**** BEGIN LOGGING AT Mon Sep 17 02:59:57 2007 Sep 17 04:02:41 * rtyler crosses fingers Sep 17 04:02:53 let's see if a fresh rebuild will work after a week of broken builds >_< Sep 17 04:07:36 good luck. Sep 17 04:17:52 ipkg can't update any more... / is full Sep 17 04:18:11 did I do something wrong? maybe too many updates all at once? Sep 17 04:21:02 maybe it's just time to reflash Sep 17 04:38:29 This constant OE breakage really can get on one's nerves fast... Sep 17 04:42:02 cesarb - sounds like it - i'll try later Sep 17 04:55:11 moin Sep 17 06:12:26 good morning Sep 17 06:45:26 <|R> /clear Sep 17 06:45:28 <|R> oops :) Sep 17 06:46:06 now we'll _all_ have to do /clear :( Sep 17 06:46:43 who spoke after |R? I just did a /clear and missed it Sep 17 06:47:36 how tykical Sep 17 06:55:14 mickey|zzZZzz: http://webcvs.freedesktop.org/gstreamer/gst-plugins-ugly/ext/mad/gstmad.c?hideattic=0&r1=1.150&r2=1.151 seems responsible for the 16->32 bit change in mad. Here's the reverse patch as reversing the original patch fails: http://abraxa.dyndns.org:81/random/gstmad_16bit.patch Sep 17 06:57:12 mickey|zzZZzz: I seem to be unable to run "filesrc ! mad ! pulsesink" because gstreamer hangs on "Setting pipeline to PAUSED" but I figure you might have more luck with that as I didn't get a successful devel image build for weeks now Sep 17 07:06:39 guten morgen :) Sep 17 07:06:49 re Sep 17 07:06:50 einen wunderschoenen.. Sep 17 07:07:29 mein devel ist endlich korrekt gebaut Sep 17 07:08:29 nur im Splash steht immer noch August :) Sep 17 07:08:37 RalphEichelberge: good to here - i used ipkg all the time and had no probs... :) Sep 17 07:08:45 s/here/hear/ Sep 17 07:08:45 emdete meant: RalphEichelberge: good to hear - i used ipkg all the time and had no probs... :) Sep 17 07:09:21 well, and I thougth you are german Sep 17 07:09:49 sry, i am... but perhaps the others want to read too ;) Sep 17 07:10:03 well you are right, of course Sep 17 07:10:23 so we will do it like the universities Sep 17 07:10:23 glaub ich nicht :p Sep 17 07:10:33 also i would prefere german because my english sucks... but... Sep 17 07:10:35 nicht alle, oder? Sep 17 07:11:02 no problem, emdete, we all suck somehow :) Sep 17 07:11:06 emdete: trust me, your english sucks much less than my german :) Sep 17 07:12:00 well something is wrong, my neo ist starting up since 10 minutes now :( Sep 17 07:12:01 cjb_ie: :D Sep 17 07:12:50 my best german phrase is nonsense from monty python: "Wenn ist das Nunstück git und slotermeier? Ja! Beierhund das Oder die Flipperwaldt gersputt!" Sep 17 07:13:56 cjb_ie: hm... Sep 17 07:14:14 told you it was nonsense :p Sep 17 07:14:58 right, arbeitszeit... catch you later Sep 17 07:15:41 RalphEichelberge: still booting? Sep 17 07:16:29 the problem seems to be, booting directly afer flashing Sep 17 07:16:52 after powering down and starting again it works fine :) Sep 17 07:17:09 - BUT Sep 17 07:17:18 there is no systemtray Sep 17 07:19:27 i first need to restart the xserver before i get the machbox-panel Sep 17 07:19:58 hm... strange. i don't use that gui anyway Sep 17 07:20:12 join #olpc Sep 17 07:21:23 rob_w|laptop: good idea Sep 17 07:21:52 emdete, yeah sorry Sep 17 07:22:32 emdete: what gui do you use than? Sep 17 07:22:37 then* Sep 17 07:22:39 my own... ;) Sep 17 07:22:52 really? - show me! Sep 17 07:23:38 :) it's plain simple. the data behind is more important and i am fihting with gsmd... not able to do sms/calls Sep 17 07:24:45 i see - i just found a bug Sep 17 07:24:53 ? Sep 17 07:25:09 do you know, when X server is starting up but it hangs with that noisy sound? Sep 17 07:25:25 oups... no. never had that. Sep 17 07:25:30 thats directly after it faild to find gsmd Sep 17 07:25:56 i use directfb for my gui but play around with x for edje and webkit.. never had a sound with x11 Sep 17 07:26:21 i am talking about the neo Sep 17 07:26:28 not a emulator Sep 17 07:26:28 me too Sep 17 07:26:32 yes Sep 17 07:26:32 ok Sep 17 07:26:51 directfb? is it on that image? Sep 17 07:26:57 i tried the emu but it was to slow here... on 1,8GHz Sep 17 07:27:13 no, not in the image but in oe Sep 17 07:27:30 just build it with `make build-package-directfb` Sep 17 07:27:43 well it automatically starts up using /etc/init.d/xserver-nodm Sep 17 07:28:12 yes. that's x11. but some other process with contact the gsmd... may that one issue a sound? Sep 17 07:29:10 no the thing is, it stalls while starting up having this startup sound Sep 17 07:29:20 and then BRRRRRRRRRRRRRRRR Sep 17 07:29:54 very strange. what does `ps x` say? Sep 17 07:29:56 ssh stops working - somehow the complete system hangs Sep 17 07:30:01 and whats in /tmp/gsm.log? Sep 17 07:30:07 lets see Sep 17 07:30:07 oupssss Sep 17 07:30:46 BRRRRRRRRRRRRRRRR? the vibrator? or sound? :D Sep 17 07:30:51 sound Sep 17 07:30:55 :) Sep 17 07:31:14 did you ever had you audio player stall? Sep 17 07:31:39 no, that worked fine, but again: madplay Sep 17 07:31:43 no gui Sep 17 07:32:37 my app plays mp3 in background but had no time to port that part for the neo Sep 17 07:33:07 s/had/i had/ Sep 17 07:33:07 emdete meant: my app plays mp3 in background but i had no time to port that part for the neo Sep 17 07:35:40 no i mean the effect when 'a' player stops, the soundcard makes a strange noise Sep 17 07:35:45 however Sep 17 07:35:58 ill try that other gui Sep 17 07:36:02 steady noise? Sep 17 07:36:09 yes Sep 17 07:36:16 strange, no, hadn't that Sep 17 07:36:57 shit gsmd - always dies with: "gsmd.c:82:alive_tmr_cb() modem dead!" Sep 17 07:37:16 i had that too with the latest build i found 9. sep Sep 17 07:37:28 aha Sep 17 07:37:36 u working on that demon? Sep 17 07:37:53 i try to use it. via libgsm Sep 17 07:38:08 but that api is very... hm. strange. Sep 17 07:40:14 i c Sep 17 07:42:43 RalphEichelberge: so what are u doing with the neo? Sep 17 07:44:07 my boss gave it to me with the words: "find out if we can do something with it" Sep 17 07:44:57 oh... may i ask which company? Sep 17 07:45:12 www.tourist-mobile.at Sep 17 07:46:24 directfb is building Sep 17 07:46:42 do u have a app for directfb? Sep 17 07:47:28 RalphEichelberge, so currently the product of the company is a sim card with cheap roaming? Sep 17 07:47:58 right# Sep 17 07:48:09 interesting concept already :-) Sep 17 07:48:24 and travel and neo's gps fit, right? :) Sep 17 07:48:34 is direct fb somehow a grahical display like X Sep 17 07:48:41 RalphEichelberge: yes Sep 17 07:48:48 faster, simple, less mem Sep 17 07:48:51 well, its new to me Sep 17 07:48:55 s/simple/simpler/ Sep 17 07:48:56 emdete meant: faster, simpler, less mem Sep 17 07:49:25 i should look how the splash screen on boot is implemented Sep 17 07:49:29 i didn't hear about yet - working on linux since 15 years now ;( Sep 17 07:49:31 * bedboi is away: sleeping Sep 17 07:49:32 but: i start to look for x because if neo gta2 arives with hw acceleration that will be supported by x early Sep 17 07:50:14 RalphEichelberge: directfb has a concept for hw acceleration but won't support that chip Sep 17 07:50:28 so better stay on X? Sep 17 07:51:04 depends on what you want to do. my app is so simple it works with directfb just fine Sep 17 07:51:23 but it is a lot of work doing more graphic Sep 17 07:51:30 so i looked for edje Sep 17 07:51:39 and now for a concept with webkit Sep 17 07:51:51 i think of using webkit for the whole gui Sep 17 07:52:56 but there is a port for gimp on directfb right? Sep 17 07:53:08 ...and have static pages local on the phone for alle the phone functions Sep 17 07:53:24 yes, gimp is ported and gtk, gdk, ... all the crap. Sep 17 07:53:45 gtk/gdk tends to be slow Sep 17 07:53:54 /and/ needs lots of mem Sep 17 07:54:45 well i see no need to switch for now Sep 17 07:55:22 you use directfb on you workstation? Sep 17 07:55:27 no Sep 17 07:56:31 direvtfb is used alot on settopboxes Sep 17 07:56:38 i am on directfb.org currently and see that they have real transparency - thats nice however :) Sep 17 07:57:09 it supports hw layers for viewing tv and having a menu window in front done by the hw Sep 17 07:57:31 that's the cause i choose dfb. i need it alot ;) Sep 17 07:58:12 so its more a multimedia thing Sep 17 07:59:31 that's the area where it is used alot. and: okay there is fusion and viewing videos included. but it's more a graphic base api on top of linux's fb Sep 17 08:02:20 if u switch your neo to directfb you are on your own... no app, no support. i would do that only if you want to replace the main gui as i am doing it here. Sep 17 08:11:10 morning Sep 17 08:11:20 hi rwhitby Sep 17 08:11:54 emdete: Ambitious project you're working on :) Sep 17 08:12:26 CM: depends... Sep 17 08:12:51 * rwhitby hears the new ringtone for the first time ... Sep 17 08:13:50 CM: it will stay very, very simple. and if i can show it perhaps some people start to help... Sep 17 08:14:03 *nod* Sep 17 08:14:10 emdete: Still cool though :) Sep 17 08:14:38 CM: what part do you mean: throwing away any other gui or using webkit as main gui? Sep 17 08:14:48 Webkit Sep 17 08:15:25 :D what do you think: how difficult is it to make the phone api (say libgsmd, ...) avaiable for js? Sep 17 08:17:06 I have no clue, but hopefully it's not too hard. Sep 17 08:17:43 emdete: I guess you've seen http://www.pyrodesktop.org ? Sep 17 08:18:12 ...no Sep 17 08:20:00 hm.. basically the same idea. so the idea may be not too bad. Sep 17 08:20:11 Just what I was thinking :-) Sep 17 08:20:44 * CM takes a coffee break Sep 17 08:20:51 * emdete joins Sep 17 08:20:57 joins Sep 17 08:21:27 Willkommen ;) Sep 17 08:22:15 * rwhitby also ran feedreader for the first time with an active bluetooth panu connection (through an Iomega NAS 100d NAS box) to the internet. Sep 17 08:23:20 so bluetooth pan networking should work nicely to an NSLU2 running Debian 4.0 Sep 17 08:24:38 meanwhile libwnck fails to build on Debian Sarge, reporting a missing gdk-pixbuf-csource Sep 17 08:24:43 ill start my hello world project :) Sep 17 08:25:33 make that fails on Debian Etch Sep 17 08:27:20 hello Sep 17 08:33:38 rwhitby: just encountered the same prob. any workarond? Sep 17 08:34:22 not that I know of yet. rebuilding that package doesn't work, I'm trying a rebuild of gtk+ now, in case it hasn't staged something properly. Sep 17 08:37:14 now trying installing libgtk2.0-dev on the host Sep 17 08:38:06 why are channel logs empty? it seems that there is a problem since 9 september Sep 17 08:40:04 Ghiottone|work: http://logs.nslu2-linux.org/livelogs/openmoko.txt and http://logs.nslu2-linux.org/livelogs/openmoko seem to be working fine - which logs are you looking at? Sep 17 08:41:09 ok, the work-around for the libwnck build failure is to install libgtk2.0-dev on the build host Sep 17 08:41:14 emdete: ^^ Sep 17 08:41:35 http://www.hentges.net/tmp/logs/irc/%23openmoko/2007/September/ Sep 17 08:41:52 I thought they were the only one. thanks rwhitby Sep 17 08:42:20 Ghiottone|work: that'll teach you to use second-rate logs from CoreDump|afk ;-) Sep 17 08:43:30 rwhitby: hm, but those are from unstable? Sep 17 08:43:47 which? Sep 17 08:44:11 libgtk2.0-dev? oh, no... hm. does not want to install Sep 17 08:44:26 rwhitby :) Sep 17 08:46:51 * rwhitby alerts openmoko-devel to the new build dependency Sep 17 08:49:41 rwhitby; morning! Im seeing my downloads in /sources getting deleted frequently - do you know why that would be happening? Sep 17 08:50:15 they should not be - how are you seeing that they are deleted? Sep 17 08:50:57 ls sources -> nothing. i even chmod 444 * to make them read-only but still they got deleted! Sep 17 08:51:06 a 'feature' of bitbake? Sep 17 08:52:55 anyone else seeing their sources directory getting cleaned out? Sep 17 08:53:56 Mine are all there Sep 17 08:55:43 ho hum another 500Mb download... Sep 17 08:56:27 jpcass: I have 4.2GB in my sources directory ... Sep 17 08:56:44 * CM has 1.4G Sep 17 08:57:14 jpcass: It's not anything that mokomakefile runs. Sep 17 09:04:13 Anyone with a neo p1 that has never experienced the modem dead stuff? Sep 17 09:06:25 CM, that pyro thing makes me think: "activedesktop" Sep 17 09:09:34 inz: Something like that. Not sure what to think of it yet Sep 17 09:14:14 sannes: a gsmd restart in /etc/matchbox/session seems to make things stable for me Sep 17 09:28:55 rwhitby: thanks, I will monitor closely and try to find out what is deleting things...gonna have to look into the bitbake fetcher I guess... Sep 17 09:42:00 someone working on java for moko? Sep 17 09:44:25 RalphEichelberge: Yes, there's some people on it. No idea how they're doing though Sep 17 09:47:45 thx CM - just courious :) Sep 17 09:48:19 RalphEichelberge: sun showed java on the neo already Sep 17 09:52:59 emdete: wasn't that just a 'shopped stock image of the Neo that they were using for publicity? Sep 17 09:54:03 hm, as far as i remember they showed java on it which crashed... Sep 17 09:57:29 baird: you are right. they show the software on a nokia... :/ Sep 17 09:58:11 roh: ping Sep 17 09:59:36 oh man! java on pda and mobile is a pain in the ass Sep 17 10:00:14 java on *FIXME: replace with anything you like* is a pain in the ass. Sep 17 10:00:19 just ask my ass Sep 17 10:00:28 he brays every time. Sep 17 10:00:42 java == overhyped, underdeveloped Sep 17 10:00:45 i see :D Sep 17 10:01:34 everything in java i have ever seen is fat, bloated, leaks like a seieve and is slow. Sep 17 10:01:40 people claim j2me is fast Sep 17 10:01:55 but then people claim many things about java Sep 17 10:02:08 a java runtime Sep 17 10:02:10 even j2me Sep 17 10:02:11 is not small Sep 17 10:02:27 raster: My belief is, when some one shows me a "fast" java app, I'll show them some thing that does very little Sep 17 10:02:36 hahahaha Sep 17 10:02:50 by sheer definition java will always be slower Sep 17 10:02:53 as even with a jit Sep 17 10:02:57 u need to compile on the fly Sep 17 10:03:04 so apps that quickly execute, do their thnig and exit Sep 17 10:03:07 Like it or not, J2ME is the lingua franca of mobile handsets. Sep 17 10:03:11 will always be much slwoer than equivalent c/c++ Sep 17 10:03:22 if its a long-lived app Sep 17 10:03:25 the cost is much less Sep 17 10:03:31 raster: This is very very true, but on the flip side, it DOES off some thing that is important. Sep 17 10:03:37 but then u h ave that wonderful garbage collector Sep 17 10:03:49 Software developers, to write some thing generic that will run on a large variety of handsets Sep 17 10:04:06 and all those java coders thaqt suddenly believe memory is no longer somethnig to worry about - the garbage collector is mean to do that for them! Sep 17 10:04:27 and that is true Sep 17 10:04:34 tho the moment u want to get imaginative Sep 17 10:04:41 and access features Sep 17 10:04:45 ...if that would only be memory...they often do not even close files and/or sockets :/ Sep 17 10:04:47 u end up using handset specific classes Sep 17 10:05:00 Whatever. There is a huge load of J2ME apps/applets out there. Not supporting J2ME would mean not being able to run all those apps. Sep 17 10:05:00 emdete: its the mindset Sep 17 10:05:06 once u become lazy Sep 17 10:05:10 it permeates all areas Sep 17 10:05:19 raster: :D Sep 17 10:05:37 Lynet_: not saying it shouldnt have support Sep 17 10:05:50 but i certainly would discourage its use Sep 17 10:05:54 for any new developers Sep 17 10:06:00 jumping on the bandwagon Sep 17 10:06:16 yes, and if java would work really on every platform the same way - that would be the only thing that it would have the right to exist Sep 17 10:06:27 raster: Ah, sorry. No argument there, I agree that Java tends to be slow bloatware. Sep 17 10:06:47 agreed - lots of stuff u want to have "just work" Sep 17 10:06:52 and let the user make the chocie of the cost Sep 17 10:06:53 ie Sep 17 10:06:57 but everyone who tryed it knows that thats just a dream! Sep 17 10:06:59 they know they need to use up X mb of flash Sep 17 10:07:02 and so on Sep 17 10:07:31 Personally, I would choose a native app over a java app every time, but some times, a native app just isn't available. Sep 17 10:07:32 i'd love to see a palm emulator work too Sep 17 10:07:37 hell Sep 17 10:07:41 an arm version of wine Sep 17 10:07:53 cut down to emulate wince/pocketpc/windows mobile Sep 17 10:07:59 etc. Sep 17 10:08:08 I also think that java shouldn't be overly criticised for lazy programmers (though the fact that they CAN be lazy is an issue). Sep 17 10:11:02 and easy languages attract bad developer... Sep 17 10:11:16 rwhitby: ah, doesn't for me, well sometimes it does, but it seems that it might actually bad connection between sim and chip, and I was wondering if such a thing could make it misbehave in that kind of way .. Sep 17 10:12:10 Sep 17 10:15:19 anyone know if bitbake logs its fetch actions somewhere? there are no do_fetch logs in build/tmp/work directories.... Sep 17 10:15:22 what about gsmd - my latest build (update at 15.9.2007) can't find a gsm-network still Sep 17 10:16:01 RalphEichelberge: what does /tmp/gsm.log say? Sep 17 10:16:13 RalphEichelberge: restart gsmd - sometimes it works Sep 17 10:16:20 i found the workaround in the init-script of gsmd allready in the project Sep 17 10:17:00 what happens, when you click on that gsm icon and power up oder down the modem? Sep 17 10:17:19 AT+CFUN=0/1 as far as i know Sep 17 10:17:32 just the "echo 1(or 0) .... into that dbus file? Sep 17 10:17:44 RalphEichelberge: dbus? Sep 17 10:19:00 dbus is not file based Sep 17 10:19:10 neither does gsmd use dbuds Sep 17 10:19:51 the icon probably goes through libgsmd-tool Sep 17 10:20:11 root@fic-gta01:~$ echo "1" > /sys/bus/platform/devices/gta01-pm-bt.0/power_on Sep 17 10:20:19 and if it issues a power-off it sends a AT+CFUN=0 Sep 17 10:20:33 RalphEichelberge: no, that is not done as far as i know Sep 17 10:20:34 what is this? Sep 17 10:20:56 look into /tmp/gsm.log Sep 17 10:25:12 there is a way to talk to the modem? Sep 17 10:25:23 cu Sep 17 10:26:12 The power_on button is only a virtual power on button Sep 17 10:26:21 a 0-1 edge soft-resets the modem Sep 17 10:26:34 0 or 1 state alone does nothing Sep 17 10:26:43 RalphEichelberge: i recommend you to read the wiki Sep 17 10:26:53 and fyi, this is not the wiki Sep 17 10:27:17 It is, in a way. Sep 17 10:27:36 ok Sep 17 10:27:38 formatting sucks though. Sep 17 10:28:05 you are the living wiki - hu? :) Sep 17 10:30:47 hi...i have problems to build 'libwnck'. How can I solve this? Sep 17 10:31:14 Depends on what the problems are, doesn't it. Sep 17 10:31:24 That's like saying "my car doesn't work, what's wrong?" Sep 17 10:32:29 Placid: but my card doesn't work! Sep 17 10:32:32 help! Sep 17 10:32:34 :) Sep 17 10:33:22 lol Sep 17 10:34:46 so look here: http://pastebin.ca/700548 Sep 17 10:36:17 configure: error: gdk-pixbuf-csource executable not found in your path - should be installed with GTK Sep 17 10:36:21 Did you read the errors at all? Sep 17 10:36:31 Is GTK installed properly? Sep 17 10:40:03 I think so...I only want to update and rebuild my rootfs with the MokoMakefile Sep 17 10:40:48 i cant find a wiki about talking to the modem using cu Sep 17 10:40:48 shouldnt it be "cu /dev/ttyS...." ? Sep 17 10:42:45 cu -l /dev/ttyACM0 Sep 17 10:42:46 err Sep 17 10:42:51 cu -l /dev/ttySAC0 Sep 17 10:43:06 ~~. to exit Sep 17 10:43:22 what are you trying to do/ Sep 17 10:43:43 just playing around with the hardware for now Sep 17 10:43:50 get the feel :) Sep 17 10:43:56 at@poff turns the modem off Sep 17 10:45:19 aha Sep 17 10:46:02 well i have stopped gsmd - but still get "root@fic-gta01:/var/volatile/log$ cu -l /dev/ttySAC0 Sep 17 10:46:02 cu: open (/dev/ttySAC0): Permission denied Sep 17 10:46:02 cu: /dev/ttySAC0: Line in use Sep 17 10:46:02 " Sep 17 10:46:22 why? Sep 17 10:47:28 chown uucp.uucp /dev/ttySAC0 Sep 17 10:47:39 or you have killed gsmd? Sep 17 10:47:56 thx - that does the trick :) Sep 17 10:52:51 Hey All, quick question, what do we use for OpenMoko's init system, is it just traditional sysv? Sep 17 10:55:37 Okay, from the wiki, it looks sysvinit is the one in use. Sep 17 10:55:41 where is the config file for sysv? Sep 17 10:55:57 Is any one looking at the upstart stuff? Sep 17 10:56:23 upstart migt be a bit overkill. but minit or cinit schould be fine Sep 17 10:57:12 i was eger to look how much of a difference a better init would make. but i have no time atm Sep 17 10:57:22 if I use "cu -l /dev/ttySA0" and type in "AT" RETURN it should give me "OK" dont it? Sep 17 10:57:24 DPThough1: Fair enough. Sep 17 10:57:49 DPThough1: The tight intergration between dbus and hal with upstart looked like some thing that I'd have thought would be good with the openmoko Sep 17 10:58:00 on the other side. upstart provides a dynamic way to launch services. Could be usefull for us too Sep 17 10:58:10 wibbit: ;) Sep 17 10:58:18 hehe Sep 17 10:58:23 Exactly what I was thinking Sep 17 11:02:18 Is any one in here that has been working on the "call profiles" in Open Moko? Sep 17 11:02:37 Placid: I have installed GTK and pixbuf is also installed. Sep 17 11:02:50 so what can I do? Sep 17 11:10:35 anyone heard or got plans about python-gconf or python-webkitgtk? Sep 17 11:18:09 openmoko: 03thomas * r2978 10/trunk/src/target/OM-2007.2/libraries/libmokoui2/ (5 files in 2 dirs): Sep 17 11:18:09 openmoko: * configure.ac: Sep 17 11:18:09 openmoko: * libmokoui/Makefile.am: Sep 17 11:18:09 openmoko: * libmokoui/moko-hint-entry.c: Sep 17 11:18:09 openmoko: * libmokoui/moko-hint-entry.h: Sep 17 11:18:10 openmoko: Add MokoHintEntry based on KotoHintEntry Sep 17 11:18:43 Am I right in understanding, that the current call profiles will allow us to specific numbers to forward calls on to? Sep 17 11:24:41 \help Sep 17 11:25:30 hey tuukkah Sep 17 11:25:49 hi Ghiottone|work! Sep 17 11:30:24 uhm, don't you think the games on the new moko images impare the developer's productivity? ;) Sep 17 11:30:47 they're sort of addictive Sep 17 11:30:58 hehehe Sep 17 11:31:11 what games? Sep 17 11:31:54 net, the one with 4 color problem and so on Sep 17 11:45:24 openmoko: 03thomas * r2979 10/trunk/src/target/OM-2007.2/applications/openmoko-today2/ (. libtaku/): * Switch to using the upstream version of libtaku via svn:externals Sep 17 11:54:54 /hello people Sep 17 11:57:06 * Placid waves Sep 17 13:07:48 configure:21316: ccache arm-angstrom-linux-gnueabi-gcc -march=armv4t -mtune=arm920t -o conftest -isystem/home/j/hak/openm Sep 17 13:07:52 oko/build/tmp/staging/arm-angstrom-linux-gnueabi/include -fexpensive-optimizations -fomit-frame-pointer -frename-register Sep 17 13:07:55 s -Os -I/usr/include/SDL -D_GNU_SOURCE=1 -D_REENTRANT -isystem/home/j/hak/openmoko/build/tmp/staging/arm-angstrom-linux-g Sep 17 13:07:58 nueabi/include -L/home/j/hak/openmoko/build/tmp/staging/arm-angstrom-linux-gnueabi/lib -Wl,-rpath-link,/home/j/hak/openmo Sep 17 13:08:01 ko/build/tmp/staging/arm-angstrom-linux-gnueabi/lib -Wl,-O1 conftest.c -lGUI -L/usr/lib -lSDL >&5 Sep 17 13:08:26 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/SDLi think its -isystem Sep 17 13:12:56 what do you try to build? Sep 17 13:56:37 openmoko: 03andrew * r2980 10/trunk/src/host/qemu-neo1973/ (gnokiigsm/at-emulator.c hw/modem.c): Emit the AT greeting on every interpreter reset. Sep 17 14:11:15 anyone tried mmcinit with qemu-neo1973 and got size != 0 ? Sep 17 14:11:31 in u-boot over (emulated) serial line Sep 17 14:26:22 AMD promised to open all their GPU specs, but did that include embedded graphics? Would be nice.. http://www.theinquirer.net/default.aspx?article=42392 Sep 17 14:26:43 CM: no Sep 17 14:26:53 XorA: Shame :( Sep 17 14:27:11 HTC Trinity 4 Ever ^^ Sep 17 14:27:22 well, better then nothing Sep 17 14:27:24 CM: personally I kind of doubt a) AMD actually hold the rights to those chips and b) whether the guys who had the docs even work there anymore Sep 17 14:28:01 XorA: *nod* Sep 17 14:28:20 sry, a friend has written this bullshit! Sep 17 14:28:44 RedDog: Ex friend? :P Sep 17 14:29:17 hmm ... died on 17.09.2007 ... ;-) Sep 17 14:29:52 Well, htc is fic too and they do have some nice hw Sep 17 14:40:33 exit Sep 17 14:40:35 bleh Sep 17 14:40:42 sorry ^^ Sep 17 14:52:14 openmoko: 03andrew * r2981 10/trunk/src/host/qemu-neo1973/ (347 files in 25 dirs): Pull changes from cvs.savannah.nongnu.org:/sources/qemu. Sep 17 14:58:38 hi! Sep 17 14:59:44 did an make update and make openmoko-devel-image and Sep 17 14:59:54 what the hell is this? --x Sep 17 14:59:59 NOTE: package libspiff-0.7.2-r0: task do_qa_staging: started Sep 17 15:00:00 NOTE: QA checking staging Sep 17 15:00:00 ERROR: QA issue: libsystray.la failed sanity test (reference to workdir or installed) Sep 17 15:00:00 ERROR: QA issue: liblauncher.la failed sanity test (reference to workdir or installed) Sep 17 15:00:00 .... Sep 17 15:00:13 sanity test? Sep 17 15:00:48 It always does sanity tests, but I don't know why they fail some times Sep 17 15:00:56 dont say "just make clobber" - that means days of compiling again Sep 17 15:01:14 Did the build stop at that? Sep 17 15:01:23 it? Sep 17 15:01:29 I just completed a build from a fresh update a few hours ago Sep 17 15:01:48 well that what i did too Sep 17 15:03:18 after the first start - it started up just fine, but that panel just wasn't here again - stopped the xserver (/etc/init.d/xserver-nodm stop) Sep 17 15:04:00 after /etc/init.d/gsmd restart - the neo was dead Sep 17 15:05:03 ok so shutdown the device and switch it on again - perfect every thing fine Sep 17 15:05:06 WHY? Sep 17 15:06:38 RalphEichelberge: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=788 Sep 17 15:08:57 hallo Sep 17 15:32:13 re Sep 17 15:40:25 * bedboi is back (gone 07:50:43) Sep 17 15:44:12 the mokomakefile fails to build webkit for me, is there a known workaround? Sep 17 15:47:47 jas4711, yup, changing SRCREV_pn-webkit-gtk to "25582" Sep 17 15:47:48 in /openembedded/conf/distro/include/sane-srcrevs.inc works Sep 17 15:48:58 Sup3rkiddo: hm, the file already contains that value Sep 17 15:49:21 how can i clean webkit and retry building it? Sep 17 15:49:31 jas4711, make clean-package- Sep 17 15:49:43 or..make rebuild-package- Sep 17 15:50:13 Sup3rkiddo: tried that, but 'No providers of build target webkit'. is the webkit package called something else? Sep 17 15:50:51 jas4711, yeah, webkit-gtk Sep 17 15:50:54 ok it is webkit-gtk Sep 17 15:51:13 trying to build again now Sep 17 15:51:14 thanks Sep 17 15:57:28 openmoko: 03mickey * r2982 10/trunk/src/target/OM-2007.2/artwork/sounds/touchscreen_click.wav: OM 2007.2: make touchscreen wav less intrusive (compromise) Sep 17 16:02:58 <|R> I was wondering, does anyone know if iGo Stowaway bluetooth keyboard will work with GTA01 / 02? It would make a perfect ssh machine ;-) Sep 17 16:07:20 rebuilding wekit-gtk failed again, with the same error. 'undefined reference to `operator delete(void*)@GLIBCXX_3.4''' while linking testkjs. ideas? Sep 17 16:08:46 i suspect the problem is the icu package, cleaning and rebuilding... Sep 17 16:46:59 so I just flashed a new build (that I did last night), and it looks like gsmd is not starting Sep 17 16:49:03 hrrm.. okay, I restarted gsmd and xserver-nodm Sep 17 16:54:27 whooohoo, I made a call Sep 17 16:55:27 :D Sep 17 16:55:40 ugh i gota get my build envirnment setup properly so I can start getting to work! Sep 17 16:59:02 now how the heck to a make a gsm data connection ;) Sep 17 16:59:18 See [[Using GPRS manually]] Sep 17 16:59:21 Or similar Sep 17 17:00:16 there's no real obvious speakerphone button either Sep 17 17:00:39 aux is the obvious one :) Sep 17 17:00:51 SpeedEvil: do you grok how the gsmd/gprs muxing is supposed to work eventually? Sep 17 17:02:03 That's how it's supposed to work. Sep 17 17:02:07 Eventually. Sep 17 17:03:26 well, it seems like the call history doesn't work ;) Sep 17 17:03:26 eventually... Sep 17 17:03:28 one day Sep 17 17:03:32 some day! Sep 17 17:03:40 call history used to work Sep 17 17:04:57 and if I select "call history - received" with my finger, and then try to go back to "call history- all", I can't get the dialog to scroll up Sep 17 17:05:09 I can do it with the stylus, but not finger Sep 17 17:05:28 sharpen your finger Sep 17 17:05:41 a large pencil sharpener or knife will probably work Sep 17 17:06:14 :P Sep 17 17:13:55 at the mo, call history does not work for me Sep 17 17:15:16 hello Sep 17 17:20:15 hey everyone, happened around a weird bug, for some packages the build system proceeds to do_install before compiling and eventually spits out a "No rule to make target: install" error..however all is well, if i clean the said paakage..anyone gettng the same thing? Sep 17 17:33:04 openmoko: 03andrew * r2983 10/trunk/src/host/qemu-neo1973/hw/usb-hid.c: Pass correct pointer to HID keyboard event handler, fixes regression from IDLE mode introduction. Sep 17 17:41:40 Hey, just being reading about openmoko, thought i'd give it a try on qemu, i got the mokomakefile and typed make qemu i get this error: U-boot failed to finish writing in 300 seconds, giving up. Sep 17 17:44:37 hmm.. I got the same error when I just gave it a try... Sep 17 17:45:16 Mek, do you think it is the mokomakefile or the uboot? Sep 17 17:45:26 Sup3rkiddo: did you update the makefile? Sep 17 17:45:45 torpor, yeah i did Sep 17 17:45:48 webjames: no idea.... I'm as new with this as you are... Sep 17 17:46:07 http://dpaste.com/19877/ this is the message i get Sep 17 17:46:31 torpor, a more detailed description if you would like..http://lists.openmoko.org/pipermail/openmoko-devel/2007-September/001535.html Sep 17 17:46:53 is there an irc client for openmoko? Sep 17 17:47:14 I have gprs working and I want to be obnoxious and irc from the phoen ;) Sep 17 17:48:05 I think that there is the package for xchat Sep 17 17:48:31 Mek, type make run-qemu Sep 17 17:48:55 hmm, still an error but at least it kinda boots Sep 17 17:49:20 Mek ill be back in a hour or two message me if you find an answer Sep 17 17:49:33 ok Sep 17 18:08:58 hozer: irssi in the terminal? :) Sep 17 18:23:39 mickeyl: how can i build an kernel modul for the squashfs?pls. point me in the right direction. Sep 17 18:25:50 i have no idea. if i were you i'd theck the squashfs homepage and see whether you need to patch the kernel or whether you can build it standalone Sep 17 18:26:03 if the latter, see one of the external module bb recipes in OE Sep 17 18:26:15 if the first, then just add the patch to SRC_URI and adjust the defconfig Sep 17 18:26:18 boom - it's done :D Sep 17 18:27:54 mickeyl++; Sep 17 18:28:53 mickeyl: i have made all this steps, but no success, s*** Sep 17 18:32:44 mmh Sep 17 18:32:50 what do you need squashs for anyways? Sep 17 18:33:37 for the german high compressed wikipedia Sep 17 18:34:00 a offline wikipedia Sep 17 18:34:02 hmm Sep 17 18:34:05 i see Sep 17 18:34:09 cool Sep 17 18:34:11 fgau: site your source! Sep 17 18:34:11 a worthwhile wish Sep 17 18:34:18 it would be really cool to have it on the neo :) Sep 17 18:34:19 fgau: Are you using mysql with it? Sep 17 18:34:21 or bunzip? Sep 17 18:34:28 err....whats your method, I have seen a couple out there... Sep 17 18:35:35 fgau: ? Sep 17 18:35:41 bkruse: no not sql, 700.000 compressed files Sep 17 18:35:55 apparently the method is to use squashfs to compress it transparently. :) Sep 17 18:35:58 fgau: what compression? what are you using to open them? how much disk space does it take? Sep 17 18:36:11 dkr: squashfs to compress the bz2? wow :] Sep 17 18:36:53 bkruse: it's not my project, the german mokopedia is compressd to 1GB Sep 17 18:37:07 fgau: wow, do you know about en? Sep 17 18:37:19 well, en is about 200 million articles, vs german 700,000 Sep 17 18:37:23 so much for that.... lol Sep 17 18:37:27 mickeyl: http://www.squashfs-lzma.org/ Sep 17 18:37:31 nearly 3.5 GB compressd Sep 17 18:37:53 fgau: yes, however, I wonder if you can access that data on the fly, being compressed Sep 17 18:37:57 to serve up in a webpage...... Sep 17 18:37:58 * bkruse thinks Sep 17 18:38:11 If I port it to something else, I will port it for the moko also Sep 17 18:39:37 bkruse: see http://wiki.openmoko.org/wiki/Mokopedia Sep 17 18:39:52 josch is the head of mokopedia :) Sep 17 18:40:11 I need to get started on mokoiax Sep 17 18:40:18 fgau: is he doing the EN port also? Sep 17 18:40:26 yes Sep 17 18:40:40 * bkruse sends him a polite message :] Sep 17 18:40:47 :) Sep 17 18:41:11 cool idea Sep 17 18:41:18 dkr: very Sep 17 18:41:31 people have taken it to the extreme being able to parse the bz2 on the fly per article Sep 17 18:41:35 via some crazy indexing.... Sep 17 18:42:01 reminds me of scifi book I just finished where they cache data for local searches when cut-off from net. (Rainbows End by Vernor Vinge :) Sep 17 18:42:01 anyways, I am going to do that also, I will have to talk to josch Sep 17 18:42:29 bkruse: you are german? Sep 17 18:42:37 nah Sep 17 18:42:41 I can still write code, however :] Sep 17 18:42:59 we are in #neo1973-germany Sep 17 18:43:04 kk Sep 17 18:50:10 Anyone dev on a windows box? Sep 17 18:50:13 I am just curious.... Sep 17 18:53:03 hmm Sep 17 18:53:06 * mickeyl CONFIG_SQUASHFS=m Sep 17 18:53:09 bbnew Sep 17 18:53:41 hello! Sep 17 18:54:09 is the guy working on XSPF integration in here? anybody know whom to talk to about that? Sep 17 18:54:09 sping: 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 17 18:55:51 abraxa_: found you name related to that - are you there? Sep 17 18:57:07 sping: He's usually around during the day Sep 17 18:57:31 CM: which timezone? Sep 17 18:57:45 I think CET, Europe that is Sep 17 18:57:45 CM: good