**** BEGIN LOGGING AT Wed Jan 07 02:59:57 2009 Jan 07 04:26:43 fso unstable feeds are syncing now for gta01, gta02 and a780. Jan 07 04:56:32 anyone familiar with callbacks for edje events in c? i have one that is not working, even thou it is almost the same as one that does work except the signal, and function name Jan 07 05:01:40 nevermind, I just can't cound to 4, aparently :-P Jan 07 05:16:23 hahahah Jan 07 05:55:55 any openmoko devs here who know the git URL I need to specify to push to git.openmoko.org? my ssh key is on the server apparently, I just need to know the push URL, cause git push ssh://git@git.openmoko.org/git/openmoko.git isn't working. Jan 07 05:56:14 (the commit is an update to conf/checksums.ini) Jan 07 06:01:37 never mind, google found it for me. Jan 07 07:31:43 when I have the debug board connected, can I still flash u-boot to NAND flash? or do I need to directly connect the FR to the PC to use it with dfu-util? Jan 07 07:41:40 Ainulindale? Jan 07 07:45:16 mbuf, not advisable Jan 07 07:45:35 hey Sargun Jan 07 07:45:39 Sargun, which one? Jan 07 07:45:44 quickdev, hey quickdev Jan 07 07:46:18 Sargun, are you going to try to fix #1024 with a resistor? Jan 07 07:47:45 quickdev, My SMT guy is out until the 24th Jan 07 07:48:01 It's a pity :) Jan 07 07:49:37 quickdev: hey, morning - no school today? Jan 07 07:50:15 mrmoku, I'd be lucky...first school day this year...but it starts a in half an hour...thus I have to go now ;) Jan 07 07:50:28 have fun ;) Jan 07 07:50:40 thanks, have a nice day Jan 07 07:50:41 * mrmoku is going to office now... Jan 07 07:54:22 Sargun: what's the resistor fix for #1024? Jan 07 07:54:35 Sargun: I see #1024 all the time and can solder smt Jan 07 07:54:53 budfive, oh, holy crap Jan 07 07:56:11 budfive, Did you read the ML post from dieter Jan 07 07:56:34 Sargun: I tell you that Joerg told that soldering 180k in parallel to R1050 (which is believed to be 220k) to get 99k. And that the can can be lifted easily but with care. And that it would help OM. Correct? ;) Jan 07 07:56:36 err Jan 07 07:56:42 what? no. I unsubscribed from community and support because of too much whining. maybe that was foolish Jan 07 07:56:53 Jan 06 13:54:44 DocScrutinizrFlu: So when i see him next time, i should tell him that Joerg told that soldering 180k in parallel to R1050 (which is believed to be 220k) to get 99k. And that the can can be lifted easily but with care. And that it would help OM. Correct? ;) Jan 07 07:56:54 checking. thanks Jan 07 07:58:14 budfive, http://lists.openmoko.org/pipermail/hardware/2008-December/000870.html Jan 07 08:00:22 i see. thanks. i'll try that tonight if a have a properly-sized resistor lying around Jan 07 08:01:36 budfive, excellent. Jan 07 08:01:41 where are you located Jan 07 08:01:44 LA Jan 07 08:02:09 is R1050 a resistor that's missing from the schematics because of the calypso not being completely open? Jan 07 08:03:25 and why is R1050 "believed" to be 220k? Nobody knows for sure? Jan 07 08:07:42 and does anybody know if dieter tried this himself? Jan 07 08:09:27 Dieter isn't able to reproduce #1024 Jan 07 08:09:35 And, it is 220k, dieter is an OM employee Jan 07 08:09:49 I would bug docScrutinizer (Joerg) Jan 07 08:10:35 ok. awesome. i'll try this tonight Jan 07 08:53:31 anybody opened the gta02 fully, exposing the component side of the circuit board? Jan 07 08:57:15 budfive: R1050 is a suspect we (om) want to investigate. It's way from considered to be a fix. Jan 07 08:57:56 yeah, yeah. i know. won't blame you if i break something. :) Jan 07 08:58:10 k :) Jan 07 08:58:26 does the case open the same way as gta01? can't seem to get the board out Jan 07 08:58:34 yup Jan 07 08:58:51 i should flex the outer case to free the headphone jack, and it all should pop out? Jan 07 08:59:03 yup Jan 07 08:59:05 there are a few clips that aren't mentioned on the wiki Jan 07 08:59:38 budfive, Do you have a resistor? Jan 07 09:00:01 i have 100k Jan 07 09:00:09 will take out the original one Jan 07 09:00:11 push slightly inside battery compartment with one finger Jan 07 09:00:54 aha. something just moved. thank you Jan 07 09:00:58 heyho Jan 07 09:01:04 joerg_42, do you know if GTA02v06 is going to become the primary unit? Jan 07 09:01:12 er, v7 Jan 07 09:02:12 budfive: we don't know why this is 220k (heritage). so maybe 100k won't work at all. Could you also scope the osc-signal? Jan 07 09:02:37 Sargun: whats "primary"? Jan 07 09:03:04 Like, when OM sends more orders to manufacturing, will it be with buzz fix Jan 07 09:03:04 joerg_42: it should be 32khz? Jan 07 09:03:13 yup Jan 07 09:03:24 joerg_42: ok. i'll do that and will take pictures Jan 07 09:03:37 and start cleanly on at%sleep=4 Jan 07 09:03:38 joerg_42: luckily #1024 happens for me with the gta02 sitting on my desk Jan 07 09:03:57 joerg_42: just tested that. it recamps about once per minute Jan 07 09:04:12 great, errr Jan 07 09:04:52 :) Jan 07 09:05:49 budfive: please use >10MR for signal and gnd of probe, not to detune osc Jan 07 09:06:34 budfive: no capacitive load!! Jan 07 09:07:02 10pF is already too much Jan 07 09:07:24 joerg_42: ok. thanks. I'm not sure if my probe is very good, but i'll check. looking at the can right now Jan 07 09:07:44 joerg_42: thoughts on opening it? it looks like it just pops out, but there's stuff on top of it. wifi module maybe Jan 07 09:07:53 pop off wlan Jan 07 09:08:08 first Jan 07 09:08:25 joerg_42: thanks. just trying to not kill anything unnecessarily Jan 07 09:08:29 lifting board connector Jan 07 09:08:39 joerg_42, feeling any better Jan 07 09:08:57 The eagle has landed. Jan 07 09:09:07 just woke, lets wait n see Jan 07 09:12:21 wait a minute. I'll try to talk werner into creating an excerpt of the schem Jan 07 09:26:11 [10:21] doc: like this ? http://people.openmoko.org/werner/r1050.png Jan 07 09:26:45 DocScrutinzrFlu: thanks. Jan 07 09:26:56 DocScrutinzrFlu: good timing, actually. i just got the can off, and see no way to get at r1050 Jan 07 09:27:08 budfive: welcome. your help appreciated Jan 07 09:27:13 DocScrutinzrFlu: it's sitting under the lip of the metal housing the can sat on Jan 07 09:27:40 DocScrutinzrFlu: ideas? I don't see any way to get at it without cutting the metal or desoldering the whole thing, neither of which i want to do... Jan 07 09:27:53 cut out a piece of the lip Jan 07 09:28:06 DocScrutinzrFlu: ok. :) Jan 07 09:28:32 hi adventurous #1024 hunters ;-) Jan 07 09:29:25 hi werner :-) thanks for the quick help Jan 07 09:32:01 doc: no problem. not sure what can be done about that part, though. well, changing R1050 is one option ... Jan 07 09:32:21 doc: would be fun as a sequel to the buzz rework (-:C Jan 07 09:32:57 dieter wants to look into the tuning code forr 32k osc Jan 07 09:33:10 on some level i hope that this doesn't work, since this is way more involved than the buzz fix Jan 07 09:33:28 wpwrak: first we need to understand the problem Jan 07 09:34:05 budfive: it's just one small component ;-)) Jan 07 09:34:53 maybe we see driver-strength settings etc inside calypso Jan 07 09:35:36 so in the end we might find a sw-fix for a hw-flaw Jan 07 09:35:44 budfive: of course, it's just under the shield, so telekinetic skills are useful for soldering :) Jan 07 09:35:58 as soon as we definitely know where to tackle Jan 07 09:36:14 doc: a configurable drive strength would be almost too much to wish for ... Jan 07 09:36:45 so you want me to scope across c1052? Jan 07 09:36:49 doc: but that would have been my next question as well :) Jan 07 09:36:57 i'm almost sure it got sth like that, for power minimization Jan 07 09:37:27 budfive: better across C1050 - less risk of messing up the signal Jan 07 09:37:33 grr, closed the pic again Jan 07 09:37:38 mompl Jan 07 09:39:05 yup C1050 Jan 07 09:41:19 check for dropouts, bursts or other anomality Jan 07 09:41:56 ok. my probe is 10Mohm, 14pF. Is that sufficient? Jan 07 09:42:01 budfive: a good way to find clock troubles is to set the trigger as many cycles as possible before the screen. that way, any speed changes cause a significant phase drift. Jan 07 09:42:17 si th14pf in parallel with 22pf may be too much... yes? Jan 07 09:42:26 analog electronics isn't my strongest suit. Jan 07 09:43:15 i already recommended R probe 2 x >10M Jan 07 09:43:22 wpwrak: thanks. that makes sense Jan 07 09:43:57 we aren't very interested in <5% abs voltage value Jan 07 09:44:10 budfive: yes, putting a scope on c1052 would almost certainly confuse the loading of the crystal, which can be good or bad. better put the scope on c1050. it's also a little easier to reach. Jan 07 09:44:23 so you can possibly create such probe urself Jan 07 09:45:01 doc: 10M R probe ? wow ... I gotta try this :) Jan 07 09:45:17 are you more concerned about the C loading or the R loading? Jan 07 09:45:42 C Jan 07 09:55:13 notice the stupid meaningles offsheet symbols [2, ...] Jan 07 09:55:51 there's sth fishy about this R Jan 07 09:56:00 @fso-team: i see you're opening the libc via a ctypes-call - for which reason you need functions of the libc? Jan 07 09:56:04 is the shield grounded? should I clip my ground there, or would that add too much parasitic capacitance? how does one attach the probe here? sorry about the questions. I don't often deal with very analog electronics Jan 07 09:56:27 DocScrutinzrFlu: i wanted to ask just what the meaning of those ;) But you're saying it's meaningless :) Jan 07 09:57:17 budfive: for 32k can gnd should be ok Jan 07 09:57:38 DocScrutinzrFlu: Why there's no answer from EE who made a decision about this R? Did he quit OM or what? Jan 07 09:57:40 Hey DocScrutinzrFlu Jan 07 09:57:50 I'm about to contact an ircop to throw you from here Jan 07 09:57:54 Go to sleep Jan 07 09:58:15 PaulFertser: this is heritage from FIC Jan 07 09:59:16 And now FIC dislikes OM enough to make access to its engineers too hard to be practical? Jan 07 09:59:27 me away for some relax, before ainu kicks me Jan 07 09:59:33 Thanks :-) Jan 07 09:59:40 Go read a book :-) Jan 07 09:59:51 Or listen to some calm music :) Jan 07 10:02:46 just one more worfd to PF: chinese way to select a component isn't completely understood yet ;-) Jan 07 10:05:03 :)))) Jan 07 10:05:31 see audio-C ;-) Jan 07 10:06:41 I've talked to one chineese manager and he told me that engineers work very hard there (harder than managers). Long work day, too few holidays, etc. No wonder there're some small mistakes :) Jan 07 10:06:50 PaulFertser: hey by the way Jan 07 10:07:01 I see you helping for gsm0710muxd, why don't you help for SHR? :-p Jan 07 10:09:15 Ainulindale: i know too little about end-user-friendly programming. And i think i sometimes can do something useful on the lower level. Moreover, i don't have a stable lower layer (FSO) on my FR yet (hopefully it'll be addressed by the FSO team), so i'm afraid to add another layer upon it. Jan 07 10:09:32 PaulFertser: what do you think I'm doing? :-) Jan 07 10:09:39 I don't like UIs and such at all Jan 07 10:09:41 Can't work with that Jan 07 10:09:48 Besides I have no taste Jan 07 10:09:56 But we need tons of stuff for this or that on SHR Jan 07 10:10:00 Even ideas are good to take :-) Jan 07 10:10:08 And I see you have some, hence my asking Jan 07 10:12:12 Ainulindale: i'm too lazy, i must admit. I haven't done all the job i should on #31 yet (i'll do it in an hour, i hope). I need to make a patch to ti_calypso/modem documentation, close the bug, open two another and so on... Not quiet timely response from me this time :( Jan 07 10:13:34 heh :-) Jan 07 10:13:48 ~. Jan 07 10:13:49 ~. Jan 07 10:15:12 paul: that gsm thingy has a very long and twisted history. the design has actually been inherited at the gerber level from some feature phones. so it may be really hard to find the source of that component change Jan 07 10:15:41 btw, does anyone with #1024 also have a gta01 ? Jan 07 10:17:37 wpwrak: great :-) I see I may take an hour off Jan 07 10:18:24 "The design inherited at the gerber level" -- that's tough :) otoh it's understood: if RF works, don't touch it :) Jan 07 10:18:39 yup Jan 07 10:19:14 certif issue ;) Jan 07 10:22:38 mirko-paroli: hey :) Jan 07 10:22:50 mirko : hey Jan 07 10:32:08 Well, I have a big post for the blog Jan 07 10:34:36 doc: hmm, just made a 2M resistive probe. i can barly see a 10Vpp signal with it :) Jan 07 10:35:09 doc: (break) yeah, better do a complete resume than spend weeks in zombie state :) Jan 07 10:37:28 wpwrak: seems i need a more clear head. why can't you see 10Vpp with a 2meg probe? what's your input impedance at Y-amp??? Jan 07 10:39:40 doc: 50R Jan 07 10:39:58 mpfff Jan 07 10:40:31 wpwrak: so i'm running my gta02, as it's exhibiting #1024, and i'm probing it Jan 07 10:40:32 ah nice .. with 1M, things get better Jan 07 10:40:46 it seems pretty close to 32khz Jan 07 10:40:56 how much drift should i be looking for? Jan 07 10:41:21 look for instabilities Jan 07 10:41:29 the signal is .8V peak-peak Jan 07 10:41:33 dropouts Jan 07 10:42:35 maybe use a .2sec/div sweep Jan 07 10:44:03 budfive: how is startup of osc when switching sleep2/4? Jan 07 10:46:00 hold on. i'm not as pro as i'd like to be. right now i'm just holding the probe to see the signal, which isn't good enough to do what you ask. should i solder a little wire to the probe point, and scope that, so i can go hands-free? Jan 07 10:46:33 good plan Jan 07 