**** BEGIN LOGGING AT Mon Dec 17 02:59:57 2007 Dec 17 03:16:52 jeremym: rwhitby.net is up Dec 17 03:17:10 thanks. Dec 17 03:17:17 it was never down Dec 17 03:17:38 we changed IP address on Saturday from .60 to .53 - perhaps you had a cached DNS which hadn't timed out Dec 17 03:24:35 rwhitby: you know how to get palm apps running on OM? Dec 17 03:24:41 * duffyd wishful Dec 17 03:24:51 port pose Dec 17 04:38:33 moo Dec 17 07:08:08 good morning Dec 17 07:13:49 * SpeedEvil hasn't decided if it's morning, or if he's going back to sleep. Dec 17 07:25:58 how long does the 'make setup' stage of mmf take? It's downloaded the 140M bz2 Dec 17 07:26:41 MokoMakefile, not Master Mind Flayer Dec 17 07:26:57 Ah - it's fallen over. Dec 17 07:27:09 Error validating server certificate for 'https://svn.openmoko.org:443': Dec 17 07:27:23 * SpeedEvil ponders Dec 17 07:29:07 moin Dec 17 07:29:11 okaay - that seems to work. Dec 17 07:29:54 grrr Dec 17 07:29:55 rwhitby: the initial svn co stage in make setup fails to run again due to the directory openmoko existing Dec 17 07:29:59 why is fakeroot now broken? Dec 17 08:31:27 SpeedEvil: better talk to cesarb about his patch ... Dec 17 09:29:39 Anyone have thoughts as to why 'ERROR: Importing sqlite3 and pysqlite2 failed, please install one of them. A 'python-pysqlite2' like package is likely to be what you need. Dec 17 09:29:44 ' might be happening? Dec 17 09:30:03 I've downloaded, and make install'd it. Dec 17 09:30:59 openmoko: 03erin_yueh * r3657 10/trunk/src/target/gsm/ (10 files in 5 dirs): gsmd: conference call support (Sean Chiang) Dec 17 09:34:31 how to start sshd on the Neo, if it doesn't start by default? "ps -ae | grep ssh" returns nothing Dec 17 09:34:42 it's not called that Dec 17 09:34:47 it's called something else Dec 17 09:34:53 ... Dec 17 09:34:57 SpeedEvil, openssh-server? Dec 17 09:34:58 dropbear? Dec 17 09:35:03 netstat -lpn | grep ':22' Dec 17 09:35:45 Ok - that problem solved - looks like I forgot to ldconfig Dec 17 09:36:01 but now it's saying bitbake openmoko-devel-image uboot-openmoko ) Dec 17 09:36:01 ERROR: Unable to open conf/bitbake.conf Dec 17 09:36:09 * SpeedEvil ponders Dec 17 09:37:47 SpeedEvil, zash, urgh, battery problems, Neo shutsdown Dec 17 09:41:05 * SpeedEvil cuts down (build) trees. Dec 17 10:12:35 rwhitby: I didn't expect the directory to exist but not be a svn checkout (both the stamps file and the entries file have to be missing for it to try to do the checkout) Dec 17 10:12:58 cesarb: that can happen if the initial checkout is aborted Dec 17 10:13:02 rwhitby: In fact, I'd expect the previous version (without a mkdir in that place) to fail as well if the directory exists Dec 17 10:13:39 rwhitby: I see. Would changing it to mkdir -p be enough? Dec 17 10:14:09 we also need to handle the case where trunk is checked out successfully, but developers is not Dec 17 10:14:15 rwhitby: (I tend to prefer having it abort if things are not "clean", but with mkdir -p would behave more like it was before) Dec 17 10:14:31 rwhitby: I thought of that, but didn't want to make the lines even longer ;-) Dec 17 10:14:42 I would split it into two sections Dec 17 10:15:00 then the entries check can be specific to the checkout (trunk or developers) Dec 17 10:15:19 (and do mkdir -p as well) Dec 17 10:15:35 rwhitby: that would work (and be EVEN MORE like the original in terms of behaviour, while using 2x the amount of lines :D ) Dec 17 10:15:46 would probably move the mkdir into it's own line too Dec 17 10:16:00 ok, I'll fix it up Dec 17 10:17:44 similar problem I was having earlier - svn failed due to certificate unknown in make setup Dec 17 10:17:49 hellu Dec 17 10:18:21 then wouldn't retry as the dir is made Dec 17 10:18:23 SpeedEvil: certificate problems usually mean you have to (on Debian-style systems) install