**** BEGIN LOGGING AT Mon Feb 23 02:59:58 2009 Feb 23 03:06:26 PaulFertser: good morning :) what's the BT issue ? Feb 23 03:12:50 :D Feb 23 04:49:20 bug #666 is indeed fixed on moko11. Feb 23 05:20:34 :o Feb 23 07:45:15 morning Feb 23 07:47:53 morning Feb 23 07:48:13 no more OpenMoko development for some time. My FR went up in flames Feb 23 07:48:37 (well at least the SIM card is partially molten and the FR has a burnt black spot) Feb 23 07:48:49 :O Feb 23 07:48:51 sorry Feb 23 07:48:58 with no hardware tinkering at all... :-( Feb 23 07:49:08 yep, sucks Feb 23 07:49:10 it's not possible to made some "forensics" ? Feb 23 07:49:17 is all burn? Feb 23 07:49:30 no clue. It's within warranty, so I'll send it back Feb 23 07:49:40 and have them do the forensics... Feb 23 08:07:56 spaetz: maybe post this to the support list, so that it gets the proper attention Feb 23 08:20:11 wpwrak: Hey :) Feb 23 08:20:31 wpwrak: Let me explain it. Feb 23 08:22:00 wpwrak: The profile for bi-directional audio exchange over bluetooth is called SCO and is supported by all BT headsets/earpieces. A2DP (for high quality stereo playback) has no relation to it and doesn't support recording so it's irrelevant here. Feb 23 08:22:49 spaetz: ouch... just like that... :( Feb 23 08:23:20 wpwrak: SCO audio data can be routed over PCM interface of the chip or over HCI (be it UART or USB). The way to activate one or another way is vendor-specific. Feb 23 08:23:22 wpwrak: sent Feb 23 08:23:35 mrmoku: http://img147.imageshack.us/img147/8655/mokoburned.jpg Feb 23 08:24:54 * mrmoku moves his FR away... a little Feb 23 08:25:15 oops... hope this does not happen to me Feb 23 08:26:08 wpwrak: usually CSR chips come pre-programmed with a firmware that allows either "compat" way of doing things, that is route every SCO exchange over PCM, or "new correct" mode where SCO is routed via HCI by default but a special command can be used to route the next SCO exchange via PCM. Feb 23 08:26:17 never did any hardware tinkering so I feel pretty innocent... Feb 23 08:27:53 spaetz: hmm, looks as if B4902 burnt. that's the filter on USB VBUS Feb 23 08:28:49 wpwrak: the best info i could find is AN107 which is old (2002) and talks about the previous chip generations. But it looks that the information still applies. According to this AN the chip in FR comes pre-programmed with "compat" (PCM) setting. Feb 23 08:28:51 would explain why it did not recharge when I plug in USB power... Feb 23 08:29:25 spaetz: yup. but not why it went up in flames Feb 23 08:29:40 well, not flames really, but it the plastic did melt Feb 23 08:29:56 spaetz: that's bad enough ... Feb 23 08:30:01 yep Feb 23 08:30:30 PaulFertser: ... but things don't work as expected ? or do you need additional information on those modes ? Feb 23 08:31:00 PaulFertser: I don't really know if anyone in openmoko actually knows much about BT ... Feb 23 08:31:03 wpwrak: i changed the setting to the "modern" one and after that SCO data started to flow over HCI. But i still get no sound in my earpiece (though exactly the same bluez version and procedure works on my laptop). I have a detailed hcidump logs from both, but it doesn't look like anything is different. Feb 23 08:32:49 wpwrak: the way to change the setting is firmware-specific so i think OM has every right to ask the vendor (who could have mass-programmed any firmware in ROM) about exact instructions/commands wrt this routing. Feb 23 08:33:23 spaetz: ... you're lucky you have warranty. Feb 23 08:37:38 wpwrak: Chip revision: 0x0030 (BC4-ROM) and Build name: odj_4hci_rom_bt2.0_19p2_0503071059_encr128 2005-03-07 if it means anything. Feb 23 08:39:10 wpwrak: I asked * Johan Hedberg Feb 23 08:39:51 from the Bluez to look at the dumps but he said everything is looking right. The only difference is that i don't hear anything with freerunner it