10:47:44 earlier when you were talking about detuning the oscillator, was that referring to something going out of calibration permanently, or just overloading the oscillator while the probe was attached? Jan 07 10:48:03 i think i may be too careful here Jan 07 10:48:30 Hey yoyo o/ Jan 07 10:48:41 np, thats just due to probe load Jan 07 10:48:58 budfive: reversible :) Jan 07 10:49:29 DocScrutinzrFlu: I thought you were supposed to sleep! =) Jan 07 10:49:43 oh. ok. i thought you were talking about something i've never heard of. i was feeling a bit like a dufus. :) Now go sleep! Jan 07 10:50:01 ive slept toooo long all the year now Jan 07 10:50:09 just relax Jan 07 10:50:37 hi Jan 07 10:52:33 budfive: those osc are tuned to very low energy. any load change my stop or mess oscilation Jan 07 10:53:07 but wont detroy anything permanently Jan 07 10:53:57 doc: or actually make it better. e.g., cypress recommend an asymmetric loading of their crystals to increase oscillator stability. crystals are just weird :) Jan 07 10:54:20 yup, damn true Jan 07 10:56:38 i tried to use a resonator for at812051 clock once. made a nice cpu-singlestep, as i had to flick the board/resonator with fingernail to trigger a few clocks :D Jan 07 10:56:54 doc: ;-))) Jan 07 10:57:07 doc: next time, try to do this intentionally :) Jan 07 10:57:45 pfoooh, that difficult for sure Jan 07 10:58:40 doc: i think i would accept defeat before even trying :) Jan 07 11:12:10 hello guys Jan 07 11:12:37 http://wiki.openmoko.org/wiki/Xoo should emulate om and @ the same time i can develop on , right Jan 07 11:27:10 Sargun: (v7) yes, we're going to manuf A7 with bigC right now afaik Jan 07 11:32:47 DocScrutinzrFlu: the oscillator should be 32768hz or 32000 hz? Jan 07 11:33:10 prolly 768 Jan 07 11:34:10 ok. i'm seeing something very close to that. no obvious dropouts Jan 07 11:34:12 budfive: though absolute exactness isnt the issue. stability is Jan 07 11:34:29 right. at%sleep=4 and =2 doesn't seem to change it Jan 07 11:34:40 i have no idea if it misses a beat somewhere, though Jan 07 11:35:41 hmm, strange. i'd expect the osc to switch down at sleep2. but dieter knows more about that Jan 07 11:36:04 well, i have everything running right now with my hands free and can run some experiments Jan 07 11:36:15 if you tell me what you want to test Jan 07 11:36:46 if it's stable and no glitches or anything, i'd NOT expect R1050 change to fix #1024 Jan 07 11:37:10 #1024 isn't happening this second, though. It is switching back and forth between two cells, but never dropping registration Jan 07 11:37:26 budfive: make sure you still see #1024 Jan 07 11:37:39 yeah. waiting for it. it'll come. give it a few minutes Jan 07 11:37:56 how can we be sure the oscillator is stable? i'm only seeing a very small chunk on my scope Jan 07 11:38:29 there could be some very short-lived dropout that causes the issue, and i can easily miss it on the scope, right? Jan 07 11:38:40 there. it just dropped Jan 07 11:39:02 budfive: switch off trigger and sync on screen with timebase finetune Jan 07 11:39:38 ok. so how would i see a short-lived dropout? sudden phase shift? Jan 07 11:40:40 budfive: watch it for an hour and meditate about it until you feel! what it's doing. (phase) yup Jan 07 11:41:11 allright. let me figure out how to do that with this scope Jan 07 11:41:20 it just dropped twice more. at least THAT is repeatable Jan 07 11:41:35 switch off trigger. just freerun and stabilize by x-finetune at scope Jan 07 11:42:22 maybe you can setup some trigger to catch and freeze a drop? Jan 07 11:43:27 i don't know what you mean by x-finetune. If i switch off triggering, the phase jumps around, which makes sense to me Jan 07 11:43:35 what's the adjustment i'm looking for called? Jan 07 11:44:11 time/div finetune? Jan 07 11:46:15 if you have a steady picture with trigger set to freerun, you don't need to adjust it. but pict should start to "move" when slightly tuning this dial Jan 07 11:46:51 well, we already spotted dropouts Jan 07 11:47:22 so we are very interested in start and stop of these now Jan 07 11:48:15 when triggering, it's solid Jan 07 11:48:21 oh! i just found something Jan 07 11:48:24 are there kinky clock-cycles during start/stop? Jan 07 11:48:29 just as it dropped Jan 07 11:48:36 ok. hold on. let me investigate Jan 07 11:49:20 how long is duration of drop? is it always same? Jan 07 11:51:32 guys anyone here know how to c code and have some time free ? Jan 07 11:51:37 try to document start of dropout signal form. duration. end of drop signal form. by screenshot Jan 07 11:52:00 budfive: for a few dropouts Jan 07 11:52:54 budfive: then next step would be 220k to 100k and do same again, then compare Jan 07 11:54:28 i'm trying to recreate what i saw. you are saying you already saw dropouts? under what circumstances? Jan 07 11:55:26 budfive: expected results: a) 100 doesnt change anything, or b) 100 fixes dropout Jan 07 11:55:46 budfive: no we didnt see any here yet Jan 07 11:55:59 didnt do this test yet Jan 07 11:56:07 budfive: You see generation dropout or network registration drop? You seem to be using the same word "drop" for different things ;) Jan 07 11:56:25 sorry about that. I see network drops about once per minute Jan 07 11:56:47 i saw something funky happen with the clock signal once, and am trying to reproduce that Jan 07 11:56:58 nothing conclusive yet Jan 07 11:57:00 but i'm looking Jan 07 11:57:14 budfive: that's cool :) Jan 07 12:04:02 a seed of an idea (for the HC-electronics ;): feed the 32k square to antenna in VHF in of a classic obsolete TV Jan 07 12:05:28 tune TV til U see a vertical white or black bar Jan 07 12:05:44 :) no tv here Jan 07 12:06:35 very convenient for checking phase jitters , dropouts etc Jan 07 12:06:44 budfive: are you russian? ;) Jan 07 12:06:58 heh? what makes you think that? Jan 07 12:07:12 budfive: ircname : dima Jan 07 12:07:38 oh yeah. lived in moscow long ago. usa now Jan 07 12:08:51 eeek, ntsc then Jan 07 12:09:07 doc: hmm, your high-Z probe has really interesting characteristics Jan 07 12:10:09 wpwrak: doesnt matter ;-) generates distinguishable signal. thats all that counts Jan 07 12:11:12 doc: let me get you a frequency sweep ... Jan 07 12:12:27 wpwrak: for c1050 any 1M/50p prope should possibly do Jan 07 12:12:56 doc: http://people.openmoko.org/werner/probe-2M.png Jan 07 12:13:09 wpwrak: nah, please no useles freq sweep for a R type probe Jan 07 12:13:32 doc: input is 1Vpp from 100Hz to 100kHz logarithmic sweep. probe is 2M with 1M termination Jan 07 12:13:40 gnnhhh, i feel soooo tired Jan 07 12:13:56 doc: ch2 rising is 100Hz, falling is 32.768KHz Jan 07 12:14:05 doc: just look at it. it's fun ;-) Jan 07 12:14:57 yeah, exactly what is expected for this kinda probe Jan 07 12:16:07 doc: didn't expect it to fall off so quickly. but yeah, a few pF matter when dealing with MOhms ... Jan 07 12:16:52 thats why i was concerned about the usual 50pF of a standard probe Jan 07 12:17:06 well sometimes 5pF Jan 07 12:17:52 keep that away from osc anyway, accept it spoiling picture instead Jan 07 12:18:47 so -> insert some meg to signal and gnd, near to osc, into probe path Jan 07 12:21:36 wpwrak: wasnt it U who had a whole blog about using R probes somewhere on your public appearance Jan 07 12:21:37 ? Jan 07 12:26:30 so here's what i'm seeing. The clock seems stable. i can't get it to predictably hiccup in anyway. I AM currently trying to trigger with a long delay to catch a phase shift. There is none. Jan 07 12:26:59 one interesting thing, is that I lose network registration aboug 20seconds after i change the at%sleep setting even if i'm setting it to 2 Jan 07 12:27:02 this is very consistent Jan 07 12:28:22 huh, so what you tell is "sleep2 wont fix #1024"? Jan 07 12:28:52 well, no. with at%sleep=2, i see one dropout about 20sec after i change the setting and then it's fine. with sleep4 it keeps happening Jan 07 12:29:05 dropout of the registration that is, not of the clock Jan 07 12:29:24 that's a GREAT clue Jan 07 12:29:54 let me time it to see if it's some round number of seconds after the setting change Jan 07 12:30:17 yeah, please Jan 07 12:34:38 Kensan: ping Jan 07 12:41:14 dcordes: hi Jan 07 12:41:27 hi mrmoku Jan 07 12:41:41 somebody has the Kensan website url? Jan 07 12:42:04 dcordes: sorry, no Jan 07 12:42:26 do you have some url explaining how to get a linux image onto a HTC phone? Jan 07 12:42:51 I deleted his shr build with frameworkd-devel for armv6-novfp Jan 07 12:42:54 ~lart dcordes Jan 07 12:42:54 * bzzbot resizes dcordes's terminal to 40x24 Jan 07 12:43:07 mrmoku: depends on the HTC phone Jan 07 12:43:15 dcordes: raphael Jan 07 12:43:26 mrmoku: I'm planning to make a nice website Jan 07 12:43:48 dcordes: ok, ping me if there is something to look at :D Jan 07 12:43:49 mrmoku: set your local.conf to htcraphael. but wait until I updated srcrevs and defconfig for the kernel Jan 07 12:44:00 ok Jan 07 12:44:07 you can boot from sd now :) Jan 07 12:44:28 mrmoku: by the way Jan 07 12:44:37 I'm talking about this HTC stuff on the next newsletter Jan 07 12:44:42 (draft, I'm waiting for bumbl) Jan 07 12:44:49 it's best to have a 1. vfat partition with the bootloader and kernel image and the rootfs on some linux filesystem Jan 07 12:45:03 mrmoku: I use 1gb vfat and 3gb ext2 partition card Jan 07 12:45:37 dcordes: 1gb for bootloader and kernel looks like enough free space ;) Jan 07 12:45:38 Ainulindale: what newesletter? I can feed you info if you like. Jan 07 12:45:55 mrmoku: well you can keep your favourite stuff to use in wince as well Jan 07 12:46:12 dcordes: so it is dual-boot? can boot to wince or linux? Jan 07 12:46:46 mrmoku: if you put it that way yes. the linux bootloader is a wince binary Jan 07 12:46:49 dcordes: the SHR newsletter, which will go on planet and ML Jan 07 12:46:50 doc: (r probes) i did some experiments with them, yes. the ones i normally use are much lower impedance, though Jan 07 12:46:56 dcordes: I'd like info yes :-) Jan 07 12:47:03 Ainulindale: ok we can talk Jan 07 12:47:14 DocScrutinzrFlu: let me revise my theory. It now seems to me that with at%sleep=4, then phone comes in and out of a state where it's prone to dropping registration. at%sleep=2 doesn't seem to pull the phone out of this state, rather it prevents it from entering that state. This is why i can still see registration dropouts with at%sleep=2. I'm saying this because I only seem to see dropouts in sleep=2 if i set sleep=2 shortly aft Jan 07 12:47:15 dropout. If registration is stable and i switch to at%sleep=2, dropouts don't seem to occur Jan 07 12:47:27 dcordes: currently I just wrote down => Jan 07 12:47:28 "Last but not least, we are seriously thinking, as soon as the first release will hit the ground, to develop software for products other than the OpenMoko devices. FSO is a transversal initiative, and there was some talks about being able to produce a full and working distro for, say, the HTC Universal. This won't distract us from our first priority, don't worry: testing on OM devices :-)" Jan 07 12:47:42 mrmoku: http://www.handhelds.org/moin/moin.cgi/HaRET Jan 07 12:48:22 i have debug board v3 connected to FR, booted to NAND flash, and on opening openocd, I get these errors, http://pastebin.ca/1302695; why is it so? i checked jtag cable connection and it is fine Jan 07 12:49:17 DocScrutinzrFlu: another possible clue: In my testing, if I at%sleep=2 when registration is stable, calypso always responds with "EXT: I" and OK. If I do this after a dropout, it only says OK. And then keeps dropping registration, at times more than once Jan 07 12:50:02 dcordes: thanks :-) Jan 07 12:50:05 if i do at%sleep=2 a second time, it does say EXT: I, but still drops registration Jan 07 12:50:23 Ainulindale: universal is a rather old pxa device. it's better to make propaganda with the newer devices. HTC Touch Pro, HTC Touch HD, HTC Diamond, Sony Ericsson Xperia X1. well I don't know about SHR but the msm devices will be maintained in OE (and frameworkd) and support and functionality will be expanded Jan 07 12:51:03 dcordes: SHR is based on frameworkd, so as long as you're able to help us to set up the buildhost, and potentially submit patches, it's good to go Jan 07 12:51:05 I'm looking forward to the day when shr goes oe. Jan 07 12:51:16 dcordes: Well that day is not today Jan 07 12:51:26 I won't commit to OE until release Jan 07 12:51:34 Ainulindale: I already mentioned this: only problem is your frameworkd-glib not supporting frameworkd-devel Jan 07 12:51:36 And we patch OE to make things work Jan 07 12:51:41 dcordes: Yeah I know Jan 07 12:51:48 If you want to help, please do Jan 07 12:51:53 I already intended to branch it Jan 07 12:52:02 branch what? Jan 07 12:52:06 libframeworkd-glib Jan 07 12:52:08 budfive: great job. could you create a summary of all this, and post it to a matching thread at hw-ml, please? so it wont get lost, and dieter egts to know about it. also it seems to me we excluded R1050 from suspects list Jan 07 12:52:08 into lfg-devel Jan 07 12:52:13 lfg ? Jan 07 12:52:15 libframeworkd-glib Jan 07 12:52:28 (far too long to write, I have an alias :-) ) Jan 07 12:52:29 lol Jan 07 12:52:33 ok nvm me Jan 07 12:52:52 yea that's a good idea. give me the project url when it's done Jan 07 12:53:12 well it's already there on git.freesmartphone.org