or dpkg-reconfigure the ca-certificates package Dec 17 10:18:26 how are things Dec 17 10:18:53 cedric: probably - I just said 'accept permenantly' manually though Dec 17 10:20:16 hmm - the non-developer checkout should be http, not https. Dec 17 10:20:22 i dont mean to be a nag, but was it the power bug alone that set the GTA02 back from Dec 2007 to March 2008? Dec 17 10:21:48 rwhitby: don't look at me, I copy-pasted the URL part of the lines ;-) Dec 17 10:21:52 No - on reflection december was unlikely to be hit. Dec 17 10:22:12 hit? Dec 17 10:22:13 but if the power bug hadn't appeared, maybe january? Dec 17 10:22:19 achieved. Dec 17 10:22:25 oh, ok Dec 17 10:22:37 is there a chance to give back gta01 due to malefunction gps-antenna and get something off when bying a gta02? :D Dec 17 10:23:15 openmoko: 03rwhitby * r172 10mokomakefile/trunk/Makefile: Fixed the error handling of the openmoko svn checkout. Dec 17 10:23:43 cesarb, SpeedEvil: give that a try please, I was unable to test it here cause I'm in the middle of an OE database download. Dec 17 10:23:55 is it difficult to make openmoko apps on a PC? Dec 17 10:24:05 * cesarb abstains from mentioning "mtn"... oops Dec 17 10:24:27 i have an idea i'd like to tryout, but i cant wait any longer for gta02 :) Dec 17 10:27:17 i wonder if there are any other "open" phones in production right now? someone said the first android phone would arrive spring 2008, which now appears to be the same period as gta02 :) Dec 17 10:27:19 rwhitby: looks sane, can't actually test here since it takes hours to do the initial checkout (due to some "small" sound files) and I will leave to work soon Dec 17 10:27:52 Blastur: heh - you're assuming an android phone will be open. Dec 17 10:28:18 Blastur: I don't know if android will be as open as the neo. In fact, I don't think there's any phone as open as the neo on the market currently. Dec 17 10:28:35 ok Dec 17 10:28:55 Blastur: except for perhaps a "make your own" gumstix or similar contraption (sp?) Dec 17 10:28:58 I expect initial android phones to be locked up tight. Especially since they are from US companies and will be marketed to people who are used to subsidies from contracts. Dec 17 10:29:00 cesarb: all current phone-uis suck also Dec 17 10:29:16 well, tbh, i dont know much about the android project .. it seems like they talk a lot about java apps, and provide a good interface to hw through tehre Dec 17 10:29:21 zash: but some suck more than the others... (I have a Razr, for instance) Dec 17 10:29:57 cesarb: my 6 yrs old 3310 has the most logical and smart ui i have ever seen in a phone Dec 17 10:30:10 thats how mutch all phones suck Dec 17 10:30:11 Blastur: that's not nearly as open as the neo. On the neo, you have FULL (down to the metal) control of the main CPU, the PMU, and the audio subsystem Dec 17 10:30:38 Blastur: with consequence that you also fully control things like the LCD, vibrator and backlight (which are solely controled by the CPU) Dec 17 10:30:55 Blastur: in fact, the only parts you don't control fully on a neo are the three radios Dec 17 10:31:29 yeah, but wont i be able to control lcd etc on a normal phone too? like on my current motorola phone, i can put pixels on the screen, make the vibrator go off Dec 17 10:32:47 im just saying that if ppl get to pick between neo, which is fully open, and an android phone which may have better hardware (still, not so open) (.. like EDGE), they may go for the android Dec 17 10:33:28 Blastur: ...and then get hooked on freedom, and end up picking a neo later? sounds like a plan ;-) Dec 17 10:34:21 probably :) Dec 17 10:34:32 how bad is the gta01 in terms of hw bugs? Dec 17 10:34:45 i may consider getting one if the gta02 is delayed until march Dec 17 10:36:21 Blastur: well, as far as I recall the only hardware "bug" (more