**** BEGIN LOGGING AT Thu Sep 13 02:59:57 2007 Sep 13 03:00:33 I've made/received a few calls with recent builds (last three days) Sep 13 03:00:52 hrrm Sep 13 03:00:57 * hozer starts off a build Sep 13 03:01:02 I'm not happy with the current gsmhandset.state, tho. For some reason, the mic is piped REALLY LOUD into the handset speaker. Sep 13 03:01:32 Make sure you read the stuff on http://wiki.openmoko.org/wiki/Phase_1_Software_Testing that other people have encountered. Sep 13 03:04:38 Does the neo send DTMF tones yet? Sep 13 04:41:19 moin Sep 13 04:52:56 moo Sep 13 04:53:17 baaaah ;) Sep 13 04:54:37 burrp Sep 13 06:22:46 good morning Sep 13 06:25:44 mornin' Sep 13 06:45:08 openmoko: 03abraxa * r2959 10/trunk/src/target/OM-2007.2/applications/openmoko-mediaplayer2/src/ (13 files): Sep 13 06:45:08 openmoko: Now informs user when appending a single file to the playlist, too Sep 13 06:45:08 openmoko: Auto-creates playlist directory now Sep 13 06:45:08 openmoko: Minor clean-ups/clarifications Sep 13 06:58:15 re Sep 13 07:02:40 abraxa_: Is the media player included in the default rootfs now? Sep 13 07:05:49 CM: Don't think so - I'll just keep working on it and let Mickey decide on when to include it Sep 13 07:06:02 Ok Sep 13 07:11:36 jo Sep 13 07:11:41 here again Sep 13 07:12:08 still trying to build that development thing Sep 13 07:12:21 You're note having any luck are you? Sep 13 07:12:32 NOTE: build 200709130907: completed Sep 13 07:12:49 last thing I did: bitbake vte Sep 13 07:13:18 Have you updated since? make update ; make setup ? Sep 13 07:13:26 i got at task some 800 the configure of vte_0.16.8.bb Sep 13 07:13:35 it cant find my ncurses Sep 13 07:13:47 it is installed Sep 13 07:14:07 ncurses of host or target? Sep 13 07:14:18 host Sep 13 07:14:21 try bitbake ncurses too Sep 13 07:14:23 for building Sep 13 07:14:31 there is? Sep 13 07:14:33 ok Sep 13 07:15:06 CM, can there also be something like ncurses-native for the host? Sep 13 07:15:19 native? Sep 13 07:15:33 as a package? Sep 13 07:15:54 yes, i think openembedded sometimes calls the host packages "native" Sep 13 07:16:10 wenn it is doing something with ncurses-native Sep 13 07:16:22 "wenn"?-) Sep 13 07:16:30 well* Sep 13 07:16:50 ok perhaps i guessed right %-) Sep 13 07:16:53 i sometimes use 'n' as a 'L' Sep 13 07:17:00 strange :) Sep 13 07:18:51 tuukkah: Yes, I had a lot of trouble with db3-native when I first tried to build the images Sep 13 07:19:05 actually all the packages in tmp/work/i686-linux are called something-native Sep 13 07:19:19 For cross building something or? Sep 13 07:19:34 cross dressing Sep 13 07:19:46 I really haven't understood bitbake, and I guess it's not as "sandboxed" as one could wish Sep 13 07:20:13 i lately got that old feeling of trying to compile a new piece of software from the internet, only to find that after having installed all the libraries that it needed, it finaly want a library that conflicts with the running system .. bummer (that was in the old days of linux...) Sep 13 07:20:20 CM, yeah it seems to be not like scratchbox Sep 13 07:21:44 it could be nice to have something like a tour of the oe directories Sep 13 07:22:06 tuukkah: Heh, sure would Sep 13 07:22:10 well that tour could last forever Sep 13 07:22:24 tmp/cross/ seems to have the cross compiler proper Sep 13 07:22:26 RalphEichelberge: Too many circular dependencies? ;) Sep 13 07:22:43 Wow, this would be nice to have in a laptop: http://www.linuxdevices.com/news/NS4331778531.html Sep 13 07:22:57 mhm " ln -s . something" :) Sep 13 07:23:27 tmp/rootfs is ... the rootfs :-) Sep 13 07:24:35 hi all Sep 13 07:24:45 hi zdanek Sep 13 07:25:19 I have a question, does somebody have a repo address where I get "Midori" from? Sep 13 07:25:57 I've been looking at scaredycat but it's only in filesystem build Sep 13 07:26:11 heh, was just about to say that Sep 13 07:26:21 bitbake it Sep 13 07:26:24 I don't want to flash all, I want to ipgk install it :) Sep 13 07:26:24 Have you checked rwhitby's feed? Sep 13 07:26:34 it's not in ipk form? Sep 13 07:26:36 * ScaredyCat checks Sep 13 07:26:36 nope - don't know the address Sep 13 07:26:47 hi ScaredyCat Sep 13 07:26:57 regards for repo Sep 13 07:27:25 it's in my repo Sep 13 07:27:29 * ScaredyCat can see it. Sep 13 07:27:31 there Sep 13 07:27:35 right there... Sep 13 07:27:38 :) Sep 13 07:27:40 * ScaredyCat points Sep 13 07:27:50 ipkg search '*midori*' shows nothing Sep 13 07:28:09 did you a) add my .conf b) ipkg update Sep 13 07:28:18 yep Sep 13 07:28:21 and updated Sep 13 07:28:26 ScaredyCat, did you update the package index ?-) Sep 13 07:28:27 uno momento Sep 13 07:28:36 according to wiki Sep 13 07:28:41 I always update the package index... prior to upload Sep 13 07:30:36 zdanek: Lots of files here http://ipkg.nslu2-linux.org/feeds/openmoko/unstable/armv4t/ Sep 13 07:30:36 doing upgrade it overwrites changes to /etc/init.d/gsmd script Sep 13 07:30:48 and it hangs again Sep 13 07:31:08 CM: thx Sep 13 07:31:55 and what's the difference between ../armv4t and ../fic-gta01 ? Sep 13 07:32:38 I think fic-gta01 is the model specific things Sep 13 07:32:58 Not general arm packages, but special for the openmoko dist Sep 13 07:33:08 * CM is guessing wildly Sep 13 07:34:11 openmoko-* packages are in armv4t though, probably because they run on other hardware too Sep 13 07:34:58 Sounds likely Sep 13 07:35:03 also, gsmd is armv4t but neod is fic-gta01 Sep 13 07:35:12 it all makes sense :-) Sep 13 07:35:56 yep :D Sep 13 07:39:09 ScaredyCat: what about your's repo and midori? Sep 13 07:40:46 hang on a tick... Sep 13 07:41:01 just in the middle of an update to the repo Sep 13 07:41:24 when it's done try an ipgk update then ipkg install midori Sep 13 07:41:32 I'll let you know when it's done Sep 13 07:42:05 it's done Sep 13 07:49:42 scaredycat: thx! Sep 13 07:50:01 np Sep 13 07:53:30 got Midori Sep 13 07:53:34 awesome :D Sep 13 08:03:26 still building ncurses -!!!! Sep 13 08:03:45 i cant believe this Sep 13 08:04:11 i hope its doing more than just that Sep 13 08:08:21 RalphEichelberge, well obviously the build dependencies need to be built first Sep 13 08:12:03 /clear Sep 13 08:19:40 Hello all Sep 13 08:20:35 is anybody here working i18n issues for OM? Sep 13 08:23:34 translations for example? Sep 13 08:35:10 Oh I have plenty i18n issues, thank you :) Sep 13 08:35:33 abraxa_, are you working on them, or reporting them? Sep 13 08:35:42 Working Sep 13 08:37:48 abraxa_, good, I just made today ru & he translation for feed reader, but for some apps in OM2007.2 cann't find some .pot or at least .op files Sep 13 08:38:18 Neato Sep 13 08:39:18 abraxa_, pity, it's takes smfn like 5-10 min to do it, but nothing to translate yet Sep 13 08:40:20 s/.op/.po/g Sep 13 08:50:13 Any_Key, . in the first part of s doesn't work like you think it does ;-) Sep 13 08:50:52 s/.o/*o/ Sep 13 08:51:17 where is apt when we need a demonstration? Sep 13 08:51:52 s/o/p/ Sep 13 08:51:52 tuukkah meant: where is apt when we need a dempnstration? Sep 13 08:52:45 or is apt .broken Sep 13 08:52:55 s/./somewhat / Sep 13 08:52:55 tuukkah meant: or is apt somewhat broken Sep 13 08:53:09 it is =-/ Sep 13 08:53:13 No, but he only does s// on your own text Sep 13 08:53:20 apt: don't you? Sep 13 08:53:42 CM, see, it doesn't interpret the first part of s as a regex Sep 13 08:54:11 That's true, should be easy to fix Sep 13 08:54:28 But not for me Sep 13 08:54:40 ~lart apt's regexp skillz Sep 13 08:54:40 * apt stamps apt's regexp skillz on the forehead with the official Troll marker Sep 13 08:54:50 LaF0rge, how high-level issues should we enter in the bugzilla vs. somewhere else? Sep 13 08:55:10 tuukkah: Example of high level issue? Sep 13 08:55:21 "The phone should take the necessary actions for the battery to never run too empty." Sep 13 08:55:47 That's definitely an issue.. It really dies suddenly Sep 13 08:56:05 this is more like a requirements item than a bug in some specific part of the system Sep 13 08:58:25 CM, i wanted to spec that we always need to be able to shut off the subsystems and negotiate the fast charge. preferably also show some indication for the user Sep 13 08:58:37 *nod* Sep 13 09:00:52 in case we're trying to boot the system, this means checking we have enough power to start with. and if we're not currently charging, we need some additional charge to bring up the system first Sep 13 09:01:58 I've got a patch that's aimed at that. Accurately predicting battery life. Sep 13 09:02:13 keeping track of battery capacity much more accurately. Sep 13 09:02:44 and very accurately predicting life, taking account of the draw of various subsystems and stuff Sep 13 09:02:51 battery aging, ... Sep 13 09:03:10 However, I need to get a build environment up to get it tested properly. Sep 13 09:03:55 As a first cut, simply a daemon that cuts off the power at 3.5V or so would be reasonable. Sep 13 09:04:02 openmoko: 03abraxa * r2960 10/trunk/src/target/OM-2007.2/applications/openmoko-mediaplayer2/src/playlist.c: Sep 13 09:04:02 openmoko: No more crashing when creating a playlist after deleting one Sep 13 09:04:02 openmoko: "Preliminary title"-flag gets auto-set now, allowing proper metadata handling on tracks whose metadata wasn't set in the playlist yet when the playlist got Sep 13 09:04:02 openmoko: saved Sep 13 09:04:18 SpeedEvil, sounds like something for neod Sep 13 09:04:47 probably. Sep 13 09:04:58 I'm doing the kernel-side bits first Sep 13 09:05:25 hmm, what will we be able to do in case we are in the sleep mode? Sep 13 09:06:20 do we periodically wake up and check if it's time to shut down completely? Sep 13 09:09:04 is it already possible to slow down the cpu like with cpufreqd? Sep 13 09:09:32 emdete, at least in uboot Sep 13 09:11:18 :D so linux will run at a lower speed all the time? saving battery power? Sep 13 09:13:03 openmoko: 03thomas * r2961 10/trunk/src/target/OM-2007.2/applications/openmoko-dialer2/ (ChangeLog src/moko-contacts.c): Sep 13 09:13:03 openmoko: * src/moko-contacts.c: (moko_contacts_lookup): Protect against empty number Sep 13 09:13:03 openmoko: value. Should fix bug 819. Sep 13 09:13:53 anyone able to check /proc/cpuinfo? my phone has too low battery to boot again, even though it was on fast charge Sep 13 09:14:59 tuukkah: what to check? Sep 13 09:15:15 does it tell the frequency? Sep 13 09:16:07 bogoMips Sep 13 09:16:17 somehow connected to freq Sep 13 09:16:24 Processor : ARM920T rev 0 (v4l) Sep 13 09:16:24 BogoMIPS : 132.71 Sep 13 09:16:29 the Bootloader page says "You can re-configure the S3C2410 PLL to generate a 266 MHz core cpu clock (rather than the 200MHz default)" Sep 13 09:16:31 half of clock Sep 13 09:17:08 so that would mean we do run at 266 MHz by default Sep 13 09:17:36 yes Sep 13 09:17:42 and it was said that some version (GTA01v4?) should work only on 266 Sep 13 09:18:00 I think it was at wiki's uboot/bootloader page Sep 13 09:18:05 zdanek, want to test what happens if you set the speed to 50 MHz in uboot? Sep 13 09:18:23 the page has " WARNING: The old hardware (Phase 0 - GTA01Bv3) only runs stable at 266MHz if you do not attach anything (esp. the FPC / Debug Board to the Debug Port" Sep 13 09:18:29 yes Sep 13 09:18:34 that what I ment Sep 13 09:18:59 hm, why doesn't /proc/cpuinfo conatain a 'cpu MHz'? Sep 13 09:19:02 I could as my wizards of electronics Sep 13 09:19:03 zdanek, i thought that means it doesn't run stable at 266 MHz Sep 13 09:19:20 !ombug 819 Sep 13 09:19:22 I does Sep 13 09:19:28 it does Sep 13 09:20:18 "does mean" or "does run stable"? the ambiguity of the language Sep 13 09:23:33 how can i switch cpufreq with uboot? Sep 13 09:23:40 tuukkah: I asked my colleagues and they said that if you lower the clock it shouldn't be hazardous for hardware Sep 13 09:24:16 it only affects everything which usues "time Sep 13 09:24:18 emdete, s3c2410 speed list Sep 13 09:24:28 emdete, s3c2410 speed set 50 Sep 13 09:24:29 " based on clock tics Sep 13 09:24:45 everything should be slower Sep 13 09:25:02 oups Sep 13 09:25:06 I don't know how it affects serial trasmissions Sep 13 09:25:13 if they're based on clock Sep 13 09:25:25 no, the serial port is gone also... ;) Sep 13 09:25:29 zdanek, ok, we might need to do some testing indeed Sep 13 09:25:57 wow, slower... isn't the right word for that :D Sep 13 09:26:16 I'd like to ask you what do you need it for? Sep 13 09:26:32 zdanek, to save the battery Sep 13 09:26:39 :D Sep 13 09:26:43 In principle, we want to slow the clock. Sep 13 09:26:44 ofcourse Sep 13 09:26:49 See the power managment page Sep 13 09:27:03 there are some 7 possible power saving techniques. Sep 13 09:27:09 no more explaination is required Sep 13 09:27:18 most of which we need to support Sep 13 09:27:52 I my company we create gsm+gps devices and we shut down gsm modem when it's not needed Sep 13 09:28:32 meant: In my company... Sep 13 09:29:09 http://wiki.openmoko.org/wiki/Neo1973_Power_Management Sep 13 09:29:16 SpeedEvil, what do you think of the state of bug #770? Sep 13 09:29:35 I think it is something I have no knowledge of Sep 13 09:29:41 !ombug 770 Sep 13 09:29:47 !om bug 770 Sep 13 09:29:56 * SpeedEvil ponders Sep 13 09:30:04 it would seem to me that someone should just add echo "AT@POFF" >/dev/ttySAC0 to gsmd stop and test that it works Sep 13 09:30:43 it does Sep 13 09:30:49 or rather Sep 13 09:30:58 I've not put it in there Sep 13 09:31:02 just used echo Sep 13 09:31:38 and the amp_mode to 'off' Sep 13 09:31:43 and it uses under 100uA Sep 13 09:31:56 for a battery life of around a year Sep 13 09:31:59 SpeedEvil, the amp mode part i would put to alsa-state stop instead Sep 13 09:32:37 so it doesn't get saved to the state that is restored after reboot Sep 13 09:32:42 tuukkah: Amp Mode to off should be added in suspend function in neo1973_wm8753.c Sep 13 09:33:17 XorA, oh ok, makes more sense. do we have a bug# ? Sep 13 09:34:02 tuukkah: no, Im due to do some work on sound suspend soon Sep 13 09:35:41 XorA, but soon probably isn't today so it might make sense for me to file the bug and put in what you said here? Sep 13 09:36:44 tuukkah: go for it Sep 13 09:36:58 tuukkah: you can even assign it to me :-) Sep 13 09:38:16 what do i put in there to assign to you? Sep 13 09:39:21 !oebug 2928 Sep 13 09:43:06 XorA, does that change take care of powering off too or only suspending? Sep 13 09:43:48 tuukkah: probably needs a mode change in the module remove stuff as well, but to be safe, shutdown scripts should load a mixer file to turn it off Sep 13 09:44:41 what about, what do you think of shutting the amp off in kernel when no sound is playing? Sep 13 09:44:52 Not an eventual solution Sep 13 09:45:10 SpeedEvil, what do you mean? Sep 13 09:45:23 the amp can be playing sound with the CPU off Sep 13 09:45:49 in what kind of situations? Sep 13 09:45:51 from the GSM modem, for example. Sep 13 09:45:56 tuukkah: thats ok, but GSM also uses the amp, and thats more difficult Sep 13 09:46:24 Ideally, we should even be doing that on calls. Sep 13 09:46:34 ultimately neod (or whatever the audio manager is) should load a state for off when nothing is using audio out Sep 13 09:46:38 turn off the CPU and screen when talking Sep 13 09:46:58 SpeedEvil, that sounds a bit extreme Sep 13 09:47:05 it should be an option Sep 13 09:47:54 even bluetooth should be possible with CPU off Sep 13 09:48:11 Heh, this was a nice new unit of measurement: mickey = ratio of computer mouse movement to onscreen cursor movement Sep 13 09:48:12 that would mean we can't do it in the kernel driver Sep 13 09:48:20 it's up to a third more talk-time Sep 13 09:48:26 (CPU off) Sep 13 09:48:56 SpeedEvil,"CPU off" here means suspend, right? Sep 13 09:49:34 no, actually off, with only RAM alive Sep 13 09:49:44 it needs some work to get it to do that rapidly though Sep 13 09:49:55 SpeedEvil, isn't that what suspend means? Sep 13 09:50:14 Well - the suspend mode uses - at the moment - a fair amount of power Sep 13 09:50:19 think about possibily to records voice call Sep 13 09:50:24 so it's not properly doing that. Sep 13 09:50:26 we need CPU for that Sep 13 09:50:33 however, not all the time Sep 13 09:50:44 It again should be an option Sep 13 09:51:04 for those that don't want to record calls Sep 13 09:51:08 SpeedEvil, yeah but that's what the suspend mode does wrt the cpu Sep 13 09:51:20 I would rather slooow down CPU clock just to react on inputs Sep 13 09:51:27 Unfortunately. Sep 13 09:51:35 slow clock may not save that much power. Sep 13 09:51:36 and if user taps screen it could start off Sep 13 09:51:43 in principle, we can wake from suspend really fast Sep 13 09:51:48 on screen touches Sep 13 09:51:54 sub-second Sep 13 09:52:08 this will take work though. Sep 13 09:53:55 do we have scripts that run on suspend and resume? Sep 13 09:54:02 I would add small phototransistor right near the speaker so we could know when user is listening to voice (so not wathing the screen) Sep 13 09:54:14 and then shut down all userinput and screen Sep 13 09:54:37 I mean during gsm call Sep 13 09:54:57 a button probably works Sep 13 09:55:05 'screen off' Sep 13 09:57:10 currently neod "Full PM" means the screen goes off automatically and we enter suspend state automatically too Sep 13 09:57:30 afaict this happens during calls too Sep 13 09:58:02 Hmmm, where does the name Neo1973 come from? Sep 13 09:58:19 Keanu Reeves making the first phone call in 1973 Sep 13 09:58:25 +gsm Sep 13 09:58:35 or +cell? Sep 13 09:58:55 -gsm+cell Sep 13 09:59:53 XorA, after this discussion, do you agree we should do the amp off in user space rather than kernel? Sep 13 10:00:24 tuukkah: well when audio driver is suspended I think Amp should go to off Sep 13 10:00:38 tuukkah: but thats a different state than halting the CPU during a phonecall Sep 13 10:01:20 how do these modes look different in the kernel? Sep 13 10:02:02 tuukkah: AFAIK kernel is either suspended or not suspended in its entirety Sep 13 10:02:06 hi all Sep 13 10:02:28 XorA, so CPU halt and suspend are the same thing? Sep 13 10:02:29 tuukkah: but the kernel can choose to halt CPU until IRQ occurs (this is called suspend sometimes in hardware land) but this is different Sep 13 10:02:41 oh ok Sep 13 10:02:44 tuukkah: too many uses for the same term Sep 13 10:03:06 XorA, so what's that called in kernel? "CPU halt"? Sep 13 10:03:23 tuukkah: Its what the idle loop does Sep 13 10:03:44 XorA, but how would we get to the idle loop during a call? Sep 13 10:04:21 tuukkah: because during call there will be nothing happening, we can turn off screen, CPU can sleep Sep 13 10:04:41 tuukkah: its not special behaviour in any way, in fact even desktop PC's already do it Sep 13 10:04:49 XorA, but there are processes that keep executing Sep 13 10:05:04 tuukkah: only if they are doing work, most processes wait on IO Sep 13 10:05:57 there are processes like clocks and gsmd that keep getting io Sep 13 10:06:47 tuukkah: but between IO they can sleep Sep 13 10:06:49 'idle' mode on the CPU is unfortunately high power. Sep 13 10:07:06 You really want it off, if you can. Sep 13 10:07:18 slow is a good alternative if you can't. Sep 13 10:07:55 SpeedEvil: I dont thinkt he CPU will ever truly be able to suspend while in action Sep 13 10:08:14 "suspend while in action"? Sep 13 10:08:18 If you don't want the screen, it's plausible. Sep 13 10:08:29 for some things at least. Sep 13 10:08:48 it will require a u-boot with a fast resume path Sep 13 10:09:19 ah that's cool. we can put part of power management in uboot Sep 13 10:10:02 (i'm specifically thinking of the "wake up periodically to check the battery isn't running too low) Sep 13 10:10:35 that sort of thing Sep 13 10:10:41 need a periodic alarm to check the battery every minute or so, so you need to weight the power costs of waking the CPU which are higher than the runnig costs against just idling the CPU in slow mode Sep 13 10:10:54 nowhere near every minute Sep 13 10:11:09 unless we know we are critically low on battery Sep 13 10:11:13 and on a call Sep 13 10:11:35 SpeedEvil: old batteries go from fine to critical in a matter of seconds Sep 13 10:11:54 Not really. Sep 13 10:12:03 with technology on Neo they do Sep 13 10:12:08 That's an artifact of insufficient understanding Sep 13 10:12:23 there isnt the fine grained reading of energy needed Sep 13 10:12:32 you can measure the battery impedence quite closely, which gives you an idea of the state of the battery Sep 13 10:12:40 age of the battery Sep 13 10:12:46 and you've got a good idea of power use Sep 13 10:13:10 does the PMIC in Neo do that measurement? Sep 13 10:13:12 fine-grained energy meters are nice. Sep 13 10:13:17 No, it doesn't Sep 13 10:13:20 alas. Sep 13 10:13:28 so stop dreaming and work with what we got :_) Sep 13 10:13:37 you don't need that though Sep 13 10:13:59 if you're not on a call, the current in use is quite tightly bound. Sep 13 10:14:10 screen brightness, CPU usage, ... Sep 13 10:14:18 sounds buggers that up Sep 13 10:14:18 and there is a good known upper limit Sep 13 10:14:31 loud sounds do effect it, yes. Sep 13 10:14:41 and impedance of headphones Sep 13 10:14:52 but you will never go from 50% battery state to 0% in minutes Sep 13 10:15:02 openmoko: 03chris * r2962 10/trunk/src/target/OM-2007.2/applications/openmoko-today2/ (ChangeLog src/today-launcher.c src/today.h): Fix bug with empty search and use MokoSearchBar in openmoko-today2 Sep 13 10:15:14 unless you're severely screwing up the sums, or assuming a linear battery curve like the current one Sep 13 10:15:20 well in the normal use case of suspend, the gsm chip is keeping contact with the base station and that takes a variable amount of power too? Sep 13 10:15:32 it's variable, but fairly low. Sep 13 10:15:36 ok Sep 13 10:15:39 30-50mA or so max Sep 13 10:15:40 tuukkah: and bluetooth is constantly polling Sep 13 10:15:49 Not while off. Sep 13 10:16:04 (at least without considerable u-boot hackery) Sep 13 10:16:10 bluetooth shouldnt be off while suspended Sep 13 10:16:12 If it's possible at all Sep 13 10:16:13 i'd expect in the normal use case bluetooth can be off Sep 13 10:16:25 No, bluetooth shouldn't. Sep 13 10:16:32 the right wires to support it are not there. Sep 13 10:16:51 SpeedEvil, shouldn't what? Sep 13 10:16:52 so what has to happen - it seems likely to be possible, but is not certain Sep 13 10:17:10 is to hack u-boot, so it supports a 'fast check' mode, which does stuff. Sep 13 10:17:59 this wakes every n seconds, turns on USB, asks the bluetooth module if it's had anything interesting happen, or if there is low battery, and then goes to sleep all within a few milliseconds Sep 13 10:18:21 it would seem to me power management is the most complex and laborous part of the whole openmoko/neo project Sep 13 10:18:24 If something interesting has happened, it goes to resume to RAM Sep 13 10:18:37 SpeedEvil: then you are calling reset on bluetooth, thats nasty and I doubt will work Sep 13 10:18:46 no, you're not Sep 13 10:19:00 Id be suprised if USB bus reset didnt reset bluetooth module Sep 13 10:19:02 the bluetooth module is on the other side of the USB bus, and doesn't have reset wired like that. Sep 13 10:19:11 why do we need to check for usb and bluetooth while in suspend? Sep 13 10:19:12 it can be told to go to sleep and turn off USB Sep 13 10:19:26 tuukkah: I press a button on my handsfree handset Sep 13 10:19:36 I don't want to dig the neo out of my pocket to get it to respond Sep 13 10:19:59 that sounds not the most important thing to worry about Sep 13 10:20:01 and as to USB - the bluetooth module is on USB Sep 13 10:20:20 no, it's something that is nice for assorted reasons though. Sep 13 10:20:46 there are a number of reasons why we want to wake up on interrupts, for short periods of time, without going all the way into linux Sep 13 10:21:14 for example, to ignore simple touches, but only wake on gestures across the screen Sep 13 10:21:44 