Jan 07 12:53:25 that's why I was telling you, if you want to help, that'd be great Jan 07 12:53:28 budfive: (r1050) as you didn't see any clock dropouts, right? Jan 07 12:53:31 I have so many things to do I can't do all of them Jan 07 12:53:35 http://git.freesmartphone.org/?p=libframeworkd-glib.git;a=summary I only see master Jan 07 12:53:39 libframeworkd-glib being one of the thing I can't follow Jan 07 12:53:45 dcordes: well as I said, I intended to branch :-) Jan 07 12:53:47 I didn't Jan 07 12:54:02 and I said give me the url when it's done and you said it's already there Jan 07 12:54:05 * dcordes is confused Jan 07 12:54:30 I thought you meant the project URL :-) Jan 07 12:54:47 budfive: of course you still can try the ->100k swap, to make absolutely sure Jan 07 12:55:09 DocScrutinzrFlu: i saw one glitch, but i couldn't reproduce it and i'm not sure if wasn't due to me bumping the probe or something Jan 07 12:55:35 yup, understood Jan 07 12:56:13 budfive: Great clues, congratulations! Jan 07 12:56:57 DocScrutinzrFlu: so i'd say i saw no glitches. I'm not going to swap the resistors. Even with the cut I made in the lip, it's still hard to get to and I don't have the best soldering equipment, so it'd take a while Jan 07 12:56:59 budfive: you used a slow X scale - like 0.2s/div - to whatch the clock for dropouts? Jan 07 12:57:01 hello, how do i disactivate wifi in a way that it consume the same amount of power than after a reboot(so no connection)? i tried ifconfig eth0 down and killall udhcpc and it didn't work Jan 07 12:57:11 it's for suspending... Jan 07 12:57:54 DocScrutinzrFlu: i was doing a bit faster than that, and was watching for phase shifts Jan 07 12:58:12 DocScrutinzrFlu: but the whole thing is still up. i'll look at it with a slower scale Jan 07 12:58:32 DocScrutinzrFlu: BTW, i've read Sean promised some presents for those brave community who perform buzz rework. I think he should also thank budfive for his work ;) Jan 07 12:58:39 please try 0.2 or 0.5 and watch for a few minutes Jan 07 12:59:00 DocScrutinzrFlu: ok. doing that now Jan 07 12:59:25 PaulFertser: :) the buzzfix was much easier than getting in this deep Jan 07 12:59:49 budfive: make sure to see #1024 while watching ;-) Jan 07 13:00:15 yep. doing that now Jan 07 13:00:29 budfive: i'm yet to perform the buzzfix. All the shops are half-working these days in Russia because of exceptionally long holidays, so i don't have necessary SMT components yet. Jan 07 13:00:58 mabe with wmiconfig? Jan 07 13:01:14 Gnutoo: why not unload the module? It's the same as suspend. Jan 07 13:01:37 Gnutoo: or if the module is compiled-in, then unbind it. Jan 07 13:01:37 PaulFertser, good idea...i'll look if the wifi driver is avaliable as a module Jan 07 13:01:57 Gnutoo: it isn't Jan 07 13:01:59 paul: (shops half-working) they must have a heck of a hangover after all those holidays ;-) Jan 07 13:02:05 dcordes: could you write down some info for the HTC stuff and mail it to me please? :-) Jan 07 13:02:17 Ainulindale: htc stuff is broad Jan 07 13:02:19 Gnutoo: I use echo s3c2440-sdi > /sys/bus/platform/devices/s3c2440-sdi/driver/unbind to unload the compiled-in driver. Jan 07 13:02:25 Ainulindale: and I don't know what you plan in shr Jan 07 13:02:30 PaulFertser, thanks a lot Jan 07 13:02:34 Gnutoo: and echo s3c2440-sdi > /sys/bus/platform/drivers/s3c2440-sdi/bind to activate it again. Jan 07 13:02:38 gnutoo: you're probably looking for wmiconfig -i eth0 --wlan disable Jan 07 13:02:50 wpwrak: We're still on holidays. Till the 11th, i think. Jan 07 13:02:58 wpwrak, i know that command but does it consume less power? Jan 07 13:03:03 dcordes: well we just plan to do stuff on top of frameworkd Jan 07 13:03:10 dcordes: so our goal is pretty much the same Jan 07 13:03:13 gnutoo: module unloading or unbinding should do the trick too Jan 07 13:03:13 guys if other languages are installed , does applications need any changes , like sms or stuff ? Jan 07 13:03:20 gnutoo: it's what atheros recommend :) Jan 07 13:03:34 wpwrak, i even tried it when the wifi didn't reconnect but it didn't work and i had to reboot Jan 07 13:03:51 wpwrak, ok i'll try it Jan 07 13:04:40 Ainulindale: well as I said, in frameworkd and OE people can expect good support of the msm7* based htc boards (which include the above listed devices) in future Jan 07 13:04:52 gnutoo: unbind/bind or unload/load reset the module while --wlan enable/disable only turns off some functions, so if the module itself is stuck somewhere, --wlan enable/disable is less likely to help than the others Jan 07 13:05:02 * DocScrutinzrFlu bored and tired. sorry. bbl Jan 07 13:05:04 Ainulindale: if you want to inform the people reading openmoko planet I can sign up for an account and update it myself Jan 07 13:05:36 wpwrak, ok thanks i'll use bind/unbind Jan 07 13:06:44 wpwrak: notice that i unbind the s3c2440-sdi driver, because i can bind/unbind the wifi driver itself, but nothing'll generate "new card connected" event. Is it ok? Jan 07 13:07:56 thanks a lot Jan 07 13:09:03 by the way how can i have a reliable wifi without wmiconfig -i eth0 --power maxperf that consumes a lot of power? Jan 07 13:09:42 gnutoo: in what way does not setting maxperf make it unreliable ? Jan 07 13:09:59 wpwrak, it often disconnect or is unreachable Jan 07 13:10:16 dcordes: pong Jan 07 13:10:42 DocScrutinzrFlu: a bit late but: happy 09 to you and get well soon! Jan 07 13:10:46 paul: hmm, i didn't even know you could unbind the wifi driver :) Jan 07 13:11:07 dcordes: well that would be a good thing indeed :-) Jan 07 13:11:46 Ainulindale: how can I send info to planet? Jan 07 13:12:30 gnutoo: interesting ... i thought it should only affect performance, not overall stability Jan 07 13:12:33 dcordes: you need an accoutn on our blog Jan 07 13:12:39 blog.shr-project.org Jan 07 13:14:00 wpwrak: I think any driver can be unbound. One can reinit some PCI devices with this trick. Or even a PCI controller itself. The new (wait, how long ago 2.4 was?) /sys system is really flexible. :) Jan 07 13:15:16 PaulFertser: good find with the circular buffer bug! :) Jan 07 13:15:52 Kensan: thanks. But it was long-long ago. Now we're waiting for a pty buffer overflow bugfix :) Jan 07 13:16:29 PaulFertser: ah just saw the commit msg last night. What ticket nr? Jan 07 13:16:42 (the pty) Jan 07 13:17:45 Kensan: mickey's already investigating it. My patch is at http://projects.linuxtogo.org/pipermail/smartphones-userland/2009-January/000874.html Jan 07 13:18:05 PaulFertser: cool. Jan 07 13:18:47 Gnutoo: can you provide any details about AP used when you see instability? It might help wpwrak if he's going to look into this issue ;) Jan 07 13:19:17 PaulFertser: hm the archive does not seem to store patches... Jan 07 13:19:32 Kensan: ltg archive sucks. :) I'll give you another link. Jan 07 13:20:01 PaulFertser, it's hostapd with libnl with bcm4318 with b43 driver....kernel 2.6.27.4 Jan 07 13:20:13 mrmoku: ok update your org.oe.dev, set machine to htcraphael, PREFERRED_PROVIDER_frameworkd = "frameworkd-devel", require conf/distro/include/fso-autorev.inc Jan 07 13:20:24 mrmoku: then you can build fso images for the raph Jan 07 13:20:34 paul: hmm, if what you bind/unbind is the sdio_ar6000 driver, then it should actually come back. i've never tried that combination, though. Jan 07 13:20:39 Kensan: http://news.gmane.org/find-root.php?message_id=%3cm2aba5pcv7.fsf%40home.pavel.comp%3e Jan 07 13:20:40 it will give you the latest kernel with sd card support Jan 07 13:20:53 dcordes: I need to reup the image. Give me another 5 mins. Jan 07 13:21:31 wpwrak: the cards need hardware re-init, performed by SDIO stack, i believe. And it's not triggered without card insert event. Jan 07 13:22:32 Kensan: thanks a lot. although sms won't work because of frameworkd-devel vs frameworkd-glib conflict I'd like to give it a go on the touch pro Jan 07 13:22:58 dcordes: thanks a lot Jan 07 13:24:19 mrmoku: in frameworkd.conf set modem to qualcomm_msm. that's all. oh and I only use frameworkd -s ogsmd so I don't get the non working other services Jan 07 13:24:56 paul: yes, the card needs to go through the detection phase. however, this should be the same as for module unload/load, and we follow the proper protocol there. the sdio stack does the init when it sees the new driver register. i suspect that unbind/bind may miss a notification inside the ar6k stack. Jan 07 13:25:54 Gnutoo: you should have hilighted wpwrak, as i've no clues how to debug atheros's driver. Though i can test it with hostapd and AR5005G. Jan 07 13:26:12 ok Jan 07 13:28:23 gnutoo: maxperf goes to my list of questions for atheros, thanks. one problem with that module is that everything happens in that firmware black box. you just send it the right magic ioctl and then something happens, but it's often unclear what :-( Jan 07 13:31:31 wpwrak, ok lol(better have a softmac firmware next time no?) Jan 07 13:31:38 dcordes: I don't have one... but now I can tell my collegue to go and get one :-) Jan 07 13:34:03 mrmoku: ok. it's fairly easy to boot linux Jan 07 13:34:57 and thanks a lot for the info on unloading the driver Jan 07 13:35:04 dcordes: do you have a raphael? Jan 07 13:35:06 gnutoo: well, we're talking with atheros about this ... (softmax) Jan 07 13:35:08 mrmoku: no Jan 07 13:35:14 seen one? Jan 07 13:35:34 looks like a nice gadget :-) Jan 07 13:35:51 gnutoo: problem is that they're not very comfortable with letting people replace their firmware. fear of FCC, i guess. Jan 07 13:36:04 mrmoku: which not everybody can afford.. Jan 07 13:36:20 yep :( Jan 07 13:36:23 mrmoku: I'd already have one if it wouldn't be so damn expensive Jan 07 13:36:39 wpwrak: but they have already opened HAL for some laptop chipsets. Aren't they softmac? Jan 07 13:36:43 gnutoo: but yes, a softmac would solve about 99% of our nightmares in that area Jan 07 13:37:37 mabe http://www.softwarefreedom.org/resources/2007/fcc-sdr-whitepaper.html is interesting Jan 07 13:37:55 paul: i think it's different groups inside atheros with different points of view on those issues Jan 07 13:37:59 mrmoku: I asked my provider (vf ger) if I can get customer rebate. even if I prolong my contract for two years, it's still almost 400 euros Jan 07 13:38:03 FCC sucks! Oh, btw, can anyone explain the meaning of those FCC messages on various devices (incl. keyboards)? Something like "this device must receive any harmful interference received" etc.? Jan 07 13:38:05 gnutoo: you can still interpret that one either way :) Jan 07 13:38:18 ok Jan 07 13:38:28 dcordes: yep, 400 euros is what I heard too... with contract Jan 07 13:38:41 PaulFertser: that print makes no sense :) Jan 07 13:39:00 mrmoku: I thought of a cheapo, second handed SE xperia x1 Jan 07 13:39:10 dcordes: But i've seen it numerous times on various devices. Not exactly the wording i used, but quiet like that. Jan 07 13:40:23 it should say "warning: this electromagnetic stuff can toast your brains" Jan 07 13:42:15 paul: maybe it means that the device shouldn't burst into flames if someone puts a cell phone next to it Jan 07 13:43:27 wpwrak: "Operation is subject to the following two conditions: (1) this device may not cause harmful interference, and (2) this device must accept any interference received, including interference that may cause undesired operation." Jan 07 13:43:50 I can't on earth figure what exactly does it mean. Jan 07 13:44:35 "may not cause"? Does it mean can't cause? MUSTN'T cause? Or just MAY not cause, and MAY cause? Jan 07 13:45:15 "must accept" How can it not accept interference received? Doesn't seem to have any meaning to me at all ;) Jan 07 13:47:28 paul: as your fcc lawyer ;-) Jan 07 13:47:36 what is the /sys value to check battery power? Jan 07 13:47:39 asK Jan 07 13:49:50 never mind, found this, http://wiki.openmoko.org/wiki/GTA02_sysfs#power_supply_battery_information Jan 07 13:55:21 yoyo is working on experimental design for shr-settings - http://scap.linuxtogo.org/files/4447c0bf5290609c26b6edee7d7899a2.png Jan 07 13:55:31 what do you, shr guys, think about it? ;) Jan 07 13:56:42 dos1: scap does not work... just for me? Jan 07 13:56:44 scap is down? Jan 07 13:56:55 dos1: you broke scap :o Jan 07 13:58:26 ups Jan 07 14:00:00 yoyo: i prefer as illume-config does Jan 07 14:00:05 with a toolbar Jan 07 14:00:50 toolbar is not ported to python-elementary Jan 07 14:00:51 :) Jan 07 14:01:03 I also prefer toolbar, but till then this is a nice replacement Jan 07 14:01:13 spank morphis Jan 07 14:05:15 Hire, dolf1074: doing this layout helped in future implementation with toolbar ;) Jan 07 14:07:22 anyone know if, using efl it is possible/what complications there would be to having somethins like this: app emits a signal -> edje program runs, looping once/sec -> time passes -> app emits another signal -> looping program stops Jan 07 14:13:27 Ainulindale, quickdev: python-elementary from new bb recipe is not upgrading, if there was older version Jan 07 14:14:09 in old recipe there was 0.1.0+svnr, now there is 0.0.0+gitr Jan 07 14:16:54 dos1: can't do naught, that's OE/morphis fault Jan 07 14:26:38 cat /sys/devices/platform/bq27000-battery.0/power_supply/bat/status says "Not Charging" when FR (ASU-200811142247)) connected via USB to PC; how to force it to charge? Jan 07 14:34:15 Ainulindale: my opkg still wants