**** BEGIN LOGGING AT Wed Nov 25 02:59:57 2009 **** ENDING LOGGING AT Wed Nov 25 03:35:07 2009 **** BEGIN LOGGING AT Wed Nov 25 03:35:39 2009 Nov 25 06:41:30 TAsn: isn't StopSound or StopAllSounds that does exactly that? (mute ringtone) Nov 25 06:44:08 PaulFertser: solved the sound issue Nov 25 06:44:15 (thanks to larsc) Nov 25 06:44:23 PaulFertser: and morning :) Nov 25 07:08:14 mornin' Nov 25 07:11:56 mrmoku, pleas read my full discussing with Doc, I explained exactly why this isn't enough. Nov 25 07:20:52 freesmartphone.org: 03mickey 07cornucopia * rd17b0e589549 10/docs/frameworkd.conf.sample: document fsodevice.accelerometer_lis302 values Nov 25 07:24:51 TAsn: hmm... can't find anything about StopSound Nov 25 07:25:08 TAsn: btw. I very much disagree about completely handling ringtones Nov 25 07:25:33 TAsn: that is what we have FSO for... to abstract the underlying device Nov 25 07:26:00 TAsn: and on gta02 we want the pure alsa player... but on something else we want something else Nov 25 07:26:03 that might handle mp3 Nov 25 07:26:08 as the hw is able to do that Nov 25 07:26:18 and _I_ don't want to take care of that in UI Nov 25 07:26:39 mrmoku, you got us wrong Nov 25 07:26:53 handling ringtones means issuing Nov 25 07:27:11 FSO.PlaySound(filename) Nov 25 07:27:16 FSO.Vibrate(on) Nov 25 07:27:25 not actually playing stuff on our own Nov 25 07:27:28 ahh Nov 25 07:27:29 ok :) Nov 25 07:27:31 so hw abstraction is still there Nov 25 07:27:46 btw, we are not really hardware agnostic Nov 25 07:27:57 because of libphone-ui's alsa handling :| Nov 25 07:28:15 Either fso will expose methods for us, or we'll have to make different configs for different devices. Nov 25 07:28:20 yeah, but we should limit that :P Nov 25 07:28:36 I completely agree. Nov 25 07:28:43 Anyhow, I gtg Nov 25 07:28:45 please test mute Nov 25 07:28:54 yup Nov 25 07:28:55 cu Nov 25 07:29:03 should work :P Nov 25 07:29:15 TAsn: btw. you should try 2.6.31... in a moment you don't need your phone ;) Nov 25 07:29:19 it really rocks Nov 25 07:29:28 mrmoku, (and I don't remember what I else I had in mind last night, but I sent you many pms) Nov 25 07:29:35 mrmoku, :) I know. !!! bah! Nov 25 07:29:42 mrmoku, but I need my phone ;( Nov 25 07:29:49 mrmoku, btw, what about module stripping? Nov 25 07:29:58 Please bump PR Nov 25 07:30:02 and build Nov 25 07:30:02 TAsn: again? Nov 25 07:30:06 nah Nov 25 07:30:10 why not? Nov 25 07:30:15 that should all already be built Nov 25 07:30:18 no? Nov 25 07:30:33 no I idea Nov 25 07:30:37 but kernel didn't update for me Nov 25 07:30:37 :) Nov 25 07:30:40 so I guess not. Nov 25 07:30:43 (opkg upgrade) Nov 25 07:30:45 hmm... thinking about it Nov 25 07:30:53 with opkg upgrade did not get anything too Nov 25 07:31:03 will take a look Nov 25 07:31:04 but I saw a major opkg upgrade Nov 25 07:31:17 mrmoku, you probably didn't bump PR or something annoying like that ;) Nov 25 07:31:21 anyhow, I'm off, ciao. Nov 25 07:31:27 (I'm already half an hour late to work) Nov 25 07:31:33 then hurry Nov 25 07:31:36 and shut up ;) Nov 25 07:37:27 morning. how can I try the .31 kernel? Just add it to the feed? Nov 25 07:40:31 spaetz: no Nov 25 07:40:38 spaetz: install it manually Nov 25 07:40:46 spaetz: download the uImage and the modules.tar.gz Nov 25 07:40:59 spaetz: put the uImage as uImage-2.6.31 into /boot Nov 25 07:41:05 spaetz: and unpack the modules Nov 25 07:41:30 spaetz: then cd /boot; ln -sf uImage-2.6.31 uImage Nov 25 07:41:46 spaetz: and run the script that I'm preparing :) Nov 25 07:42:26 sounds good. downloading... Nov 25 07:46:47 uhh just 1.8MB for uImage and 239kb for modules? small Nov 25 07:47:05 isn't our vmlinuz image 18MB or something like that? :) Nov 25 07:47:47 spaetz: no, the modules for 2.6.29 are big Nov 25 07:47:53 because there are many modules Nov 25 07:48:02 for 2.6.31 its just sound + vfat Nov 25 07:50:24 ahh. sure Nov 25 07:56:51 spaetz: http://build.shr-project.org/tests/mrmoku/switch-for-kernel Nov 25 07:57:11 call it with parameter new for 2.6.31 or without for the old one Nov 25 07:57:46 will extend it to fiddle with alsa state files too Nov 25 07:57:59 and then the last missing piece (I think) is fsodeviced Nov 25 07:58:04 mrmoku: modules for 2.6.29 are not that big now no? :) Nov 25 07:59:07 mrmoku: maybe add depmod to your script Nov 25 07:59:16 JaMa: heh, yeah actually they shrinked from 31M to 2.5M Nov 25 07:59:19 mrmoku: but I'm not sure, maybe modules.dep is tarbaled Nov 25 07:59:39 JaMa: does update-modules not depmod too? Nov 25 08:00:05 JaMa: there is some version problem with the kernel though Nov 25 08:00:19 ahh... localcount Nov 25 08:00:31 that's why we don't get kernel updates :P Nov 25 08:00:42 mrmoku: for current kernel yes.. Nov 25 08:00:59 mrmoku: but if you're switching with 2.6.29 running i think it want call depmod for 2.6.31 Nov 25 08:01:21 JaMa: ahh... hmm yeah Nov 25 08:01:47 JaMa: btw. for sound to work you have to load an alsa state Nov 25 08:01:51 after that all is fine Nov 25 08:02:01 (if you do it with an adjusted state file for 2.6.31) Nov 25 08:02:11 mrmoku: with alsactl? Nov 25 08:02:17 JaMa: for example Nov 25 08:02:30 JaMa: will do new-format state files for 2.6.31 later Nov 25 08:02:39 ah.. that's probably why it worked for me once.. Nov 25 08:02:45 anyone tested zoff's navit script on the freerunner? Nov 25 08:02:46 now I have to go to the waste-disposal... otherwise my wife will beat me :P Nov 25 08:02:53 yesterday I tried and got the same message as you Nov 25 08:02:53 bbl Nov 25 08:02:57 yeah Nov 25 08:03:06 JaMa: just take the statefiles from openwrt Nov 25 08:03:16 mrmoku: ok thanks Nov 25 08:03:25 https://dev.openwrt.org/browser/feeds/phone/freerunner-alsa-scenarios/files Nov 25 08:03:57 JaMa: does fltk2 work for you now? Nov 25 08:04:05 playyaWorking__: I want before http://trac.navit-project.org/ticket/506 Nov 25 08:04:16 s/want/won't/ Nov 25 08:04:16 JaMa meant: playyaWorking__: I won't before http://trac.navit-project.org/ticket/506 Nov 25 08:04:36 blindcoder: yes, pushed, thanks.. but I solved it a bit different id the end.. Nov 25 08:05:01 blindcoder: http://cgit.openembedded.net/cgit.cgi/openembedded/commit/?id=dc8558fc70d80b845548a46bb8f46cc154bd90cc Nov 25 08:05:10 i fixed the script. i cannot test the menu because i can't configure my touchscreen Nov 25 08:05:45 JaMa: so do_configure overrides the autotools_do_configure from 'inherit autotools'? Nov 25 08:05:47 playyaWorking__: it will segfault in menu, I'm pretty sure.. tested here on 3 different architectures.. Nov 25 08:05:56 ok Nov 25 08:06:09 blindcoder: no.. do_configure_autotools is called only with inherit autotools Nov 25 08:06:23 blindcoder: with autotools_stage is not Nov 25 08:06:31 JaMa: aaah, okay. gotta remember that Nov 25 08:07:18 blindcoder: i changed it "full" autotools inherit to get call to generate configure, but after that there was that problem with X11 detection broken.. Nov 25 08:07:43 JaMa, have you tested it with ? Nov 25 08:07:52 playyaWorking__: yes on all 3 Nov 25 08:08:14 playyaWorking__: and that bug is probably only with internal Nov 25 08:08:30 playyaWorking__: as it segfaults in gui_internal .. Nov 25 08:09:30 oh yes. i just saw libgraphics_gtk_drawing_area Nov 25 08:11:42 playyaWorking__: btw i don't like to have 2 exactly the same navit launchers, where you can specifi "theme" and "resources" with param and have just one file shipped with navit Nov 25 08:12:20 what's the navitD stuff? Nov 25 08:12:20 playyaWorking__: also 2 new .bb files are not needed.. we should integrate them to navit_svn.bb files with PACKAGES += Nov 25 08:12:34 playyaWorking__: navit with Display requested Nov 25 08:12:41 ok Nov 25 08:12:54 maybe we can make KaZeR to add fso support Nov 25 08:14:43 does QI try to load uImage first? Nov 25 08:14:48 or does it use uImage-GTA02.bin Nov 25 08:15:17 afaik uImage-MACHINE.bin Nov 25 08:15:19 -GTA02.bin iirc Nov 25 08:15:54 k Nov 25 08:16:18 mrmoku's script hardlinks the 2.6.31uImage to /boot/uImage Nov 25 08:16:31 spaetz: that's right.. Nov 25 08:16:40 spaetz: there is link uImage-GTA02.bin -> uImage Nov 25 08:16:43 ahh uImage-GTA02.bin -> uImage Nov 25 08:16:52 had not noticed that :) Nov 25 08:16:55 and uImage is adjusted by opkg when you update kernel etc Nov 25 08:19:50 ooh for fucks sake, the Freerunner can be FAST! Nov 25 08:20:10 Oh man, I did not know this was possible Nov 25 08:20:20 spaetz: Freerunner and fast in one sentence? mad? Nov 25 08:20:23 * spaetz understands now why mrmoku wants to switch asap Nov 25 08:20:37 Defiant-: I say fast, and I MEAN REALLY FAST Nov 25 08:20:46 well, relatively speaking :) Nov 25 08:20:49 spaetz: with 2.6.31? Nov 25 08:20:53 yes Nov 25 08:21:04 budfive_: do you have an image that can suspend? Nov 25 08:21:19 for the first time, scrolling actually deserves its name Nov 25 08:21:52 budfive_: no, I just installed the kernel in the current shrU image. And I haven't tested anything but the UI Nov 25 08:22:09 and there were issues, that's why we haven't switched yet Nov 25 08:22:24 spaetz: I see. :) Interestingly, the old unstable was very fast and responsive with 2.6.31 and x11-16. The new one, not so much. yet. Nov 25 08:22:32 2.6.29 that is Nov 25 08:23:07 budfive_: yep, there is something flawed there Nov 25 08:23:13 and nobody knows why. Nov 25 08:23:40 but the old shrU is nothing compared to the UI responsiveness I am seeing now... Nov 25 08:27:22 spaetz, you booted a 2.6.31. here a pages that wants an update: http://wiki.openmoko.org/wiki/GTA02_sysfs Nov 25 08:29:01 playyaWorking__: but not today. day work deadline... Nov 25 08:29:28 but some stuff seems to have shifted, at least ousaged didn't have a WiFi resource and stuff Nov 25 08:29:40 I am back into 2.6.29 for now Nov 25 08:29:42 you have to adjust your priorities Nov 25 08:29:53 :) that's what my boss tells me Nov 25 08:31:57 weiss: found some interesting patches on codeaurora.org including drm/gem support for the next generation integrated ati gpu and blitting from gem for the mdp chip (the 2d chip) on msm Nov 25 08:32:26 playyaWorking__: and OMG that page is a chaos Nov 25 08:32:34 yes Nov 25 08:32:38 and quite old Nov 25 08:32:43 and don't tell me: "fix it, its a wiki" :) Nov 25 08:35:59 playyaWorking__: navit bug #506 is because of old config, cp15 just said that to me.. Nov 25 08:36:14 ok. Nov 25 08:36:29 playyaWorking__: no more navit.xml shipped as oe patch without proper using of xi:include.. Nov 25 08:36:30 anyone already updated the wiki page? Nov 25 08:36:34 Hi, all! I have question about bitbake recipe. Some developer provide sources in file named program.bz2 which is actually tarball, but bitbake can't extract files correctly without ".tar" part in filename. how to say bitbake that it is tarball, not just bzipped file? Nov 25 08:37:17 parilo: probably #oe knows an answer to that Nov 25 08:37:51 playyaWorking__: no idea.. he just suggested to try new navit.xml (doh I should think about it myself before..) and it works Nov 25 08:38:03 spaetz: I'm searching through their doc... but still no answere Nov 25 08:38:17 ok. it try to update it for OE Nov 25 08:38:29 playyaWorking__: he said that there will be error message instead of crash Nov 25 08:38:39 playyaWorking__: if there is no menu definition in navit.xml Nov 25 08:38:52 playyaWorking__: after this commit I'll bump navit srcrev in oe Nov 25 08:41:38 playyaWorking__: and if you'll play with xml configs.. I haven't checked how exactly xi:include works there, but would be really nice to have standard definitions included from /usr/share config, then separate config for layout, maps, and then merge this 3 includes in navit-bike.xml and navit-car.xml or what ever.. Nov 25 08:43:03 I have been asking for separate .xml files for ages in #navit Nov 25 08:43:21 I wnat to be able to customize the menu but get the vanilla updated style rules for example Nov 25 08:43:40 no success so far. That would really make sense. Than one big monlithic .xml is a pain Nov 25 08:44:04 Can I turn a request for a bittorrent client package in the feed down? Nov 25 08:44:22 FR is not reallysuited to be a file sharing device, is it? Nov 25 08:44:30 or do we have a working recipe? Nov 25 08:44:50 spaetz: file sharing device? Nov 25 08:45:12 lindi-: people are asking me for a bittorrent client, such as "transmission" Nov 25 08:45:22 spaetz: why wouldn't that work? Nov 25 08:45:26 spaetz: and have you tried xi:include? Nov 25 08:45:35 spaetz: i don't think those do anything hardware specific Nov 25 08:45:40 JaMa|Gone: I think I have back then. Nov 25 08:45:46 09:33:16 <@cp15> Ok, the menu structure has been exported recently to navit.xml, so I guess you either have to use an xi:include in your navit.xml or copy data Nov 25 08:45:51 lindi-: not hardware specific. I know it would work in general Nov 25 08:46:02 spaetz: this looks like it should work somehow.. Nov 25 08:46:06 spaetz: ok so make it "SHR is not really suited for ..." :) Nov 25 08:46:12 but is it worth the effort for SHR to maintain a bittorrent client? Nov 25 08:46:35 spaetz: imho maintaining your own distribution is really large task Nov 25 08:46:37 spaetz: I'll try today (xi:include) if I won't be fired from work :) Nov 25 08:46:51 JaMa|Gone: hehe Nov 25 08:47:01 I think it worked. but not sure anymore Nov 25 08:47:11 Its just standard xml... Nov 25 08:52:50 morning Nov 25 08:53:23 morning Nov 25 08:53:34 up early today mickey :) Nov 25 08:53:43 yeah, just came from the airport Nov 25 08:53:55 mrmoku, JaMa: I got one +1 for reverting the navit.launcher script Nov 25 08:54:03 brought someone who's flying to TPE now Nov 25 08:54:05 will do that in OE.dev. Nov 25 08:54:09 ahh :) Nov 25 08:54:23 spaetz, i'll update it later Nov 25 08:54:49 * mickey|office works on org.freesmartphone.Device.Audio.SaveScenario( s ) now Nov 25 08:55:15 it uses pgrep ^navit$ to figure out if navit is already running. in a script called navit ... Nov 25 08:57:08 spaetz: seen that, congratz :) Nov 25 09:01:30 JaMa|Gone: it's nice to get positive feedback for once :) Nov 25 09:01:55 mickey|office: we should really work to get 2.6.31 working Nov 25 09:02:07 have you seen the UI responsiveness on an FR with it? Nov 25 09:02:10 no Nov 25 09:02:18 you all report miracles :) Nov 25 09:02:29 i guess i should try for myself Nov 25 09:02:30 I haven't tried as I thought there would be no miracles Nov 25 09:02:47 but I did not know this is possible with an FR Nov 25 09:03:01 1st time that scrolling actually deserves that name Nov 25 09:03:03 stop teasing, the FR is at home Nov 25 09:03:05 :D Nov 25 09:03:06 hehe Nov 25 09:03:14 mrmoku did a switcher script Nov 25 09:03:22 so it's easy to swtich back and forth Nov 25 09:03:28 good Nov 25 09:03:29 with a regular shrU image Nov 25 09:17:03 can someone fix ffalarms to work directly with fso.deviced please? that stupid atd-over-fso stuff is horrid Nov 25 09:17:14 and it seems to have gotten uninstalled here again. Nov 25 09:20:21 freesmartphone.org: 03mickey 07specs * rdb37cc7bff3f 10/org.freesmartphone.Device/org.freesmartphone.Device.Audio.xml.in: org.freesmartphone.Device.Audio: add SaveScenario(s) -> () Nov 25 09:25:09 TAsn: wake up lazy guy :-). your new API is there... Nov 25 09:27:20 ahh, its the spec only :) Nov 25 09:28:34 freesmartphone.org: 03mickey 07libfso-glib * r7c89d1a0eccb 10/src/ (4 files): regen Nov 25 09:28:36 what's wrong with atd-over-fso? Nov 25 09:29:00 you won't make that vanish, as we need it anyway Nov 25 09:30:32 DocScrutinizer-8: it doesn't get installed. :). And fsodeviced sometimes blocks /dev/rtc0 when anything is scheduled Nov 25 09:31:00 sounds like moot points Nov 25 09:31:05 which maked atd-over-fso fail it seems Nov 25 09:31:12 made Nov 25 09:31:30 if sth is broken, then fix it. rather than discarding it Nov 25 09:32:05 I have enough fixing on my plate already... Nov 25 09:32:06 or, more precisely, rather than discarding systems that base on it Nov 25 09:32:19 but I agree with your attitude in general :) Nov 25 09:32:52 you're licky, almost got me started ;-D Nov 25 09:33:14 hehe Nov 25 09:33:28 spaetz: i'm on fso-apm Nov 25 09:34:00 tasn uttered something with similar notion last night, so I'm still upset ;-P Nov 25 09:34:15 freesmartphone.org: 03mickey 07specs * r58a1210fb69e 10/ (2 files in 2 dirs): update xml for org.freesmartphone.Device.Audio Nov 25 09:34:45 freesmartphone.org: 03mickey 07libfso-glib * r61a7170f1385 10/src/freesmartphone-device.vala: regen Nov 25 09:35:28 spaetz: oh atd-over-fso, we should remone atd from the feed, then there shouldn't be any Problem Nov 25 09:39:02 mickey|office: can we add a RPROVIDES = "apm" to fso-apm? Without that PREFERRED_PROVIDER isn't working... Nov 25 09:40:16 