SpeedEvil, last you were saying we don't need to wake up every minute, now you're saying we need to wake up every n seconds where n is very small? Sep 13 10:21:54 for different tasks Sep 13 10:22:19 if you are simply caring about battery use, you can adapt it to the battery charge state Sep 13 10:22:36 for example, on a fully charged battery, the first poll might be 2 hours. Sep 13 10:22:52 yeah Sep 13 10:23:21 you'd only want a minute if you are on a call, with the battery really low. Sep 13 10:23:29 so we'd have another pm mode "suspend with handsfree bluetooth polling" Sep 13 10:24:10 Something like that. Sep 13 10:24:11 Probably a stupid Q, but.. Is the battery smart enough to cut out by itself before it drains enough to damage the battery, or does that have to be handled by sw? Sep 13 10:24:20 it's hardware Sep 13 10:24:29 there is a thingy in the battery that cuts off at 2.8Vish Sep 13 10:25:21 bonus for first person to set fire to Neo Sep 13 10:25:43 SpeedEvil, before 2.8V there's no harm done or there's no dramatic harm done? Sep 13 10:25:46 openmoko: 03thomas * r2963 10/trunk/src/target/OM-2007.2/applications/openmoko-today2/ChangeLog: * Add a modeline to the ChangeLog to keep Chris happy Sep 13 10:25:49 No harm at all Sep 13 10:26:04 It may not be 2.8V - it's around there though Sep 13 10:26:15 ok good, so the software side is not that critical Sep 13 10:26:35 it's only to make fast charge etc. possible Sep 13 10:26:48 s/possible/always available/ Sep 13 10:26:48 tuukkah meant: it's only to make fast charge etc. always available Sep 13 10:27:04 In the totally flat battery case, we will start to charge at the 'qual' rate - which is about 10mA Sep 13 10:27:08 I think for 5 min Sep 13 10:27:14 this is _really_ suboptimal Sep 13 10:27:31 ok so s/fast/faster/ Sep 13 10:27:36 this is another thing we want the 'fast wake' for. Sep 13 10:27:44 it can do charger turning on. Sep 13 10:27:50 yeah Sep 13 10:28:26 can we wake up automatically to uboot when the battery isn't totally flat anymore? Sep 13 10:28:50 Possibly - I have to re-read the PMU docs Sep 13 10:28:53 (150 pages) Sep 13 10:29:06 or the user needs to wait some minutes, press the power button and then uboot can turn on the charger Sep 13 10:29:31 We never, ever want to run down the battery all the way basically Sep 13 10:29:51 SpeedEvil, but it can happen, there's no way around it Sep 13 10:29:56 Of course. Sep 13 10:30:01 99% perhaps but not 100% Sep 13 10:30:43 it's not *that* serious if we get 10mA for 5 min Sep 13 10:30:44 Sure - even if you reserve 5% power - you will lose that power over a month due to 'off' current. Sep 13 10:31:27 It means that the user has to wait 10 min before the phone will actually meaningfully start to charge Sep 13 10:32:07 why do they need to wait, doesn't the change to 100 mA happen automatically? Sep 13 10:32:30 this is another reason for fast-wake. If you can turn on the charger in under 1s, then it means you can start that process a lot earlier Sep 13 10:32:44 Alas, it's not 100mA, but 40mA Sep 13 10:33:18 do we have these documented anywhere? Sep 13 10:33:29 and yes, but if the battery is completely dead - 0V - as it will be if the internal chip has cut it off - Sep 13 10:33:39 then the charger starts out at a very low rate Sep 13 10:34:00 is this different from what you said already? Sep 13 10:34:06 I'm planning on going over the power managment page. Sep 13 10:34:20 we need more than one page Sep 13 10:34:33 I have a self-consistent internal model of the power system of the neo. I may not be expressing it well. Sep 13 10:34:59 Most of what's been said belongs on that page. Sep 13 10:35:12 jo Sep 13 10:35:38 i have completed the bitmake vte Sep 13 10:36:06 initiating "make openmoko-devel-image" Sep 13 10:36:07 there's info about the chips, the configuration we use them in, the functions we have in the kernel, the functions we have in user-space... Sep 13 10:36:45 and again it cant find the package vte Sep 13 10:36:49 True. Sep 13 10:36:52 then we have another set of info about how we'd *like* it to be, the requirements and decisions made about them Sep 13 10:37:14 I was planning on first making the page one sane large page. Sep 13 10:37:22 third set is documentation for end-users Sep 13 10:37:28 then breaking it out into a few. Sep 13 10:38:23 makes sense that i made a separate page about Power management menu, right? Sep 13 10:38:39 suppose, yes. Sep 13 10:38:41 I haven't seen it Sep 13 10:38:52 http://wiki.openmoko.org/wiki/Power_management_menu Sep 13 10:39:29 any feedback welcome :-) Sep 13 10:43:39 openmoko-terminal2_svn.bb' failed Sep 13 10:44:02 checking for MOKOUI... configure: error: Package requirements (libmokoui2 vte) were not met: Sep 13 10:44:10 No package 'vte' found Sep 13 10:44:23 after completed "bitbake vte" Sep 13 10:44:30 I DONT GET THIS Sep 13 10:45:03 you should perhaps check that vte is really done and there Sep 13 10:45:13 how? Sep 13 10:45:28 the scary directories under tmp Sep 13 10:45:47 ok Sep 13 10:46:00 tmp/deploy/glibc/ipk/armv4t/vte/vte_0.16.8-r2_armv4t.ipk at least Sep 13 10:46:12 * CM tries a bitbake -c rebuild openmoko-terminal2 Sep 13 10:47:41 tmp/work/armv4t-angstrom-linux-gnueabi/vte-0.16.8-r2/ Sep 13 10:47:42 openmoko@openmoko-celsius:~/moko/tmp/staging/arm-angstrom-linux-gnueabi/lib$ ls libvte.* Sep 13 10:47:42 libvte.a libvte.la libvte.so libvte.so.9 libvte.so.9.2.12 Sep 13 10:48:47 openmoko@openmoko-celsius:~/moko/tmp/deploy/glibc/ipk/armv4t$ ls vte* -l Sep 13 10:48:47 -rw-r--r-- 1 openmoko openmoko 274216 2007-09-13 12:10 vte_0.16.8-r2_armv4t.ipk Sep 13 10:48:47 -rw-r--r-- 1 openmoko openmoko 51232 2007-09-13 12:10 vte-dbg_0.16.8-r2_armv4t.ipk Sep 13 10:48:47 -rw-r--r-- 1 openmoko openmoko 369340 2007-09-13 12:10 vte-dev_0.16.8-r2_armv4t.ipk Sep 13 10:48:47 - Sep 13 10:49:13 i ll try "bitbake -c rebuild openmoko-terminal2" Sep 13 10:49:52 tmp/staging/pkgmaps/debian/libvte Sep 13 10:50:02 mjr: #openmoko.fi Sep 13 10:51:00 openmoko@openmoko-celsius:~/moko/tmp/staging/pkgmaps/debian$ ls libvte Sep 13 10:51:00 libvte Sep 13 10:51:04 is here Sep 13 10:51:42 there's vte stuff all over the place, hard to tell what's critical Sep 13 10:52:02 tmp/staging/pkgdata/armv4t-angstrom-linux-gnueabi/runtime/vte Sep 13 10:52:07 tmp/staging/pkgdata/armv4t-angstrom-linux-gnueabi/runtime/libvte Sep 13 10:53:37 well there is no libvte and in my runtime dir Sep 13 10:53:47 could it be that? Sep 13 10:54:51 * tuukkah is making a bold guess: it could be that Sep 13 10:54:52 RalphEichelberge: bitbake -c rebuild of both openmoko-terminal2 and vte works just fine here, so I wonder what's messed up on your comp Sep 13 10:54:59 openmoko: 03thomas * r2964 10/trunk/src/target/OM-2007.2/applications/openmoko-dialer2/ (ChangeLog src/moko-dialer.c src/moko-keypad.c): Sep 13 10:54:59 openmoko: * src/moko-dialer.c: (on_keypad_dial_clicked): Update status string Sep 13 10:54:59 openmoko: * src/moko-keypad.c: (on_dial_clicked): Close bug 814 - Number is Sep 13 10:54:59 openmoko: cleared when unable to dial Sep 13 10:55:47 i wonder too, believe me Sep 13 10:56:15 No package 'libmokoui2' found Sep 13 10:56:27 package openmoko-terminal2-2.1.0+r2879: failed Sep 13 10:56:44 so you got further? Sep 13 10:57:12 is thomas wood on this channel? Sep 13 10:57:13 no I did that bitbake -c rebuild openmoko-terminal2 Sep 13 10:57:18 RalphEichelberge: Do you have symlinks in the directory? Sep 13 10:57:27 no Sep 13 10:57:34 tuukkah: thos Sep 13 10:57:36 i didt nothing Sep 13 10:57:46 thos, ping :-) Sep 13 10:57:56 yesterday i "rm -r moko" Sep 13 10:58:08 weget the Makefile Sep 13 10:58:20 RalphEichelberge, perhaps bitbake -c mrproper does more that -c rebuild Sep 13 10:58:22 and started to compile with make Sep 13 10:58:31 RalphEichelberge: It's very odd, since no one else have had these problems when starting from scratch Sep 13 10:58:47 i dont know, why that should be nesesary Sep 13 10:59:07 RalphEichelberge, well i had problems about vte and terminal, i think it was because of parallel bitbake Sep 13 10:59:25 yeah, i should just build, period Sep 13 10:59:49 parallel? can i switch that Sep 13 10:59:58 RalphEichelberge: You have a dualcore? Sep 13 11:00:23 maybe - its not my PC Sep 13 11:00:23 RalphEichelberge, PARALLEL_MAKE and BB_NUMBER_THREADS Sep 13 11:00:44 off by default and safer that way Sep 13 11:01:21 tuukkah: But if he's running the default makefile, that stuff should be off shouldn't it? Sep 13 11:01:26 i didnt change anything Sep 13 11:01:33 so it should be off still Sep 13 11:01:41 ok Sep 13 11:02:12 i ll try to "bitbake libmokoui2" now Sep 13 11:02:36 but why dont bitbake do this automaticaly Sep 13 11:02:37 yes, that's the next step you got to that i meant Sep 13 11:03:05 RalphEichelberge, the dependency metadata can have bugs you know Sep 13 11:03:19 i build a lot of kernels in the old days but i had never such trouble Sep 13 11:03:37 oh i see Sep 13 11:03:55 there can be more bugs, there can be less bugs. the bugs might affect you, they might not Sep 13 11:03:56 shouldnt we report it somewhere? Sep 13 11:04:02 mooh Sep 13 11:04:14 meeeh Sep 13 11:04:15 RalphEichelberge: Better to figure out what is wrong first Sep 13 11:04:19 RalphEichelberge, good point, if we understand enough about what's going on Sep 13 11:04:27 i c Sep 13 11:04:39 well its compiling again Sep 13 11:04:43 tuukkah: Great minds think alike, or how is it the saying goes? Sep 13 11:04:47 ;) Sep 13 11:04:50 i will do my brake now Sep 13 11:04:58 CM =) Sep 13 11:05:11 reporting back lager Sep 13 11:05:14 later* Sep 13 11:05:23 RalphEichelberge: No, we like lager ;) Sep 13 11:06:12 if this is the correct bitbake file at least it doesn't include the dependencies: org.openembedded.dev/packages/openmoko-apps/openmoko-terminal_svn.bb Sep 13 11:06:42 tuukkah: Isn't it a 2 missing? Sep 13 11:07:17 ok Sep 13 11:07:23 so org.openembedded.dev/packages/openmoko2/openmoko-terminal2_svn.bb Sep 13 11:07:49 which does have DEPENDS = "libmokoui2" RDEPENDS = "vte" Sep 13 11:10:04 where are DEPENDS and RDEPENDS documented? not in bitbake manual at least Sep 13 11:11:41 tuukkah: isn't it the same as portage, which bitbake was based on? Sep 13 11:12:02 ok so do i need to learn portage now :-) Sep 13 11:12:22 ok "DEPENDS variable, which contains a space seperated list of “package names”" Sep 13 11:12:39 that much is what the bitbake manual says Sep 13 11:13:05 source package names for compile-time dependencies, that much i can infer from the context Sep 13 11:13:44 i think i read somewhere that RDEPENDS is runtime dependency Sep 13 11:14:43 tuukkah: RalphEichelberge please update your meta data its seriously out of date Sep 13 11:15:52 guys, what about sending DTMF tones while calling? Sep 13 11:16:04 important especially while in voice mail Sep 13 11:17:29 zdanek: How well can you whistle? :-) Sep 13 11:17:34 !ombug 814 Sep 13 11:18:00 borg_: give up, the bot has been dead for weeks ;-) Sep 13 11:18:08 :\ Sep 13 11:19:01 I can't whistle bi-tonally :D Sep 13 11:19:28 zdanek, imo it would be cool, if dtmf-tones were interpreted during a call, the numbers pressed would end up on the screen or something Sep 13 11:19:57 zdanek, easy way to transfer phone numbers ;9 Sep 13 11:20:34 inz: that would be niceness! Sep 13 11:20:35 I meant to push number and to send tone via gsm channel Sep 13 11:20:44 but what you said is also very good Sep 13 11:20:47 XorA, that change (DEPENDS = "vte libmokoui2") can't fix it all since libmokoui2 was already in DEPENDS still the build was missing that... Sep 13 11:21:31 zdanek: AT+VTD=X is the GSM comand for DTMF tones Sep 13 11:21:56 i just wish for more gsmd stability.. everything else seems less important. Sep 13 11:21:59 XorA, where X is one digit? Sep 13 11:22:08 tuukkah: yes Sep 13 11:22:28 cool, should be easy to add in dialer then Sep 13 11:23:44 yep Sep 13 11:23:51 XorA: thx Sep 13 11:24:01 so the keypad stays as a tab during calls Sep 13 11:24:03 I should learn more AT commands :) Sep 13 11:24:22 What, we can't manage with atdt and atz? Sep 13 11:24:41 is adding gsmd to inittab a smart idea? Sep 13 11:24:44 Lynet, the issue was on-call dtmf Sep 13 11:25:10 folken, it's in /etc/init.d like daemons typically are Sep 13 11:25:11 tuukkah: I know, that was an (admittedly botched) attempt at a joke. Sep 13 11:25:22 tuukkah: yes.. but it dies at random. Sep 13 11:25:23 Lynet_: atdt isn't DTMF, so you fail at the joke ;-) Sep 13 11:25:43 (it's some ISDN-based thingy instead) Sep 13 11:25:57 Lynet_, oh you just forgot a smiley then and it would have been a good joke :-) Sep 13 11:26:49 how about an incall menu with "send dtmf" then popup dialer, with a close button which returns to the call screen? Sep 13 11:26:51 what we really need is send 2400+2600 tone to seize lines old skewl style Sep 13 11:27:07 XorA: lol Sep 13 11:27:28 folken, i'd say tabs are better. what would a menu and a popup gain over that? Sep 13 11:27:52 tuukkah: hmm actually you are right.. tabs is the better idea. Sep 13 11:28:05 s/is/are/ Sep 13 11:28:06 folken meant: tuukkah: hmm actually you are right.. tabs are the better idea. Sep 13 11:28:23 folken, *one* idea *is* Sep 13 11:28:26 ;-) Sep 13 11:29:39 that's what one gets for correcting himself. Sep 13 11:33:10 folken: do you have the latest ipks with the console= removal and the gsmd restart in matchbox session? That's the path to gsmd stability. Sep 13 11:33:50 rwhitby: did all that. it still crashes after some tinkering (e.g. normal usage with thumb.) Sep 13 11:34:13 n00b question: How long is make setup supposed to take? Sep 13 11:34:25 anywhere between 2 and 6 hours Sep 13 11:34:33 folken, have you noticed that it has become very difficult to select items in Today with thumb? Sep 13 11:34:41 re Sep 13 11:34:55 I think the style of the user gnu is somewhat questionable Sep 13 11:34:55 (mostly depending on your network speed and monotone version Sep 13 11:35:12 tuukkah: yes. the full row should be selectable imo. Sep 13 11:35:25 svn: Unable to find repository location for 'http://svn.openmoko.org/trunk/src/target/OM-2007.2/artwork/themes/openmoko-standard-2' \ Sep 13 11:35:26 in revision 1 Sep 13 11:35:26 folken, it isn't ?! Sep 13 11:35:26 folken, it is Sep 13 11:35:34 hmm - looks like a SRCREV is missing Sep 13 11:35:35 tuukkah: acuttally it is selectable.. but it requires about 3 times tabbing. Sep 13 11:35:35 hello thos :-) Sep 13 11:35:47 test hhf423 Sep 13 11:35:48 tuukkah: sorry false alarm. :) Sep 13 11:35:57 folken, it's an issue with the jitter on the display and the scrolling Sep 13 11:35:58 i thought it just isn't sensitive somehow Sep 13 11:36:11 thos, could more jitter be tolerated? Sep 13 11:36:11 thos: I too have a faulty thumb :-) Sep 13 11:36:23 thos: all works perfect with styluys though Sep 13 11:36:26 tuukkah, possibly, mickey already added some compensation Sep 13 11:36:33 thos: according to debug out it recieves the click Sep 13 11:36:34 XorA, exactly :-) Sep 13 11:36:48 thos, are you working dialer still? Sep 13 11:36:49 we just need to get thumbs surgically improved Sep 13 11:36:59 working on :-) Sep 13 11:37:02 tuukkah, yes Sep 13 11:37:11 XorA, they need to be more pointy ;-) Sep 13 11:37:26 remove skin back past bone, grind bone to sharp point and card with stainless steel Sep 13 11:37:44 folken, it will recieve a "mouse down, move, mouse up" which trigers scrolling rather than a click Sep 13 11:37:44 thos, could you check that the command line options -d 12345 and --dial 12345 work and the number appears on the widget? Sep 13 11:38:08 thos: ah, ok. Sep 13 11:39:17 thos, only --dial=12345 worked when i tried to test and even then the number didn't appear on the widget Sep 13 11:39:44 whats interesting is that the buttons in the main menu (e.g. dialer contacts) work witouth problems. Its only the selection in the "application manager" where the running tasks are shown that it is difficult to select an application. Sep 13 11:39:49 tuukkah, hmm, all I get is "Not connected" Sep 13 11:39:57 another n00b question: Is the qemu emulation good enough to get started poking around in openmoko and devel/contribute stuff or is real hardware required? Sep 13 11:40:50 Lynet: for UI stuff, absolutely Sep 13 11:41:01 for low level drivers, no. Sep 13 11:41:30 thos, the dialog? so perhaps it works now Sep 13 11:41:58 tuukkah, it doesn't enter the number into the display though Sep 13 11:42:10 tuukkah, I would rather people used the dbus interface if they could Sep 13 11:42:27 thos, doesn't -d use dbus? Sep 13 11:42:54 tuukkah, use the dbus interface instead of calling openmoko-dialer Sep 13 11:43:06 what does -d use then? Sep 13 11:43:16 you are still calling openmoko-dialer Sep 13 11:43:25 calling/executing Sep 13 11:43:29 sure, i'm just curious Sep 13 11:43:52 tuukkah, openmoko-dialer *is* a dbus service Sep 13 11:44:02 (and i'm shell scripting so i don't know how to use the dbus interface) Sep 13 11:44:14 right Sep 13 11:44:25 now why on earth are you shell scripting the dialer? ;-) Sep 13 11:44:49 i made this funny app that lets you call a number from any app Sep 13 11:45:00 funny app? Sep 13 11:45:09 written as a shell script? Sep 13 11:45:21 funky app. Sep 13 11:45:31 by selecting the number and launching the script via the neod aux menu Sep 13 11:46:07 tuukkah, write it in python then ;-) Sep 13 11:46:27 thos, i can do that, but it's not the simplest thing. that's why i'm curious about the reasons Sep 13 11:46:50 the reason we have a dbus interface is exactly so you can write that sort of application Sep 13 11:47:09 writing a shell script may be easy, but it is far from ideal Sep 13 11:47:22 but you haven't told me what dialer -d does and why it doesn't do the dbus thing Sep 13 11:47:37 does anyone have a favourite basic introduction to dbus tutorial ? Sep 13 11:47:56 thougt: use uri's.. like callto: and some urihandlingthingy ? Sep 13 11:48:24 tuukkah, you are talking about two different things Sep 13 11:48:35 tuukkah, executing the dialer, and using dbus Sep 13 11:48:41 yes? Sep 13 11:48:48 tuukkah, yes executing the dialer will start using dbus Sep 13 11:48:53 but that misses the point Sep 13 11:48:58 you don't need to execute the dialer Sep 13 11:49:12 infact, that's the whole point of the dialer being a dbus service Sep 13 11:49:33 you can call the method through dbus and it will start the dialer only if necassary Sep 13 11:50:21 thos, dialer -d also seems to call using an existing dialer process if there is one Sep 13 11:50:41 tuukkah, of course it will, but you are still forking and creating a new process Sep 13 11:50:54 it doesn't matter, does it? Sep 13 11:51:12 well what's the point, since it will use the existing process anyway Sep 13 11:51:25 you might as well use the existing process directly through dbus Sep 13 11:51:47 thos, you're telling me this is an important efficiency to achieve versus the simplicity of being able to call a command in shell? Sep 13 11:52:38 tuukkah, i'm saying the reason we added a dbus service was so applications could communicate with the dialer Sep 13 11:53:06 tuukkah, calling the dialer through the command line is totally inferior way of accessing it Sep 13 11:53:15 tuukkah: it's analogous to the reason why you don't start a new Apache process for every single http request ... Sep 13 11:53:55 think of the dialer as a call initiation server, not as an executable Sep 13 11:54:03 rwhitby, you can do that, if there aren't too many requests per second for your system to handle the forks Sep 13 11:54:31 sure you can. you can reboot your system between http requests too ;-) Sep 13 11:54:36 dbus also makes it a whole lot easier to create apps with different languages Sep 13 11:55:12 man you people aren't understanding my point at all. this is one tradeoff between performance and ease-of-use. it's a tradeoff Sep 13 11:55:33 tuukkah, it's a trade off between that and "doing things the right way" Sep 13 11:55:48 so using an executable and handling it is easier than using the provided dbus interface? Sep 13 11:56:04 guaqua, yes, in the shell Sep 13 11:56:19 so you are going to tell me the phone is going to be used from the shell? Sep 13 11:56:22 guaqua: it probably is for someone who isn't familiar with dbus (like me). Sep 13 11:56:38 but if the dialer author tells me I should be using dbus to access it, then I learn dbus. Sep 13 11:56:48 rwhitby: i'm not familiar with dbus, either, but it can't be harder than inventing the whell again with process handling Sep 13 11:56:49 guaqua, i was making a shell script which is accessed from gui Sep 13 11:57:02 guaqua, there already is the -d option Sep 13 11:57:23 i wasn't even saying it shouldn't be removed at this point, i was just *asking* Sep 13 11:57:36 "deutschsparchige bull user society" i guess dbus.org isn't it. Sep 13 11:57:57 off course, you can even make a command line phone - this is an open source project Sep 13 11:58:37 http://dbus.freedesktop.org/doc/dbus-tutorial.html Sep 13 11:58:45 :| Sep 13 11:59:01 thos, thanks for telling me -d is an obsolete or otherwise wrong way to do this, could you help me get started with dbus by telling what options to pass dbus-send to achieve the same functionality? Sep 13 11:59:05 http://www.freedesktop.org/wiki/Software/dbus Sep 13 11:59:18 tuukkah, i'm just looking that up myself actually ;-) Sep 13 11:59:27 thos, cool :-) Sep 13 11:59:58 i'll promise to change to using that right away and change to python call later on Sep 13 12:00:31 zash_se, Lynet_: thx Sep 13 12:02:18 * rwhitby fixes the build for openmoko-theme-standard2 and openmoko-terminal2 Sep 13 12:03:01 rwhitby, was there still something wrong with the openmoko-terminal2 bb file? Sep 13 12:03:13 missing SRCREV Sep 13 12:03:24 (in sane-srcrevs.inc) Sep 13 12:03:58 which meant that we weren't always building a stable version? Sep 13 12:05:17 up until the last 4 hours or so, we've never been building a stable version Sep 13 12:05:55 ok Sep 13 12:06:16 and unfortunately, the "require conf/distro/include/moko-autorev.inc" line in openmoko.conf (which I say should *not* be there) means we will still not be building stable versions of openmoko-* packages. Sep 13 12:07:37 I'm not sure whether it was koen or mickey|zzZZzz who decided to put that in there, but I say it shouldn't be there. The OpenMoko developers should be forced to change the revision in sane-srcrevs each time they intentionally want to release a new (rudimentally tested) version of a package to the community for further testing Sep 13 12:08:05 I am considering making MokoMakefile patch out that line in protest. Sep 13 12:08:38 (or add known good versions to preferred-openmoko-versions.inc) Sep 13 12:08:42 rwhitby: Sounds reasonalbe to me Sep 13 12:09:54 Many people here, including me, are total oe and bitbake newbies and just want the image to build. Sep 13 12:10:36 tuukkah, ok, try this: dbus-send --type=method_call --print-reply --dest="org.openmoko.Dialer" /org/openmoko/Dialer org.openmoko.Dialer.Dial string:'450' Sep 13 12:11:10 tuukkah, infact, only this is required: dbus-send --dest="org.openmoko.Dialer" /org/openmoko/Dialer org.openmoko.Dialer.Dial string:'450' Sep 13 12:12:02 hi...where can I set the font size in the applikations? Sep 13 12:15:22 thos, i made that into a script openmoko-dialer-dial and it seems to work. thanks! Sep 13 12:16:00 tuukkah, cool :-) Sep 13 12:16:20 thos, would it be bad to include that script in the package? Sep 13 12:16:37 tuukkah, which package? Sep 13 12:16:45 openmoko-dialer Sep 13 12:17:22 tuukkah, how long is it? Sep 13 12:17:32 should be a one-liner ;-) Sep 13 12:17:44 that's what I thought too :-) Sep 13 12:18:02 thos, only "#!