to install 0.1.0 version of pyelementary from bearstech and it gets 404 error Jan 07 14:34:31 Ainulindale: can you regenerate package index? Jan 07 14:35:32 on shr-unstable of course Jan 07 15:24:29 arghhh Jan 07 15:24:55 mickeyl: Hi Jan 07 15:25:36 python-elementary in e svn has not toggle.state_set fixed :/ Jan 07 15:25:49 hi kensan Jan 07 15:26:31 mickeyl: not on a train today? ;) Jan 07 15:26:55 PaulFertser, /sys/bus/platform/devices/s3c2440-sdi/driver/unbind does not exist on my kernel (2.6.24) Jan 07 15:29:22 GNUtoo: find a corresponding node yourself with the help of find Jan 07 15:29:26 ok Jan 07 15:29:33 that's what i was doing Jan 07 15:29:34 Kensan: hehe. thankfully not. I need to get some rest Jan 07 15:29:55 * mickeyl cleaning up for the rest of this week -- processing inbox, patches-to-review, etc. Jan 07 15:30:12 and probably some light oe work Jan 07 15:30:45 the correct file is /sys/bus/platform/drivers/s3c2440-sdi/bind and /sys/bus/platform/drivers/s3c2440-sdi/unbind on 2.6.24 Jan 07 15:31:26 mickeyl: ah the overflowing inbox... Jan 07 15:32:03 mickeyl: saw mareks mail to the ML: seems like the meeting was productive. Jan 07 15:33:11 mickeyl: btw, what's the status of abraxa, if I may ask. Is he back in the game or busy with other things? Jan 07 15:34:09 dos1, ask morphis about python-elementary Jan 07 15:34:10 MIA again. didn't do the tasks he promised me until end of last year Jan 07 15:34:22 i will have to launch plan B Jan 07 15:34:37 not relying on nor waiting for him Jan 07 15:34:41 *shrug* Jan 07 15:34:49 mickeyl: ah I see... makes planning hard I guess. Jan 07 15:34:56 quickdev: he isn't here... Jan 07 15:35:03 yeah..let's wait for him Jan 07 15:35:09 q Jan 07 15:35:13 i diffed version from linuxtogo and e repo Jan 07 15:36:00 yeah, he did good work, but if i can't co-work without relying on people, so we need to go on on our own Jan 07 15:36:09 s/if// Jan 07 15:36:09 mickeyl meant: yeah, he did good work, but i can't co-work without relying on people, so we need to go on on our own Jan 07 15:37:47 mickeyl: yeah, absolutely. I for one think you gave him enough "lead-time" to pick up "his" project again... so it's he who didn't seize the opportunity... Jan 07 15:54:09 bumbl: hey Jan 07 15:54:20 bumbl: need what you gave me as a blog entry and your image Jan 07 15:54:25 bumbl: I'm going to post it today it's almost ready Jan 07 15:56:10 one moment Jan 07 15:59:20 Ainulindale: i don't have the blogpost anymore because rafb.net has deleted it Jan 07 15:59:43 and i thought you don't need it - i will upload the pic though Jan 07 16:00:12 bumbl: well then I'll show you mine and you'll give me your opinion Jan 07 16:00:20 bumbl: I need a small text on organization within the team too Jan 07 16:00:26 If you can help me :-) Jan 07 16:00:37 i'll try Jan 07 16:00:49 even if I might be off-on-off all the time Jan 07 16:01:21 i am at the moment trying to fix a problem concerning openrc (sysvinit replacement) Jan 07 16:01:46 http://img3.imagebanana.com/img/igaa2a8w/iwantyouforshrdev.jpg Jan 07 16:02:35 Thanks Jan 07 16:03:12 bumbl: http://rafb.net/p/XhgiZI14.html Jan 07 16:03:42 The only bit missing is the team organization bit IMHO Jan 07 16:03:50 If you have ideas, please tell me :-) Jan 07 16:03:57 (the image is for the organization bit) Jan 07 16:04:40 Ainulindale: good foto :) hahahahha :) Jan 07 16:04:50 yoyo: that's bumbl's :-) Jan 07 16:08:07 that's funny. There was a very similar poster made during the russian civil war. http://www.ruslania.com/pictures/big/4605000010941.jpg wonder which was first Jan 07 16:12:03 budfive: http://www.loc.gov/exhibits/treasures/trm015.html Jan 07 16:12:06 Ainulindale: what do you mean with team organisation? Jan 07 16:12:23 bumbl: globally, what we should do to formalize stuff Jan 07 16:12:28 e.g. a team to handle tickets Jan 07 16:12:33 join #openmoko-cdevel or irc and ask for the almighty Ainulindale or quickdev Jan 07 16:12:35 a team to follow OE changes Jan 07 16:12:44 a team to do QA Jan 07 16:12:47 rwhitby ping Jan 07 16:12:47 s/or irc/or mailinglist/ Jan 07 16:12:49 bumbl meant: join #openmoko-cdevel or mailinglist and ask for the almighty Ainulindale or quickdev Jan 07 16:12:50 etc. Jan 07 16:12:56 Ainulindale: bug-wrangler Jan 07 16:13:04 Kensan: what ? Jan 07 16:13:35 Ainulindale: one person dedicated to triaging bugs and making sure none gets lost forgotten etc. Jan 07 16:13:40 bumbl: you are a 'bilder-ideen-klauer', hrrrr :D Jan 07 16:13:43 Yeah, but what was that? :-) Jan 07 16:13:58 budfive, you go to caltech? Jan 07 16:14:00 Kensan: is that a "I want this position!"? :-) Jan 07 16:14:02 Ainulindale: a job description ;) Jan 07 16:14:15 Ainulindale: no not me, Jan 07 16:14:25 Ainulindale: I suck at that *heh* Jan 07 16:14:42 bumbl: we also need blog authors *hint* Jan 07 16:14:51 *nudge nudge* Jan 07 16:14:55 Say no more! Jan 07 16:15:04 Ainulindale: it's good if you have a couple "open spots" and not just state: "We need you". Jan 07 16:15:07 lol Jan 07 16:15:14 *winkwink* Jan 07 16:15:29 Kensan: that's what I intended to do :-) Jan 07 16:15:33 Kensan: hence the task writing Jan 07 16:15:40 Ainulindale: trés bien. Jan 07 16:15:52 Ainulindale: btw, totaly forgot: happy 09 to you too. Jan 07 16:16:14 :-) Jan 07 16:16:16 thanks Jan 07 16:42:49 bumbl, I know you like those dumb but lucky people's videos: http://www.youtube.com/watch?v=3WWqxO64Krw Jan 07 16:52:46 irefox Jan 07 16:53:03 bah Jan 07 16:54:04 TAsn: btw. you are the bad opressors now in some of our newspapers :( Jan 07 16:54:20 I know. Jan 07 16:57:56 TAsn: i don't find that point of view fair because it is black/white Jan 07 16:58:37 I don't find it fair either, though shit happens ;\ Jan 07 16:58:52 freesmartphone.org: 03mickey 07specs * rfc28c25dbefe 10/ (3 files in 2 dirs): fix bug with some cross references not appearing & regen docs Jan 07 16:58:52 freesmartphone.org: 03mickey 07specs * rf77848094654 10/ (4 files in 3 dirs): otapi: add org.freesmartphone.GSM.Monitor interface Jan 07 17:00:08 Ainulindale: is the assignment of bugs to MS1 in trac complete - or would a bug monkey have to assign them correctly? Jan 07 17:01:00 at least a friend told me that on a local social network site (similar to facebook) there already is a group Israeli victory - for victory and peace in the holy land which is quite popular Jan 07 17:01:10 Well, it leads to discussion as to the fact that the bug are truly meant to be corrected before MS1 Jan 07 17:01:30 bumbl: hey, you didn't tell me if you wished to write that missing part in the end :-) Jan 07 17:03:04 Ainulindale: so non-showstopper bugs or bugs that will get fixed in fso would block MS1? Jan 07 17:03:58 Ainulindale: at the moment trying to find the right words Jan 07 17:06:19 mrmoku_: that's subject to discussion :-) Jan 07 17:17:26 freesmartphone.org: 03mickey 07framework * r456d077428e7 10/ (3 files in 3 dirs): ogsmd: add stubs for monitoring API org.freesmartphone.GSM.Monitor.Get{Serving|Neighbour}CellInformation Jan 07 17:26:18 howevet Jan 07 17:26:22 *however Jan 07 17:26:30 i get echo with SHR Jan 07 17:48:08 fgau: please answer me in private when you'll be available Jan 07 18:22:45 for ho send diff file for framework rule? Jan 07 18:32:26 yoyo: yea, ho's must get some rules ;) Jan 07 18:33:31 PaulFertser: I use gta01 and I need have diffrent values in rules.yaml file :/ Jan 07 18:34:19 I can't gate correct to make working with filters :/ Jan 07 18:36:05 yoyo: s/ho/who/ :DDDD Jan 07 18:36:46 PaulFertser, how are you? Everything fine? Jan 07 18:37:28 quickdev: I'm fine. It's Joerg who's ill, unfortunately. Jan 07 18:37:35 si?? Jan 07 18:37:46 yoyo: no! ;) Jan 07 18:37:59 PaulFertser, yes. I'm very curious to know if a resistor can fix 1024 :) Jan 07 18:38:10 PaulFertser: I talking about code not spaling Jan 07 18:38:14 quickdev: Haven't you read the report of Dima Kogan? Jan 07 18:38:28 yoyo: wait a minute, i'll try to explain. Jan 07 18:40:23 quickdev: I wasn't the one who was going to perform a rework, because i can't reproduce #1024. I was trying to be a mediator :) And Dima did a lot of cool testing, noticed some interesting things. Now we'll wait for Dieter to investigate it further. I think we finally got what's needed to debug it. Dieter, the gsm reverse-engineering expert and Dima, experienced EE with equipment who can easily reproduce #1024. So, the victor Jan 07 18:41:27 yoyo: sorry for making fun of your spelling. ;) Could you please describe again, what problems do you have? Jan 07 18:42:01 PaulFertser: I using gta01 and I have problem with brightness Jan 07 18:42:11 its one problem from many Jan 07 18:42:15 meny Jan 07 18:42:18 fiev Jan 07 18:42:28 yoyo: Oh, backlite node is called differently. Are you using MS4? Jan 07 18:42:38 yoyo: in frameworkd-devel it's solved afaik Jan 07 18:42:56 PaulFertser: shr-unstable Jan 07 18:43:19 which is MS4/stabilization Jan 07 18:43:26 PaulFertser: in ./oevents/rules.yaml is SetDisplayBrightness("pcf50633_bl", 90) Jan 07 18:43:44 yoyo: ok, that's MS4. You have to change pcf50633_bl to gta01_bl, i guess Jan 07 18:43:51 in gta02 shoud by pcf50633_bl and in gta01 gta01_bl Jan 07 18:44:06 yoyo: Yes, that's correct. Jan 07 18:44:12 dos1: is there already a pacakge for the settings app? Jan 07 18:44:44 mrmoku: yes, i was posting url to it here many times ;) Jan 07 18:44:45 PaulFertser: there is alsov a problem with a toggle headset Jan 07 18:45:03 dos1: I mean in the shr repo :-) Jan 07 18:45:14 mrmoku: i don't know Jan 07 18:45:18 yoyo: so, you mean you solved several problems with the rules on gta01 and want to suggest a patch? Jan 07 18:45:24 mrmoku: i have only access to git :p Jan 07 18:45:24 is there a recipe (.bb) Jan 07 18:45:41 dos1: if there is a recipe a could build it :-) Jan 07 18:46:28 PaulFertser, do you think that's enough to finally debug it? I'm in doubt :) just read the mailing list post Jan 07 18:46:29 mrmoku: i sended to Ainulindale old recipe (with svn address), i don't know what happened later Jan 07 18:46:43 PaulFertser: I don't know where to fix it. or in rules.yaml or in fso_actions.py Jan 07 18:47:56 dos1: I think Ainulindale is too busy... Where can I get it from? Jan 07 18:48:38 mrmoku: old recipe is here: Jan 07 18:48:40 http://openmoko.opendevice.org/build/shr-unstable/shr/openembedded/packages/shr-settings-new/shr-settings-new_svn.bb Jan 07 18:48:58 i don't know if it is 100% correct, but it works Jan 07 18:50:08 mrmoku: ok, will try that now Jan 07 18:50:31 yoyo: please describe your problem in more detail :) Jan 07 18:50:33 mrmoku: Ainulindale wanted to replace old shr-settings by it in repo Jan 07 18:50:43 No Jan 07 18:50:50 hehe :-) Jan 07 18:50:50 mrmoku: we need to bb files Jan 07 18:50:58 shr-settings new version will have a superior PV Jan 07 18:51:05 But default version for testing will be the old one for now Jan 07 18:51:09 On unstable, the new one Jan 07 18:51:50 oh :D Jan 07 18:51:51 does anyone else have problems with the shr git repo? Jan 07 18:52:13 PaulFertser: yyy one problem is not correctly working rules.yaml Jan 07 18:52:32 http://n2.nabble.com/building-shr-locally-td2116229.html Jan 07 18:52:34 so i didn't undestand it correctly before ;) that's good Jan 07 18:52:45 Ainulindale: ok, so the new one gets something like SHR_PV = "0.1.0" ? Jan 07 18:53:07 yes Jan 07 18:53:07 yoyo: And what exactly not working except backlight? Because backlight is already fixed in FSO HEAD. Jan 07 18:53:24 PaulFertser: rules.yaml it in on present time write not multi platform and not for diffrent devices that part of seting path for example brightnes is beed Jan 07 18:53:24 but how shall the bb be named? the package needs to be called shr-settings for both. Jan 07 18:53:30 can you set PN explicitely in the .bb? Jan 07 18:53:38 yoyo: that's easy to do Jan 07 18:53:51 mrmoku: Well yes but that'd be a bad idea Jan 07 18:53:54 mrmoku: why would you do that? Jan 07 18:53:59 yoyo: backlight is multiplatform in FSO HEAD. Jan 07 18:54:11 Ainulindale: yes but if I make upgrade all my handes will by gone Jan 07 18:54:24 Nope Jan 07 18:54:27 You'll have the choice to merge Jan 07 18:54:32 Ainulindale: hmm, how does the package name be determined? by the name of the .bb I thought Jan 07 18:54:44 yoyo: and we can currently ship different rules.yaml dependong on the device Jan 07 18:54:55 upgrade don't have good diff config updater Jan 07 18:54:58 mrmoku: why do you need to determine pn? Jan 07 18:55:16 ok Ainulindale it is idea Jan 07 18:55:35 I think beter will by to make fso_actions.py more flexible Jan 07 18:55:48 Ainulindale: you said two .bb files one for old settings, one for new. But both will be called shr-settings (the package name) Jan 07 18:55:54 it is first adres prom path org.freesmartphone.Device.Display Jan 07 18:55:54 yes Jan 07 18:55:57 just set PV in the bb file name Jan 07 18:56:04 check midori for example Jan 07 18:56:09 yoyo: what exact suggestions do you have except backlight? The backlight is already multi-platform. Jan 07 18:56:23 I have to go for now, I'll be back after dinner Jan 07 18:56:53 PaulFertser: yes but it demon witch running in background don't work currectly Jan 07 18:57:34 two option 1. send diffrent file rules.yaml to deffrent devices Jan 07 18:58:12 yoyo: do you read what i'm saying? You're using MS4. And it's fixed in FSO HEAD. Do you