**** BEGIN LOGGING AT Fri Dec 07 02:59:57 2007 Dec 07 03:01:43 wrong time of day? Dec 07 03:01:44 :) Dec 07 03:01:47 quiet Dec 07 03:02:52 most people are in europe Dec 07 03:03:02 I'm on the west coast of the US Dec 07 03:03:18 OK Dec 07 03:03:31 Is there a pictograph spec in openmoko? Dec 07 03:04:04 but we're here, and awake, just not talking. Dec 07 03:05:42 Ok :) Dec 07 03:06:02 what's the wiki page for wishlists so I can peruse and see if it was added Dec 07 03:07:40 I was just hearing about a method to have a phone locked by a pictograpgh instead of a PIN which I thought should probably be looked at for openmoko Dec 07 03:28:10 ScaredyCat, ping Dec 07 05:05:42 ScaredyCat, ping Dec 07 05:06:19 unrelated: anyone know how to make the /etc/resolv.conf not reset? Dec 07 05:57:53 the new lock screen is awesome Dec 07 06:00:57 :) Dec 07 06:58:45 Web 2.0 humor: http://www.youtube.com/watch?v=fi4fzvQ6I-o Dec 07 07:03:01 ok, this is really weird Dec 07 07:03:30 Neo was totally switched off, yet the sim card was active - I could hear it ring from my other phone! Dec 07 07:04:28 I could hear the sounds from my speaker, that an incoming call is coming to a device nearby Dec 07 07:04:55 No, that's not so weird.. Dec 07 07:05:14 There's something broken with the shutdown, or at least it used to be. Not sure if it's fixed already Dec 07 07:05:24 CM, it's not :-( Dec 07 07:05:33 heh Dec 07 07:05:37 Earlier at least, when you powered off the Neo, the gsm modem was not turned off correctly Dec 07 07:05:45 And it will drain your battery :( Dec 07 07:05:45 thats new to me Dec 07 07:05:59 ahven, your phone wasn't switched off completely - you have to remove the battery for a few seconds to do so Dec 07 07:06:06 o Dec 07 07:06:06 You need to take the battery out ~a minute Dec 07 07:06:09 +k Dec 07 07:06:48 josch: Yeah, seconds is right, but 1 min + 1h charge to get it back running seems to be easier for people to remember ;) Dec 07 07:07:01 CM, right ^^ Dec 07 07:07:42 mwester had some script that powered off the gsm, wonder why it's not used.. Dec 07 07:09:25 but even if gsm is turned off - there is still the audio chip remaining :-/ Dec 07 07:22:01 Hi, when I update build 2007.2 using make update it will gives me error like mtn: misuse: database /home/qvantel/moko/OE.mtn is laid out according to an old schema Dec 07 07:22:01 mtn: misuse: try 'mtn db migrate' to upgrade Dec 07 07:22:01 mtn: misuse: (this is irreversible; you may want to make a backup copy first) Dec 07 07:22:01 mtn: misuse: database /home/qvantel/moko/OE.mtn is laid out according to an old schema Dec 07 07:22:01 mtn: misuse: try 'mtn db migrate' to upgrade Dec 07 07:22:03 mtn: misuse: (this is irreversible; you may want to make a backup copy first) Dec 07 07:22:05 make: *** [update-mtn] Error 1 Dec 07 07:23:44 CM, Any Idia Dec 07 07:24:11 atul: Did you try to migrate? Dec 07 07:24:13 :) Dec 07 07:24:48 atul: You might have to use mtn --db OE.mtn db migrate Dec 07 07:24:57 Have you updated monotone? Dec 07 07:25:04 cm, No Dec 07 07:25:26 CM, I dont have any idea about how to migrate Dec 07 07:25:36 Just run that command :) Dec 07 07:25:52 But if I remember correctly it's fairly slow.. Dec 07 07:25:52 CM, ok am trying Dec 07 07:26:03 CM, Lets see Dec 07 07:27:00 CM, It seems very slow you are right Dec 07 07:27:29 Maybe better to download a new OE.mtn from openembedded? Dec 07 07:29:56 cm, how to download it ? Dec 07 07:30:54 atul: wget http://downloads.openmoko.org/OE/snapshots/OE.mtn.bz2 :) Dec 07 07:31:11 CM, Thanks Dec 07 07:34:20 moin Dec 07 07:39:33 do also non-germans say "moin" for goof morning or what?? Dec 07 07:39:43 s/goof/good/ Dec 07 07:40:24 only if you repeat it often enough ;) Dec 07 07:40:51 blindcoder: It sounds finnish to me ;) Dec 07 07:42:08 I only know this word from nerthern germans living at the north sea :-D Dec 07 07:42:19 ack... s/nerthern/northern/ Dec 07 07:43:11 * CM says: Morn Dec 07 07:48:16 CM, Is the new build is working fine did you have any idea ? Dec 07 07:49:12 You mean the 2007.11 image? Dec 07 07:49:57 I've not been able to test myself yet, but I've heard there's a problem with pin codes in the dialer in it Dec 07 07:50:03 That have been fixed in later images Dec 07 07:53:38 CM, Not mean that i mean to say build 2007.2 is working fine or not means i am able to download it or not ? Dec 07 07:54:23 hehe Dec 07 07:55:07 i deed someone with debugboard from Poland Dec 07 07:55:12 CM: just wanted to ask about the pin codes :P Dec 07 07:55:30 s/deed/need/ Dec 07 07:55:30 NineX meant: i need someone with debugboard from Poland Dec 07 07:55:32 NineX: You didn't got the FPC for debug board? Dec 07 07:55:39 sometimes it works ok, sometimes it creates 4-5 requests Dec 07 07:56:39 ahven, +1 to this bug Dec 07 07:57:02 but if it is fixed in the later images, then all is well Dec 07 07:58:41 and once it started ringing just when I unlocked the screen - and no incoming call window Dec 07 07:58:48 are sed regexes PCRE compliant? it seems not Dec 07 08:03:40 NineX: There are some boards in Poland, but I'm not sure where Dec 07 08:04:01 CM: i will speek with hrw Dec 07 08:04:07 :) Dec 07 08:04:16 i have 2 phones bricked right now Dec 07 08:04:36 Ouch, how did that happen? Dec 07 08:04:46 Sure it's not just drained batteries? Dec 07 08:05:01 CM: i'm sure Dec 07 08:05:33 :( Dec 07 08:05:44 Flashed bad uboots on both? Dec 07 08:09:41 CM: 1 phone is from my friend Dec 07 08:45:06 gneeeh! Dec 07 09:21:33 CM, Now i get this error, mtn: error: database /home/qvantel/moko/OE.mtn lacks some cached data Dec 07 09:21:33 mtn: error: run 'mtn db regenerate_caches' to restore use of this database Dec 07 09:21:33 mtn: error: database /home/qvantel/moko/OE.mtn lacks some cached data Dec 07 09:21:33 mtn: error: run 'mtn db regenerate_caches' to restore use of this database Dec 07 09:21:33 make: *** [update-mtn] Error 1 Dec 07 09:21:34 any Idia Dec 07 09:23:40 atul: How about "mtn db regenerate_caches", as mtn suggests? Dec 07 09:24:56 abrasx_, Yes I tried this command but it gives me error like atull@rautl-desktop:~/moko$ mtn db regenerate_caches Dec 07 09:24:56 mtn: regenerating cached rosters and heights Dec 07 09:24:56 mtn: misuse: no database specified Dec 07 09:29:51 atul: you have to specify db, something like --db Dec 07 09:30:04 atul: --db OE.mtn Dec 07 09:32:07 CM, ok Dec 07 09:32:30 CM, it says command not found Dec 07 09:33:02 The full would of course be: mtn --db OE.mtn db regenerate_caches Dec 07 09:33:23 CM, thanks Dec 07 09:39:34 there needs to be an opie image :> Dec 07 09:42:11 is there an image w. qtopia & ipkg ? Dec 07 09:47:57 xcasex: ipkg for qtopia? what for? Dec 07 09:48:07 xcasex: they do not provide feeds Dec 07 09:49:57 hrw: i think you misunderstand me :) ipkg for the shell and ssh access to the phone. i dont care much if the ipkg feeds are exposed via the gui :) Dec 07 09:50:21 xcasex: ? can you explain more Dec 07 09:51:24 hrw: alright i'll elaborate. ipkg does package management, right? i want ipkg on the moko, like i get when i run the stock moko, but i want to run qtopia. and the qtopia image seemingly lacks ipkg :) Dec 07 09:52:10 You want to install programs using ipkg, basically? Dec 07 09:52:33 xcasex: but what for you will use ipkg with qtopia? Dec 07 09:52:46 xcasex: qtopia cames as images without feeds Dec 07 09:52:50 qtopia images have ipkg on them. Dec 07 09:53:43 CM: precisely Dec 07 09:54:02 ljp: aaah thanks and i can access the moko via ssh as usual? Dec 07 09:54:35 xcasex: ScaredyCat builds current images if there's no new official snapshot out Dec 07 09:54:47 CM: of qtopia Dec 07 09:54:47 ? Dec 07 09:54:53 yers Dec 07 09:54:54 i know he does the moko ones Dec 07 09:55:01 awesome CM and theyre on? Dec 07 09:55:38 xcasex: Same folder, http://buildhost.automated.it/OM2007.2/ Dec 07 09:55:39 http://buildhost.automated.it Dec 07 09:55:52 Latest is from Nov 28th I think Dec 07 09:56:39 ty CM :D Dec 07 10:01:39 xcasex: the qtopia images are like the openmoko images but with qtopia :) Dec 07 10:02:08 :) Dec 07 10:03:24 so they are better :) Dec 07 10:05:02 how much of carsten heitzlers work has made it downstream? Dec 07 10:06:45 ljp: :P ;) Dec 07 10:07:30 xcasex: You mean to OpenMoko? The e libs are mostly there, but I don't think he's been allowed to work on anything moko specific yet Dec 07 10:10:24 CM: oic. hmm did he get hired to do moko specific tasks or was he relegated to some other fic project? Dec 07 10:11:42 I think he said some other task right now, but I don't know the details Dec 07 10:12:59 * CM goes home to fill the trailer and move to Stockholm Dec 07 10:14:27 good luck CM Dec 07 11:13:44 http://www.theunwired.net/?item=review-vodafone-lg-ks20-windows-mobile-smartphone - another nice formfactor ;( Dec 07 11:16:44 I don't like it... Dec 07 11:29:26 hi Dec 07 11:33:09 hello PBeck Dec 07 11:36:26 hallo OJW Dec 07 12:00:21 hmm..about that gsm problem...if switch bt and gsm on and then hold down the power button until it forces it to shut down w/o going through the shutdown scripts, then the gsm & bt modules are still using power, while the rest of the device is powered off? Dec 07 12:03:10 sounds like there's some power management ASIC somewhere that wants to be programmed Dec 07 12:30:13 i have question. if i fully erase nandflash, i found how to recreate bad block tables and reflash u-boot, but i don't found noting about splash. where i founs splash to put into nand ? Dec 07 12:34:07 NineX: check out the openmoko/flash.sh sript in qemu-neo1973 Dec 07 12:47:55 ninex: loook at the download.sh and flash.sh of qemu-1973 Dec 07 12:48:04 hm, my neo dosn't charge anymore when