like lack of a feature) is that the BT wakeup line is not brought to a CPU interrupt pin Dec 17 10:36:43 but it should also have things like power bug, right? Dec 17 10:36:49 like the gta02 suffers from Dec 17 10:36:53 Blastur: all the gta01 "bugs" (or lack of useful features like... power management) I can recall so far are on software Dec 17 10:37:01 oh Dec 17 10:37:07 Blastur: ah, there's one true hardware bug: the 850MHz issue Dec 17 10:37:07 it doesnt have pwr management? Dec 17 10:37:16 im not in a 850 mhz area anyway Dec 17 10:37:50 Blastur: and the corresponding software bug (but on the GSM modem side which is a hell get authorization to ship a program to do the update), the 3G SIM bug Dec 17 10:38:40 Blastur: it has power management (very poor) right now. some time ago, it didn't. I'm planning on hacking a bit on the kernel to add more power management. Dec 17 10:39:22 And I'm hoping someone makes the battery charge meter be less unrealistic (shutting off with half the bar on the UI is not my idea of "fun") Dec 17 10:40:08 Blastur: there's another hardware bug I'm recalling right now, that in some units the GPS antenna cable gets pinched by the case Dec 17 10:46:15 cesarb: I've got a patch that hopefully should be going in in a few days to do better battery managment Dec 17 10:46:28 at least better capacity estimation. Dec 17 10:46:42 Finally got stuff unpacked off my backup hd, and a OM build kicked off Dec 17 10:46:50 SpeedEvil: yay Dec 17 10:47:59 SpeedEvil: I'm going to try writing a patch to first do the dynticks stuff (longer IDLE times), and then make SLOW work (which will be an annoying amount of work, since if I'm reading it correctly it changes the clock for EVERYTHING, including things like the UARTs) Dec 17 10:48:18 SpeedEvil: since I don't have much free time, it's gonna take some time to write all that... Dec 17 10:49:05 And I don't even know if I will succeed (mostly because of the UARTs, which could get confused by the clock transition... and I can't hack gllin to compensate. *sigh* closed source...) Dec 17 10:49:08 lol @ uart timing changes Dec 17 10:49:11 I think slow is an order of magnitude more power saving though Dec 17 10:49:26 you can pause gllin Dec 17 10:49:37 while stuff stabilises Dec 17 10:49:46 it's not hard-real-time Dec 17 10:49:51 SpeedEvil: the kernel doesn't know about gllin. the kernel shouldn't know about gllin :| Dec 17 10:50:17 SpeedEvil: the GSM modem should be easier, if I can just drop CTS for a moment while the clock changes Dec 17 10:51:24 And there are all the odds and ends (I2C, I2S, I2etc) which can also get confused but I didn't read much about yet Dec 17 10:56:17 Well, gtg Dec 17 11:16:01 * * OM Bug 1136 has been created by chris(AT)openedhand.com Dec 17 11:16:02 * * GSM panel applet keeps on displaying 'Connected to GSM network' notification Dec 17 11:16:03 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1136 Dec 17 11:32:24 too bad that the Neo headset isn't compatible with other cellulars - my N6120C plays music in the left speaker only Dec 17 11:33:27 inverted also Dec 17 12:11:17 openmoko: 03chris * r3658 10/trunk/src/target/OM-2007.2/applications/openmoko-messages2/ (ChangeLog src/sms-contacts.c src/sms.h): Dec 17 12:11:17 openmoko: * src/sms-contacts.c: (delete_all_added_cb), Dec 17 12:11:17 openmoko: (delete_all_progress_cb), (delete_all_clicked_cb): Dec 17 12:11:17 openmoko: Implement delete-all functionality in the contacts view Dec 17 12:11:17 openmoko: * src/sms.h: Dec 17 12:11:19 openmoko: Remove unused variable Dec 17 12:16:52 mornin Dec 17 12:17:18 hm Dec 17 12:17:22 is there a file manager? Dec 17 12:17:30 or -browser Dec 17 12:18:16 there will be eventually Dec 17 12:18:44 any idea of how miller columns would be on a small screen? Dec 17 12:18:49 +touch Dec 17 12:19:00 miller? Dec 