does anyone else get the frameworkd python exception when receiving an incoming call when suspended? Nov 25 09:40:20 http://dpaste.com/124884/plain/ Nov 25 09:40:34 it rings briefly and then stops. Nov 25 09:42:50 Heinervdm: RPROVIDES_${PN} for that should be possible, it should be communicated on the oe mailing list though, since other distros might have to set PREFERRED_PROVIDER_apm = apm Nov 25 09:43:01 ah, scratch that Nov 25 09:43:04 that was wrong Nov 25 09:43:08 Heinervdm: (remove atd) ack Nov 25 09:43:09 the 2nd sentence at least Nov 25 09:44:33 this seems to be the error: KeyError: "Can't register the object-path handler for '/org/freesmartphone/Phon Nov 25 09:44:36 e/GSM/0': there is already a handler" Nov 25 09:45:31 someone needs to fix ophoned Nov 25 09:45:46 it's been broken for long Nov 25 09:45:57 really? Nov 25 09:46:00 uuh Nov 25 09:46:40 any pointers of bug tickets? Nov 25 09:46:45 or bug tickets? Nov 25 09:48:03 mickey|office: ok, then we can apply this patch: http://patchwork.dev.bearstech.com/patch/536/ Nov 25 09:49:57 TAsn: thanks mate, I appreciate that. Nov 25 09:50:16 DocScrutinizer: it's already removed from all future images, because the recipe isn't parsed anymore, but the package was build once and still is in the directory to download Nov 25 09:56:49 freesmartphone.org: 03mickey 07specs * r14de522adbea 10/ (2 files in 2 dirs): org.freesmartphone.Device.Audio.SaveScenario(s) throws FreeSmartphone.Error Nov 25 09:57:12 freesmartphone.org: 03mickey 07libfso-glib * r9a627aa1c33f 10/src/freesmartphone-device.vala: regen Nov 25 09:59:42 mickey|office, Heinervdm: then other distros will get multiple providers for "apm" and maybe fso-am will be picked on non-fso distro, no? or does this happen only for PROVIDES (without R)? last time I tried this patch I got fso-apm in angstrom.. :/ Nov 25 10:01:52 can't we just make ffalarms depend on atd-over-fso? rather than atd? Nov 25 10:02:05 It's the only thing that pulls in atd in our image it seems Nov 25 10:02:11 at least the lite image. Nov 25 10:02:56 spaetz: atd is blacklisted in our tree Nov 25 10:03:13 so it can't be pulled in any image Nov 25 10:03:36 ahh, right. but ffalarms DEPENDS on atd Nov 25 10:03:56 if we made it depend on atd-over-fso everything would be ok Nov 25 10:04:42 spaetz: but with DEPENDS on atd it's useable for other distris Nov 25 10:04:57 I know... :-( Nov 25 10:05:24 I wish BB supported: RDEPENDS: atd|atd-over-fso Nov 25 10:05:42 then we wouldn't have to introduce too many virtual/* packages Nov 25 10:08:26 whats wrong on virtual packages? Nov 25 10:08:59 i think that this is exactly the case why virtuals exist, isn't it? Nov 25 10:09:57 only problem I see is that in OE isn't some file like sane-srcrevs included by all distros, where default providers can be set for all Nov 25 10:10:32 JaMa|Gone: we need to many of them. atd, apm, all kind of basic functionality :) Nov 25 10:10:55 no but there is nothing wrong with it in general. besides adding additional configuration ballast to every other distro Nov 25 10:11:11 JaMa|Gone: yes that would be very useful Nov 25 10:11:24 spaetz: and what's wrong with that? we have multiple implementations for basic functionality stuff Nov 25 10:11:25 to have one common include where default providers for everything are set Nov 25 10:11:40 1:10 < spaetz> no but there is nothing wrong with it in general. besides adding additional configuration ballast to every other distro Nov 25 10:11:51 :) Nov 25 10:11:59 which one common include would solve Nov 25 10:12:12 then we should use already existing mechanisms to select which implementation is right for us.. Nov 25 10:12:28 ok then we should propose sane-providers.inc on oe-devel :) Nov 25 10:12:39 probably sane-providers would make most sense Nov 25 10:12:42 agreed Nov 25 10:15:59 freesmartphone.org: 03mickey 07cornucopia * r18e671904ba5 10/libfsoframework/fsoframework/alsa.vala: libfsoframework: alsa: move BunchOfMixerControls to here Nov 25 10:16:19 freesmartphone.org: 03mickey 07cornucopia * rc61f3950affa 10/fsodeviced/src/plugins/alsa_audio/plugin.vala: fsodeviced: implement org.freesmartphone.Device.Audio.SaveScenario() Nov 25 10:16:58 JaMa: don't confuse RPROVIDES w/ PROVIDES Nov 25 10:17:06 mutliple providers you'll only get w/ PROVIDES Nov 25 10:17:16 bbiab, breakfast Nov 25 10:44:13 spaetz: I see you like it too (kernel) :) Nov 25 10:46:06 mrmoku: all I can say is WOW Nov 25 10:46:15 I never thought the FR can do that Nov 25 10:46:27 I regained a bit confidence, actually :) Nov 25 10:46:44 so let's get that piece working ASAP :) Nov 25 10:47:07 yup Nov 25 10:48:39 mrmoku: seen my mail with regard to ophoned crash on incoming call while suspended? Nov 25 10:48:45 do you have that too? Nov 25 10:48:58 think I have seen something like that in frameworkd.log Nov 25 10:49:03 we should disable it Nov 25 10:49:09 ophoned? Nov 25 10:49:11 yup Nov 25 10:49:13 is it needed? Nov 25 10:49:16 no Nov 25 10:49:29 ahh, then we should disable it. No clue what it is for Nov 25 10:49:35 at least we don't need it Nov 25 10:49:41 all it seems to do is to dispatch calls Nov 25 10:49:42 and I don't think anybody else is using Nov 25 10:49:43 it Nov 25 10:49:46 as it does not work :P Nov 25 10:49:56 and send some signals on call events Nov 25 10:50:08 ok, trying to disable it locally will see what breaks Nov 25 10:50:14 good Nov 25 10:50:25 * mrmoku back to digging into sysfs changes :| Nov 25 10:51:28 mrmoku, please update: http://wiki.openmoko.org/wiki/GTA02_sysfs Nov 25 10:52:26 playyaWorking__: might eventually do that... but before I have to dig into what actually changed :P Nov 25 10:53:17 mrmoku: are you taking an "ls" snapshot and comparing systematically or simply digging manually through stuff? Nov 25 10:53:41 spaetz: did a find /sys for both kernels Nov 25 10:53:47 and diffing them now Nov 25 10:53:53 ok, makes sense Nov 25 10:54:19 though in first place I'm interested in what changed that is relevant for fsodeviced Nov 25 10:55:20 mrmoku: yep Nov 25 11:01:27 mrmoku: maybe very first thing to do was to revert some really insane changes in /sysfs ? Nov 25 11:02:02 1st thing to do is find out what has changed though :) Nov 25 11:03:51 well, I heard vibrator now is called gta02::vibrator (resp. gta01*) for absolutely no good reason, and to my understanding not in accordance with LED-developer mastermind Nov 25 11:06:21 DocScrutinizer-8: but to discuss about what changes are insane and which are not... we have to know the changes first, no? Nov 25 11:06:44 hmm, yes - maybe ;-P Nov 25 11:07:22 which inotify are people using these days? Nov 25 11:07:32 my reasoning about sequence was not to edit wiki prior to that discussion though Nov 25 11:07:40 library, that is Nov 25 11:08:00 mrmoku, mrmoku: disabling ophoned makes the exception go away, but I still have only one ring and then incoming call gets released when suspended... Nov 25 11:08:24 hmm Nov 25 11:08:42 sorry for the ugly linebreaking: frameworkd.log: http://pastey.net/129425 Nov 25 11:09:11 JaMa|Gone: possible that some drivers are missing in 2.6.31 config? Nov 25 11:09:11 and by now I suspend it is because oeventsd seems to get everything three times and gets confused about what action to trigger Nov 25 11:09:44 my logs contain always three idle notifier events for the same thing within a few millisecs for example Nov 25 11:10:01 and also the InComing call signals seem to come multiple times Nov 25 11:10:06 mickey|office: btw. I had no luck in building the dump util from alsa-scenarios Nov 25 11:10:17 ok Nov 25 11:10:23 had autofoo problems with it :| Nov 25 11:10:28 my newest commit should obsolete it Nov 25 11:10:35 oh, in how far? Nov 25 11:10:42 ahh SaveScenario :) Nov 25 11:10:45 yep Nov 25 11:11:03 note that it only existing scenarios are supported though Nov 25 11:11:07 mickey|office: finished already? Nov 25 11:11:08 can't write to an arbitrar file Nov 25 11:11:12 spaetz: I seem to remeber to have seen similar on 0808+ Nov 25 11:11:22 well, i just commited it and it seems to work fine on my laptop Nov 25 11:11:41 pasting 3 lines of frameworkd.log excerpt now: Nov 25 11:11:42 12:02:18.222 oeventsd.fso_triggers INFO Receive CallStatus, status = incoming Nov 25 11:11:45 12:02:19.175 oeventsd.fso_triggers INFO Receive CallStatus, status = incoming Nov 25 11:11:48 12:02:19.201 oeventsd.rule WARNING Untrigger for 'CallListContains(active)' called, but not yet triggered. Not untriggering Nov 25 11:12:17 see how the Receive CallStatus comes in twice within a second and it comes in a third time a bit later again Nov 25 11:13:16 and Receive CallStatus, status = release also comes three times wtihin 10 millisecs Nov 25 11:14:24 anyway, lunch and meeting. BBL Nov 25 11:15:09 and me kindergarten and lunch :) Nov 25 11:15:20 mickey|office: will build and try the SaveScenario thing after lunch Nov 25 11:15:21 bbl Nov 25 11:16:05 oki Nov 25 11:16:07 bon appetit Nov 25 11:16:35 spaetz, mickey|office: INFO Receive IdleState, status=idle two times with a time distance of 15ms Nov 25 11:17:18 without a 'busy' in between? Nov 25 11:17:27 yes Nov 25 11:17:32 d'oh that's very wrong. never seen that Nov 25 11:19:49 status=idle_prelock two times, timegap 18 millisec Nov 25 11:20:05 perhaps it's just printed 2 times? Nov 25 11:20:34 status=lock, imegap 14ms Nov 25 11:21:38 can you check whether this signal comes from the same idlenotifier? Nov 25 11:21:46 perhaps there's multiple active Nov 25 11:23:11 I'm getting every signal 3 times Nov 25 11:23:26 wow, that's clearly broken Nov 25 11:23:40 what's the paramter to tell mdbus to listen on system bus? Nov 25 11:24:09 mdbus -s -l Nov 25 11:24:40 ah that is -s standing for ;) Nov 25 11:25:02 ok Nov 25 11:25:24 mdbus shows the signal once, but it appears in frameworkd.log 3 times Nov 25 11:25:52 mickey|office: you got mail Nov 25 11:26:27 frameworkd.log Nov 25 11:26:49 thanks Nov 25 11:27:05 should contain a inbound call that succeeded as well Nov 25 11:27:22 this is 0808+ Nov 25 11:27:56 only oeventsd.fso_triggers prints the lines 3 times Nov 25 11:29:53 mickey|office: may I sggest all fso subsystems rgister with version/build in frameworkd.log on startup Nov 25 11:30:26 ...or is that there already and I missed it? :-) Nov 25 11:32:47 good point, i'll do that end of this month when I change config files Nov 25 11:33:06 frameworkd.conf to individual files Nov 25 11:33:11 bbiab Nov 25 11:37:22 is there any support for QT4 on SHR? Nov 25 11:38:00 Heinervdm, any updates on xf86-input-tslib? Nov 25 11:38:35 playyaWorking__: no Nov 25 11:38:40 will try to build it Nov 25 11:38:55 and then sent a patch to add it to the feed Nov 25 11:39:12 hackable.1 seems to be a little bit better: if i touch top left corner i hit 4 instead of 7 on the numpad :) Nov 25 11:39:50 :) Nov 25 11:40:08 ok build fine Nov 25 11:42:34 spaetz: can you apply #537 and #538 ? Nov 25 11:44:24 Heinervdm: does transmission build? Nov 25 11:44:30 spaetz: yes Nov 25 11:44:44 will apply 537 Nov 25 11:44:44 rtorrent doesn't build because of libtrorrent Nov 25 11:46:08 spaetz: and 538? Nov 25 11:47:08 xf86-input-tslib does build too, and some users with uncalibrated screens need it to get a calibrated screen Nov 25 11:47:16 * DocScrutinizer muses about which is the most weird app for FR, torrent, openoffice, or gimp Nov 25 11:47:29 Heinervdm: haven't looked at it yet. Nov 25 11:47:34 let me do 537 first Nov 25 11:47:48 torrent might be useful Nov 25 11:47:59 spaetz: i thought you have a script for that ;) Nov 25 11:48:09 but a ooo viewer might be nice Nov 25 11:48:57 iPhone, Pre and Android are having a viewer too Nov 25 11:49:22 playya_: I want a SMP-kernel ;-) and beowulf Nov 25 11:49:57 DocScrutinizer-8: if there is a working recipe for it we will add it :P Nov 25 11:51:36 DocScrutinizer, and support for 8192 cores? Nov 25 11:56:23 4069 should suffice to put the stock in Suzhou to purpose ;-P Nov 25 11:57:14 I have a script. But I need to push to OE.dev and merge into shr/merge then Nov 25 11:57:56 aso :) Nov 25 11:58:07 and navit failed to merge, for example :) Nov 25 11:58:36 what a pity Nov 25 11:59:43 thanks Heinervdm for the patch Nov 25 11:59:43 * DocScrutinizer-8 hopes for results on project tomake a decent FPU from glamo Nov 25 12:00:16 I'm in a meeting will look at 538 later Nov 25 12:00:16 mrmoku|away: yes.. I've added only (sound+ts+mci-already in git repo) +ubi+drm to defconfig Nov 25 12:00:31 JaMa|Gone: g_ether please :) Nov 25 12:00:41 spaetz: then JaMa|Gone can do it ;) Nov 25 12:02:22 spaetz: its included Nov 25 12:02:29 spaetz: it was in defconfig before.. Nov 25 12:02:39 spaetz: I listed only added later Nov 25 12:03:09 538 pushed Nov 25 12:03:29 thx :) Nov 25 12:03:52 mrmoku|away: ping, please rebuild the feed ;) Nov 25 12:03:55 now get someone to build the feed :) Nov 25 12:05:16 we need a second proxy to build the feed Nov 25 12:05:45 everytime mrmoku is |away everybody is lost Nov 25 12:06:05 playyaWorking__: or a cronjob for this Nov 25 12:06:34 * * * 5 rebuild feed Nov 25 12:06:43 :D Nov 25 12:06:59 no, but 2 times a day would be ok ;) Nov 25 12:08:34 or once every German night, when mrmoku sleeps :) Nov 25 12:09:28 and when we he's on the way to pick up his son from the kindergarten Nov 25 12:10:08 hehe, lets put a GPS tracker on mrmoku. whenever he enters kindergarden the feed is rebuild Nov 25 12:10:20 :D Nov 25 12:10:32 9 a.m. 5 p.m. and 0 am Nov 25 12:10:37 3 times a day Nov 25 12:10:52 spaetz, screensaver Nov 25 12:11:08 2009.11.25 12:02:24.227 oeventsd.fso_triggers INFO MrMokuKindergardenStatus = entered Nov 25 12:11:26 hehe Nov 25 12:13:59 actually a GPS trigger would be cool Nov 25 12:21:12 JaMae, already tested zoff's new patch? Nov 25 12:22:27 the huge one? be careful, there are pieces in it that should not go in Nov 25 12:23:13 ok Nov 25 12:23:29 that's what i wanted to say Nov 25 12:23:40 e.g. navit starts twice Nov 25 12:24:01 * DocScrutinizer-8 injects a cron-virus into mrmoku|away's screensaver Nov 25 12:24:39 and it does ugly stuff like export LC_ALL='' Nov 25 12:24:59 etc Nov 25 12:25:05 eeek Nov 25 12:25:14 there is more things I don't like there, then what I like :) Nov 25 12:25:21 as I said before.. Nov 25 12:25:29 and no time.. Nov 25 12:28:02 i don't know whats his problem Nov 25 12:28:21 illume exports de_DE.UTF_8 on my phone Nov 25 12:28:41 he Nov 25 12:28:59 i just added it to my config Nov 25 12:29:01 en_US.UTF-8 here Nov 25 12:29:20 playyaWorking__: probably needs an shr-settings setting :) Nov 25 12:29:27 yes Nov 25 12:29:40 even for both, LC-ALL and LC-MESSAGES o.O Nov 25 12:30:13 wonder what's *that* for Nov 25 12:31:17 spaetz, i always build my images with de lang support. Then i can choose it in the first run wizzard Nov 25 12:31:44 mmh Nov 25 12:31:45 eve LANG and LANGUAGE - wtf?! Nov 25 12:33:09 DocScrutinizer-8: you can consult http://iki.fi/lindi/getenv.png to easily see what environment variables programs use :) Nov 25 12:35:11 hehe, fun Nov 25 12:37:42 lindi-: man 7 locale :: ""LC_ALL All of the above."" So it seems to me LC_MESSAGES is absolute nonsense Nov 25 12:39:08 LANG Provide a default value ... LC_ALL If set to a non-empty string value, override the values of all the other Nov 25 12:39:25 form 'man 1p locale' Nov 25 12:41:38 DocScrutinizer: hmm. maybe LANG is for setting default value to all categories and LC_ALL is for forcing a specific value for all categories? Nov 25 12:43:15 AIUI LC_ALL overrides LC_*, and LANG is even lower precedence than any LC_* Nov 25 12:43:35 on my desktop there's only $LANG Nov 25 12:43:37 yep Nov 25 12:43:43 no LC_* at all Nov 25 12:48:02 well, as I get it exporting LC_ALL='' is the sanest thing you can do to sanitize the mess I found in 0808 Nov 25 12:49:45 basically to me it seems LC_* never should be set in first instance (usually) Nov 25 12:50:58 LANG should suffice Nov 25 12:54:52 you *might* want to set LC_TIME, LC_NUMERIC, and LC_MONETARY to your own preferences Nov 25 12:55:24 I currently only set LC_CTYPE Nov 25 12:56:45 LANG Provide a default value for the internationalization variables that are unset or Nov 25 12:56:46 null. (See the Base Definitions volume of IEEE Std 1003.1-2001, Section 8.2, Nov 25 12:56:48 Internationalization Variables for the precedence of internationalization vari- Nov 25 12:56:49 ables used to determine the values of locale categories.) Nov 25 12:58:05 LC_ALL If set to a non-empty string value, override the values of all the other interna- Nov 25 12:58:07 tionalization variables. Nov 25 12:58:39 so quite obviously the setup found in 0808+ is fishy Nov 25 13:01:15 DocScrutinizer: only setting in the new shrU is: LANG=en_US.UTF-8 Nov 25 13:01:21 no LC_* settings at all Nov 