seems :) Feb 23 08:40:59 PaulFertser: hmm, let me check where marcel holtmann hangs out ... Feb 23 08:41:27 wpwrak: the same place i talked with Johan (jhe) #bluez on freenode Feb 23 08:41:35 :) Feb 23 08:41:39 PaulFertser, bluez4> Feb 23 08:41:40 ? Feb 23 08:41:44 shoragan: sure Feb 23 08:41:48 did you change the SCORouting? Feb 23 08:42:06 shoragan: it doesn't affect the routing itself, only a hint to other programs. Feb 23 08:42:15 ah, ok Feb 23 08:42:19 shoragan: i knew it from the sources and jhe confirmed it's like that. Feb 23 08:42:41 shoragan: bluez doesn't and is not going to support chip-specific mechanisms to affect SCO routing. Feb 23 08:42:48 can't we route it via the codec? Feb 23 08:42:50 That's what he told me. Feb 23 08:43:06 PaulFertser: or the "devel" list Feb 23 08:43:10 shoragan: we can route (like you obviously do), but we loose a whole lot of possibilities that way Feb 23 08:43:56 wpwrak: but don't forget that it depends on the firmware fully. How can Marcel know what firmware in our chips do? ;) Feb 23 08:45:21 wpwrak: FYI: logs sound working on laptop: http://pastebin.com/m6046adf6 and dump: http://pastebin.com/mf9e09ef . SCO working but no sound on freerunner: http://pastebin.com/m63870a3a and dump: http://pastebin.com/m7ba0182f Feb 23 08:45:45 PaulFertser: hmm, dunno :) okay, i'd suggest you post it on the list and let people try to figure it out Feb 23 08:46:01 wpwrak: on our devel list? Feb 23 08:46:17 wpwrak: why don't you just write an e-mail to vendor? ;) Feb 23 08:46:45 PaulFertser: if this doesn't work, write a mail we can then pass to CSR or whatever our contact is (could be an intermediary). but i don't really know what level of support we have here. Feb 23 08:46:58 shoragan: routing via HCI may be useful for capturing GSM conversations. Or to use BT headset to record a voice note (unrelated to GSM, thinking about general functionality). Or to use BT headset for VoIP. Feb 23 08:47:02 PaulFertser: yes, the openmoko devel list Feb 23 08:47:34 PaulFertser: i remember that some paths are unavailable but i don't know the details. i never touched BT. Feb 23 08:48:29 PaulFertser, i think all this should also be possilbe via the codec Feb 23 08:48:59 wpwrak: when we route sound from BT via PCM to the WM8753, there's practically no way to connect ADC to SoC, that's it. Our hardware doesn't have necessary switches to sync both interfaces in the same way. Feb 23 08:49:25 shoragan: please take a look at the sound routing diagram and DAI modes, you'll see what i'm talking about. Feb 23 08:49:53 ah, ok :( Feb 23 08:50:12 I'd like to be proven wrong here, but it doesn't seem it can happen. Feb 23 08:50:48 I looked at the datasheet, timing diagrams and such. Feb 23 08:51:09 Also Doc told me it's not possible with the hardware we have. Feb 23 08:52:47 wpwrak: anyway i think we have any right to request firmware-specific information from the vendor. It's unavailable now, but it should be available as OM is after non-NDA'd documentation for all. Feb 23 08:53:32 wpwrak: after all, they might have programmed any weird behaviour they liked in our chips. Feb 23 08:55:04 PaulFertser: understood. it's just one vendor channel i'm completely unaware of. the two i've "reopened" have been painful enough :) Feb 23 08:55:55 spaetz: did you use the wall charger or a pc ? Feb 23 09:01:04 wpwrak: I don't know exactly when it broke. I used Wall charger, I had it plugged into a comp, and in the car I had plugged it into a 12v->USB adapter (which I have been using for the N800 for a couple of months without problems) Feb 23 09:12:47 spaetz: thanks ! Feb 23 09:28:21 wpwrak: looks like Marcel Holtmann is availbale via IRC atm, so i'm going to first ask his opinion on my BT issues. Feb 23 09:33:05 PaulFertser: thanks ! i'm sure he's a lot more useful for this than i can be :) Feb 23 