/bin/sh" and the line you gave with '450' replaced by "$1" Sep 13 12:18:11 i guess an example directory might be neat Sep 13 12:18:59 tuukkah, I haven't looked yet, but it should be possible to process a return value Sep 13 12:19:16 tuukkah, for example you could know if there is already a call in progress or something Sep 13 12:19:43 that would be cool to get as a return value from the script Sep 13 12:19:47 actually, there should at least be a method to check the dialer status anyway Sep 13 12:22:06 thos, is the dialer process meant to always be there even if no window is open? Sep 13 12:23:46 | checking for MOKOUI... configure: error: Package requirements (libmokoui2 vte) were not met: No package 'vte' found Sep 13 12:23:46 | Sep 13 12:25:25 tuukkah, it's a service, so yes Sep 13 12:25:39 tuukkah, if it's not running, dbus is supposed to start it when you call into it Sep 13 12:25:45 sorry, but I am trying to compile this thing now for over 2 weeks Sep 13 12:25:54 still not working Sep 13 12:25:55 thos, yeah it seems to even with this script Sep 13 12:26:23 i just did, what i found in the wiki page for MakeMoko or shuch Sep 13 12:27:18 hi! is the python-gtkhtml2 binding in openembedded? Sep 13 12:27:27 any of you guys own a GTA01? Sep 13 12:27:33 thos, but not always :-( i'm trying to understand what's going on Sep 13 12:27:33 bahadunn, yes Sep 13 12:27:49 thos: what do you think of it? Sep 13 12:27:49 tuukkah, yeah, i just tried it - it seems to start the dialer but also report missing method Sep 13 12:27:58 wget http://www.rwhitby.net/files/openmoko/Makefile Sep 13 12:28:04 bahadunn, um, what for? Sep 13 12:28:18 thos, then on second try it works Sep 13 12:28:20 josch_, probably, but we are using webkit in openmoko Sep 13 12:28:35 tuukkah, did you kill the dialer in between? Sep 13 12:28:43 that is the source of all i have on my workstations Sep 13 12:28:48 thos, not between the tries Sep 13 12:29:16 tuukkah, oh sorry, mine had an error connecting to gsmd Sep 13 12:29:21 thos: so where can I find webkit for python? Sep 13 12:29:22 RalphEichelberge: there are hundreds of people who have been able to build openmoko at various times using MokoMakefile, so instead of just repeating the error, you need to go back to basics and check all assumptions about your environment, Sep 13 12:29:31 josch_, no idea Sep 13 12:29:31 thos: I am thinking about selling them here when the GTA02 comes out and I am trying to get a feel for how the it is Sep 13 12:29:31 thos, yeah saw that one too at some point Sep 13 12:29:46 bahadunn, GTA02 is a whole different kettle of fish Sep 13 12:29:55 thos: I know Sep 13 12:29:59 ok, so where to start - which page? Sep 13 12:30:34 http://wiki.openmoko.org/wiki/MokoMakefile ? Sep 13 12:31:07 tuukkah, eek, I seem to have a zombied dialer process :-) Sep 13 12:31:53 thos, killall openmoko-dialer. the first time running the script after that, the dialer complains about gsmd and exits. the second time, it compiles about the interface but the process stays. the third time, it works Sep 13 12:32:17 tuukkah, i'll investigate after lunch Sep 13 12:32:24 RalphEichelberge: yep Sep 13 12:32:25 tuukkah, I think the dialer initialisation process could do with a little clean up Sep 13 12:32:42 thos, about the zombie, /etc/matchbox/session must be starting it the wrong way Sep 13 12:32:43 tuukkah, (which is one reason -d wasn't going to work) Sep 13 12:32:50 RalphEichelberge: and use a 32-bit Debian Etch or Ubuntu 7.04 host for most reliable results. Sep 13 12:32:58 thos, so you can remove -d now? Sep 13 12:33:19 tuukkah, hope so :-) Sep 13 12:34:03 thos, it would be really nice for hackability to include the shell script on the phone though, so people can quickly find how to access the dbus interface Sep 13 12:34:22 i am using ubuntu 7.04 Sep 13 12:34:38 just set up for that only thing openmoko Sep 13 12:34:44 openmoko-dial, and you see there's an openmoko-dialer-dial. try that, less that, go :-) Sep 13 12:35:07 i did all apt-get install for the openembedded.org Sep 13 12:35:24 and I really don't know what to do nest Sep 13 12:35:29 next* Sep 13 12:35:32 RalphEichelberge: 32-bit? Sep 13 12:35:36 yes Sep 13 12:35:43 ist a laptop so Sep 13 12:36:06 ok, I build on OM on xubuntu 7.04 regularly. Sep 13 12:39:22 ok - so - what should i do to finaly get a working development enviroment Sep 13 12:42:15 did you know that some guys broke iPhone's SIMLock? Sep 13 12:42:30 http://www.engadget.com/2007/08/24/iphone-unlocked-atandt-loses-iphone-exclusivity-august-24-2007/ Sep 13 12:42:48 hey again: where can I change the font size? In my OpenMoko build for the TX all fonts are _very_ small Sep 13 12:42:51 http://www.youtube.com/watch?v=U09iNOqnBL0 Sep 13 12:43:33 XorA, do forgot to tell me what to put in bugzilla to assing the bug to you? Sep 13 12:43:44 s/do/you/ Sep 13 12:43:44 tuukkah meant: XorA, you forgot to tell me what to put in bugzilla to assing the bug to you? Sep 13 12:44:04 tuukkah: assign to graeme.gregory@wolfsonmicro.com Sep 13 12:45:10 http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=842 "Whenever possible, Amp Mode should be Off" Sep 13 12:45:30 tuukkah: cheers Sep 13 12:46:08 is something happened to mailllists? I haven't received nothing today Sep 13 12:48:04 s/mailllists/maillists/g Sep 13 12:48:05 Any_Key meant: is something happened to maillists? I haven't received nothing today Sep 13 12:48:11 hi, is there any news about the bug #666? do we'll have a firmware upgrade? Sep 13 12:48:12 model name : Intel(R) Pentium(R) III Mobile CPU 1133MHz Sep 13 12:49:06 niclone: latest message on the list indicates a GSM firmware upgrade is required, but no information on whether that means a return to vendor to get it done, and the firmware upgrade has not been developed yet. Sep 13 12:49:33 rwhitby: ok :/ thanks Sep 13 12:49:42 zefanja: is it yours finger? :) Sep 13 12:49:58 jpp...too big ;-) Sep 13 12:50:38 ? Sep 13 12:50:52 nope, it's okay :) Sep 13 12:51:34 niclone: I tell you this bug #666 is pure evil :D Sep 13 12:51:53 What does that mean? Sep 13 12:51:53 NOTE: preferred version 0.5.0+svn20070913 of osb-browser not available (for item osb-browser) Sep 13 12:52:02 zdanek: i know :( Sep 13 12:52:44 zdanek: i'm about considering finding another sim card, but it's not easy to know if it will work or not (i'm in france) Sep 13 12:54:57 niclone:ask provider for non 3g card Sep 13 12:55:07 RalphEichelberge: my build (which started 2 days ago, but hasn't been running continuously all that time) just completed. Sep 13 12:55:08 if you have such possibility Sep 13 12:56:20 RalphEichelberge: that message means that somewhere, someone said that a specific version of that package is preferred for that DISTRO. It will choose the latest version instead and try and build that. If it builds, then it's usually not a problem. Sep 13 12:56:51 I must be going Sep 13 12:56:56 best regards to all Sep 13 12:56:56 bye Sep 13 13:03:04 well, every providers now give 3G card :/ Sep 13 13:07:30 every providers? Sep 13 13:07:33 where? Sep 13 13:08:13 amusingly, UK providers were at this time last year using 2g SIMs in 3g phones Sep 13 13:11:51 El_Salvador: in france Sep 13 13:12:10 El_Salvador: afaik Sep 13 13:13:28 El_Salvador: i'm looking at all operators, they all talk about 3G... so i suppose the sim card is 3G Sep 13 13:15:21 bye..I have to go Sep 13 13:15:56 rwhitby: thanks for mokomakefile, i was studying it this morning... Sep 13 13:17:39 just wondering, is there a reason you use " [ ! -e filename ] || " rather than " [ -e filename ] && ? Sep 13 13:19:48 at least some typical idioms are "wanted-condition || fix" and "pre-condition && proceed" Sep 13 13:21:25 i read it as "if filename doesnt exist, dont do the following" Sep 13 13:23:14 umm with || it means "if filename doesn't exist, do the following" Sep 13 13:24:25 and naturally with && "if filename exists then do the following". no unnecessary negations Sep 13 13:25:18 i thought && meant proceed if condition true and || proceed if condition false? Sep 13 13:25:31 anyone! Sep 13 13:25:48 jpcass, isn't that what i say? Sep 13 13:26:05 is ther a other way to develop for openmoko other than buidling that MokoMakefile Sep 13 13:26:09 ? Sep 13 13:26:26 jpcass, of course the direct meaning of the operators is boolean "and" and boolean "or" Sep 13 13:26:47 (with short-circuit evaluation) Sep 13 13:27:01 jpcass: those are just my personal idioms I guess. Sep 13 13:27:12 RalphEichelberge, you can use bitbake without mokomakefile Sep 13 13:27:16 [ ! -e filename ] is true if filename doesnt exist. Sep 13 13:27:33 so [ ! -e filename ] && xyz says "do xyz if fiename doesnt exist" Sep 13 13:27:41 I usually make the first clause mean what I want it to mean, and then adjust the operator and second clause to match Sep 13 13:27:44 and [ ! -e filename ] || xyz says "do not do xyz if fiename doesnt exist" Sep 13 13:27:53 jpcass, sure. i was talking conditions in general. Sep 13 13:28:19 rwhitby: ok, thanks - I just thought I might have been missing something... Sep 13 13:28:23 jpcass, it's "filename is missing or otherwise xyz" Sep 13 13:28:37 I tend to use || more than && for some reason Sep 13 13:29:00 rwhitby: i noticed - thought it might be because they look a bit like pipes. Sep 13 13:29:08 ;-) Sep 13 13:29:08 I think I usually put the second part in first, and then put the first part in there to inhibit it Sep 13 13:29:31 tuukkah: if i do - ala bitbake world - it says nothing to do Sep 13 13:29:50 jpcass: it also has to do with what the resulting complete expression will return to make - you don't want it to return false, or else make will stop with an error Sep 13 13:29:58 actually I think that is the main reason Sep 13 13:30:22 do someone know how to find that problem in configure Sep 13 13:30:23 ? Sep 13 13:30:33 RalphEichelberge, you should follow the instructions from start: http://wiki.openmoko.org/wiki/OpenMoko2007.2#Building_2007.2 Sep 13 13:31:16 rwhitby, very good point :-) Sep 13 13:31:17 rwhitby: ! ah! never thought of that - that must be why some expressions end in " || true " Sep 13 13:31:26 ye Sep 13 13:31:27 p Sep 13 13:31:54 there is method to the madness ;-) Sep 13 13:32:07 i thought so, its nice to understand the method! Sep 13 13:32:58 rwhitby, do you know "BUILD_ALL_DEPS" and do we use it? Sep 13 13:33:57 tuukkah: the openmoko-feed target uses it Sep 13 13:34:36 in general, "meta" packages in OE define it. Sep 13 13:35:03 (openmoko-feed.bb is a "meta" package) Sep 13 13:35:07 could openmoko-devel-image use it too so that we wouldn't miss any RDEPENDS in the rootfs? Sep 13 13:35:17 or would it work like that? Sep 13 13:35:35 t/wi26 Sep 13 13:35:36 oops Sep 13 13:35:41 tuukkah: RDEPENDS inclusion in the image is not affected by BUILD_ALL_DEPS Sep 13 13:36:05 if an RDEPENDS of the image is missing from the image, then it's a bug. Sep 13 13:36:19 ok, i just understood some old message on bitbake like that Sep 13 13:36:34 If a DEPENDS (note the missing 'R') is not in an image, then that's not a problem - if it needs to be in the image then it should be an RDEPENDS Sep 13 13:36:42 rwhitby, so there's no way for us to enforce that? Sep 13 13:37:11 tuukkah: what do you mean by enforce. it's doing the right thing already, so we would be enforcing the wrong thing. Sep 13 13:37:37 rwhitby, i mean it would be the right thing to make sure that all RDEPENDS of the image are in the image Sep 13 13:37:46 tuukkah: that is the case already Sep 13 13:37:53 rwhitby, there have been bugs Sep 13 13:38:12 it would be better if it was enforced by the build tools or if it was automatic Sep 13 13:38:25 it is automatic. based on RDEPENDS. Sep 13 13:38:52 rwhitby: mokomakefile can apply for openmoko, openembedded and bitbake but my svn checkout has no files in ~moko/patches... Sep 13 13:39:03 are there just none needed at the moment? Sep 13 13:39:09 jpcass: that is correct Sep 13 13:39:16 rwhitby, you said "if an RDEPENDS of the image is missing from the image, then it's a bug." do you mean a bug in bitbake or bug in the image description? Sep 13 13:39:29 a bug in the OE metadata, not bitbake Sep 13 13:39:36 yes Sep 13 13:39:43 (i.e. someone writing a .bb recipe didn't set RDEPENDS correctly) Sep 13 13:39:48 so that's what i meant by it not being automatic. you can make a bug Sep 13 13:40:02 rwhitby: that facility is just for providing quick temporary fix without having to push upstream sometimes? Sep 13 13:40:12 if you're talking about something that generates the correct values for RDEPENDS automatically, then that's a completely different question Sep 13 13:40:15 rwhitby, i thought didn't define the image correctly to include all RDEPENDS? Sep 13 13:40:30 jpcass: yes, for cases where we know it's wrong, but can't find someone to fix it upstream. Sep 13 13:41:16 tuukkah: as far as I know, if you define RDEPENDS correctly in each of the packages, and in the image recipe, then bitbake will correctly build and install all those referenced packages in the image rootfs. Sep 13 13:42:14 if, however, someone doesn't set RDEPENDS correctly for one of the packages, and no other recipe RDEPENDS on it, then that package will (rightly) not end up in the image Sep 13 13:42:16 rwhitby, but if openmoko-terminal2 has RDEPENDS="vte" and openmoko-devel-image has RDEPENDS="openmoko-terminal2", is vte going to be in the rootfs? Sep 13 13:42:28 yes it will Sep 13 13:42:48 it's a transitive relationship Sep 13 13:43:03 thanks, finally i got it :-) Sep 13 13:43:23 ok, you get to document it for the next guy. Sep 13 13:43:37 my skills in this language seem to be ~0 Sep 13 13:43:51 what's your native language? Sep 13 13:43:54 rwhitby, can you just suggest where at? Sep 13 13:44:00 rwhitby, finnish Sep 13 13:44:30 well, your english is infinitely better than my finnish, so I don't think you should be the one apologising Sep 13 13:44:31 tuukkah, tried openmoko-fi ? Sep 13 13:45:01 mbuf, we're very quiet and small bunch there - just like Finland in general ;-) Sep 13 13:45:13 rwhitby, ok thanks :-) Sep 13 13:45:45 Has anyone been able to build emotion lately? Sep 13 13:45:54 tuukkah, everything starts small Sep 13 13:46:04 * rwhitby is eternally grateful that so many people in the world go out of their way to speak english in IRC channels. Sep 13 13:46:21 rwhitby: Do you speak any other languages? Sep 13 13:46:41 CM: I built some emotion today at work ... ;-) Sep 13 13:47:06 rwhitby: I get this, both yesterday and today: http://rafb.net/p/lH6XGn29.html Sep 13 13:47:06 is there recommended documentation/wiki to build a .ipkg binary so it can be released at projects.openmoko.org? Sep 13 13:47:08 CM: I speak enough of a couple of languages to be able to not be mistaken for an american when I visit other countries ;-) Sep 13 13:47:23 rwhitby: Hehe.. Sep 13 13:47:25 rwhitby: surely the accent gives it away? Sep 13 13:47:36 Oy mate! Sep 13 13:47:43 XorA: yeah, i do have that benefit too Sep 13 13:47:48 i would like the sources to be able to build for target qemu as well for ARM for OM2007.2; Sep 13 13:48:09 AFAIK only the kiwis hate the ozzies, so your fairly safe abroad :-) Sep 13 13:48:23 CM: really only about 50 words in french, and maybe 10 in a couple of other languages. Sep 13 13:48:28 English and accent... Sep 13 13:48:34 XorA: how is my accent? Sep 13 13:48:42 for some reason everyone thinks Im South African Sep 13 13:48:54 Hehe Sep 13 13:49:09 hrw: well it would have been weird apart from the 10,000 polish living in Edin :-) Sep 13 13:49:11 XorA: the kiwi-aussie thing is a friendly rivalry, not a hatred. Sep 13 13:49:23 aj hav a sviidisch accent.. :( Sep 13 13:49:34 rwhitby: I know, used to have a Kiwi flat mate Sep 13 13:49:38 CM, pratar du finska ?-) Sep 13 13:49:48 tuukkah: Ei.. ;) Sep 13 13:50:08 tuukkah: Bara lite, det r svrt. Sep 13 13:50:26 CM, you probably know our finglish and finland-swedish accent :-) Sep 13 13:50:40 tuukkah: Sure do :D Sep 13 13:51:27 tuukkah: Used to have a gf in Tampere, so I've been quite a lot to finland Sep 13 13:52:08 * XorA has lived in Scotland for 20 years and still cant roll his Rs Sep 13 13:52:40 XorA: Watch more Simpson episodes with grounds keeper Willie ;) Sep 13 13:52:48 XorA, you know children go to therapy to learn that Sep 13 13:52:50 Scottish accents are cool Sep 13 13:53:08 That's not a very good scottish accent. Sep 13 13:53:15 Though I have heard worse. Sep 13 13:53:17 XorA, try to repead Craig Ferguson's accent, its cool Sep 13 13:53:24 CM: my openmoko-feed build is up to gnome-vfs (2069 of 3575) Sep 13 13:54:01 so I'll be able to report when I wake up. Sep 13 13:54:02 rwhitby: Ok. I'll see if I can figure something out with this emotion error Sep 13 13:54:08 :) Sep 13 13:54:26 NOTE: package emotion-0.0.1+cvs20070819-r0: task do_compile: completed Sep 13 13:54:30 it's past emotion Sep 13 13:54:36 Hmm.. Sep 13 13:55:23 rwhitby, i think i found the place where to explain DEPENDS and RDEPENDS and BUILD_ALL_DEPS :-) linked from the front page: http://www.openembedded.org/wiki/Bbfile Sep 13 13:55:34 I've got a good feeling about this build today ... finally a locked down SRCREV build. Sep 13 13:55:38 night all Sep 13 13:56:10 night Sep 13 13:56:15 night Sep 13 13:56:27 Sep 13 13:57:52 jpcass: ~ugt ;) Sep 13 13:58:36 mbuf: Im betting after last night, Parisians can all sing Flower of Scotland in true Scottish accent :-) Sep 13 14:02:01 any one see the news about that terranet mobile technology (mesh peer to peer using handsets) Sep 13 14:03:15 Mobile system promises free calls --> http://news.bbc.co.uk/2/hi/technology/6987784.stm Sep 13 14:03:34 I told my colleagues that opejnmoko could do that ... :-) Sep 13 14:03:40 likely Sep 13 14:05:44 http://blog.haerwu.biz/2007/09/13/openmoko-20072-testing/ Sep 13 14:06:00 worth reading test of OM 2007.2 Sep 13 14:07:05 webar7: I would think everyone who wanted WiFi in the Moko was anticipating something like that... Sep 13 14:07:26 hrw: Nice writeup :) Sep 13 14:07:38 people might find this interesting: http://www.openembedded.org/wiki/DirectoryStructure Sep 13 14:10:04 CM: next part probably in October Sep 13 14:11:45 hrw: Is this your review? Sep 13 14:12:23 wibbit: yes Sep 13 14:13:16 hrw: quite nice Sep 13 14:15:41 hrw: I just came on--would you report the URL? Sep 13 14:17:16 s/report/repost/ Sep 13 14:17:16 DukeOfURL meant: hrw: I just came on--would you repost the URL? Sep 13 14:17:24 http://blog.haerwu.biz/2007/09/13/openmoko-20072-testing/ Sep 13 14:17:58 thx Sep 13 14:20:43 hrw: Btw, why not a full post rss feed? Is that a feedburner setting or something? Sep 13 14:26:35 CM: I do not have time to check how to tell Wordpress to make short ver for homepage and full for rss Sep 13 14:27:45 could anyone with OpenEmbedded username help with their wiki? i'm not able to create an account or anonymously do some edits i need to do Sep 13 14:50:11 tuukkah: Ask around in #oe. The wiki surely needs updates Sep 13 14:50:38 i'll do that Sep 13 14:50:39 tuukkah: AFAIK OE wiki shouldnt need a login Sep 13 14:51:19 XorA, there's some problem with anonymous editing, and also i'd need to create new pages Sep 13 14:52:03 tuukkah: Just preview first Sep 13 14:52:24 I just edited http://www.openembedded.org/wiki/DirectoryStructure , adding a "your" to the text Sep 13 14:52:24 CM, it said something like "someone else changed the page already" Sep 13 14:52:47 let's continue on #oe ok? Sep 13 14:56:12 hi Sep 13 14:57:31 SpeedEvil: ping Sep 13 14:58:50 OLPC project has their resume down to 160ms Sep 13 14:58:56 * hhf423 yawns Sep 13 14:59:05 Writchie: yowsers Sep 13 14:59:39 Writchie: nice! :) Sep 13 15:20:38 re Sep 13 15:23:39 the bitbake method can be used to build a .ipkg to run on the target as well to run it on qemu? Sep 13 15:26:51 mbuf: qemu is supposed to be practically identical to the target Sep 13 15:27:41 polz, so it understands an ARM binary .ipkg and runs it? Sep 13 15:28:26 if it runs entire flash images, why wouldn't it run an .ipkg ? Sep 13 15:28:44 you just have to get it onto the running emulator Sep 13 15:28:59 polz, ok, thanks Sep 13 15:35:29 hi Sep 13 15:48:32 Interesting. Recommended - todays material world on speech synthesis. http://www.bbc.co.uk/radio4/science/thematerialworld.shtml Sep 13 15:48:39 (won't be up yet Sep 13 15:49:13 Including how latest modeling is done, and the revelation that you can put electrodes on each side of the vocal cords, and pick up the excitation of them. Sep 13 15:49:26 Which could be fun. Sep 13 15:54:28 hmm made two outgoing calls.. now gsmd reports: "gsmd_alive timer expired" "Modem dead!" Sep 13 15:54:35 (on the third call) Sep 13 15:56:01 is there a webkit browser for openmoko already in the works? Sep 13 15:56:15 guaqua, at least you can use midori already Sep 13 15:56:21 quaqua its allrready running. Sep 13 15:56:54 is LaF0rge avail for discussing the libgsmd / gsmd synchronisation issues ? Sep 13 15:57:12 how can i restart openmoko-dialer so that it asks for the pin again? Sep 13 15:57:20 sorry, Hi everybody should be first ... Sep 13 15:58:06 folken, the pin will be asked by the sim card only for the first power up. if you want it to ask it again, you must power cycle the modem. Sep 13 15:58:17 folken, guaqua, of course midori isn't the openmoko-specific one planned on the Web Browser page Sep 13 15:58:59 midori seems to be spending a lot of screen estate for just displaying a static menu Sep 13 15:59:04 i.e. poor gui Sep 13 15:59:05 daltona: but shouldn't gsmd tell dbus.. yoh! i need pin again. And next time dialer starts: "Enter pin" Sep 13 15:59:46 the pin would be needed only if the sim card has lost power then aplied power again. Sep 13 16:00:07 while the SIM is powered and PIN entered once, it will not ask it again. Sep 13 16:00:20 daltona: 17:53 < folken> hmm made two outgoing calls.. now gsmd reports: "gsmd_alive timer expired" "Modem dead!" Sep 13 16:00:33 17:53 < folken> (on the third call) Sep 13 16:01:00 folken, that is a well known issue in gsmd. Sep 13 16:01:46 sometimes the modem is on a state where it can not answer to a kind of keep-alive command, and gsmd considers that the modem is dead, when actually it's not dead Sep 13 16:02:06 daltona: hmm i just hung up on a call. Sep 13 16:02:38 I feel that that keep alive command is sometimes sent whan the modem is not in a 'good' state to answer it. Sep 13 16:03:47 anyone's build hanging on trying to download openmoko-standard-2 theme ? [http://pastebin.ca/695885] Sep 13 16:04:16 jpcass, have you updated the bb files recently? Sep 13 16:04:17 jpcass: mine is hanging on libgsmd0 version Sep 13 16:04:47 updated bb files with 'make update' ? yes Sep 13 16:06:51 packages/openmoko2/openmoko-theme-standard2_svn.bb has a wierd SRC_URI : Sep 13 16:07:18 some dude sitting next to me has an iphone which is interesting since I'm in canada Sep 13 16:07:19 its svn://svn.....;proto=http:// (is that normal?) Sep 13 16:08:40 wonder if he'll lend it to me to put openmoko on it. Sep 13 16:09:27 and for some reason its trying to fetch a .tar.gz from downloads.openmoko.org which is not mentioned in the .bb file Sep 13 16:13:32 OK its failing on an attempt to svn co the files from http://svn.openmoko.org Sep 13 16:14:12 the reason its failing is because it has -r 1 specified in svn co, if I remove that, I can svn co the stuff... Sep 13 16:15:04 anyone know why it would use -r 1 ??? Sep 13 16:16:13 why would bitbake svn fetcher use -r 1? i guess there is a variable being set somewhere? Sep 13 16:19:23 anyone know what variables affect the bitbake svn fetcher code? specifically revision!! Sep 13 16:24:14 hi, any idea why I can't mount /media/card ("special device /dev/mmcblk0p1 does not exit")? Sep 13 16:24:40 openmoko: 03thomas * r2965 10/trunk/src/target/OM-2007.2/applications/openmoko-dialer2/ (ChangeLog src/dialer-main.c src/moko-dialer-panel.c): Sep 13 16:24:40 openmoko: * src/dialer-main.c: (main): Delay gtk+/mokostock initialisation unless really required Sep 13 16:24:40 openmoko: * src/moko-dialer-panel.c: (moko_dialer_panel_class_init): Remove some extra debug messages Sep 13 16:28:38 im seeing the same problem with openmoko-terminal2, trying svn co -r 1 ... Sep 13 