25 13:01:47 spaetz: great :-) Nov 25 13:04:22 mrmoku|away: disabling ophoned, I can make and receive calls/USSD just fine. So I propose to disable it by default Nov 25 13:04:55 just fine = ringing stops after first ring when the phone was suspended Nov 25 13:07:01 that's not exactly "just fine", no? :-/ Nov 25 13:07:07 * mickey|office announces a vala update Nov 25 13:07:19 I uploaded 0.7.8.2 to www.fso Nov 25 13:07:26 mickey|office: found a bug? :-D Nov 25 13:07:38 not really, but some additions to bindings Nov 25 13:07:47 aah Nov 25 13:07:50 i prefer to do that upstream rather than maintaining a seperate branch Nov 25 13:08:10 (not that there wouldn't be hundreds of bugs in Vala...) Nov 25 13:08:14 but those i can't fix Nov 25 13:09:36 mickey|office: do you have a debian package for libeflvala? Nov 25 13:11:20 freesmartphone.org: 03mickey 07cornucopia * r650af1111f0d 10/libfsoframework/ (5 files in 3 dirs): libfsoframework: create INotifier dispatcher object. NOTE: We need Vala >= 0.7.8.2 now Nov 25 13:11:39 now fsotimed is compilable again :) Nov 25 13:13:49 Heinervdm: i'm afraid not. never created a debian package. how does that work? Nov 25 13:14:08 mickey|office: do you use debian? Nov 25 13:14:25 Heinervdm: ubuntu karmic Nov 25 13:15:07 ~curse buntkuh Nov 25 13:15:08 May the fleas of a thousand camels infest your most sensitive regions, buntkuh ! Nov 25 13:15:37 mickey|office: then you have to name the directory like ${PN}-${PV} and run dh_make to create the debian directory, then you have to tweak these files and then run dpkg-buildpackage Nov 25 13:16:13 mickey|office: but i have no idea how to tweak the files... Nov 25 13:16:24 well, i don#t think i should go into the packaging business Nov 25 13:16:28 can't our pkg-fso folks do that? Nov 25 13:16:29 hehe Nov 25 13:16:48 i always steal the debian/ dir from other packages and twaek them Nov 25 13:17:18 well probably they can Nov 25 13:18:29 when there is a package for libeflvala i will begin to look into vala and efl Nov 25 13:20:10 freesmartphone.org: 03mickey 07cornucopia * r44ceb4c8612a 10/libfsoframework/configure.ac: libfsoframework: post release version bump Nov 25 13:21:01 excellent Nov 25 13:23:04 mickey|office: i'd like to have my phone go immediately back suspend if it receives a gprs packet. any idea what's the most sane way to do this? (it should be done fast to save energy) Nov 25 13:23:06 hmm what lib contains ecore_main_loop_glib_integrate Nov 25 13:24:30 lindi-: seems we discussed that in the past? Nov 25 13:24:40 DocScrutinizer-8: to some degree Nov 25 13:24:44 Heinervdm, do you have any up-to-date e packages for ubuntu? Nov 25 13:25:15 playyaWorking__: i'm using the e packages from enlightment, and i will never use ubuntu :P Nov 25 13:25:42 ok. the ubuntu ones seem to be outdated Nov 25 13:25:56 http://packages.enlightenment.org/debian Nov 25 13:26:03 maybe i write a script to compile the svn versions Nov 25 13:26:14 most of the already have a debian/ Nov 25 13:27:29 lindi-: on my 0808+ fso sets state to idle on modem-irq, so if I set idle-dim, idle-lock, and suspend to very short values, it suspends in max 4sec after inbound gprs packet Nov 25 13:27:40 hmm but the packages are quiet old, from august Nov 25 13:29:55 DocScrutinizer-8: is that enough time to receive SMS? Nov 25 13:30:09 I's guess yes Nov 25 13:30:43 still, i'd like to make it 0.5 s Nov 25 13:31:17 (that's about what I get with usb disconnect event currently) Nov 25 13:31:36 but tbh with a scheme like that any app that needs to do some processing should allocate CPU resource, then do it's foo, and then release CPU, to avoid premature suspend Nov 25 13:32:39 though maybe a nice -15 for sms-app / irc-app etc might somewhat help Nov 25 13:32:59 DocScrutinizer-8: i'm hoping the muxer could signal immediately what channels received data Nov 25 13:33:12 and based on that decision I could go immediately back to sleep or continue Nov 25 13:33:47 you could do that, but won't match the concept of FSO Nov 25 13:33:56 DocScrutinizer-8: but dbus is a shared resource with no way to set priorities so nice -15 is only of marginal help Nov 25 13:33:58 so you for sure need to patch Nov 25 13:34:07 yep Nov 25 13:38:08 lindi-: the more nasty problem is to stop senseless suspends during e.g. download of a huge webpage Nov 25 13:40:33 mmh, redirecting mdbus -s -l to a file doesn't seem to work Nov 25 13:41:52 mmh, caching issue? Nov 25 13:41:55 spaetz_: 2>&1 ? Nov 25 13:42:03 yep, doing that Nov 25 13:42:19 probably it's just not written out Nov 25 13:42:29 due to delayed writing Nov 25 13:42:41 yes Nov 25 13:43:31 seen strange things like "foo | cat >xxx" Nov 25 13:43:42 hehe Nov 25 13:44:25 freesmartphone.org: 03mickey 07cornucopia * r690dc75efcc8 10/fsodeviced/ (configure.ac src/plugins/alsa_audio/plugin.vala): (log message trimmed) Nov 25 13:44:25 freesmartphone.org: fsodeviced: alsa_audio: use inotify to watch for scenario changes Nov 25 13:44:25 freesmartphone.org: Reload the scenario cache (for the modified scenario only) whenever Nov 25 13:44:25 freesmartphone.org: a change is being reported. This also works, if the changed scenario Nov 25 13:44:26 freesmartphone.org: is the current scenario. If the scenario being modified is not the Nov 25 13:44:28 freesmartphone.org: current scenario, be aware that audible artefacts can occur during Nov 25 13:44:30 freesmartphone.org: reloading. Nov 25 13:44:38 ok, this should be what TAsn wants Nov 25 13:44:49 awaiting your feedback Nov 25 13:46:13 wtf??? Nov 25 13:48:11 mickey|office: nice. BTW with respect to multiple signals and non-working incoming calls when suspended Nov 25 13:48:27 I looged all dbus system signals and there are 2 incoming call signals Nov 25 13:48:48 [SIGNAL] org.freesmartphone.GSM.Call.CallStatus from :1.17 /org/freesmartphone/GSM/Device Nov 25 13:48:51 { 'direction': 'incoming', Nov 25 13:48:58 this is actually send twice over dbus Nov 25 13:49:08 so it is not just a logging glitch Nov 25 13:49:10 yep Nov 25 13:49:12 see http://git.freesmartphone.org/?p=specs.git;a=blob_plain;f=html/org.freesmartphone.GSM.Call.html;hb=HEAD#CallStatus Nov 25 13:49:21 it might come dozens of time Nov 25 13:49:27 ahh Nov 25 13:49:28 until you pick up the call Nov 25 13:49:40 k. let me continuue reading the log then :) Nov 25 13:49:53 but this behaviour hasn't changed since months Nov 25 13:49:59 so i'm really wondering what goes wrong now Nov 25 13:50:04 now TAsn want's to write his mute crap into paermanent storage??? Nov 25 13:50:19 mickey|office: I tried buidling current fsodeviced... but it fails Nov 25 13:50:27 in do_install Nov 25 13:50:53 the sbin change=? Nov 25 13:51:24 DocScrutinizer-8: don't be that pessimistic ;) Nov 25 13:51:28 builds fine here Nov 25 13:51:32 pastebin? Nov 25 13:51:37 he wants it for persistant volume changes :) Nov 25 13:51:39 mickey|office: moment Nov 25 13:51:51 mrmoku: I'm realist Nov 25 13:52:07 mmmh, dbus log gives no hint: 4 lines of log: Nov 25 13:52:12 [SIGNAL] org.freesmartphone.Usage.ResourceChanged from :1.4 /org/freesmartphone/Usage('GSM', True, { 'policy': 0, 'refcount': 1}) Nov 25 13:52:15 [SIGNAL] org.freesmartphone.Usage.ResourceChanged from :1.4 /org/freesmartphone/Usage('Display', True, { 'policy': 0, 'refcount': 1}) Nov 25 13:52:18 [SIGNAL] org.freesmartphone.GSM.Call.CallStatus from :1.17 /org/freesmartphone/GSM/Device (1, 'release', { 'status': 'release'}) Nov 25 13:52:21 [SIGNAL] org.freesmartphone.PIM.Calls.NewCall from :1.8 /org/freesmartphone/PIM/Calls Nov 25 13:52:24 ('/org/freesmartphone/PIM/Calls/10',) Nov 25 13:52:26 it just releases the call without any intervention... Nov 25 13:52:45 perhaps it goes to the mailbox? Nov 25 13:52:56 err, voicebox, that is Nov 25 13:53:02 mickey|office: http://build.shr-project.org/tests/mrmoku/log.do_install.23370 Nov 25 13:53:03 mrmoku: seen mickeys commit? it needs newer vala Nov 25 13:53:15 hmm, no Nov 25 13:53:23 mickey|office: no, it rings once. and I have the voicebox compeletely disabled Nov 25 13:53:45 mrmoku: looks good, where's the error? Nov 25 13:53:58 no idea Nov 25 13:54:03 it bails out there Nov 25 13:54:07 libfsoframework: create INotifier dispatcher object. NOTE: We need Vala >= 0.7.8.2 now Nov 25 13:54:08 mrmoku: aiui tasn manages to adjust alsa controls by himself, but has to write back the changes in a way we did with "alsactl -f foo STORE" formerly. The new finctions of fsodeviced seem to be more like complementary to this Nov 25 13:54:28 ERROR: Task 8 (/home/mok/src/other/openmoko/shrbuild/shr-unstable/openembedded/recipes/freesmartphone/fsodeviced_git.bb, do_install) failed Nov 25 13:54:50 hmm, let me build it on my buildhost Nov 25 13:55:05 mrmoku: got your permission to disable ophoned in our frameworkd.conf? Nov 25 13:55:20 I made sure making and recieving calls/USSDs is not worse than now... Nov 25 13:55:36 JaMa|Gone: ar6000 is in kernel (not module) by defconfig Nov 25 13:55:40 spaetz_: yep Nov 25 13:55:59 mickey|office: do you like rfkill? :P Nov 25 13:56:26 mrmoku: not much, t.b.h. But that could be due to the way I was introduced to it (/me remembers lengthy heated discussions on openmoko-kernel...) Nov 25 13:56:34 i'm open to change my mind Nov 25 13:56:53 might be the best thing since bread came sliced Nov 25 13:56:57 I can't fine the -spi thing to power wifi... but there is a rfkill interface now Nov 25 13:57:01 s/fine/find/ Nov 25 13:57:01 mrmoku meant: I can't find the -spi thing to power wifi... but there is a rfkill interface now Nov 25 13:57:08 SHR: 03Sebastian 07shr-themes * r4c5166ae3dd5 10/frameworkd/frameworkd-config-shr/om-gta02/frameworkd.conf: frameworkd.conf: turn off ophoned. It's just crashing on incoming calls anyway. Nov 25 13:57:08 you really loving .31, right? Nov 25 13:57:12 :) Nov 25 13:57:22 it's really fast :) Nov 25 13:57:26 * mickey|office will try when he comes back home Nov 25 13:57:34 I'm loving it Nov 25 13:57:54 who needs phone when scrolling is that smooth, right? :P Nov 25 13:58:10 mrmoku: is it question? Nov 25 13:58:10 * mickey|office hears the McD* jingle in his head Nov 25 13:58:20 JaMa|Gone: ar6000? Nov 25 13:58:24 hehe. BTW; they just changed their logo color in Europe Nov 25 13:58:28 no question... noticed that Nov 25 13:58:28 Yellow on Green Nov 25 13:58:36 really? Nov 25 13:58:40 yep Nov 25 13:58:40 mrmoku: yes Nov 25 13:58:44 SHR root@gojama ~ $ zcat /proc/config.gz | grep 6000 Nov 25 13:58:44 CONFIG_AR6000_WLAN=y Nov 25 13:58:48 that's dumb. changing an established logo is almost always dumb Nov 25 13:58:56 brand recognition is one of the most important values Nov 25 13:59:00 *shrug* Nov 25 13:59:08 "It reflects our increased quality standards, greenness, blah blah" Nov 25 13:59:18 US stays yellow-on-red Nov 25 13:59:49 mickey|office: another thing I can't find is usb host mode :| Nov 25 14:00:29 aah, sysfs pretty fucked up :-/ Nov 25 14:01:50 perhaps it's not exported yet Nov 25 14:01:54 larsc should know Nov 25 14:02:13 CONFIG_S3C_DEV_USB_HOST=y Nov 25 14:02:43 who needs host mode when you can have fast scrolling ;) Nov 25 14:02:43 SCNR Nov 25 14:02:45 larsc: do you know by heart how to turn on usb host mode with 2.6.31? Nov 25 14:02:49 seems we're getting closer to a working image every day Nov 25 14:03:22 who needs SHR if you can have maemo? Nov 25 14:03:33 who wants maemo in the first place ;) Nov 25 14:03:59 DocScrutinizer-8: feel free to port maemo to the FR and I might switch to your distro. Until then I stick with SHR ;P Nov 25 14:04:16 spaetz: my feeling would be it boots a lot faster too... have to measure that though Nov 25 14:04:20 well, I prefer closed source drivers to missing sysfs nodes Nov 25 14:05:09 cool, boot in zero time to a totally borked system :-/ Nov 25 14:05:27 nah, did not say anything about zero time :P Nov 25 14:05:53 well, you could argue a system with that degree of stability really needs fast boot Nov 25 14:06:03 DocScrutinizer-8: well, I don't Nov 25 14:07:49 mickey|office: btw. I'm doing a local build... dos the buildsystem need anything outside fsodeviced? Nov 25 14:07:53 maybe that is my problem.. Nov 25 14:08:40 no, that should be self contained Nov 25 14:08:49 S is directly pointing to fsodeviced as well Nov 25 14:08:54 ok Nov 25 14:12:01 mrmoku: I timed boot Nov 25 14:12:03 0:00 WSOD 0:30 boot screen goes away 0:48 e appears 0:53 desktop icons done 1:40 SIM PIN Nov 25 14:12:34 spaetz: non-KMS? Nov 25 14:12:34 but that might well be an effect that there are no modules to evaluate. So we might get similar stuff if we remove the 1,000s of modules we have by default Nov 25 14:12:58 Weiss: mrmoku/test/2.6.31 whatever that is. Let me grep kernel log Nov 25 14:13:46 KMS without fbcon enabled would probably give that effect Nov 25 14:14:05 (nothing to provide anything to display until X comes along) Nov 25 14:14:51 Weiss: grep -i kms doesn't show anything in kernel log Nov 25 14:15:09 spaetz: you ned to grep for "DRM" :) Nov 25 14:15:24 tons of [glamo-drm] Falied to allocate object Nov 25 14:15:30 eek, that's bad Nov 25 14:15:40 could you post the log somewhere? Nov 25 14:17:37 Weiss: will do, but it seems the log buffer was too small so I think the start is missing Nov 25 14:18:01 Weiss: but yes its 2.6.31+your patches.. Nov 25 14:18:26 Weiss: with GLAMO_DRM enabled of course :) Nov 25 14:18:42 logread (busybox's dmesg buffer) starts with those [glamo-drm] Failed to allocate object messages Nov 25 14:19:01 so all the previous log has already been discarded :-( Nov 25 14:19:40 someone needs to increase the buffer size and reboot. No time right now Nov 25 14:19:42 sorry Nov 25 14:19:43 spaetz: hmm.. does "dmesg -s 1000000" help? Nov 25 14:19:52 Weiss: with KMS Nov 25 14:20:11 spaetz: how about using klogd? Nov 25 14:20:20 odd.. this worked the last time I tested it :S Nov 25 14:21:18 lindi-: if I had time now to install klogd yes... Nov 25 14:21:25 :) Nov 25 14:21:54 spaetz: you just write the commands to test it and mail the results. then you put the phone to your pocket and let the phone do the tests on its own :P Nov 25 14:22:26 Weiss: I don't have any "Falied to allocate object" on same kernel, but built here.. Nov 25 14:22:29 Weiss, spaetzWrk: I don't have those errors Nov 25 14:22:39 :) Nov 25 14:22:39 same config? Nov 25 14:22:51 Weiss: same kernel build as spaetzWrk Nov 25 14:23:04 interesting Nov 25 14:23:16 mrmoku: as in, same binary? Nov 25 14:23:19 yup Nov 25 14:23:45 spaetzWrk: those messsages are definitely after a fresh boot? Nov 25 14:24:00 cleanup of old objects doesn't really work in many circumstances Nov 25 14:26:37 spaetzWrk: I guess there's no chance of seeing your /proc/dri/0/gem_objects? Nov 25 14:26:50 (it can wait if you're busy - I'm working too) Nov 25 14:52:04 back Nov 25 14:52:10 rebooting. will let you know Nov 25 14:56:20 Weiss: I don't get those messages until startup completed. Nov 25 14:56:59 The last normal entry is: kernel: glamo-fb glamo-fb.0: glamo fb_cmd_mode(on=0) Nov 25 14:57:36 and then the SIM PIN dialog pops up (using e-wm gry theme here). Not sure if it's related, but that's the only graphic thing that had happened Nov 25 14:58:19 and then/now I get the [glamo-drm] Failed to allocate object error Nov 25 14:58:51 it occurs at least dozens of times per second Nov 25 14:59:01 the sim pin dialog is still up. let me close that Nov 25 14:59:35 closed the dialog, but the error messages still occur Nov 25 14:59:39 lots of time per second Nov 25 15:01:09 /proc/dri/o/gem_objects: Nov 25 15:01:35 231 objects 8183808 object bytes (the rest of the rows all 0) Nov 25 15:02:12 so... somehow you ran out of VRAM :S Nov 25 15:02:23 no swap file, if that is of any intereste :) Nov 25 15:02:38 nope, the VRAM allocation is very naive and simple Nov 25 15:02:51 but if you run out, X will fall back to software rendering Nov 25 15:02:59 (for that pixmap) Nov 25 15:03:51 normally I don't get anywhere near running out :S Nov 25 15:04:13 mmh, this is a standard shrU, with the gry theme and the SIM PIN dialog popping up Nov 25 15:04:36 and the vala-terminal thingie to get console output (usb networking doesn't work) Nov 25 15:05:32 that should be OK Nov 25 15:05:54 outch. tried suspending WSOD ... Nov 25 15:06:43 not surprised :) Nov 25 15:06:53 heh Nov 25 15:06:55 complete and total banishment of WSoD forever is this weekend's job Nov 25 15:07:05 yay :) Nov 25 15:07:08 go for it :) Nov 25 15:07:25 (I like to set myself difficult challenges :) Nov 25 15:08:36 mmh, rebooted, SIM PIN doalog popped up, no errors so far Nov 25 15:08:46 let me try to enter my PIN Nov 25 15:09:14 no, just switching between my windows caused it to start Nov 25 15:17:26 we should have some safe patches to improve perf in 2.5.29 until wer upgrade :) Nov 