understand that? Jan 07 18:58:13 2. fix / change fso_actions.py to by more flexible on making "actions" Jan 07 18:58:40 fso fso fos fso not working on gta01 Jan 07 18:58:49 yoyo: are you using FSO HEAD? Jan 07 18:59:02 yoyo: no Jan 07 18:59:17 shr-unstable Jan 07 18:59:43 yoyo: answer on -pl Jan 07 18:59:45 yoyo: If you did, you'd found that the backlight is compatible among 2.6.24, 2.6.28 and GTA01, don't you think so? Jan 07 19:00:09 s/-pl/.pl/ Jan 07 19:00:09 dos1 meant: yoyo: answer on .pl Jan 07 19:01:04 dos1: you stole him from me :-/ Jan 07 19:01:06 dos1: ;) Jan 07 19:01:43 PaulFertser: i want to explain him what you mean in his native language ;) Jan 07 19:01:46 PaulFertser: funny Jan 07 19:04:02 mwester, mwester-laptop: ping-combo Jan 07 19:04:04 quickdev: meanwhile Ainulindale is eating... is there a fix for the multiple repo problem? Jan 07 19:04:35 mrmoku, not that I know..please ask him about it Jan 07 19:04:57 PaulFertser: yoyo said that sound scenarios for using headset are not compatible with gta01 Jan 07 19:05:47 Yes there's a solution Jan 07 19:05:48 dos1: Then he should probably file a bugreport on trac, http://trac.freesmartphone.org , the only way his report won't get lost. Jan 07 19:05:58 For repos which will contain several packages Jan 07 19:06:02 Additional bb class Jan 07 19:06:08 For repo for an app (pyphonelog, settings) Jan 07 19:06:13 SUpersede SRC_URI in the bb file Jan 07 19:06:20 I'm really off Jan 07 19:06:24 Ainulindale: ok, thanks Jan 07 19:06:31 dos1: it seems that LED handling is not flexible enough too. It should really be handled with different configs for different devices. Jan 07 19:06:47 dos1: Then he should probably file a bugreport on trac, http://trac.freesmartphone.org , the only way his report won't get lost. Jan 07 19:08:28 PaulFertser: hehe, gta01 doesn't have any leds :D Jan 07 19:08:49 haha :) Jan 07 19:09:48 yoyo: the best way to file a bugreport on trac, and mention all incompatibilities there. I guess someone from FSO team will sort it out. Jan 07 19:12:31 anyone can login at shr trac? Jan 07 19:12:43 i get: site is linking to it self Jan 07 19:12:48 in firefox & konqueror Jan 07 19:13:53 works for me Jan 07 19:13:58 pops up login prompt Jan 07 19:14:20 yeha i see the htaccess too Jan 07 19:14:23 but can you login? Jan 07 19:15:01 just did Jan 07 19:15:07 -_- strange Jan 07 19:15:51 are you blocking javascript? Jan 07 19:16:25 ssweeny: that is not javascript ;p Jan 07 19:16:56 well i got noscript but all disabled Jan 07 19:17:17 and no noscript at konqueror so its a site problem i guess Jan 07 19:37:30 freesmartphone.org: 03mickey 07framework * r41985fee5153 10/ (4 files in 4 dirs): Jan 07 19:37:30 freesmartphone.org: ogsmd: Implemented org.freesmartphone.GSM.Monitoring.* as stub Jan 07 19:37:30 freesmartphone.org: erroring out with org.freesmartphone.GSM.UnsupportedCommand. Jan 07 19:37:30 freesmartphone.org: ogsmd: [TI CALPYSO] Implemented the new monitoring API: Jan 07 19:37:31 freesmartphone.org: org.freesmartphone.GSM.Monitoring.Get{Serving|Neighbour}CellInformation Jan 07 19:37:36 stefan_schmidt_: good evening Jan 07 19:37:41 Kensan! Jan 07 19:37:49 Dave: hey there Jan 07 19:37:56 :) Jan 07 19:38:10 Dave: How's 09 so far for you? Jan 07 19:38:33 What? It's 2009? Jan 07 19:38:35 Already? Jan 07 19:38:41 I have accumulated 4 failing disks so far. Jan 07 19:38:46 :o Jan 07 19:39:02 Dave: ah so you slept through the whole christmas, newyears thing. Jan 07 19:39:59 Pretty much :) Jan 07 20:08:57 Alinermedia: ping Jan 07 20:09:08 ah sorry Jan 07 20:09:11 Ainulindale: ping Jan 07 20:09:27 http://rafb.net/p/cAVskg42.html <- something like that you want me to do? Jan 07 20:11:40 bumbl: s/The guys from openmoko.pl/The guys from #openmoko.pl/ - we unfortunatly don't have this domain :( Jan 07 20:12:00 bumbl_: s/The guys from openmoko.pl/The guys from #openmoko.pl/ - we unfortunatly don't have this domain :( Jan 07 20:12:45 dos1: yeah Jan 07 20:13:00 i would have written your names but i only know of you dos and yoyo Jan 07 20:13:22 bumbl: dos, yoyo and hiciu Jan 07 20:13:26 ;) Jan 07 20:14:17 ah ok Jan 07 20:14:48 anyway tell that Ainulindale - what I have written is only a draft - i don't know if it's that what he likes Jan 07 20:19:35 bumbl: mrmoku with u in the end - has nothing to do with mokos :-) Jan 07 20:19:54 mr mokus Jan 07 20:20:37 Dave: there is just one of me - I'm unique ;) Jan 07 20:20:37 sry Jan 07 20:20:45 :D Jan 07 20:20:46 bumbl: np Jan 07 20:21:09 Hokus mokus? Jan 07 20:21:46 Mokus pokus! Jan 07 20:22:03 simsalabim ;) Jan 07 20:24:24 * mrmoku wishes to be a wizzard Jan 07 20:24:37 * bumbl too Jan 07 20:24:42 but life is a beast Jan 07 20:24:49 and a bitch Jan 07 20:24:50 and so everyone is what he is Jan 07 20:25:47 I installed the FSO distro (fso-testing) )and I tried to install an opkg package on the phone but I get some errors. Is there any problems with opkg on FSO or is should be supported ^ Jan 07 20:26:03 hm... in the very first MokoNinja movie mickey_bbl and LaF0rge get introduced as "The Doctor" and "The Wizard". Jan 07 20:26:09 *heh* Jan 07 20:26:20 haha Jan 07 20:26:39 is Harald still working for swisscom? Jan 07 20:27:19 bumbl: I wouldn't know. Jan 07 20:28:15 at least they can gant him gsm licenses Jan 07 20:28:25 * bumbl wants to have a bts too Jan 07 20:29:14 bumbl: I doubt that, and obviously not in germany. Jan 07 20:30:18 heyho Jan 07 20:30:24 well at least he "has to" do that for his job ;) Jan 07 20:30:31 hi morphis Jan 07 20:31:12 woohooo! they fixed the building's main power :) Jan 07 20:31:27 DocScrutinizer2: no more flu? Jan 07 20:31:49 sure, wont vanish that quick Jan 07 20:31:56 DocScrutinizer2: hm :( Jan 07 20:32:04 DocScrutinizer2: but feeling better I hope. Jan 07 20:32:05 bumbl, what's a bts? Jan 07 20:32:15 Gute besserung joerg Jan 07 20:32:27 quickdev: base transceiver station Jan 07 20:32:33 not exactly. more like "light again" Jan 07 20:32:48 bumbl: thanks Jan 07 20:33:06 DocScrutinizer2: hm ok... Hope you get well soon... Jan 07 20:33:42 my landlord drilled a nice hole in the main power cable of the building Jan 07 20:34:39 DocScrutinizer2: is he still alive or did he use non isolated tools Jan 07 20:35:07 hey morphis Jan 07 20:35:13 nothing happened, he's well Jan 07 20:35:40 morphis! Jan 07 20:35:45 hehe Jan 07 20:35:53 anybody here who has a glofiish m800 phone? Jan 07 20:35:54 morphis, I saw your patch concerning the segfault. I think you misunderstood something. The problem is that argv IS free'd. The problem is not, that it IS NOT free'd. :) Jan 07 20:36:06 just R, S and T main fuse blew Jan 07 20:36:09 morphis: stefan_schmidt_ does Jan 07 20:36:13 quickdev: ok Jan 07 20:36:29 Kensan: thanks Jan 07 20:36:32 morphis, that's why I removed the free call Jan 07 20:36:45 morphis: and another issue - toggle.state_set doesn't work in pyelementary from e repo Jan 07 20:36:45 dos1: http://scap.linuxtogo.org/files/4447c0bf5290609c26b6edee7d7899a2.png <- shr settings new? if yes: please different icons Jan 07 20:36:49 *use Jan 07 20:36:55 morphis, what about "dos1: python-elementary in e svn has not toggle.state_set fixed :/" Jan 07 20:37:05 ah..he already answered Jan 07 20:37:05 morphis: from linuxtogo it works Jan 07 20:37:49 hm .. I moved the last linuxtogo rev to e-svn Jan 07 20:37:55 * DocScrutinizer2 will try to get a decent 1KW UPS Jan 07 20:38:33 DocScrutinizer2, UPS? Jan 07 20:38:35 DocScrutinizer2: heh Jan 07 20:38:40 morphis: i diffed, pyelm from e repo is using pointer, pyelm from linuxtogo is using functions Jan 07 20:38:57 dos1: I am currently working on the toolbar, but I a have currently a problem with implementing the callback function for added items Jan 07 20:39:24 dos1: ok, I will fix it now Jan 07 20:39:34 morphis: ok, thanks Jan 07 20:39:36 dos1: I guess that will cost a decent bit Jan 07 20:40:06 bumbl: that's yoyo's experimental design Jan 07 20:40:40 shr-settings is modular, so different layouts can be used with the same modules Jan 07 20:40:47 dos1: hm it needs icons - these look so ... 20th century :p Jan 07 20:41:13 bumbl: say it to yoyo Jan 07 20:41:23 http://en.wikipedia.org/wiki/Uninterruptible_power_supply Jan 07 20:41:34 mickey: please add "opim" to the specs Makefile Jan 07 20:43:43 dos1: please update :) Jan 07 20:44:35 dos1: the screenshot shows your approach while waiting for the toolbar? :) Jan 07 20:46:14 morphis: not mine - yoyo's ;) Jan 07 20:46:22 dos1: ok :) Jan 07 20:47:45 dos1: I think it tooks longer as expected to implement the toolbar - especially the callback functions Jan 07 20:48:27 dos1: maybe I have to ask raster, to change something in c elementary for that Jan 07 20:48:46 morphis: hmm, ok Jan 07 20:49:06 morphis: what about hoversel? Jan 07 20:49:30 dos1: currently I have a misterious bug when adding items to the toolbar - the whoole programm freezes and I do nothing else than the test.c app Jan 07 20:50:35 dos1: I don't watch a it Jan 07 20:50:56 morphis: Clock.time_get() would be usefull right now too Jan 07 20:51:27 dos1: uups :) Jan 07 20:51:29 make setup still isn't working for me. I'm getting the same error Jan 07 20:52:12 dos1: will be there in some minutes Jan 07 20:52:20 morphis: ok :) Jan 07 20:53:54 jhass840: still have the link to the error? Jan 07 20:54:07 dos1: you have the current version installed? Jan 07 20:54:37 morphis: i'm installing now your version from few minutes ago :p Jan 07 20:54:53 mrmoku: Hold on, I'll see if I can find it Jan 07 20:55:50 morphis: ok, i have it installed Jan 07 20:55:51 jhass840: ok... do you have MAKEFLAGS set? Jan 07 20:55:58 morphis: and toggles work Jan 07 20:56:16 mrmoku: http://pastebin.com/m7c2527e0 Jan 07 20:56:54 dos1: ok, can you run the test.py and check the toolbar test Jan 07 20:57:29 dos1: I want know if the app freezes if you show the toolbar-window and click on one item on the toolbar Jan 07 20:58:17 morphis: Segmentation fault when i clicked on "Toolbar" button in main window Jan 07 20:58:38 hmm Jan 07 20:58:53 morphis: sorry, it may be my fault Jan 07 20:59:44 raster: here? Jan 07 20:59:46 dos1: problem? Jan 07 20:59:47 jhass840: what git version are you using? Jan 07 20:59:58 dolf1074: ja Jan 07 21:00:00 kind of Jan 07 21:00:31 raster: I've copied the elm_entry.c and it is now working good, but what's the best/easiest way to change fontsize? Jan 07 21:00:34 morphis: i have segfault with all buttons in test.py :( Jan 07 21:00:40 raster: I have a short question about the api of the elementary toolbar Jan 07 21:00:54 ok segfault is even ok, can you give me a backtrace? Jan 07 21:01:08 mrmoku: 1.5.6.3 Jan 07 21:01:11 dolf1074: can I close http://shr-project.org/trac/ticket/60 Jan 07 21:01:21 morphis: ok, give me one minute Jan 07 21:01:37 jhass840: hmm, strange... I just retried myself and it works fine Jan 07 21:01:39 mrmoku: You may close, but the work is still in progress Jan 07 21:01:46 dolf1074: there pretty much is none other than theme Jan 07 21:01:59 dolf1074: have a todo item for setting scaling per-widget Jan 07 21:02:05 raster: why the callback functions for the toolbar items pass the toolbar object and not the toolbar item object? Jan 07 21:02:14 ie "scale this widget 2x size" Jan 07 21:02:21 which is realyl what you want :) Jan 07 21:02:35 raster, but now, Can I change the size in my c program on the fly? Jan 07 21:02:51 dos1: ok Jan 07 21:03:00 mrmoku: Perhaps it's on my end...I'll dig around for a bit and see what I find Jan 07 21:03:03 morphis: toolbar item is not an object Jan 07 21:03:08 jhass840: ok Jan 07 21:03:39 same with hoversel item Jan 07 21:03:41 or list item Jan 07 21:04:16 raster: ok, but I currently have a little problem while binding it to python Jan 07 21:04:21 raster: how is your elementary list coming along? progress? Jan 07 21:04:37 mrmoku: the beginnings of a working one are committed Jan 07 21:04:40 defintiely not finished Jan 07 21:04:44 raster: I tried 'edje_text_class_set("textblock_style", NULL, 20);' but that's is not working Jan 07 21:04:44 Raster Jan 07 21:04:58 raster: I can't decide if the the callback function is called to which item the call belongs Jan 07 21:05:29 raster: looking forward to it :-) Jan 07 21:05:36 raster: is there any possibility to get the toolbar item within the callback function? Jan 07 21:06:04 dolf1074: as i said. there's a todo item - you can't change font size. so don't try. Jan 07 21:06:14 Dave: boo Jan 07 21:06:18 okey :( Jan 07 21:07:12 raster: also nothing that makes the fontsize change? Jan 07 21:07:25 i'll pass it in with event_info Jan 07 21:07:28 since that isn't used Jan 07 21:07:55 raster: in the current svn-version? Jan 07 21:08:33 there Jan 07 21:08:38 is now in svn Jan 07 21:08:45 Committed revision 38495. Jan 07 21:09:17 raster: wonderful Jan 07 21:09:24 dolf1074: nothing makes font size change other than globally in the app with scaling changes - intended for when a user modifies their ui scaling preferences, resolutions change (dpi changes) etc. Jan 07 21:10:19 raster: okey, I will then work with 2 edje parts, one with a big font and one with a small font and switch between them. Jan 07 21:11:10 mickey: Have you seen the Dieter's answer about %Nxxxx commands? He can't give a definitive answer, unfortunately. Jan 07 21:12:42 dolf1074: if i were u - just wait until the feature exists instead of working around it Jan 