usb is connected and linux is running (i mean it charges in u-boot with that menu entry from the wiki). how can i determine if it charges? the voltage is going down... Dec 07 12:48:08 ninex: it contains the url and the conveter command Dec 07 12:48:30 emdete: it's documented on the kernel page, there is a sysfs entry 'chgmode' and 'state' Dec 07 12:48:57 chgmode is 'idle'. state is fast_enabled Dec 07 12:49:24 i started with voltage 4,2 or something, now its 3.7 Dec 07 12:58:16 LaF0rge: the kernel page says battemp is in centigrades. what's that? my shows 15... what would that be in celsius? Dec 07 12:58:39 centigrade = celsius Dec 07 12:59:52 15 grad celsis? i dont think so :( the battery is more than hand-warm... Dec 07 13:00:47 also chgcur is 17127 while battemp is decline... Dec 07 13:01:22 http://en.wikipedia.org/wiki/Centigrade redirects to Celsius Dec 07 13:01:43 cjb_ie: thnx ;) Dec 07 13:02:31 cjb_ie: but do you agree that 15 can't be? Dec 07 13:03:06 15 sounds pretty wrong alright Dec 07 13:03:22 cjb_ie: agree. also 17127 is a little hight ;) Dec 07 13:03:35 what's that meant to be, milliamps? Dec 07 13:04:07 cjb_ie: as i understand the wiki, that makes 17.127 A? Dec 07 13:04:16 bahaha Dec 07 13:04:40 hehe :) Dec 07 13:04:42 LaF0rge: any idea? Dec 07 13:04:50 500mA max from usb Dec 07 13:05:03 as far as I know Dec 07 13:05:23 yup, that's right Dec 07 13:05:28 and 100mA without negotiation Dec 07 13:05:29 ahven, cjb_ie what does your chgcur say? Dec 07 13:05:46 Ahven: spec is supposed to be that but I've found a few cases where it differs Dec 07 13:06:00 anyway 17A can't ever be Dec 07 13:06:03 device at home, currently at work Dec 07 13:06:59 depends ... if its 17AH that would mean its roughly 1/4 of a normal car accu Dec 07 13:07:15 seems somewhat overkill for a cellphone Dec 07 13:07:18 leila: we talk about 17A Dec 07 13:07:29 hi Dec 07 13:07:31 ...charging current Dec 07 13:07:35 LaF0rge: all spi devices are gone under kernel 2.6.24-rc4 Dec 07 13:07:42 emdete, i see... like no way :) Dec 07 13:07:55 maybe it uses a scale and it needs be calcualted ? Dec 07 13:08:07 but neo's kernel tells me so with a temp of 15 degree celsius... Dec 07 13:08:38 leila: the wiki states milliamperes Dec 07 13:09:11 ok, my vote would be that its off :) Dec 07 13:09:13 171mA? Dec 07 13:09:23 17127mA == 17A Dec 07 13:10:49 LaF0rge: any idea? Dec 07 13:11:56 does anybody have openmoko working on a treo 650? Dec 07 13:13:02 dam i forgot my notes . anyone remember me witch modules sholud be compiled into kernel to use debugboard v3 ? Dec 07 13:18:02 emdete: i think the 15 is hardcoded or something Dec 07 13:18:07 Does anybody know when the orange neo is obtainable? Dec 07 13:18:25 iirc there was some bug in the temperature calcultion and they decided it was unneeded anyway Dec 07 13:18:48 i want the orange too :/ Dec 07 13:20:37 * balrog-kun will paint a neo orange for as little as EUR 50 Dec 07 13:21:40 :P Dec 07 13:28:56 uni bremen ;D Dec 07 14:06:31 openmoko: 03thomas * r3611 10/trunk/src/target/OM-2007.2/libraries/moko-gtk-engine/ (4 files in 3 dirs): Dec 07 14:06:31 openmoko: * src/moko-draw.c: (moko_gradient): Dec 07 14:06:31 openmoko: * src/moko-utils.c: (moko_shade_colour): Dec 07 14:06:31 openmoko: Change how shading operates to allow shading up from black. Dec 07 14:06:31 openmoko: * data/gtk-2.0/gtkrc: Dec 07 14:06:32 openmoko: Change the default fg colour to black (used as bg colour in some Dec 07 14:06:36 openmoko: widgets, e.g. toolbars) Dec 07 14:09:42 Hi all. I'm going to need a reliable GSM phone that lets me make and receive calls by the 15th of January. I don't need any other features. I would really like to spend my money on a neo1973, but in the event that it still doesn't meet my needs by then, what do you all recommend? Ideally, I'd like to support open hardware with my dollars. Dec 07 14:10:09 erpo: just make/receive calls? Dec 07 14:10:30 hrw: Yeah. That's all I need. I don't need scheduler apps or SMS or anything like that. Dec 07 14:10:41 hrw: I want those things, but they aren't hard requirements. Dec 07 14:10:42 I would grab something lowend like nokia 1600/2300/3200 Dec 07 14:11:28 yeah, it's still a developer board :) Dec 07 14:11:48 gta02 will be devboard too Dec 07 14:12:15 nokia 3310 ftw :D Dec 07 14:12:35 what about qtopia? Dec 07 14:12:48 on the neo1973 Dec 07 14:13:12 I am not fan of Qtopia so will not tell Dec 07 14:15:22 In summary, Qtopia on the Neo1973 is much closer to being ready than OpenMoko as of this writing. However, sound and suspend support prevent it from being at all useful as a cell phone (as a landline that doesn't ring but instead vibrates, it works ok). Dec 07 14:29:50 erpo: you can get a reliable gsm phone for making calls for like $20 (then have more money left over for when a smartphone comes along that you like) Dec 07 14:31:38 OJW, hrw : I was hoping not to have to spend money on a closed phone, but your suggestions seem to make the most sense. Dec 07 14:34:31 hmmm. i could put /usr on the card Dec 07 14:34:35 Hmm. Dec 07 14:36:52 erpo: or don't spend money at all, just get the old phone from someone who just bought a new one Dec 07 14:37:25 erpo i have the nokia 6100 Dec 07 14:37:29 thats very nice Dec 07 14:38:17 I don't know anyone with an old GSM phone. I'll be moving to an area I've only visited before, and the only reliable cell service there is GSM. They have CDMA too, but it sucks. Dec 07 14:38:19 Why do all phones suck nowdays? Dec 07 14:38:24 small andindestructible Dec 07 14:38:26 ;D Dec 07 14:38:58 nokia 1110 go for 15EUR here Dec 07 14:39:00 opo|away: my 6+ yrs old 3310 seems to be indestructible too Dec 07 14:39:08 of course Dec 07 14:39:10 :) Dec 07 14:39:16 old nokia rules Dec 07 14:39:25 5110 Dec 07 14:39:26 ;D Dec 07 14:39:42 zash: I threw my 3330 across the room in a rage once. I found all the little plastic bits, put them back together and the bloody thing still worked Dec 07 14:39:44 zash: Probably because cost and coverage are how consumers pick cell service, so they get hosed on everything else. Dec 07 14:39:56 i open my beer with my handy Dec 07 14:39:59 V500 broke though with only minor abuse Dec 07 14:40:11 erpo: or exterior design Dec 07 14:41:00 zash: I don't think most people are quite that shallow. People can be remarkably sensible when it comes to saving money. Dec 07 14:41:00 Will I get kicked if I say: android? Dec 07 14:41:11 * XorA kicks drx Dec 07 14:41:13 Yes. Dec 07 14:41:28 just kidding Dec 07 14:41:30 What's wrong with android? Everything I've heard about it sounds pretty good. Dec 07 14:41:31 what about NokiaN800 Dec 07 14:41:32 ;) Dec 07 14:41:43 yep it's da bomb Dec 07 14:41:51 N800 is a fabulous device. But it's not a phone, is it? Dec 07 14:41:54 erpo: one word: girls Dec 07 14:41:55 finally a unified platform for application developers Dec 07 14:42:03 based on java Dec 07 14:42:07 kinda Dec 07 14:42:10 it's not open, but most people won't care Dec 07 14:42:12 a java based platform Dec 07 14:42:16 It is open Dec 07 14:42:20 nope Dec 07 14:42:21 Wait, I thought it was open. Dec 07 14:42:28 You can get the source right now Dec 07 14:42:30 It's java. Cobol for the new century. Dec 07 14:42:31 nope Dec 07 14:42:39 ?? Dec 07 14:42:47 show me Dec 07 14:43:09 dtx: you can get the source for things that they have to make open Dec 07 14:43:22 but not for the jvm itself i.e. Dec 07 14:43:38 there is no jvm Dec 07 14:43:48 dvm Dec 07 14:43:55 yes, whatever ;) Dec 07 14:44:08 i.e. the libs? Dec 07 14:44:32 I haven't done java stuff in ages so bear with me Dec 07 14:44:38 ScaredyCat: re Dec 07 14:44:55 So the virtual machine isn't open AND the libs aren't open? Dec 07 14:44:57 So what is open? Dec 07 14:45:06 I thought it was all going to be BSD licensed Dec 07 14:45:22 dtx: yes, thats what they say Dec 07 14:45:26 "going to be" Dec 07 14:45:26 why do we even care about android aside from branding? Dec 07 14:45:35 dunno someone mentioned it Dec 07 14:45:38 :D Dec 07 14:45:38 portability Dec 07 14:45:45 portability my ass. Dec 07 14:45:53 write good code and it'll be portable Dec 07 14:45:54 pseudo portability Dec 07 14:46:02 indeed Dec 07 14:46:07 common platform Dec 07 14:46:23 honestly I wish they would've just gone with this project and thrown some money behind it Dec 07 14:46:41 I would much rather have a platform with native apps than have to run a JVM/JIT Dec 07 14:46:46 Wasn't that tried with j2me - it only sort of worked Dec 07 14:47:00 ScaredyCat: could you queue up a new qtopia build w. cruxus ? Dec 07 14:49:27 Dtx: I only agree with that cause native apps should run faster then on a vm and til mobiles have multicores or more speed its the only benifit Dec 07 14:50:08 me too Dec 07 14:50:50 I do have to admit that Java has a hell of a set of OSS libs though Dec 07 14:50:56 in Apache Jakarta or whatever Dec 07 14:51:00 Any chance someones looking @ those new dualcore arm chips I heard about Dec 07 14:52:15 Or better yet a company that's licesnced the technology Dec 07 14:54:09 * mwester wonders why oh why is the solution to throw more power-hungry technology at a problem, when that problem is created by the use of bloated inefficient languages in the first place. Dec 07 14:54:41 Actually I was thinking from a multitasking perspective Dec 07 14:54:54 So to put Android on a device do you have to get a license from Google? Dec 07 14:54:56 The single cpu is more than adequate Dec 07 14:55:17 I would assume so since its not all open source and there are so many chipsets it would have to run on Dec 07 14:55:20 Well Dec 07 14:55:24 yeah I guess Dec 07 14:55:39 I'm listening to music, reading a