question, i have 20:55 here :-) Sep 17 18:57:50 Same here Sep 17 18:57:52 :-) Sep 17 18:58:20 bkruse: Yes, some do, but then running linux in vmware ;) Sep 17 18:58:54 CM: ahh! Sep 17 18:58:55 meh, why does my neo insist on rebooting suddenly :/ Sep 17 18:58:58 hacks! ;] Sep 17 18:59:11 do you guys that dev run your image in qemu, or dev away from the image and cross compile for your native Sep 17 18:59:23 (specfically gtk+ based apps) Sep 17 19:02:06 bkruse: I haven't done any developing on my own yet, but people seem to work nativly first, then just recompile and try in qemu or on the neo Sep 17 19:02:29 CM: I was about to say, trying a quick code fix that would take hours Sep 17 19:02:36 of course, writing a gtk app, you want to see it in its envirnment :/ Sep 17 19:03:22 why would it take hours? Sep 17 19:03:42 bitbake the program, then just copy the ipk over and install Sep 17 19:03:49 CM: not hours, my apology, I meant of overall time Sep 17 19:03:55 Ah, sure :-) Sep 17 19:04:06 But building the rootfs from scratch sure takes hours.. *sigh* Sep 17 19:05:07 are there known problems with building on a 64bit OS? Sep 17 19:05:21 guaqua: right Sep 17 19:05:30 CM: thats what I was referring to, but I will get a decent bitbake/dev setup to make it as quick as possible Sep 17 19:05:58 bkruse: it was a quiestion Sep 17 19:06:26 guaqua: I've been building on 64-bit only from the start Sep 17 19:06:36 CM: oh, I thought you meant with cross compiling Sep 17 19:06:40 CM: okay, nm then Sep 17 19:06:43 Some small problems every now and then, but mostly it's worked every day Sep 17 19:06:50 no, it wouldnt, because gcc doesnt care (as a binary compiling other things) Sep 17 19:07:00 bkruse: Yes, crosscompiling with the MokoMakefile or bitbake Sep 17 19:07:43 bkruse: The problem is that it does care, since sometimes building the cross-build tools themselves use system libs Sep 17 19:08:12 CM: gcc should take care of all that if specified right? Sep 17 19:09:10 bkruse: Yes, but not for runtime dependencies. Anyway, it almost always works :D Sep 17 19:09:37 CM: nono, you make a good point, your right Sep 17 19:10:31 Hmm.. My neo has been unplugged and powered off during the weekend, and now it's dead.. Sep 17 19:10:41 * CM puts the usb cable in and waits an hour Sep 17 19:11:43 lol Sep 17 19:14:43 anyone know where bitbake logs its "do_fetch" commands, and where general progress/errors are logged Sep 17 19:14:56 (i know specific commands are logged in the work dir - but do_fetch is not...) Sep 17 19:16:11 jpcass: Maybe someone in #oe can give you a hint on that one? :) Sep 17 19:16:29 CM: thanks...let me try... Sep 17 19:26:29 hmm Sep 17 19:26:30 no go Sep 17 19:26:32 | CC [M] fs/squashfs/inode.o Sep 17 19:26:32 | fs/squashfs/inode.c: In function 'init_once': Sep 17 19:26:32 | fs/squashfs/inode.c:2301: error: 'SLAB_CTOR_VERIFY' undeclared (first use in this function) Sep 17 19:26:32 | fs/squashfs/inode.c:2301: error: (Each undeclared identifier is reported only once Sep 17 19:26:33 | fs/squashfs/inode.c:2301: error: for each function it appears in.) Sep 17 19:26:35 | fs/squashfs/inode.c:2301: error: 'SLAB_CTOR_CONSTRUCTOR' undeclared (first use in this function) Sep 17 19:27:14 that sounded bad; hope it was mounted readonly Sep 17 19:27:32 sorry, i don't have time looking into that. but at least the patch applied. once that gets fixed, it can be included in the kernel build Sep 17 19:32:42 mickeyl: To repeat jpcass' question earlier, is do_fetch commands logged somewhere when using bitbake? Sep 17 19:33:43 no, i'm afraid the fetcher doesn't spit out logs Sep 17 19:33:57 it spits out directly to the console Sep 17 19:34:05 err, outputs, even Sep 17 19:34:54 ok, thanks Sep 17 19:34:57 jpcass: got that? :) Sep 17 19:36:23 what's the easiest way to go about changing the kernel config used to build an openmoko kernel? Sep 17 19:37:29 use make menuconfig to change the kernel in the source dir Sep 17 19:37:48 then copy it over the defconfig in the build dir Sep 17 19:37:51 rebuild Sep 17 19:38:15 does it end up mattering that there's a pile of patches that need to be added in? Sep 17 19:40:46 from within the .bb? Sep 17 19:40:55 don't think so, should just work Sep 17 19:41:03 you might also manually enable what you want in the defconfig Sep 17 19:41:06 but take care Sep 17 19:41:16 i always use make menuconfig and then diff Sep 17 19:41:17 ok, to be honest, i'm still not sure how the kernel is getting built Sep 17 19:42:06 i kinda wish I could get bitbake to just spit out the entire list of commands it plans on doing, and where its getting each step from Sep 17 19:44:45 is it just be or is the OE/OM tree hosed right now? Sep 17 19:44:58 must be you Sep 17 19:45:02 why? Sep 17 19:45:15 build failure, let me clobber and retry Sep 17 19:45:53 builds slow, i should use the 8-way machines Sep 17 19:46:06 * cw666 grumbles Sep 17 19:46:24 well, as you can see in http://buildhost.openmoko.org/OM2007.2/tmp/deploy/glibc/images/fic-gta01/, i'm doing regular builds. hardly fails here Sep 17 19:47:11 mickeyl: actually, that probably makes more sense right now, you have one <8hrs old? Sep 17 19:47:14 * cw666 checks Sep 17 19:49:24 mickeyl: how can i tell how recent things are? i want to know if this is post revision 2724 Sep 17 19:49:37 (the timestamps would suggest it is) Sep 17 19:49:53 iiirc i built just a couple of minutes ago Sep 17 19:49:58 however we no longer build HEAD Sep 17 19:50:02 orpheus: bitbake -g openmoko-devel-image and then example *.dot Sep 17 19:50:06 cw666: I have one that's 11 hours ago that worked fine Sep 17 19:51:57 although there are some other experiences, OpenEmbedded builds are not crashing every minute Sep 17 19:53:41 mickeyl: ok, so if it's not trunk you're building how can i find the revision/tag you built so i can check something? Sep 17 19:54:05 the source revisions are encoded in the packages Sep 17 19:54:10 e.g. see ipkg Sep 17 19:54:12 directory Sep 17 19:54:32 i'm always building w/ OE HEAD though, if that's what you mean Sep 17 19:54:38 org.oe.dev, that is Sep 17 19:54:57 there were toms egsmd fixes in the last day, i'm wanting to retest with those Sep 17 19:55:04 i think one of them is even yours Sep 17 19:55:18 hmm gsmd fixes? Sep 17 19:55:28 anyone know what the problem is with the libsdl-net configure stage? why is it picking up /usr/include in the configure step? Sep 17 19:56:05 cw666: gsmd has 2983 Sep 17 19:56:06 torpor: me having same problem Sep 17 19:56:16 Writchie: doh! Sep 17 19:56:17 cw666: compare w/ what's lying in ipkg Sep 17 19:57:01 torpor: can you analyse which autotools test drags /usr/include in? Sep 17 19:57:03 mickeyl: that should be fine, btw, is there are recent uboot somewhere? Sep 17 19:57:15 i can launch a build for you Sep 17 19:57:19 mickeyl: from the config.log you mean? Sep 17 19:57:32 mickeyl: i tried make rebuild-package-uboot and failed here Sep 17 19:57:39 mickeyl: what magic incarnation should i be using? Sep 17 19:57:41 torpor: yes, from a combination of looking at config.log, Makefile and Makefile.am you should be able to deduce the offender Sep 17 19:58:04 cw666: don't know about magic, but bitbake uboot-openmoko works fine for me here Sep 17 19:58:05 mickeyl: okay i'll try a bit of a grok Sep 17 19:58:49 sping: I'm here now, what's up? Sep 17 19:58:57 cw666: how does your build break? do you actually try to look into the cause or just give up when it doesn't work? Sep 17 19:59:10 cw666: if the first, we could try to help Sep 17 20:00:48 mickeyl: i'm rebuidling now, at the time i had to run out so i did a clobber and and restarted it then ran out the door Sep 17 20:01:14 abraxa_: awesome, hello! Sep 17 20:01:22 ok. note that i don't know much about mokomakefile though..., so i don't know 100% what all these fancy things are doing :D Sep 17 20:01:33 i'm one of those crazy guys using bitbake commands directly Sep 17 20:01:34 mickeyl: Did you see my message about mad's 16->32 bit patch? Sep 17 20:01:42 abraxa_: no, must have missed that one. Sep 17 20:01:51 i noticed you wrote an enhance libSpiff C wrapper for OpenMoko Sep 17 20:02:06 what is libspiff? Sep 17 20:02:17 torpor: XSPF support library Sep 17 20:02:38 Tronic: xspf.org - playlist format of tomorrow :-) Sep 17 20:02:44 ah i see. Sep 17 20:02:46 cool! Sep 17 20:03:12 Tronic: sorry, that was for torpor :-) Sep 17 20:03:37 abraxa_: first way cool to see this! Sep 17 20:04:17 abraxa_: i was wondering in what way you are using libspiff and what you needed more in a wrapper (which is a good sign to me, not a bad one) Sep 17 20:05:06 abraxa_: did you succeed w/ that patch? Sep 17 20:05:31 mickeyl: The patch that changed mad from 16 to 32 bit can't be reversed properly so I reversed it by hand and made a patch: http://abraxa.dyndns.org:81/random/gstmad_16bit.patch - it seems to work fine according to gst-inspect, though I can't test it properly as my image is old and I didn't have time to flash a new gstreamer yet Sep 17 20:06:00 abraxa_: and be sure to update to libSpiff 0.8.0 released a few minutes ago Sep 17 20:06:01 mickeyl: I'm going to flash and test again asap Sep 17 20:06:11 abraxa_: pretty cool. keep me posted Sep 17 20:06:42 sping: Are you related to XSPF in some way? Sep 17 20:07:16 abraxa_: yes, libSpiff devel, online validator author and a few minor things :-) Sep 17 20:07:21 sping: And yeah, I wrote the wrapper because I need access to the track's "author" field which the simple wrapper doesn't permit Sep 17 20:07:30 sping: Ohhh, awesome :D Sep 17 20:10:05 abraxa_: on the phone, please wait a sec Sep 17 20:10:22 Take your time Sep 17 20:19:14 abraxa_: back Sep 17 20:19:55 abraxa_: author, i see Sep 17 20:20:48 abraxa_: what's the current that of XSPF support in OpenMoko? Is there a release with it already? Sep 17 20:23:30 abraxa_: i'm also talking to EdSchouten right now, the author of spiff_c - he's just about adding a creator field Sep 17 20:24:27 abraxa_: i think anything from Ed and me adding your wishes to spiff_c/libSpiff to you gaining svn access would be possible Sep 17 20:25:03 wow - quite busy :) Sep 17 20:25:28 as some of you may know, openmoko uses libspiff for XSPF parsing Sep 17 20:25:43 sping: Yeah, the current player uses XSPF already Sep 17 20:25:47 EdSchouten: It's almost only abraxa_ who knows ;) Sep 17 20:25:58 Heya EdSchouten :) Sep 17 20:26:10 i am one of the commiters on the libspiff project and i just merged back one of the features that openmoko needed, which caused the project to fork the c bindings :) Sep 17 20:26:15 EdSchouten: hey :-) Sep 17 20:26:25 abraxa_: it looks like you need the following stuff: Sep 17 20:26:29 - support for the creator field Sep 17 20:26:38 EdSchouten: It seems I should have talked to you guys before forking, heh :) Sep 17 20:26:39 - support for private data Sep 17 20:27:00 abraxa_: well, no problem. the changes are so small, we can figure it out ;) Sep 17 20:27:25 EdSchouten: Yep, that's it for now, though I do plan on using more XSPF features later Sep 17 20:27:26 abraxa_: very cool! Sep 17 20:29:11 abraxa_: let's team up to find a good API when you need it, how about that? Sep 17 20:29:18 abraxa_: what do you think about