25 15:17:33 eg. that So I turned the FIFO up to "8 stages" (from 1) to see what happened Nov 25 15:17:38 looks interesting :) Nov 25 15:17:49 * spaetzWrk reads Weiss' blog Nov 25 15:18:17 how do you tweak that? :) Nov 25 15:23:52 my first mail on qi-hardware maillist is about internet explorer 6... Nov 25 15:23:56 isn't is strange? ;P Nov 25 15:28:55 too much birtday party, i'd say ;) Nov 25 15:40:17 spaetzWrk: http://git.bitwiz.org.uk/?p=kernel.git;a=commitdiff;h=01e4e0c2a1c39d7e4556f8996f21dfaf73b2869c - that 0x4000 Nov 25 15:41:00 Weiss: thnks. might try over the weekend Nov 25 15:41:45 0x0000 is 1 stage, 0x4000 is two stages, 0x8000 is four stages, 0xC000 is eight stages Nov 25 15:41:50 but at 8 stages it's really not pretty.. Nov 25 15:41:53 four stages is OK Nov 25 15:42:23 hehe /me scared a bit of being fired today, because of working on other issues than todays project deadline, just signed salary increase :) Nov 25 15:43:04 :) Nov 25 15:43:07 :D Nov 25 15:57:50 DocScrutinizer-8, not mute crap Nov 25 15:57:55 DocScrutinizer-8, volume changes Nov 25 15:58:09 I wrote the script we talked about that will change the scenarios Nov 25 15:58:17 mickey|zzZZzz, thanks a lot :P just saw the commit. Nov 25 15:58:26 mrmoku, mind rebuild fso_deviced? Nov 25 15:59:20 spaetzWrk, what were you talking about? TAsn: wake up lazy guy :-). your new API is there... Nov 25 15:59:53 the SaveSoundAPI Nov 25 15:59:53 :) Nov 25 16:00:01 by mickey. Nov 25 16:00:06 spaetzWrk, yeah. Nov 25 16:00:10 saw it :) Nov 25 16:00:13 thanks. Nov 25 16:00:16 don't bump libfsoframework too far though Nov 25 16:00:18 NOTE: We need Vala >= 0.7.8.2 now Nov 25 16:00:28 650af1111f0d3dc2ea2fa98b41d96a8100ba4ddd Nov 25 16:00:38 you need to bump it to master Nov 25 16:00:42 you also want the inotify stuff Nov 25 16:00:51 otherwise there's no way to reload scenario files Nov 25 16:00:57 (except restarting) Nov 25 16:00:59 sybren, I appreciate your bug reporting :P Nov 25 16:01:16 yep. Nov 25 16:01:32 mickeyl, well, if I have the SaveSoundAPI Nov 25 16:01:45 I can get around the inotify need Nov 25 16:01:55 but still, it's best just to go to master :P Nov 25 16:02:05 mrmoku, please get latest fsodeviced and change config to use aplayer plugin. :) Nov 25 16:02:41 TAsn: stil Nov 25 16:02:44 oops Nov 25 16:02:54 inconming calls while suspended still fail though Nov 25 16:03:09 it rings once and then it gives up Nov 25 16:03:15 I failed to debug it Nov 25 16:04:46 is there a way to prevent SHR from locking the screen while in a call? Nov 25 16:06:09 ask, dos1 and he will tell you it's by design. :) Nov 25 16:06:20 ask me and I call it a horrible bug... Nov 25 16:06:52 lol Nov 25 16:06:59 I'll ask you then Nov 25 16:07:12 I'd also like it to unlock upon an incoming call Nov 25 16:07:30 it already takes enough time to figure out someone's calling me, get my phone from my pocket, and click an accept button Nov 25 16:08:02 not sure about the unlocking thing. I don't want the FR to automatically accept or reject calls when I carry it around in my pocket Nov 25 16:08:28 The real solution is to make shr-today really quick here, or to present an "accept call slider" :) Nov 25 16:08:29 sybren, are you sure you want this? Nov 25 16:08:41 spaetzWrk, it'll be really quick here Nov 25 16:08:50 accept call slider: please don't :P Nov 25 16:09:13 TAsn: why not? I'd like it Nov 25 16:10:06 sybren, Will make things a bit more complicated for us to handle Nov 25 16:10:16 ATM we assume we are the only one handling calls Nov 25 16:10:16 ah you lazy bastard :P Nov 25 16:10:19 a bit naive, but still. Nov 25 16:10:34 well shr-today is soon going to be "we", right :-) Nov 25 16:10:41 so "we" can handle it :) Nov 25 16:11:11 * mickeyl votes for an accept call slider Nov 25 16:11:13 well, things like volume management would be more important to me than a fast shr-today Nov 25 16:11:17 but don't forget a reject call button Nov 25 16:11:17 +1 Nov 25 16:11:24 iphone forgot that... Nov 25 16:11:35 if locked, then you can not reject an incoming call on iphone Nov 25 16:11:39 * mickeyl chuckles Nov 25 16:11:53 iphone users are not supposed to reject calls :) Nov 25 16:12:04 it gives the mobile phone operators a slice of the profit :) Nov 25 16:12:11 lol Nov 25 16:12:29 hahaha Nov 25 16:12:30 righto Nov 25 16:12:32 "slide clockwise to accept, anticlockwise to reject" Nov 25 16:12:45 an evil Steve Jobs plan with the mobile phone operators Nov 25 16:12:51 spaetzWrk, that's actually true. Nov 25 16:13:08 sybren, volume management works... Nov 25 16:13:19 sybren: awesome idea Nov 25 16:14:00 mickeyl: did it build for you? Nov 25 16:14:02 mickeyl, got any idea what may have caused new shr to be so amazingly slow? Nov 25 16:14:09 connecting to ssh takes ages Nov 25 16:14:30 TAsn: including storing the volume so that it's reused on the next call? Nov 25 16:14:30 mrmoku: yes Nov 25 16:14:32 hmm Nov 25 16:14:33 sybren, yes. Nov 25 16:14:35 mickeyl: thanks :) Nov 25 16:14:37 TAsn: no idea, what's htop saying? Nov 25 16:14:40 TAsn: great! Nov 25 16:14:42 (at the moment mrmoku updates fsodeviced) Nov 25 16:14:46 * mrmoku bumps rev and tries on buildhost... Nov 25 16:14:46 TAsn: waht's powertop saying? Nov 25 16:14:49 mickeyl, that's the thing, nothing! Nov 25 16:14:51 (top) Nov 25 16:14:56 powertop say 900 wakeups/sec Nov 25 16:15:00 powertop, sec. Nov 25 16:15:00 mickeyl: just fsodeviced, right? Nov 25 16:15:04 mickeyl, ^ Nov 25 16:15:12 400 from accel even if nobody ready them... Nov 25 16:15:16 mrmoku: vala,libfsoframework Nov 25 16:15:30 that would be your slowdown Nov 25 16:15:30 and fso-specs Nov 25 16:15:36 mrmoku: and libfso-glib ;) Nov 25 16:15:40 :) Nov 25 16:15:51 just rebuild or bump rev? Nov 25 16:15:51 actually its lower now Nov 25 16:15:56 mickeyl, does libfso glib support both async and sync ? Nov 25 16:16:11 205.9 wakeups here. it finally seems to work Nov 25 16:16:24 and those are S3C2410 Timer Tick Nov 25 16:16:43 mrmoku, have you started using the callback to update volume bars in ui-shr? Nov 25 16:17:04 mrmoku, and btw, don't forget to update fsodeviced config to use alsa player instead of libcanberra Nov 25 16:17:19 60.9% (409.2) : lis302dl Nov 25 16:17:19 TAsn: no... I cleaned our garage and braught the old cruft to the waste disposal :P Nov 25 16:17:26 TAsn: unfortunately only async. if you want to call sync, use dynamic DBus Nov 25 16:17:27 ... Nov 25 16:17:28 (from Vala) Nov 25 16:17:39 mickeyl, async is good enough ;) Nov 25 16:17:42 Wakeups-from-idle per second : 660.2 Nov 25 16:17:56 207 here Nov 25 16:18:01 i'd say the kernel is broken then Nov 25 16:18:05 dos1: lis302 ? Nov 25 16:18:22 mickeyl, that's what I think Nov 25 16:18:23 mrmoku: rebuidling should do Nov 25 16:18:27 I think it's because of the modules Nov 25 16:18:30 not being stripped Nov 25 16:18:35 mrmoku: well, to master, of course Nov 25 16:18:36 (kernel modules) Nov 25 16:18:47 TAsn: stripping shouldn't affect the number of wakeups Nov 25 16:18:49 something is wrong with accels Nov 25 16:19:34 is Qi honoring the "boot flag" of partitions or will it not care Nov 25 16:19:56 spaetzWrk, not number of wakeups, but at least number of caching and rechaching of the modules into the processor's cache. Nov 25 16:19:59 mickeyl: it does not care... at least I never have set it and it boots fine Nov 25 16:20:06 mrmoku: ok, thanks. trying a new SD now Nov 25 16:20:18 Wakeups-from-idle per second : 612.8 Nov 25 16:20:33 64.7% (397.4) : lis302dl Nov 25 16:20:33 32.7% (200.5) : S3C2410 Timer Tick Nov 25 16:20:41 nasty Nov 25 16:20:43 Suggestion: Enable the CONFIG_NO_HZ kernel configuration option. Nov 25 16:20:43 This option is required to get any kind of longer sleep times in the CPU. Nov 25 16:20:48 that's for sure accounting for a great slowdown Nov 25 16:21:02 Suggestion: Enable the CONFIG_HPET_TIMER kernel configuration option. Nov 25 16:21:02 Without HPET support the kernel needs to wake up every 20 milliseconds for Nov 25 16:21:02 some housekeeping tasks. Nov 25 16:21:06 hm... Nov 25 16:21:10 ignore the hints Nov 25 16:21:16 :P Nov 25 16:21:16 those are x86isms by large Nov 25 16:21:22 TAsn: I had that too, although weirdly enough not now. Nov 25 16:21:34 accel producing 400 interrups/sec Nov 25 16:21:53 so... what are we going to do? Nov 25 16:22:05 mickeyl, isn't there a way to turn off accels? Nov 25 16:22:22 TAsn: yup, there is Nov 25 16:22:22 just tested Nov 25 16:22:27 or at least make the kernel ignore when no one is reading from /dev ? Nov 25 16:22:27 cat /dev/input/event2 Nov 25 16:22:27 and then ctrl+c Nov 25 16:22:48 and then amount of interups falls down to 200 Nov 25 16:22:48 and then Nov 25 16:22:52 cat /dev/input/event3 Nov 25 16:22:56 and then ctrl+c Nov 25 16:23:00 and amount of interups falls down to 0 Nov 25 16:23:08 so there is kernel bug :P Nov 25 16:23:24 yeah, 0 here as well. Nov 25 16:23:32 97.0% (200.4) : S3C2410 Timer Tick Nov 25 16:23:36 what's that though? Nov 25 16:23:44 TAsn: no, the kernel is supposed to disable them when noone reads Nov 25 16:23:53 i don't know any further control means from userland Nov 25 16:24:03 ah Nov 25 16:24:07 so the initial irq state is wrong Nov 25 16:24:22 should be easy to fix Nov 25 16:24:40 is this with current andy-tracking, or 2.6.31? Nov 25 16:24:46 andy-tracking Nov 25 16:24:48 mickeyl, does the default irq state revert on resume? Nov 25 16:24:56 ah, I thought I noticed my battery life being down a bit.. Nov 25 16:25:02 suspend/resume Nov 25 16:25:16 ok Nov 25 16:25:24 this is the main reason for the slow down Nov 25 16:25:27 phone is now almost reasonable :) Nov 25 16:25:46 after suspend/resume accels are generating interrups again Nov 25 16:25:56 TAsn: that's beyond my kernel knowledge Nov 25 16:26:03 it should not though Nov 25 16:26:10 since if you have it open on suspend Nov 25 16:26:15 TAsn, mickeyl: ^^^^^ Nov 25 16:26:16 it should still give you data on resume Nov 25 16:26:26 dos1, yeah Nov 25 16:26:29 just tested myself Nov 25 16:26:31 goes back to default :| Nov 25 16:26:45 ok Nov 25 16:26:46 there has been just one accell commit since long iirc Nov 25 16:26:53 that might have introduced the regression Nov 25 16:27:02 so who we should spank for that? :P Nov 25 16:27:05 git blame.. Nov 25 16:27:39 awesome... with this SD Qi boots Nov 25 16:27:50 funny enough the one that's not working was the one in the FR box Nov 25 16:27:52 *chuckle* Nov 25 16:28:52 nelson is offline on jabber :x Nov 25 16:28:59 PaulFertser: maybe you can help us? Nov 25 16:29:33 I have 200 interrupts on s3c-mci, 200 on "S3C2410 Timer Tick" and 100 on "ksdioirqd/mmc1 : schedule_timeout (process_timeout)", is that normal? Nov 25 16:29:55 and I get "Suggestion: Enable the CONFIG_NO_HZ kernel configuration option." from powertop Nov 25 16:30:09 unfortunately balrog dropped the ball on NO_HZ Nov 25 16:30:10 it once worked :( Nov 25 16:30:15 and made a great difference Nov 25 16:30:19 it did? Nov 25 16:30:19 ahh Nov 25 16:30:23 would have been nice Nov 25 16:30:42 and CONFIG_HPET_TIMER? Nov 25 16:30:49 never heard about that one Nov 25 16:30:51 must be a new thing Nov 25 16:31:21 sybren: ignore those suggestions Nov 25 16:31:22 apparently - according to powertop - without it the kernel will wake up every 20ms to do some stuff Nov 25 16:31:34 sybren: it can be x86-only Nov 25 16:31:39 ok, fair enough Nov 25 16:31:57 hmm Nov 25 16:31:59 but those interrupt counts I get - are they normal? Nov 25 16:32:02 we'd be happy to only wake up every 20ms Nov 25 16:32:08 why did you configure SHR to not dim Nov 25 16:32:15 but rather go off immediately? Nov 25 16:32:20 200wakeups/sec are 5ms anyway Nov 25 16:32:21 sybren: yup Nov 25 16:32:42 mickeyl: well, there is no way to that sanely now in oeventsd Nov 25 16:32:45 mickeyl: historical raisins from when we didn't have shr-today Nov 25 16:32:49 dos1: that's good to know, then I have something to compare against when I want to check out powertop again ;-) Nov 25 16:33:14 hmm, too bad, i always thought that adds class Nov 25 16:33:22 we can fix it again with a new oeventsd Nov 25 16:33:40 dimming would be nice Nov 25 16:33:47 mickeyl: as now in oeventsd it's always "overwriting" (couldn't find better word :P) brightness value Nov 25 16:33:48 and suspend rather than blank Nov 25 16:33:58 so when i have backlight on 10% Nov 25 16:34:12 although i want to blank when I prevent suspend while eg taking tracks with tangogps Nov 25 16:34:22 oeventsd will change it to, say, 50%. and then, on busy state, back to 100% Nov 25 16:34:39 so it's ignoring previous settings Nov 25 16:34:58 dos1: can't we save brightness level with opreferences Nov 25 16:35:05 and pull it up again somehow? Nov 25 16:35:21 spaetzWrk: yup, but it still has to be implemented Nov 25 16:35:46 mrmoku, any news about fsodeviced/registering to volume change callback? :P Nov 25 16:35:48 as current oeventsd doesn't allow you to do such advanced things :P Nov 25 16:40:01 TAsn: fsodeviced is building... and I'm eating :) Nov 25 16:40:18 bbl Nov 25 16:40:55 SHR: 03tom 07libphone-ui * r81ac9ea38869 10/src/ (phoneui-utils-sound.c phoneui-utils-sound.h): added the basic stuff for registering callbacks for mute events as well Nov 25 16:41:02 mrmoku, ciao. ;P (you haven't answered about using the volume_changed callbacks) Nov 25 16:41:03 :P Nov 25 16:43:32 TAsn, mickeyl: on boot accels are disabled Nov 25 16:43:44 they are enabled only after suspend and resume Nov 25 16:44:32 dos1, WOW. that's an interesting find. Nov 25 16:45:39 TAsn: and i already noticed slow down after first suspend/resume cycle ages ago Nov 25 16:45:43 but noone believed me :P Nov 25 16:45:55 hehe. Nov 25 16:46:06 I just noticed amazing slowdown in general! Nov 25 16:46:12 TAsn: me too Nov 25 16:46:48 mickeyl, btw, did you find the opimd_redesign notes? :) Nov 25 16:47:55 mrmoku: not possible Nov 25 16:48:32 TAsn: i'll do some housecleaning later today, including searching my bag Nov 25 16:48:41 mickeyl, :P Nov 25 16:48:42 tahnks. Nov 25 16:48:45 thanks* Nov 25 16:48:53 dos1, ! Nov 25 16:49:06 oops, I just did an upgrade and now I get: Nov 25 16:49:06 * Package libpam-base-files wants to install file /etc/pam.d/polkit Nov 25 16:49:06 But that file is already provided by package * policykit Nov 25 16:49:11 dos1, I added /etc/phoneprefix to libphone-utils's package Nov 25 16:49:45 please use that to populate the list we talked about (for selecting country in wizard) Nov 25 16:50:01 TAsn: ok, you already told me that ;P Nov 25 16:50:06 btw, phoneutils does not yet accept '|', so you'll have to clean it before passing it to me Nov 25 16:50:07 dos1, really? Nov 25 16:50:12 :\ Nov 25 16:50:28 * TAsn is sad, because he's getting old and loosing his memory. Nov 25 16:52:06 mrmoku: attention, you are building libpam-base-files (there was some commit about that recently) whcih doesn't exist yet on the FR Nov 25 16:52:18 mrmoku: and there was a report on the ML that there is a file conflict Nov 25 16:52:27 commuting home now Nov 25 16:52:46 mrmoku: file conflict should be fixed or at least I've seen patch for it Nov 25 16:53:45 delete mode 100644 recipes/pam/libpam-base-files/pam.d/polkit Nov 25 16:53:58 already in shr/merge.. at least I had conflict with this one Nov 25 16:54:19 now looking at powertop on 2.6.31 Nov 25 16:54:19 ;) Nov 25 16:54:25 TAsn: I'm not happy. Abaddon_ told me you borked his phone, and he forced me to pay him lunch. Nov 25 16:54:30 Fix it! :-) Nov 25 16:55:04 Ainulindale, it's not me! Nov 25 16:55:04 it's that damn kernel! Nov 25 16:55:16 :-) Nov 25 16:55:19 mrmoku: did you build some newer 2.6.31 kernel? Nov 25 16:55:20 Excuses, excuses... Nov 25 16:55:25 dos1, if it's ok in .6.31 it's maybe a good idea to just migrate Nov 25 16:55:32 (after some missing issues will be fixed) Nov 25 16:55:52 hi guys Nov 25 16:55:57 so where is that shiny new kernel? Nov 25 16:56:35 found it Nov 25 16:56:39 I miust admit the 20/11 image was kinda glitchy on the phone side :) Nov 25 16:57:22 the 24/11 looks better though Nov 25 16:59:21 mrmoku: I haven't time to do really exhaustive testing, but so far the only problem I've seen after upgraded yesterday aft was the problem with not being able to answer a call when waking up from suspend... Should I open an issue for that on Trac? Nov 25 17:01:22 with 2.6.31: Nov 25 17:01:24 no ACPI power usage estimate available Nov 25 17:01:25 No detailed statistics available; please enable the CONFIG_TIMER_STATS kernel option Nov 25 17:01:30 This option is located in the Kernel Debugging section of menuconfig Nov 25 17:01:30 (which is CONFIG_DEBUG_KERNEL=y in the config file) Nov 25 17:01:30 Note: this is only available in 2.6.21 and later kernels Nov 25 17:15:54 hi, is it still recommended to