07 21:13:38 If I have to wait on every feature that I want, I will have to wait long ;) It's not a big deal, It easily done and can be changed afterwards ;) Jan 07 21:14:40 I'm someone who wants result now ;) Jan 07 21:14:55 ok ok :) Jan 07 21:18:57 raster: btw, how is it going with the elm list? Jan 07 21:19:37 mrmoku: just asked that a few lines up Jan 07 21:19:38 :) Jan 07 21:19:52 I will have to learn to read Jan 07 21:20:39 raster: moo! Jan 07 21:20:49 Kensan: burrrrp! Jan 07 21:21:58 raster: what's the temperature in Au? Jan 07 21:22:19 20 Jan 07 21:22:40 raster: right, and I bet you think that's cold Jan 07 21:22:48 here it is the same temperature, except with a minus :p Jan 07 21:23:04 hmmpf Jan 07 21:23:27 Do they sell heated keyboards? Jan 07 21:23:30 Kensan: for summer - like january.. hell yeah it is! Jan 07 21:23:32 morphis: ok, i need help - how to get backtrace? :D i installed gdb Jan 07 21:24:01 raster: heh Jan 07 21:24:13 dos1: run "gdb python" in gdb type "run test.py" all in the tests/ dir Jan 07 21:24:29 than you open the toolbar window, clicks on some items Jan 07 21:24:43 than the app should freeze or segfault'd Jan 07 21:24:51 back in gdb type "backtrace" Jan 07 21:25:41 morphis: ok, i have it Jan 07 21:25:45 morphis: http://pastebin.com/m667f20b2 Jan 07 21:26:09 git clone git://git.openmoko.org/git/u-boot.git u-boot Jan 07 21:26:13 uups Jan 07 21:26:53 dos1: you get the segfault while starting test.py? Jan 07 21:27:27 morphis: no, when i clicked button Jan 07 21:28:04 the toolbar button? Jan 07 21:28:20 morphis: yep, but it segfaults on others too Jan 07 21:28:49 dos1: hm, when do you updated the e libs last? Jan 07 21:29:07 morphis: i have e libs from shr-unstable Jan 07 21:30:54 dos1: which svn rev does shr-unstable have? Jan 07 21:31:34 morphis: it looks like 38274 Jan 07 21:32:21 raster: you know that e is segfaulting, when pressing the euro sign in the number keypad? Jan 07 21:32:39 mrmoku: that.. is not good! Jan 07 21:33:06 dos1: hm Jan 07 21:33:26 raster: the euro sign and german sz make it go BOOM BANG ;) Jan 07 21:33:33 dos1: can you try it on your desktop? Jan 07 21:34:22 morphis: on desktop i don't have recent e installed ;x Jan 07 21:34:28 only some old from debian repo Jan 07 21:34:41 and there is no elementary :( Jan 07 21:35:14 dos1: ok Jan 07 21:35:23 quickdev: some time for me? Jan 07 21:42:21 dos1: puuuh, I fixed the bug Jan 07 21:43:23 raster: if I set the same instance of an icon for more than one toolbar item, the process of the app hangs Jan 07 21:43:56 raster: but all items are displayed with the correct icon Jan 07 21:44:54 raster: it seems there is something wrong with handling the items and there icons within elementary Jan 07 21:46:29 you take the same icon object and put it in each toolbar item? Jan 07 21:46:50 or you create multiple copies of the same icon object Jan 07 21:46:50 ? Jan 07 21:47:18 I take the same icon object and put it in each toolbar item Jan 07 21:47:20 hmm Jan 07 21:47:28 * raster spanks morphis Jan 07 21:47:30 duh Jan 07 21:47:32 no wonder Jan 07 21:47:47 elementayr should work ok - if used right/sanely Jan 07 21:47:58 the moment u deviate it has very little "save your arse for you" code Jan 07 21:48:04 haha Jan 07 21:48:17 don't do that Jan 07 21:48:23 how do u ever expect that to work sanely Jan 07 21:48:29 ? Jan 07 21:48:38 an object (widget) has 1 instance Jan 07 21:48:43 it lives in 1 place Jan 07 21:48:55 if u go putting the 1 thing in multiple places Jan 07 21:49:03 things are going to go haywire Jan 07 21:49:21 raster: ok Jan 07 21:49:22 as it triest to steal the sub-widget from another part of itself Jan 07 21:49:24 tries Jan 07 21:49:45 if u want multiple copies of the icon Jan 07 21:49:51 make multiple icon objects Jan 07 21:49:57 just with the same file Jan 07 21:49:59 raster: I will write this in the documentation : Jan 07 21:50:04 dont worry - the lower levels are smart Jan 07 21:50:10 they re-use the same image data loaded off disk Jan 07 21:50:30 great Jan 07 21:50:35 they dont go load the same image 10 times and make 10 copies of the pixel data (which is the really memory=eating stuff) Jan 07 21:56:06 dos1: the next commit brings the toolbar to you :) Jan 07 21:56:49 morphis: yay! :D Jan 07 21:57:41 dos1: callback now works Jan 07 21:57:44 mrmoku: Okay, I fixed the problem Jan 07 21:58:06 jhass840: what was it? Jan 07 21:59:09 mrmoku: apparently git doesn't like the objects prefixed by ad, so I downloaded them all manually and placed them into the proper directory during the make process. Jan 07 22:00:00 jhass840: heh, strange... so that would depend on the git version used... hmm Jan 07 22:00:07 anyway good building :-) Jan 07 22:00:21 thanks! :) Jan 07 22:08:49 morphin: when will be next commit? ;> Jan 07 22:09:25 Raster Jan 07 22:09:38 Dave Jan 07 22:09:39 alarm Jan 07 22:10:11 morphis*, btw: with 38496 i still have segfaults on test.py Jan 07 22:10:28 brinb bring Jan 07 22:14:42 raster, can a program in an edje collection be run when the app emits a signal, run continuously ( 1 do_something(); 2 sleep(1); 3 goto 1 ), and then stop when the app emits another signal? Jan 07 22:14:57 w/o causing other problems? Jan 07 22:15:03 yes Jan 07 22:15:10 but there is no sleep Jan 07 22:15:16 the p[rogram will need to schedule a timer Jan 07 22:15:19 raster, I got to play with a friend's treo today Jan 07 22:15:28 it was nice, as usual, but I was disappointed with the alarm app Jan 07 22:15:30 andn store the return handle somewhere Jan 07 22:15:33 it could have been a LOT cooler Jan 07 22:15:36 and emit another would delete the timer Jan 07 22:15:47 Dave: hahhaahhaha Jan 07 22:15:52 it gave me a lot of ideas; so I figure I'll either write my own, or hopefully merge my ideas into yours :) Jan 07 22:15:58 i was making elementary's widgets work Jan 07 22:16:07 not trying to make a hyper-cool alarm app Jan 07 22:16:10 its a demo app Jan 07 22:16:11 although it's honestly more of a different thing Jan 07 22:16:12 right Jan 07 22:16:20 so I may as well do my own, then show it off, eh? :) Jan 07 22:16:24 it's really fucking geeky :p Jan 07 22:16:43 raster, what else happens when a group of young engineers get together for a party ;) Jan 07 22:16:55 raster, cool, and I assume the app can pass a variable to the edje collection as well? Jan 07 22:18:05 also, where did the detailed api references that i used to be able to find online go? :) Jan 07 22:18:30 AntonTakk: yes. there's a whole edje messaging api Jan 07 22:18:39 u can send messages with ints, strings and so on Jan 07 22:18:51 (and edje can send messages back to the app too) Jan 07 22:20:03 and are there any good api references with this stuff (other than the code it's self)? Jan 07 22:20:12 dos1: I will now commit the toolbar stuff - only dealing direct with the toolbar items (delete, select) is currently not implemented Jan 07 22:20:30 morphis: ok Jan 07 22:21:05 raster, are you a master edje hacker? Jan 07 22:21:17 Dave: .... Jan 07 22:22:00 i know nothing about edje Jan 07 22:22:04 totally clueless Jan 07 22:22:07 :) Jan 07 22:22:09 xD Jan 07 22:22:22 AntonTakk: there's examples - e uses it for the battery for example Jan 07 22:22:25 there he said it! Jan 07 22:22:29 and cpufreq Jan 07 22:22:40 raster & edje : :O Jan 07 22:22:46 Don't listen to raster, he's totally clueless! Jan 07 22:22:53 yeah Jan 07 22:22:55 dont listen to me Jan 07 22:23:00 i know nothing Jan 07 22:23:02 You crazy old cook Jan 07 22:23:06 just some raving lunatic Jan 07 22:23:09 exactly Jan 07 22:23:15 but we like you, nonetheless Jan 07 22:23:16 raster: cool, i'll dig those out of svn/etc tonight, thanks :) Jan 07 22:23:32 The Zen of E Jan 07 22:23:44 LET E SHOW YOU THE WAY Jan 07 22:23:49 dos1: it is commited Jan 07 22:23:50 lol Jan 07 22:23:57 wait, that sounded wrong Jan 07 22:24:05 morphis: thanks :) Jan 07 22:24:11 so gn8 everybody Jan 07 22:24:14 cya Jan 07 22:26:30 * raster giggles at the whole gta02 expensive brick emails Jan 07 22:28:11 raster: community ml? Jan 07 22:28:18 yeah Jan 07 22:28:40 raster: ah am have given up on that a couple months ago Jan 07 22:28:41 more widgets please Jan 07 22:28:56 i wasnt even reading it - until now Jan 07 22:29:18 sadly the negative-mails... are mostly accurate and true Jan 07 22:29:31 and a lot of the "community defence" is just wishful thinking Jan 07 22:29:37 :\ Jan 07 22:29:45 at least it's not heartbreaking Jan 07 22:30:23 for some people it is Jan 07 22:30:33 poor people :( Jan 07 22:30:37 poor users Jan 07 22:31:01 for some peolpe - they spent $400 and yes fr was marketed as "Ready to go" Jan 07 22:31:06 verye xplicitly at the start Jan 07 22:31:19 then the "ready for consumers" wording was removed - making it neutral Jan 07 22:31:35 yeah :( Jan 07 22:31:36 no big warning "this may not even boot and may eat your children" added Jan 07 22:31:39 I just ran into one of them Jan 07 22:31:52 the user came bawling into my arms today, completely fed up. Jan 07 22:32:04 Thinking they were cheated out of their money, and after pulling out most of their hair. Jan 07 22:32:09 so understandably a lot of people who werent totally clued into the dev cycle and what develoeprs were bitchnig about daily Jan 07 22:32:15 haha Jan 07 22:32:20 thought it'd be a usable phone - possibly with the odd rough bit Jan 07 22:32:22 right Jan 07 22:32:33 and what is your position now, raster? Jan 07 22:32:46 those totally clued into the grapevine of course knew what they were getting Jan 07 22:32:58 i understand and feel for both sides Jan 07 22:33:04 a community that wants thgns to succeed Jan 07 22:33:15 as otherwise their love and devotion would have been misspent Jan 07 22:33:18 and just where do you fit in, raster? ;) Jan 07 22:33:22 and so its like throwing good money after bad Jan 07 22:33:36 ie throw mroe love/attention/defence at your pet to save its life Jan 07 22:33:40 it can work Jan 07 22:33:48 the nay-sayers are just behaving like a normal consumer Jan 07 22:33:58 i was marketed X - i expect X Jan 07 22:34:06 i paid for X, now where is X? Jan 07 22:34:09 :( Jan 07 22:34:15 good point, bad thing Jan 07 22:34:23 now I'm disappointed again Jan 07 22:34:34 raster is the gui crack dealer of those who still have hope ;) Jan 07 22:34:36 i feel for the community - i do, but i have to side with the "consumers" Jan 07 22:34:49 * Dave runs his send sean hatemail sean script Jan 07 22:34:58 ahaha Jan 07 22:34:59 no no Jan 07 22:35:02 theres reality Jan 07 22:35:05 gta01 was dropped Jan 07 22:35:06 why? Jan 07 22:35:08 uh huh Jan 07 22:35:10 ¬_¬ Jan 07 22:35:13 only like 1000 or so exist Jan 07 22:35:22 its a siny # of devices to suport Jan 07 22:35:23 yet in some respects it was actually better - haha Jan 07 22:35:28 yeah yeah Jan 07 22:35:29 and frankly there are scant few of them around om's offices Jan 07 22:35:33 so its HARD to support Jan 07 22:35:42 and it was sold as a developer device Jan 07 22:36:02 the freerunner should have been the same Jan 07 22:36:04 so thats a valid "oh well - we did warn you - but you have the source so you can support yourselves - unlike closed vendors" Jan 07 22:36:10 I mean, really, what the fuck DID they expect? Jan 07 22:36:17 and there is a little too much "om is dumping ta02 support" stuff going around Jan 07 22:36:26 gta02 can't afford to drop support Jan 07 22:36:28 it HAS to work Jan 07 22:36:31 no matter the effort Jan 07 22:36:43 and likely it will end up losing om more than it makes in the process Jan 07 22:36:45 right! Jan 07 22:36:53 as the continual software effort is a big cost Jan 07 22:36:57 BUT Jan 07 22:37:03 they need to to legitimise the platform Jan 07 22:37:12 duh Jan 07 22:37:18 if they fail on gta02 and say "well community - the soruce is yours - you do it" Jan 07 22:37:24 or else they're just ruining/dooming themselves, all-around, as far as I can tell. Jan 07 22:37:32 Right, exactly :p Jan 07 22:37:34 that simply justify everything the nay-sayers say Jan 07 22:37:42 haha, true Jan 07 22:37:58 and despite a core group of fanbois - u're going ot have a lot of negative press and feedback Jan 07 22:38:02 when people reserahc gta03 Jan 07 22:38:15 "should i get it from this little known taiwanese vendor?" Jan 07 22:38:23 hah Jan 07 22:38:26 they will see a lot of fanbois going ":yay yay yay" Jan 07 22:38:32 I don't believe that Jan 07 22:38:33 and a lot of consumers going "fuck that shit" Jan 07 22:38:36 YEAH Jan 07 22:38:37 in the end Jan 07 22:38:37 exactly Jan 07 22:38:52 they will see a lot of positive fans and negative "realists" Jan 07 22:38:56 Seriously, I think OM shit on a lot of people's hopes, dreams, and aspirations. Jan 07 22:39:00 all those who did buy the gt02 where those who believed in opensource Jan 07 22:39:02 and probabyl decide the risk isnt worth the $ to spend on it Jan 07 22:39:10 now are they disappainted, so I fair for om Jan 07 22:39:14 or better yet, their faith in their product, structure, and vision. Jan 07 22:39:23 and then gta03 will collapse as it absically has a saturated market than alreayd spent its available $ on gta02 Jan 07 22:39:26 dolf, ja Jan 07 22:39:35 a few of them can afford to pay for a gta03 upgrade Jan 07 22:39:40 but not the majority Jan 07 22:39:47 haha, can we have an upgrade program? ;) Jan 07 22:39:50 and you have the other half of the market scared off Jan 07 22:39:57 