09:33:58 wpwrak: but jhe already told me that logs look ok, so i don't count much on Marcel to have more info. Feb 23 09:34:20 OTOH he's the man behind bluez SCO support :) Feb 23 09:50:02 raster: thanks for exporting elm_win_inwin_content_set :-) Feb 23 09:51:22 raster: inwins are supposed to work correctly? or are they still work in progres... Feb 23 09:51:32 +s Feb 23 10:01:10 mrmoku: they work for me Feb 23 10:01:11 :) Feb 23 10:01:48 raster: fine... I had a problem with an entry not resizing correctly... Feb 23 10:01:59 might as well be some 'don't know how to do it' problem :P Feb 23 10:02:15 well elementary-sms uses them Feb 23 10:02:17 and for entries too Feb 23 10:02:30 ok... will look there how you do it then :-) Feb 23 10:03:16 now back to checkbox Feb 23 10:03:21 then radio Feb 23 10:03:22 woots Feb 23 10:03:34 i need to create a bunch of genlist item styles Feb 23 10:03:39 the current one is just mostly testing Feb 23 10:03:50 (default) Feb 23 10:03:55 uhh, yeah... list is another thing... :-) Feb 23 10:04:04 theres list for quick dirty short ones Feb 23 10:04:04 list of key/value pairs... Feb 23 10:04:16 and genlist for when its going to be either more flexible (different style entrties) Feb 23 10:04:17 or longer Feb 23 10:04:24 yep the quick/dirty list has widget/label/widget as fixed content? Feb 23 10:04:29 moin Feb 23 10:04:42 yes Feb 23 10:04:49 so I could use the first arbitrary widget as label for a key and the label for the value? Feb 23 10:04:50 tho the widget and widget are optional Feb 23 10:04:55 ok Feb 23 10:04:58 and the theme can shuffle them around Feb 23 10:05:00 if it wants Feb 23 10:05:08 but it has 3 elements per list item Feb 23 10:05:12 an icon, an end and a label Feb 23 10:05:17 icon and end are widgets Feb 23 10:05:22 joerg_42: yo Feb 23 10:05:30 so icon can be a label Feb 23 10:05:41 it could Feb 23 10:05:47 but it'd look straneg when selected Feb 23 10:05:52 strange Feb 23 10:06:09 or more accurately.. it wouldnt "look" Feb 23 10:06:12 it'd be black on black Feb 23 10:06:13 :) Feb 23 10:06:18 ouch :-) Feb 23 10:06:54 so... label for key and end for value would "look"? Feb 23 10:07:41 same with end Feb 23 10:07:43 like icon Feb 23 10:07:45 black on black Feb 23 10:07:48 coming right up Feb 23 10:08:05 you could put the label in a frame... Feb 23 10:08:33 hmm, might as well take a closer look at genlist then... Feb 23 10:08:48 genlist is using the same list theme items right now for default Feb 23 10:08:54 widget/label/widget Feb 23 10:08:58 but its meant to allow multiple styles Feb 23 10:09:04 so one with multiepl labels Feb 23 10:09:10 and multiple "icons" Feb 23 10:09:25 and "states" (ie check or radio elements than turn on/off) Feb 23 10:09:35 i havent tested it all - but it should work as its generic Feb 23 10:10:08 ok Feb 23 10:10:37 as i said - i need to do more styles to "show them off" Feb 23 10:10:56 and provide a base set of them that cover a useful palette of items people would want to pack in Feb 23 11:00:35 PaulFertser_: any tips on how to debug multithreaded stuff with gdb? Feb 23 11:01:50 mrmoku: i know that gdb stops all threads when the process is stopped. So, you place the breakpoint and then look backtrace for all threads, that's how i see it. Feb 23 11:02:22 mrmoku: info threads shows threads, 'thread X' switches to thread X Feb 23 11:02:48 ahh, so it would be thread X to switch into it and the bt to get the backtrace for that thread... Feb 23 11:03:47 mrmoku: yep Feb 23 11:04:02 warning: Can't attach process 2559: No such process Feb 23 11:04:06 can I ignore that one? Feb 23 11:04:38 and I get Cannot find new threads: generic error Feb 23 11:04:43 though info threads works Feb 23 11:05:12 every time u spawn a thread... god kills a kitten Feb 23 11:05:49 lindi-: here's inkscape source for you: http://wiki.openmoko.org/wiki/Image:WM8753_routing_diagram.svg but it doesn't help much as only alsa control names are vector, the initial diagram is raster (;)), placed in the separate layer. Feb 23 11:05:49 every time you try and beat evas's software with hardware rendering.. god also kills a kitten it seems Feb 23 11:06:14 raster: still struggling with hardware GL magic slowdown? :) Feb 23 11:06:19 raster: well... I'm trying to debug that damn segfault in our sms app... and that might be related to the multithreaded nature of it Feb 23 11:06:38 * mrmoku would prefer to not have multithreading and more kittens instead... my wife might differ ;) Feb 23 11:07:37 PaulFertser_: i see one large red X as a sign of missing bitmap :) Feb 23 11:07:39 mrmoku: dont say i didnt warn u about the kittens :) Feb 23 11:07:50 * raster avoids threads like the plague.. for precisely this reason Feb 23 11:07:54 bugs happen Feb 23 11:07:55 lindi-: you'd download it first Feb 23 11:07:56 thats life Feb 23 11:07:57 raster: have to tell that quickdev and Ainulindale ;) Feb 23 11:07:59 PaulFertser_: '** (inkscape:22024): WARNING **: Unable to open linked file: /home/pavel/openmoko/WM8753_control_diag.png' Feb 23 11:08:19 lindi-: shit, my inkscape skills suck :) Feb 23 11:08:24 bugs that are triggered by threads being bad and doing things like forgetting to lokc thigns and accessing shared data are bitch-hard to track and reproduce and fix Feb 23 11:08:30 they suck like 10x the effort of a normal bug Feb 23 11:08:43 PaulFertser_: should 'Image:WM8753_routing_diag.png' be in the wiki somewhere? Feb 23 11:08:53 at leats in my experience Feb 23 11:09:08 the only way to living kittens and threads is to pretty mcuh treat threads like processes Feb 23 11:09:17 lindi-: that's the same file that on the wiki, http://wiki.openmoko.org/wiki/Image:WM8753_control_diag.png Feb 23 11:09:24 keep them totally isolated in their own wor5lds and have very well abstracted high level comms between them Feb 23 11:09:37 PaulFertser_: yeah. i am just beginnign to think "thats how it is" Feb 23 11:09:47 mind u i did find badness in the gles driver for the 6410 Feb 23 11:10:03 i can double its swapbuffers just by stoppingg to do things in a silyl way Feb 23 11:10:30 they do too many copies to get pixels to the screen Feb 23 11:10:42 but thats not the real major problem Feb 23 11:10:48 the real one is just that drawing is slow Feb 23 11:10:56 software beats it in spedd Feb 23 11:11:01 even in "Regular usage" Feb 23 11:11:19 expedite is a bit of a nasty artificial benchmark designed to find the nasties in a rendering pipeline Feb 23 11:11:30 but even elementary under gles is smoother with software Feb 23 11:11:55 and i'm fresh out of ideas on how to improve it other than droppign quality (rgb565/rgb4444 textures, 16bpp) Feb 23 11:12:23 but thats just wrong as evas software renderingin 32bpp with 32bpp image srcs and dest is doing better. Feb 23 11:12:59 a 3d hw unit should at LEASt keep up (eg if software is basically pushing the memory bandwidth limits of the system then i doubt a 3d unit can do better in the simpler ops) Feb 23 11:13:03 but it should keep up Feb 23 11:13:49 IBar in illume top shelf works good ;D Feb 23 11:14:29 PaulFertser_: ok Feb 23 11:14:31 dos1: how? Feb 23 11:14:38 mrmoku: http://scap.linuxtogo.org/files/ecac8b7e18652707dabbf1c68de9f176.png Feb 23 11:15:33 weee Feb 23 11:15:35 16bit engine Feb 23 11:15:40 ewwwwwwwwwwwwwwww Feb 23 11:15:41 yuk yuk Feb 23 11:16:09 * raster shrinks away Feb 23 11:17:34 raster: works really faster and on gta02 screen lower quality isn't that noticeable as on pc :] Feb 23 11:18:07 him notice it Feb 23 11:18:23 so your argument is invalid Feb 23 11:18:52 Hire: no, i said "it isn't (...) as on pc" - i said that on pc it is very noticeable :P so you are wrong Feb 23 11:18:55 ;] Feb 23 11:19:13 dos1: i've seen it