website, on multichat client, and irc @ this moment on my phone Dec 07 14:55:41 In fact, the only way to have reasonable battery life is if the CPU runs in short bursts and sleeps in between -- the less work it does, the longer the battery lasts. Dec 07 14:56:01 StarKnight: Yep, and still the CPU is probably mostly idel. Dec 07 14:56:05 er, idle Dec 07 14:56:31 Probably true memory latency then Dec 07 14:56:41 We need faster memory Dec 07 14:59:07 Though doesn't a multicore system allow you to switch between programs with less noticeable overhead Dec 07 15:01:08 * Stephmw sighs at the 'discussion' Dec 07 15:01:36 Correct me if I'm wrong - not a computer major Dec 07 15:02:14 I understand a little about the penalty when switching threads but that's about it Dec 07 15:03:19 StarKnight, do you mean human-noticeable overhead? At the machine level it makes a small difference, but not in a useful way. IIUC. Dec 07 15:03:27 StarKnight: it's not you specifically... it's overbroad statements regarding any one solution... I've had many discussions in the past based around correcting assumptions with my commercial experience - but those sap my will to have many of them Dec 07 15:04:16 openmoko: 03thomas * r3612 10/trunk/src/target/OM-2007.2/libraries/moko-gtk-engine/ (ChangeLog src/moko-draw.c): Dec 07 15:04:16 openmoko: * src/moko-draw.c: (moko_draw_vline), (moko_draw_hline), Dec 07 15:04:16 openmoko: (moko_draw_style_class_init): Dec 07 15:04:16 openmoko: Add hline and vline functions, but don't actually draw anything Dec 07 15:04:42 So, is anyone here using a Neo as their day-to-day phone yet? Dec 07 15:04:44 And yes I was refering to human noticeable overhead Dec 07 15:05:09 StarKnight, in that case it will have no noticeable effect on program switching to be multicore or not. Assuming the software isn't totally appalling, of course :) Dec 07 15:07:05 So on my current phone or even my computer when I switch between my apps the wait cursur is because? Memory? Dec 07 15:07:16 Because the software is poorly written. Dec 07 15:07:55 Don't get me wrong -- on current mainstream systems it's often impossible to write high-quality UI software. Dec 07 15:07:59 (practically speaking) Dec 07 15:08:29 It will take an entirely new kind of system to change things. Mobile computers (cellphones) might be it. Dec 07 15:08:58 So on a technical lvl what causes that? Dec 07 15:09:14 rushed software Dec 07 15:09:27 lazy coders... bad coders... the problem is human Dec 07 15:09:59 I was thinking like code that doesn't execute properly and causes faults or cashe misses Dec 07 15:10:10 If you're interested in the topic, you may find http://www.vpri.org/html/work/ifnct.htm and in particular http://www.vpri.org/pdf/NSF_prop_RN-2006-002.pdf interesting Dec 07 15:10:35 StarKnight: right - but those can be addressed if the coder has enough time/will to do so Dec 07 15:10:41 The problem is definitely human, but not something an individual can solve within the system, IMO Dec 07 15:10:59 So cellphones offer an opportunity for revolution :-) Dec 07 15:11:02 StarKnight: development times in the embedded software world are incredibly rushed and *sales driven* Dec 07 15:11:08 (as does Ian Piumarta's COLA project) Dec 07 15:11:12 StarKnight: so as tonyg says, it's the system at fault Dec 07 15:11:15 So like say with the api then? Dec 07 15:11:28 Specifically, the *economy* we operate within. Dec 07 15:11:34 tonyg: fair point Dec 07 15:11:52 Given time, people would be able to "collapse the software stack", and work in a small, comprehensible system Dec 07 15:12:04 Since noone's given time, the stack keeps growing taller and dirtier... Dec 07 15:12:28 * Stephmw washes his hands in the comments Dec 07 15:12:38 So its starts with how the OS handles the hw and moves up from there Dec 07 15:12:51 StarKnight: the hardware suffers from the same rushing Dec 07 15:12:53 StarKnight, exactly Dec 07 15:13:29 I'm starting to think Chuck Moore wasn't completely wrong when he said that the software problem was solved... by FORTH ;-) Dec 07 15:13:38 haha Dec 07 15:14:15 I've been building an openmoko userland over the past week or so in Erlang, and it's amazing what a good choice of language working very close to the hardware (or in my case, the kernel) can accomplish Dec 07 15:14:28 StarKnight: which phone is it you're using Dec 07 15:14:35 So wouldn't the quickest and easiest way to handle the problem be to remove the OS and have the program directly interface the hw Dec 07 15:14:39 StarKnight, yes. Dec 07 15:14:40 tonyg: any plans on attending FOSDEM? Dec 07 15:14:43 Blackberry 8100 Dec 07 15:14:44 * tonyg googles Dec 07 15:14:59 StarKnight: Back to assemblers then? Dec 07 15:15:16 * zedstar needs to take a look at Erlang one day Dec 07 15:15:18 Stephmw, hadn't considered it... maybe? Dec 07 15:15:22 How about DOS? Dec 07 15:15:24 Well it works on my picmicros Dec 07 15:15:27 erpo: EEEK Dec 07 15:15:55 tonyg: I'm planning an informal, mini-presentation on the realities of the embedded/mobile development world ;) Dec 07 15:16:01 excellent Dec 07 15:16:05 I understand why there is an OS and api though so everyone doesn't need to write the same thing and simplify coding Dec 07 15:16:09 tony btw do u know about mobile monday in london? the other week somebody gave an erlang talk Dec 07 15:16:20 Stephmw, that'd be a good thing for the first london openmoko UG too Dec 07 15:16:23 zedstar: I've been meaning to attend those ;) Dec 07 15:16:30 tonyg: when's that planned for? Dec 07 15:16:34 no plan yet Dec 07 15:16:40 sometime after the GTA02 release? Dec 07 15:16:43 haha Dec 07 15:16:47 so, the new year, practically Dec 07 15:16:48 stephmw i just signed up but last one was on voice which didnt interest me but will go soon if good topic Dec 07 15:16:48 uh Dec 07 15:16:53 unless there's something I don't know? Dec 07 15:16:57 is there a date set for a non-dev edition? Dec 07 15:17:08 zedstar, I was at that momo Dec 07 15:17:10 tonyg: no, I'm just cautious about release dates Dec 07 15:17:19 Stephmw, commendable conservatism ;) Dec 07 15:17:36 tonyg: problem with momo is that it clashes with my climbing nights Dec 07 15:17:52 tonyg how was it are they worth going? Dec 07 15:17:56 It wasn't a very technical crowd Dec 07 15:18:03 it is just down the road from me Dec 07 15:18:09 I'm not sure. Unless there's a topic of particular interest... dunno. Dec 07 15:18:27 If I hadn't been coming down with a cold I'd probably have got more out of the mingling after the talks Dec 07 15:18:28 * Stephmw throws a stone at the MI6 building Dec 07 15:18:44 * XorA watches Stephmw get bundled into a van Dec 07 15:19:02 XorA: I'm pretty sure they have my drive imaged by now :) Dec 07 15:19:03 yeh chatting to interesting people appeals to me Dec 07 15:19:16 zedstar: you could also sign up for the Google Dev Jams Dec 07 15:19:21 Stephwm or Tonyg whoever posted the links earlier could you email them to starknight83@Gmail.com I would greatly appreciate it (they already dropped from my clients history) Dec 07 15:19:32 bassinboywk: No. Dec 07 15:19:38 stephmw yeh i need to do this i missed last one it was full quickly Dec 07 15:19:51 zedstar: the mailing list gets first dibs Dec 07 15:19:56 bassinboywk: Originally I thought it was last October, but all references to that seem to have been erased. Dec 07 15:20:02 a thanks for the tip Dec 07 15:20:04 zedstar: I didn't end up going - was hurting too much ;) Dec 07 15:20:20 StarKnight, done Dec 07 15:20:24 not a huge google fan tho ;) Dec 07 15:20:26 Ty Dec 07 15:20:38 zedstar: likewise - but hey, it's free pizza Dec 07 15:21:08 yeh....i did manage to get a free drink out of some google recruitment drone after that future of mobile event Dec 07 15:22:46 * zedstar counting the hours left til office christmas party Dec 07 15:23:25 erpo: strange, huh? Dec 07 15:23:58 zedstar: indian lady? Dec 07 15:24:05 stephw yeh she was Dec 07 15:24:13 zedstar: I know her - very smart Dec 07 15:24:31 zedstar: didn't manage to hook up with her at that conf Dec 07 15:24:38 bassinboywk: Yeah, it's disappointing. But at some point you have to choose an interim solution and just check in every couple of months to see if it's finished. Dec 07 15:25:05 i didnt realise she was from google when i was bitching about android.....should have read her t-shirt with google written all over it Dec 07 15:25:24 bwahaha Dec 07 15:26:01 I met up with another friend and we ended up having lunch on the southbank instead of IMAX Dec 07 15:26:32 the free bar finished very quickly Dec 07 15:27:22 zedstar, tonyg, what do you work on? mobile? Dec 07 15:28:15 i work on mobile on my personal projects but work as a programming drone for an isp for cash Dec 07 15:29:38 and yourself stephmw? Dec 07 15:30:07 I'm an embedded software engineer... so I piss lines for phones ;) manufacturers, operators, etc Dec 07 15:30:21 C/J2ME/J2SE Dec 07 15:30:52 u working in london also? Dec 07 15:31:32 yup Dec 07 15:31:43 depending on contract, of course Dec 07 15:31:58 currently cooling my heels at consultancy HQ waiting for the next contract to start Dec 07 15:32:13 wherabouts in london? Dec 07 15:32:27 and I tells ya... reviewing CVs for technical content is... worse than pissing lines Dec 07 15:32:37 zedstar: Waterloo/Vauxhall Dec 07 15:32:58 you? Dec 07 15:33:05 ah ok......going there today for christmas party....im in holborn Dec 07 15:33:22 oh, not far at all then ;) Dec 07 15:33:36 I'm heading to Manor House later for my regular climbing session Dec 07 15:33:50 * Stephmw can't wait for the London Bridge place to open Dec 07 15:33:55 not far....going to some gaming