use shr-unstable? Nov 25 17:16:18 naeg: of course Nov 25 17:16:34 naeg: depends how brave you are... :-) Nov 25 17:16:48 i am using an old installation, and some of the messages just don't arrive at other people Nov 25 17:16:53 's phones Nov 25 17:17:14 * mickeyl boots 2.6.31 Nov 25 17:17:17 known problem? Nov 25 17:17:29 somehow e-wm, e-wm-utils and policykit depend on libpam. i thought libpam is only included if it's added to distrofeatuers... Nov 25 17:18:34 naeg: I've been using the latest version since yesterday, and it seems pretty good. The only flaw I've found so far is that if you get a call while the phone is suspended, it can't be answered... Nov 25 17:19:39 wjbaird: really? this is working for me with an old installation Nov 25 17:19:54 i need this function Nov 25 17:19:56 naeg: How old is your current SHR? I was previously using a build from sept, and hadn't noticed problems with text messages not arriving. Nov 25 17:20:13 naeg: it's expected to be fixed in a day or two, I believe mrmoku said he'd look at it today... Nov 25 17:20:30 how can i get rid of Nov 25 18:19:56 om-gta02 user.info kernel: [glamo-drm] Failed to allocate object Nov 25 17:20:32 ? Nov 25 17:20:34 naeg: It was broken as part of the recent 'big bang' of changes... Nov 25 17:20:37 that's kind of swamping my dmesg Nov 25 17:21:20 wjbaird: the messages doesn't arrive at other peoples phones. i get all mesasges(afaik) but the others don't get mine Nov 25 17:21:29 i don't know how old the installation is Nov 25 17:22:18 mickeyl: switching DRM off will get rid of it most quickly Nov 25 17:22:25 nelson is online, i'll ask him Nov 25 17:22:32 Weiss: how would i do that? Nov 25 17:22:59 mickeyl: you'd need to change the kernel config. hmm.. are you using mrmoku's kernel as well? Nov 25 17:23:05 you're the second person with those messages :S Nov 25 17:23:21 yeah, didn't bother to build it myself, using mrmoku's Nov 25 17:24:17 JaMae, do you have a wikipage for the new navit features Nov 25 17:24:19 mickeyl: 2.6.31 or 2.6.29(andy-tracking) Nov 25 17:24:21 +? Nov 25 17:24:26 2.6.31 Nov 25 17:24:57 naeg: I haven't noticed that with my old shr... I think I was running a build from around the begining of Sept. Nov 25 17:25:25 mickeyl: ah.. noone should be using 2.6.31+DRM right now :S.. Nov 25 17:25:56 ok, so i ignore the messages for now Nov 25 17:26:46 mrmoku: the speed difference _is_ visible Nov 25 17:26:47 amazing Nov 25 17:26:50 larsc: ohh, no more usb host mode? Nov 25 17:26:52 mickeyl: yup :) Nov 25 17:27:37 * mickeyl reads up on rfkill Nov 25 17:28:04 dos1: newer than when? Nov 25 17:28:08 is the same speed improvement with 2.6.31 visible without DRM? Nov 25 17:28:25 mrmoku: than i tested it the first time :P Nov 25 17:29:00 dos1: which was when? ... hmm IIRC it was before KMS, then yes Nov 25 17:29:05 dos1: do you have kms? Nov 25 17:29:21 mrmoku: i think it was before kms ;) Nov 25 17:29:26 ok, i'll upgrade Nov 25 17:29:39 hmm Nov 25 17:29:50 * mrmoku should have read spaetz's comment about pam things earlier :| Nov 25 17:30:29 Weiss: sorry, no idea what I'm actually using. 2.6.31 w/ mrmoku config Nov 25 17:31:07 mickeyl: it is the gta02-defconfig from the kernel tree with some patches applied Nov 25 17:31:24 from recipes/linux/linux-openmoko-2.6.31/ Nov 25 17:31:35 mickeyl: if it's 2.6.31 and you're getting anything that mentions DRM, you have 2.6.31+DRM, which noone should be using yet (I only got it to compile a few days ago) Nov 25 17:31:48 TAsn: fsodeviced is in the feed Nov 25 17:32:00 spaetz: just upgraded and had no conflicts Nov 25 17:32:39 bbiab Nov 25 17:33:08 mrmoku, THANKS. :) Nov 25 17:33:18 mickeyl: not that I'm complaining about the testing :() Nov 25 17:33:21 :)* Nov 25 17:33:49 can't see anything about DRM since dmesg is floded with glamo complaints ;) Nov 25 17:33:52 flooded, even Nov 25 17:34:15 * mickeyl can't stop to pop up and down the shelve Nov 25 17:34:16 mrmoku: nobody cared to reimplement it yet. Nov 25 17:34:17 it's soooooo faaaaaaast Nov 25 17:34:35 :D Nov 25 17:34:49 i sent mail to openmoko-kernel, as nelson wanted Nov 25 17:35:03 feel free to add something Nov 25 17:35:09 mickeyl: those complaints *are* DRM :) Nov 25 17:35:28 Weiss: oi Nov 25 17:35:29 oki Nov 25 17:35:33 then we know at least that :) Nov 25 17:35:58 I'm curious about why it's so fast.. because those DRM warnings indicate that (likely) something in X has been pushed out of VRAM Nov 25 17:37:05 I was wondering about speeding up (say) GTK by just refusing to let small pixmaps go in VRAM Nov 25 17:37:21 our biggest limiting factor at the moment is that X tries to accelerate way too much Nov 25 17:37:37 (in fact, this is a major limiting factor even on "real" graphics hardware) Nov 25 17:37:38 mrmoku: http://build.shr-project.org/tests/mrmoku/2.6.31/images/om-gta02/ , yup? Nov 25 17:42:34 kernel panic :x Nov 25 17:44:25 larsc: ic Nov 25 17:44:52 dos1: yup Nov 25 17:45:04 dos1: had that on first boot too... just retry Nov 25 17:45:13 retried and it works Nov 25 17:45:15 ;) Nov 25 17:50:55 Weiss: not meassured feeling was that 2.6.31 was a bit faster even without DRM Nov 25 17:54:43 TAsn: hmm... does fsodeviced work for you? Nov 25 17:54:54 mrmoku, haven't tested yet Nov 25 17:55:03 sec, will test Nov 25 17:55:06 (already installed it) Nov 25 17:55:06 ahh... gives me time to do the config :P Nov 25 17:55:09 mrmoku, where is libsdl-mixer? Nov 25 17:55:29 mrmoku, ok, I'll wait ;P Nov 25 17:55:31 F4t: no idea :) Nov 25 17:55:35 I'm busy with something :) Nov 25 17:56:16 opkg is faaaaaast Nov 25 17:56:54 i never thought that i will say something like this... Nov 25 17:56:57 dos1, :) Nov 25 17:57:03 that's because you are used to the broken Nov 25 17:57:05 but newest upgrade did it really faster :) Nov 25 17:57:08 bla. Nov 25 17:57:20 libapm probably consists of nothing :> Nov 25 17:57:50 previously it was processing packages with nothing about 20 times longer :P Nov 25 17:57:57 and libphone-utils also was processed faster Nov 25 17:58:05 well, i have to check if nothing is broken now ;D Nov 25 17:58:17 it was too fast :P Nov 25 17:59:34 TAsn: hmm... have to do some maths with my son first Nov 25 17:59:37 mrmoku, okie. Nov 25 18:10:07 mrmoku: rfkill is not too shabby Nov 25 18:11:10 mickeyl: good to hear :) Nov 25 18:11:25 (and learned a new word too :P) Nov 25 18:11:50 hehe Nov 25 18:12:11 opkg fast? wow, that's new indeed Nov 25 18:12:17 actually i think i quite like it Nov 25 18:12:29 wow Nov 25 18:12:30 since it reduces openmoko_powercontrol and thinkpad_powercontrol to kernel26_rfkill Nov 25 18:12:58 mickeyl: do you want me to rename the vibrators sysfs node to neo1973::vibrator on both devices or can you live with (gta01|gta02)::vibrator? Nov 25 18:13:13 larsc: i can live with 01|02 Nov 25 18:13:17 but thanks for asking Nov 25 18:13:20 good Nov 25 18:13:37 SHR: 03mok 07shr-themes * r67b98be048bd 10/ (3 files in 3 dirs): Merge branch 'master' of git+ssh://shr.bearstech.com/shr-themes Nov 25 18:13:46 SHR: 03mok 07shr-themes * r1ed3d0f675db 10/frameworkd/frameworkd-config-shr/om-gta02/frameworkd.conf: e-wm-config: make the syscon Lock button set IdleState to lock Nov 25 18:14:05 mickeyl: btw pwm is back for the power button leds Nov 25 18:14:18 larsc: ooh, that's quite amazing Nov 25 18:14:29 already in the version i have here? Nov 25 18:14:37 or just hit master? Nov 25 18:14:54 oh well... i might as well just go ahead and try :) Nov 25 18:15:01 only on my local machine Nov 25 18:15:35 ah :) Nov 25 18:16:51 larsc: you mean being able to set for instance blue led to 50% brightness? Nov 25 18:18:14 yes Nov 25 18:18:16 mrmoku, tell me when to upgrade. Nov 25 18:18:24 minute Nov 25 18:18:38 i still need to finetune the pwm period before commiting. Nov 25 18:18:55 DocScrutinizer-8: maybe you have an input on what should be used there Nov 25 18:19:11 ermm, context? Nov 25 18:19:24 pwm? Nov 25 18:19:30 hw-pwm? Nov 25 18:19:33 dim leds Nov 25 18:19:39 nice! Nov 25 18:20:18 DocScrutinizer-8: pwm for the leds Nov 25 18:21:08 * mickeyl wants apple-style swelling LEDs Nov 25 18:21:15 TAsn: feed is prepared to conquer (and break) your phone :) Nov 25 18:21:19 what kind of pwm? hw or sw-fiq? and what is the exact question? Nov 25 18:21:34 okie :) Nov 25 18:21:56 hw pwm. the question is: what should i consider when choosing the pwm period. Nov 25 18:22:04 where are the state files now ..which were previously at /usr/share/openmoko/scenarios/ Nov 25 18:22:22 UberNeo: /etc/freesmartphone/alsa/default Nov 25 18:22:35 UberNeo: if you're using fsodeviced Nov 25 18:22:36 larsc: please elaborate your question Nov 25 18:22:53 mrmoku, oterwise Nov 25 18:23:10 DocScrutinizer-8: hm... Nov 25 18:23:58 mrmoku, nothing to update :| Nov 25 18:23:58 on "pwm for leds" I can't answer as it's no question at all ;-P Nov 25 18:24:37 the question is whats the optimum period time Nov 25 18:24:49 i know that it works fine for example with 50ms Nov 25 18:25:07 lower limit Nov 25 18:25:22 25ms would be optimum Nov 25 18:25:27 ok, thanks Nov 25 18:25:38 and for the vibrator? Nov 25 18:25:41 yw Nov 25 18:25:51 100ms Nov 25 18:26:07 though probably 25ms won't hurt Nov 25 18:26:16 TAsn: update... then upgrade :P Nov 25 18:26:49 DUH. Nov 25 18:26:53 nothing still :) Nov 25 18:26:59 UberNeo: on hackable? no idea... Nov 25 18:27:39 TAsn: I know there is... as I updated myself... hmm Nov 25 18:28:28 mrmoku: [19:13] SHR: mok shr-themes * r1ed3d0f675db /frameworkd/frameworkd-config-shr/om-gta02/frameworkd.conf: e-wm-config: make the syscon Lock button set IdleState to lock Nov 25 18:28:30 hmm? Nov 25 18:28:35 frameworkd-config-shr vs. e-wm-config? Nov 25 18:29:06 mrmoku, what pkg? Nov 25 18:29:37 TAsn: lot's of pam... fsodeviced... opkg Nov 25 18:29:45 (or libopkg) Nov 25 18:30:01 larsc: btw this was for fiq Nov 25 18:30:23 mrmoku: s/fsodeviced/broken fsodeviced/ Nov 25 18:30:25 for hw-pwm you easily can go up to 1ms period Nov 25 18:30:26 *** glibc detected *** fsodeviced: malloc(): memory corruption: 0x000486c0 *** Nov 25 18:30:27 Aborted Nov 25 18:30:32 dos1: on 2.6.31? Nov 25 18:30:35 mrmoku, OH. Nov 25 18:30:38 mrmoku: no, on 2.6.29 Nov 25 18:30:39 fsodeviced ships it's own config? Nov 25 18:30:41 hmm Nov 25 18:30:44 I alreadygot fsodeviced Nov 25 18:30:51 I thought you were talking about config. Nov 25 18:30:52 no own config Nov 25 18:30:57 still in /etc/frameworkd.conf Nov 25 18:31:04 that will change in december :) Nov 25 18:31:18 dos1: yeah, have that too (on 2.6.31) Nov 25 18:31:40 hmmmm. nutella Nov 25 18:33:12 dos1: hmm... old libfso-glib :| Nov 25 18:33:39 larsc: and for the renaming: please add compatibility symlinks Nov 25 18:34:08 dos1: hmm... did not sync it... rebuilding Nov 25 18:34:29 well, i'm still using tests/mrmoku/unstable :P Nov 25 18:35:20 DocScrutinizer-8: (compatibility symlinks) the most important devices will be changed so much that userspace would need to treat them differently anyway. Nov 25 18:35:54 that's really deprecated anyway Nov 25 18:37:15 DocScrutinizer-8: what was for fiq? Nov 25 18:37:32 the mentioned 25ms Nov 25 18:38:00 * naeg is wondering why there is this new iphone-like screensaver-thingy :/ Nov 25 18:38:16 but since hw pwm should even more smooth then fiq it should be ok, right? Nov 25 18:39:06 err, please rephrase Nov 25 18:39:24 dos1: I expressed myself badly... it does not hit tmp/deploy/ipk Nov 25 18:39:27 builds fine though Nov 25 18:39:30 * mrmoku cleaning Nov 25 18:39:36 if we could have hw-pwm, I'd strongly prefer that Nov 25 18:39:39 DocScrutinizer-8: the problem and the main reason why userspace devs argues with the kernel devs about the nodes is that, first of all, it's assumed that the user can update the kernel without updating the userspace. Which was the case with qtopia e.g. But kernel devs feel easy about changing nodes because in reality it's not much of a hassle to change names for the devs on either side. Also there's one problem that userspace devs try to suppo Nov 25 18:39:55 freesmartphone.org: 03mickey 07cornucopia * rb3836f6d1242 10/fsodeviced/ (5 files in 4 dirs): fsodeviced: add skeleton for new plugin kernel26_rfkill Nov 25 18:39:55 ... many different kernel versions. But they need to anyway because they support different devices... Nov 25 18:39:55 mrmoku, so I didn't get your framework.conf update yet! Nov 25 18:40:01 TAsn: better for you ;) Nov 25 18:40:05 PaulFertser: omhacks library will fix this problem by offering a stable API on top of the kernel interface :) Nov 25 18:40:26 PaulFertser: Nov 25 18:40:26 DocScrutinizer-8, is it possible to break a number to it's core sections? i.e to COUNTRY_PREFIX, AREA_CODE, etc... ? Nov 25 18:40:36 *COUNTRY_CODE Nov 25 18:41:30 TAsn: sure, normalization does exactly that Nov 25 18:41:35 DocScrutinizer-8: I probably just didn't understand what you wanted to say and responded with some nonsense ;) Nov 25 18:41:46 I mean Nov 25 18:41:53 but only for known areas countries Nov 25 18:41:56 is it possible that a country will have Nov 25 18:42:01 country prefix of Nov 25 18:42:05 country code of Nov 25 18:42:07 97 Nov 25 18:42:16 and somewhere with an area code of 2 Nov 25 18:42:21 which will result in a number like: Nov 25 18:42:25 +972XXXXX Nov 25 18:42:33 and a country with country code of 972? Nov 25 18:42:46 probably not. Nov 25 18:43:06 larsc: please use hw-pwm if possible, rather than FIQ. If you have HW-pwm it's no problem to choose any period between 25mS and 0.1ms Nov 25 18:43:39 mickeyl, i wrote this small script to use system sounds for speech in navit: http://pastebin.ca/1687120 but it always throws an error: (Can't play song /tmp/speech.Sm57VG.wav: File or data too large) Nov 25 18:44:02 obviously not, tasn Nov 25 18:44:07 my build is ~2 days old Nov 25 18:44:23 playya_: d'oh, try player_alsa Nov 25 18:44:26 DocScrutinizer-8: ok Nov 25 18:44:28 you need a recent build though Nov 25 18:44:29 TAsn: country codes have to be distinct Nov 25 18:44:35 ok Nov 25 18:44:36 playya_: check config docs Nov 25 18:44:44 there's a new field to select the alsa player Nov 25 18:44:51 DocScrutinizer-8, I'm talking about country codes + area code Nov 25 18:44:53 vs country code Nov 25 18:45:12 TAsn: yes, me too Nov 25 18:45:14 mrmoku, so what about a new frameworkd.conf for fsodeviced? Nov 25 18:45:17 DocScrutinizer-8, okie, thanks. Nov 25 18:45:35 playya_: why? there is already a script (maybe even lots of them) Nov 25 18:45:42 TAsn: as US has 1, ther will be *no other* country starting with 1 Nov 25 18:45:58 ok, thanks. :) Nov 25 18:46:01 zoff99, because it emits the SoundStatus signal Nov 25 18:46:22 and as germany has 49 there will be no others like 495 or whatever Nov 25 18:46:29 playya_: ok. what is the soundstatus signal? Nov 25 18:46:54 zoff99, http://git.freesmartphone.org/?p=specs.git;a=blob_plain;f=html/org.freesmartphone.Device.Audio.html;hb=HEAD#SoundStatus Nov 25 18:48:50 playya_: how does the soundstatus sound? Nov 25 18:49:28 it uses alsa, so no difference Nov 25 18:49:56 playya_: what would u want to do? Nov 25 18:50:00 but it emits the signal and the rest of the system (e.g. my music daemon) can stop playing sound Nov 25 18:50:35 or oevents stops playing Nov 25 18:50:42 playya_: ok i understand Nov 25 18:50:49 freesmartphone.org: 03mickey 07cornucopia * r0360c1c3a4ed 10/fsodeviced/vapi/rfkill.vapi: fsodeviced: add rfkill vapi (scheduled for inclusion in linux.vapi before next release) Nov 25 18:53:09 DocScrutinizer-8: I think I'm going to stick with 1ms for the leds. gives way better results than 25ms Nov 25 18:53:30 DocScrutinizer-8: for the vibrator neither 100ms or 25ms seem to be ok Nov 25 18:54:07 larsc: what type of pwm???? Nov 25 18:54:38 hw Nov 25 18:55:19 you're aware vib uses a different counter/register set? Nov 25 18:55:26 yes Nov 25 18:55:28 even different clock source? Nov 25 18:55:34 yes Nov 25 18:55:43 what's the problem? Nov 25 18:56:03 lots of noise, no so much motion. Nov 25 18:56:31 that's strange Nov 25 18:56:47 100ms should work just fine Nov 25 18:57:10 how would 2sec work? just for test Nov 25 18:57:44 on a 50% it should be qsec on, 1 off then Nov 25 18:57:57 1sec on Nov 25 18:58:44 larsc: make sure you're not mixing ms amd us Nov 25 18:59:09 im pretty sure i'm not Nov 25 18:59:41 so please try "blink" the vib by using a period of 2000ms Nov 25 18:59:58 mrmoku, you didn't answer me yet, what about a frameworkd.conf for alsa player? I can't get it from opkg upgrade! :) Nov 25 19:00:13 TAsn: and you don't need it yet Nov 25 19:00:20 why not? Nov 25 19:00:25 TAsn: as fsodeviced with alsa player doesn't work at all ;p Nov 25 19:00:33 i c :) Nov 25 19:01:42 DocScrutinizer-8: hm ok apperently something is wrong here since it doesn't work Nov 25 19:02:08 