17 12:19:36 http://en.wikipedia.org/wiki/Miller_Columns Dec 17 12:19:53 oh that Dec 17 12:20:21 not interested in doing anything like that myself Dec 17 12:20:52 i just don't much like them Dec 17 12:21:07 given lack of screen space they dont really seem to lend themselevs to a small screen Dec 17 12:28:49 raster: rage. Dec 17 12:31:01 * * OM Bug 1137 has been created by dodji(AT)openedhand.com Dec 17 12:31:02 * * need to add recipe to OE/OM to build expedit on the image Dec 17 12:31:03 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1137 Dec 17 12:39:06 xcasex: what about rage? Dec 17 12:39:46 raster port it. Dec 17 12:40:16 hell no Dec 17 12:40:32 its intended mostly for hd video Dec 17 12:40:34 its way too heavy Dec 17 12:40:40 video will crawl at 1fps Dec 17 12:40:42 or less Dec 17 12:43:19 there will be no vidoe accel for rage Dec 17 12:43:23 thus its software only Dec 17 12:43:36 software only on the neo... playing fullscreen video Dec 17 12:43:37 hahahhahaha Dec 17 12:44:29 Alt Sex Stories Text Repository Dec 17 12:44:51 ah Dec 17 12:45:40 * blindcoder hands SpeedEvil a EWINDOW point :-) Dec 17 12:48:04 Ah, I'd thought that Apple had patented 'single-column miller columns', but the patent was unsucessful: http://forums.appleinsider.com/showthread.php?t=56754 Dec 17 12:50:09 I am trying to develop a simple X app for a test with the neo Dec 17 12:50:16 I am not so experienced with linux Dec 17 12:50:26 any suggestions for the easiest tool chain Dec 17 12:50:42 yes, python Dec 17 12:50:51 sorry has to be C++ Dec 17 12:50:55 forgot to mention Dec 17 12:51:06 I tried for python an my boss shot me down Dec 17 12:51:26 Tek_Grml1, the toolchain supports C++ Dec 17 12:52:05 On what grounds? Dec 17 12:52:10 I know there is C++ support, but I am just wondering what other people use for the complete toolchain Dec 17 12:52:12 c or c++ is fine Dec 17 12:52:14 eg eclipse Dec 17 12:52:18 python is not unreasonable Dec 17 12:52:19 toolchain you mean Dec 17 12:52:21 or toolkit? Dec 17 12:52:21 + etc Dec 17 12:52:32 whole thing Dec 17 12:52:37 The MokoMakefile is a popular way to get a toolchain up and running Dec 17 12:52:55 look on the wiki Dec 17 12:52:57 but it is huge is it not, and requires hours to build Dec 17 12:53:01 I have Dec 17 12:53:10 I am just looking for hints to speed things up Dec 17 12:53:20 Not that many hours - but it is a 'come back tomorrow' in many cases. Dec 17 12:53:28 my experience with the wiki is it is often incomplete for linux newbies such as myself Dec 17 12:53:45 On a related matter - how do I tell it not to build qemu Dec 17 12:53:48 I just want to build a simple app transfer to phone and run Dec 17 12:54:11 like I would in ubunto but compile for the neo Dec 17 12:54:44 that also assumes it builds Dec 17 12:54:52 which it very often does not Dec 17 12:54:52 :) Dec 17 12:55:11 Tek_Grml1: do you actually need to compile for the neo? Dec 17 12:55:27 I don't know exactly I am just starting Dec 17 12:55:28 you can display windows on the neo from the desktop Dec 17 12:55:42 which is completely fine for 99.9% of concept stuff Dec 17 12:55:54 raster: hehe evil :p Dec 17 12:56:38 SpeedEvil: true - assuming your usb connection works sanely anddoesnt have 100ms ping times Dec 17 12:56:39 :) Dec 17 12:56:52 raster: I can only say it works for me Dec 17 12:57:05 but cant I compile it somehow on my desktop and just transfer the binary Dec 17 12:57:09 hehehe Dec 17 12:58:11 Tek_Grml1: yes you can, if you've setup the appropriate build environment. This is non-trivial though. Dec 17 12:58:23 this is why I am looking for hints Dec 17 12:58:27 The other alternative is to use a portable bytecode or interpreted language. Dec 17 12:58:42 no no, must be C++ Dec 17 12:58:43 no choice Dec 17 12:59:07 right now you basically have no choice Dec 17 