16:32:14 http://img301.imageshack.us/my.php?image=image425bq6.jpg Sep 13 16:34:28 ok, there is a variable called SVNREV used in make openmoko .bb files which is defaulting to 1 ;-( Sep 13 16:35:48 jpcass: I think this is part of the changes to rev based build Sep 13 16:35:49 it derives from SRCREV .... Sep 13 16:36:09 i think was this was completed some hours ago. Sep 13 16:36:23 Yes, especially that 1 thing Sep 13 16:36:49 whic is set in classes/patch.bbclass --> srcrev = bb.data.getVar('SRCREV', d, 1) Sep 13 16:37:04 jpcass: Before when I had a problem with qemu I did a PREFERRED_VERSION_qemu-native = "0.9.0+cvs20070701" Sep 13 16:37:05 you think its been sorted? i should try updating again Sep 13 16:37:07 hopefully this change when completed will stop the nonsense of changing the tires with the car rolling down the road Sep 13 16:37:43 a simple update isn't doing it for me. Sep 13 16:38:06 unfortunately, I think its going to be necessasry to rebuild it all Sep 13 16:39:08 ive been doing that all day! Sep 13 16:39:36 jpcass: my build is handling on wrong version for libgsm but the built version is correct so the problem is somewhere else. Sep 13 16:39:49 jpcass: i've been rebuilding for days and days Sep 13 16:40:00 [ and nights and nights ] Sep 13 16:40:03 i can catch it right only once per week Sep 13 16:40:29 but of course, the version I catch is almost broken Sep 13 16:40:33 So, sounds like the life of an openmoko dev isn't too dissimilar to that of gentoo users. Sep 13 16:41:17 ok i edited classes/openmoko2.bbclass and made SVNREV = "rHEAD" ... looks like its working (no fail yet...) Sep 13 16:41:19 its all management issues or lack thereof that go with the territory. Sep 13 16:41:24 you can't fire volunteers Sep 13 16:43:28 wow, sound in openmoko-mediaplayer Sep 13 16:43:52 I just built openmoko, but can't seem to run any apps when I click on them Sep 13 16:43:56 how can I get to a terminal Sep 13 16:44:22 cndougla, clicking is difficult Sep 13 16:44:25 cndougla, ssh? Sep 13 16:44:42 ah, should be more precise, using make run-qemu Sep 13 16:44:47 so within qemu Sep 13 16:44:49 no, teminal2 still tries for -r 1 and fails... Sep 13 16:44:54 cndougla, still should be able to ssh? Sep 13 16:45:01 to what address? Sep 13 16:45:01 cndougla: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 13 16:47:28 also added rev=HEAD to the default SRC_URI provided in classes/openmoko2.bbclass - trying again... Sep 13 16:47:54 thos: where would I ssh to? Sep 13 16:48:15 there's no wifi on the gta01, which is what the default makefiles build for Sep 13 16:48:32 and no real bluetooth under qemu Sep 13 16:49:38 cndougla - connect the neo to your pc with the usb cable and use usbnet Sep 13 16:49:53 jpcass, I'm running through qemu Sep 13 16:50:40 cndougla: sorry. i think there's a way to emulate this usb connection, never done it myself but i vaguely remember info on the wiki... Sep 13 16:50:57 been searching some, but not coming up with anything Sep 13 16:51:06 is there no easy way to bring up the console on this? Sep 13 16:51:27 start a terminal :P Sep 13 16:52:03 but I can't start a terminal Sep 13 16:52:13 this is like an endless look Sep 13 16:52:15 ;oop Sep 13 16:52:18 ... Sep 13 16:53:22 http://wiki.openmoko.org/wiki/OpenMoko_under_QEMU , specifically #Setting_up_USB_connection and #Connecting_to_the_serial_port.2FCommunicating_with_u-boot Sep 13 16:53:44 re Sep 13 16:54:00 how did I miss that... Sep 13 16:54:01 thanks Sep 13 16:55:46 Can someone help me? I'm still trying to build the openmoko-devel-image for x86.... And I have a new error now... http://nopaste.org/p/avAzknjAd Sep 13 16:55:53 cndougla: Once you have USB up and running, http://wiki.openmoko.org/wiki/USB_Networking Sep 13 16:56:01 ok Sep 13 16:57:27 does the latest OM builds allow successful GPRS connection? Sep 13 16:57:39 I never had one Sep 13 16:58:56 I don't have any experience with svns, can someone help me with the error? http://nopaste.org/p/avAzknjAd Sep 13 16:59:28 I've done it with 2007.1 - but just manually Sep 13 17:02:34 SpeedEvil, I wrote pppd script but I had some errors Sep 13 17:02:55 not with AT commands but with syncing to serial port Sep 13 17:03:43 zdanek: did you start from scratch or after gsmd Sep 13 17:05:28 what do you mean? Sep 13 17:06:47 "after gsmd"? Sep 13 17:07:31 did gsm initialize the modem, if it did there are unsolicited reports coming back Sep 13 17:10:53 Hi Sep 13 17:10:57 hi Sep 13 17:11:06 i tried to compile an image with mokomakefile Sep 13 17:11:16 for xephyr following the wiki Sep 13 17:11:32 SRCDATE_eds-dbus = "now" <-- this seems to be wrong Sep 13 17:12:03 make can't wget "http://downloads.openmoko.org/sources/trunk_svn.o-hand.com_.repos.eds-dbus__now.tar.gz" Sep 13 17:12:07 yes, they are changing the way they handle revisions in openembedded right now Sep 13 17:12:33 hi, wasn't there an easy way to change the screen to 640x480 instead of 480x640 in the newer builds? anyone care to tell me what it is ...? Sep 13 17:12:37 so should i just pick the latest? Sep 13 17:13:04 What is the default value? Sep 13 17:13:18 err... Sep 13 17:13:39 Try building without that setting Sep 13 17:13:41 It's wrong Sep 13 17:13:58 ok thanks Sep 13 17:14:07 Writche: it's hard to answer your question, I wrote pppd scripts as said on wiki Sep 13 17:14:35 launched it with and without gsmd Sep 13 17:15:50 Writche: I tell you what error I had in the moment Sep 13 17:16:29 I launched pppd call gprs Sep 13 17:16:39 and got [LCP ConfReq id=0x1 ] Sep 13 17:16:39 > Sep 13 17:16:41 so many people in this chan .... just want to know if someone managed to connect to a phone network from the gui. I used to manage to register to the network and get the gui to enter pin code but since last month and lots of images behind ... i still can't do it anymore. Is it a known issue ? am i the only one ? Sep 13 17:17:12 sunix: did you restart gsmd in your matchbox session ? Sep 13 17:17:15 sunix: did you add gsmd restart in mathcbox Sep 13 17:17:16 eheh Sep 13 17:17:28 restart ? :) Sep 13 17:17:33 eh Sep 13 17:17:36 from the gui ? Sep 13 17:17:38 yep Sep 13 17:17:49 when you restart gsmd all goes well Sep 13 17:18:07 add /etc/init.d/gsmd stop Sep 13 17:18:09 is that something like "start gsm antena" ? Sep 13 17:18:11 or Sep 13 17:18:11 and ... start Sep 13 17:18:13 ok Sep 13 17:18:15 i'll try Sep 13 17:18:17 :) Sep 13 17:18:23 in /etc/matchbox/session Sep 13 17:18:26 before dialer Sep 13 17:18:38 no it isn't Sep 13 17:18:55 yep ok ... wait a sec :) booting Sep 13 17:18:55 it's something like local workaround for gsdm errors Sep 13 17:19:05 but Sep 13 17:19:24 it should hang if you didn't patch gsmd script :D Sep 13 17:19:43 what do you mean by ... "hang" :p Sep 13 17:20:44 if it doesn't - don't ask :) Sep 13 17:21:04 i'm trying :) Sep 13 17:24:05 did it start? Sep 13 17:24:17 root@fic-gta01:~$ /etc/init.d/gsmd stop Sep 13 17:24:17 Stopping GSM daemon: stopped /usr/sbin/gsmd (pid 1137). Sep 13 17:24:17 gsmd. Sep 13 17:24:17 root@fic-gta01:~$ /etc/init.d/gsmd start Sep 13 17:24:34 lol Sep 13 17:24:36 alech, http://wiki.openmoko.org/wiki/Functions_menu Sep 13 17:24:46 and ... nothing ... Sep 13 17:24:54 nothing -what? Sep 13 17:25:02 didn't blow? Sep 13 17:25:17 i thing my moko is dead :p Sep 13 17:25:21 think Sep 13 17:26:03 may be the "hang" you were talking about Sep 13 17:26:49 yep Sep 13 17:27:10 http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=788 Sep 13 17:27:21 solution is at the bottom Sep 13 17:27:26 zdanek: thx :) going to dig that :) Sep 13 17:27:44 and then add restart lines in matchbox session Sep 13 17:27:53 :) Sep 13 17:28:23 I have a question regarding security. are we always going to run with root privileges? or will this change by the GTA02 release? Sep 13 17:28:34 Not GTA02 prolly Sep 13 17:28:43 but maybe by mass market time Sep 13 17:28:50 certainly, rather Sep 13 17:29:50 SpeedEvil: thanks. it'd be a bit dangerous to mistype typing at the terminal when you're on the go and see the filesystem wiped clean :) Sep 13 17:30:13 xaid: hopefully there will be some discussion on security model and approach instead of implement first design later Sep 13 17:31:40 SpeedEvil: did you catch my comment ealier on OLPC guys now having resume < 160ms Sep 13 17:33:03 Greetings #OpenMoko Sep 13 17:33:32 xaid: I did that yesterday - cleaned whole filesystem Sep 13 17:33:51 I was wondering if anyone had any problems building uicmoc. Sep 13 17:34:12 I ran gllin with chroot and had mounted whole fs to subdirectory Sep 13 17:34:25 then forgot to unmount before deleting dir Sep 13 17:34:28 poof! Sep 13 17:34:45 Writchie: this would be ideal. im surprised there isn't much about this on the wiki. Sep 13 17:35:35 zdanek: this would be bad if it had your address book...etc. a backup solution for the basic PIM applications would be a life saver Sep 13 17:37:02 there was little discussion when i asked on the list - its low priority - but there is some discussion in PM and should be a wiki entry soon to discuss possible approaches Sep 13 17:37:36 Does anyone have any suggestions about troubleshooting uimoc build problems? Sep 13 17:37:37 Casten: See http://wiki.openmoko.org/wiki/FAQ (and also SH1, SH1_FAQ, Neo1973, Developer_preview and Neo1973_Hardware#GTA01Bv4 pages) (Hopefully these links answers your question. This autotrigger for new nick and question during first 15 minutes done only once.) Sep 13 17:38:08 Thanks autotrigger Sep 13 17:38:23 Casten: this was a problem a few back. are you on ubuntu 7.04 Sep 13 17:38:25 hi, any way to simulate a right click? I'd really like to play some scummvm, but apparently the game needs right clicks a lot ... Sep 13 17:39:01 yes, I'm running FF Sep 13 17:39:27 Is there a better distro to use? I'll happily change. Sep 13 17:39:51 My goal is fewest problems, I'm not tied to any distro/ver. Sep 13 17:40:04 when this problem appeared before it was solved by added a symlink - for libXext.so IIRC Sep 13 17:40:05 Casten: 32-bit debian Sep 13 17:40:25 I'm on 64-bit gentoo, and I run into issues all the time :D Sep 13 17:40:50 twice the bits 4 times the pain Sep 13 17:40:53 Yeah, I'm using 32bit Ubuntu 7.04 Sep 13 17:41:03 OpenMoko is based on Debian anyway :-) Sep 13 17:41:27 Casten: And about your problem: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=747 Sep 13 17:41:30 ok, so symlink for libXest.so Sep 13 17:42:13 God this IRC client is bad.. no copy/paste Sep 13 17:42:31 hmm, disregard Sep 13 17:42:32 * CM hands irssi to Casten Sep 13 17:42:38 Casten: iPhone? ;) Sep 13 17:43:20 what, they can install software on the iPhone now? ;-p Sep 13 17:44:02 * hhf423 does the ipkg upgrade dance Sep 13 17:44:26 actually, the problem I am having doesn't look like 747's description Sep 13 17:44:40 guys! I just ran ipkg upgrade which rewrties gsmd script and reenables bug #788 Sep 13 17:44:49 but that's not important Sep 13 17:45:04 while doing upgrade my neo rang Sep 13 17:45:06 It seems to not be recognizing some trailing command options Sep 13 17:45:08 .obj/release-static-emb-x86/qfsfileengine_iterator_unix.o -lz -ldl Sep 13 17:45:09 | /usr/bin/ld: cannot find -lz Sep 13 17:45:27 and I was chatting with my wife while neo hung because of gsmd restart Sep 13 17:46:02 but phone call survived Sep 13 17:46:13 zdanek: thats fun :-) Sep 13 17:46:19 now my wife end call and neo is dead with dialer's 'talking' message Sep 13 17:46:25 it freezed Sep 13 17:46:36 very lol Sep 13 17:47:29 Anyone have any ideas about the problematic -lz option? Is this something worthy af posting to openmoko-devel? Sep 13 17:47:47 Casten: install zlib Sep 13 17:47:54 what's problematic about -lz? Sep 13 17:48:02 .obj/release-static-emb-x86/qfsfileengine_iterator_unix.o -lz -ldl Sep 13 17:48:02 | /usr/bin/ld: cannot find -lz Sep 13 17:48:46 install zlib-dev/zlib-devel depending if you deb or rpm based Sep 13 17:48:57 tyvm, trying Sep 13 17:51:28 zdanek: :) Sep 13 17:52:33 * hhf423 yawns Sep 13 17:53:07 zdanek: thank you i managed to have the dialer asking for my pin code :D Sep 13 17:53:19 going to lunch :) see you later guys :) Sep 13 17:53:30 (-lunch + dinner) Sep 13 17:53:34 bye sun Sep 13 17:53:38 :) Sep 13 18:16:37 mailinglist broken again? Sep 13 18:23:22 fgau: looks like that did it, thank you Sep 13 18:24:23 Casten: enjoy the build process :) Sep 13 18:25:04 yes, it seems lengthy. Hopefully full rebuilds aren't required often Sep 13 18:26:25 on my laptop 6 hours, duo core Sep 13 18:26:54 Is fetching the process of getting the file from svn? Sep 13 18:27:16 I'm getting a fetch failure, when that happens should I just retry? Sep 13 18:28:33 Had I known it was so slow, I would have gotten a faster laptop Sep 13 18:29:30 Casten: http://wiki.openmoko.org/wiki/MokoMakefile#Tips Sep 13 18:29:33 can't you use distcc and compile on multiple machines? :) Sep 13 18:29:45 I think my initial build has beengoing for 12 hours now, so don't complain. ;-) Sep 13 18:30:11 try PARALLEL_MAKE and BB_NUMBER_THREADS Sep 13 18:31:26 I'll do that, good tip. Sep 13 18:35:08 is this working for anyone here: http://wiki.openmoko.org/wiki/Application_Development_Crash_Course ? Sep 13 18:36:55 I just did small car route to my grandma with GPS pos storring Sep 13 18:36:59 wanna see? Sep 13 18:37:01 http://www.gpsvisualizer.com/display/1189708426-06913-83.17.74.122.html Sep 13 18:37:15 started from right-bottom Sep 13 18:37:34 zdanek: slick, watch out for wolves on the way to see her though Sep 13 18:37:43 then to north and returned after receiving delicious dinner :D Sep 13 18:37:43 zdanek: how did you get the data from the chip ? Sep 13 18:38:04 zdanek: with Neo GPS? Sep 13 18:38:08 yep Sep 13 18:38:15 I'm talking about that Sep 13 18:38:18 oh, how did you do that? Sep 13 18:38:22 is there a daemon available for the new distro ? Sep 13 18:38:51 zdanek: hooow ? :) Sep 13 18:39:00 nope nope Sep 13 18:39:04 via gllin Sep 13 18:39:06 with chroot Sep 13 18:39:09 http://wiki.openmoko.org/wiki/Gllin Sep 13 18:39:34 can you download gllin even? Sep 13 18:39:44 zdanek: did that little bit of documentation work for you? Sep 13 18:39:51 nope Sep 13 18:40:14 I got whole scripts from my colleague which found it on the net Sep 13 18:40:16 wait Sep 13 18:40:46 i figured someone somewhere put the binary up, but ive not really tried hard to find it Sep 13 18:41:37 it's a mix with that link above (to run gllin OABI) Sep 13 18:41:40 and http://wiki.openmoko.org/wiki/GPS_scripts Sep 13 18:41:42 I did search a bit, but did not find it. did not spend much time on it though because the documentation did not look like it would be easy to use Sep 13 18:41:46 to collect nmea frames Sep 13 18:42:29 zdanek: so that worked for you? Sep 13 18:42:29 I will upload whole chroot package on my OM page soon Sep 13 18:42:41 but I can't upload it with gllin Sep 13 18:42:54 i think gllin is what people are mostly after Sep 13 18:43:18 |\/|ail me Sep 13 18:43:23 * hhf423 'd be interested in gllin and the chroot package Sep 13 18:43:42 at bartek(DOT)zdanowski(ATTT) gmail(DOT)c0m Sep 13 18:44:20 hhf423: yes it worked Sep 13 18:44:25 as you see Sep 13 18:44:38 on that map-page Sep 13 18:44:44 zdanek: btw, how did you make that plot? Sep 13 18:44:54 is there an interface to feed NMEA goop into some google maps machup? Sep 13 18:44:57 mashup Sep 13 18:46:49 can google earth plus not read NMEA? Sep 13 18:47:29 nope Sep 13 18:47:34 SpeedEvil: did you not say that there would be a GPS related announcement soon Sep 13 18:47:34 you have to use GPSBABEL Sep 13 18:47:38 outstanding soft Sep 13 18:47:51 hhf423: yes Sep 13 18:47:53 and put it into gpsvisualizer.com form Sep 13 18:48:23 I will put an article on my Neo blog Sep 13 18:48:31 I'll do it right now Sep 13 18:50:58 Is anyone working on a color-blind friendly theme yet? since its Gtk im pretty sure I can make something up when the mass market release is near, but I was curious if anyone has started working on different color themees in general. Sep 13 18:50:59 nice Sep 13 18:51:36 I was in Warshaw once, a while ago Sep 13 18:52:19 still have polish money somewhere. I hope I can spend it before you guys join the euro :-) Sep 13 18:52:58 I hope we will get euro soon :) Sep 13 18:53:03 xaid|work: did not hear nothing about a colorblind theme Sep 13 18:53:17 some mails went out... Sep 13 18:53:36 SpeedEvil: you still in good hope for that? not loosing faith yet? Sep 13 18:54:37 * hhf423 wonders what to do with the P1 one Neo when the P2 comes out. Keep it? Sell it? Blend it? Sep 13 18:55:15 hhf423: I can help reviewing themes if someone needs the opinion of a color blind person :) Sep 13 18:55:41 xaid|work: you should put that into the wiki somewhere .-) Sep 13 18:56:17 woohoo my contacts work. :) Sep 13 18:57:05 hhf423: I'll add it later today :) Sep 13 19:00:11 hhf423 did you enjoy somehow Poland? Sep 13 19:01:59 zdanek: I was there only short, on a business trip, but I enjoed it, yes. only the way people drive I found a bit scary Sep 13 19:02:12 zdanek: only moscow and italy is worse :-) Sep 13 19:02:36 thanks to all :) i managed to make my first moko call :D Sep 13 19:02:47 * hhf423 congratulates sunix Sep 13 19:02:51 :D Sep 13 19:03:10 sunix: I hope you called someone meaningfull :-) Sep 13 19:03:19 libmokoui has changed in OM2007.2 ? Sep 13 19:03:43 i called my self :p Sep 13 19:03:44 sunix: congrats Sep 13 19:04:05 hhf423: yes, Warsaw is some kind on "wild-west" on the streets Sep 13 19:04:19 drivers are quite in-hurry Sep 13 19:04:35 I bought a car 2 months ago Sep 13 19:04:38 you could say that :-) Sep 13 19:04:40 code earlier built with OM2007.1 builds fine with OM2007.2, but, when I run it with OM2007.2 image, I get 'cannot find libmokoui.so.0: no such file or directory' error message Sep 13 19:04:46 but now I'm not a virgin ;) Sep 13 19:04:51 mbuf: different binary format Sep 13 19:05:07 It's a quick grow-up routine Sep 13 19:05:30 cjb, so how do i fix it? Sep 13 19:05:51 recompile? Sep 13 19:06:22 cjb, no, i meant i recompiled the code with OM2007.2 build environment (arm-angstrom-linux-gnueabi), and running it with OM2007.2 images on the phone Sep 13 19:08:07 xaid: i'm colourblind myself, and have colourblind friends. colourblind varies from person to person. Sep 13 19:09:41 yeah, there are tens of colorblindness possibilities. Sep 13 19:10:15 it's because it's a recent mutation, only humans and monkey have the different red/green receptor genes, and because it's so recent there are many ways for the genes to misalign. Sep 13 19:11:25 (I hadn't realized until reading about it that only humans *aren't* colorblind, all other animals (except great apes) are. Sep 13 19:11:27 ) Sep 13 19:15:21 cjb: it depends, some animals are monochromats, some are dichromats, some are trichomats and some are are tetrachomats (some human women might be partly this too) Sep 13 19:16:15 cw666, yeah tetrachomat women rule :-) Sep 13 19:16:24 yes, but only humans are trichromats+. Sep 13 19:17:14 cjb, really? i don't remember it like that Sep 13 19:17:44 http://en.wikipedia.org/wiki/Trichromatic_color_vision Sep 13 19:18:08 "Some species of insects (such as honeybees) are also trichromats" Sep 13 19:18:44 "Tetrachromacy is expected to occur in some birds, fish, amphibians, reptiles, arachnids and insects." Sep 13 19:18:57 that's not RGB, though. Sep 13 19:19:24 the red-green differentiation in gene was recently unique to monkeys, which is why only monkeys and humans have it. Sep 13 19:21:49 ok so in other animals it's something else than RGGB Sep 13 19:23:42 tuukkah: i sometimes wonder about my kid, she draws things in colors that look almost the same to me Sep 13 19:23:47 juri_: you're right. but i think the common type is the red/green one, but then again, there are different degrees so you're right. there wont be a neutral theme that works for all CB people :| Sep 13 19:23:55 it's probably just because she's a girl and like all women can see colors better than men Sep 13 19:24:04 tuukkah: yes, infra-red or ultraviolet Sep 13 19:24:08 is there an equivalent for libmokoui/moko-ui.h (from OM2007.1) that can be used for OM2007.2? Sep 13 19:24:44 i tell people who use words like mauve and turquoise that "im a guy, i only know the seven colors from the rainbow. go get me a beer" Sep 13 19:24:56 and it's six colors really Sep 13 19:25:01 indigo and purple are the same thing Sep 13 19:25:07 violet Sep 13 19:25:09 whatever Sep 13 19:25:41 the OM-2007/examples/openmoko-stylus-demo still usings libmokoui/moko-ui.h in OM2007.2 build; Sep 13 19:26:42 cw666: i'm officially blue/green colourblind, but also have problems in the browns. and i can see slightly outside the normal "red" spectrum. in some "black" paints, i see a ... mix of a different colour. Sep 13 19:27:08 i'm a mutant. :) Sep 13 19:27:25 juri_: so if you weren't CB you'd be tetrachromat? :-) Sep 13 19:27:26 :) Sep 13 19:27:38 gah. damn irc client forgetting my nick Sep 13 19:28:01 i have issues with the green and red. i can't differenciate between the red and orangle/yellow of the traffic lights at nighttime :) Sep 13 19:28:05 m1: something like that. :) Sep 13 19:28:17 *orange Sep 13 19:28:19 xaid: the green lights are white to me. scary. :) Sep 13 19:28:49 juri_: i see them a bit grayish with maybe a hint of green. im not sure if they're supposed to be dark green :D Sep 13 19:28:55 juri_: does that include monochromatic green LED lights? Sep 13 19:28:56 heard that only 6-7 bands of color would be useful Sep 13 19:29:16 cjb: yes, it does. Sep 13 19:29:26 xaid|work: that's why they put them in different boxes :) Sep 13 19:29:27 juri_: yeah, niki likes to draw on white paper with yellow, i can barely see what she's doing unless im really close but she seems to have no problems with it Sep 13 19:29:43 cjb_ie: augh, nick collisipn Sep 13 19:29:49 cjb_ie: get a better nick! Sep 13 19:30:05 cjb: this _is_ my better one :/ Sep 13 19:30:07 that should allow to see in nature exact material for example Sep 13 19:30:16 cjb_ie: /me disagrees Sep 13 19:30:55 i used to be cjb in most of my irc sessions Sep 13 19:30:58 aloril: that's what I find so interesting about red-green distinction being limited to humans Sep 13 19:31:11 there's so much of nature that only we see, and it isn't even meant for us Sep 13 19:31:34 so mokomakefile causes a svn checkout of the entire of svn.openmoko.org but how is this used? Sep 13 19:31:36 cjb: the problem is at night, i can't see the black boxes that they are in so i can't figure out which box is actually lit :D Sep 13 19:31:51 openembedded just goes and checks out what it wants again anyway for the build? Sep 13 19:32:08 cjb: hmm.. yeah, anyway more is possible with suitable goggles (though not at same time, but looking at suitable subset at the time) Sep 13 19:32:21 sandos: ah. in most countries they solve that by yellow = red and yellow on at the same time Sep 13 19:32:22 xaid|work: in sensible countries the boxes are outlined in retroreflective rectangles Sep 13 19:32:24 ie.. a bit like looking at 6-7 dimensional object using 3d viewer Sep 13 19:32:26 (and stores it in the sources as .tar.gz) Sep 13 19:32:29 but for colors Sep 13 19:32:33 s/sandos/xaid|work/ Sep 13 19:32:33 cjb meant: xaid|work: ah. in most countries they solve that by yellow = red and yellow on at the same time Sep 13 19:32:38 cjb: i thought that was only UK and germany Sep 13 19:33:07 (and only for immediately before green) Sep 13 19:33:53 in .ch it is green yellow red (nothing in between) Sep 13 19:34:02 zdanek: did you upload the OABI chroot anywhere? Sep 13 19:34:18 i'm working on that Sep 13 19:34:28 I'm writing an article Sep 13 19:37:27 just uploaded my first home-brew kernel and rootfs, it boots it boots! Sep 13 19:37:46 jpcass: good for you! :-) Sep 13 19:38:17 * Writchie returns from lunch to meet the 2837th borked build Sep 13 19:39:11 cjb: canada doesn't do it in the province that i'm in :| i think quebec