on gta02 Feb 23 11:19:18 and i notice it. Feb 23 11:19:19 ewwww Feb 23 11:19:40 and it makes me go... Feb 23 11:19:43 ewwwwwwww Feb 23 11:21:32 raster: btw.... elementary_test is segfaulting Feb 23 11:22:04 YAY! Feb 23 11:22:35 yay you know that... or yay how nice? :P Feb 23 11:22:43 yaya how nice Feb 23 11:22:54 :)wprking here without complaint under valgrind Feb 23 11:22:58 it segfaults when closing the list test Feb 23 11:23:00 at least window coems up and i can scroll... Feb 23 11:23:05 mv elementary_test elementary_crash Feb 23 11:23:06 list or genlist? Feb 23 11:23:08 problem solved Feb 23 11:23:13 list Feb 23 11:23:31 ok Feb 23 11:23:31 got it Feb 23 11:23:40 will need to look into it Feb 23 11:23:50 its all three list tests... Feb 23 11:24:16 nice to have a problem you can reproduce :-) Feb 23 11:27:27 mrmoku: what about cpu hogging btw? ;) Feb 23 11:28:04 PaulFertser_: hoggidihoggidi... thats why people complain SHR is slow... :( Feb 23 11:30:28 Sometimes happens to my (Ex-)FR and sometimes not. Feb 23 11:30:39 It's really a weird thing, that CPU hog Feb 23 11:34:10 mm thats how i like them Feb 23 11:34:19 mrmoku: i mean raster reproduced it :) Feb 23 11:34:19 mrmoku: thats how i like themn Feb 23 11:34:20 :) Feb 23 11:34:35 PaulFertser_: the shr image i built does it Feb 23 11:34:40 but i frankly have had time to go back to it Feb 23 11:35:37 raster: please don't think that anybody pushes you or smth, just don't forget to finally find some time to look into it, ok? :) Feb 23 11:36:03 time will come maybe in may Feb 23 11:36:10 until then.. i'm booked solid Feb 23 11:36:20 with timelines and a work queue Feb 23 11:36:29 of course u guys benefit from that queue Feb 23 11:36:34 raster: hope you'll enjoy it :) Feb 23 11:36:34 but not the cpu usage bug Feb 23 11:44:57 spaetz: btw, when usin the 12v adapter, did you manually adjust the charge current ? Feb 23 11:46:08 nope. Running SHR, but no further fiddling with stuff Feb 23 11:46:35 spaetz: so the car charger charged only at 100mA ? Feb 23 11:47:36 no clue. it's a 12V output (max 500mA rated) charger, nothing fancy, but it worked fine with all my other devices... Feb 23 11:48:45 err, make that 5V output, of course :) Feb 23 11:49:07 spaetz: that explanation would have been just too convenient :) Feb 23 11:49:37 hehe, I guess so Feb 23 12:09:14 joerg_42: most people reported that when they don't turn off the Amp they can hear the sound from both headphones (and it's not too loud) and speaker at the same time. Feb 23 12:10:16 from datasheet of amp it seems impossible Feb 23 12:10:55 hm, maybe right headphone isn't switched off Feb 23 12:11:08 wpwrak: btw, if GTA03 will have headphones/headset it'd be cool to wire up the headset in a way allowing independent control of the speaker (imagine i listened to the music and forgot to take out the headphones from the jack but took them out from my ears. Now i won't be able to hear the incoming call indication sound). Feb 23 12:11:50 yup I mentioned that long ago. won't come though it seems Feb 23 12:12:27 joerg_42: won't have a headset or won't have a sane wiring? Feb 23 12:12:40 Maybe wpwrak will be able to bring attention... Feb 23 12:13:19 joerg_42: but really, sound from both the speaker and headset is possible, i think i even heard that myself, gotta test later. Feb 23 12:13:24 we should have swapped earpiece and speaker, as wm8753 is same setup powering down speaker for headphones. earpiece would be a better choice Feb 23 12:14:04 gta03 won't use this nonsense additional amp anymore ;-) Feb 23 12:14:32 I always wondered why we needed it for GTA02 Feb 23 12:16:41 PaulFertser_: there allegely is a screwed way to connect headset to GTA03. sane wiring though: no not really Feb 23 12:17:16 gta03 will instead use yo momma Feb 23 12:17:27 joerg_42: sane and OM seem to be more and more