right Jan 07 22:40:03 I think more than you think cannot affort the gta03 upgrade Jan 07 22:40:09 Man, an upgrade program would be sweet, but totally impossible. Jan 07 22:40:15 totally. Jan 07 22:40:27 it'd be a loss maker for om to do an upgrade program Jan 07 22:40:30 i cant say Jan 07 22:41:04 but they'd need to weigh up that loss against the possible marketing/pr gains Jan 07 22:41:05 but om HASt to get gta02 right Jan 07 22:41:05 I think google with the android will get all the opensource fanboys Jan 07 22:41:05 as not doing it will simply kill gta03 Jan 07 22:41:10 huhu Jan 07 22:41:16 dolf, no kidding Jan 07 22:41:22 google has a bigger fanboi space Jan 07 22:41:25 I mean, it's "Google", right? :D Jan 07 22:41:27 they love anything google Jan 07 22:41:29 same with apple Jan 07 22:41:39 OM has two fangirls over here :) Jan 07 22:41:39 gogole could shit on a plate Jan 07 22:41:45 and it would be hot, right? Jan 07 22:41:48 and they'd love it and say google revolutionised the world Jan 07 22:41:52 hahaha Jan 07 22:41:58 but I think all the ones who has so much trouble with the gta02 will eventually buy some android gsm Jan 07 22:42:01 It gives new meaning to the word craptacular Jan 07 22:42:12 dolf, misschien, ja Jan 07 22:42:15 googles fanboise who are to that extreme probably outnumber om's entire market for gta02 Jan 07 22:42:16 :) Jan 07 22:42:27 yeah Jan 07 22:42:31 Is there any point in releasing gta03 if the software for gta02 isn't even finished yet? Jan 07 22:42:36 so sadly true Jan 07 22:42:47 dolf1074: or someone competes with om in the "make an open linxu handset that ISNT android" space Jan 07 22:43:04 ie they may go "ok - since its not om - i might give it a second shot" Jan 07 22:43:04 freesmartphone.org: 03mickey 07framework * r83356edeaa14 10/framework/subsystems/ogsmd/ (4 files in 3 dirs): ogsmd: move dbus error module one level up Jan 07 22:43:05 haha, I can't say much, given who I've worked for hahaha Jan 07 22:44:03 so i hope om doesnt drop gta02 in favor of just focusing on gta03 Jan 07 22:44:09 * Dave coughs Jan 07 22:44:15 tho really only part of the software stack cares if its gta02 or gta03 Jan 07 22:44:17 Raster, and if they do? Jan 07 22:44:19 other bits dont give a flying hoot Jan 07 22:44:33 Will your opinion of OM change? Jan 07 22:45:47 i think if om can get a reasonably usable ui on top of a mostly finished and stable fso BEFORE the gta03 is released, things will improve greatly Jan 07 22:46:18 gta02 will ending up working either way -- gta01 has been working great works as a daily, and it's been "unsupported" for a long time -- really it's supported, just by FSO in software and community hackers in kernel Jan 07 22:46:18 well, yes, and then once the GTA03 debuts, they can simply dump it on it. Jan 07 22:46:23 The software must be finished Jan 07 22:46:32 i mean, as i understand it, the first nes kinda floped, and look at nintendo now Jan 07 22:46:36 The spice must flow! Jan 07 22:46:56 Dave: sucky tho it is - my money is probably riding on craptacular implosion :) if i had to lay bets (and wanted to not lose) Jan 07 22:47:04 freesmartphone.org: 03fercerpav 07framework * r587f0f9bb547 10/docs/ogsmd/ti_calypso/modem: Jan 07 22:47:04 freesmartphone.org: Add information about Calypso's flow control handling Jan 07 22:47:04 freesmartphone.org: It was discovered that Calypso doesn't respond to the flow control as it Jan 07 22:47:04 freesmartphone.org: should, i.e. it doesn't stop output immediately and instead transfers the Jan 07 22:47:06 freesmartphone.org: rest of the message, which may lead to UART buffer overflows. See mwester's Jan 07 22:47:08 freesmartphone.org: description at http://docs.openmoko.org/trac/ticket/1376 . Jan 07 22:47:20 Raster, yeah, same here, unfortunately Jan 07 22:47:24 I fear they will drop the ball. Jan 07 22:47:26 Sadly. Jan 07 22:47:45 Let's hope that doesn't come to pass. Jan 07 22:47:46 :) Jan 07 22:47:46 theres this fine line Jan 07 22:47:50 of "business reality" Jan 07 22:48:08 and "being principled and reliable to your word" Jan 07 22:48:14 Oh gee, yes Jan 07 22:48:23 Something I fear they don't quite understand, even now :( Jan 07 22:48:34 invariably someone makes an argument of "fuck it drop this - its losing money - we need to to stay in business" Jan 07 22:48:39 * DocScrutinizer2 ducks away Jan 07 22:48:40 then that makes you look like an arse Jan 07 22:48:53 and your customers run off as u dont stick bny principles (or are reliable) Jan 07 22:49:02 so you lose on the business side anyway Jan 07 22:49:22 haha JOERG Jan 07 22:49:25 but by sticking to principles u have to fend off all the arguments of "but thats not the reality of business" Jan 07 22:49:27 YOU CANNOT ESCAPE US! Jan 07 22:49:38 its a fine line to draw - which one basically make you more money Jan 07 22:49:41 thats the question Jan 07 22:49:46 oh god Jan 07 22:49:48 its a tough question Jan 07 22:49:52 Raster, that reminds me... of... FIC Jan 07 22:49:55 * Dave laughs Jan 07 22:49:57 personally, i think i intend to buy a gta03 when it becomes available, reguardless of the state of the software stack, and hope things turn out for the best Jan 07 22:49:58 its voodoo magic Jan 07 22:50:04 but if it were me Jan 07 22:50:07 i'd stick to ptinciples Jan 07 22:50:10 raster: yo Jan 07 22:50:20 AntonTakk: Do you have a gta02? Jan 07 22:50:25 as i can justify "bsuiness realities" on principles too Jan 07 22:50:27 dolf1074: yes Jan 07 22:50:33 raster: how are you getting on with your sms app? Jan 07 22:50:33 but to me at least being a man of your word Jan 07 22:50:42 AntonTakk, I will inevitably end up with a GTA03 whenever it is manufactured, and will invariably end up destroying that, too :) Jan 07 22:50:42 "doing the right thing" is more important than money Jan 07 22:50:51 Can you also buy a gta03 for me when it becomes available. I don't have money left :p Jan 07 22:50:56 and chances are that in doing so u'll make more than if u didnt Jan 07 22:51:08 you will re-inforce the lyoalty of customers you do havbe Jan 07 22:51:25 you will bring in less-loyal customers ont he fringe into the "i love you guys" camp Jan 07 22:51:25 gta03 will just been more software for gta02 (that's what it'd done for gta01 anyway) Jan 07 22:51:40 dolf1074: hehe, $500 i can muster fairly easily, 1k not so much Jan 07 22:51:43 and you will now create an image of solid dependability that will earn you yet more customers later Jan 07 22:52:03 mato: yo. um - doing a list widget now for my contact selector Jan 07 22:52:04 dolf :P Jan 07 22:52:04 :) Jan 07 22:52:17 then u can select the contact (and type in numbers) Jan 07 22:52:19 You can always try :p Jan 07 22:52:26 :) Jan 07 22:52:32 then a simple matter of the entry being a proper one that saves the result oat the end Jan 07 22:52:43 and send actually packages up the editings, the desitination number Jan 07 22:52:48 and passes it off to fso Jan 07 22:52:54 (i'll do the pass-off last) Jan 07 22:52:59 good stuff Jan 07 22:53:12 Well, when I bought my freerunner I decided that the more I put into it the more I'd get out of it. So I'm setting up my development evironment and I plan to help finish the software Jan 07 22:53:37 DocScrutinizer2: hahahah Jan 07 22:53:49 raster: /me will have a fso simulator for you, in the next couple of weeks Jan 07 22:53:57 raster: I think the multi-device angle can help OM. Jan 07 22:54:04 yeah Jan 07 22:54:13 mato: werd! Jan 07 22:54:31 mato: i'd be interested in that as well Jan 07 22:54:32 mato: i just go slowly on the sms app as i need to translate needs into elementary widgets appropriately Jan 07 22:54:38 not as nasty hacks Jan 07 22:54:45 also it's beenn xmas/new years Jan 07 22:54:46 raster: if people start using the software on say HTC phones or other devices word spreads quickly. But for that to work out the software stack has to be very good and I see a chance at that. Jan 07 22:54:46 :) Jan 07 22:55:05 raster: same here, first clients screaming bloody murder, then xmas :) Jan 07 22:55:14 but all quiet now, it seems :) Jan 07 22:55:41 will get onto the simulator right after i fix the bugs in^W^W^W^Wrewrite gsm0710muxd Jan 07 22:55:41 heheheheh Jan 07 22:56:09 mato: what bugs? Jan 07 22:56:16 might make it a kernel module while you're there Jan 07 22:56:32 mickeyl: i wish, but i don't feel up to that right now Jan 07 22:56:44 mickeyl: but someone else can take that up at some stage Jan 07 22:57:02 PaulFertser: well, for starters, alloc/dealloc of a channel is not reliable Jan 07 22:57:10 *nod* laf0rge has this task on his agenda since ages Jan 07 22:57:26 jhass840: i have come to a similar conclusion, thou i am evilly duplicating some effort, as I don't find the official direction i see coming particularly pleasing to myself Jan 07 22:57:28 hard to predict when he'll be able to do it though Jan 07 22:57:55 PaulFertser: so i can get gsm0710muxd into a state where it says (iirc) write to channel not open or similar Jan 07 22:57:58 and gets confused Jan 07 22:58:28 PaulFertser: 2nd, the logging is totally useless Jan 07 22:58:30 You just write too fast probably. Before the channel is really opened. Jan 07 22:58:47 mato: i don't agree, it clumsy, but i found a bug because of the logging. Jan 07 22:58:53 mickeyl: opim is missing in subdirs variable of fso-spec Makefile Jan 07 22:59:05 welp, later all, heading home Jan 07 22:59:47 mickey, who? what? Jan 07 22:59:48 PaulFertser: 3rd, at least in the fso git version, there is gta02-specific stuff in there (sysfs powermanagement) i'd like to abstract out Jan 07 23:00:11 PaulFertser: since gsm0710muxd is useful in itself for e.g. anyone with a data card Jan 07 23:00:11 Dave: kernel muxe Jan 07 23:00:38 mato: but mickeyl's planning to move upstream anyway, you'd better look at the upstream version, i guess. Jan 07 23:00:39 Who's doing it? Jan 07 23:00:41 elektranox: true. since there is no xml.in docs i don't see a point in adding it now though Jan 07 23:00:56 Dave: noone, but it was/is on laf0rge's list of things Jan 07 23:01:19 elektranox: the one in charge promised me docs until end of 08 Jan 07 23:01:25 bah Jan 07 23:01:26 alas, we're in 09 and nothing happened Jan 07 23:01:35 like that'll get done in '09... *rolls his eyes* Jan 07 23:01:56 well Jan 07 23:01:58 well there are 2 xml.in files in there Jan 07 23:02:08 as soon as you move to 3G you really want the muxer in kernel space Jan 07 23:02:15 PaulFertser: 4th, i can't seem to get it to close mux mode reliably, at least on my data card Jan 07 23:02:18 besides I'd love having a "real" serial emulation Jan 07 23:02:21 not just ptxes Jan 07 23:02:47 mickeyl: "real" emulation could be faked by adding some dbus methods to send MSC Jan 07 23:02:56 at least in the interim Jan 07 23:03:26 sure, we could do that Jan 07 23:03:33 but that screams "ugly ugly ugly" :) Jan 07 23:03:46 UGLY Jan 07 23:04:11 sure, but better than waiting for a kernel module for ages Jan 07 23:04:24 what to I do about TOPDIR not defined? Jan 07 23:04:24 mickeyl: specs are under gpl, aren't they? Jan 07 23:04:33 mato: i agree with 1) (i'm not sure how you want to do that though. Async? That would change API) and 4) (an investigation is needed, as it seem to work on calypso) :) Jan 07 23:04:36 elektranox: nah, i'd say MIT or BSD Jan 07 23:04:43 ok Jan 07 23:05:05 or even CC Jan 07 23:05:21 should be the same as the ones from freedesktop.org Jan 07 23:06:10 PaulFertser: for 1) i need to investigate more, but really i looked at the code and it's (imo) a horrible mess (been hacked all over the place) Jan 07 23:06:28 PaulFertser: so i'd like to give it a major once-over Jan 07 23:06:34 mato: i agree. It was designed as a temporary hack while waiting for kernel level muxer. Jan 07 23:07:07 speaking from experience, temporary hacks tend to turn into permanent solutions :) Jan 07 23:07:51 anyhow, expect some shiny new code from me next week and hopefully everyone will be happy Jan 07 23:08:03 mato: are you really going to rewrite it? Jan 07 23:08:44 PaulFertser: liberally using the old code but more or less, yes Jan 07 23:09:02 call it refactor and spring clean if you want Jan 07 23:09:13 * Kensan welcomes the new gsm muxing overlords Jan 07 23:09:41 before you really dive in, please crosscheck with the new one Jan 07 23:09:41 mato: great. But please base it on upstream version. Don't forget to implement buffering for the pty's. And i don't understand how you can solve 1) without changing API, really. Jan 07 23:09:52 http://git.openmoko.org/?p=zecke/qtopia-muxer.git;a=summary is almost complete Jan 07 23:10:10 should get us a way to compare "bugs" Jan 07 23:10:41 mickeyl: that is a different implementation? Jan 07 23:10:46 yes Jan 07 23:10:54 this one is based on qtopia's code Jan 07 23:11:24 for some reason it doesn't write to the ptys, i was planning to debug that Jan 07 23:11:30 but it talks fine to the modem Jan 07 23:11:41 and strace shows it's receving answers ok Jan 07 23:12:23 PaulFertser: for 1) we may not be talking about the same thing... let me pull out the data card and look at the exact problem... 