has shapes along with the traffic lights Sep 13 19:39:13 is the dialer supposed to auto register with the network, or do i need to intervene on the console? Sep 13 19:39:14 mickey|zzZZzz: I wonder if OpenMoko is one of the invited mentoring organisations? Sep 13 19:40:04 no sign of gsmd with ps :-( Sep 13 19:40:10 jpcass: did you change the boot parameters? did you put gsmd stop/start into matchbox/session? then it should automatically ask for pin Sep 13 19:40:11 apt: that would make it easier to distinguish, especially for flashing red/yellow Sep 13 19:40:39 hhf423: no, where is that? Sep 13 19:42:11 anyway, that is what some future NeoXXXX will do for me, arbitrary color bands Sep 13 19:42:11 i just did /etc/init.d/gsmd stop and the machine has frozen (console and X) Sep 13 19:42:57 jpcass: removed "console=ttySAC0,115200" from bootargs_base Sep 13 19:43:04 re Sep 13 19:43:06 jpcass: in uboot Sep 13 19:43:24 how many plp have wasted more than 10 times as much time on the build system as on doing any real development? Sep 13 19:43:24 jpcass: yes, thats a classic now Sep 13 19:43:39 bug 744 sor something Sep 13 19:44:04 Writchie: count me. Sep 13 19:44:35 i'm not counting testing prebuild images as development Sep 13 19:44:58 Writchie: >10 - have you read this? http://www.openembedded.org/%7Ekoen/usermanual.pdf its very nelpful Sep 13 19:47:26 what, there was a manual? Sep 13 19:47:33 ...i.e. advertise it Sep 13 19:48:37 zdanek: when I do ./gllin -low 5 I get "not found" Sep 13 19:48:48 its not complete but the bits that are there are very good. Sep 13 19:48:53 I wish I knew why the App. devel. crash course isn't working for me Sep 13 19:49:22 hhf423: thanks I will check that out! Sep 13 19:49:31 if I link build/conf/local.conf, bitbake finds my helloworld.bb. If I don't, it doesn't Sep 13 19:49:42 jpcass: why should everyone wanted to development have to debug the bb files for 4000 packages Sep 13 19:50:08 amen brother Sep 13 19:50:44 a core developer makes a two line patch to libgsm and the wheels come off the build ! Sep 13 19:51:15 how do I get bitbake to use the site.conf in local instead of the one in build ? Sep 13 19:51:39 if I'm using mokomakefile ? Sep 13 19:52:04 polz: I think you just put your own there Sep 13 19:52:07 * tonyg . o O ( wow, mtn is cpu-intensive ) Sep 13 19:52:21 mokomakefile writes a default one to start with Sep 13 19:52:21 okay, article ready Sep 13 19:52:25 http://www.openmoko.org.pl/node/55 Sep 13 19:52:33 making gllin work on 2007.2 Sep 13 19:53:01 polz: but mokomakefile does not change it when its defaults change Sep 13 19:53:09 from example for rm_work Sep 13 19:55:51 hi all. can someone tell me what bootargs_base should be? I think I messed it up when I tried to remove the "console"-part. Sep 13 19:57:19 mine (original) reads: rootfstype=jffs2 root=/dev/mtdblock4 console=ttySAC0,115200 console=tty0 loglevel=8 Sep 13 19:57:20 as for people who will use chroot package - please unmount all chroot directories before deleting whole dir structure - or you will delete root fs as I accidently did... Sep 13 19:57:33 whats the syntax for setenv anyone? Sep 13 19:57:39 jpcass: thanks! Sep 13 19:57:54 jpcass: I ll try that. Sep 13 19:59:10 is there available port of MC on OpenMoko? Sep 13 19:59:41 midnight comander? Sep 13 20:00:07 CM yep Sep 13 20:00:10 zdanek: thanks, let me try it out ... for some reason GPS doesn't seem to work inside even near a window so ill prob go for a walk Sep 13 20:00:31 did you run gllin already? Sep 13 20:00:45 zdanek: I tried to search for it in oe, but didn't find it Sep 13 20:00:53 pity Sep 13 20:00:59 zdanek: satscan doesnt work indoors Sep 13 20:01:07 that's the only test code ive tried do far Sep 13 20:01:44 yes, I've heard about this program, where did you get it from? Sep 13 20:02:24 http://members.shaw.ca/mmontour/satscan/ Sep 13 20:02:43 zdanek: mc_4.6.0.bb is there Sep 13 20:03:03 CM sorry, where? Sep 13 20:03:06 zdanek: Using the makefile? Sep 13 20:03:25 nope Sep 13 20:03:46 Ok, then just bitbake mc or? Sep 13 20:03:52 I don't even have dev environment set up Sep 13 20:04:08 Ah, ok Sep 13 20:04:12 Anyway, it's there Sep 13 20:04:14 mc Sep 13 20:04:17 I have no time for that :((( but I'll try to set it up in next weeks Sep 13 20:04:23 oe includes a lot :) Sep 13 20:04:31 it might take that long ;-) Sep 13 20:04:38 jpcass: Hehe Sep 13 20:05:05 that's why I waste my time with articles for people how to use neo Sep 13 20:05:25 to spread the idea of OM Sep 13 20:05:38 but I'd like to write some apps Sep 13 20:06:15 CM is there a way to prepare a mc package for me (and others)? Sep 13 20:06:37 zdanek: Sure, just build the ipkg Sep 13 20:06:42 I'm building right now Sep 13 20:09:06 i also built it for fun ipk is 260K Sep 13 20:12:32 zdanek: you can download it from http://www.johnkerry.plus.com/mc_4.6.1-r2_armv4t.ipk Sep 13 20:12:50 (untested!) Sep 13 20:13:30 jpcass: You beat me to it :-) Sep 13 20:14:08 CM: eager to try out my only recently working build system ;-) Sep 13 20:14:36 still cant dial anywhere though... Sep 13 20:14:40 Me neither Sep 13 20:14:49 I've only ever been able to call 3 times Sep 13 20:15:07 Was with scaredycats image from september 3rd Sep 13 20:15:08 jpcass: thank you Sep 13 20:15:08 about the same for me - using 2007.1 about 3 weeks ago! Sep 13 20:15:19 So with 2007.2 Sep 13 20:15:37 over and over :) Sep 13 20:15:37 2007.1 seemed loads faster - booting and reacting to UI Sep 13 20:16:03 CM & jpcass did you do trick with reloading gsmd in matchbox/session? Sep 13 20:16:05 zdanek: no worries, let us know if it works! Sep 13 20:16:14 ok Sep 13 20:16:26 zdanek: I haven't tried that yet. Been too lazy Sep 13 20:16:36 zdanek: no; i just removed the console from bootargs... what is the reload idea? Sep 13 20:16:42 Hoped that some om guy would have enough time to fix it properly Sep 13 20:17:01 I can dial with this fix Sep 13 20:17:27 how can I force ipkg to install mc from .ipk file? Sep 13 20:17:40 (not from repo) Sep 13 20:17:59 zdanek: Isn't it in any repo? Sep 13 20:19:01 I mean from address that jpcass provided Sep 13 20:19:06 zdanek: i guess you need to download it (onto Sd card?) and run ipkg on the file itself Sep 13 20:20:07 | psplash.c:260: error: 'HAND_IMG_WIDTH' undeclared (first use in this function) Sep 13 20:20:11 mh Sep 13 20:20:14 I did that Sep 13 20:20:35 and it's works because it's downloading additional packages Sep 13 20:20:35 did someone at least compile this before he commitet it? ;) Sep 13 20:20:37 thx Sep 13 20:20:50 WOW! Sep 13 20:20:54 MC works! Sep 13 20:20:56 :D Sep 13 20:21:09 zdanek: Hehe.. Nice :-) Sep 13 20:21:21 thank you guys :D Sep 13 20:23:55 great ;-) Sep 13 20:25:20 zdanek: ok, i tried it out ... it's not working Sep 13 20:26:39 i guess i need a full OABI userspace to poke about Sep 13 20:28:25 what's wrong? Sep 13 20:28:40 I could make a mistake while writing it Sep 13 20:28:51 I've installed it on my neo some time ago Sep 13 20:29:13 zdanek: you need to +x the binaries and the libs Sep 13 20:29:22 i exepcted the tar was +x already but apparently not Sep 13 20:29:46 yes, you're right Sep 13 20:29:59 I broke it while unpacking and packing again Sep 13 20:30:10 I'll try to fix it Sep 13 20:32:44 is the panel brolen at the momment? Sep 13 20:32:56 s/brolen/broken/ Sep 13 20:32:56 borg_ meant: is the panel broken at the momment? Sep 13 20:34:01 anyone know if the gps antenna sensitivity on the neo is typical compared to other devices? Sep 13 20:35:00 I could tell you that is quite good Sep 13 20:35:14 s/could/can Sep 13 20:35:17 s/could/can/ Sep 13 20:35:18 ok, i see only A1 events Sep 13 20:35:18 zdanek meant: s/can/can Sep 13 20:35:25 damn machine ;) Sep 13 20:35:40 open window Sep 13 20:35:43 go to balcony :) Sep 13 20:35:51 wait... 15 minutes Sep 13 20:36:05 it needs some time for first boot Sep 13 20:36:19 so, gps is working, and media player? awesome. Sep 13 20:36:42 juri_: shame we cant make a call ;-) Sep 13 20:36:57 jpcass: i didn't buy my neo to be a phone. :P Sep 13 20:37:04 zdanek: it's really 15 minutes? Sep 13 20:37:48 yes, because it gets satellite states or someting with very low speed Sep 13 20:38:18 now if I turn on GPS on open space it needs max 1 minute to start Sep 13 20:38:38 cm666 I put proper package Sep 13 20:39:11 zdanek: so 15 minutes to sync up if i start it w/o good signal and 1 minute otherwise? Sep 13 20:40:16 no, I meant if you turn on GPS after long period of not using (especially for the first time) Sep 13 20:40:41 look for frames Sep 13 20:40:42 $GPGSA,A,1,05,24,30,,,,,,,,,,2230.0,1000.0,2000.0*00 Sep 13 20:41:00 05,24,30 means that it sees 3 satellites Sep 13 20:41:16 satellites number 05, 24 and 30 Sep 13 20:41:21 and it's listening to them Sep 13 20:42:41 $GPGSA,A,1,,,,,,,,,,,,,89.2,40.0,80.0*0F Sep 13 20:42:44 over and over Sep 13 20:42:53 i'll fiore up bnep and see if i can get details from here Sep 13 20:43:00 i just have to find my bluteooth dongle Sep 13 20:44:35 so you don't see any satellites Sep 13 20:44:52 go to window Sep 13 20:44:54 i do now, only 3 at most Sep 13 20:44:56 or go outside Sep 13 20:44:57 and those go away Sep 13 20:44:59 it's is outside Sep 13 20:45:01 on the balcony Sep 13 20:45:04 but there is one above Sep 13 20:45:07 so maybe ill go for a walk Sep 13 20:45:09 what is the sky? Sep 13 20:45:14 cloudy Sep 13 20:45:18 trees? Sep 13 20:45:28 another balcony at your head Sep 13 20:45:43 blue sky Sep 13 20:45:50 another balcony above yeah Sep 13 20:45:55 should satscan appear in Applications list? Sep 13 20:46:01 zdanek: no Sep 13 20:46:09 cw666... go for a walk :D Sep 13 20:46:28 zdanek: what accuracy can i expect ones it is working? Sep 13 20:46:48 quite good Sep 13 20:46:58 did you see my link to google maps with my gps trace? Sep 13 20:47:05 yeah Sep 13 20:47:22 im asking this because i want to be able to locate the nearst bus-stop to me fairly quickly Sep 13 20:47:33 SpeedEvil: did you get a measurement on how much the GPS draws? Sep 13 20:47:37 you should have map uploaded Sep 13 20:47:50 and stick gps trace to nearest road Sep 13 20:48:13 that what we do at my work while programming gps navigation boxes Sep 13 20:48:32 can the gps chip keep track of all this w/ the daemon? or enough of this to recover posn quickly if we suspend the CPU? Sep 13 20:48:47 Several things in that. Sep 13 20:48:56 A) this all may change in GTA02 Sep 13 20:49:16 B) the GPS has its own very accurate clock that you'd really like to keep running. Sep 13 20:49:34 C) the current software doesn't really support suspend Sep 13 20:49:58 D) in principle, you can do interesting things with rapidly aquiring a position - but again the current software does not. Sep 13 20:50:19 so we're at the mercy of how crappy the closed driver will be? Sep 13 20:50:29 Well. Sep 13 20:50:45 It's not been revealed what the alternate GPS engine is. Sep 13 20:51:12 SpeedEvil: if it's nmea, ntpd has a driver to be stratum-1 with that clock... Sep 13 20:51:24 That's quite irrelevant Sep 13 20:51:30 SpeedEvil: i thought it would be the same chip and an EABI version of the same binary Sep 13 20:51:37 Possibly not Sep 13 20:51:38 i hope it's a library if it has to be closed Sep 13 20:51:51 i *really* dont want to have to deal with some bloated dbus enabled fucko app if possible Sep 13 20:51:51 ...probably overkill, but maybe some folks would get a kick out of having a stratum-1 time server in their phone ;P Sep 13 20:52:14 There is a VCXO that is the accurate clock for the GPS Sep 13 20:52:24 oh, pps? Sep 13 20:52:34 this and the hammerhead chip when idle uses around 15mA, or 45mW Sep 13 20:52:38 no, not PPS Sep 13 20:52:53 it's a fast clock which the psuedoranges are measured against Sep 13 20:53:24 oh Sep 13 20:53:34 http://wiki.openmoko.org/wiki/Hammerhead/Protocol may be of use Sep 13 20:53:38 - for the current solution Sep 13 20:54:05 See Enhancements on that page Sep 13 20:54:38 http://wiki.openmoko.org/wiki/Hardware:AGPS has some information on interesting stuff that the hammerhead can do precisely because it is so dumb Sep 13 20:54:50 SpeedEvil: i hoped i could leave GPS on mostly 24/7 and have a usable device Sep 13 20:54:53 sounds unlikely Sep 13 20:55:11 You can, but 24/1.5-2 maybe Sep 13 20:55:38 If you are willing to take a position every 180s or so, then the power use goes quite low. Sep 13 20:55:46 If 'slow' mode works as advertised. Sep 13 20:56:16 SpeedEvil: well, if i power down the CPU it's still a lot Sep 13 20:56:19 like 2x for GSM only Sep 13 20:56:35 It's not quite that simple :/ Sep 13 20:57:14 SpeedEvil: gsm should be ~20mA or less, GPS that too Sep 13 20:57:20 CPU near zero once there is usable PM Sep 13 20:57:20 If you keep the clock on and the chip hot, that means that you get a nearly trivial position probably up to several minutes later. Sep 13 20:57:34 CPU is not near zero - unless it's off. Sep 13 20:57:48 the powerdown states are not especially low power Sep 13 20:58:03 right, off, which is really suspended, samsung call it something confusing Sep 13 20:58:56 if you powerdown the chip, and the clock, then you need a clock which can provide time to within about a millisecond ideally to get a fast position on resume Sep 13 20:59:12 SpeedEvil: the idea was/is to geep the GPS chip on Sep 13 20:59:19 s/geep/keep/ Sep 13 20:59:19 cw666 meant: SpeedEvil: the idea was/is to keep the GPS chip on Sep 13 20:59:20 cw666: PM is going to suck until we do a linux suspend which requires that all drivers proper support suspend and resume Sep 13 20:59:25 The GPS chip isn't a conventional chip Sep 13 20:59:39 it requires the CPU to do all the sums. Sep 13 20:59:45 It does not spit out NMEA Sep 13 20:59:53 SpeedEvil: right, it's more limited, that's why i asked how much external support was needed to get a fast fix quickly Sep 13 20:59:59 This is good in principle, as it means you can do really interesting things with OS drivers. Sep 13 21:00:06 if the GPS is on, how long after wake the CPU can i get usable/sane results is what im wondering Sep 13 21:00:36 cw666: if everything is done right in all the drivers resume can be < 200ms Sep 13 21:02:03 SpeedEvil: IIRC we've lost time because the counter for the RTC is not accessable Sep 13 21:02:37 SpeedEvil: although maybe we can catch the 1 second tick Sep 13 21:02:37 gah Sep 13 21:02:47 That was what I was assuming. Sep 13 21:03:13 time resolution is still 1/32 ms Sep 13 21:03:27 that's plenty probably Sep 13 21:03:41 but we have to wait /5 seconds average for a tick Sep 13 21:03:58 RTC stops when the CPU is suspended? Sep 13 21:04:09 There are eleventy RTCs in the neo. Sep 13 21:04:10 RTC is IIRC on PM chip now Sep 13 21:04:10 fine, I got connection error and disconnected but my nick still here Sep 13 21:04:14 Well - 3 Sep 13 21:04:39 I'm unsure which ones can actually in principle be used. Sep 13 21:04:49 which one's have the crystal? Sep 13 21:04:58 good question Sep 13 21:05:02 i guess if GPS is on we can get a value from that Sep 13 21:05:13 on resume Sep 13 21:05:16 Probably Sep 13 21:05:23 if it maintains counters Sep 13 21:05:30 would end up being userspace funkies though, might be ugly Sep 13 21:05:43 user space can be running in <200ms Sep 13 21:05:56 And a stub in u-boot much faster Sep 13 21:06:02 much much faster Sep 13 21:06:11 but then your at the bit level Sep 13 21:06:15 yeah Sep 13 21:06:52 i'm feeling better about power now that OLPC has measured 160ms resume Sep 13 21:07:00 of course they are i386 Sep 13 21:07:14 but it should be similar for arm Sep 13 21:07:35 they had problems with USB which were recently solved and I suspect we may have the same Sep 13 21:07:55 Oh - the USB host mode works Sep 13 21:08:02 someone had it going earlier. Sep 13 21:08:19 The patch in SVN plus that patch I posted on the ML Sep 13 21:08:55 SpeedEvil: somebody had USB suspend working? Sep 13 21:10:08 Writchie: OLPC had problems one the USB resume from CPU off Sep 13 21:10:17 s/one/on/ Sep 13 21:10:17 Writchie meant: Writchie: OLPC had problems on the USB resume from CPU off Sep 13 21:10:45 getting confused between cold start and resume IIRC Sep 13 21:11:04 * cjb perks up. Sep 13 21:11:06 oh, my phone doesn't start if I press that power button anymore Sep 13 21:11:08 anyway they got it working fully on their platform Sep 13 21:11:14 unplugged everything, took out battery etc etc Sep 13 21:11:23 waited a few minutes while usb-cable is plugged in, retried Sep 13 21:11:24 etc Sep 13 21:11:27 is it broken? ;) Sep 13 21:11:30 * cjb is right now working on OLPC's unsolved usb resume problems. Sep 13 21:11:42 pvanhoof: few=? Sep 13 21:11:44 I just received it, have started it up twice and other than that it has been lying on my desk Sep 13 21:11:48 it may need to be 30 Sep 13 21:11:52 SpeedEvil, ~5 minutes Sep 13 21:12:00 hi folks, how can I clean all? Sep 13 21:12:02 I got bad news: 160ms resume was without USB. Sep 13 21:12:09 oh ok, then I'll just let it charge a few more minutes Sep 13 21:12:12 with mokomakefile Sep 13 21:12:14 USB is only bluetooth though Sep 13 21:12:17 in the undocked mode Sep 13 21:12:36 of course, that's one of the interesting things Sep 13 21:13:05 so, how do I know its charging? Sep 13 21:13:13 to led? :( Sep 13 21:13:13 you don't Sep 13 21:13:16 s/to/no Sep 13 21:13:19 ok Sep 13 21:14:12 Writchie: I don't know where you get this idea that our problems are fixed from ;-) Sep 13 21:14:39 i think it was your boss's blog ;) Sep 13 21:15:10 ah, that'd do it :) Sep 13 21:15:28 so they are fixed - but what's the real story? Sep 13 21:15:36 they aren't fixed Sep 13 21:15:44 the tests jg reported on were booted nousb Sep 13 21:15:55 other than usb its working? Sep 13 21:16:04 but the problems we're having (probably) aren't kernel-specific Sep 13 21:16:25 i'm sure we will have a different set of problems with our hardware Sep 13 21:16:44 At least we have JTAG Sep 13 21:16:45 but it it's promising to know there that the kernel framework is close Sep 13 21:16:50 SpeedEvil: so do we Sep 13 21:16:57 I know Sep 13 21:16:59 yeah, the kernel's behaving well, I think Sep 13 21:17:08 I'm just saying - without JTAG, it could be painful Sep 13 21:17:25 did you do every driver twice? Sep 13 21:17:38 hm? Sep 13 21:18:22 cjb: we do have a potential out of band solution for USB but only on GTA01 Sep 13 21:18:45 what is being done out of band? Sep 13 21:19:02 we can fiddle the BT chip into standby and then resume Sep 13 21:19:21 it will work if the USB state on the SoC works to specs Sep 13 21:19:28 which is probably unlikely Sep 13 21:19:38 OHCI specs Sep 13 21:19:42 how does usb work at all while you're in suspend? isn't the host controller powered down? Sep 13 21:20:07 yes it's powered down in suspend. Sep 13 21:20:36 the bluetooth ahs a seperate PSU Sep 13 21:20:41 there is no USB power Sep 13 21:20:49 so you're letting a separate bluetooth chip run during suspend, and then there's some wire that it asserts to wake up the EC when it wants to, and then you just reenumerate usb as normal when the kernel wakes up? Sep 13 21:21:01 Alas, no Sep 13 21:21:02 yup Sep 13 21:21:05 there is no wire Sep 13 21:21:10 at least in this version Sep 13 21:21:17 we have to poll it Sep 13 21:21:24 oh, the EC polls it? Sep 13 21:21:31 No, the main CPU Sep 13 21:21:41 ... Sep 13 21:21:45 which is in suspend? :) Sep 13 21:21:47 we can do it the Steppingstone trick if we have to Sep 13 21:21:47 so confused. Sep 13 21:21:48 there needs to be nastiness with an early checker in u-boot probably Sep 13 21:22:07 turn on every n seconds on RTC Sep 13 21:22:13 whoa Sep 13 21:22:16 poll the USB, turn off - in u-boot Sep 13 21:22:17 that is insane Sep 13 21:22:20 there is no EC Sep 13 21:22:22 but we don't bring up the whole kernel Sep 13 21:22:28 oh, in uboot Sep 13 21:22:29 just take a peek Sep 13 21:22:31 wow Sep 13 21:22:44 how's the battery life doing? :) Sep 13 21:22:45 welcome to workaround world Sep 13 21:22:46 Many of us would like an EC Sep 13 21:22:55 cjb: not brilliant Sep 13 21:23:03 most of stuff is not implemented Sep 13 21:23:06 http://www.digg.com/linux_unix/Openmoko_Phone_linux_based_New_UI_Screenshots Sep 13 21:23:34 Lowest power I've gotten it at is about 300mW Sep 13 21:23:36 at 100Mhz Sep 13 21:23:40 and undervolted Sep 13 21:23:45 is don't see us being 16 hour usable until we can suspend Sep 13 21:23:50 and of course resume Sep 13 21:24:09 Nobodies gotten it running in the mythical 'slow' mode Sep 13 21:24:14 12Mhz Sep 13 21:24:21 it sounds like it doesnt save much power anyhow Sep 13 21:24:30 what did you want to wake from BT for, anyway? what sort of BT device? Sep 13 21:24:32 It goes down to about 50mW or so Sep 13 21:24:38 from what? Sep 13 21:24:43 for example, wake on press of hold button on headset Sep 13 21:24:46 press button on headset to like may a call Sep 13 21:24:51 ratehr than digging it out of pocket Sep 13 21:24:58 or purse Sep 13 21:25:32 What we really need is OPPC. :) Sep 13 21:25:44 gsmd really hates me and is being a little bitch today Sep 13 21:26:23 * Writchie dropped the bitch Sep 13 21:27:22 cjb: what are you using USB for in OLPC Sep 13 21:29:12 Writchie: our wireless chip presents over USB. Sep 13 21:29:13 SpeedEvil: the steppingstone hack doesn't have to be all that ugly Sep 13 21:29:20 and is always running, even while the main CPU is powered down. Sep 13 21:29:36 cjb: does it have wakeup from Lan Sep 13 21:29:40 or WLAN Sep 13 21:29:41 yes Sep 13 21:29:52 which chip Sep 13 21:29:55 Writchie: in principle, no. It's annoying to have to develop it for the want of a wire though Sep 13 21:29:59 wireless->ed Sep 13 21:30:00 ec Sep 13 21:30:13 no polling :) Sep 13 21:30:42 cjb: cool which we had something like that Sep 13 21:30:43 EC would enable stuff like wake on signature shake Sep 13 21:30:50 if connected to the accel Sep 13 21:30:55 or .. Sep 13 21:31:02 Writchie: sounds like gta01 has some design problems :) Sep 13 21:31:34 well, at least the one Sep 13 21:31:41 a few - but major unplanned additions are causing the problems. Sep 13 21:31:59 some things not well thought and designed reviewed Sep 13 21:32:23 The bluetooth problem was as the primary designer hadn't really used bluetooth Sep 13 21:32:38 so diddn't think that wake on bluetooth was interesting Sep 13 21:32:43 which would have been caught instantly at design review by anyone from the mobile space Sep 13 21:34:13 with lates build manual GSM connection doesn't work Sep 13 21:34:17 cu hangs Sep 13 21:34:33 even with reenabling crtscts Sep 13 21:34:51 zdanek2: i think cesarb has had similar problems for on OM2007.2 for a couple of weeks Sep 13 21:35:01 I can dial Sep 13 21:35:19 but I can't run AT commands manually, aaargh Sep 13 21:35:30 I'm trying to run damn GPRS for weeks Sep 13 21:35:33 and nothing Sep 13 21:36:15 you need to stay on original 2007.1 to play with gsm Sep 13 21:36:32 serious regressions in 2007.2 Sep 13 21:36:44 with no end in sight Sep 13 21:37:10 Writchie: gprs works for me in 2007.2 Sep 13 21:37:16 Writchie: at least did last i tried it Sep 13 21:37:42 cw666: from an image or a build Sep 13 21:37:45 cw666 can you check gprs for me? Sep 13 21:37:59 and if it works - send me your scripts? Sep 13 21:38:48 Writchie: from a build Sep 13 21:38:57 how recent? Sep 13 21:39:10 Writchie: a couple of days, not tested on a recent build Sep 13 21:39:48 maybe there is a window where it worked or maybe its even been fixed. Sep 13 21:40:03 latest libgsm changes have borked the build Sep 13 21:40:15 Writchie: i finally got a build this morning Sep 13 21:40:25 i think given how crappy/slow OE im about to give up on it Sep 13 21:40:30 and move to makefiles Sep 13 21:40:34 screw pretty packages for now Sep 13 21:40:39 cw666: me too Sep 13 21:40:41 SpeedEvil, it starts! Sep 13 21:40:46 I guess indeed the battery was too down Sep 13 21:40:52 pvanhoof: :) Sep 13 21:40:57 Writchie: ive done makefiles and crossbuilds for years w/o