place... http://www.namcostation.co.uk/ u know it? not been before Dec 07 15:35:51 haha, yeah :) Dec 07 15:35:59 haven't been there yet though Dec 07 15:36:06 hi guys, just a question: is anybody here using the Neo as his every-day-phone? Dec 07 15:36:14 free drink and dodgem cars apparently Dec 07 15:37:51 ??? Dec 07 15:38:43 christian_: german? Dec 07 15:38:49 Or in other different words: is there a chance that it will be suitable for every day use by the end of febraury Dec 07 15:38:53 yes Dec 07 15:38:56 naja Dec 07 15:39:04 ich benutz es nicht als alltagsgerät Dec 07 15:39:07 hab nen zweites Dec 07 15:39:09 ^^ Dec 07 15:39:19 kannst natürlich telefonieren und sowas Dec 07 15:39:34 geht das jetzt schon zuverlässig? Dec 07 15:39:43 warte noch nen bißchen wenn du nicht programmieren kannst Dec 07 15:40:07 ist nicht so das ich mir jetzt des gta01 kaufen will, will nur wissen wie die entwicklung so voranschreitet Dec 07 15:40:41 ja es werden immer neue projekte gestartet Dec 07 15:41:13 also nen halbes jahr denk ich mal dann isses für den normalverbraucher nutzbar Dec 07 15:41:22 ja das hab ich auch gesehen, aber ich hab fast das gefühl das viele von denen im sand verlaufen Dec 07 15:41:33 jo... Dec 07 15:41:53 kann man halt bei opensource nicht so pauschal sagen wanns fertig ist Dec 07 15:42:03 also jetzt würde ich es nicht im alltag nutzen Dec 07 15:42:10 wie siehts mit na gps navigation aus? Weiste ob sich da was tut? Dec 07 15:42:17 jo is in arbeit Dec 07 15:42:23 mit spritzefranz alles macht spass und fünf kartoffel Dec 07 15:42:34 ne alpha version steht glaub ich shcon Dec 07 15:42:34 Stephmw, I work for lshift.net; openmoko is a personal project, really Dec 07 15:42:35 ROFL Dec 07 15:43:27 kanst ich spritze aus ehren bösen mit meine grosse schvanze? (can i have a burger to go please, no, really ;) ) Dec 07 15:43:29 christian_, I'm very close to having it in a state where it could be my daily phone Dec 07 15:43:36 tonyg: that's no shame - I didn't start out in mobiles Dec 07 15:43:38 I just need to improve the battery lifetime Dec 07 15:44:04 und weil des display ja doch kleiner ist als beim iphone: kann man da noch gescheit tippen? Dec 07 15:44:13 tonyg: last good place I worked at did software for network root-cause-analysis Dec 07 15:44:18 christian_: wahre arbeit, wahre sauerkraut Dec 07 15:44:26 tonyg: So you can make and receive calls without using the command line? Dec 07 15:44:54 * Stephmw heads out - laters guys Dec 07 15:45:41 erpo, yeah, but I'm using my own userland Dec 07 15:45:58 tonyg: Ah. :( Dec 07 15:46:03 the openmoko build, plus some packages on top of the standard kernel Dec 07 15:46:13 tonyg: How's the battery life as it stands? Dec 07 15:46:17 essentially my own replacements for neod, gsmd and dialer Dec 07 15:46:26 erpo, with the stock userland, I don't know Dec 07 15:46:32 with mine, it's about 5-6 hours Dec 07 15:46:42 tonyg: Standby time or talk time? Dec 07 15:46:44 there are *big* wins to be made in power management Dec 07 15:46:47 standby :( Dec 07 15:47:02 tonyg: If you can double that, I'll buy an openmoko to replace the phone I have now. Dec 07 15:47:16 yeah, doubling it is my target for my own personal use Dec 07 15:47:19 (really, 8h would do) Dec 07 15:47:24 tonyg: I get 12 hours standby time and that's if I don't talk at all. Dec 07 15:47:56 I've written a tiny amount about my userland on my wiki page, btw http://wiki.openmoko.org/wiki/User:TonyGarnockJones Dec 07 15:48:17 tonyg: what do you mean by "very close"? could you tell what is already working and what does not? Dec 07 15:48:51 tonyg: Do you think your code will get folded back into openmoko? Or is this just for fun? Dec 07 15:48:56 christian_, power management is the big issue. I have some (horrible) UI for SMS send/recv, dialer, addressbook cooked up Dec 07 15:49:06 erpo, just for fun. Dec 07 15:49:12 well, for my own personal use perhaps more. Dec 07 15:49:48 * zash wants IM-style sms-app! Dec 07 15:49:54 tonyg: What's the underlying widget set? GTK+? Dec 07 15:50:09 yes. It's erlang, using erlang's "gtknode" to do the gui Dec 07 15:50:38 ignore the gui, it's just something I threw together atop the REPL-based application core Dec 07 15:50:45 is erlang purely functional? Dec 07 15:50:53 zedstar, no Dec 07 15:50:57 well, kind of :) Dec 07 15:51:01 mostly? Dec 07 15:51:12 ah ok good Dec 07 15:51:20 it's applicative-order, with IO effects and very restricted mutable state Dec 07 15:51:25 and communication primitives, of course Dec 07 15:51:56 thing i like about erlang is its ASN.1 encoding stuff Dec 07 15:52:02 zedstar, yes that's cool Dec 07 15:52:09 not needed for this project tho :) Dec 07 15:52:19 i wrote my own packed encoding rules stuff was pain Dec 07 15:52:40 PER is a pain indeed. I'm puzzled as to why BER has fallen out of fashion, though Dec 07 15:52:50 Perhaps it was just that ASN.1 was such an arcane *concrete* language Dec 07 15:52:55 too verbose ;) Dec 07 15:53:03 BER? or ASN.1? Dec 07 15:53:07 BER Dec 07 15:53:20 interesting. To me, its verbosity is a win... Dec 07 15:53:26 TLV = pleasant Dec 07 15:53:53 * tonyg gets food Dec 07 15:53:54 bbl Dec 07 15:54:36 Yet another question guys: will there be an integrated IM capable of handling multiple protocols (icq, msn, jabber, maybe Skype (I know its not open source)) Dec 07 15:55:16 christian_: Skype is a definite no, others could be implemented using libpurple or jabber, of course Dec 07 15:56:04 Telepathy Dec 07 15:56:13 will do all of those, except Skype :) Dec 07 15:56:28 nice UI-independent API, designed for embedded devices Dec 07 15:56:39 abraxa_: is the "no" due to the politics of openmoko or because of some dependencies? Dec 07 15:56:49 <_diego_> skype is undocumented, and worse, the use trick just to avoid to be disclosed Dec 07 15:57:00 * Robot101 is waiting for a reply from Sean/Mickey/Wolfganf about it Dec 07 15:57:07 <_diego_> s/the/they Dec 07 15:57:18 s/nf/ng/ Dec 07 15:57:22 christian_: unless Skype themselves do a port onto the platform, there's no way to do it Dec 07 15:57:33 even the N800 Skype client checks its running on N800 hardware Dec 07 15:57:33 christian_: As _diego_ said, it's proprietary, majorly protected against reverse engineering and there's no binary blob for ARM Dec 07 15:58:20 ok I see. what a shame... Dec 07 15:59:03 would it even work to use Voip via GSM? Dec 07 15:59:36 Verizon sued over GPL violations on BusyBox Dec 07 15:59:58 <_diego_> christian_: it should be possible Dec 07 16:00:38 _diego_: thanks Dec 07 16:00:40 <_diego_> but with low quality, probably Dec 07 16:01:29 <_diego_> i expect 0.5 second (at least) of lag Dec 07 16:01:58 _diego_: what would you say. Its unlikely that GTA02 will be released until febrauary, isn't it? Dec 07 16:02:26 <_diego_> i think so Dec 07 16:03:03 <_diego_> (but i'm not more informed of anyother person reading planet.openmoko.org ^^) Dec 07 16:14:21 * Writchie OM really screwed the pooch on GSM 850. It is now missing a wave of "Open Frenzy" spin from US operators. Dec 07 16:15:41 Hello Dec 07 16:15:47 hey Dec 07 16:15:59 I was asking yesterday if anyone knows about a pictograph spec for openmoko? Dec 07 16:20:07 * hhf423__ yawns Dec 07 16:21:47 or at least a link to the wishlist so I can peruse and see if it's been suggested already :) Dec 07 16:24:15 Writchie, any specific reasons why they choose the TI calypso chip? Dec 07 16:24:30 <_diego_> DaSkreech, look at wiki.openmoko.org Dec 07 16:26:27 iirc it's not the chip that's at fault... Dec 07 16:26:35 it's the wiring of it in the neo.. Dec 07 16:27:17 hello. is the gta02 already running on qemu? Dec 07 16:27:23 ScaredyCat, also, the calypso project came to a point where TI, Bangalore was about to scrap it Dec 07 16:27:29 with the moko makefile Dec 07 16:28:29 mbuf: the calypso has a rather complete and mostly standard compliant mux. Dec 07 16:28:36 Writchie, ok Dec 07 16:28:58 so it's probably the pinnacle of an open radio ATM Dec 07 16:29:57 i.e. for GSM/GPRS non-edge Dec 07 16:29:59 _diego_: Wow that's a lotta wishes :) Dec 07 16:30:40 openmoko: 03andrew * r3613 10/trunk/src/host/qemu-neo1973/ (7 files in 2 dirs): Quick sync with upstream changes. Dec 07 16:30:48 openmoko: 03andrew * r3614 10/trunk/src/host/qemu-neo1973/openmoko/ (README linux-gta02-pseudo.patch): Provide a kernel for the pseudo GTA02 board and a README. Dec 07 16:32:19 thomasgruebler: balrog-kun could answer that but from what I know some of the new hardware is not being emulated yet Dec 07 16:32:40 but can i boot it? Dec 07 16:32:50 because some weeks ago it wasn't possible Dec 07 16:32:58 right, gta02 is not emulated Dec 07 16:33:00 and some weeks befor some weeks it was also not possible Dec 07 16:33:03 thomasgruebler: boot what? Dec 07 16:33:19 boot the openmoko in qemu Dec 07 16:33:28 to have the software/gui of gta02 Dec 07 16:33:52 the software is the same as for gta01 Dec 07 16:34:03 the software/gui is the same on all devices running OpenMoko Dec 07 16:34:07 only the kernel differs Dec 07 16:34:53 the gta02 kernel probably won't run in current qemu as gta02 parts are not yet emulated, mostly Dec 07 16:34:54 ok Dec 07 16:35:16 so when i compile now the environment the whole night, i can also use it for more than compiling Dec 07 16:35:30 because i don't have a neo yet Dec 07 16:35:46 yes, but we all know compiling is the best part of it Dec 07 16:36:01 ;) Dec 07 16:36:49 ^^ Dec 07 16:36:56 * mwester *loves* compiling. It's cold, very cold, here and anything that makes the CPU work harder for very long periods of time is appreciated. Dec 07 16:37:45 * mwester can warm his hands on the top of his build system... Dec 07 16:38:43 mwester: you use gentoo? Dec 07 16:39:17 Fedora. I just rebuild openmoko and SlugOS regularly. Dec 07 16:39:49 right..