mutually exclusive :( Feb 23 12:17:56 a wiring resembling the one of e.g. samsung omnia isn't sane Feb 23 12:18:47 joerg_42: have you seen Andy's reply about my diagram on community list? Feb 23 12:18:59 nope Feb 23 12:19:26 and probably I don't want to see it Feb 23 12:20:27 Then for those who want: http://article.gmane.org/gmane.comp.handhelds.openmoko.community/37100 and see my reply to it. Basically he said "Matt and Candy here in Taipei are making a spreadsheet about this too" Feb 23 12:20:49 No single word from them about it, what a lack of communication! Feb 23 12:21:30 nah Feb 23 12:21:46 freesmartphone.org: 03jluebbe 07python-helpers * rf5060958b6ee 10/pycd/ (pyc.c pycd.py): pyc: add a unix socket interface Feb 23 12:21:47 freesmartphone.org: 03jluebbe 07python-helpers * r382de6e12a42 10/pycd/ (pyc.c pycd.py): pyc: support waiting until the child is done Feb 23 12:21:48 freesmartphone.org: 03jluebbe 07python-helpers * r79019eca43bc 10/pycd/TODO: pyc: update TODO Feb 23 12:21:49 freesmartphone.org: 03jluebbe 07python-helpers * r79042eaa15a2 10/pycd/ (pyc.c pycd.py): pyc: update license Feb 23 12:21:51 Can someone tell me how to disable otimed - the @##$%# thing wants to overwrite my timezone with Feb 23 12:22:00 one 300km away ... Feb 23 12:22:19 sorry, not 300, 3000+km away Feb 23 12:22:21 billk: look at frameworkd.conf at the ZoneSources Feb 23 12:23:34 PaulFertser_: abandon any high expectations on that (commu and spreadsheet) Feb 23 12:24:22 either Ive got a part frameworkd.conf or its a default and not mentioned - will search some more Feb 23 12:25:29 btw sounds like a typical answer. "nice work but we don't need it, we already started doing same/better project" Feb 23 12:26:51 joerg_42: can't imagine any useful spreadsheet for that stuff anyway. Feb 23 12:27:13 some guys feel really unhappy on finding the don't have full control about any detail Feb 23 12:28:28 PaulFertser_: tkx, found it on the wiki - added it to my file. Feb 23 12:30:22 PaulFertser_: you would be amazed finding what different types of documents can be done abusing spreadsheets Feb 23 12:31:21 PaulFertser_: spreadsheets are about the most abused document creation apps on the planet Feb 23 12:31:27 or for that matter anything-creation Feb 23 12:31:34 anyway for me that means I have to speed my RFC ;-) Feb 23 12:31:43 peolpe do anything from presentations to comapny finances (all of them) Feb 23 12:31:47 to stock trading apps Feb 23 12:32:50 to word type text editors by using formulas and macros and 1char per cell. No joke! Feb 23 12:33:15 hmm Feb 23 12:33:18 maybe Feb 23 12:33:20 both are abused tho Feb 23 12:33:25 siemens guy Feb 23 12:33:36 In shr-unstable contacts, I cant add a new contact because I cant Feb 23 12:33:49 move to the number field to add the number Feb 23 12:34:00 is there a magick wat to do this? Feb 23 12:34:12 nah, you didn't get it. this siemens guy was an excel crack and built a editor out of excel Feb 23 12:34:25 billk: tap number field many times :P Feb 23 12:35:20 dos1: tkx, was magick after all :) Feb 23 12:44:34 (excel editor) well I mustn't frown, I created even an interpreted programming language running off a hiden table of a spreadsheet document ;D Feb 23 12:45:19 incl variables and structure elements Feb 23 12:46:54 search for XMLMaker, maybe it's mentioned somewhere still Feb 23 12:50:07 hmm seems someone hijacked the name, mine was 2000~2001 Feb 23 12:51:39 PaulFertser_: (no comm) well, they mentioned it in time to save you from doing wasted work :) Feb 23 12:52:11 PaulFertser_: (audio) hmm yes, that's a sensible use case Feb 23 12:53:22 (xmlmaker) keywords: "XMLMaker excel gsdc" Feb 23 12:53:58 xml! eek! Feb 23 12:54:03 lol Feb 23 12:54:04 * raster hides from the xml monster Feb 23 12:54:20 wellformed, wellformed Feb 23 12:54:42 XD Feb 23 12:55:01 I am starting on