adding a simple void * that application writers can use as they wish? Sep 17 20:29:19 EdSchouten: 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 17 20:29:46 EdSchouten: Sounds just fine to me Sep 17 20:30:16 sping: Sure thing :) Sep 17 20:35:47 EdSchouten/sping: What's the best way to contact you guys? Sep 17 20:35:57 abraxa_: i hope libspiff as of rev 354 should suit your needs Sep 17 20:36:08 are the AR6001 drivers being used in the GTA02 based on a OSS project, or did they come directly from Atheros? Sep 17 20:36:20 abraxa_: if it's about the C bindings, you'd better contact me :) Sep 17 20:36:24 email will do fine Sep 17 20:36:47 cbrake: neither. they're being written from scratch Sep 17 20:37:01 mickeyl: wow Sep 17 20:37:02 cbrake: probably based on some other gpl code, no idea 'bout details Sep 17 20:37:03 EdSchouten: Okay. Thanks a bunch :) Sep 17 20:37:09 you're welcome :) Sep 17 20:37:23 abraxa_: openmoko is the second project, using libspiff btw :) Sep 17 20:37:35 sping: right? :) Sep 17 20:37:45 the first one was Herrie (http://herrie.info/) Sep 17 20:37:55 EdSchouten: stop your spam ;-) Sep 17 20:38:05 rbu: sst ;) Sep 17 20:38:30 Greetings Sep 17 20:39:37 EdSchouten: I'll remove my own binding then as soon as the changes have found their way to OpenEmbedded Sep 17 20:40:02 great! :) Sep 17 20:40:31 well, i thought it was pretty funny seeing my name inside the openmoko svn, but we'd better work together ;-) Sep 17 20:43:13 Hey EdSchouten Sep 17 20:43:27 last time we met at the freebsd booth on linuxtag :) Sep 17 20:43:36 ah, cool :) Sep 17 20:43:45 I'm currently looking into booting a FreeBSD kernel on the Neo :) Sep 17 20:43:50 Maledictus: what's your name? :) Sep 17 20:43:59 Marius :) Sep 17 20:44:04 hey marius! :) Sep 17 20:44:07 hehe Sep 17 20:44:24 we meet everywhere now and then :D Sep 17 20:44:38 When I read your name here I thought you want to help out with FreeBSD on the neo ;) Sep 17 20:45:20 lol :) Sep 17 20:45:24 sorry, not enough time for that ;) Sep 17 20:45:39 only a couple of months and i'm going on my final internship for school Sep 17 20:45:40 how sad Sep 17 20:45:48 ah ok Sep 17 20:46:03 hopefully i'm going to work at a really big commercial linux company in the UK Sep 17 20:46:23 sound interesting. how long will you go? Sep 17 20:47:22 half a year Sep 17 20:47:49 EdSchouten: linux? must be hard for you Sep 17 20:48:00 rbu: well, i can live with it ;) Sep 17 20:48:16 whois nbd Sep 17 20:49:12 Maledictus: still taking a long time, those BSDA results Sep 17 20:49:41 yep, I mailed axel bout them. They have to completly finish the Beta first... Sep 17 20:50:09 nbd: how is the AR6001 driver coming? I'm getting hardware for an industrial project soon and looking for options to start with. Sep 17 20:50:12 yeah, i got an email about that some time ago Sep 17 20:50:35 well, good hearing from you, Maledictus Sep 17 20:50:43 i have to go now -> have to wake up very early :) Sep 17 20:50:50 cbrake: i'm still having trouble with my card and i don't know if it's the controller, the stack or the card itself Sep 17 20:50:55 yeah, cu, and have fun in the uk :9 Sep 17 20:50:55 abraxa_: thanks for your cooperation :) Sep 17 20:50:57 :) Sep 17 20:51:01 bye :) Sep 17 20:51:04 bye Sep 17 20:51:23 cbrake: it starts doing the driver init and then the card locks up after a second Sep 17 20:51:38 Is there public documentation for the atheros chip available? Sep 17 20:51:44 no Sep 17 20:51:54 how's the driver written? Sep 17 20:52:23 Maledictus: there's free software code for it Sep 17 20:53:16 how does that help? people trying to use it and trying to fix bugs by choosing other random values for registers? Sep 17 20:53:21 EdSchouten: Thank *you* ;) Sep 17 20:54:16 Maledictus: someone did an RE'd HAL for it, there is an openish wrapper around that Sep 17 20:54:25 nbd: I am working on integrating the Codeplex WinCE driver on a CE project. But I'm not sure if we'll get it finished before my customer switches to Marvell as I'm still working on the SDIO stack portion. Sep 17 20:54:27 Maledictus: i think someone may have done informal specs Sep 17 20:54:58 nbd: we're still up in the air which module we'll go with Sep 17 20:56:14 same HAL as for the other atheros wireless stuff? from reyk? Sep 17 20:56:41 no Sep 17 20:56:54 the 6k has equivalent functionality in its firmware Sep 17 20:56:56 hi nbd Sep 17 20:57:15 hey mickeyl Sep 17 20:59:09 mjr: oh, i didnt know that... does the firmware need to be uploaded on init or is it in flash? Sep 17 21:02:10 flash, iirc Sep 17 21:06:42 mickeyl: you were on mrmcd, right? Sep 17 21:07:54 nbd: yeah, but just shortly. you were surrounded by guys, so i didn't want to drag you out ;) Sep 17 21:08:35 :) Sep 17 21:25:24 wow, people are quiet today Sep 17 21:25:37 busy hacking, i guess ;) Sep 17 21:25:49 or debugging broken builds as usual Sep 17 21:27:39 I often get the error "svn: PROPFIND request failed on '/' Sep 17 21:27:39 svn: PROPFIND of '/': Could not resolve hostname `svn.openmoko.org': No address associated with hostname (http://svn.openmoko.or Sep 17 21:28:05 anyone have any idea why that happens? Sep 17 21:28:18 ping svn.openmoko.org? Sep 17 21:28:59 hmm... Sep 17 21:29:20 i get a response from sita Sep 17 21:29:24 first attempt failed to resolve, second worked fine Sep 17 21:29:30 ... Sep 17 21:29:35 maybe my dns is hosed Sep 17 21:29:43 that happened to me too... i assumed i was going crazy Sep 17 21:30:31 hmmm.. maybe it's not just me.. when I build it often stops in the middle due to this dns problem Sep 17 21:31:23 oddly, from a remote site, it resolved just fine Sep 17 21:31:57 make that two remote sites Sep 17 21:33:19 maybe we're both going insane Sep 17 21:34:22 could be, my Linux skills are pretty weak, that's not helping me at all and I hate to ask dumb Linux questions here Sep 17 21:35:01 well, don't feel too bad, i ask dumb openembedded questions all the time Sep 17 21:35:26 like for instance how to downgrade my glibc to 2.5, I've tried several things with no success Sep 17 21:35:41 openmoko: 03andrew * r2984 10/trunk/src/host/qemu-neo1973/ (usb-linux-gadget.c vl.c): Force SIGALARM based timers when -usbgadget is used, for now (until gadgetfs does AIO). Sep 17 21:36:41 casten: what distro are you running on? Sep 17 21:37:11 Ubuntu 7.04 Sep 17 21:39:24 well Sep 17 21:40:27 it seems like you ought to be able to go into some package manager and tell it that you want glibc2.5, if you can tell apt where to find all the packages it needs Sep 17 21:40:59 but i'm not sure the regular 7.04 sources will include anything about glibc2.5 Sep 17 21:42:15 yeah, Synaptic Package Manager doesn't seem to list it, I guess I could try te find something else Sep 17 21:43:38 i end up using aptitude, but they ought to all kinda show the same things Sep 17 21:44:16 you should be able to get everything working by using only apt-get install Sep 17 21:44:44 7.04 is feisty, right? Sep 17 21:44:53 hmm, looking at Synaptic, I see it reports libc6 is installed with version 2.5, so now I am wondering why the build says it's not 2.5 Sep 17 21:45:02 yeah Sep 17 21:45:06 yeah, 7.04 is feisty Sep 17 21:45:16 i was about to say, feisty looks like it has 2.5 right now Sep 17 21:45:22 http://packages.ubuntu.com/feisty/libs/libc6-i686 Sep 17 21:45:32 yeah, apt-get install should work, but it just doesn't for me Sep 17 21:46:01 well, it seems like as far as apt is concerned, you have glibc2.5 installed Sep 17 21:46:21 brb Sep 17 21:47:08 yeah, looks like I do, I'm just wondering why the build thinks it's not.. maybe it is looking in the wrong place Sep 17 21:47:10 Casten: why? how? Sep 17 21:47:22 what does apt-get update && apt-get upgrade tell you? Sep 17 21:47:42 guaqua: please disregard my comment about apt-get not working Sep 17 21:48:18 what kind of errors do you run into? Sep 17 21:50:04 First I get a warning that says "Preferred version 2.5 of glibc not available..." Sep 17 21:50:27 warnings are okay Sep 17 21:51:27 I'm running into problems while building libsdl-net Sep 17 21:51:48 I'm getting an error bout GUI_Create missing. any advice? Sep 17 21:52:18 does google tell you soemthing? Sep 17 21:53:07 I got far in previous attempts to build, but at the end there was an error about incompatible library.. that was when I tried figuring out about the 2.5 problem. Sep 17 21:53:48 Ghiottone: i'm running into the same problems. Sep 17 21:54:13 looks like bb files are pulling files from the host system. Sep 17 21:54:23 s/files/paths/ Sep 17 21:54:23 Writchie meant: looks like bb paths are pulling files from the host system. Sep 17 21:55:19 mmm Sep 17 21:55:19 ghiottone: was able to do a make openmoko-devel-image without problems - make update all fails Sep 17 21:55:31 the bb file is very simple Sep 17 21:55:36 no clue from it Sep 17 21:55:39 too simple to work Sep 17 21:55:48 look at the libsdl-x11* Sep 17 21:57:28 oh. got it Sep 17 21:57:37 I think you're right. Sep 17 21:57:44 I'll have a better look at it Sep 17 21:58:00 i think there are still lots of things being picked up from the host system Sep 17 21:58:12 which is why testing on the developer's system is useless Sep 17 21:58:29 well... Sep 17 21:58:48 maybe not useless Sep 17 21:59:01 but not representative of build platforms in the field Sep 17 21:59:18 i didn't run into big problems building Sep 17 21:59:34 the advice given in the wiki and OE website is pretty accurate Sep 17 22:00:13 quaqua: this is a change introduced in last day or so Sep 17 22:01:12 mmh...okay Sep 17 22:02:13 there should be few reasons for a cross development environment picking up things from the local system Sep 17 22:02:42 there should be *zero* things the cross environment is picking up from the host.. Sep 17 22:03:04 hozer: there may be a few things needed for building correct natives Sep 17 22:04:00 like what? Sep 17 22:04:18 hi. this might be a FAQ , sorry but I have not found the answer, will there be a route planner in the phone? I mean, it has GPS :) Sep 17 22:04:20 Writchie, I'm now trying with --disable-gui (that whose configure was giving me problem) Sep 17 22:04:37 but the update has brought me some other package to recompile before it :/ Sep 17 22:05:33 Casten: By the way, upgrading or downgrading your libc is not a dumb Linux newbie question. It's not recommended even for wizards, unless they're glibc wizards. Problem is that every process in the system is loading glibc as a shared library. Also, if the new one is broken, all shell commands stop working, so it's hard to fix. Sep 17 22:06:06 as in: ls -> segfault! Sep 17 22:06:47 been into that situation thanks to a brave night playing with prelink with all executables around... Sep 17 22:06:54 Casten: I suspect that the warning about "preferred glibc 2.5 not found" relates to the libc that's being installed in the OE image -- not on the host. Sep 17 22:07:38 Ghiottone: Does the OE image use prelink? I didn't find it among the executables, but it's very important in embedded machines without swap, since it reduces the number of dirty pages created by library relocations. Sep 17 22:07:39 gnu: ok, good to know, thanks, I'm too new to know what I don't know ) Sep 