5mins Jan 07 23:15:18 PaulFertser: got it... Jan 07 23:16:21 PaulFertser: this is what i mean: gsm0710muxd.c:570:pseudo_device_read(): Write to a channel which wasn't acked to be open. Jan 07 23:16:29 note i have not actually debugged this yet Jan 07 23:16:35 but it's trivial to reproduce Jan 07 23:17:05 reproduction of fail? Jan 07 23:17:10 start the muxer, open a channel e.g. by using dbus-send from the shell Jan 07 23:17:43 then do e.g. cat /dev/pts/XX and ^C the cat process Jan 07 23:17:55 muxer will correctly register that it should close the channel Jan 07 23:18:13 if you then start e.g. mickeyterm, you'll get a new channel Jan 07 23:18:18 but it won't actually work Jan 07 23:18:30 * mickeyl nods Jan 07 23:18:33 this has been a bug from day 1 Jan 07 23:18:43 Alas :( Jan 07 23:19:25 mato: It seems that's another issue... Jan 07 23:19:44 PaulFertser: yes? Jan 07 23:20:41 mato: if you're going to debug it tomorrow, well, go for it :) Jan 07 23:21:04 aha, sorry, i misread what you wrote Jan 07 23:21:21 yes, tomorrow... Jan 07 23:25:11 mickeyl: upstream for the muxer are the pyneo guys? Jan 07 23:25:21 mato: but don't you think there's another problem: a race when the muxer returns a pty to a not-yet-opened channel? Jan 07 23:25:25 mickeyl: yes Jan 07 23:25:28 mato: yes Jan 07 23:25:39 mato: emdete Jan 07 23:26:20 PaulFertser: please explain - why can't the muxer open a channel and only then reply to the dbus method? Jan 07 23:26:58 mato: i guess dbus methods are sync. But i'm not sure. Jan 07 23:28:52 mato: yes, muxer-upstream is mdt@pyneo.org Jan 07 23:29:44 thx Jan 07 23:29:51 ok, i'm off, will be in touch Jan 07 23:29:53 'nite Jan 07 23:29:59 mato: good luck! :) Jan 07 23:31:06 mickeyl: what would you say if the muxer is changed to open channels async? Jan 07 23:33:11 you mean just when the corresponding ptx is opened rather than on allocation time? Jan 07 23:33:34 i like that. Jan 07 23:33:56 btw., if you're hacking on it, there's one bug that has been driving me crazy since ages Jan 07 23:34:08 the muxer will happily allocate dozens of channels, even if the modem don't support more than 'n' Jan 07 23:34:29 if the 07.10 protocol doesn't support checking that number, I'd love to have a modem db for that Jan 07 23:34:37 since that will improve debug reports from users Jan 07 23:35:19 latest SHR looks really nice, thanks for the hard work guys Jan 07 23:37:30 mickeyl: i think we have to wait what mato'll come with soon :) As to async: no, i meant you request a new channel, but muxer doesn't answer immediately. He has to make sure the modem has opened the channel. For now it just sends him a command to open a channel and doesn't wait for answer, returns you the new pty. If you're too fast, you can happen to write to it before the modem confirmed the new channel is ready. Jan 07 23:38:24 * PaulFertser notices how he call the modem "he" like it's alive and strange human being Jan 07 23:39:02 if u liked the modem.. you'd call it a "she" :) Jan 07 23:39:23 raster: oh, btw, could you possibly explain, since when the user is called "she"? Jan 07 23:39:51 PaulFertser: ah, i see Jan 07 23:39:53 i like that behaviour Jan 07 23:39:56 more predictable Jan 07 23:39:58 less racy Jan 07 23:40:14 raster: and one more question: do you know the meaning of those 2 FCC rules written on almost all consumer electronics? Jan 07 23:40:39 mickeyl: but the API will change. You'll probably have to wait for a signal that a new channel is open. Is it ok? Jan 07 23:41:21 nah Jan 07 23:41:26 PaulFertser: nfi Jan 07 23:41:28 we have async dbus for a reason Jan 07 23:41:37 though i think more or less they mean Jan 07 23:41:39 PaulFertser: just delay the answer as long as necessary Jan 07 23:42:02 once the client receives the answer it knows it's good to go Jan 07 23:42:34 Raster, what sort of cults do you belong to? :) Jan 07 23:42:35 mickeyl: i see... Somehow the main glib should be called until the new channel is ready and a timeout implemented in case the modem somehow silently fails to allocate a channel? Jan 07 23:42:35 "this device will not infereer with other thgns you have nearby by sending out any RF signals ... OTHER than within certain licenses or limited bands as functinality and laws specify and otherwise any other rf interference (or electrical field) will be within certain standardised limits" Jan 07 23:42:47 PaulFertser: exactly Jan 07 23:42:55 what exactly the list of rf bands exit Jan 07 23:43:05 and what youa re allowed to output on all of them Jan 07 23:43:15 and other limtits and their levesl -i dont know Jan 07 23:43:23 elecrtical and magentic fields too Jan 07 23:43:56 raster: but exact wording they used is exceptionally strange and unclear imho. Jan 07 23:44:07 sounds like FCC 15 Jan 07 23:44:17 Dave: exactly. Jan 07 23:44:37 :) Jan 07 23:44:50 PaulFertser: as i said - i dont know the details Jan 07 23:44:56 all i know is the gist of what they are meant to mean Jan 07 23:45:08 :) Jan 07 23:48:52 freesmartphone.org: 03mickey 07framework * r9e32a58cebfc 10/framework/subsystems/ogsmd/modems/abstract/mediator.py: Jan 07 23:48:52 freesmartphone.org: ogsmd: gather context status from pdp object, not from network Jan 07 23:48:52 freesmartphone.org: this makes it inline with the signal and the actual logical state Jan 07 23:48:52 freesmartphone.org: closing #289, as all parts should be in place now. Jan 07 23:50:30 mickey, RE: #289, when did you "fix" it? Jan 07 23:50:42 btw. is something like a night modus planned for FSO? Jan 07 23:50:58 Sargun: over the last couple of weeks Jan 07 23:51:16 Ah, ok Jan 07 23:51:22 Sargun: the last bits were the netlink stuff Jan 07 23:51:27 everything should be in sync now Jan 07 23:51:31 and even without polling Jan 07 23:51:36 We're using netlink now? Jan 07 23:51:41 yep Jan 07 23:51:44 Hello Sargun. Jan 07 23:52:11 hello Dave Jan 07 23:53:44 when I mapped some streets three weeks ago by bike I was blindet by the brightness of the FR's display. Jan 07 23:54:03 hah! Jan 07 23:55:22 reducing the brightness to let's say 40% was better, but it will go back to 100% after blanking the display and then reenabling it via a touch Jan 07 23:56:16 elektranox: that's despite changing rules.yaml? Jan 07 23:56:44 odeviced could use some configuration for brightness Jan 07 23:56:53 open a ticket and i'll do it Jan 07 23:56:58 g'night now Jan 07 23:57:50 PaulFertser: the problem is, that 40% is not enough on sunny days Jan 07 23:58:34 elektranox: sure. I was surprised you say it goes back to 100%. Jan 07 23:59:37 ;) I just called the dbus function to reduce backlight Jan 07 23:59:40 what u really want is a photoreceptor Jan 07 23:59:48 to guage the ambient lighting Jan 07 23:59:55 ... but dont have one of those Jan 08 00:00:57 this would be cool of course, but it would already be helping to switch mode by hand Jan 08 00:01:46 similar to the switching of ring-sounds in illume perhaps Jan 08 00:02:05 ophonekitd is acting a bit strange Jan 08 00:02:16 what else is new -_- Jan 08 00:02:21 mrmoku: What's supposed to be in topdir.conf? Jan 08 00:02:24 I think such sensor should be integrated. Quite an overkill to use USB for one photo-sensitive Q or whatever. Jan 08 00:03:04 PaulFertser: ACK Jan 08 00:04:15 I did killall -9 ophonekitd; sleep 3; ophonekitd; Jan 08 00:04:15 http://om.quatrox.org/om/gprs/opkd.log Jan 08 00:04:18 I've seen it in an iPhone some days ago, they also used it to recognize if the phone is @ear Jan 08 00:04:27 so that they can disable the screen and save power Jan 08 00:05:23 PaulFertser: no need for more than one GPIO for a simple sensor like that Jan 08 00:05:58 elektranox: http://wiki.openmoko.org/wiki/Wish_List_-_Hardware#Light_Sensor Jan 08 00:06:37 quatrox: and bin-bang it?.. Nah, i'd better use SPI or I2C bus. Jan 08 00:07:16 it has SPI? Jan 08 00:08:18 I have a lot of sensors at home. They only have a variable voltage drop or variable currrent throughput Jan 08 00:08:50 but are about 5% of the price also Jan 08 00:09:06 would be better to use that for mass production Jan 08 00:10:02 quatrox: GPIO != ADC pin. For a simple sensor you have to use integrated ADC or an external one. Or might be PMU has a free output. Jan 08 00:10:26 if i ever need a quick laugh http://wiki.openmoko.org/wiki/Wish_List_-_Hardware always brings me up ;D Jan 08 00:11:29 PaulFertser: depends on the sensor and what you want. Just to sense if the phone is near your ear, a GPIO is good enough Jan 08 00:11:49 quatrox: we talked about being adoptive to ambient light. Jan 08 00:12:42 ohh... sorry for interrupting in the middle Jan 08 00:12:48 quatrox: for your ophonekitd problem i suggest you ping Ainulindale or quickdev. They'll at least tell you whom to ask. Jan 08 00:13:16 thanks:) Jan 08 00:13:30 quickdev? Jan 08 00:13:55 quatrox: i thought you just like arguing no matter what (about photosensor) ;) Jan 08 00:14:35 I argue too much Jan 08 00:14:46 I know Jan 08 00:14:57 sorry Jan 08 00:15:11 elektranox: you should really propose to add a sensor on hardware list. I don't think it's too hard to implement, therefore the devs may take that into account. If they read hardware mailing list, of course. Jan 08 00:15:33 quatrox: no need to be sorry. I like to do it even more than you, i bet ;) Jan 08 00:16:43 anyway, I got a box full of stuff I planned to integrate... just didn't find the time. One cool thing is a wireless spy cam Jan 08 00:16:58 spy cams are small Jan 08 00:17:14 xD Jan 08 00:17:26 wireless makes it hard to implement without interference:( Jan 08 00:19:10 yeah Jan 08 00:19:31 quatrox, FYI, I totally don't see the Freerunner being used as any sort of practical spy equipment. Jan 08 00:19:59 And what should James Bond do? Jan 08 00:20:07 He also wanted a freerunner Jan 08 00:21:23 but I made one cool thing: I replaced my light switches with solid state relays connected to a computer that has BT Jan 08 00:21:48 so when any of my BT things gets within reach, the light goes on Jan 08 00:21:57 :D Jan 08 00:22:34 works for FR, BT handsfree, BT keyboard, other phones Jan 08 00:22:42 very nice Jan 08 00:22:50 but computer must be on the whole time, so I hope it's not a 500W ultra pc ;) Jan 08 00:22:52 and very simple Jan 08 00:23:15 I need to have my server running anyway Jan 08 00:23:24 :) Jan 08 00:23:54 8 disks, ++ Jan 08 00:24:00 it sucks power Jan 08 00:24:07 and that is great Jan 08 00:24:18 because it is cold in Norway Jan 08 00:24:35 can you also move the shutters of your windows remotly? :D Jan 08 00:24:37 so I look at it like free heating Jan 08 00:24:43 hmm Jan 08 00:24:44 lol :) Jan 08 00:24:45 no Jan 08 00:24:51 I do not have shutters Jan 08 00:24:54 so you adjust room temperature by renice? :) Jan 08 00:25:04 lol Jan 08 00:25:07 xD Jan 08 00:25:08 (sorry for chipping in, but idling around and reading, yeah :) Jan 08 00:25:10 from my phone Jan 08 00:25:30 ssh quatrox.org Jan 08 00:25:53 :o Jan 08 00:25:57 then renice -7 Jan 08 00:26:00 and then break the dd if=/dev/sda of=/dev/null Jan 08 00:26:12 no need Jan 08 00:26:27 emerge -uD world Jan 08 00:26:33 ah :D Jan 08 00:26:36 :) Jan 08 00:27:11 I will also try to compile SHR Jan 08 00:27:20 to generate more heat Jan 08 00:27:48 hmm Jan 08 00:27:57 emerge shr Jan 08 00:32:12 someone make dev-virtual/shr ebuild :P Jan 08 00:34:15 :D with Ussd I can request remaining cash via ssh Jan 08 00:34:40 wow Jan 08 00:35:26 will the cash then be sent to you though regular postal mail? Jan 08 00:36:24 ow I meant I can ask how much EUR is still available for the phone Jan 08 00:36:32 reminds me that samba has support for ordering pizza Jan 08 00:36:56 elektranox: I know:P Jan 08 00:37:48 a pity, that there is no AT command to force the FR using a specific BTS Jan 08 00:38:02 then one would be able to locate the phone via gsm Jan 08 00:39:26 can't you do that anyway? Jan 08 00:39:41 well you can get the CID Jan 08 00:39:57 and the TA (Timing Advance) Jan 08 00:40:08 but you cannot kick it from the BTS? Jan 08 00:40:20 no Jan 08 00:40:28 :( Jan 08 00:40:30 you can only tell it to use a specific provider Jan 08 00:40:44 but not the BTS Jan 08 00:40:46 AFAIK Jan 08 00:41:54 and you cannot fake bad signal on a BTS since the firmware is locked down Jan 08 00:41:54 mh Jan 08 00:42:07 right ;) Jan 08 00:49:27 elektranox: let's build home sms server! plug gsm modem into computer, connect coffee-express to it and send sms "coffee" when you are returning to home ;D Jan 08 00:50:06 (well, i forgot how to type coffee. is it right?) Jan 08 00:50:17 I think so Jan 08 00:50:35 but I guess the bluetooth version of quatrox is better for this task :P Jan 08 00:51:17 i was trying "coffe", "cofie" but that versions didn't look good for me ;D Jan 08 00:52:10 :P in german its kaffee, so I guess it's coffee in english Jan 08 00:52:26 oh, i wanted to say it to quatrox ;D Jan 08 00:52:47 :P Jan 08 00:52:54 I read it anyway Jan 08 00:52:58 elektranox: in polish simply "kawa" ;) Jan 08 00:53:48 much more easy in Norwegian Jan 08 00:53:54 kaffe Jan 08 00:54:26 :D Jan 08 00:56:27 Kafi Lutz for everyone Jan 08 00:57:15 That's coffee with a shot of high-degree Schnaps Jan 08 00:57:40 karsk Jan 08 00:57:46 yes Jan 08 00:59:04 quatrox: Norwegian eh? Aquavit it is! Jan 08 00:59:05 :) Jan 08 00:59:41 yes Jan 08 00:59:48 Linje Aquavit Jan 08 01:00:17 which has crossed equator Jan 08 01:00:42 hehe Jan 08 01:00:48 do they still do that? Jan 08 01:00:59 yes Jan 08 01:01:11 but in huge tanks Jan 08 01:01:37 they load it on board Jan 08 01:01:50 send it on a trip Jan 08 01:01:59 and get it back some months later **** ENDING LOGGING AT Thu Jan 08 02:59:57 2009