learning the C program language and want to have some existing code to pour over, SHR is developed mostly in C correct? Feb 23 12:55:24 there are some B2B standards using XML Feb 23 12:56:34 projektdotnet: I'd discourage you to start learning C based on that though ;-) Feb 23 12:57:59 joerg_42: i don't intend to learn all off that code, just look at the code for curiosity sake, reading books to learn but some examples don't show real world examples Feb 23 12:58:42 maybe find the sourcecode of some simple shellcommand for example Feb 23 12:58:54 joerg_42: any you'd specifically reccomend? Feb 23 12:59:03 raster: what does priority do in E_Config_Module in illume config? Feb 23 12:59:18 nah, take any, like less, or ls Feb 23 12:59:28 joerg_42: thanks :) Feb 23 12:59:55 * mrmoku is diffing illume config from shr with the original one... Feb 23 12:59:57 maybe even more simple: cat Feb 23 13:00:03 affects ordering of what modules are loaded first Feb 23 13:00:43 will take a look, thanks :) Feb 23 13:00:52 original for gadman is -100 shr sets it to 100... hmm... ??? Feb 23 13:01:08 shouldnt make a difference Feb 23 13:01:12 nothnig uses gadman anyway Feb 23 13:01:18 ahh, ok :-) Feb 23 13:01:22 but.. it can cause issues if u had gadgets on the desktop Feb 23 13:01:25 mrmoku: while you are in the illume default config. Mind dropping the default framerate somewhat? Feb 23 13:01:30 or tee Feb 23 13:01:30 ie gadman comes up after the moduels that create those gadgets Feb 23 13:01:35 Just to shut down all that nagging in the mailing lsits? Feb 23 13:01:46 s/dropping/decreasing/ Feb 23 13:01:50 spaetz: default is 30fps whihc si pretty normal Feb 23 13:01:53 which Feb 23 13:01:55 is Feb 23 13:02:12 unless u mean the cpu use Feb 23 13:02:23 and al u do is recduce that proportionally Feb 23 13:02:30 raster: CPU use Feb 23 13:02:40 raster: use predictive kbd? ;D Feb 23 13:02:41 u arent really going to make a dent in it without heavily affecting smoothness/usability Feb 23 13:02:43 another difference is shr adding lots of orient int:0 Feb 23 13:02:45 people on the mailing list state it helps CPU usage. Feb 23 13:03:00 joerg_42: using normal kbd man.. not an n810 like u Feb 23 13:03:07 lol Feb 23 13:03:19 XD Feb 23 13:03:23 raster, and people keep on nagging about it: http://article.gmane.org/gmane.comp.handhelds.openmoko.community/37132 Feb 23 13:03:24 spaetz: and will slow down other stuff too Feb 23 13:03:32 mmh, k Feb 23 13:03:36 again wont help much until the actual cause is found Feb 23 13:04:01 but up to you Feb 23 13:04:08 it wont stop people nagging Feb 23 13:04:09 trust me Feb 23 13:04:13 unles u drop it to like 1fps Feb 23 13:04:16 then u'll get new nags Feb 23 13:04:24 like "why is it so slow?" Feb 23 13:04:24 :) Feb 23 13:04:58 * mrmoku refuses to set software_16 as default ;) Feb 23 13:05:09 mrmoku: good boy! Feb 23 13:05:10 :) Feb 23 13:06:30 :) Feb 23 13:07:21 well... it looks dead ugly... Feb 23 13:07:37 :) Feb 23 13:09:03 raster: what about adding orient for gadcon clients where in the original is no orient... no problem? Feb 23 13:10:02 mostly it wont make a diff Feb 23 13:10:06 except some gadgets Feb 23 13:10:10 like ibar and ibox Feb 23 13:10:17 they will want to irient the wrong way Feb 23 13:10:19 orient Feb 23 13:10:29 actually wait Feb 23 13:10:33 oh no Feb 23 13:10:35 never mind Feb 23 13:10:39 that should get changed runtime Feb 23 13:24:47 freesmartphone.org: 03mickey 07python-helpers * r41bcdfdfa608 10/mickeyterm/mickeyterm: add debug log for mickeyterm Feb 23 13:32:41 freesmartphone.org: 03mickey 07mickey/ogsmd-new-timeout * r5509afd1d414 10framework/framework/subsystems/ogsmd/modems/abstract/unsolicited.py: ogsmd: don't deactivate a pdp object that is not yet existing Feb 23 13:32:42 freesmartphone.org: 03mickey 07mickey/ogsmd-new-timeout * reed87f3e3287 