17 22:07:47 weeeelll, that's nice Sep 17 22:08:07 my ATT gophone SIM seems to end up hanging the kernel Sep 17 22:09:26 hozer: last I heard gsmd was hanging the kernel - don't know if its been fixed Sep 17 22:09:29 Ghiottone, don't know. that episode happened some years ago. it is not related to openmoko or openembedded Sep 17 22:09:33 ops :) Sep 17 22:09:51 gnu, don't know. that episode happened some years ago. it is not related to openmoko or openembedded Sep 17 22:09:54 sry for the spam Sep 17 22:10:19 Ghiottone: ok Sep 17 22:10:58 hozer: I've had better luck after doing "ipkg update && ipkg upgrade" in the last few days. GSM daemon problem seemed to be worked out (at least once!) Sep 17 22:11:06 gnu, why the reduction? shouldn't it be better with shared libraries as the total memory requirement should be lower? Sep 17 22:11:15 Writchie: yeah, that's definitely what's going on Sep 17 22:11:30 Writchie: any idea why? It seems to be related to sound Sep 17 22:11:34 hozer: I believe it was kernel problem Sep 17 22:11:51 gnu: did you upgrade the kernel? Sep 17 22:11:58 hozer, the shutdown of the gsm chip triggered the kernel lock, if this is what you are refering to Sep 17 22:12:10 i'm still testing with OM2007.1 Sep 17 22:12:13 a change in boot environment helped Sep 17 22:12:27 don't have an image good enough for OM2007.2 Sep 17 22:12:59 Ghiottone: change in uboot? Sep 17 22:13:07 Ghiottone: Libs that are shared between programs (w/mmap) do reduce total mem requirement. Prelink doesn't change that. What it does is statically relocates each shared lib to a different address, and fixes up the pointers among shared libraries to assume that they're linked at those addresses. Thus when the libs are mmap'd in, if they end up at the pre-picked addrs, no relocation is needed. Sep 17 22:13:24 hozer, not exaclty. in the uboot environment. it is a parameter passed to the kernel Sep 17 22:13:45 Ghiottone: Since no relocation is done by the runtime linker, the pages mapped in from the filesystem are never modified, so they can all be shared among different processes that use the library. Sep 17 22:13:46 Ghiottone: any references? Sep 17 22:14:20 hozer, the problem is that when the gsm chip is turned of, its serial line is given to a serial console. but as it has hardware flow control active, this locks the kernel into a forever loop Sep 17 22:14:23 iirc Sep 17 22:14:31 AH Sep 17 22:14:35 hozer, sure, give me a second Sep 17 22:14:38 i think its in the Bugzilla Sep 17 22:14:52 Ghiottone: Also, and more important: Once a copy-on-write mmap'd page has been written to, it can never be discarded until the process dies. Unwritten pages can be thrown away when we run out of phys RAM, and read back in from the filesystem as needed. Sep 17 22:14:52 heh, I just duplicated that bug then ;) Sep 17 22:15:09 bug 788 Sep 17 22:15:34 hozer: I never did update my kernel. ipkg claimed to bring in a new kernel, but it doesn't change what gets booted from the boot partition. Sep 17 22:15:44 hozer, the boot variable is bootargs_base= byrootfstype=jffs2 root=/dev/mtdblock4 console=ttySAC0,115200 console=tty0 loglevel=8 Sep 17 22:15:56 hozer, you have to remove console=ttySAC0,115200 Sep 17 22:16:24 ah, so if you have a debug board connected, you get serial console? Sep 17 22:16:36 hozer: If the new kernel is important, somebody (who knows how) needs to make a new kernel-partition that folks like me can just reflash with. Sep 17 22:17:22 abraxa_: strange. even with your patch all i get is "Internal Data Flow Error" when trying to connect mad to pulsesink w/o audioconvert and auidoresample Sep 17 22:17:46 hozer: you don't need to change kernels to circumvent the kernel-hang bug; instead, the /etc/init.d scripts have to disable hardware flow control on that UART. Sep 17 22:17:49 I think there is a way to boot a new kernel from SD without the need to reflash it Sep 17 22:18:10 hozer, actually I don't know if it is the debug board console or the usb one. Sep 17 22:18:49 * hozer reads up on the bug Sep 17 22:19:03 gnu, if libraries gets pre-relocated at fixed address, how is the process virtual memory space managed? Sep 17 22:19:14 gnu, could the image of the process overlap with the library? Sep 17 22:20:25 gnu, also, writes for relocation should be needed only in the pic table of the shared object, whose size should be no more than a few pages, I think Sep 17 22:29:53 Ghiottone: Libraries remain relocatable, but rather then being linked as if they go at address 0 (which means every ref will have to be relocated), they are linked as if they go at some unique large address. Thus if mapped at that addr, no runtime reloc is needed. If that addr is already occupied, then the lib goes elsewhere and gets relocated. Sep 17 22:31:05 gnu, ah I see now. pretty simple Sep 17 22:31:11 Ghiottone: There are relocations for finding the data segment of the library. Also, a few pages of written PIC table in every process means hundreds of pages total that are stuck in RAM forever. Sep 17 22:31:51 Ghiottone: There's a good paper written up about prelink by its author (a Redhat emp and former Cygnus emp). Sep 17 22:31:57 gnu, yeah I see the point. Sep 17 22:32:18 gnu, interesting. do you have a link at hand? Sep 17 22:33:23 Ghiottone: http://people.redhat.com/jakub/prelink.pdf , also: http://en.wikipedia.org/wiki/Prelink Sep 17 22:33:51 gnu, thanks! Sep 17 22:44:15 it seems that extra_oeconf is ignored Sep 17 22:45:11 someone was talking about gsm/pppd muxing earlier, right? Sep 17 22:45:23 is the plan to integrated ppp as a library into gsmd? Sep 17 22:46:12 How does does the modem tell you there's an incoming call during a GPRS ppp session? Sep 17 22:46:52 * mjr has no clue of the plan, but a pipe pair would probably do... Sep 17 22:47:09 a pty would make more sense Sep 17 22:47:22 since that's essentially what they're for... Sep 17 22:50:49 hozer: when gsm mux is implemented pppd will be just be able to open a pty and dial Sep 17 22:52:00 hozer: an incoming (or outgoing) call will suspend the current gprs session Sep 17 22:52:47 ppp, dbus, or both for gsmd? Sep 17 22:53:04 mickeyl: I'll flash a new image tonight and keep working on it - thanks for trying it, too :) Sep 17 22:53:16 :) Sep 17 22:53:28 i couldn't resist :) Sep 17 22:54:00 hehe Sep 17 22:54:15 Writchie: I'm thinking about it and the mux will cost extra context switches and power in the kernel Sep 17 22:55:20 hozer: the mux is designed for being powered down with active connections. Sep 17 22:55:22 gnu, a prelink package seems to be in openembedded. I'm trying to build it Sep 17 22:55:54 hozer: it will wake on any activity on any of the DLC's Sep 17 22:56:27 Writchie: I was thinking of the ttySAC0 -> kernel -> gsmd -> kernel -> pppd -> kernel -> userspace networking app Sep 17 22:56:38 hozer: the actual mux/dmux can be done in a kernel or tty line discipline or in user space. Sep 17 22:58:30 ah, okay. Sep 17 22:58:32 hozer: I am currently working on a libgsm/gsmd alternative that uses gsm to establish or connect to sessions for gprs, fax, or other legacies. Sep 17 22:58:46 s/gsm/d-bus/ Sep 17 22:58:46 Writchie meant: hozer: I am currently working on a libd-bus/gsmd alternative that uses gsm to establish or connect to sessions for gprs, fax, or other legacies. Sep 17 22:59:11 dbus will also be used for the other telephony functions incl. SMS etc. Sep 17 22:59:42 through d-bus, you can connect bluetooth serial sessions to gprs or fax Sep 17 23:00:06 Ghiottone: cool! Sep 17 23:00:08 ooh, scary ;) Sep 17 23:00:10 it will also support neo to neo encrypted voice sessions using CSD Sep 17 23:00:24 what's csd Sep 17 23:00:31 circuit switched data Sep 17 23:00:44 okay. Could this also use wifi voip ? Sep 17 23:00:54 Ghiottone: BTW, best way to see the improvement in prelink RAM usage is to look at /proc/*/smaps and look for Dirty pages in particular. Sep 17 23:00:54 who has the libsdl-net problem? put a tarball of ${WORKDIR} online and I'll take a look Sep 17 23:01:23 with appropriate network yes Sep 17 23:01:49 hrrm.. what's it take to terminate a CSD connection .. say I want to go from my neo to something that has native IPv6 routing Sep 17 23:01:51 CSD has very predictable delays - but similar performance can be achieved with QoS Sep 17 23:02:07 CSD is just like a phone call. Sep 17 23:02:17 hozer: I think ISDN Sep 17 23:02:37 CSD will interwork with GSM or analog modems Sep 17 23:02:45 or ISDN Sep 17 23:02:57 for ISDN it remains in digital domain Sep 17 23:03:03 ditto for GSM Sep 17 23:03:23 Writchie, sounds good, what you're working on Sep 17 23:03:30 okay, cool.. I was just messing with gprs, and getting 800ms or so ping times Sep 17 23:03:35 is the latency any lower with CSD? Sep 17 23:03:43 hozer: yes Sep 17 23:03:48 gsm to gsm Sep 17 23:04:14 what's it generally billed as by the providers? Sep 17 23:04:19 voice call Sep 17 23:04:21 data call Sep 17 23:04:29 which is often more expensive Sep 17 23:04:43 and may not be easily enableable on some accounts Sep 17 23:04:44 hmmm. i'll check my bill Sep 17 23:04:55 tmo usa charges same per minute IIRC Sep 17 23:05:00 probably safe to say 'varies with locality' Sep 17 23:05:06 hrrm Sep 17 23:05:07 usually against the plan Sep 17 23:05:39 most in US don't even know about CSD and fax calls to and from their handsets Sep 17 23:05:45 now, if I can just find some way to make a CSD call to my server at home Sep 17 23:05:53 just need a modem Sep 17 23:05:57 hozer, hook up a modem Sep 17 23:05:58 yeah Sep 17 23:06:07 ick.. crappy latency though ;) Sep 17 23:06:19 hozer: not really Sep 17 23:06:53 what latency do you get? Sep 17 23:06:55 it travels digital to the inter-working port Sep 17 23:07:11 then its a circuit switched analog call Sep 17 23:07:19 converted to digital for transport Sep 17 23:07:24 i tested mobile to mobile csd with some phones i had lying around Sep 17 23:07:25 and pack to digital by your modem Sep 17 23:07:34 a very solid half duplex 9600bits/s Sep 17 23:07:53 it should be find for an encrypted session Sep 17 23:08:02 use a codec of your choice Sep 17 23:08:12 voip codecs should be even better Sep 17 23:08:24 yeah.. I was thinking more of something that has predictable latency for SSH sessions Sep 17 23:08:25 because they handled delay better than gsm codec Sep 17 23:08:26 Yeah. There will likely be a noticable, but not overly horrid voice latency. Sep 17 23:08:48 hozer: i don't think you'll even notice anything on SSH session. Sep 17 23:08:55 however, the clock is running Sep 17 23:09:05 I notice anything over 100ms Sep 17 23:09:06 you do, actually, but it's noticably better than gprs of course Sep 17 23:09:25 but peak rate is 9.6kbps? Sep 17 23:09:41 yes Sep 17 23:09:45 you can negotiate for more Sep 17 23:09:55 depends on network Sep 17 23:09:56 is it possible to have a gprs connection and a CSD connection open? Sep 17 23:10:01 no Sep 17 23:10:03 No. Sep 17 23:10:05 i'm not sure we are HSCSD capable Sep 17 23:10:12 same as voice and gprs Sep 17 23:10:32 ah, but EDGE devices can do both, right? Sep 17 23:10:35 need class A or UMTS for simultaneous packet and CSD/voice Sep 17 23:10:45 hozer, no Sep 17 23:10:58 (hmm, 