(silently takes note of "regularly") Dec 07 16:39:52 i'm sure that's be enough Dec 07 16:39:55 s/be// Dec 07 16:39:55 balrog-kun meant: i'm sure that's enough Dec 07 16:40:56 hi, when i try to enable the usb slave with the latest version in QEMU I have getting the following error. gadget_read: event error: 51 Dec 07 16:40:56 s3c_udc_handle_packet: EP0 overrun Dec 07 16:40:56 s3c_udc_handle_packet: EP0 overrun Dec 07 16:40:56 gadget_read: event error: 4 Dec 07 16:41:15 * Sup3rkiddo raises his hands for the same error Dec 07 16:41:40 any reasons/suggestions? Dec 07 16:42:42 no idea, i thought it was my weirdly setup that was causing the problem.. Dec 07 16:43:42 and it's not working i understand? Dec 07 16:43:56 did that change with some recent update? Dec 07 16:43:57 balrog-kun: nope, usb0 isn't getting created Dec 07 16:44:25 balrog-kun: been like that for a long time, maybe about 4-5 weeks Dec 07 16:44:59 balrog-kun: hope the following dmesg output will help [ 1698.549230] dummy_udc dummy_udc: dequeued req efc2cc80 from ep-c, len 4096 buf ef607000 Dec 07 16:45:00 [ 1698.553223] dummy_udc dummy_udc: dequeued req efc2ca80 from ep-a, len 0 buf ef490be0 Dec 07 16:45:00 [ 1698.557217] dummy_udc dummy_udc: dequeued req efc2cc80 from ep-c, len 4096 buf ef607000 Dec 07 16:45:00 [ 1698.561210] dummy_udc dummy_udc: dequeued req efc2ca80 from ep-a, len 0 buf ef490be0 Dec 07 16:45:00 [ 1698.565205] dummy_udc dummy_udc: dequeued req efc2cc80 from ep-c, len 4096 buf ef607000 Dec 07 16:45:01 [ 1698.569198] dummy_udc dummy_udc: dequeued req efc2ca80 from ep-a, len 0 buf ef490be0 Dec 07 16:45:03 [ 1698.573192] dummy_udc dummy_udc: dequeued req efc2cc80 from ep-c, len 4096 buf ef607000 Dec 07 16:45:05 [ 1698.577186] dummy_udc dummy_udc: dequeued req efc2ca80 from ep-a, len 0 buf ef490be0 Dec 07 16:45:09 [ 1698.581186] dummy_udc dummy_udc: dequeued req efc2cc80 from ep-c, len 4096 buf ef607000 Dec 07 16:45:11 [ 1698.585174] dummy_udc dummy_udc: dequeued req efc2ca80 from ep-a, len 0 buf ef490be0 Dec 07 16:46:54 balrog-kun: just a hunch, but the gadgetfs.h header in the kernel had been moved some place from 2.6.22 or so...i had to bring the header from an older kernel to get it built... Dec 07 16:47:12 techindian: not really :p , but i'm seeing the same issue on mine Dec 07 16:47:23 :) Dec 07 16:47:24 i'll try to do something with it later Dec 07 16:47:47 Dec 7 18:48:38 [kernel] usb 3-1: can't set config #1, error -110 Dec 07 16:47:53 balrog-kun: oops, sorry udb_gadgetfs.h Dec 07 16:48:46 Sup3rkiddo: but i hope "make" didn't error out when the header was missing? Dec 07 16:49:08 ./configure should detect that and disable gadgetfs Dec 07 16:49:13 balrog-kun: it did Dec 07 16:49:32 were you rebuilding without re./configuring? Dec 07 16:49:34 btw, are there any everyday-users of Neo-s GSM near? Dec 07 16:49:56 balrog-kun: does, make clobber-qemu count.. Dec 07 16:50:09 Sup3rkiddo: oh, i don't know :p Dec 07 16:50:33 balrog-kun: just to be sure, lemme remove that header, clean, and rebuild and pastebin the errors Dec 07 16:50:35 lol Dec 07 16:51:03 but it was "undefined/undeclared/unitialized" sort of errors Dec 07 16:51:35 Sup3rkiddo: ok, but i don't know if clean does reconfigure Dec 07 16:52:01 balrog-kun: hmm..alright, i will reconfigure it then Dec 07 16:54:39 2 people, whom I have called, have complained of hearing themself from me, echo Dec 07 16:55:27 ahven: this usually indicates your audio settings are messed up. Dec 07 16:55:42 alsa state files for handset Dec 07 16:55:47 or headset Dec 07 16:55:54 depending on which you are using Dec 07 16:55:57 I haven't changed anything in that matter Dec 07 16:56:01 handset Dec 07 16:56:11 last public build Dec 07 16:57:25 balrog-kun: ah, sorry i forgot...i had enabled gadgetfs support which was otherwise disabled...by modifying certain macros, i think otherwise it would have built correctly but without gadgetfs support Dec 07 16:59:41 Sup3rkiddo: ok, cool Dec 07 16:59:57 i hope they stop constantly changing the usb header path soon Dec 07 17:01:19 balrog-kun: yeah me too, i am still having a bad time searching around where they moved it... Dec 07 17:03:01 ScaredyCat, you said earlier that it was the wiring of the calypso in the neo that was a problem. Can you elaborate? Were you referring to the serial-console/GSM-modem switchover issues, or something to do with 07.10, or something else again...? Dec 07 17:03:20 Wake-on-GSM-event, for instance? Dec 07 17:04:42 Sup3rkiddo: now you are able to use the usb connection?? Dec 07 17:04:45 * balrog-kun will bbl, gotta go pay a fine -_- Dec 07 17:05:26 techindian: nope, got the same set of warnings/errors...btw you are from India? Dec 07 17:05:36 hmm. Dec 07 17:06:42 hmm? is that a "yes" or a "no" :P Dec 07 17:06:56 :) yes.. Dec 07 17:07:35 techindian: me too :) thats why asked Dec 07 17:08:10 grt.. Dec 07 17:08:24 * arw1 hello people Dec 07 17:09:17 hello Dec 07 17:09:52 Sup3rkiddo: update me if you find the solution. I will also try.. Dec 07 17:10:24 tonyg: it was somehting to do with the chip being 1uad band but the current layout on the pcb only allowed the use of 3 .. Dec 07 17:10:36 hence the firmware would alow 850 + 2 other Dec 07 17:10:36 techindian: sure...i think its the header, but btw, which kernel are you running on Dec 07 17:11:06 ubuntu 2.6.20.16 Dec 07 17:11:18 ah, ok. So it was to do with the actual GSM-to-network featureset, rather than the interface-to-the-CPU side? Dec 07 17:11:36 as in, it's perfectly *controllable* :-) just doesn't support 850? Dec 07 17:12:16 techindian: ok no..atleast it built for you, newer kernels dont have the linux/usb_gadgetfs.h at the place where qemu searches for, fyi Dec 07 17:13:44 tonyg: the RF out has only three LNA's, two for hi-bands (1800/1900) and one for low bands (850/900). In order to support both 850 and 900 a different type of filter component is required. 850 or 900 can be supported by changing filters of the same general type (and form factor). Dec 07 17:14:00 s/RF out/RF in/ Dec 07 17:14:00 Writchie meant: tonyg: the RF in has only three LNA's, two for hi-bands (1800/1900) and one for low bands (850/900). In order to support both 850 and 900 a different type of filter component is required. 850 or 900 can be supported by changing filters of the same general... Dec 07 17:14:20 experiment Dec 07 17:14:24 s/periment/cellent/ Dec 07 17:14:24 tonyg meant: excellent Dec 07 17:15:01 tonyg: true quadband chipsets have 4 LNA's - not three. Dec 07 17:15:09 Writchie, right, that makes sense. Dec 07 17:15:14 * hhf423 yawns some more Dec 07 17:15:24 Being UK based, it's of less relevance to me personally :) Dec 07 17:16:18 for those marketing US is basically game-over until it is confirmed at least an 850/1800/1900 triband variant Dec 07 17:16:29 * tonyg nods Dec 07 17:16:32 That's a shame. Dec 07 17:18:15 oo, the gta02 will have a 400MHz CPU. What will that do to power consumption? Dec 07 17:19:06 WiFi +400MHz + Graphics Accel will make it a hot potato ;) Dec 07 17:19:17 but it has a larger battery Dec 07 17:20:41 And the CPU can be downclocked when the performance is not needed, I assume? Dec 07 17:21:18 AVee, in theory. The current codebase for gta01 doesn't do so yet. Dec 07 17:21:23 AVee: in theory, ARM's are pretty much linear mW/mips Dec 07 17:22:03 in theory tickless kernel should also help. Dec 07 17:22:14 It's just occurred to me to ask: by writing 0 into /sys/class/backlight/*/brightness, the display appears to switch off. Is it *actually* off? Is there an even more off state I can get to? Dec 07 17:22:15 but you have to get basic PM working well first Dec 07 17:26:02 <_diego_> looks like there's no cpu freq scaling atm on neo Dec 07 17:27:21 _diego_: No, as it's difficult to implement and takes a significant amount of time to alter Dec 07 17:27:28 alter the speed, I mean Dec 07 17:27:57 abraxa__, I had a look at the PDF spec for the CPU that is linked to from the wiki. I couldn't find the part where it talked about the pause for changing the PLL. Dec 07 17:28:09 Or at least, it mentioned 150 microsecs, but that seems fast, not slow... Dec 07 17:28:15 tehre is no explicit spec Dec 07 17:28:16 fast enough, anyway Dec 07 17:28:28 but yes, from memory that's the order of magnitude Dec 07 17:29:03 It's the kind of thing that could be done by some daemon at the time the display is dimmed. On the order of tens of seconds, or minutes apart. Dec 07 17:29:24 tonyg: have you seen the power managment nice things page I wrote on the wiki? Dec 07 17:29:34 Yes, that was a good page. The one with the tables? Dec 07 17:29:48 yeah Dec 07 17:30:07 btw, SpeedEvil, you might be able to answer my question: "by writing 0 into /sys/class/backlight/*/brightness, the display appears to switch off. Is it *actually* off? Is there an even more off state I can get to?" Dec 07 17:30:13 Hasn't the kernel recently gon no;2~_hzx_? Dec 07 17:30:23 gone to notick? Dec 07 17:30:24 <_diego_> abraxa__, that's a pitty. At least a 100Mhz mode it would be nice Dec 07 17:30:42 ciao Dec 07 17:30:44 As I understand it, that's wrong (abraxa) Dec 07 17:30:51 * tonyg . o O ( was that line noise?? ) Dec 07 17:30:56 you can change the divisor on the fly. Dec 07 17:31:03 tonyg: keyboard noise. Dec 07 17:31:11 :) Dec 07 17:31:25 tonyg: the backlight is off. Dec 07 17:31:29 however. Dec 07 17:31:35 damn. I was hoping for a big saving :) Dec 07 17:31:47 I think the display controller is still active, as is the LCM Dec 07 17:31:51 aha Dec 07 17:31:51 ok Dec 07 17:31:53 SpeedEvil: Yes, however