10framework/framework/subsystems/ogsmd/ (3 files in 2 dirs): ogsmd: remove per-command timeouts Feb 23 13:32:44 freesmartphone.org: 03mickey 07mickey/ogsmd-new-timeout * r6495c7196893 10framework/ (3 files in 2 dirs): (log message trimmed) Feb 23 13:32:44 freesmartphone.org: ogsmd: Rewritten timeout handling and unsolicited message handling. Feb 23 13:32:46 freesmartphone.org: We now maintain a list of valid prefixes for every command Feb 23 13:32:48 freesmartphone.org: in the Queue. Once the parser processes a complete line, it checks whether Feb 23 13:32:50 freesmartphone.org: this line contains expected results. With this, it's safe to operate modems Feb 23 13:32:52 freesmartphone.org: that mix unsolicited and solicited responses. It's also a safeguard against Feb 23 13:32:54 freesmartphone.org: race conditions when talking to the modem over buffered lines. Feb 23 13:34:42 freesmartphone.org: 03mickey 07mickey/ogsmd-new-timeout * r63a030a47c2a 10framework/framework/subsystems/ogsmd/modems/ti_calypso/channel.py: Feb 23 13:34:42 freesmartphone.org: ogsmd: [TI CALYPSO] send %CUNS=0 on startup. It now sends unsolicited responses Feb 23 13:34:42 freesmartphone.org: directly once they're generated, which leads to potentially lots of mixed Feb 23 13:34:42 freesmartphone.org: solicited/unsolicited responses. This should be a nice test-case for our new Feb 23 13:34:46 freesmartphone.org: handling. Feb 23 13:47:59 freesmartphone.org: 03mickey 07mickey/ogsmd-new-timeout * r2072eae52f1a 10framework/framework/subsystems/ogsmd/gsm/channel.py: Feb 23 13:47:59 freesmartphone.org: ogsmd: improve AUTOPREFIX handling. We can compute them for combined AT commands now Feb 23 13:47:59 freesmartphone.org: as well (AT+FOO;+BAR;...) Feb 23 14:05:05 freesmartphone.org: 03mickey 07mickey/ogsmd-new-timeout * r9acdee268f9d 10framework/framework/subsystems/ogsmd/gsm/parser.py: ogsmd: fix new parser for solicited messages with PDUs (+CMGL, what else?) Feb 23 15:30:51 wpwrak: (no comm from TPE, wasted work, mentioned in time) actually, not. I still have no idea what are they doing and what are they going to do and i'm unsure that they're doing anything worthy at all wrt alsa, to be honest. Feb 23 15:32:37 anyone know where i can get ecalc? http://scap.linuxtogo.org/files/2b59f0f8e666d6478fbe3c5b213c3cb9.png Feb 23 15:33:31 wpwrak: (wlan) so you basically decided that we can wait for atheros's fixes forever and implemented a workaround to promptly reset a module instead? Well, that'll work better than nothing. Thanks for trying to solve all that issues with stupid-too-smart firmware :) Feb 23 15:37:27 s/can/can't/ Feb 23 15:37:28 PaulFertser_ meant: wpwrak: (wlan) so you basically decided that we can't wait for atheros's fixes forever and implemented a workaround to promptly reset a module instead? Well, that'll work better than nothing. Thanks for trying to solve all that issues with stupid-too-smar... Feb 23 15:53:14 bzzbot needs splitlong.pl *g* Feb 23 16:11:29 juri_: btw, dos1 mentioned today a way to switch to contact name field. It's tapping on number enough times :) Feb 23 16:12:57 15:33 < billk> In shr-unstable contacts, I cant add a new contact because I cant move to the number field to add the number Feb 23 16:13:03 15:34 < dos1> billk: tap number field many times :P Feb 23 16:26:06 many times. my blanket worked! :) Feb 23 16:27:37 * mwester wishes for a magic blanket that can make openmoko products work too... Feb 23 17:24:05 raster: btw. ... e consuming CPU is *not* a SHR specific problem... FSO does it also. (just retried). Feb 23 17:24:18 raster: just wanted to mention that while you're asleep ;) Feb 23 18:02:33 anyone knows how to contact John Beaven, the developer of MOFI-Wifi? Feb 23 18:10:24 say i want to wake my phone from suspend every 10 mins, whats the best way to do it? Feb 23 18:11:28 heyho Feb 23 18:32:28