'least I don't think so) Sep 17 23:10:59 hozer: no - EDGE or GSM doens't matter - both phone and network must support Dual Mode Sep 17 23:11:14 ah, okay. Sep 17 23:11:15 Thanks Sep 17 23:11:20 Dual Transfer Mode DTM IIRC is the term Sep 17 23:11:24 anyway yeah, if we have hscsd it can go up to 14k4, with 9k6 as the baseline Sep 17 23:11:44 so how I map 'CellID = 5279' to a provider Sep 17 23:11:58 (I don't think we have multiple time slots for that, but I could be mistaken) Sep 17 23:12:15 mjr: i think HSCSD has multiple slots Sep 17 23:12:21 mjr: 9kbps x 6 channels? Sep 17 23:12:26 but i'm not sure how its billed Sep 17 23:12:42 The spec has them, yes. Does the TI? Sep 17 23:13:00 i don't think we have HSCSD Sep 17 23:13:16 IIRC i think I looked at this before Sep 17 23:13:32 but I looked at so many things I can't keep them straight anymore ;) Sep 17 23:13:53 mickeyl, the sdl-net tarball is at http://aiurlano.netsons.org/workdir-libsdl-net.tar.gz Sep 17 23:13:58 hozer: the cell ID is not unique, what you want is the provider ID Sep 17 23:14:14 hozer: the real number is three parts: provider, LAC, and cell ID Sep 17 23:14:31 mickeyl, I'm try to change configure with extra_oeconf, but with no effect Sep 17 23:14:32 (the provider is in fact two numbers itself, country and provider...) Sep 17 23:14:47 well, s/provider/carrier/ Sep 17 23:15:31 s/try/trying/ Sep 17 23:15:31 Ghiottone meant: mickeyl, I'm trying to change configure with extra_oeconf, but with no effect Sep 17 23:17:01 hozer, 14k4x4 max for hscsd, if we have it, but I'll defer to Writchie's guess that we don't Sep 17 23:17:48 not that interesting anyway, what with bills for each timeslot... Sep 17 23:18:40 Ghiottone: aha. it picks up the wrong sdl-config binary Sep 17 23:19:19 Ghiottone: try adding the following line to your .bb Sep 17 23:19:21 export SDL_CONFIG = "${STAGING_BINDIR_CROSS}/sdl-config" Sep 17 23:19:24 then rebuild Sep 17 23:20:25 mickeyl, in a particular section? Sep 17 23:21:29 Ghiottone: top level, i.e. after the S assignment Sep 17 23:22:59 mickeyl, building... how did you see that it was getting the wrong dir? Sep 17 23:23:26 i inspected the generated Makefile Sep 17 23:23:29 and spotted the following: Sep 17 23:23:41 Makefile:SDL_CFLAGS = -I/usr/include/SDL -D_GNU_SOURCE=1 -D_REENTRANT Sep 17 23:23:41 Makefile:SDL_CONFIG = /usr/bin/sdl-config Sep 17 23:23:41 Makefile:SDL_LIBS = -L/usr/lib -lSDL Sep 17 23:24:08 and since these flags are computed during configuring by calling sdl-config, it was clear who was to blame Sep 17 23:24:27 usually it should have picked up our sdl-config though Sep 17 23:24:50 so now trying to override the path to sdl-config Sep 17 23:25:46 mickeyl, it seems to stop at the same place as before Sep 17 23:26:15 ok, so that line doesn't keep it from finding the wrong sdl-config Sep 17 23:26:27 good. i'll try to reproduce by putting a bogus sdl-config into my /usr/bin Sep 17 23:29:15 mickeyl, it seems that EXTRA_OECONF has no effect. is it possible? Sep 17 23:29:26 depends on the options you give Sep 17 23:29:33 some options are just not honored Sep 17 23:29:41 that's depending on the autofoo Sep 17 23:29:56 libsdl has been built, right? Sep 17 23:29:58 (just checking) Sep 17 23:30:04 yes Sep 17 23:30:13 I'll do it again to be sure Sep 17 23:30:29 I did the -native as well Sep 17 23:32:14 gnu, prelink seems to build fine after I corrected some warning (that was being treated as error) from elfutils Sep 17 23:32:34 sounds like a gr8 project Sep 17 23:32:42 will openmoko have built-in right-to-left language support? Sep 17 23:32:42 DB42: 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 17 23:32:50 atleast in stuff like SMS and so ? Sep 17 23:33:25 Ghiottone: great! Should be possible to just try it on your phone. If it totally hoses the phone, you can always reflash from scratch! Sep 17 23:34:39 gnu, mmm that sound not that good ;) Sep 17 23:34:42 DB42: nobody has built that part of the project yet (SMS support), so if you want right-to-left, you could make it happen! Sep 17 23:34:52 gnu, I think I can put it online in my repo Sep 17 23:34:54 no SMS yet ? hmm... Sep 17 23:35:02 ok, in the contacts list, is there RTL ? :) Sep 17 23:35:11 DB42: low level support exists, but no GUI for sms yet. Sep 17 23:35:33 DB42: don't know, haven't tried, don't even have a gui to set the language or country yet. Sep 17 23:35:53 c Sep 17 23:35:54 k Sep 17 23:35:55 Ghiottone: I'm happy to run it on my phone. Sep 17 23:36:05 dont need the language to be RTL, just the text :) Sep 17 23:43:45 Ghiottone: try applying the patch from http://pastebin.ca/701292 Sep 17 23:44:53 will there be a gecko browser? Sep 17 23:45:45 nah Sep 17 23:45:48 gecko is _fat_ Sep 17 23:46:02 so which one ? Sep 17 23:46:06 * mickeyl does the webkit-dance Sep 17 23:47:15 webkit is slimimer ? Sep 17 23:47:20 webkit is what safari is based on, yes ? Sep 17 23:47:22 opera mini? ;) Sep 17 23:47:49 http://scap.linuxtogo.org/files/f2d2c88cd2b88c0ca7a2ec550eb0eab7.png Sep 17 23:48:23 yay for nav software Sep 17 23:49:05 Well, it's really slow at the moment. Sep 17 23:49:13 But it is a start. Sep 17 23:50:48 DB42: eventually there will support for SMS in RTL languages and input systems for Arabic, Chinese, etc. Sep 17 23:51:00 DB42: yes, safari. Sep 17 23:51:21 Writchie|dinner: cool Sep 17 23:51:31 mickeyl: how smaller is webkit then gecko/trident? Sep 17 23:52:19 i always thought gecko is slim Sep 17 23:52:44 mickeyl, netincludedir variable cleared but got the same error Sep 17 23:54:54 Ghiottone: hmm, *sigh*. i'm a bit out of ideas now. just remove it from the feed then... Sep 17 23:55:13 so I've been playing with two different SIM cards, and it's a little annoying to change them all the time Sep 17 23:56:33 it seems to complain about a missing libGUI Sep 17 23:56:41 what if I disable it? Sep 17 23:56:53 but why extra_oeconf seems not to work? Sep 17 23:57:08 laters, good project btw ! Sep 17 23:57:18 Ghiottone: are you using live device or QEMU Sep 17 23:57:33 live device Sep 17 23:57:53 i think you can just build openmokoe-devel-image Sep 17 23:57:56 that worked for me Sep 17 23:58:29 yeah. but I'm needing libdsl-net for a game I was writing the bb recipe for Sep 17 23:58:39 ahh Sep 17 23:59:13 thus I made : Sep 17 23:59:15 err Sep 17 23:59:16 gnu, prelink and needed packags should be downloadable at my repo. look at my userpage Sep 17 23:59:36 gnu, http://wiki.openmoko.org/wiki/User:Alessandro Sep 18 00:00:24 http://wiki.openmoko.org/wiki/Hardware:Neo1973:Alternate_Cases:Dual_SIM Sep 18 00:04:44 mickeyl: i dont know if it´s solved already but you can fing a working squashfs.patch at http://celtune.morb-design.com/squashfs.patch or a kernel-module.ipk at http://celtune.morb-design.com/ipk/fic-gta01/kernel-module-squashfs_2.6.22.5-moko11+svnr2937-r2_fic-gta01.ipk Sep 18 00:05:03 but not tested the ipk Sep 18 00:05:49 thanks Marcel_M Sep 18 00:06:02 np Sep 18 00:06:35 mickeyl, I needed to patch elfutils for it to compile. a trivial fix. a gcc warning treated as error stopped the whole process Sep 18 00:06:44 mickeyl, should I create a bug on OE? Sep 18 00:06:48 Ghiottone: yes, please Sep 18 00:12:14 morning Sep 18 00:13:26 hello rwhitby Sep 18 00:18:50 hello rwhitby Sep 18 00:20:46 if anyone has a moment, any ideas what is causing this, I'm not quite experienced enough to figure out build failures Sep 18 00:20:55 casten@dell:~/moko$ make openmoko-devel-image Sep 18 00:20:55 ( cd build && . ../setup-env && \ Sep 18 00:20:55 bitbake openmoko-devel-image uboot-openmoko ) Sep 18 00:20:55 NOTE: Handling BitBake files: \ (4687/4687) [100 %] Sep 18 00:20:55 NOTE: Parsing finished. 4460 cached, 0 parsed, 227 skipped, 0 masked. Sep 18 00:21:13 outch Sep 18 00:21:25 sorry for the spam Sep 18 00:21:38 pastebin? Sep 18 00:22:14 hmm, guess that would be sendfile? Sep 18 00:22:47 use something like http://pastebin.ca/ Sep 18 00:25:00 ok, here's the log http://pastebin.ca/701354 Sep 18 00:25:57 night Sep 18 00:25:59 nice hostname, my computer is an equally unimaginitave 'sony' Sep 18 00:26:11 hehe Sep 18 00:26:53 actually, this is a preinstalled Ubuntu from dell, even less creative than you think Sep 18 00:27:05 wow Sep 18 00:27:16 it's cool that they really started doing that Sep 18 00:27:40 mickeyl, bug for elfutils created with fix attached. number is 3023 Sep 18 00:27:46 yeah, perfect for this project, now if I could just get it to finish a build Sep 18 00:27:53 yeah Sep 18 00:28:24 have you tried doing an update on the mokomakefile stuff (just in case) Sep 18 00:28:33 yes, several times Sep 18 00:29:04 i've restarted the process many times from the beginning as per the moko page Sep 18 00:29:08 so that works, and a 'make setup' works... Sep 18 00:29:14 yes Sep 18 00:30:08 do you have all the required software installed? Sep 18 00:30:11 /home/casten/moko/build/tmp/work/armv4t-angstrom-linux-gnueabi/ltrace-0.4-r1/temp/run.do_compile.5043: 38: INSTALL: not found Sep 18 00:30:14 looks a bit strange to me Sep 18 00:30:27 as far as I know, everything is installed Sep 18 00:30:50 is INSTALL a tool? I wasn't sure what that meant. Sep 18 00:30:52 mickeyl: i built everything alright, and i don't have an INSTALL Sep 18 00:31:27 well, INSTALL in caps would be a sign of a placeholder Sep 18 00:31:42 which in turn makes me suspect something can't be found Sep 18 00:33:09 casten: you've 'apt-get install'ed everything from that list, right? Sep 18 00:33:31 bb is running: oe_runmake LDFLAGS=${TARGET_LDFLAGS} LIBS="-lsupc++ -liberty -Wl,-Bstatic -l Sep 18 00:33:31 elf -Wl,-Bdynamic" ${EXTRA_OEMAKE} ARCH= Sep 18 00:33:42 yes Sep 18 00:33:45 errk, that's didn't work Sep 18 00:33:48 *that Sep 18 00:33:49 ok Sep 18 00:34:14 I see it complains about libiberty also Sep 18 00:35:14 yeah Sep 18 00:36:38 you know where it first dies Sep 18 00:36:56 build/tmp/work/armv4t-angstrom-linux-gnueabi/ltrace-0.4-r1/temp/log.do_compile.5043 Sep 18 00:37:00 line 38 Sep 18 00:37:20 can you drop that file into another pastebin? Sep 18 00:37:28 sure Sep 18 00:37:38 my line 38 in the same file is actually blank Sep 18 00:37:47 so i think it's line 37 that's really dying Sep 18 00:37:54 but i kinda want to see what the differences are Sep 18 00:40:22 pastebin.com/d1633e873 Sep 18 00:41:12 oh Sep 18 00:41:12 looks like a liberty complaint Sep 18 00:41:16 sorry Sep 18 00:41:50 i meant, can you paste the actual file 'moko/build/tmp/work/armv4t-angstrom-linux-gnueabi/ltrace-0.4-r1/temp/run.do_compile.5043', into a pastebin Sep 18 00:42:26 'cause that's where it's hitting the INSTALL command that it doesn't know how to handle Sep 18 00:44:32 pastebin.ca/701372 Sep 18 00:45:29 goodnight all Sep 18 00:45:38 that's... odd Sep 18 00:45:38 night Ghio Sep 18 00:45:54 your line 38 is exactly the same as mine Sep 18 00:46:08 hmm Sep 18 00:47:06 so I guess INSTALL is being defined on my build, but not on yours Sep 18 00:47:25 Any idea where that gets defined? Sep 18 00:47:27 * orpheus is rebuilding ltrace just to make sure it works Sep 18 00:47:33 not really... Sep 18 00:47:47 what's the syntax for an emote? Sep 18 00:47:58 :( ? Sep 18 00:48:21 * casten tries to do an emote. Sep 18 00:48:36 