LaF0rge said a couple days ago that other subsystems are affected in non-trivial ways (e.g. LCM) Dec 07 17:31:54 this is quite significant Dec 07 17:31:57 and the controller draw? Dec 07 17:32:02 * tonyg looks at the wikipage again Dec 07 17:32:23 abraxa__: as I understand it, that's not completely right - you can change the divisor of the CPU only. Dec 07 17:32:42 which does not get the power savings that are possible. Dec 07 17:32:49 tonyg: possibly quite significant. Dec 07 17:32:54 SpeedEvil: this glitch of GPIO's at power-off doesn't seem to make sense - the GPIO outputs are separately powered and maintained during the sleep mode. If the GPIO is sources from the parts that are powered down then yes they may glitch but not from hard GPIO states. If they all glitch this is a chipset problem not fixable without something like a CPLD between the wake-up sources and the mis-behaved SoC. Do you know any more about this Dec 07 17:32:59 <_diego_> abraxa__, on gp2x they do it in user space (changing cpu freq) Dec 07 17:33:18 SpeedEvil: Okay, that can't comment on as I haven't read that part of the specs Dec 07 17:33:20 tonyg: it's not only the controller and the LCD, it's the RAM bandwidth Dec 07 17:33:28 SpeedEvil, good observation Dec 07 17:33:32 * tonyg thinkgs Dec 07 17:33:42 Writchie: hang on, I'll look at the schematics. /me remembers he has no schematics Dec 07 17:33:44 hi folks, can someone help me about the openmoko projects developer svn access. DAV does not ask for password and SSH method says permission denied(public key). Dec 07 17:33:48 so ideally, I'd set the backlight to 0 = off, and then tell the LCD controller to stop refreshing for a while Dec 07 17:33:55 ;) Dec 07 17:34:19 tonyg: yes - I seem to remember power benefits from killing X Dec 07 17:34:23 I need to check again. Dec 07 17:34:30 excellent. right. so that's a good clue. Dec 07 17:34:44 is there an easy way to kill the display refresh? (too much to ask?) Dec 07 17:34:45 Also - going to 8 bit or mono mode will significantly reduce RAM power draw. Dec 07 17:34:53 if not, where should I start learning in the kernel tree, or the patchsets? Dec 07 17:34:54 SDRAM cycle times is very big power factor Dec 07 17:35:15 I'm seeing 5-6 hours with GSM on, cpu more-or-less idle, and backlight=0 Dec 07 17:35:26 if I can get it to 8... or 10!... it's usable as my cellphone Dec 07 17:35:53 tonyg: that's ballpark - you can get it to 8-12 by clocking the CPU to 100Mhz - see the wiki - and undervolting (arguably) the CPU Dec 07 17:35:58 tonyg: the GSM chip is not going into gsm idle. This is a 40mw hit Dec 07 17:36:18 the previous version of the spec-sheet said that 1.8V at 100Mhz was just fine. Dec 07 17:36:26 Writchie, with 5-6 hours I'm seeing something like 680mW, so 40mW is almost in the noise :) Dec 07 17:36:28 gsm idle takes you to 24 hrs Dec 07 17:36:30 this version doesn't mention that. Dec 07 17:36:35 doing ipkg upgrade I get conflicts between libgsmd and gsmd -- do I need to remove gsmd? Dec 07 17:36:47 (that 680 is a BotE number, not a measurement) Dec 07 17:37:11 hmm, interesting Dec 07 17:37:16 Sleep mode should be 10mw to 50mw Dec 07 17:37:19 I measured a minimum of about 90mA Dec 07 17:37:19 I was in a middle of ipkg update Dec 07 17:37:21 and reboot :) Dec 07 17:37:23 at 4V or so Dec 07 17:37:23 conflicts between openmoko-session2 and openmoko-today2-folders with matchbox-common Dec 07 17:37:26 also Dec 07 17:37:29 at 100Mhz Dec 07 17:37:31 SpeedEvil, how do I change the clock speed programmatically? Dec 07 17:37:41 will I need to hack on the kernel? Dec 07 17:37:50 tonyg: yues Dec 07 17:37:50 <_diego_> gabaug, it happened to me too.... Dec 07 17:37:54 or u-boot Dec 07 17:37:55 or is there a /sys interface?... oh, ok. damn Dec 07 17:37:55 :) Dec 07 17:38:04 it's an option to set it to 50/100 or so Dec 07 17:38:07 _diego_: did you find a solution? Dec 07 17:38:22 <_diego_> gabaug, i suggest to install a new image ^^' Dec 07 17:38:30 * tonyg guesses switching off the display refresh will be the same kind of hackery Dec 07 17:38:37 tonyg: pretty much Dec 07 17:38:40 ok. Dec 07 17:38:49 <_diego_> gabaug, after forcing the install, a get stuck in splash screen Dec 07 17:38:59 <_diego_> s/ a / I Dec 07 17:39:09 there are probably hacks that are doable with one-two-liners of kernel code Dec 07 17:39:28 to basically disable stuff without taking into account the existing drivers Dec 07 17:39:43 this would not actually resume for example, but would enable measurement Dec 07 17:39:50 * tonyg nods. Dec 07 17:40:03 (I need to buy a multimeter) Dec 07 17:40:13 hi folks, can someone help me about the openmoko projects developer svn access. DAV does not ask for password and SSH method says permission denied(public key). Dec 07 17:40:21 SpeedEvil: IIRC there are some better PM play hooks in later kernels Dec 07 17:40:24 (although I have to say, working just above the kernel/userland boundary suits me much better...) Dec 07 17:40:45 Writchie: I haven;'t done much kernel code. Dec 07 17:40:50 Writchie, not by any chance "later" meaning the november-build...? :) Dec 07 17:40:58 or svn trunk Dec 07 17:40:58 23 or 24 Dec 07 17:41:49 i'm getting ready to cross compile an out of tree kernel module now. Dec 07 17:42:40 SpeedEvil: Me neither until just recently Dec 07 17:43:42 SpeedEvil: from what I see interrupts are disabled for long periods of time all over the place. Dec 07 17:43:50 tick Dec 07 17:44:02 I'm amazed this even works as an embedded device. Dec 07 17:44:13 :) Dec 07 17:44:21 what's "lowlevel_foo"? Dec 07 17:44:27 * tonyg is downloading the nov snapshot Dec 07 17:44:40 u-boot startup stuff IIRC Dec 07 17:45:05 will I need it to flash a new u-boot, should I wish to do so? Dec 07 17:45:16 do you have a debug board? Dec 07 17:45:20 nope :) Dec 07 17:45:40 i wouldn't fool with u-boot without one Dec 07 17:45:57 yeah. I am still running the one that shipped with my phase1 device. Dec 07 17:46:21 tonyg: lowlevel_foo is to boot from RAM (rather than NAND flash) when using the debug board. Dec 07 17:46:24 tonyg: the shuipped uboot works Dec 07 17:46:29 (to change speed) Dec 07 17:46:37 loadable kernel modules may be safest way to play. Dec 07 17:46:57 hence your comment a page or so ago about an out-of-tree kernel module? Dec 07 17:47:02 very interesting! Dec 07 17:47:08 kind of a REPL for the kernel :-) Dec 07 17:47:09 yup Dec 07 17:47:24 a really clunky, dangerous repl Dec 07 17:47:56 (once upon a time a guy I knew put a scheme interpreter into the kernel. great for rapid driver development and low-level experimentation) Dec 07 17:48:07 (it was a tinyscheme fork, iirc) Dec 07 17:49:00 http://abstractnonsense.com/schemix/ Dec 07 17:51:32 why does Neo throw trash at me, when I connect with cu ? Dec 07 17:52:15 never-ending stream of garble Dec 07 17:52:20 only the first few lines are ok Dec 07 17:52:25 is it in boot menu? Dec 07 17:52:28 yep Dec 07 17:52:36 ./dfu-util.i386 --list Dec 07 17:52:36 dfu-util - (C) 2007 by OpenMoko Inc. Dec 07 17:52:36 This program is Free Software and has ABSOLUTELY NO WARRANTY Dec 07 17:52:36 Found Runtime: [0x1457:0x5119] devnum=10, cfg=0, intf=2, alt=0, name="USB Device Firmware Upgrade" Dec 07 17:52:36 Found Runtime: [0x0a5c:0x2101] devnum=2, cfg=0, intf=3, alt=0, name="UNDEFINED" Dec 07 17:52:40 maybe because of this? Dec 07 17:52:45 use 1457 one Dec 07 17:53:03 no. Dec 07 17:53:05 sorry Dec 07 17:53:16 maybe you flashed a bogus uboot/ Dec 07 17:53:28 nope, uboot should be fine Dec 07 17:53:34 err, wait Dec 07 17:53:37 checking something Dec 07 17:53:43 well, so I don't know. Dec 07 17:54:10 hi folks, can someone help me about the openmoko projects developer svn access. DAV does not ask for password and SSH method says permission denied(public key). Dec 07 17:55:01 is there some easy way to run openmoko in qemu or similar without emulating ARM? Dec 07 17:55:27 you can compile most of the applications for your host. Dec 07 17:55:38 that unknown was my bluetooth dongle Dec 07 17:56:07 i have a second device but works fine, that shouldn't be the problem. Dec 07 18:02:14 hmm. even if there's no way of, from userland, switching off the display refresh, I can still use switchvt to switch away from X? Dec 07 18:02:17 would that help? Dec 07 18:02:23 or is the chargen in software? Dec 07 18:08:01 openmoko: 03chris * r3615 10/trunk/src/target/OM-2007.2/applications/openmoko-messages2/ (7 files in 2 dirs): (log message trimmed) Dec 07 18:08:01 openmoko: * src/sms-contacts.c: (contacts_added_cb), (contacts_changed_cb), Dec 07 18:08:01 openmoko: (contacts_removed_cb), (contacts_iter_compare_func), Dec 07 18:08:01 openmoko: (sms_contacts_page_new): Dec 07 18:08:01 openmoko: * src/sms-contacts.h: Dec 07 18:08:03 openmoko: * src/sms-notes.c: (global_note_added_cb), Dec 07 18:08:07 openmoko: (global_note_modified_cb), (global_note_removed_ghrfunc), Dec 07 19:20:51 uh. Who was I talking to yesterday about apm's reporting of battery levels? Was it you, SpeedEvil? Dec 07 19:24:46 in what package I can find "urllib" for python? Dec 07 19:25:16 trying to explore pyroute Dec 07 19:29:09 nm, found the logs from yesterday (it was SpeedEvil) Dec 07 19:35:58 yes, it was me. Dec 07 19:50:34 Is it a recent change that /sys/.../chgstate now is to return "charger_present" and "usb_present" independently? The november-snapshot kernel seems to say simply "present" Dec 07 19:52:01 * * OM Bug 1102 has been created by balrogg(AT)gmail.com Dec 07 19:52:02 * * Qemu gadgetfs support broken Dec 07 19:52:03 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1102 Dec 07 19:55:31 does it happen to be a usb/hardware guru here? Dec 