anywhere near the pain/slowness of this Sep 13 21:41:01 it Sep 13 21:41:21 cw666: check your gprs for me Sep 13 21:41:23 OE is stagerringly retarded for my purposes, i want a semi-hackable phone, not gnome-bloat-in-my-pocket Sep 13 21:41:33 zdanek2: i just tried, i can't find the scripts right now Sep 13 21:41:39 damn Sep 13 21:41:39 cw666: buildroot is started to look good Sep 13 21:41:48 Writchie: buildroot? Sep 13 21:41:58 Writchie: i dont follow OE/OM mailing lists or irc much Sep 13 21:42:18 buildroot = older embedded toolchain generator Sep 13 21:42:48 Writchie: fwiw, ive done several crossbuild envs in the past (it's actually a part of my job that im trying to get a replacement person for) Sep 13 21:42:57 cw666: OE is probably a good thing - but not the way its being managed at present Sep 13 21:43:00 so im certainly not against doing one for the neo hacking Sep 13 21:43:05 Writchie: OE is not a good thing Sep 13 21:43:13 Writchie: it's an insanely large amount of shitty bloat Sep 13 21:43:27 cw666: welllll yeah Sep 13 21:43:45 Writchie: really, i want something a lot more minimal, faster to boot and more responsive Sep 13 21:43:59 all the OE/OM people can mock me that i dont have all their games, i dont really care Sep 13 21:44:31 like i said, i want a hackable phone, not 27 different gnome games in my pocket Sep 13 21:44:34 for practical purposes OE=OM Sep 13 21:44:59 cw666: I agree that the focus should be on building from the bottom up. Sep 13 21:45:03 and not from the top down Sep 13 21:45:07 i dont care much for either, ill steal the bits i can use (the UIs aren't bad, those prob worth stealing) Sep 13 21:45:32 Writchie: well, that's all i can do, because in RL im a kernel guy and do some low-level userspace Sep 13 21:45:32 Here's a build question, probably easy.. What does the -liberty flag do? I'm getting an error, do I need to install something? Sep 13 21:45:35 i can't do GUIs Sep 13 21:45:41 i suck do badly at that stuff Sep 13 21:46:00 Casten: it means to link with libiberty Sep 13 21:46:21 is that a package? Sep 13 21:46:22 which is a semi-generic GNU library Sep 13 21:46:34 http://gcc.gnu.org/onlinedocs/libiberty/ Sep 13 21:47:22 Is anyone around who has commit access to OE? Sep 13 21:47:53 My Free42 recipe (RPN Calculator) has been languishing in the bugtracker for a week-plus: http://bugs.openembedded.org/show_bug.cgi?id=2920 Sep 13 21:47:55 Writchie: if you really are interested in an alternative (ie. minimalist as hell and very sucky to start with) ill probably have something in a week or so Sep 13 21:48:09 cw666: I am interested. Sep 13 21:48:17 quicksand: ironically, free42 is one of the few apps i want from the OE/OM stuff Sep 13 21:48:27 And it's such a nice thing to have on my Neo that it would be nice to get it in the database. Sep 13 21:48:31 basic cross development for kernel + basic apps at shell level Sep 13 21:48:37 x optional Sep 13 21:48:42 Writchie: it's 99% going to be a debian chroot (64 bit initially) with crosstools and some makefile goop Sep 13 21:48:52 Writchie: focus will be kernel and libc stuff first Sep 13 21:49:04 Writchie: x and everything else ill need help with, to steal from the OE/OM tree Sep 13 21:49:13 oh, and ill likely not change stuff that often once it's working Sep 13 21:49:25 im a bit of a bitch like that, unless gcc is broken, i wont mess with it Sep 13 21:49:33 need BT and d-buss Sep 13 21:49:33 cw666: Hmmm, the page says to put libiberty in your sources.. so which particular place should I put it? Sep 13 21:49:34 * Writchie ducks Sep 13 21:50:21 Writchie: neither of those are that horrnedous though Sep 13 21:50:33 very true Sep 13 21:50:39 Writchie: dbus and BT are both fairly small sane c-ish things Sep 13 21:50:44 the bloat is added above Sep 13 21:51:02 e-calendar-factory is FITH Sep 13 21:51:04 quicksand: mind if I rename the .bb file to free42-vga ? Sep 13 21:51:13 gconfd-2 too Sep 13 21:51:18 root 1703 0.7 8.1 21332 10384 pts/1 S 19:40 0:06 openmoko-today Sep 13 21:51:24 rwhitby: just adding the hyphen? fine with me. Sep 13 21:51:28 how much of that resident size is unique to that process i have no idea Sep 13 21:51:50 quicksand: testing a build now Sep 13 21:52:04 * quicksand crosses fingers Sep 13 21:52:06 anyhow, run about w/ GPS going time to see what happens Sep 13 21:52:09 I presume it's been well tested on the neo, right? Sep 13 21:52:25 I'm using the .tar.bz file from the bugtracker Sep 13 21:52:42 Well, on my Neo at least! Yeah, it runs fine, no errors, no wildly crazy computation results. ;) Sep 13 21:53:43 cm666: any chances for GPRS scripts anyway? Sep 13 21:54:23 cw666: I see that my libiberty is incompatible, and now I notice that I am not using the preferred version of glibc, I guess I should switch to 2.5 Sep 13 21:55:11 quicksand: why is the resulting executable called free42dec, not just free42? Sep 13 21:55:15 Casten: honestly, i never looked into the internals of OE/OM and failures as it irritates me Sep 13 21:55:28 Casten: i know a build from this morning did complete though Sep 13 21:55:32 * cw666 wanders off Sep 13 21:55:55 zdanek: have you looked -> http://wiki.openmoko.org/wiki/Manually_using_GPRS Sep 13 21:56:02 could someone please point me to an explanation of where exactly bitbake looks for .bb files ? Sep 13 21:56:13 The original package -- from source -- builds free42bin and free42 dec. *bin uses binary math, while *dec uses decimal math, which while slower, is more accurate and more consistent with the original calculator. Sep 13 21:56:19 polz: i think the short answer is everywhere Sep 13 21:56:19 especially if it's invoked by mokomakefile Sep 13 21:56:50 One of my patches causes the *bin version not to be built. I didn't think we needed both. Sep 13 21:56:57 Writchie: well, I've tried doing what the application development crash course in the wiki told me to do... Sep 13 21:56:58 and it doesn't seem to be working Sep 13 21:56:59 cw666: Maybe once I am done I can add my notes to the build requirements. So far glibc 2.5 and zlib were two surprises. Sep 13 21:57:50 unless I replace build/conf/site.conf with the one from my local "overlay", it doesn't find the helloworld.bb Sep 13 21:57:58 which is also in local Sep 13 21:58:00 polz: sorry I can't help, I spend past couple of hours trying to follow dependencies for to debug the present borked build Sep 13 21:58:13 polz: have you read the bitbake and OE manuals? Sep 13 21:58:45 rwhitby: do I have to ? Those docs are huge :/ Sep 13 21:59:12 My philosophy is if you are going to use a tool, then you should read it's documentation cover to cover at least once. Sep 13 21:59:48 Otherwise you are not respecting the author of the tool and the work they put into documenting it. Sep 13 21:59:48 cw666: If openmoko folks would like to widen adoption, an idea might be to have a premade Ubuntu distro with everything building. something like that would ahve and I'm sure will save me countless hours trying to get a vanilla system set up for building. Sep 13 21:59:56 you should, but you shouldn't have to... Sep 13 21:59:57 writchie: I did and it never worked Sep 13 22:00:30 rwhitby: I'd much rather start writing useful little apps than learn about the internals + the local configuration of something as flexible as bitbake Sep 13 22:01:13 polz: well, unfortunately, the option of not using bitbake/OE is not yet offered by OpenMoko, so that's not an option on the table, right? Sep 13 22:01:34 I agree with you that application developers shouldn't be using bitbake/OE/MokoMakefile at all. Sep 13 22:02:11 rwhitby: I guess you're right. I'll waste a few more afternoons and read through the docs... Sep 13 22:02:18 polz: I agree with rwhitby on reading docs thoroughly. But just this once: http://bitbake.berlios.de/manual/ch04s02.html#id870544 Sep 13 22:02:23 OE is a distro build tool, not an application SDK. Sep 13 22:02:57 rwhitby: but it is not building a distribution Sep 13 22:02:59 although I'm pretty sure I'll have to also read half of the configuration files provided by openmoko... and the makefile which is supposed to make the process of starting development easier... Sep 13 22:03:23 rwhitby: and openmoko recipes are now in OE Sep 13 22:03:26 OE=OM Sep 13 22:03:37 polz: yep, I'm afraid that's the case. Sep 13 22:03:54 Writchie: OE>>OM imho Sep 13 22:04:13 * rwhitby adds free42 to OE Sep 13 22:04:32 rwhitby: I don't mean to be disrespectful towards the smart people who built the distribution, but if I have to study the system in detail just to use it, it's crap. Sep 13 22:04:35 * Writchie which does what to my build in progress Sep 13 22:04:43 \me offers rwhitby a beer or beverage of choice if he happens to visit that San Francisco area. Sep 13 22:05:00 rwhitby: i have seen a screenshot of free42, looks nice Sep 13 22:05:08 is there an ipk somewhere already? Sep 13 22:05:11 polz: I expect the people who built the distribution will agree with you. Sep 13 22:05:22 the barrier to entry should be as low as possible... This whole thing reminds me of Zope + Plone :( Sep 13 22:05:51 polz: fully agree Sep 13 22:06:06 polz: The documentation is sorely lacking, to be sure. The free42 package I have been discussing with rwhitby just now is the result of hours and hours of reading .bb and .bbclass files from the OE tree. Sep 13 22:06:12 AND the manuals. :) Sep 13 22:06:16 rwhitby: ehh, well, at least it's something. This is still way better than having to build gcc-cross on my own. Sep 13 22:06:19 there should be a whole completely different SDK for application developers Sep 13 22:06:30 polz: i'm not so sure about that Sep 13 22:07:04 rwhitby: well i don't care about the apps and i've been blocked for days at a time. Sep 13 22:08:00 A seperate SDK for app developers vs. system developers? Sep 13 22:08:32 Casten: yes - usually the SDK is done *after* the system has been built, by a completely different development group. Sep 13 22:09:04 rwhitby: I beleive simple instructions on how to set up a working overlay with some additional packages would suffice Sep 13 22:10:18 polz: rwhitby: what good is documentation if there is no working frozen build to start with?? Sep 13 22:10:22 rwhitby: something like "run 'make local' to create an overlay", plus some example packages - one with autoconf, one with qmake and one with a custom makefile Sep 13 22:10:33 polz: ok, now you just have to bribe quicksand to write that set of simple instructions Sep 13 22:10:51 Writchie: I'm working on the frozen bit. Sep 13 22:10:57 hehe Sep 13 22:11:13 i though that was done - sorry? Sep 13 22:11:13 But mickey|zzZZzz is still zzZZzz, so I haven't been able to harrass and convince him yet. Sep 13 22:11:22 rwhitby: I think someone has already attempted that - there are some instructions in the wiki Sep 13 22:11:24 openmoko-* packages are still floating Sep 13 22:11:33 unfortunately, they're not working for me Sep 13 22:11:53 rwhitby: i assume that's why depencies are broken Sep 13 22:12:23 rwhitby: How did you know that I take bribes? I will write something when I understand it well enough to do so. That's why I'm here. Sep 13 22:12:42 rhtitby: Well, it seems this project is a little unusual. If the system builds, I don't see why an app developer would mind, once they got past the first build. I feel a few basic samples and API's should suffice, assuming one can get the system running. Sep 13 22:13:20 quicksand: pushing free42 now Sep 13 22:13:30 it should be in the feeds next time mickey|zzZZzz runs the builder Sep 13 22:13:46 fsck... I have to go to bed... another day GONE without a simple app. ARGH! Sep 13 22:13:52 rwhitby: your work is very much appreciated ;) Sep 13 22:14:07 polz: Right now the instructions in the wiki are somewhat like a magic incantation. As in: "Do this -- it works -- nobody knows why. Nobody who has lived to tell about it, at least. Mwahahahaha!" Sep 13 22:14:26 Awesome. Thanks rwhitby. Sep 13 22:14:45 thx for packaging it - I'll be a user myself. Sep 13 22:15:24 My pleasure. As my page on the wiki says, I've been using it for a long time. First on a PalmOS device, then on Windows Mobile. Sep 13 22:15:28 atm the moment i think this project is seriously loosing out on dev help from the community due to the f*ck u read the bb docs mentality of the core developers Sep 13 22:16:14 Writchie: To be fair, the core guys have a lot of things to deal with right now, and documentation/tutorials haven't been high on the list. Sep 13 22:16:15 Writchie: I think it's simpler a matter of available time, rather than an attitude, for most of the people you are referring to. Sep 13 22:16:27 I thought the core developers were working on the hardware and drivers... Sep 13 22:16:38 i'm not talking about documentation - i'm talking about a working tool chain and build Sep 13 22:16:39 I doubt there is anyone who's job is to write documentation for OM right now. Sep 13 22:16:57 and just crossing their fingers that perhaps some people will be able to make sense of the build system and write an app or two Sep 13 22:17:05 for example, what is jserv using what is harald using ? Sep 13 22:17:10 and a big phone manufacturer would have a team of 5 doing the doco for the SDK. Sep 13 22:17:30 rwhitby: well they'd at least have 1 manager Sep 13 22:17:33 I like OE, quite a bit, from what little I know about it so far. But unfortunately activity on the .dev branch is really high, and sometimes things get broken. Sep 13 22:17:56 And there's not much OM can do about it, except deal with the consequences until it gets fixed. Sep 13 22:18:10 going to sleep Sep 13 22:18:13 goodnight Sep 13 22:18:17 but OM - OE are the same people Sep 13 22:18:19 quicksand: that should change now that SRCREV support is in, as long as I can convince mickey|zzZZzz to freeze the OM packages too Sep 13 22:18:29 'nite zdanek Sep 13 22:18:50 Writchie: no they are not - there are *many* more OE people than OM people, and most of those OE people don't build the openmoko distro. Sep 13 22:19:16 rwhitby: fair enough -- there has been improvement already, actually, now that most things aren't getting rebuilt every day. Sep 13 22:19:17 Writchie: I myself do all development using locally-built libgsmd/libmokoui only and don't have any troubles with that setup at all - I just use OE for cross-compiling and testing on the actual device. Sep 13 22:19:19 I see - they just break the OM distro Sep 13 22:19:39 Writchie: yep, and sometimes OM breaks their distro. Sep 13 22:20:12 and sometimes core bitbake or OE changes break all distros Sep 13 22:20:13 Writchie: I was an Angstrom distribution user before OM, and that's OE too. And the NSLU2 stuff, etc. Sep 13 22:20:14 abraxa_: does the dialer actually work for you ? Sep 13 22:20:36 polz: I'm not working on those apps Sep 13 22:20:50 My point is that you can do application development without OE, too Sep 13 22:20:51 quicksand i think this one will get that distro stability to get better http://www.openembedded.org/openmoko-hires-developer-devoted-to-openembedded Sep 13 22:21:56 abraxa_: if there is a recipe for doing what you are doing that can be automated, I'd be happy to add it to MokoMakefile Sep 13 22:22:02 roh: Darn, I thought you might have linked to the actual announcement! ;) I'm waiting on that name. And yes, I'm sure that will help a great deal. Sep 13 22:22:53 cworth: btw, have you been successful in mtn->git conversion? Sep 13 22:24:21 and now.. packing up.. bbl.. need to repair my bike tonight Sep 13 22:24:51 rwhitby: My way is kinda hack-ish since I'm overwriting the sources OE uses to compile my .bb package so I wouldn't recommend it for widespread usage. As there however already are non-hackish ways (e.g. local overlays) I think others are better off adopting these Sep 13 22:25:06 so with an alsa audio device on the phone -- is it possible to playback music on a phone conversation, while being able to talk? Sep 13 22:25:44 I mean, that's some pretty collaborative snazzy stuff right there -- or popping a link to the person at the other end of the convo, so that they can get grab the same music you are listening too Sep 13 22:27:06 vallor: I believe so -- but the block diagram on the wiki makes my brain spill out my ears: http://wiki.openmoko.org/wiki/Neo1973_Audio_Subsystem Sep 13 22:27:19 maybe I've had too much caffeine today, but it seems like a way to differentiate the product Sep 13 22:28:29 I'm envisioning what an ad would look like, showing two people grabbing an mp3 in seconds, and playing it back together, on a phone convo... tagline: "You know what you want.(tm)" Sep 13 22:28:46 I dunno, just spitballing Sep 13 22:28:53 Pushing illegal things may cause problems in some markets. Sep 13 22:29:15 doesn't have to be illegal, if they're both licensed for the music Sep 13 22:29:16 SpeedEvil: Nice choice of words. Sep 13 22:29:32 hmm... Rhapsody, maybe? Sep 13 22:29:37 vallor: That would be cool, but I think most people around here would be pleased with a phone that actually rings and can be used as a voice communication device, first. Sep 13 22:29:39 doing anything may cause legal problems in some markets Sep 13 22:29:45 or not doing something Sep 13 22:31:02 quicksand: i agree with you Sep 13 22:31:19 :) Sep 13 22:31:34 quicksand: that's a good point, but it's going to do that eventually -- btw, I tried yesterday to build with Mokomakefile, still falls apart with the binutils -- haven't looked at gsmd since day before yesterday, but it does seem to be well put-together, it's just hard (for me) to debug because its all about callbacks, pretty much... Sep 13 22:32:13 vallor: Does this help re binutils: http://bugs.openembedded.org/show_bug.cgi?id=2928 Sep 13 22:32:14 on the other hand, I haven't run gsmd with gdb yet Sep 13 22:32:18 * vallor looks Sep 13 22:32:39 There's a workaround patch there, and I used it myself successfully. Sep 13 22:32:52 quicksand: !!!!! Sep 13 22:32:54 sweet! Sep 13 22:33:32 Thank cesarb for it, not me. We all benefit. Sep 13 22:33:38 :) Sep 13 22:33:48 thank you, cesarb :) Sep 13 22:33:59 * rwhitby built from scratch two days ago, and hasn't had any build problems since Sep 13 22:35:05 * quicksand has been having good luck lately too. Sep 13 22:35:56 * Writchie built ok 3 days ago and downhill from there Sep 13 22:36:10 simple update = 3 days of pain Sep 13 22:36:36 * hrw built yesterday Sep 13 22:37:49 rwhitby: am I wrong is or a successfull build just a roll of the dice as to what changes you hit Sep 13 22:38:31 Writchie: it used to be - now with SRCREV it should be much more predictable Sep 13 22:38:50 (since the last two days or so) Sep 13 22:38:54 but recipes are still using SRCDATE Sep 13 22:39:06 Writchie: fewer and fewer Sep 13 22:39:33 there should be very few DATE based ones, and they all should be moved to SRCREV, so point them out for people to fix. Perhaps an email to the list. Sep 13 22:40:26 * rwhitby goes to work Sep 13 22:42:19 congrats on getting on digg again Sep 13 22:42:20 :) Sep 13 22:43:07 quicksand: it does seem to be building (heavily knocking on wood...) Sep 13 22:43:44 rwhitby: quicksand: task-openmoko-phone seems pretty basic and its breaking libgsm Sep 13 22:45:23 * cesarb notices the Konversation icon blinking Sep 13 22:45:41 cesarb: yes, I thanked you for your x86_64 patch Sep 13 22:45:44 Come on, they didn't fix that yet? Sep 13 22:45:46 cesarb: thank you :) Sep 13 22:46:12 cesarb: if they did, it within approximately the last 24 hours Sep 13 22:46:31 * vallor tried the mokomakefile yesterday, but not today... Sep 13 22:54:33 vallor: oh, btw, if you want to debug gsmd, you could try the two debug patches I attached to ombug 766, which add some extra strategic debug output Sep 13 22:54:46 ah, thank you sir Sep 13 22:55:32 vallor: it's not much, but the one which prints exactly what was sent to the modem (and when) should be the most useful Sep 13 22:56:26 vallor: the other one should never print, else you got unlucky and hit missing code for an uncommon case (which I was too lazy to create a patch to actually fix) Sep 13 22:56:40 cesarb: did you ever get cu to work? Sep 13 22:56:48 Writchie|dinner: nope Sep 13 22:57:03 someone reported similar problems earlier Sep 13 22:57:07 Writchie|dinner: only gsmd's passthrough, which ended up being enough for everything I wanted Sep 13 22:57:38 when I get a working om2007.2 i'll do a side by side with om2007.1 Sep 13 22:57:38 cesarb: this is the first "real world" stress test for this new machine I put together -- I've got the mokomakefile building (with 4 cores), and also a kernel (-j2 so it won't mess with the mmf as much as it might) Sep 13 22:58:21 vallor: watching the heat output in lm-sensors? ;-) Sep 13 22:58:22 oh, to get cu to work, there's a trick with stty Sep 13 22:58:33 cesarb: yeppers, hehe -- using gkrellm Sep 13 22:59:00 Mine, for some absurdly random reason, increases the temperature for the _motherboard_, but not the CPU, when I'm compiling OE (such that the MB temp ends up higher than the CPU temp!) Sep 13 22:59:33 the individual cores themselves get hot, but cpu temp is nailed at 50C Sep 13 22:59:36 http://ponzo.net/newshawk/hot_mama.jpg Sep 13 22:59:40 I'm guessing either the sensor is near the CPU power stuff on the motherboard, or else it's either miswired or misconfigured Sep 13 23:00:13 could be -- who is the manufacturer? Sep 13 23:00:20 (for the motherboard) Sep 13 23:00:46 anyone seen this? http://roughlydrafted.com/RD/TechQ307/Entries/2007/8/23_Apple_iPhone_vs_the_FIC_Neo1973_OpenMoko_Linux_Smartphone.html Sep 13 23:00:50 It's an ASUS M2V Sep 13 23:00:58 "The phone wasn't custom designed for Linux. It is a Windows Mobile unit codeveloped by the Chinese government as a mass produced people's phone." Sep 13 23:01:21 * vallor sharpens his knives... Sep 13 23:01:26 indeed Sep 13 23:01:38 (btw, lm-sensors is IMO a kludge... something like IPMI is WAY better) Sep 13 23:01:57 * cesarb wonders if he'll ever see IPMI on a non-server Sep 13 23:02:04 that post is pretty damning Sep 13 23:02:34 doc|work: well, it's a free country -- neo1973 folks have the satisfaction of being correct :) Sep 13 23:02:43 and that advantage, too Sep 13 23:03:30 doc|work: saying it wasn't custom designed for linux is a half-lie Sep 13 23:03:58 doc|work: didn't harald say the GSM chip was chosen for having the most standard 07.10/07.10/xx.yy implementation? Sep 13 23:04:41 doc|work: not to forget the BT chip (chosen for its use of hci-usb...) Sep 13 23:04:50 cesarb: sure, GSM and GPS are closed source. the rest is from scratch using off the shelf stuff. It's definitely not a windows phonw Sep 13 23:04:57 hmmm, BT support closed too? Sep 13 23:05:52 doc|work: BT is the phone's third processor (yeah, three processors... AP, GSM, and BT) Sep 13 23:06:03 doc|work: it runs closed-source bluecore stuff Sep 13 23:06:22 doc|work: which afaik is a very common BT stack Sep 13 23:06:23 cesarb: ok, but I mean "codeveloped by the chinese government"? :D Sep 13 23:07:04 ''When you buy a Neo1973, you pay $450 for hardware from FIC, and have no commitment to anyone but the GSM provider “of your choice.” In the US, that means AT&T. The only other US GSM provider is T-Mobile, which doesn't provide GSM coverage on standard frequencies. T-Mobile may possibly be able to resell AT&T GSM service to you on slightly different terms, but that's the extent of your real “freedom” in terms of service.'' Sep 13 23:07:07 doc|work: I have no idea where the chinese government stuff came from... Smells like it was pulled out of thin air, to put it mildly Sep 13 23:07:18 yeah Sep 13 23:07:30 vallor: ah, that's the (very annoying IMO) USian attitude of "all the world is the US" Sep 13 23:07:42 doc|work: do they mean the taiwanese "chinese" government? Sep 13 23:08:03 vallor: face it, most of the world uses GSM... that particular paragraph shows the author's, to put it mildly, biad Sep 13 23:08:06 s/biad/bias Sep 13 23:08:12 vallor: are the taiwanese (who some call china) government involved in FIC? Sep 13 23:08:16 the tone of the piece is ridiculous Sep 13 23:08:17 there are only 2 providers offering gsm sevices in USA? Sep 13 23:08:26 its actually AT&T that uses the non-standard GSM frequencies which where added to GSM for them. Sep 13 23:08:42 cesarb: "FIC is really nothing to brag about. The company is a knockoff hardware cloner infatuated with Microsoft." it's the fucking pc industry. everyone's a hardware cloner. Sep 13 23:08:44 Writchie|dinner: wait... you contradicted yourself Sep 13 23:08:56 Writchie|dinner: if they added them to the GSM standard, they cannot be nonstandard! Sep 13 23:09:10 "The “free and open Anti-iPhone” rhetoric surrounding the FIC phone is therefore marketing drivel to sell a Chinese Windows Mobile device as a hobbyist kit for phone hackers." Sep 13 23:09:14 they are now standard Sep 13 23:09:15 I smell an iphone fanboi Sep 13 23:09:38 oh brother Sep 13 23:09:38 * cesarb finds this "anti-iphone" talk to be an annoying diversion Sep 13 23:10:08 cesarb: just like gsm 450 Sep 13 23:10:24 doc|work: I guess it's also nice that I've never heard of "roughlydrafted.com" -- nor will I ever hear of them again, if that's the quality of their "research".... Sep 13 23:10:43 Writchie|dinner: they were used before being added to the GSM standard? Sep 13 23:11:03 vallor: I think I saw that site once before on slashdot Sep 13 23:11:30 cesarb: the bands were used for AMPS and US style TDMA Sep 13 23:11:43 cesarb: very true, the false dichotomy draws me in -- I'm not much of an iphone fan -- but ultimately, it can be stated that roughdraft is "wrong." Sep 13 23:11:46 vallor: with any luck Sep 13 23:11:52 vallor: http://slashdot.org/tags/roughlydrafted Sep 13 23:11:59 and leave it at that -- unless someone wants to flame 'em Sep 13 23:12:14 Writchie|dinner: yeah, but were they used for GSM before being added to the standard? Sep 13 23:12:16 er, "correct" them ;P Sep 13 23:12:40 it's stupid that they're comparing beta stage hardware (missing some of the stuff he's even complaining about - wifi) with alpha software, which hasn't been released, against the iphone which had all of apple's resources behind it Sep 13 23:12:40 i don't think so. Sep 13 23:12:41 is there a way to ask the phone what frequency, channel, etc it's on? Sep 13 23:13:00 dual-band used to mean 900/1800 and tri-band added 1900 Sep 13 23:13:13 1900 was added for the US IIRC Sep 13 23:13:27 I wouldn't be surprised if roughlydrafted worked for apple. Sep 13 23:13:36 GSM folds tried for years to get the US carriers to adopt GSM Sep 13 23:13:39 vallor: the problem with focusing on being the anti-iphone is that you lose time chasing after the "competitor"'s strenghts, instead of focusing on your own strong points Sep 13 23:13:45 but they didn' Sep 13 23:13:50 like the idea of an open phone Sep 13 23:14:21 cesarb: that's a good point Sep 13 23:14:34 eventually, AT&T wireless decided to go along with GSM as well as Cingular Sep 13 23:14:38 screw the iphone, I want a decent, open, phone that has wifi and a browser, at that stage, everything else is a bonus Sep 13 23:14:54 the alternative was CDMA 2000 Sep 13 23:15:32 I want a phone (in fact, a gadget, but my phone already uses one gadget "slot") which I can program to do what I want Sep 13 23:15:38 I want to be able to go on my own crazy tangents Sep 13 23:15:49 ditto Sep 13 23:16:13 Without having to put up with random limitations of whatever OS is there, or with the poorer stability of a reverse-engineering-based linux port Sep 13 23:16:14 cesarb: I'd talking about earlier doing collaboration with (say) your girlfriend, playing music for her while talking on the phone Sep 13 23:16:40 cesarb: sure, I'm sure there'll be lots of stuff that people develop that I think "cool! I'll be having that thanks!" :) but the basic requirements for me in an ideal phone were browser, wifi,calls + sms Sep 13 23:16:44 something that I think might be possible -- but certainly isn't possible with most mobile phones... Sep 13 23:17:08 AFAIK, the Zire 72 port (at least last time I read about it) disabled the reset button... because if you used it while on linux, it crashed in such a way you had to wait until the non-removable battery drained before being able to reboot Sep 13 23:17:28 ugh Sep 13 23:17:39 vallor: I want, to, for instance, share parts of my work environment between my computer and phone... Sep 13 23:17:59 do you have a smartphone already? Sep 13 23:18:14 vallor: so I might be reading a slashdot article, the time goes to get the bus to the ferry to work, and I can continue on the phone, automatically Sep 13 23:18:34 wifi on the ferry? very snazzy Sep 13 23:18:56 or just caching? Sep 13 23:19:02 vallor: nope, only a (closed, slow, crashy JVM) Razr V3 Black and a (closed, annoying to program for, extremly crashy JVM) Zire 72 (the blue one, that which has flaking paint) Sep 13 23:19:27 vallor: no wifi... just _move_ the memory representation of the page I'm reading to the phone Sep 13 23:19:39 I've been moderately happy with my treo 700p, but it has its "quirks" Sep 13 23:20:20 we just started providing wifi on airport express from santa rosa Sep 13 23:21:07 there's a wifi gadget that takes one of those pcmcia data cards for a carrier Sep 13 23:21:22 the lappy modem card, don't know what they are called Sep 13 23:21:52 I wonder how SF would feel about open wifi on their ferries? :) :) Sep 13 23:26:54 finally I got e17 running on my neo :) Sep 13 23:27:55 :) congrats Sep 13 23:30:28 cool :) is that good ? Sep 13 23:30:30 thomasg: scap or it didn't happen ;-) Sep 13 23:30:47 tomorrow I'll take a video :) Sep 13 23:30:52 :D Sep 13 23:31:18 erm, today I mean (1.30 AM here) Sep 13 23:31:58 e17 = enlightenment, right? Sep 13 23:32:18 i think so Sep 13 23:34:10 is there someone from France there ? Sep 13 23:34:42 * vallor is in California Sep 13 23:34:49 when i received my neo, i got to pay a airport tax ... i was wondering if it's normal Sep 13 23:35:03 * cesarb last used enlightenment back when it was (IIRC) the default gnome window manager... a looong time ago... Sep 13 23:35:10 sunix, I had to pay £42 customs charge for delivery to the UK Sep 13 23:35:23 :S Sep 13 23:35:59 about 70€ for my airport tax Sep 13 23:36:01 sunix: be happy the price didn't almost double for you (60% import tax over the whole price including the delivery, then about 18% sales tax OVER the resulting total) Sep 13 23:37:20 man, that character on rough-trade, or whatever that site's name was, reminds me of this one guy, Russinovich, who once claimed in "NT" magazine that Linux didn't have threads Sep 13 23:37:41 sorry, I missed the link? Sep 13 23:37:46 vallor: url? Sep 13 23:37:48 but it was at a time that, not only did it have a threads, but it also had this new "clone()" call Sep 13 23:38:29 cesarb: where are u from ? Sep 13 23:38:38 http://roughlydrafted.com/RD/TechQ307/Entries/2007/8/23_Apple_iPhone_vs_the_FIC_Neo1973_OpenMoko_Linux_Smartphone.html Sep 13 23:39:01 sunix: Brazil Sep 13 23:39:20 roughlydrafted must be some sort of theonion in disguise, I can find no hter explanation for it. Sep 13 23:39:31 s/ther/other/ Sep 13 23:40:08 Lynet: well, some people get these wild hairs, and get caught up with the "process", forgetting to (or not knowing to) check their facts Sep 13 23:40:47 cesarb: yep so i'm happy :) Sep 13 23:40:54 Heh, doesn't even render properly in firefox. Sep 13 23:42:32 you realise that this makes all of you communists, right?! Sep 13 23:42:37 COMMUNISTS! Sep 13 23:42:42 *ahem* Sep 13 23:43:12 223€for the neo - 53€ for shipping - 70€ for the customs/airport Sep 13 23:43:28 Wasn't it taiwanese windows-supporters supported by the big red scare? Sep 13 23:43:48 Lynet: no, chinese windows-supporters Sep 13 23:43:55 no mention of taiwan at all Sep 13 23:44:07 taiwan > china Sep 13 23:44:30 while at the same time, taiwan ~= china, funny how that works :) Sep 13 23:44:36 http://maps.google.com/maps?f=q&hl=en&geocode=&q=1001+Page+Street,+San+Francisco,+CA&sll=38.466105,-122.715947&sspn=0.006552,0.02281&ie=UTF8&ll=37.772699,-122.436651&spn=0.003307,0.011405&z=17&om=1&layer=c&cbll=37.773485,-122.437494 Sep 13 23:44:50 that's where roughlydrafted.com says they are Sep 13 23:44:53 "OpenMoko was started last year inside of FIC, a Windows PC maker located in Taiwan." Sep 13 23:44:53 the author's address? Sep 13 23:44:55 hahaha Sep 13 23:45:01 * doc|work hands vallor a pitchfork Sep 13 23:45:15 Lynet: ah, ok, my bad Sep 13 23:45:34 looks like a one-man band Sep 13 23:45:37 I have to admit, I scanned, laughed, scanned, laughed, cried a little Sep 13 23:45:41 then posted it here Sep 13 23:45:50 scanned a little more, then closed it Sep 13 23:47:23 http://www.roughlydrafted.com/RD/Journal/7DBAEF76-AE98-4D20-BB8E-4D82D8713D2E.html Sep 13 23:47:34 there's a pic of the guy doing humanitarian work, or something Sep 13 23:47:48 yeah Sep 13 23:48:10 woopdy doo, nice to kids but can't research for shit Sep 13 23:48:28 * juri_ used to use enlightenment before miguel or nat were using linux. Sep 13 23:48:52 ^ old Sep 13 23:48:55 * doc|work runs Sep 13 23:49:01 doc|work: well, hmm -- seems he also got hit by a truck Sep 13 23:49:08 vallor: before or after? Sep 13 23:49:09 "oh, so that's what happened to him" Sep 13 23:49:11 2004 Sep 13 23:49:16 http://www.roughlydrafted.com/sharingairspace.html Sep 13 23:49:26 explains the bitterness to some extent Sep 13 23:52:06 ''As mirrors and signals and other vehicle peripherals exploded below and behind me in an orderly progression, I tried to recall what all is on the back of a tow truck, and the likelihood of hitting any more of it as my forward progression plateaued and I began my descent approach to the pavement in the middle of Mission Street between Seventh and Eighth, just past the construction site for the grand new Federal Building.'' Sep 13 23:52:13 wow. Sep 13 23:52:16 I'll cut the guy a break, heh Sep 13 23:52:32 poor guy Sep 13 23:53:26 quote: Hi, I’m Daniel Eran Dilger, a tech consultant and writer in San Francisco, California. I ride a motorcycle and I like to work on art projects. Sep 13 23:53:29 loool Sep 13 23:53:36 he likes motorcyles Sep 13 23:54:02 wait, he's still on them? I mean, it'd be one thing if he was disabled or something Sep 13 23:54:05 maybe I could post a comment: "numnumnum, what does a tow truck taste like?" Sep 13 23:54:06 how fucking quaint Sep 13 23:54:24 actually, no, you know what? even disabled people should do research, fuck him :) Sep 13 23:54:46 vallor: hah Sep 13 23:55:44 http://www.roughlydrafted.com/about.html <-- but there is how he wants to be known, hmm Sep 13 23:55:53 brb' Sep 13 23:56:09 I'm actually going to email him tonight Sep 13 23:56:22 a nice respectful email Sep 13 23:58:30 it's probably best, doc|work Sep 13 23:58:49 this guy is a douche Sep 13 23:58:53 there's already been an email exchange with him, and it was copied to the mailing list Sep 13 23:59:04 yet another blogoretard with a paypal donation account Sep 13 23:59:07 oh! Sep 13 23:59:07 this is all so "last week" ;-) Sep 13 23:59:12 rwhitby: :( Sep 13 23:59:21 let me guess, that didn't get anywhere? Sep 13 23:59:24 * doc|work will have to look at the mailing list later :) Sep 14 00:00:29 This the thread start? http://www.mail-archive.com/community@lists.openmoko.org/msg09067.html Sep 14 00:00:32 his only technical skill seems to be apple certified technician Sep 14 00:00:51 ah ha! apple fan boi! I knew it Sep 14 00:00:58 http://roughlydrafted.com/resume.html Sep 14 00:03:06 now there's a new wiki page trying to give a high-level view on power management: http://wiki.openmoko.org/wiki/Power_management_requirements#Power_saving Sep 14 00:03:11 :) Sep 14 00:03:18 mike hodson's repsonse++ Sep 14 00:03:26 s/#Power_saving/ Sep 14 00:03:31 / Sep 14 00:04:15 his comments remind me a little of my own linux zealotry Sep 14 00:04:33 sometimes you get into a row with someone and try to dig up a lot of nonsense just to justify your own position Sep 14 00:04:55 wow, that's long.... Sep 14 00:05:01 http://maps.live.com/default.aspx?v=2&cp=q8vt624sx6qq&style=o&lvl=2&tilt=-90&dir=0&alt=-1000&scene=13286728&sp=Point.q8vtnz4sx632_1001%20Page%20St%2C%20San%20Francisco%2C%20CA%2094117%2C%20United%20States___~Point.q8vtwz4sx677_1001%20Page%20St%2C%20San%20Francisco%2C%20CA%2094117-2218%2C%20United%20States___&encType=1 Sep 14 00:05:08 I KNOW WHERE YOU LIVE, BIATCH! Sep 14 00:05:10 hehe Sep 14 00:05:15 lol Sep 14 00:05:23 just being silly Sep 14 00:05:42 to closed to your home to send a nuclear bomb :p Sep 14 00:05:52 to -> too :) Sep 14 00:08:36 o fsck.. Sep 14 00:08:52 http://www.digg.com/linux_unix/Openmoko_Phone_linux_based_New_UI_Screenshots - 270 diggs and 76 comments on my blog post Sep 14 00:09:01 thats yours? Sep 14 00:09:03 right on Sep 14 00:09:06 nice placement Sep 14 00:09:57 home time Sep 14 00:10:51 hrw, is the page down already? Sep 14 00:11:01 i get a blank page Sep 14 00:11:02 tuukkah: no - its dugged Sep 14 00:11:56 tuukkah: Cool, another "requirements" page... I have mine on http://wiki.openmoko.org/wiki/User:CesarB/Requirements Sep 14 00:13:07 tuukkah: mine is more generic, however... Sep 14 00:13:14 during normal days my apache log was ~100KB. now it is 5.5MB Sep 14 00:13:56 cesarb, we should have some process for requirements management so these wouldn't be personal opinions anymore Sep 14 00:14:45 http://blog.haerwu.biz.nyud.net:8080/2007/09/13/openmoko-20072-testing/ is mirror if someone want to test Sep 14 00:14:52 s/test/look Sep 14 00:14:57 tuukkah: we have a wiki... isn't that enough? we can just make a [[Requirements]] page and start shaping it Sep 14 00:15:01 hrw: are you looking for a place to host the images? Sep 14 00:15:08 vallor: no Sep 14 00:15:14 kk Sep 14 00:15:41 tuukkah: mine was more intended as a personal checklist (the first item on it is gsmd - so I started trying to make gsmd work reliably... I failed so far, but at least I tried ;-) ) Sep 14 00:16:30 cesarb, indeed, like you see that's what i did wrt power management requirements ;-) but wiki has the problem that there is not explicit process to get people to accept the results Sep 14 00:17:43 tuukkah: no need for any "acceptance", it can be just a focus for further development Sep 14 00:17:48 cesarb, i think your power management requirements are mostly covered by my page, except for "Tone when battery low" which is about ui Sep 14 00:18:07 cesarb, but it's no use if people don't look at it Sep 14 00:18:11 tuukkah: if you have something as "must have", you will try to work on it earlier than a "nice to have" one... Sep 14 00:19:08 tuukkah: yeah, the tone is UI, but it's very important (it tells you you should recharge soon or lose your phone for a while... it's even more important in the middle of a call, where you can tell the other side "I'll call you later, my battery is almost empty") Sep 14 00:20:01 tuukkah: my page was written from an "end-user" POV, so whether something is UI or plumbing or voodoo doesn't make a difference Sep 14 00:20:20 hmm perhaps my page should get third category: ui :-) Sep 14 00:20:51 "The phone should keep the user informed of power management state and events" Sep 14 00:21:04 hrw: nice review, gives 'em a taste :) Sep 14 00:21:12 tuukkah: precisely. nice way to put it. Sep 14 00:22:08 quick question: does the calendar accept vcard data via bluetooth? I might try feeding it my contact list from the treo... Sep 14 00:23:50 "The “free and open Anti-iPhone” rhetoric surrounding the FIC phone is therefore marketing drivel to sell a Chinese Windows Mobile device as a hobbyist kit for phone hackers." interesting POV Sep 14 00:26:18 6000 hits in 2 hours... Sep 14 00:26:20 * rwhitby notes never to blog with pictures that might cause it to get dugg Sep 14 00:26:32 normally it takes 4-6 days for such amount Sep 14 00:28:23 rwhitby: or place pictures on flickr Sep 14 00:28:49 NOTE: build 200709131540: completed :) Sep 14 00:29:06 eh - I never write anything interesting enough to digg anyway. Sep 14 00:29:11 tuukkah: I added a category Sep 14 00:29:36 about an hour and forty minutes to build Sep 14 00:30:00 rwhitby: I didn't thought that such post can be so popular Sep 14 00:30:45 so throw some google adsense ads on there ;P Sep 14 00:31:30 vallor: I do not like ads Sep 14 00:31:42 I was half-kidding ;P Sep 14 00:33:33 hrw: ah, but Google's are Not Evil(tm) Sep 14 00:34:41 cesarb, the category's a good idea Sep 14 00:35:01 i'm celebrating my second year in a row turning down a sysadmin job at google. i'm a programmer, not a sysadmin. DAMNIT. Sep 14 00:35:37 tuukkah: I just wasn't so sure about the category's category Sep 14 00:35:39 juri_: it's the same sport that i practice. refusing jobs Sep 14 00:35:52 * hrw refused google offer too Sep 14 00:35:54 juri_: btw i'm italian. Italians do not work. Sep 14 00:36:10 tuukkah: I sorta picked a reasonably-looking one, but I probably chose the wrong one Sep 14 00:36:17 what's the point offering a job to an italian guy Sep 14 00:36:29 tuukkah: I have always been BAD with categories... Which is why I let dealing with them to the other wikipedia admins Sep 14 00:36:39 cesarb, the categories are blurry in any case. i think it's more important to put links in good places Sep 14 00:36:40 bedboi: hurts a lot for me; google would pay ~3X what i'm making now. Sep 14 00:36:41 juri_: have you ever had a sysadmin job? Sep 14 00:36:57 cesarb: unfortunately, most of my carreer. Sep 14 00:37:06 juri_: if you had, then you already lost - once a sysadmin, always a sysadmin. You'll sooner or later get back to it. Sep 14 00:37:16 juri_: whatever i accept would pay at least 5x :) Sep 14 00:37:31 juri_: (something like that is one of the scary devil monastery's mottoes) Sep 14 00:37:39 2.5 years at my current job. Sep 14 00:37:51 well, I work, but my (theoretically current) r&d project is fiddling with a meraki -- so the whole embedded linux world(s) is exciting, new, and "work-related" (if you squint a certain way) ... Sep 14 00:38:04 * cesarb had a sysadmin job in the past... went to a programmer job... and is back at sysadmining again :D Sep 14 00:38:21 its hard to emphasize that my programming skills are why i make a good sysadmin, not the other way arround. :) Sep 14 00:38:24 sysadmin has to be one of the best jobs out there Sep 14 00:38:29 if you are a linux sysadmin that is Sep 14 00:38:33 now a job developing openmoko would be something... :-) Sep 14 00:38:39 once you set things up, all you do is twiddle your thumbs most of the time Sep 14 00:38:43 holycow: it depends. Sep 14 00:38:57 holycow: sysadmin is only one of the best jobs if you are one step ahead of the trouble Sep 14 00:39:00 holycow: for me it would be the worst Sep 14 00:39:10 cesarb: thats true too Sep 14 00:39:13 holycow: which is easier with linux, of course Sep 14 00:39:14 i wrote remote desktop printing code over the last few weeks. is that sysadmining? :) Sep 14 00:39:22 evening. Sep 14 00:39:27 hi mickeyl Sep 14 00:39:28 hi mickeyl Sep 14 00:39:29 holycow: if you aren't one step ahead... it can be one of the most stressing computing jobs Sep 14 00:39:30 bedboi: sure everyone has their preferences, yous is? Sep 14 00:39:38 mickeyl: good night. Sep 14 00:39:39 cesarb: yup that too Sep 14 00:39:43 holycow: C coding. Sep 14 00:39:49 * cesarb is against the silly "ugt" Sep 14 00:39:51 holycow: and research Sep 14 00:39:53 of course Sep 14 00:39:54 :) Sep 14 00:39:55 hrw: do you mind if I pass on that mirrored link to our local lug's mailing list? Sep 14 00:40:07 i like research too, not too many people want to pay me to google all day long tho Sep 14 00:40:14 ... oh you mean REAL research :) Sep 14 00:40:24 holycow: yep that stuff, conferences, journals Sep 14 00:40:30 vallor: only if you pass original url + mirror one Sep 14 00:40:46 holycow: they call it academia. Sep 14 00:40:56 I can give them just the original url -- what is it? :) Sep 14 00:40:57 actually i could dig that too Sep 14 00:41:12 mickeyl: posting OpenMoko screenshots can be deadly Sep 14 00:41:27 hrw: heh. too much load on your site? :D Sep 14 00:41:47 mickeyl: there was moment when there was no site at all Sep 14 00:41:51 wow Sep 14 00:42:11 had a whole day of fighting w/ WLAN today Sep 14 00:42:16 mickeyl: since midnight I got nearly 7000 views. Sep 14 00:42:23 mickeyl: and its 2:42 here Sep 14 00:42:31 i know, same timezone ;) Sep 14 00:42:36 hrw: what? Sep 14 00:42:40 mickeyl: fighting on a neo or not? ;-) Sep 14 00:42:46 mickeyl: can I remove moko-autorev.inc from openmoko.conf and then bleeding edge developers just add it to local.conf? Sep 14 00:42:50 i'll +1 to your page views Sep 14 00:42:52 mickeyl: normally I am getting less then 1000 per day Sep 14 00:43:19 and we bump SVN revs in sane-srcrevs to indicate that a new version of an openmoko app is *intentionally* available for testing by the community Sep 14 00:43:45 (see discussion in #oe about 11 hours ago) Sep 14 00:43:53 hrw: be glad you didn't get them on ashdotslay Sep 14 00:43:56 mickeyl: I am also for removing it. nevermind that I added it there Sep 14 00:44:04 cesarb: hehe, unfortunately not. I don't have the WiFi-enabled version yet. After changing my DSL router to an ADSL2+ one, I had to reconfigure my wlan infrastructure. Sep 14 00:44:23 rwhitby: yeah, sounds reasonable. go for it Sep 14 00:44:34 hrw: you were maintaining the status quo, which was appropriate for the period before we could discuss with OM core team Sep 14 00:44:39 mickeyl: woohoo!!! Sep 14 00:45:30 i will probably stop making bleeding edge builds available on the web as well Sep 14 00:45:35 there's too much breakage Sep 14 00:45:46 mickeyl: here I have cablemodem->router/AP->machines. you have other config? Sep 14 00:46:39 ~praise mickeyl Sep 14 00:46:40 All hail mickeyl! Sep 14 00:46:45 hrw: yeah. cablemodem/router -> fixed ETH machines | AP downstairs (WDS) | wireless machines -> AP upstairs (WDS) -> fixed ETH machines | wireless machines Sep 14 00:46:47 ~worship mickeyl Sep 14 00:46:48 * apt genuflects before mickeyl. "I'm not worthy!" Sep 14 00:48:36 heh, that'll learn me -- tried to use the volume control on the mp3 player app -- woopsy, now it won't go back up... Sep 14 00:49:35 vallor: Not going up at all or not to the max? Sep 14 00:50:06 vallor: can't you just reload the alsa state? Sep 14 00:50:11 * rwhitby pushes the remove of moko-autorev.inc Sep 14 00:50:21 fwiw, i spent the whole night on gstreamer. i understand a lot more now :) Sep 14 00:50:34 mickeyl: Oh? Anything worth mentioning to me? :) Sep 14 00:50:48 abraxa_: it goes up, then moves back down when I let go Sep 14 00:50:53 mickeyl: could you kick off a build based on the current sane-srcrevs as a baseline? Sep 14 00:51:01 cesarb: I just loaded an alsactl state, didn't work Sep 14 00:51:25 abraxa_: yeah. 32bit support in pulseaudio will be a lot of work. we better tackle mad. the fun thing is the mad plugin once worked w/ 16 bit (in the gstreamer 0.8 days), so we might be able to steal some code from that Sep 14 00:51:37 i'm also 100% sure that we don't like dmix :D Sep 14 00:51:56 vallor: That's caused by the slider if you pull it up too high and the 'cursor' leaves the slide area - try raising it to 80% or so and it should stay Sep 14 00:52:04 rwhitby: sure Sep 14 00:52:17 let me push as well here... Sep 14 00:52:49 mickeyl: Very interesting. Makes me wonder why 16-bit support was removed in the first place Sep 14 00:53:03 abraxa_: because hifi people complained about the quality Sep 14 00:53:10 lol? Sep 14 00:53:32 well, they have a point. mad does a lot to extrackt clean 32 bit int and gstreamer just threw the 16 bits away, not even dithering :D Sep 14 00:53:40 but for us, this would be perfect Sep 14 00:53:59 so we want 16bit back :D Sep 14 00:54:11 Yeah but using mp3 and complaining about 16 bit not providing enough