larsc: btw for leds there should be absolutely no diff to tell between 10ms and 0.1ms Nov 25 19:02:55 l Nov 25 19:02:57 larsc: one of the timers is used for FIQ timer Nov 25 19:03:14 so will collide with HW PWM Nov 25 19:03:33 dunno which one is used for FIQ though Nov 25 19:04:02 i'm using 2 for fiq and 3 for pwm Nov 25 19:04:04 so please check Nov 25 19:04:29 duh, I guess we don't have 5 timers? Nov 25 19:04:35 not sure though Nov 25 19:04:55 lemme check Nov 25 19:05:20 has 5 timers Nov 25 19:05:41 but only 2 clock sources Nov 25 19:05:55 maybe 2 and 3 share a prescaler or something Nov 25 19:06:10 and the fiq driver messes up the vibrator pwm Nov 25 19:06:10 vibrator shares clock source with the 2 non-PWM timersa Nov 25 19:06:40 FIQ messes up prescaler divider I guess Nov 25 19:07:01 mrmoku: how's fixing going? Nov 25 19:07:40 dos1: cleaned and rebuilt and synced libfso-glib Nov 25 19:07:48 larsc: check TCKL1 divider and prescaler settings Nov 25 19:07:55 TCLK1 Nov 25 19:08:21 TAsn: have some problems with building... Nov 25 19:08:43 OR more like CLK1 Nov 25 19:09:33 super github down :( Nov 25 19:09:37 JaMa: help Nov 25 19:09:44 mrmoku: nothing upgradable here Nov 25 19:09:52 mrmoku, okie. Nov 25 19:10:15 *** glibc detected *** fsodeviced: malloc(): memory corruption: 0x0004a0d8 *** Nov 25 19:10:15 Aborted Nov 25 19:10:16 :| Nov 25 19:10:28 TAsn: i told you ;p Nov 25 19:10:33 oh Nov 25 19:10:37 you weren't joking ;( Nov 25 19:10:45 ffs ;) Nov 25 19:10:56 mrmoku: what? Nov 25 19:11:09 JaMa: did you check out my navit themes? Nov 25 19:11:30 larsc: TCFG0-prescaler1 has to be same value for FIQ and HW-PWM as they share it Nov 25 19:11:37 zoff99: yes.. read irc log from morning.. Nov 25 19:11:37 JaMa: somehow it builds the correct version of packages... but does not put them in deploy/ipk Nov 25 19:11:53 JaMa: have to manually clean and rebuild to get them there :| Nov 25 19:11:54 mrmoku: is there newer version of the same package? Nov 25 19:12:02 mrmoku: check morgue.. Nov 25 19:12:02 mrmoku: check if it isn't moved to morgue Nov 25 19:12:06 JaMa: where can i read the irc-log? Nov 25 19:12:11 zoff99: Nov 25 19:12:13 !logs Nov 25 19:12:13 Channel logs for #openmoko-cdevel are archived at: Nov 25 19:12:14 what is morgue for? Nov 25 19:12:15 http://hentges.net/tmp/logs/irc/%23openmoko-cdevel Nov 25 19:12:16 Live-logs are available at Nov 25 19:12:17 http://hentges.net/tmp/logs/irc/livelogs/%23openmoko-cdevel.livelog Nov 25 19:12:19 See ?? help-logs for usage instructions Nov 25 19:12:19 TCFG0 Bit Description Initial State Nov 25 19:12:21 Reserved [31:24] 0x00 Nov 25 19:12:22 mrmoku: older versions Nov 25 19:12:22 Dead zone length [23:16] These 8 bits determine the dead zone length. The 1 unit time 0x00 Nov 25 19:12:22 mrmoku: for old versions of packages :P Nov 25 19:12:24 of the dead zone length is equal to that of timer 0. Nov 25 19:12:25 Prescaler 1 [15:8] These 8 bits determine prescaler value for Timer 2, 3 and 4. 0x00 Nov 25 19:12:52 larsc: timer2 is vib Nov 25 19:12:53 mrmoku: I've seen similar behavior when building older version of some package intentionally.. Nov 25 19:13:07 DocScrutinizer: timer3 is vib Nov 25 19:13:22 mrmoku: I had to remove ipks from deploy dir to make newly built ones there.. and in package-index Nov 25 19:13:24 nah, not really... Nov 25 19:13:32 heyho Nov 25 19:13:33 JaMa: I update some revs and it rebuilds them with the same localcount Nov 25 19:13:48 0 and 1 are LED, 2 is vib, 2 and 4 are free Nov 25 19:13:53 JaMa: hmm... that is bad for automatic building :P Nov 25 19:13:54 hi, is there any progress on 2.6.31 kernel? It is very fast but fsodeviced doesn't work Nov 25 19:14:05 DocScrutinizer: 0, 1, 2 are LED Nov 25 19:14:06 s/2 a/ 3 a/ Nov 25 19:14:06 DocScrutinizer meant: 0 and 1 are LED, 2 is vib, 3 and 4 are free Nov 25 19:14:18 larsc: oops you're right Nov 25 19:14:35 mrmoku: same localcount? do you still have any bitbake patches applied? Nov 25 19:15:06 mrmoku: but even with them it shouldn't hurt.. Nov 25 19:15:34 JaMa: hmm for frameworkd-config-shr it is localcount-1 Nov 25 19:15:54 JaMa: still have your patches... should I reset? Nov 25 19:15:54 mmm... fsodeviced of today segfaulting with 2.6.29 Nov 25 19:16:00 larsc: anyway the problem is FIQ spoiling the prescaler setting Nov 25 19:16:47 mrmoku: "-1" really? Nov 25 19:17:18 JaMa: the newer one has localcount-1 respective the older one, yes Nov 25 19:17:26 DocScrutinizer-8: i disabled the fiq and now the vibrator doesn't do anything :/ Nov 25 19:17:38 lol Nov 25 19:17:40 mrmoku: and sqlite says what for that url? Nov 25 19:18:27 DocScrutinizer-8: oh well, like every 30 seconds it's turned on for 1 second Nov 25 19:18:33 mrmoku: is it ${last_localcount} - 1, or localcount == '-1' ? Nov 25 19:18:52 mrmoku: because I don't see how negative number can get to localcount.. Nov 25 19:18:53 JaMa: hmm Nov 25 19:18:57 larsc: I could only tell you how to set each register value correctly. but I have no idea about kernel arrogance to do the calculations and arbiting itself Nov 25 19:19:10 frameworkd-config-shr has that freaking SRCREV_FORMAT thing Nov 25 19:19:31 and thus has two localcounts Nov 25 19:19:45 larsc: 30? not 25? Nov 25 19:19:53 larsc: 25s is the CC poll period Nov 25 19:19:59 i did not count Nov 25 19:20:41 JaMa: frameworkd-config-shr - 0.9.5.9+gitr1723+d48bd43c430322b26744e31829f9c76e9b725819-231+cb4159e653d770da8e8f9fc9a65135839533d4c5-r7.4 Nov 25 19:20:48 is what is installed on my FR Nov 25 19:21:08 frameworkd-config-shr-0.9.5.9+gitr1722+d48bd43c430322b26744e31829f9c76e9b725819-235+67b98be048bd0f6e702678451573643d89dd93f0-r7/ Nov 25 19:21:11 is what it builds Nov 25 19:21:15 larsc: modprobe -r bq27k-foo Nov 25 19:21:32 JaMa: so frameworkd localcount went -1 Nov 25 19:21:38 wtf :/ Nov 25 19:21:41 JaMa: not the one of the rev I changed though Nov 25 19:22:09 git:git.shr-project.org.repo.shr-themes.git-frameworkd-config-shr_rev|67b98be048bd0f6e702678451573643d89dd93f0 Nov 25 19:22:12 git:git.shr-project.org.repo.shr-themes.git-frameworkd-config-shr_count|235 Nov 25 19:22:14 mrmoku: did you remove some items from sqcache? Nov 25 19:22:15 is the db Nov 25 19:22:20 DocScrutinizer-8: it already was disabled Nov 25 19:22:21 not recently Nov 25 19:22:46 larsc: well anyway seems you're close to the problem Nov 25 19:22:50 mrmoku: hmm I'm not sure how it works in sqlite cache for packages with more srcrevs :/.. I need to check it first.. Nov 25 19:23:07 mrmoku: first one is FSO_REV, yes? Nov 25 19:24:18 yup Nov 25 19:24:59 JaMa: so i read the whole day, still dont know anything. except that you dont like my patch Nov 25 19:25:26 and the "pgrep" you need, and unset LC_ALL also, its even on the navit wiki page Nov 25 19:27:13 zoff99: did you read about missing menu definition in your xml examples? Nov 25 19:27:28 zoff99: about broken old releases with "new" .inc file Nov 25 19:27:37 zoff99: about xi:include? Nov 25 19:27:40 larsc: larsc you need to set prescaler and divider to get lowest possible clock, then adjust FIQ timer autoreload/counter value to a compensated value to keep the timing. make sure PWM and FIQ driver agree on that clock. Then you can sanely use PWM for Aux and vib Nov 25 19:27:45 JaMa: nope, about xi_include yes Nov 25 19:28:04 zoff99: hmm than something maybe wasn't logged :) Nov 25 19:28:30 JaMa: also wanting to put it into navit.inc , but i especially didnt want to do that. navit will most likely change more often Nov 25 19:29:09 JaMa: and the patch didnt apply because spaetz pushed something up just today, deleted a file :) Nov 25 19:29:27 yes.. I just merged it before few sec.. Nov 25 19:29:39 JaMa: anyway if i bitbake navit, i segfaults. Nov 25 19:30:34 zoff99: which revision? Nov 25 19:30:55 JaMa: also having to almost exact same navit.lauchner scipts. thats is fine. as you have 2 pkg. also in the future someone will want to change launcher for e.g. bike Nov 25 19:31:37 zoff99: why not ship one launcher with navit itself and modify just params there in call from Exec=? Nov 25 19:31:44 JaMa: -rwxr-xr-x 1 root root 413784 Nov 25 17:54 navit_0.1.0+svnr2760-r6.0.4_armv4t.ipk Nov 25 19:32:09 does it segfault while starting or when entering menu? Nov 25 19:32:12 JaMa: because someone will tomorrow want to put "do_somthing.sh;echo "blabla" into launcher for bike Nov 25 19:32:30 JaMa: segfaults when starting reading from gpsd Nov 25 19:32:31 do you use navit.xml shiped with navit itself? Nov 25 19:32:39 nope Nov 25 19:32:48 larsc: seems with lowest possible clock the FIQ granularity is ~75us/count. PWM will probably work fine if you tune prescaler to get 50us or 25us Nov 25 19:33:12 (if that's possible to tune the divider so finegrained) Nov 25 19:34:00 DocScrutinizer-8: i'm not touching any registers the pwm driver should handle all this automagically. but it probably has some bugs somewhere Nov 25 19:34:13 JaMa: oh its not navit that segfaults, its some lib.so file Nov 25 19:34:42 larsc: the problem is PWM driver and FIQ driver share some bits of hw Nov 25 19:35:16 fsodeviced crashing with 2.6.31 too Nov 25 19:35:17 :( Nov 25 19:35:37 JaMa: if i take just the navit binary it works (with the shr lib.so files) Nov 25 19:35:53 DocScrutinizer-8: the fiq driver should use the pwm driver Nov 25 19:36:13 here what navit has to say about LC_ALL: navit:main_init:Warning: LC_ALL is set, this might lead to problems (e.g. strange positions from GPS) Nov 25 19:36:14 larsc: so whatever you do, you probably won't be able to avoid tweaking FIQ and PwM driver to collaborate on the divider/prescaler setting. If you ever want PWM put to work Nov 25 19:37:08 larsc: so check for the granularity or resolution setting in FIQ driver Nov 25 19:37:38 mrmoku: do you have BB_GIT_CLONE disabled, right? Nov 25 19:38:04 JaMa: yeah, disabled Nov 25 19:38:07 probably if you set this to 10us it will just work Nov 25 19:38:37 freesmartphone.org: 03mickey 07cornucopia * r20f435265ef1 10/fsodeviced/ (3 files in 2 dirs): fsodeviced: kernel26_rfkill: do not use sysfs; objects are created based on what /dev/rfkill reports Nov 25 19:38:39 time4 is used for the system timer. and guess what, it also access the raw registers and prescaller values Nov 25 19:38:50 mrmoku: i just tried to rebuild and still 0.9.5.9+gitr0+d48bd43c430322b26744e31829f9c76e9b725819-8+67b98be048bd0f6e702678451573643d89dd93f0-r7.4 Nov 25 19:39:23 JaMa: hmm 0 for FSO and 8 for SHR Nov 25 19:39:44 mrmoku`: yes.. maybe first one is not incremented while rebuilding.. Nov 25 19:39:53 JaMa: that SRCREV_FORMAT must die :) Nov 25 19:40:13 mrmoku`: and your gitr1723 is from time when you had bitbake with bug (incrementing count grom wc -l Nov 25 19:40:22 anyway I have to do some university stuff first. Nov 25 19:40:22 s/grom/from/ Nov 25 19:40:22 JaMa meant: mrmoku`: and your gitr1723 is from time when you had bitbake with bug (incrementing count from wc -l Nov 25 19:40:22 mrmoku`: why for there is such SRCREV Nov 25 19:40:24 JaMa: yes Nov 25 19:40:25 +_FORMAT? Nov 25 19:40:33 mrmoku`: bump it in sqlite manually Nov 25 19:40:36 dos1: I don't know why Nov 25 19:40:43 JaMa: ok Nov 25 19:40:52 to 1724 Nov 25 19:40:52 JaMa: but I had similiar problem with libfso-glib too Nov 25 19:41:10 JaMa, how do you calculate the version for navit-icons? Nov 25 19:41:17 mrmoku`: but there is only one SRCREV isn't it? Nov 25 19:41:19 JaMa: hmm which one do I have to bump? frameworkd? Nov 25 19:41:27 JaMa: yup there is only one Nov 25 19:41:27 playya_: its from sane-srcrev.. Nov 25 19:41:33 playya_: bumped manually Nov 25 19:41:36 ah. ok Nov 25 19:41:47 bbl Nov 25 19:41:54 playya_: before bumping navit rev, you should check if there is something new in xpm Nov 25 19:42:00 git:git.freesmartphone.org.framework.git-frameworkd-config-shr_count|1722 Nov 25 19:42:03 this one Nov 25 19:42:22 mrmoku`: yes Nov 25 19:42:41 mrmoku`: and I'll check what's happening if there are 2 srcrevs in SRCREV_FORMAT Nov 25 19:43:05 mrmoku`: what does frameworkd-config-shr use from FSO? Nov 25 19:43:26 mrmoku`: because if its only from shr repo and we always want newest rev there.. Nov 25 19:43:41 mrmoku`: then se should probably use only rev from shr git, no? Nov 25 19:43:43 JaMa: I _think_ it once used something... don't know now Nov 25 19:43:47 * mrmoku` checks recipe Nov 25 19:44:12 JaMa: ooh, now it uses Nov 25 19:44:28 larsc: i reckon all devices will be turned off by default when you boot the kernel, right? i.e. rfkill devices will report SOFT_BLOCK for both on startup? Nov 25 19:44:38 JaMa: stuff like networks.tab Nov 25 19:44:45 s/now/no/ Nov 25 19:45:04 mrmoku`: ah ok.. then I should check it in bitbake :) Nov 25 19:45:18 mrmoku`: but probably not today :/ Nov 25 19:45:52 mickeyl: yes Nov 25 19:46:00 larsc: excellent Nov 25 19:46:00 JaMa: ok Nov 25 19:46:01 thanks Nov 25 19:46:13 PaulFertser: (sysfs changes) if we need new sysfs nodes or other functions to control whatever hw, we definitlely should keep the old sysnodes as well, as long as they're not completely buggy or the pragma changed that much you could impossibly keep them while using the new stuff Nov 25 19:46:22 * mrmoku` wonders if mickeyl got some motivation by trying the kernel :) Nov 25 19:46:38 * mickeyl smiles viciously Nov 25 19:47:11 mrmoku: SRCREV_pn-frameworkd-config-shr ?= "cb4159e653d770da8e8f9fc9a65135839533d4c5" Nov 25 19:47:14 SRCREV_pn-frameworkd-config-shr_SHR_REV = ${SRCREV_pn-frameworkd-config-shr} Nov 25 19:47:38 PaulFertser: and it's kernel devels damn duty to support a compatibility layer when developing new drivers Nov 25 19:47:50 mrmoku: is it AUTOREV? then not sure if it works for SRCREV_pn-frameworkd-config-shr_SHR_REV or if its not then why do you have that revision you pasted in sqlite Nov 25 19:48:58 JaMa: frameworkd-config-shr is autorev Nov 25 19:49:17 SRCREV_pn-frameworkd-config-shr = "${AUTOREV}" Nov 25 19:49:17 SRCREV_pn-frameworkd-config-shr_FSO_REV = "${AUTOREV}" Nov 25 19:49:17 SRCREV_pn-frameworkd-config-shr_SHR_REV = "${AUTOREV}" Nov 25 19:49:26 is what we have in shr-autorev.inc Nov 25 19:49:50 PaulFertser: app devels maybe have only few patches they need to do, but you don't know how hard they learnt to understand the old sysnode semantics. and if they ever care to learn the new ones Nov 25 19:51:40 PaulFertser: see kernel-ml quote of linus Nov 25 19:54:09 PaulFertser: changes in kernel api basically render useless the whole repository of apps for all users. And you nevr know how long it will take until all app devels even notice the breakage and needed fixes Nov 25 19:55:05 DocScrutinizer-8, another question Nov 25 19:55:15 is it allowed for a number to start with other countries CC Nov 25 19:55:28 i.e, if the US's CC is 1 Nov 25 19:55:33 is it ok for a number to be: Nov 25 19:55:38 +972 - 3 - 12345 ? Nov 25 19:55:40 or Nov 25 19:55:48 +972 - 1 - 23456 ? Nov 25 19:56:13 PaulFertser: you never change the parameters or semantics of a lib function, rather you add another function foo_new(). or you keep the whole lib and bump version so you basically get a whole new one. Same rationale should apply for sysfs Nov 25 19:57:20 sure, TAsn. why not? Nov 25 19:57:49 DocScrutinizer-8, just was interested. Nov 25 19:58:43 JaMa: I have a similiar problem with libfso-glib though Nov 25 19:59:01 JaMa: libfso-glib0_0.2.2+gitr64+3630315a498f91106b4187e41608613520d63090-r0.4_armv4t.ipk is in deploy Nov 25 19:59:16 libfso-glib-1_0.2.2+gitr64+9a627aa1c33f5a46ae4316fc274126ca5f37e979-r0 got build after rev bump Nov 25 19:59:25 and it does not put it in deploy Nov 25 20:00:22 JaMa: puts it directly in morgue Nov 25 20:01:48 * mrmoku bumped localcount manually for that Nov 25 20:02:21 now for real ->doc away Nov 25 20:06:37 mickeyl: I got those DRI failures too. Welcome in the club. But .31 IS faster, right :-) ? Nov 25 20:07:08 mrmoku: checked bitbake and it uses different url and key in database so it should be safe.. I can give you patch for more verbose output.. then you need to update BB_URI_LOCALCOUNT set value='blah' where key like '%frameworkd-config-shr_rev'; to force rev change it gives me http://shr.pastebin.com/m69732e40 Nov 25 20:07:14 mrmoku: but there is filename different.. Nov 25 20:07:14 mrmoku: libfso-glib wont upgrade with localcount bump Nov 25 20:07:23 spaetz: it is :) Nov 25 20:07:43 JaMa: why won't it upgrade? Nov 25 20:07:46 libfso-glib0 - libfso-glib Nov 25 20:07:46 second has higher version (even when first one is 64 instead of 63 because of that bug before.. Nov 25 20:07:46 mrmoku: http://jama.homelinux.org/org.openembedded.shr/0001-fetcher-make-sqlite-verbose-a-lot.patch Nov 25 20:08:20 JaMa: nah, that is because one is the ipk and the other the workdir in tmp :) Nov 25 20:09:04 spaetz: I extended my switcher script :) Nov 25 20:09:46 mickeyl: ERROR:inotifier.c:120:fso_framework_inotifier_construct: assertion failed: (self->priv->fd != (-1)) Nov 25 20:09:55 TAsn, dos1: upgrade Nov 25 20:10:13 THAKN.S Nov 25 20:10:14 mrmoku, any progress on themeability? ( iknow broken fsodeviced has priority...) Nov 25 20:10:27 JesusMcCloud: sorry, not yet Nov 25 20:10:53 mrmoku: checking what changed... I can't see a difference Nov 25 20:11:08 spaetz: nah, did not yet put it only, sorry :P Nov 25 20:11:22 spaetz: I added 2.6.31 compatible alsa scenarios Nov 25 20:11:25 ahh :). But it is quite useful, thanks for it Nov 25 20:11:25 mrmoku: can you strace that? Nov 25 20:11:34 mickeyl: moment Nov 25 20:11:58 might want to disable the other modules to not clutter output Nov 25 20:11:58 g_ether would be nice in the new kernel Nov 25 20:12:13 mickeyl: SYS_360(0, 0xdd000, 0x49e10, 0x1, 0x49e00) = -1 ENOSYS (Function not implemented) Nov 25 20:12:23 heh Nov 25 20:12:35 CONFIG_INOTIFY = y Nov 25 20:12:43 .31 party, yay Nov 25 20:13:00 I'll need to guard it Nov 25 20:13:14 CONFIG_INOTIFY=y Nov 25 20:13:14 # CONFIG_INOTIFY_USER is not set Nov 25 20:13:20 where is our kernel config for .31? Nov 25 20:13:23 is what we have Nov 25 20:13:38 spaetz: in kernel tree and some patches in oe Nov 25 20:15:18 thanks, found it Nov 25 20:16:21 spaetz: updated the script now Nov 25 20:16:40 spaetz: it will download and install 2.6.31 compatible alsa state files Nov 25 20:16:48 spaetz: and then link them to this or that Nov 25 20:17:49 freesmartphone.org: 03mickey 07cornucopia * r37ddf9af4c1e 10/libfsoframework/fsoframework/inotifier.vala: libfsoframework: inotifier: do not bail out if inotify subsystem can't be initialized; report the error instead Nov 25 20:18:07 nice Nov 25 20:18:32 but shouldn't CONFIG_INOTIFY=y be enough? Nov 25 20:18:42 * spaetz has no clue what CONFIG_INOTIFY_USER does Nov 25 20:19:05 and g_ether is CONFIG_USB_USBNET, right? Nov 25 20:19:18 hmm Nov 25 20:19:29 can you enable NFS, while you're there? Nov 25 20:20:03 I am not there and I don't have any commit rights anyway :) Nov 25 20:20:11 just making a wishlist... Nov 25 20:20:20 spaetz: CONFIG_USB_ETH=y Nov 25 20:20:34 USBNET is the counterpart on the other side I think Nov 25 20:20:37 ahh, so it should work, theoretically Nov 25 20:20:45 spaetz: it does for me Nov 25 20:20:56 spaetz: note I got a different eth though Nov 25 20:21:08 ahh, that might be, need to check that Nov 25 20:21:08 on 2.6.29 it was eth1 now it's eth2 Nov 25 20:21:53 mickeyl, vala 0.7.8.2 stages vala-1.0.pc only into the hosts path. it's still 0.7.8.1 in targets staging Nov 25 20:22:17 spaetz: it does for me too.. and I said its included and you didn't believe me? :) Nov 25 20:22:39 No, I never believe you unless I see a prove :P Nov 25 20:22:39 playya_: did you build 0.7.8.1 for the target? Nov 25 20:22:54 will have to check what goes wrong then Nov 25 20:22:55 can't see why anything should have changed Nov 25 20:23:03 it's pretty much the same buildwise Nov 25 20:23:15 only some fixes in the bindings Nov 25 20:23:33 yes. but it wasn't intenional Nov 25 20:23:39 ah Nov 25 20:23:41 perhaps that confuses it Nov 25 20:24:03 i'll launch a fresh build from scratch how to see what happens Nov 25 20:24:06 s/how/now/ Nov 25 20:24:06 mickeyl meant: i'll launch a fresh build from scratch now to see what happens Nov 25 20:24:07 but a build should override the file. strange Nov 25 20:24:10 * mrmoku is rebuilding libfsoframework Nov 25 20:24:19 need to fix linux-libc-headers anyways Nov 25 20:24:21 for fsodeviced Nov 25 20:28:04 spaetz: hmm how sad :P Nov 25 20:29:03 mrmoku: I don't understand that about tmpdir.. Nov 25 20:29:20 mm dinner Nov 25 20:29:25 Say Y here to enable inotify support for userspace, including the associated system calls. Inotify allows monitoring of both files and directories via a single open fd. Events are read from the file descriptor, which is also select()- and poll()-able. Nov 25 20:29:31 you need CONFIG_INOTIFY_USER Nov 25 20:29:42 or rather, we Nov 25 20:29:48 more cheess than spaghetti :) Nov 25 20:30:26 who has access to that def config and who do we need to ask for permission? Nov 25 20:30:38 or does gta02 wants to fork its own defconfig anyway? Nov 25 20:30:49 meatballs! (cheese is optional) Nov 25 20:30:50 well Nov 25 20:31:05 OE's policy usually is to keep their own defconfig Nov 25 20:31:46 since distro and machine policies may have an effect on the defconfig Nov 25 20:31:54 yep Nov 25 20:32:02 spaetz: I sent patch to lars and he applied that.. Nov 25 20:32:39 spaetz: but we have defconfig patch in OE for ubi.. and DRM so we can first test it from there.. Nov 25 20:34:35 JaMa: I copy & pasted the url from tmp/work/... which does not contain PE Nov 25 20:34:49 JaMa: just to show localcounnt + git hash Nov 25 20:35:02 spaetz: and I'm using patches agains kernel git in OE, not own defconfig.. just to prepare easier upstream inclusion Nov 25 20:35:21 mrmoku: ah, but then why is not the second higher version? looks like higher to me Nov 25 20:36:14 JaMa: that's what I don't understand... the git has is higher and thus it should have taken that one Nov 25 20:36:21 mickey|bbl: nice try... still no go :) Nov 25 20:36:31 ** ERROR **: inotifier.vala:59: Can not init the inotify subsystem: Function not implemented; some features will not work Nov 25 20:36:40 after that it aborts Nov 25 20:37:42 * mrmoku switches kernel to see if it works on 2.6.29 Nov 25 20:39:39 on 2.6.29 it works Nov 25 20:39:39 ** (process:2007): DEBUG: inotifier.vala:63: inotifier created Nov 25 20:39:39 ** (process:2007): DEBUG: inotifier.vala:112: wd = 1 Nov 25 20:39:39 ** Message: inotifier.vala:120: b Nov 25 20:39:44 ** (process:2007): DEBUG: inotifier.vala:122: inotifier watch added, total 1 Nov 25 20:39:44 and then typical kobjectnotifier lines Nov 25 20:39:51 <[Rui]> hi Nov 25 20:40:09 <[Rui]> I can't git clone git.shr-project.org/repo/libphone-ui-shr.git => fatal: The remote end hung up unexpectedly Nov 25 20:41:43 <[Rui]> am I using the wrong url? Nov 25 20:42:14 [Rui]: you want to clone to commit or just to see? Nov 25 20:42:42 <[Rui]> mrmoku: well, I'd like to integrate what I did with sss into liphone-ui-shr Nov 25 20:43:00 <[Rui]> mrmoku: but I don't mind not having commit access, I can just as well send a diff Nov 25 20:43:02 [Rui]: then send me your public ssh key Nov 25 20:43:10 <[Rui]> mrmoku: ok Nov 25 20:43:16 [Rui]: nah. that would just be more work for me ;) Nov 25 20:43:23 <[Rui]> I will generate a new one, then Nov 25 20:43:31 oh! Nov 25 20:43:31 mickey|bbl: ok, on 2.6.29 it does not abort Nov 25 20:43:38 sound isn't choppy! Nov 25 20:43:40 :) Nov 25 20:43:50 dos1: with the new alsa player plugin? Nov 25 20:44:01 yup Nov 25 20:44:01 works nicely here too Nov 25 20:44:10 <[Rui]> mrmoku: 3f:9c:8d:27:9b:55:ef:5c:47:16:50:eb:8e:ac:a3:6f (will mail the key) Nov 25 20:45:01 [Rui]: ok Nov 25 20:45:59 <[Rui]> mrmoku: to mnet-online? Nov 25 20:46:00 tztzztz, I seem to have told "use alsa, use aplay" Nov 25 20:46:06 [Rui]: yup Nov 25 20:48:48 <[Rui]> mrmoku: sent Nov 25 20:49:32 <[Rui]> also signed with my pgp key, should you want to check, 0x7A34FADE Nov 25 20:50:19 [Rui]: now, you should give him a call, or meet him at Munich main station to confirm the key ID ;) Nov 25 20:50:22 SCNR Nov 25 20:50:45 <[Rui]> spaetz: indeed :) we could make a signing party if we meet at fosdem :) Nov 25 20:51:25 isn't it already signed by heise.de? ;-) Nov 25 20:52:02 <[Rui]> mrmoku: what's the usual commit policy, specially for new comers? ask first in shr-devel? Nov 25 20:52:28 mrmoku: don't break it or you'll get whipped Nov 25 20:52:33 err, [Rui] Nov 25 20:52:39 <[Rui]> lol Nov 25 20:54:13 I'll send my army of bats Nov 25 20:56:28 and we implemented a selfdestruction to git Nov 25 20:56:35 *yawn* Nov 25 20:56:42 with remote trigger Nov 25 20:56:55 * [Rui] rolls on the floor laughing Nov 25 20:57:35 *update* Nov 25 20:57:50 hehe Nov 25 21:03:32 [Rui]: hmm first barrier is your key... still got no mail :P Nov 25 21:03:48 <[Rui]> mrmoku: you have greylisting? Nov 25 21:03:55 <[Rui]> mrmoku: can I priv-paste you? Nov 25 21:04:17 [Rui]: possible that my provider is greylisting... Nov 25 21:04:31 [Rui]: no idea how priv-pasting works :P but just try Nov 25 21:04:50 [Rui]: or better still Nov 25 21:04:54 <[Rui]> mrmoku: yeah, it's greylisted... Nov 25 21:04:57 send it to the other address Nov 25 21:05:02 <[Rui]> just checked on my server Nov 25 21:07:50 [Rui]: ok, now git clone git+ssh://git@git.shr-project.org/libphone-ui-shr.git Nov 25 21:08:17 brb Nov 25 21:08:20 <[Rui]> I've added that host to my ssh config Nov 25 21:08:49 <[Rui]> mrmoku: got a permission denied Nov 25 21:14:54 hey, mokonnect scan cannot find my wlan while "iwlist scan" can Nov 25 21:14:57 dammit Nov 25 21:17:16 [Rui]: hmm Nov 25 21:17:51 [Rui]: permission denied? or a hangup? Nov 25 21:17:59 <[Rui]> permission denied Nov 25 21:18:08 <[Rui]> with hangpu Nov 25 21:18:26 <[Rui]> s/pu/up/ Nov 25 21:18:26 [Rui] meant: with hangup Nov 25 21:18:42 [Rui]: just tried the command I pasted... that works Nov 25 21:18:50 <[Rui]> mrmoku: problem with the key? Nov 25 21:18:54 <[Rui]> is it all in one line? Nov 25 21:18:58 yup Nov 25 21:19:26 <[Rui]> hms... I wonder if it's using my default key rather than the new one? Nov 25 21:25:53 spaetz: yeah, had that often too Nov 25 21:26:03 <[Rui]> should I put sss.c in src/phoneui-quick-settings.c or in some other subdir? Nov 25 21:26:09 <[Rui]> or other name? Nov 25 21:26:38 [Rui]: what does it contain? Nov 25 21:27:04 <[Rui]> mrmoku: the program for simple settings I have been working on Nov 25 21:27:08 [Rui]: there is two parts... phoneui-quick-settings.c and views/quick-setting-view.c Nov 25 21:27:31 <[Rui]> hms? what should be in each? Nov 25 21:27:40 [Rui]: the first one is just the wrapper to call stuff from libphone-ui Nov 25 21:27:48 the views are the actual UI ones Nov 25 21:27:56 so I guess view/quick-setting-view.c Nov 25 21:28:21 <[Rui]> ok, so then it should #include "../phoneui-quick-settings.h" ? Nov 25 21:28:43 hmm Nov 25 21:29:33 what do you have in the .h Nov 25 21:29:35 ? Nov 25 21:29:58 [Rui]: well just put it somewhere... we can still clean it up Nov 25 21:30:19 [Rui]: or take a look at phoneui-dialer.c and views/dialer-view.c Nov 25 21:32:04 <[Rui]> ah so it's there as well? :) Nov 25 21:32:30 the dialer? Nov 25 21:32:31 <[Rui]> interesting... as I have an angry finger to point at it (closing completely after a call) Nov 25 21:32:36 <[Rui]> yeah :) Nov 25 21:32:50 and now you know where to fix it :P Nov 25 21:33:04 [Rui]: I asked tasn already Nov 25 21:33:08 ? Nov 25 21:33:09 he told me it is by design Nov 25 21:33:10 ... Nov 25 21:33:14 Not? Nov 25 21:33:16 what? Nov 25 21:33:25 dialer closing after a call Nov 25 21:33:26 dialer closing Nov 25 21:33:30 by design Nov 25 21:33:32 yeah. Nov 25 21:33:41 I'll make it configurable Nov 25 21:33:48 after I'll finish this cool libphone-utils Nov 25 21:33:49 <[Rui]> TAsn: grr! Nov 25 21:33:51 improvement :) Nov 25 21:33:55 don't know if that's needed Nov 25 21:33:57 [Rui], I didn't design that. Nov 25 21:33:57 if we make it cached Nov 25 21:34:01 <[Rui]> TAsn: *g* Nov 25 21:34:10 [Rui]: what do you need it open for? redial? Nov 25 21:34:19 <[Rui]> redial, make a new call, etc... Nov 25 21:34:21 for the record, I don't need to keep it opened Nov 25 21:34:41 I was just asking tasn about whether it's intentional :) Nov 25 21:35:42 TAsn: did you upgrade and does it work now? Nov 25 21:36:58 no new fsodeviced Nov 25 21:37:02 is that intentional? Nov 25 21:37:17 ok Nov 25 21:37:18 yeah Nov 25 21:37:23 sec, testing Nov 25 21:37:39 mrmoku, please bind you ui-shr function to the volume changed callback Nov 25 21:37:44 it's really important Nov 25 21:37:54 and easy to do. Nov 25 21:38:37 TAsn: will look at that now Nov 25 21:38:43 yay, thanks :) Nov 25 21:38:45 btw Nov 25 21:38:51 mrmoku, also added "mute" function Nov 25 21:38:58 so please also make use of that Nov 25 21:39:04 (with the toggle0 Nov 25 21:40:20 <[Rui]> I'm guessing that what I used to have as main should now be a similar function to all those others in views.h, yes? Nov 25 21:40:20 by the way, just upgraded my SHR tree, as told on http://www.shr-project.org/trac/wiki/Building%20SHR, and tried to build shr-unstable, and after three days it still fails because wgetting http://sources.openembedded.org//matchbox-panel-2_svn.o-hand.com_.repos.matchbox.trunk_2098_.tar.gz fails with 404 Nov 25 21:41:07 Wonka: not true, that failure is not critical. It will try to download stuff via SVN if there is no mirror Nov 25 21:41:25 delete the downloads/svn.o-hand directory Nov 25 21:41:37 (or however it is called) and retry Nov 25 21:41:55 oh, you're right... Nov 25 21:42:07 /usr/bin/env svn update -r 2098 failed with signal 1, output: Nov 25 21:42:07 svn: Failed to add directory 'applets': an unversioned directory of the same name already exists Nov 25 21:42:20 TAsn: ohh... you have to fix the callback Nov 25 21:42:24 I need userdata :) Nov 25 21:42:29 mrmoku, oops :) Nov 25 21:42:30 sec. Nov 25 21:44:02 spaetz: but trying to "find . -type d -name applets" returns nothing... Nov 25 21:44:53 aargh. because it's behind a symlink. Nov 25 21:44:57 *sigh* Nov 25 21:45:30 * Wonka needs more sleep Nov 25 21:47:23 TAsn: another thing... Nov 25 21:47:29 TAsn: why long? Nov 25 21:47:36 I hope you don't pass me the raw value :P Nov 25 21:47:50 hm.. If I do, it's not on purpose. Nov 25 21:47:53 sec, changing to int. Nov 25 21:49:39 mrmoku, give it a go Nov 25 21:49:43 I haven't even compiled :) Nov 25 21:49:51 SHR: 03tom 07libphone-ui * r494710f72a5e 10/src/ (phoneui-utils-sound.c phoneui-utils-sound.h): changed the longs to ints (type of volume in the volume_changed callback). Nov 25 21:50:02 SHR: 03tom 07libphone-ui * r14ace6a162e0 10/src/ (phoneui-utils-sound.c phoneui-utils-sound.h): added the forgotten userdata to the callbacks Nov 25 21:50:44 mrmoku, I can't reliably test fsodeviced without this change of your Nov 25 21:50:45 s Nov 25 21:50:47 so let me know. Nov 25 21:50:49 :) Nov 25 21:52:00 <[Rui]> hms... how do I use a local git repo? I copied libphone-ui-shr recipe to another file, but I don't want to commit anything that doesn't work, so I need to build it first... Nov 25 21:53:02 <[Rui]> git:///home/me/git/libphone-ui-shr;protocol=file... but it seems it was a naive url Nov 25 21:53:21 skip the protocol= Nov 25 21:53:31 <[Rui]> what do you use? Nov 25 21:53:37 use branch=master (probably default) Nov 25 21:53:47 git:///home/me/git/libphone-ui-shr;branch=master Nov 25 21:53:50 <[Rui]> ok... Nov 25 21:54:02 works fine here Nov 25 21:54:04 <[Rui]> but that's what fails: /opt/openmoko/shr-unstable/downloads/git/.home.rms.git.libphone-ui-shr Nov 25 21:54:07 <[Rui]> fg Nov 25 21:54:12 TAsn: sure... building locally now Nov 25 21:54:33 <[Rui]> SRC_URI = "git:///home/rms/git/libphone-ui-shr;branch=master" Nov 25 21:55:01 <[Rui]> it's the only thing I changed Nov 25 21:55:28 [Rui]: use the local-builds.inc... no need to fiddle with SRC_URI in the recipe Nov 25 21:55:45 [Rui]: in shr-unstable/conf/local.conf uncomment the include local-builds.conf Nov 25 21:55:55 and then in local-builds.conf add the following: Nov 25 21:56:02 [Rui]: see local-builds thingie on trac Nov 25 21:56:04 http://trac.shr-project.org/trac/wiki/Getting%20started%20developing%20SHR Nov 25 21:56:06 err... local-builds.inc Nov 25 21:56:13 its described Nov 25 21:56:16 SRC_URI_pn-libphone-ui-shr = "file:///home/mok/src/other/openmoko/libphone-ui-shr" Nov 25 21:56:20 SRCREV_pn-libphone-ui-shr = "LOCAL" Nov 25 21:56:22 SRCPV_pn-libphone-ui-shr = "LOCAL" Nov 25 21:56:23 <[Rui]> I guess I should go back to that... Nov 25 21:56:25 S_pn-libphone-ui-shr = "${WORKDIR}/${PN}" Nov 25 21:56:26 and adjust the path in SRC_URI Nov 25 21:57:33 SHR: 03mok 07libphone-ui-shr * rd2e13b6fa517 10/src/view/call-active-view.c: call-active-view: add callback for volume changes Nov 25 21:57:50 <[Rui]> ok.. Nov 25 21:57:53 mrmoku, great. :) Nov 25 21:58:03 TAsn: building on buildhost now Nov 25 21:58:06 mrmoku, plesae build both on buildhost :) Nov 25 21:58:14 did not try it... but at least it builds :) Nov 25 21:58:14 mrmoku, have you tested it? Nov 25 21:58:16 :P Nov 25 21:58:17 lolz. Nov 25 21:58:26 TAsn: can't ... I'm 2.6.31 user ;) Nov 25 21:58:59 oh :| Nov 25 21:59:03 let me know. Nov 25 21:59:14 yup Nov 25 22:00:40 <[Rui]> oh my, out of date cache... this might take a bit long Nov 25 22:01:26 [Rui]: do you have BB_GIT_CLONE thing removed? Nov 25 22:01:53 (in local.conf) Nov 25 22:02:33 TAsn: we have to decide when we want to tag a 0.0.1 for max_posedon :) Nov 25 22:03:08 mrmoku, :) Nov 25 22:03:16 