07 19:56:44 zash: I know quite a lot of electronics and USB outside of its implementation on the neo, and a fair bit on that Dec 07 19:56:52 not so much on the software side. Dec 07 19:57:03 hm. has the power management chip changed from the gta01 to the gta02? Dec 07 19:57:33 yes Dec 07 19:57:43 it's now the 50633? and was the 50606? Dec 07 19:58:09 dunno Dec 07 19:58:22 * mwester thinks the biggest difference is the the newer pm chip can supply +5v for the usb on the GTA02 Dec 07 19:58:26 SpeedEvil: i have solderd togeather a usb male and a 3.5mm 4 pole plug Dec 07 19:58:32 tonyg: yes Dec 07 19:58:39 zash: why? Dec 07 19:58:40 I'm figuring out what are permitted chgstate values by reading the patches Dec 07 19:58:48 thanks SpeedEvil , balrog-kun Dec 07 19:59:44 tonyg: i think there's perhaps some generic charger class in linux and it probably changed the values upstream, rather than in any particular driver Dec 07 19:59:57 oh. ok. Dec 07 20:00:08 they're different in an odd, ad-hoc kind of a way Dec 07 20:02:00 * * OM Bug 1103 has been created by balrogg(AT)gmail.com Dec 07 20:02:01 * * Qemu gadgetfs needs to use asynchronous IO Dec 07 20:02:02 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1103 Dec 07 20:02:16 interesting...in my debian package the urllib.py is in the main python package Dec 07 20:02:25 but it isn't in openmoko Dec 07 20:02:33 ahven, yes, I'd thought it was a core module... Dec 07 20:03:02 and is there a way to search through packages eg. apt-file search ? Dec 07 20:03:49 don't know, I'm afraid Dec 07 20:06:12 ScaredyCat: hey Dec 07 20:14:12 The X screensaver - does it actually turn off the LCD? Dec 07 20:14:20 as in, the refresh, not just the backlight Dec 07 20:15:11 that may do it Dec 07 20:15:24 USB tends to be a bit more robust Dec 07 20:15:33 USB? Dec 07 20:16:35 tonyg: mischan Dec 07 20:16:44 :) Dec 07 20:18:45 tonyg: i believe yes, it stops the transfer of data between framebuffer and display Dec 07 20:18:56 !seen Dec 07 20:19:21 ah! I wonder how it does that? Dec 07 20:19:51 tonyg: one easy way to check is run it in qemu and see if the clock updates when the backlight is off according to the console messages from qemu Dec 07 20:22:25 easy, if qemu were running ok for me :-( Dec 07 20:22:54 well, maybe it will work. the problems I was having were to do with the usb-keyboard support Dec 07 20:23:35 oh, related to the AUX and POWER buttons? Dec 07 20:23:44 so what should I look for in the qemu output? Dec 07 20:24:19 uh, sort of - AUX and POWER are CR and SP, and I can't type into an xterm Dec 07 20:24:23 but on my other dev box, I can. Dec 07 20:24:25 weird Dec 07 20:24:33 it'll say something like "LCD Backlight now 0/64" when the screensaver kicks in Dec 07 20:24:53 and when the transfer of data stops? Dec 07 20:25:07 hm, maybe I'll need to use the actual qemu monitor? Dec 07 20:25:14 you will only know if you see that the screen is not updating Dec 07 20:25:15 as opposed to reading qemu's stdout? Dec 07 20:25:28 no, it's on stdout Dec 07 20:25:29 ah, ok, gotcha. Dec 07 20:25:35 can you add the keyboard issue to the bugzilla? Dec 07 20:25:52 balrog-kun, I'll just now try to reproduce. Dec 07 20:26:30 there were no changes in anything that has to do with keyboard recenetly Dec 07 20:26:32 hm, actually the pointer seems off, too Dec 07 20:26:43 never mind. I'll just watch the clock now :) Dec 07 20:27:47 (how long is the default screensaver?) Dec 07 20:28:21 like five or three minutes Dec 07 20:30:21 btw, why is the background of the panel items usually black, but sometimes white? Dec 07 20:30:22 bug? Dec 07 20:31:20 i think there are images in the background? nt one solid colour? Dec 07 20:31:57 ah panel.. Dec 07 20:33:33 http://scap.linuxtogo.org/ Dec 07 20:33:34 577 Dec 07 20:34:17 ahven: the gpe blub? :D Dec 07 20:34:23 *blurb Dec 07 20:34:27 yes, some matchbox glitch Dec 07 20:34:31 or theme glitch Dec 07 20:34:40 ok Dec 07 20:40:00 wow. the clock is really not at all realtime :) Dec 07 20:40:14 balrog-kun: can you elaborate why python doesn't have urllib.py in the core package? Dec 07 20:40:20 ~1.5 real seconds per second... Dec 07 20:40:40 don't have a clue, what package to install for getting it Dec 07 20:41:04 ahven, you could investigate the *.bb bitbake scripts for python in the openembedded directory of your build Dec 07 20:41:26 ahven: no idea, i haven't written a single line of python in my life Dec 07 20:41:35 and i don't use packages Dec 07 20:41:39 ok Dec 07 20:44:34 finally. the LCD Backlight now at 0/64 message... Dec 07 20:44:41 ...aha! and I saw the digits update Dec 07 20:44:46 I'll wait for another one to be sure Dec 07 20:45:19 ahven: urllib is in python-netclient. found out by grepping through packages/python/python-2.5-manifest.inc Dec 07 20:45:33 oh :) Dec 07 20:45:53 tonyg: ouch, that would mean the framebuffer driver doesn't suspend Dec 07 20:46:17 there it goes again. yeah, looks that way. Dec 07 20:46:23 ok, good to know. Dec 07 20:46:32 uhmm.. which is true looking at the driver code :( Dec 07 20:46:41 doh :) Dec 07 20:46:43 it ony resets the palette but doesn't stop transfer Dec 07 20:46:55 is there a programmatic way to stop transfer from userland? Dec 07 20:47:00 which file are you looking at? Dec 07 20:47:25 drivers/video/s3c2410fb.c:s3c2410fb_blank() Dec 07 20:47:31 from userspace, i doubt Dec 07 20:47:56 or maybe.. Dec 07 20:51:24 when it's in text mode, ie VT 1 for instance, is the character generator in software or hardware? If it's in software, switching to text mode may not save any RAM accesses Dec 07 20:51:35 because the resolution is the same after all Dec 07 20:52:07 yes, it's in software Dec 07 20:52:31 damn. Dec 07 20:53:03 SpeedEvil, how does that line up with your memory of power benefits from killing X? Dec 07 20:53:46 i'm gonna enter the keyboard bug into bugzilla because i need to log out soon Dec 07 20:54:10 greetings all Dec 07 20:54:14 well, the bug where you can't use AUX and POWER after linux boots, only in u-boot Dec 07 20:54:23 I couldn't reproduce it, because qemu wouldn't let me click on the terminal emulator entry in the app startup window Dec 07 20:54:31 balrog-kun, no, aux and power worked ok Dec 07 20:54:36 it was the usb keyboard that didn't Dec 07 20:54:42 ... or so I *assume* Dec 07 20:54:43 question, how well has the pcf50606 worked out for the neo? reliable? Dec 07 20:54:53 really, I don't know enough to be able to file a good bug report Dec 07 20:55:00 tonyg: they worked in Linux? Dec 07 20:55:16 balrog-kun, I don't know. I'm sorry. Dec 07 20:55:23 they work on the real hardware ;-) Dec 07 20:55:43 tonyg: i'm gonna add that as a bug so i don't forget to check it on monday :) Dec 07 20:55:45 If I manage to pin it down, I'll file a bug Dec 07 20:55:47 okay :) Dec 07 20:56:09 pH5: could you find also where does cgi.py hide itself? according to debian, also should be in the core package :) Dec 07 20:56:33 * tonyg removes a chvt hack, since it isn't going to save any power anyway Dec 07 20:56:54 anyone know if mtn uses the $OMDIR in setup-env to locate the OE.mtn database ? Dec 07 20:57:53 just upgraded to gutsy from Fiesty ubuntu and all my external drives have been renamed by automount Dec 07 20:59:04 nm it doesn't. Dec 07 20:59:18 sry 4 the noise Dec 07 21:01:01 * * OM Bug 1104 has been created by balrogg(AT)gmail.com Dec 07 21:01:02 * * GTA01 Hardware keys don't work under Linux 2.6 in qemu Dec 07 21:01:03 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1104 Dec 07 21:10:04 pH5: found the place, no matter now :) Dec 07 21:26:57 evenin' Dec 07 21:27:54 pyroute's hotpoints for clicking are weird kind of Dec 07 21:28:05 can't find the right spot :P Dec 07 21:29:16 not very useful then :) Dec 07 21:29:36 yup Dec 07 21:30:00 only had to install almost all the python packages to get it working :) Dec 07 21:30:50 .. Dec 07 21:31:05 O_o Dec 07 21:31:24 yes, well Dec 07 21:34:17 http://pygps.org/ <--- excellent Dec 07 21:34:29 tonyg: you know your couple of packages for a UI-less OM... are they in the tree? Dec 07 21:34:57 no, very much external to the tree :-) Dec 07 21:35:09 If you'd like to experiment (how's your erlang?), I can upload my erlang .ipks Dec 07 21:35:14 or you can build your own Dec 07 21:35:44 Once you have those, it's a slightly manual process to get the userland running, but not too bad. Basically mkdir /.matchbox, ln -s /media/card/erlang-openmoko/priv/*.session /.matchbox/. Dec 07 21:35:50 and restart the X server Dec 07 21:36:02 Oh, I've been putting the userland on the SD card for ease of development Dec 07 21:36:11 but it could go anywhere if you change the hardcoded paths in the code ;-) Dec 07 21:37:05 (it has the ui started as part of it at the moment, but I'm considering ditching that (who wants to do GUI programming anyway? yuck) and implementing the GSMD protocol) Dec 07 21:37:30 er. or rather, *ahem*, why do you ask? :-) Dec 07 21:39:38 erlang works? Dec 07 21:39:47 univac, sure. it works great :) Dec 07 21:39:50 just like on my desktop Dec 07 21:39:55 (the gs libs aren't there) Dec 07 21:40:11 Hello = fun () -> io:format("hello world!