate/deactivarte Sep 21 18:52:01 panicking: probably it's android that doesn't follow established Linux practicies? Sep 21 18:54:42 Well I don't find a lot of accelerometer Sep 21 18:54:44 in linux Sep 21 18:54:52 can you point to other one Sep 21 18:54:52 ? Sep 21 18:55:25