12:59:10 you need to build openmoko Dec 17 12:59:14 the whole tree Dec 17 12:59:21 then create a .bb recipie for yur app Dec 17 12:59:25 then build a package for it Dec 17 12:59:27 then install that Dec 17 12:59:35 if you want it on the phone and running Dec 17 12:59:41 for non-X stuff, you can install the compiler natively, and do that Dec 17 12:59:46 I haven't got X stuff working Dec 17 12:59:51 that seems pretty retarded Dec 17 13:00:09 You are welcome to create a non-retarded toolchain. Dec 17 13:00:49 you don't need to create a .bb recipee for your app Dec 17 13:01:39 well u can run it by hand Dec 17 13:01:40 once you have build the whole tree you can just cross-compile anything you like with it without having to try to figure out how to create a .bb recipee Dec 17 13:01:46 if u set your environment variables right etc. Dec 17 13:01:54 if it is any sufficiently complex app that needs to link to libs Dec 17 13:01:57 and build many files Dec 17 13:02:00 u'll want a recipe Dec 17 13:02:12 nah, I've build large parts of kde4 without a recipe Dec 17 13:02:15 if its a single .c file uwiht no libs needed u might get away with manual compiling Dec 17 13:02:16 :) Dec 17 13:02:18 All we need is x86-qemu on the phone. Dec 17 13:02:22 so everytime I want to test my code I have to wait 6 hours for it to build????? Dec 17 13:02:23 Then this problem goes away Dec 17 13:02:28 Tek_Grml1: no Dec 17 13:02:32 it only builds your app Dec 17 13:02:39 ok Dec 17 13:02:43 then it is ok Dec 17 13:02:43 Tek_Grml1: oh no Dec 17 13:02:46 you built it all once Dec 17 13:02:48 SpeedEvil: yeah! that'll be fast to! Dec 17 13:02:51 then just rebuild your app Dec 17 13:03:04 you only need to wait for just your app to rebuild Dec 17 13:03:11 phew Dec 17 13:03:11 Zanoi_: ustilising JTL technology. (Just Too Late) Dec 17 13:03:40 that is not so retarded then Dec 17 13:03:48 just me Dec 17 13:27:22 openmoko: 03chris * r3659 10/trunk/src/target/OM-2007.2/applications/openmoko-messages2/ (ChangeLog src/sms-notes.c): Dec 17 13:27:22 openmoko: * src/sms-notes.c: (delete_clicked_cb): Dec 17 13:27:22 openmoko: Implement delete functionality in notes view (allows deleting of Dec 17 13:27:22 openmoko: individual messages) Dec 17 13:28:31 yeah!!we're growing...not so fast but we're growing!!! Dec 17 14:07:12 anyone tryed to make cals from neo my calling partner hear himself (echo) . someone tell me what i need to change in alsa settings to avoid this? Dec 17 14:10:06 16 16:08:57 < KrisAbsinthe> It was quite a while ago about the audio: the echo is caused by the phone. If you manually change the settings in alsamixer turn down the amplification and up the actual audio volume to 1 red bar it sounds great, gets rid of that crappy airy sound too Dec 17 14:10:11 16 16:20:32 < KrisAbsinthe> Yes, i edited the alsa settings wile on the phone with my wife, if you do what i typed earlier it ends the echo Dec 17 14:15:23 openmoko: 03chris * r3660 10/trunk/src/target/OM-2007.2/applications/openmoko-messages2/ (ChangeLog src/sms-contacts.c src/sms.h): Dec 17 14:15:23 openmoko: * src/sms-contacts.c: (search_toggled_cb), Dec 17 14:15:23 openmoko: (search_text_changed_cb), (search_combo_changed_cb), Dec 17 14:15:23 openmoko: (contacts_visible_func), (sms_contacts_page_new): Dec 17 14:15:23 openmoko: * src/sms.h: Dec 17 14:15:24 openmoko: Add contacts filtering via search string Dec 17 14:23:17 What's the 'proper' way to remove a package from the om build?I tried cd openembedded/packages/;rm -r qemu Dec 17 14:25:19 but it's now taken about an hour of CPU and done nothing Dec 17 14:25:30 stalled on NOTE: Parsing finished. 