~n") end. Hello(). Dec 07 21:40:12 ;d Dec 07 21:40:21 :) Dec 07 21:40:49 univac, see http://www.lshift.net/blog/2007/09/16/erlang-on-neo1973-cellphone Dec 07 21:42:47 tonyg: hmmm, no, my erlang is nonexistant :) Dec 07 21:43:12 tonyg: I'm just not familiar on how to hook into the boot sequence to replace everything myself Dec 07 21:43:23 all I did was create the /.matchbox/session script Dec 07 21:43:33 and then copied the default and chopped out what I didn't want Dec 07 21:43:40 (neod, openmoko-today etc) Dec 07 21:44:12 It will check $HOME/.matchbox/session before /etc/matchbox/session Dec 07 21:44:18 and $HOME is empty when it boots Dec 07 21:44:38 (warning: it isn't empty when you ssh in as root, so I've also symlinked /home/root/.matchbox to /.matchbox) Dec 07 21:44:44 xcasex: did you compile pygps yourself? Dec 07 21:44:52 * tonyg watches the netsplit repair itself Dec 07 21:45:33 why would you need to compile pygps? Dec 07 21:45:44 tonyg: ahh. I was thinking of also removing X :) Dec 07 21:46:11 nelson: http://familiar.handhelds.org/releases/v0.7.2/x/armv4l/pygps_0.06-1_all.ipk fails installing Dec 07 21:46:25 ERROR: Cannot satisfy the following dependencies for pygps: Dec 07 21:46:25 libgtk1.2 libglade0 gdk-imlib1 Dec 07 21:47:21 hrm. Yeah, you need the one that pavelm has (hopefully) packaged. Dec 07 21:48:12 only gtk2 packages in openmoko repos Dec 07 21:49:05 nelson: wth. shouldnt you be out there giving that gpe guy a hard time :p Dec 07 21:49:22 j/k Dec 07 21:51:01 * * OM Bug 1105 has been created by tonyg(AT)kcbbs.gen.nz Dec 07 21:51:02 * * Blanking GTA01 LCD doesn't switch off display refresh Dec 07 21:51:03 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1105 Dec 07 21:51:48 Stephmw, woo, hardcore! Dec 07 21:52:53 tonyg: in fact... I wonder if I can staticly link libc into my main executable :D Dec 07 21:53:05 tonyg: but that's not really Neo-related Dec 07 22:00:06 Stephmw, you're not thinking of doing anything tricksy like running squeak as the userland are you? ;-) Dec 07 22:09:21 tonyg: n...ooooo Dec 07 22:09:34 tonyg: I'm not quite that twisted Dec 07 22:09:45 (I considered it) Dec 07 22:09:50 (erlang is a much better choice) Dec 07 22:10:08 (it does run! ... a bit slowly, but that's just bloated smalltalk code rather than a slow vm) Dec 07 22:54:06 woot - out of tree kernel module builds, inserts and removes without crashing ! Dec 07 22:54:14 :) Dec 07 22:54:16 yoy Dec 07 22:54:37 any progress on the wifi driver ? Dec 07 23:06:46 CVirus: It works, that's what I know Dec 07 23:07:04 any progress on selling the gta02 to me yet? Dec 07 23:07:16 s/the/a Dec 07 23:08:08 yeah .. any progress on that too ? Dec 07 23:08:25 zash: it was talked some time ago and it should come in about 3 plutonian months Dec 07 23:09:10 and the pricetag will go a little up or stays the same? :) Dec 07 23:09:34 too early to talk about that though Dec 07 23:17:36 The price will be about 300,000 plutonian dollars. Dec 07 23:18:15 * mwester wonders why people seem to be happier when they have a fixed number for something, even if that number is meaningless ;) :D Dec 07 23:18:17 about the same in US dollars by the time its out Dec 07 23:18:39 given current FED policies Dec 07 23:19:32 well, the masses tend to get more interested in smaller numbers Dec 07 23:19:36 like $299.99 :) Dec 07 23:19:41 bah. It isn't the Fed. It's the idiots who mortgaged themselves way out of what they could afford with adjustable rate mortgages, and now suddenly find they can't pay it anymore. Dec 07 23:20:04 but its the fed bailing them and banks out. Dec 07 23:20:30 yay for fixed rate (so i know what it'll cost) and trying my damned-est to pay it off ASAP Dec 07 23:20:42 That part is partially true, anyway. Enough are failing to wake people up, it'll take time. And a lot of newly-homeless families. Dec 07 23:21:33 it will be easier to payback the $trillions in US debt to Chinese et all with the new $100,000 bills Dec 07 23:21:39 But what really gets me are the number of folks who are on the verge of losing everything, but still go out and rack up credit debt buying xmas gifts they can't afford. Dec 07 23:21:53 The Chinese are too smart for that. They'll want land. Dec 07 23:22:20 ha property owners Dec 07 23:22:34 when its all over they'll go back to keeping gold in their mattress Dec 07 23:23:18 * mwester is rather pleased with how well his gold investment in his retirement fund has been doing ;) Dec 07 23:23:37 * Writchie can't wait for his new Arabian Express Card Dec 07 23:23:51 from Citibank of Arabia Dec 07 23:23:54 * unknown_lamer can't wait for his new my beard and rifle let me steal shit from you card Dec 07 23:23:58 The Saudis will give you credit? Dec 07 23:24:06 it doesn't matter at all. USD is falling so the "price" is higer, even it didn't change much at all compared to other currency Dec 07 23:24:41 Asia is mostly dollar based Dec 07 23:24:47 but not for much longer Dec 07 23:24:58 hehe! that's the truth! Dec 07 23:25:10 But will they even consider the euro? Dec 07 23:25:22 rmb will be king Dec 07 23:25:31 eventually Dec 07 23:25:33 screw the euro, SEK ftw Dec 07 23:26:24 zash speaks the truth Dec 07 23:26:54 or we could just trade with cows Dec 07 23:27:24 16" wafers Dec 07 23:27:30 We already do. Dec 07 23:27:31 what would you do if your money..erhm cows catch BSE ? Dec 07 23:27:46 Penalties for non-delivery of milk of cows are very severe. Dec 07 23:32:08 * mwester calculates roughly how many gallons of milk would buy a GTA02 in the state of Illinois. Dec 07 23:32:38 no fair - subsidies all over the place Dec 07 23:33:01 they probably pay you to dump the milk Dec 07 23:34:26 Writchie: I was meaning - trading with cows - those cows that don't have any milk to trade face severe sanctions. Dec 07 23:34:26 well since 25% of americans think the sun orbits the milk, I wonder how many think milk comes from citrus trees Dec 07 23:34:30 * mwester computes distance to a Wisconsin dairy farm and calculates fuel costs vs local supermarket costs for milk. Dec 07 23:35:10 s/the milk/the moon/ Dec 07 23:35:10 Writchie meant: well since 25% of americans think the sun orbits the moon, I wonder how many think milk comes from citrus trees Dec 07 23:35:22 s/moon/earth/ Dec 07 23:35:35 * mwester realizes that he neglected to calculate the rather-significant weight of sufficient milk to purchase a neo, and wonders how one calculates shipping costs on the payment as well as shipping costs for the product... Dec 07 23:35:39 s/the milk/the earth/ Dec 07 23:35:39 Writchie meant: well since 25% of americans think the sun orbits the earth, I wonder how many think milk comes from citrus trees Dec 07 23:36:20 don't forget the carbon credits for the cows gasious byproducts Dec 07 23:36:37 * mwester scowls as this is becoming rather complicated now. Dec 07 23:36:42 Writchie: first of all americans aren't so stupid and it isn't inteligent to just sling mud at them. Second it's impolite to attack any nation at all Dec 07 23:37:42 Mononoke: since I'm American I can do it - and the 25% figure was just reported yesterday. Dec 07 23:38:02 ~blame Writchie Dec 07 23:38:03 * apt blames Writchie (and Canada) for all the evil in the world Dec 07 23:38:04 it's actually a true number of adults and I think they are voters Dec 07 23:38:23 the kids in school aren't that smart ;) Dec 07 23:38:40 statistics... everyone with some math skills can say that statistics is pain in math's butt Dec 07 23:38:46 Did the report say how many of those reject the science, vs how many are truly ignorant? Dec 07 23:39:54 actually the survey was 1 in 5 two years ago and now 1 in 4. Dec 07 23:40:08 Ouch Dec 07 23:41:23 20% - 25% are "scientifically savvy and alert" Dec 07 23:42:29 mwester: on the upside all these idiots feel good about themselves. Dec 07 23:42:33 Or don't understand the question and answer randomly. Dec 07 23:44:41 I think I found that article. It's quoting someone at Northwestern University right here in Chicago. There are crackpots all over, but generally NWU is reputable. Dec 07 23:44:55 my kids have now seen "An Inconvenient Truth" three times in class this year. Dec 07 23:44:58 Dunno if the link requires registration: http://www.nytimes.com/2005/08/30/science/30profile.html?_r=1&oref=slogin Dec 07 23:45:12 that's the one from a couple of years ago Dec 07 23:45:40 Writchie: Dunno if mine have, it's sad that that is what passes for scientific education in some schools. Dec 07 23:46:21 it's banned from schools in UK as propaganda IIRC Dec 07 23:46:46 * Writchie has to run BBL Dec 07 23:47:08 "Fewer than a third can identify DNA as a key to heredity" -- makes the thought of a jury scary, given the prevelance of DNA evidence anymore. Dec 07 23:48:15 aye... talk about dumbing down science in UK schools... Dec 07 23:48:43 they're now studying 'general trends'... no more practical science in classrooms Dec 07 23:51:28 mwester, you didn't find it scary before? Dec 07 23:51:42 Never thought about it before. Dec 07 23:51:54 luckily no juries as in US here Dec 07 23:52:09 * mwester shudders Dec 07 23:52:23 What, the cops or the military do the honors? Dec 07 23:52:25 mjr: 20 more years and the judges will be the children of today Dec 07 23:52:28 mjr: scared? Dec 07 23:52:37 mjr: err.. ok, 40 years :) Dec 07 23:54:31 mwester, you know, there's judges. (There are a couple of, hmh, well, let's say jurors assisting at least in district court, but with much more limited power than your US jury) Dec 07 23:56:01 ...laymember of the court, whatever Dec 07 23:59:03 * Mononoke snickers "Bruce Schneier's mail server only sends him the emails' hashes, just to make things a little more interesting for him." Dec 08 00:06:18 bye Dec 08 00:21:38 ScaredyCat: is your u-boot image from http://buildhost.automated.it/u-boot-good-for-sd-boot-r13_0_2632_0.bin SDHC-capable? Dec 08 02:19:58 hi Dec 08 02:20:33 why am i getting this error when using the USB networking in qemu "dummy_hcd dummy_hcd: Unlink after no-IRQ? Controller is probably using the wrong IRQ." Dec 08 02:21:35 its works fine with the old version (machine type "neo") **** ENDING LOGGING AT Sat Dec 08 02:59:57 2007