when we'll get a decent, usable UI Nov 25 22:03:27 I can't wait for years! Nov 25 22:03:27 [Rui]: your mail arrived meanwhile :P Nov 25 22:03:27 (i.e all buttons on screen work) Nov 25 22:03:29 it builds and you touched README. that's a 0.0.1 Nov 25 22:03:33 max_posedon, ^ Nov 25 22:03:53 <[Rui]> mrmoku: hehe Nov 25 22:04:06 playya_: oohhh.. then we have to decide when to tag a 0.99.9999 Nov 25 22:04:23 and I think UI usable now enough, to name it 0.0.1 Nov 25 22:05:02 with a stable enlightnemnet release Nov 25 22:05:20 <[Rui]> BB_GIT_CLONE_FOR_SRCREV ? Nov 25 22:05:38 yep, remove it Nov 25 22:05:59 <[Rui]> ok, let's see where this will lead me to... Nov 25 22:06:00 playya_: max_posedon won't like that ;) Nov 25 22:06:10 [Rui]: to faster parsing :) Nov 25 22:06:29 <[Rui]> I thought it wasn't very recommended as you could get to use older software versions as it wouldn't update Nov 25 22:06:44 <[Rui]> I suppose uncommenting once a month should be enough... Nov 25 22:07:02 [Rui]: no, BB_GIT_CLONE_FOR_SRCREV is evil Nov 25 22:07:35 it will very git repo in OE to be able to do git rev-list Nov 25 22:07:40 <[Rui]> ig Nov 25 22:07:40 s/very/clone every/ Nov 25 22:07:40 mrmoku meant: it will clone every git repo in OE to be able to do git rev-list Nov 25 22:08:13 TAsn: prepare you FR Nov 25 22:08:17 (index building) Nov 25 22:08:44 cool. Nov 25 22:08:55 sec. Nov 25 22:09:07 * spaetz types "opkg update" and hovers over the enter key Nov 25 22:09:26 TAsn: ok now Nov 25 22:09:34 spaetz: finger down Nov 25 22:09:44 ENTER Nov 25 22:10:20 Activate! Nov 25 22:10:26 SHR: 03seba.dos1 07shr-settings * r965d270755a5 10/shr_settings_modules/shr_usb.py: [usb] port to new UsbHost resource interface Nov 25 22:10:27 Make it so! Nov 25 22:10:29 booooom Nov 25 22:10:54 "Running task 5410 of 7454" *wait* Nov 25 22:13:12 mrmoku, ok, trying. Nov 25 22:13:55 39.6% (200.6) : S3C2410 Timer Tick Nov 25 22:13:55 39.5% (200.4) : s3c-mci Nov 25 22:13:55 19.8% (100.2) ksdioirqd/mmc1 : schedule_timeout (process_timeout) Nov 25 22:13:56 btw Nov 25 22:14:11 changed Nov 25 22:15:13 still way too many Nov 25 22:15:25 what is s3c-mci? Nov 25 22:15:32 the SD? Nov 25 22:15:36 yep Nov 25 22:15:47 or wlan Nov 25 22:16:25 prolly uSD is glamo-mci Nov 25 22:17:35 ohh... :) Nov 25 22:17:43 TAsn: so... works? Nov 25 22:17:59 hehe Nov 25 22:18:00 sec Nov 25 22:18:05 forgot to press "i" Nov 25 22:18:10 for it to install the new config. Nov 25 22:18:15 :P Nov 25 22:18:35 ~lart opkg Nov 25 22:18:35 * apt blames opkg for all the evil in the world Nov 25 22:18:47 policykit? Nov 25 22:19:08 that on a device that runs everything as root 8-) Nov 25 22:19:23 lol Nov 25 22:19:53 ok, testing. Nov 25 22:20:02 i thought dos would come up with a script to change that Nov 25 22:20:15 btw Nov 25 22:20:20 fsodeviced crashes on boot Nov 25 22:20:29 but works if I restart it afterwards Nov 25 22:20:39 TAsn: yeah, had that too :| Nov 25 22:20:45 oh, actually also dies Nov 25 22:20:56 TAsn: but that is of no use as you will not have all resources registered properly then Nov 25 22:21:17 ok, now it runs Nov 25 22:21:18 after fsodeviced restart one always needs frameworkd restart too :-( Nov 25 22:21:22 it jsut dies from the init script. Nov 25 22:22:15 no Nov 25 22:22:19 mrmoku, doesn't work. Nov 25 22:22:28 mrmoku, it sets both bars to 0 Nov 25 22:22:30 something happens Nov 25 22:22:38 just not what intended :) Nov 25 22:22:40 spaetz: I'm missing a tool to do that Nov 25 22:23:24 yep Nov 25 22:23:25 SHR: 03seba.dos1 07shr-wizard * rfd3c0f01d542 10/python/shr-wizard.py: python: add initial page Nov 25 22:24:46 spaetz, startig it manually Nov 25 22:24:49 but not from init script Nov 25 22:24:50 works Nov 25 22:24:53 (fsodeviced) Nov 25 22:25:34 SHR: 03seba.dos1 07shr-settings * r9791fa4246a1 10/shr_settings_modules/shr_phoneutils.py: [phoneutils] fix typo Nov 25 22:25:35 yep, but weird, where is the difference Nov 25 22:25:47 TAsn: don't you think it's more about timing/sequence than about whether it's an interactive or a script shell? Nov 25 22:25:58 DocScrutinizer-8, if I start manually Nov 25 22:26:00 using init script Nov 25 22:26:01 it fails Nov 25 22:26:16 all /etc/init.d/fsodevice does is calling /usr/sbin/fsodeviced basically Nov 25 22:26:21 that's weird indeed Nov 25 22:26:40 /usr/bin/... Nov 25 22:26:55 ooohhh here we are ;) Nov 25 22:27:01 fsodeviced moved to /usr/sbin ;) Nov 25 22:27:09 hehe Nov 25 22:27:14 it does not segfault... it just does not get started :P Nov 25 22:27:18 ooh, I got it intuitively right :) Nov 25 22:27:21 lol. Nov 25 22:27:22 :) Nov 25 22:27:27 lol Nov 25 22:27:43 spaetz: you have some quite interesting qualities :P Nov 25 22:27:59 I could feel the bad vibrations from that initscript Nov 25 22:28:18 * spaetz should go to bed soon Nov 25 22:28:44 bad vibrations? You need to listen to some Beach Boys,then. Nov 25 22:29:05 * mwester should get some sleep too... starting to think silly... Nov 25 22:29:18 good suggestion, but not shortly before going to bed. My wife kills me when I hum Beach Boys in bed Nov 25 22:29:48 btw, DocScrutinizer Nov 25 22:29:55 I added that cool thing we talked about (libphone-utils) Nov 25 22:30:25 I can now identify numbers with different prefixes as the same one Nov 25 22:30:30 mwester: does busybox start-stop-daemon need an absolute path to the binary? Nov 25 22:30:34 for instance, here for calling abroad we use either Nov 25 22:30:36 00 Nov 25 22:30:36 or Nov 25 22:30:37 0123 Nov 25 22:30:40 *012 Nov 25 22:30:41 or 013 Nov 25 22:30:42 etc Nov 25 22:30:47 mrmoku: don't think so Nov 25 22:30:52 so I just let libphone-utils to know all are options Nov 25 22:31:03 although it can't hurt Nov 25 22:31:05 mrmoku, I don't think so, but the $PATH used by the initscripts is very limited, IIRC. Nov 25 22:31:05 and it identifies all 00XXXX 012XXXX 013XXXXX Nov 25 22:31:07 as XXXXX Nov 25 22:31:09 as it should :) Nov 25 22:31:50 same goes for area codes/country codes or whatever (for countries that have 2 CCs when one is deprecated) Nov 25 22:31:53 your numbers are all XXXses? Nov 25 22:31:54 and national prefix Nov 25 22:32:02 spaetz, israel is a weird country :) Nov 25 22:32:03 how do you type these many X with a numpad? Nov 25 22:32:09 * spaetz should really go to bed now Nov 25 22:32:11 sorry Nov 25 22:32:12 :) Nov 25 22:32:15 spaetz, spell israeli only cheat :P Nov 25 22:32:18 TAsn: I hope you simply used a regex for the international prefix in configfile Nov 25 22:32:34 nah, I could have Nov 25 22:32:37 but it's too simple not to Nov 25 22:32:40 so I just didn't. Nov 25 22:32:50 TAsn: my sister-in-law learns hebrew at college. I get enough israel lessons by now :) Nov 25 22:32:51 (it's just | delimited text) Nov 25 22:32:57 spaetz, hehe. WHY?! Nov 25 22:33:09 DocScrutinizer-8, so no need for fancy regex Nov 25 22:33:21 aha Nov 25 22:33:35 well, she learns arabic and philosophy too, so she might just be a weird ;) Nov 25 22:33:56 she is. Nov 25 22:34:17 TAsn: but it's enough for going to Yale to get your PhD :) Nov 25 22:34:19 anyway, I am off. I'll lurk a bit from my N800 and then go to sleep Nov 25 22:34:38 spaetz, omg, cool! Nov 25 22:34:41 I know hebrew Nov 25 22:34:43 SHR: 03tom 07libphone-utils * r0d8a016c0c8e 10/src/ (phone-utils.c phone-utils.h.in): added support for multiple values for each field. for example: CC1|CC2|CC3 Nov 25 22:34:46 maybe I can get a phd at yale. Nov 25 22:34:47 :) Nov 25 22:34:56 hehe Nov 25 22:35:11 night Nov 25 22:35:28 night. Nov 25 22:40:07 ok Nov 25 22:40:16 hm.. Nov 25 22:42:35 TAsn: my example script will use regex for any of the CC IP NP AC constants ootb Nov 25 22:42:58 TAsn, if you have enough money, you can get a phd at any of the American ivy-league schools... political connections work pretty well too... Nov 25 22:43:01 you added multichoice support? Nov 25 22:43:07 mwester, :) Nov 25 22:43:21 so if your implementation doesn't support that and needs a tweak to the code it's obviously flawed Nov 25 22:43:37 mwester, are you implying spaetzN800's sister isn't that good? :P Nov 25 22:43:54 Yeah, whatever you do, don't ever make the mistake of thinking that someone with a PHD from Harvard or Yale is actually any smarter than anyone who went to a normal school -- that's just not the case. Nov 25 22:44:31 * spaetzN800 slaps mwester before falling asleep :-) Nov 25 22:44:47 i agree with you though Nov 25 22:45:03 seen enough dummb phd´s. Nov 25 22:45:11 It's sad - I think once a degree from such a university meant something... :)( Nov 25 22:45:19 :( , I mean. Nov 25 22:45:30 heck, i am one and i can´t even spell Nov 25 22:45:43 hehe Nov 25 22:46:09 * mwester has a brother-in-law who has a master's degree -- and cannot hold a full-time job. Nov 25 22:46:35 My sister had the sense to divorce him, finally. Nov 25 22:47:04 .oO( haven't seen spaetz's sister yet... ) Nov 25 22:47:12 hehe Nov 25 22:48:12 spaetzN800: N810 greets N800 Nov 25 22:48:51 fixed fsodeviced in the feed Nov 25 22:49:00 ;-) Nov 25 22:49:17 quadCoreDesktopAmd greets poor N8(0|1)0 :P Nov 25 22:49:37 ah one zero missing Nov 25 22:49:54 :) You need to change your nick, JaMa! Nov 25 22:50:08 doh! Nov 25 22:50:29 /nick mwester_tiny-little-intel-atom-in-a-netbook Nov 25 22:51:45 /nick JaMaPOWERsavingMode :) gnite all Nov 25 22:52:04 SHR: 03tom 07libphone-ui * rab428ea2373c 10/src/phoneui-utils-sound.c: Added a check before printing a useless thing that fills tho logs Nov 25 22:52:18 <[Rui]> no way... it compiled at first? :) (of course it just had some pruning down in order to try to fit...) Nov 25 22:52:20 ./nick WonkaGTA02A5 Nov 25 22:52:56 Cool. What client, Wonka ? Nov 25 22:53:36 openssh. screen and irssi remote. Nov 25 22:54:15 but I think one could run irssi directly on the FR Nov 25 22:57:29 Wonka: xchat Nov 25 22:57:46 I actively dislike xchat Nov 25 22:57:54 mrmoku, alive? Nov 25 22:57:57 or pidgin - whic I didn't like as much Nov 25 22:58:13 TAsn: no Nov 25 22:58:21 :-D Nov 25 22:58:32 RIP Nov 25 22:58:47 hehe Nov 25 22:58:57 TAsn: brb Nov 25 22:59:06 or 1 mrmoku (zombie) Nov 25 22:59:34 DocScrutinizer: I am planning to write my own IRC/IM client, probably in python. will be some client-server-stuff. server on some host with constant internet access, clients for X, ncurses, ... Nov 25 22:59:40 which btw is impossible ;-) Nov 25 23:00:20 as 1 is the only process without real PPID Nov 25 23:00:21 mrmoku, :) Nov 25 23:01:34 Wonka: sounds like it's exactly what I'd like most :-) Nov 25 23:02:13 DocScrutinizer: ack... I just need to find some time... Nov 25 23:02:21 Wonka: could even *filter* the annoying join/quit msgs always resuming for no good reason Nov 25 23:02:57 DocScrutinizer: ack. Nov 25 23:02:59 would strech my IRC online time on FR beyond 16h easily Nov 25 23:03:27 possibly hitting the magic 24h barrier Nov 25 23:03:37 irssi in screen via ssh produces too much traffic Nov 25 23:04:15 when you switch back to a channel you've been to before, most of the stuff dragged though the connections has been though before. Nov 25 23:04:49 well with current traffic on #openmoko I could also hit the 24h limit I guess :-/ Nov 25 23:05:06 Wonka: ack Nov 25 23:06:36 waah. what do I need nano for, in shr-lite-image? Nov 25 23:07:02 i'd rather take vim (which is there anyway, afair) Nov 25 23:08:50 TAsn: tell me :) Nov 25 23:08:53 but fast Nov 25 23:08:59 mrmoku, I have nothing to say. Nov 25 23:08:59 :) Nov 25 23:09:02 because soon I will RIP Nov 25 23:09:03 ahh Nov 25 23:09:14 any idea about the sliders? Nov 25 23:09:34 SHR: 03tom 07libphone-ui * rf3f005417baf 10/src/phoneui-utils-sound.c: fixed a typo in the last commit and added more debug info Nov 25 23:10:55 mrmoku, no :| Nov 25 23:11:00 for some reason they are set to 0 Nov 25 23:11:02 debugging Nov 25 23:11:32 ok... will let you debug n peace then and rest in peace myself :) Nov 25 23:11:40 gnight all Nov 25 23:12:10 night. Nov 25 23:15:42 SHR: 03tom 07libphone-ui-shr * r893177f69fd7 10/src/view/message-list-view.c: commented out a very annoying useless debug message Nov 25 23:43:15 darn, opimd is borked once again. no more storing of contacts. might reboot to rule out it's due to the phonelog segfault of 3h ago Nov 25 23:45:02 <[Rui]> gnigh Nov 25 23:45:10 aah, and of course phonelog still fails to log outbound calls (or last calls, or random calls, dunno) Nov 25 23:45:19 rui: n8 Nov 25 23:46:04 Hi, I'm trying to build an shr image, as explained on http://trac.shr-project.org/trac/wiki/Building%20SHR , but it fails when I do "make image Nov 25 23:46:04 " Nov 25 23:46:34 it tells me http://pastebin.com/fa157b5a Nov 25 23:46:58 I've just done what is said on the page at the moment Nov 25 23:48:52 don't worry. Even devels are doing a happy dance for every successfull build Nov 25 23:49:17 :) Nov 25 23:50:15 piti: "Please install following missing utilities: help2man,texi2html,makeinfo,gawk" do you have them? Nov 25 23:51:59 ooops, I guess I had to check that before asking Nov 25 23:57:56 so, the message about the dependencies have gone, but it still fails, and "Using dash as /bin/sh causes various subtle build problems, please use bash instead." is strange because I'm sure that I'm on bash and not dash Nov 25 23:59:19 pini, this message isn't about your shell Nov 25 23:59:30 its about /bin/sh symlink Nov 26 00:00:43 right Nov 26 00:03:11 mrmoku|away, still no idea :( Nov 26 00:03:29 opimd-contacts "New" foo "Save" definitely NoOp Nov 26 00:03:40 :-( Nov 26 00:05:30 err "Edit" foo "Save" Nov 26 01:44:16 SHR: 03tom 07phoneuid * rfdc6be45af06 10/ (5 files in 3 dirs): Added config file and support for different log levels (set from config). ATM default log level is DEBUG, should probably change to INFO Nov 26 01:49:20 SHR: 03tom 07phoneuid * r3f6a676bb6e4 10/src/phoneuid.c: changed a g_message to a g_warning Nov 26 01:53:50 DocScrutinizer-8: apperently it's a problem for all pwms connected to the second prescaler Nov 26 01:54:29 larsc: yes, as second prescaler is messed up by FIQ Nov 26 01:55:29 counters 234 share same prescaler and devider Nov 26 01:55:44 DocScrutinizer: the fiq doesn't do anythink since it's code is commented out Nov 26 01:56:06 hmm, strange then, probably a bug in the driver? Nov 26 01:56:07 wasn't there a tool somewhere which dumps all pwm registers? Nov 26 01:56:17 dunno Nov 26 01:56:44 I know of s3c2442-gpio Nov 26 01:57:11 but this tool probably can't access the PWM registers Nov 26 02:03:15 DocScrutinizer: /proc/kmem can access the full memory space, so likely you can look at the PWM settings that way Nov 26 02:03:54 yup, but the 2442-gpio tool doesn't have the syntax :-P Nov 26 02:05:09 of course that's easy to patch Nov 26 02:05:32 DocScrutinizer-8: heh? maybe I'm not understanding what you're doing. Nov 26 02:05:52 DocScrutinizer: you can dd /proc/kmem to read arbitrary memory. no special utility needed Nov 26 02:05:54 aah, sorry got you wrong Nov 26 02:07:22 tbh I have no idea if S3C registers are memory mapped or special IO opcode Nov 26 02:07:39 DocScrutinizer: I've used http://secretsauce.net:5050/peek.c to look at timer settings before. This are shared with pwm Nov 26 02:08:26 budfive_: these are exactly the registers we're interested in Nov 26 02:08:39 yeah. use that utility. Nov 26 02:08:49 it dumps the whole page onto stdout Nov 26 02:09:24 so if you ask for the memory at 0x50000000, you'll get all the registers Nov 26 02:09:25 I seem to remeber a tool named peek or poke of werner as well Nov 26 02:09:31 0x51000000, I mean Nov 26 02:09:53 ok, so regs are memmapped Nov 26 02:09:59 DocScrutinizer: I thought I wrote this thing, but it was a while back, so maybe not... Nov 26 02:10:46 well, you won't stop werner from doing it again anyway ;-) Nov 26 02:10:59 wpwrak: any comments? Nov 26 02:13:59 SHR: 03tom 07phoneuid * rb70ff1c871e2 10/src/phoneuid.c: added fallback if can't open log Nov 26 02:14:49 SHR: 03tom 07phoneuid * rc9e604ae97fb 10/src/phoneuid.c: fixed a debug message Nov 26 02:15:30 http://svn.openmoko.org/developers/werner/poke/ Nov 26 02:15:49 larsc: ^^^ Nov 26 02:16:25 DocScrutinizer: poke also takes care of access sizes. plus, it can also change things. Nov 26 02:16:38 fprintf(stderr, "usage: %s [-8|-16|-32] hex_address [value]\n", name) Nov 26 02:16:43 :-)) Nov 26 02:16:45 yup :) Nov 26 02:18:17 how's core? Nov 26 02:18:44 access sizes matter if you have, say, a FIFO. each read/write has a side effect. so if it's > 8 bits wide, you want to make sure you don't access it byte-wise. (actually, i don't know what exactly happens in such cases, but i would be a bit surprised if it didn't cause problems) Nov 26 02:18:45 duh, missed on the chan Nov 26 02:19:19 ah yes, still got a few mails to answer. **** ENDING LOGGING AT Thu Nov 26 02:59:56 2009