resolution? Ah well. Sep 14 00:54:32 And yeah, I'll look into that Sep 14 00:55:47 hehe, right. there's a certain irony in that Sep 14 00:55:56 * mickeyl likes uncompressed audio Sep 14 00:56:33 anyone know about the vtable problems with the current source (icu problems)? Sep 14 00:56:34 what about .flac? ;P Sep 14 00:56:50 yeah flac is interesting Sep 14 00:57:01 mickeyl: when Neo will get SID/MOD/S3M/XM player? Sep 14 00:57:03 it's just that the numbers are not impressive enough to do work on that Sep 14 00:57:20 hrw: does mikmod work on it? Sep 14 00:57:22 hrw: when i have time to dive back into application programming again Sep 14 00:57:27 yes, mikmod runs Sep 14 00:57:33 hrw: because IIRC mikmod is split into a UI and a library... Sep 14 00:57:35 re Sep 14 00:57:37 but a nice GUI-based player for tracking stuff would be nice Sep 14 00:57:38 wb roh Sep 14 00:57:47 Isn't there a gstreamer plugin for modules, too? Sep 14 00:57:48 It should be easy to add a new UI on top of it Sep 14 00:58:00 I tried to rebuild icu but that didn't help Sep 14 00:58:03 abraxa_: yes, but gstreamer doesn't export the channel/sample/note info Sep 14 00:58:14 that's why modplayers can provide a more rich interface Sep 14 00:58:25 Point taken :) Sep 14 00:58:41 although there's probably not much of a difference when all you want is listen ;) Sep 14 00:59:08 mickeyl: just make it work with .zip files like mikmod does ;-) Sep 14 00:59:15 undefined reference to `__gxx_personality_v0@CXXABI_1.3; did gcc version or option change tha requires rebuilding everything? Sep 14 00:59:16 cesarb: :) Sep 14 00:59:20 hm... what was that about dmix? thats dead and dreprecated Sep 14 00:59:25 -r Sep 14 00:59:34 roh: yes, that's what we said Sep 14 00:59:44 tnb: news to me. changed your host system? Sep 14 01:00:18 no, just did 'make update; make' -- used to build Sep 14 01:00:39 (moko makefile) Sep 14 01:00:54 wow, you can switch the thing to landscape mode :-o Sep 14 01:01:41 vallor: :D Sep 14 01:02:43 shoot, I was hoping it would make the kbd bigger Sep 14 01:02:59 kbd font is hardcoded atm. :/ Sep 14 01:03:08 ay carumba Sep 14 01:03:11 going to pester alphaone to fix kbd Sep 14 01:03:17 he's going to visit me next week Sep 14 01:03:25 we're having a two days moko hack fest Sep 14 01:03:29 at my place Sep 14 01:03:38 zecke will also be there Sep 14 01:03:57 sounds swizzy :) Sep 14 01:03:57 hmm... need to clean my room before they come :) Sep 14 01:04:12 mickeyl: I'll clean out tmp/ and rebuild everything overnight and see what happens Sep 14 01:04:25 tnb: good plan Sep 14 01:04:49 * mickeyl hacks linux.inc to support NFSROOT more easily Sep 14 01:05:19 mickeyl: which of your machines use linux.inc? Sep 14 01:05:38 hrw: ezx and fic Sep 14 01:06:11 yes, linux-openmoko is not commited yet. needs some more clean up, but i'm on it Sep 14 01:08:16 * hrw -> bed Sep 14 01:08:20 bye all Sep 14 01:09:36 n8 hrw Sep 14 01:09:38 hmm Sep 14 01:09:43 what does sed -e '/CONFIG_AEABI/d' \ do? Sep 14 01:09:46 i.e. the /d Sep 14 01:10:37 mickeyl: find and del? Sep 14 01:10:42 mickeyl: I'd guess it's the "delete" command to sed Sep 14 01:10:51 looks like it Sep 14 01:10:54 mickeyl: the prefix selects the lines, and the command does the act... Sep 14 01:10:55 hmm Sep 14 01:11:01 then i don't understand linux.inc :D Sep 14 01:11:05 first it's added, then removed? Sep 14 01:11:19 isn't there a name for that ;) Sep 14 01:11:21 aah Sep 14 01:11:25 nevermind, got it Sep 14 01:11:41 mickeyl: it should be: delete the line (irrespective of whether it is enabled or not), and then add a new line to enable or disable Sep 14 01:11:42 we're sedding it out of the defconfig and concatenate to our .config Sep 14 01:11:47 right Sep 14 01:12:38 I can make the media player stutter by running alsamixer, then leaning on the right-arrow Sep 14 01:12:44 mickeyl about that svnrev/autorev stuff ive seen in the changelogs.. can you write 5 lines what that means for the people out there? Sep 14 01:12:47 (or left arrow) Sep 14 01:12:54 network interrupts confusing it? Sep 14 01:13:09 mickeyl: salvation Sep 14 01:13:09 roh: sure. will do it tomorrow Sep 14 01:13:35 roh: world peace Sep 14 01:13:50 it's as simple as that ;-) Sep 14 01:14:08 well, it also means less testing and if i was cynical i would claim it means stagnation Sep 14 01:14:17 (just to bring some counterarguments) Sep 14 01:14:20 any rummors about GTA02 release date? Sep 14 01:14:26 mickeyl super. anything important to keep in mind? ive seen that one line in conf/distro/openmoko.conf.. i have to remove the hash and then it builds trunk, elsewise it builds fixed svnrevs, right? Sep 14 01:14:49 roh: yes, that's basically it. remove the # to get the old floating SRCDATE behaviour back Sep 14 01:14:51 mrcucumber when its done. Sep 14 01:14:54 vallor: gstreamer causes CPU load > 100% so the GUI responds very sluggishly Sep 14 01:15:07 mickeyl: I think the drive to get access to new features mentioned in commits that CIA reports here will be enough to drive people to bump SRCREV Sep 14 01:15:08 oh hmm Sep 14 01:15:16 mickeyl wonderful. Sep 14 01:15:16 rwhitby: hopefully :) Sep 14 01:15:16 roh: hehe, I know, doesn't mean I can't look forward to it right? Sep 14 01:15:18 if you're after testing you can always release as many binaries as you want - if you want to multiply real development SRCREV should help do that a lot Sep 14 01:15:30 mrcucumber ;) Sep 14 01:15:46 abraxa_: just limit the networking to 10000 packets per second Sep 14 01:15:50 at this point of time i think the 2d driver will be the hardest part Sep 14 01:16:01 roh: as the comment says, I would suggest putting the line in local.conf so you don't accidentally check it back in Sep 14 01:16:01 cesarb: huh? Sep 14 01:16:25 rwhitby ah.. good idea. Sep 14 01:16:41 Writchie: agreed Sep 14 01:16:56 roh: So still looks like october? Sep 14 01:17:31 abraxa_: http://blog.rlove.org/2007/08/those-dang-dpcs-clogging-mmcss.html Sep 14 01:18:07 cesarb: To be frank I'm not quite sure why you're showing me this Sep 14 01:18:18 abraxa_: it was an attempt at an obscure joke Sep 14 01:18:25 [22:13:53] vallor: gstreamer causes CPU load > 100% so the GUI responds very sluggishly Sep 14 01:18:28 mrcucumber i dunno. as long as ive not got my preseries prototype its not even in reach Sep 14 01:18:42 cesarb: In that case I am guilty for not getting it ;) Sep 14 01:18:43 * cesarb seems to have failed at the obscure joke Sep 14 01:19:10 roh: yeah, as a college student I couldn't really afford to get the dev moddel. Sep 14 01:19:16 abraxa_: it's been discovered recently that windows vista limits the networking stack to 10000 packets per second while playing multimedia, to avoid stuttering Sep 14 01:19:40 cesarb: Oh! Okay, that makes a whole lot more sense now :) Sep 14 01:19:41 abraxa_: which is more than 100Mbit but way less than 1Gbit Sep 14 01:19:54 abraxa_: obviously, slashdot was all over it ;-) Sep 14 01:19:55 cesarb: Shows how important Vista is in this channel, hehe Sep 14 01:21:22 abraxa_: http://it.slashdot.org/it/07/08/21/1441240.shtml and http://blogs.technet.com/markrussinovich/archive/2007/08/27/1833290.aspx and http://slashdot.org/articles/07/08/28/0256208.shtml Sep 14 01:21:48 cesarb: does it limit packets, or interrupts? Sep 14 01:21:57 vallor: packets Sep 14 01:22:01 blah Sep 14 01:22:16 anyone knowing offhand how the kernel parses CMDLINE wrt. to double entries? will the first one or last one be significant? Sep 14 01:22:22 vallor: it has been experimentally determined (heh) that 10000 packets would avoid the stuttering Sep 14 01:22:30 mickeyl: I think both will Sep 14 01:22:48 mickeyl: the final effect would depend on the code for that particular option Sep 14 01:22:55 hmm Sep 14 01:22:56 that sucks Sep 14 01:23:00 so it's unpredictable Sep 14 01:23:05 i.e. when given two rootfs statements Sep 14 01:23:08 mickeyl: the neo itself already has two console= lines... Sep 14 01:23:16 right Sep 14 01:23:40 mickeyl: two rootfs will probably use the last one, since the code probably can only deal with one and thus will overwrite (you mean root= right?) Sep 14 01:24:35 rootfstype and root, yea Sep 14 01:25:50 vallor: "MMCSS’ glitch-resistant mechanisms were therefore extended to include throttling of network activity. It does so by issuing a command to the NDIS device driver, [...] that causes NDIS to “indicate”, or pass along, at most 10 packets per millisecond (10,000 packets per second)." -- Mark Russinovitch Sep 14 01:27:44 cesarb: lol, I'm so glad I don't use that stuff Sep 14 01:28:16 and you'll remember I mentioned Russinovitch before -- you'll see me flaming him on linux-kernel, if you go back far enough ('98?) Sep 14 01:28:46 vallor: the rumours I've read around say people are tending to leave Vista for XP Sep 14 01:29:15 he's the character that said Linux didn't have threads, then had to backtrack on his site and say: oh, well, it may have "threads", but they don't act like NT threads, so I'm not going to call them "threads" Sep 14 01:29:35 vallor: and the people around me seem to also use XP way more often than Vista (I think I only actually touched a box with Vista once, to show how to use the ssh setup I designed on putty to a coworker..) Sep 14 01:29:57 I'm not touching Vista -- too much "contacting the mother ship" for my blood Sep 14 01:30:13 vallor: the first I heard of him was the Sony Rootkit thing, where he did a wonderful job, so I didn't had quite that negative image of him... Sep 14 01:30:53 vallor: until this recent ridiculous "10000 packets per second stuff", where the tone on his blog was too accepting of such a bizarre decision Sep 14 01:31:34 vallor: XP already annoys me enough with its Windows Genuine Annoyance Sep 14 01:31:47 vallor: the last good Windows version was 2000 Sep 14 01:31:56 actually OS/2 Sep 14 01:32:01 vallor: but the last Windows version I actually used at home was 3.1 Sep 14 01:32:10 NT beta 5 was fine for years Sep 14 01:32:24 NT 5 beta IIRC Sep 14 01:32:54 * cesarb gave up on Windows when I noticed I was using my Linux install almost all the time, instead of the DOS/Win3.x combo Sep 14 01:33:12 * technil is happy he has never had to use windows at home.. Sep 14 01:33:26 unfortunately you can't give it up in the consumer space when 95% of people use it. Sep 14 01:33:41 technil: my first box is from 92, I didn't even know linux existed back then Sep 14 01:33:42 ask apple how many windows itunes vs. mac itunes Sep 14 01:34:38 http://uwsg.indiana.edu/hypermail/linux/kernel/9901.1/1243.html Sep 14 01:34:39 technil: I started using Linux only in 97, after wanting a real Unix so Emacs would actually multitask during compiles, unlike how it did on DJGPP Sep 14 01:34:40 cesarb: my first box was cp/m and microsoft supplies the M80 macro assembler and L80 linker Sep 14 01:34:45 there's mark's response Sep 14 01:34:50 cesarb: last decent software ever to come out of M$ Sep 14 01:34:57 (ironically, after starting using linux, I completely abandoned emacs for vi) Sep 14 01:35:12 cesarb: how come? Sep 14 01:35:15 btw, his case goes to show that PhD's don't always mean someone has their head screwed on straight ;P Sep 14 01:35:17 you miss the real coolness Sep 14 01:36:26 bedboi: dunno why, it just happened Sep 14 01:36:35 no!joe is the One True Editor!!!!!@!! Sep 14 01:36:39 it's a sad story indeed Sep 14 01:36:44 vallor: don't even mention that. Sep 14 01:36:49 ;P Sep 14 01:37:00 bedboi: the GEmacs/XEmacs split doesn't help (which one should I use? they are identical but very different...) Sep 14 01:37:03 vallor: btw it's a dangerous topic. talking about editors, languages and distros Sep 14 01:37:32 cesarb: that's why you have to use vim Sep 14 01:37:32 :) Sep 14 01:38:03 i'll stop, it is really a flame-prone discussion. Sep 14 01:38:04 bedboi: vim is what I use the most (both pure text-mode vim and gvim)... even when I have to use windows Sep 14 01:38:16 bedboi: it's the only way to avoid writing ":wq" by accident on notepad Sep 14 01:39:46 But yeah, how far has windows come in all this years... In the past, it was so obviously inferior to linux, it was laughable Sep 14 01:40:41 Nowadays you can compare both (it doesn't have 16-bit DOS as its base anymore, for instance!) Sep 14 01:42:08 i don't know quite anything about windows Sep 14 01:42:11 OTOH, in the past windows had something linux didn't: cool sounds and cute icons (see http://worsethanfailure.com/Articles/The-Cool-Cam.aspx to understand why it's important) Sep 14 01:42:30 btw, here was the guy's conclusion -- btw, this was in 1999, so he was definitely full of it: Sep 14 01:42:32 Thus, there are a number of critical areas that must be addressed before Sep 14 01:42:32 Linux can be considered to have a real (competitive) kernel-mode threads Sep 14 01:42:32 implementation and a scalable SMP architecture. And until it has these Sep 14 01:42:32 things Linux cannot compete in the enterprise with UNIX rivals or with Sep 14 01:42:32 Window NT. Sep 14 01:42:43 internally XP and Vista is like thousands of chroots for different apps Sep 14 01:42:46 -Russinovich, Jan 1999 Sep 14 01:42:52 with hidden foo Sep 14 01:44:22 vallor: he probably knew only about NT and "traditional" Unix, and didn't know enough about Linux (and yeah, by basing his conclusions on partial knowledge, in such a certain tone, he showed he was full of it) Sep 14 01:45:45 vallor: I'm a bit the opposite... I know a bit about linux kernel internals (hey, I rewrote a NIC driver last year!), but nothing about windows kernel internals (other than IRQL_NOT_LESS_OR_EQUAL means bad hardware) Sep 14 01:47:04 OS2 threads were quite innovate at the time they were first introduced, compared to *nix's of the day Sep 14 01:47:14 and the rediculous fork Sep 14 01:47:37 unfortunately the MS IBM OS2 wars set the industry back by 10 years Sep 14 01:47:45 vallor: which is why I want to learn how to program for windows... to be exposed to different programming paradigms (same reason I learned Haskell last year) Sep 14 01:48:01 Writchie: I gotta ask... are you related to "Dennis"? :) Sep 14 01:48:08 vallor: unfortunately, so far I failed to find a good "Win32 programming for Unix programmers" Sep 14 01:48:37 vollar: no Sep 14 01:48:53 dang! You would have been an instance celeb ;P Sep 14 01:49:21 I am already a celebrity Sep 14 01:49:36 And their API seems to change completely every couple of years (now the recommended way is THIS way, oops now it's THAT way) Sep 14 01:49:39 very cool :) Sep 14 01:50:50 cesarb: even MS people don't know the prefered paradigm du jour Sep 14 01:52:17 * cw666 returns from walking about collective GPS data Sep 14 01:53:28 Writchie: on my last job, I started reading a tiny bit on it... and found out that to OPEN a file _or directory_, you call a funcation called _CreateFile_... W T F Sep 14 01:53:44 Writchie: just that tiny bit was more than enough to make my head hurt Sep 14 01:53:52 you can open a file about 8 different ways Sep 14 01:54:21 ahaha, what about java then Sep 14 01:54:27 uggh Sep 14 01:54:29 you have 8bilion ways Sep 14 01:54:39 i'd take macro assembler over java Sep 14 01:54:41 just one for each extension. Sep 14 01:54:42 g'night Sep 14 01:55:02 sleep tight mickey Sep 14 01:55:05 Writchie: for all its failures, Unix's chamaleon two/three-argument open() and two-argument creat() are a beautiful gem compared to win32's madness Sep 14 01:55:31 hey all, anyone know who to contact at fic about plans for them to approve the 850 band operation in the US? Sep 14 01:55:42 (for the neo 1973) Sep 14 01:55:45 it wasn't really madness in the original API because it gave you a lot of IOCTL like options Sep 14 01:56:04 however, i agree that simplicity is worth more. Sep 14 01:56:25 but i wouldn't call pselect() that elegant Sep 14 01:56:44 Writchie: ioctl is the worst Unix API wart, and even it is elegant compared to the 30+-argument functions I've seen win32 use Sep 14 01:56:58 multi-threaded programming was much easier originally on OS/2 which became NT which became XP Sep 14 01:57:49 i still remember puking the first time I encounter fork() Sep 14 01:58:07 one look and goodbye Sep 14 01:58:08 Writchie: nah, fork() is cute... vfork() is the ugly one Sep 14 01:58:21 it's a cute trick Sep 14 01:58:40 but incredible pig Sep 14 01:59:14 how can i get at the console on the neo when i mux the gsm module out? Sep 14 01:59:21 the debug board has no rs-232 port (i assumed it would) Sep 14 01:59:30 unfortunately, it's hard to evaluate anything except in the context of history at the time Sep 14 01:59:30 it is a happy think in Linux, as it uses copy-on-write pages Sep 14 01:59:52 cw666: the debug board has a serial and a jtag port, both connected to the ftdi chip Sep 14 02:00:03 cw666: the debug board has usb - serial Sep 14 02:00:04 cw666: which is on the usb Sep 14 02:00:25 cw666: so, you can get to both (serial and jtag) via the usb on the debug board Sep 14 02:00:33 vallor: fork and threads in linux are now fine - but it wasn't always that way Sep 14 02:00:50 copy on write bailed out fork Sep 14 02:01:04 vallor: copy-on-write was implemented by linus way back on the 0.x days... Sep 14 02:01:15 oh ok, that makes sense ... does usbnet also work? Sep 14 02:01:29 cw666 the debug board itself has a serial on cmos 3.3V levels. Sep 14 02:01:32 vallor: just find the one where it says he can now use several bash instances ;-) Sep 14 02:01:50 roh: access to the uart on the samsung chip via any sane means is fine, that sounds usable Sep 14 02:01:52 sudo modprobe ftdi_sio vendor=0x1457 product=0x5118 Sep 14 02:02:02 then plug the debug board. Sep 14 02:02:06 cw666: you should plug the neo's usb to the debug board's builtin hub, and then you can get to the neo's normal usb networking Sep 14 02:02:16 roh: ftdi_sio is in mainline? Sep 14 02:02:17 then it generates /dev/ttyUSB0 and USB1 Sep 14 02:02:21 cw666 japp Sep 14 02:02:26 okies, ill try that later Sep 14 02:02:52 cw666 the first one vanishes as soon as openocd starts and claims the hardware for the jtag, the second should be your serial console Sep 14 02:02:56 i have to find my torque thangs to opeen the neo up Sep 14 02:02:59 cw666: http://wiki.openmoko.org/wiki/Debug_Board has _all_ the instructions Sep 14 02:03:13 thanks Sep 14 02:04:54 cw666: and you _must_ read http://wiki.openmoko.org/wiki/Disassembling_Neo1973 _before_ opening the neo, since there are a couple of important tricks Sep 14 02:05:10 cw666: you must use the guitar pick correctly so as to not break the pins, for instance Sep 14 02:07:07 cesarb: I find my thumb nail works better than the pick Sep 14 02:07:20 but my son uses the pick np Sep 14 02:08:05 Writchie: I didn't open mine yet... didn't see the need, and I'm sort of scared of a very small piece of paper with a colored cross in the middle Sep 14 02:08:34 cesarb: it voids the 14 day warrantee ;) Sep 14 02:09:38 * cesarb ponders if it has been 14 days already Sep 14 02:09:38 cesarb: at least its not a skull and crossbones Sep 14 02:09:42 Writchie: lol Sep 14 02:09:53 heh, neither of my neos have the screws in them, cause they get opened up too often Sep 14 02:10:17 screws are for lusers Sep 14 02:10:53 rwhitby: yeah, right, and then your neo _pops open in the middle_ (you alreayd said so in the wiki) Sep 14 02:10:58 anyway, we need to carry a lead acid battery to be portable at the moment Sep 14 02:11:01 rwhitby: the screws are there for a reason Sep 14 02:13:00 verizon going to court to block 700mhz open network - spends $2B advertising $1B lobbying and $2 billion on layers - how about spending some of that on the freakin network Sep 14 02:13:12 s/layer/lawyers/ Sep 14 02:13:13 Writchie meant: verizon going to court to block 700mhz open network - spends $2B advertising $1B lobbying and $2 billion on lawyerss - how about spending some of that on the freakin network Sep 14 02:14:18 Writchie: "Change is dangerous" -- The Oracle Sep 14 02:14:23 i'm begining to root for the patent troll NTP cleaning their clock Sep 14 02:16:03 * cesarb goes to wikiquote Sep 14 02:16:20 "The Architect: You played a very dangerous game." "The Oracle: Change always is." Sep 14 02:16:43 * Writchie salutes cesarb Sep 14 02:17:42 Writchie: guess I'll have to find someone with the DVDs, my memory of the movies seems to be getting degraded Sep 14 02:18:45 i got the HDDVD but the disk of the final was bad Sep 14 02:18:54 big crack thru the HDDVD Sep 14 02:19:20 however, http://xkcd.com/307/ Sep 14 02:20:21 Writchie: well, I'm in a bit of the dark ages... just last month we bought our first DVD player, and I've never viewed a real DVD on my computer (which has had a DVD drive since some years ago) Sep 14 02:21:56 * cesarb probably will take very long before buying a HDDVD/BluRay/whatever... the inane DRM BS drove me away Sep 14 02:21:59 cesarb: i'm the opposite, i'm on my 4th plasma, 2nd HDDVD with a library of prolly 1000 vcd's Sep 14 02:22:49 personally, i like vcd's shot in the theatre the best - good laughs from the crowd Sep 14 02:23:56 Writchie: plasma... lol... we bought our first widescreen LCD TV this year, before that we used a conventional non-flatscreen one (and, yeah, my main monitor _is_ a flat-screen 1600x1200 17" lead-like monitor) Sep 14 02:24:23 * cesarb simply isn't much into movies/TV Sep 14 02:24:24 my first plasma is now 10 years old Sep 14 02:24:29 still working fine Sep 14 02:24:39 Writchie: wow... 10 years ago, plasma TVs were very expensive Sep 14 02:24:46 $10K Sep 14 02:24:59 couldn't resist Sep 14 02:25:01 Writchie: ouch. You could buy a car with that :P Sep 14 02:25:28 * cesarb is usually in the trailing edge... the neo is perhaps the first time I have been in the very bleeding edge Sep 14 02:25:50 damn i have makefile problem. I have a tree similar to A/1 A/2 B/3 B/4 ... The problem is that A/2 relies on B/3 Sep 14 02:26:06 btw i don't want to put all in one directory Sep 14 02:26:51 bedboi: the traditional solution is to enforce ordering of the subdirectories in the base makefile Sep 14 02:27:10 cesarb: i just returned a defective 42" plasma whose 3 year warrantee was just up and got a 52" in its place for $0 Sep 14 02:27:30 bedboi: i.e. for every target, it calls the subdirectories in a specific order Sep 14 02:27:43 bedboi: the non-traditional solution is to have a single makefile at the root ;-) Sep 14 02:28:04 cesarb: i have the ordering Sep 14 02:28:12 btw it's impossible... Sep 14 02:28:31 cesarb: river TAM beats up everyone - I can't wait Sep 14 02:28:34 Writchie: ours is a 32" IIRC Sep 14 02:28:51 cedric: i can't specify A/1 B/3 A/2 B/4 Sep 14 02:29:07 actually, i think XVGA in the classes is the best picture Sep 14 02:29:21 i've been waiting 10 years since I saw them in the lab Sep 14 02:29:25 src/Makefile.am:1: directory should not contain `/' Sep 14 02:29:33 Writchie: the problem is that electronics (especially when imported) tend to be absurdly expensive here in .br (sometimes it gets about 3x the price!) Sep 14 02:29:50 yup Sep 14 02:30:03 you need to be in Asia Sep 14 02:30:25 Writchie: ah, automake... you have to use the ordering then (and you will need one makefile in A and one in B, calling _their_ subdirectories in order) Sep 14 02:30:27 china is much more advanced than US Sep 14 02:32:51 cesarb: i'm doing the ordering. The problem is that i have cross references between A/ and B/ Sep 14 02:32:57 lol, and our DVD player/recorder just CRASHED... and we had to reboot it :D Sep 14 02:33:02 A needs B and B needs A Sep 14 02:37:03 bedboi: cyclic dependencies are usually signs of a design problem somewhere... Sep 14 02:41:23 u Sep 14 02:41:27 tokesn Sep 14 02:46:08 the icon on the Dial Button really needs to be a bit bigger http://img513.imageshack.us/img513/9481/dialer1pq5.png Sep 14 02:47:52 cesarb: is not that are cyclic dependencies Sep 14 02:48:13 it's just a matter of organization of the code **** ENDING LOGGING AT Fri Sep 14 02:59:56 2007