so, ltrace is a package that's something like half way through the build, right? Sep 18 00:49:02 I imagine, I can't be sure since I've never completed a build. Sep 18 00:49:38 I got my phone a few weeks ago and haven't been able to build and flash it up. Sep 18 00:49:47 Pretty sad. Sep 18 00:49:50 "NOTE: Running task 2670 of 3476" Sep 18 00:50:29 so, yeah, it's not like your build environment is really broken, otherwise one of the first 2669 tasks should have failed Sep 18 00:50:39 hmm... I'm at 695 of 4795 :P Sep 18 00:50:45 heh Sep 18 00:50:58 casten: you do know there are pre-built images, right? Sep 18 00:51:12 yeah, looks like task 2669 is the failure point for me Sep 18 00:51:39 yeah, but what I am going to be doing will require me to modify the system image Sep 18 00:51:53 ok Sep 18 00:52:17 as in, something that can't be done after the first boot? Sep 18 00:52:24 makes sense Sep 18 00:52:48 I suppose I could just do some app testing for now with a prebuilt image, but I would prefer to have everything building first, maybe I should just give up. Sep 18 00:53:02 well, it oughta work Sep 18 00:53:06 one sec Sep 18 00:53:11 my build just died now... Sep 18 00:53:42 hmmm, really? Sep 18 00:54:03 looks like it was for a different reason Sep 18 00:54:20 "sh: oe_sha256sum: command not found" Sep 18 00:55:41 it's been a while since I updated everything... Sep 18 00:56:35 The built files are up at ftp.openmoko.org or something like that? Sep 18 00:56:48 urm, sounds right Sep 18 00:57:14 it's linked somewhere from one of the "so you just got your phone and want to flash it" pages Sep 18 00:58:28 what is the fastest build anyone has done or seen? Sep 18 00:59:02 5 days Sep 18 00:59:36 notserpe: about 10 hours Sep 18 00:59:54 what sort of machine was it on? Sep 18 01:01:31 amd64 3500+ 1GB RAM Sep 18 01:02:05 running in 32 bit mode Sep 18 01:04:03 I have an old 8 processor Proliant server in the back, 800 Mhz with 3 10krpm scsi disks, I wonder how fast that would do it. Sep 18 01:07:14 awesome Sep 18 01:12:35 sweet. Sep 18 01:13:04 i dont have any 8cpu boxen myself. just 2 and 4 core. Sep 18 01:13:42 Ok, I'm deleting everything except the OE.mtn and starting over Sep 18 01:14:12 The thing roars like a hurricane when you fire it up. Sep 18 01:15:22 i try and keep my livingroom at a low roar. Sep 18 01:20:30 openmoko: 03zecke * r2985 10/trunk/src/target/OM-2007.2/applications/openmoko-browser/: Add a meaningless directory. Sep 18 01:22:21 casten: sorry about that, phonecall Sep 18 01:22:23 openmoko: 03zecke * r2986 10/trunk/src/target/OM-2007.2/applications/ (openmoko-browser/ openmoko-browser2/): Version 1 is complete, let us start the 2nd one Sep 18 01:22:32 haha Sep 18 01:22:39 awesome Sep 18 01:22:47 RAD Sep 18 01:22:50 rapid application development Sep 18 01:23:25 casten: i've been doing all my building in ~/moko2; ~/moko contains a first attempt that didn't work, but i wasn't ready to give up and delete it yet. Sep 18 01:23:46 orpheus: np Sep 18 01:24:05 hi! Sep 18 01:24:15 I gave up, nuked everything and am starting from near scratch Sep 18 01:24:17 also, what kinda machine do you have? are they all 8 xeon's? Sep 18 01:24:37 that's actually quite a fair bit of horsepower Sep 18 01:24:45 For this I am using a core duo laptop. Sep 18 01:24:45 Can the Neo1973 in its current form be used as an actual cell phone? Sep 18 01:24:46 t00na: 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 18 01:25:05 aloril: botsnack Sep 18 01:25:41 The proliant is a very old machine, only 800mhz long before xeon Sep 18 01:25:41 just saying, many build systems (i don't know about bitbake) can do magic parallalism with 8 processors Sep 18 01:26:01 ah Sep 18 01:26:07 if it's a proliant, it'll never die Sep 18 01:26:42 mine's actually quad 700's, and they're all xeon's... not the modern xeon's of course, just the pIII-based ones Sep 18 01:26:48 still might be able to keep with a coreduo though, but I'm not going to waste any time with it, I'd like to get it building on a normal machine Sep 18 01:27:01 makes sense Sep 18 01:27:07 bt the way Sep 18 01:27:23 i can't get ltrace to build now either Sep 18 01:27:44 hmmm, want to submit a bug? Sep 18 01:28:02 well, i've actually got some other problem Sep 18 01:28:15 and even after an update of everything, it's still toasted Sep 18 01:28:28 oe_sha256sum is my problem, to be exact Sep 18 01:28:36 it should be in my path, but is not Sep 18 01:29:26 With a build system this large, it would be nice to have daily fully automated builds, but there's probably not resources. Sep 18 01:29:45 hello? Sep 18 01:29:53 Hi t00na Sep 18 01:29:54 hi Sep 18 01:30:03 Casten: http://ipkg.nslu2-linux.org/feeds/openmoko/ is a continuous openmoko builder Sep 18 01:30:16 Hi, can the current Neo1973 be used as an actual cell phone? Sep 18 01:30:18 someone else was saying that they ... yeah. that. Sep 18 01:30:32 ah, well, maybe it will choke in the next 10 hours Sep 18 01:30:50 t00na: there is currently some issues with certain SIM cards, let me find the link Sep 18 01:31:01 okay Sep 18 01:31:11 t00no: I think there is a faq that says wait for the next version ina few months. Sep 18 01:31:25 great. Sep 18 01:31:32 Casten: why do you think it will choke? Sep 18 01:31:48 And at some point, will these end up being sold by other retailers? Sep 18 01:31:52 namely Amazon? Sep 18 01:32:32 rwhitby: Oh, orpheus said they too had problems with ltrace Sep 18 01:33:16 t00na: I wouldn't bet on it t00na, not for a while.. but I am not an authority, just my guess. Sep 18 01:33:42 oh. okay. Sep 18 01:34:19 Casten: nope, the only failure on the autobuilder is the libsdl-net one Sep 18 01:34:19 t00na: at the same time, openmoko did well enough about getting mine here Sep 18 01:34:45 cool. Sep 18 01:35:04 Well as soon as it's practical for someone like me to do so, I'll snatch one up as soon as I have the money. Sep 18 01:35:58 rwhitby: Ok, I am trying a clean build, it's hard for me to determine my problem. Orpheus just said they had a problem also. Sep 18 01:36:32 t00na: they are fun to hack on: I've already gotten a usb hub plugged into mine (using the usb port on the side as a usb root hub) Sep 18 01:36:59 Hm... I didn't know that they used GTK Sep 18 01:37:07 why GTK? Sep 18 01:37:21 openmoko: 03zecke * r2987 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (15 files in 4 dirs): (log message trimmed) Sep 18 01:37:21 openmoko: 2007-09-07 Holger Freyther Sep 18 01:37:21 openmoko: Add buildsystem and stubs for the OpenMoko browser Sep 18 01:37:21 openmoko: * AUTHORS: Added. Sep 18 01:37:21 openmoko: * COPYING: Added. Sep 18 01:37:23 openmoko: * Makefile.am: Added. Sep 18 01:37:27 openmoko: * NEWS: Added. Sep 18 01:37:29 openmoko: 03zecke * r2988 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (.gitignore po/.gitignore src/.gitignore src/main.c): Sep 18 01:37:32 openmoko: 2007-09-16 Holger Hans Peter Freyther Sep 18 01:37:34 openmoko: Ignore certain autogenerated files. Sep 18 01:37:36 openmoko: * .gitignore: Added. Sep 18 01:37:38 openmoko: * configure.ac: Sep 18 01:37:40 openmoko: * po/.gitignore: Added. Sep 18 01:37:42 openmoko: * src/.gitignore: Added. Sep 18 01:37:44 openmoko: 03zecke * r2989 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (5 files in 2 dirs): (log message trimmed) Sep 18 01:37:47 openmoko: 2007-09-16 Holger Hans Peter Freyther Sep 18 01:37:49 openmoko: Create the browser skeleton. It compiles but is not Sep 18 01:37:51 openmoko: doing anything useful. Sep 18 01:37:53 openmoko: * configure.ac: Sep 18 01:38:02 what the hell? Sep 18 01:38:05 openmoko: * src/Makefile.am: Sep 18 01:38:07 openmoko: * src/browser-data.h: Added. Sep 18 01:38:08 CIA-23: try pastebin Sep 18 01:38:09 openmoko: 03zecke * r2990 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (5 files in 3 dirs): Sep 18 01:38:12 openmoko: 2007-09-16 Holger Hans Peter Freyther Sep 18 01:38:14 openmoko: Document some of the bits that were discussed with Sep 18 01:38:16 openmoko: Ellen last Monday. Sep 18 01:38:18 openmoko: * documentation/design/GuiDesign: Added. Sep 18 01:38:20 openmoko: * documentation/design/MainTasks: Sep 18 01:38:22 openmoko: 03zecke * r2991 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (6 files in 2 dirs): (log message trimmed) Sep 18 01:38:25 openmoko: 2007-09-16 Holger Hans Peter Freyther Sep 18 01:38:27 openmoko: Start implementing the "Current Page". We start with Sep 18 01:38:29 t00na: it looks like a subversion-commit bot Sep 18 01:38:31 openmoko: providing the default actions for navigation. Sep 18 01:38:33 openmoko: * src/Makefile.am: Sep 18 01:38:33 * mjr smells a misguided bot Sep 18 01:38:35 openmoko: * src/browser-data.h: Sep 18 01:38:37 openmoko: * src/current-page.c: Added. Sep 18 01:38:39 openmoko: 03zecke * r2992 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (ChangeLog src/main.c): Sep 18 01:38:42 openmoko: 2007-09-17 Holger Hans Peter Freyther Sep 18 01:38:43 * orpheus too Sep 18 01:38:46 openmoko: Use g_new0 to initialize the BrowserData. Sep 18 01:38:48 openmoko: * src/main.c: Sep 18 01:38:50 openmoko: (main): Sep 18 01:38:51 somebody needs to kick the bot Sep 18 01:38:52 openmoko: 03zecke * r2993 10/trunk/src/target/OM-2007.2/applications/openmoko-browser2/ (4 files in 2 dirs): (log message trimmed) Sep 18 01:38:52 t00na, hm, USB host mode on the Neo working? Sep 18 01:38:53 should be over soon Sep 18 01:38:55 (76 lines omitted) Sep 18 01:39:00 mrj Sep 18 01:39:01 yep Sep 18 01:39:03 eight more revs :} Sep 18 01:39:03 CIA-23, well thanks for that! Sep 18 01:39:26 CIA-23: thanks for omitting the extra 76 lines :) Sep 18 01:39:28 mrj: i've got a keyboard and mouse plugged into mine Sep 18 01:39:30 orpheus, oh yeah, I meant you of course Sep 18 01:39:36 orpheus, nice Sep 18 01:40:04 can you switch host/gadget on the fly yet? Sep 18 01:40:27 http://wiki.openmoko.org/wiki/Neo1973_USB_host Sep 18 01:40:32 hurmm? Sep 18 01:40:58 (i submitted the updated patch as a bug, since i don't know the right place to put it) Sep 18 01:41:07 switch host/gadget? Sep 18 01:42:12 meaning does the driver now know how to switch between usb host and gadget modes without, say, rebooting Sep 18 01:42:44 * mjr seems to recall the N770 needed that (according to some web page or something) Sep 18 01:42:54 CIA-23: botsnack Sep 18 01:43:48 do none of the bots in here have botsnack? Sep 18 01:43:51 !botsnack Sep 18 01:43:51 mjr: it should go from host to device (haven't actually really tried), but going from device to host is as simple as an 'echo' to a sys knob Sep 18 01:43:52 * cdbot2 devours the botsnack Sep 18 01:44:01 orpheus, handy Sep 18 01:44:04 yep Sep 18 01:44:20 i'm consoled in over ip-over-bt Sep 18 01:44:58 and i've been trying to get a kernel compiled with support for my usb camera for about 3 days Sep 18 01:49:44 anyways, nice chatting with all Sep 18 01:49:51 but it's going-home time for me Sep 18 01:49:59 night Orph Sep 18 01:50:23 casten: good luck building, since now neither of us can.... :( Sep 18 01:50:30 i'm leaving a build running Sep 18 01:50:38 ok, you too Sep 18 02:01:12 g'night Sep 18 02:06:24 good night mickey|zzZZzz **** ENDING LOGGING AT Tue Sep 18 02:59:59 2007