4685 cached, 0 parsed, 238 skipped, 0 masked. Dec 17 14:25:34 normal? Dec 17 14:25:44 Or did I break something Dec 17 14:28:23 ls Dec 17 14:28:28 ooops Dec 17 14:34:03 apt-get install sl ;d Dec 17 14:34:58 I decided to change the System_boot.png image, does any one know how to do so? Dec 17 14:36:29 Tek_Grml1: drwx------ 2 Tek_Grml1 Tek_Grml1 4096 Dec 16 14:04 goatse.jpg Dec 17 14:36:39 Tek_Grml1: drwx------ 2 Tek_Grml1 Tek_Grml1 4096 Dec 16 14:05 lovingwithneo.jpg Dec 17 14:36:51 ah, damn, left the folder attribute Dec 17 14:37:09 I was just wondering why you have folders named .jpg... Dec 17 14:37:34 Mek: yeah, I should've done it in an editor, rather than direct in irssi Dec 17 14:39:15 from where could I get package names and their function? Dec 17 14:55:50 openmoko: 03chris * r3661 10/trunk/src/target/OM-2007.2/applications/openmoko-messages2/ (5 files in 2 dirs): (log message trimmed) Dec 17 14:55:50 openmoko: * configure.ac: Dec 17 14:55:50 openmoko: Add AC_GNU_SOURCE for strcasestr Dec 17 14:55:50 openmoko: * src/sms-contacts.c: Dec 17 14:55:50 openmoko: Include config.h first Dec 17 14:55:51 openmoko: * src/sms-notes.c: (search_toggled_cb), (search_text_changed_cb), Dec 17 14:55:53 openmoko: (search_combo_changed_cb), (notes_visible_func), Dec 17 15:04:12 Hmm, 2h cpu Dec 17 15:04:59 SpeedEvil: make clobber-qemu Dec 17 15:05:14 ah, thanks Dec 17 15:05:55 Should remove everything except the source files Dec 17 15:07:12 how to rebuild the a package Dec 17 15:07:15 ? Dec 17 15:07:38 mohamed: If you use the makefile, make rebuild-package- Dec 17 15:07:48 Why rebuild it? :) Dec 17 15:08:42 CM: From where to know package name? Dec 17 15:09:15 Heh, if you don't know what you want to rebuild it's hard to rebuild it Dec 17 15:09:36 mohamed: find . -name \*.bb Dec 17 15:10:16 mohamed: You can also try searching here: http://www.angstrom-distribution.org/repo/ Dec 17 15:11:00 what's the current kernel for the neo? is 2.6.22.5-moko11 up to date? Dec 17 15:11:24 I think so, it's the latest one the makefile builds Dec 17 15:11:25 i ask because 2.6.22.5-moko11 doesn't activate the vibrator anymore... Dec 17 15:11:50 CM: I want to change the system boot image, can I? Dec 17 15:12:08 mohamed: You mean the splash image? Yes, you can. Look in the wiki Dec 17 15:12:09 while u-boot does it (i opened the case recently and wasnt shure if i broke it - gps does not work at all since then) Dec 17 15:12:37 CM: I guess make clobber qemu is not the same as make clobber-qemu... Dec 17 15:12:46 CM: do you know of any problems with vibrator? Dec 17 15:12:58 SpeedEvil: Ouch.. :( Dec 17 15:13:05 Clobbers everything.. Dec 17 15:13:21 emdete: No, but I haven't tried a later build than 071214 Dec 17 15:13:23 CM: it had only gotten up to qem. Dec 17 15:15:03 CM: oups, sorry, my fault. took code from greenphone that does on/off only. the neo has different vibrate intensity! wow!! Dec 17 15:15:09 Hehe Dec 17 15:15:24 emdete: Don't enjoy the vibrations too much.. :P Dec 17 15:15:26 CM: i wrote 1 to it, it needs at least 100 to start... phew, thats cool Dec 17 15:15:45 CM: i do. i don't like sound from gadgets Dec 17 15:15:46 CM: sorry, but could you tell me where in the wiki, I mean the image which I have to wait for loading named System_boot.png I think Dec 17 15:16:39 mohamed: You can use this search engine to find OpenMoko things easier: http://www.google.com/coop/cse?cx=018430699993342716089%3Aszsaurhronw Dec 17 15:17:11 mohamed: The first hit when searching on psplash there is: http://wiki.openmoko.org/wiki/Splash_screen Dec 17 15:19:03 hi Dec 17 15:22:29 i want me openmoko!! Dec 17 15:23:04 linuxxr: openmoko is free Dec 17 15:23:15 with moko makefile :P Dec 17 15:25:48 Can someone point me to a reference to build kernel modules for OM? Dec 17 15:26:59 jmichel: I'm afraid you'll have to search the openembedded wiki for that Dec 17 15:34:17 jmichel: Custom modules? If not then you can find them in the modules .gz Dec 17 15:34:47 jmichel: e.g. http://buildhost.automated.it/OM2007.2/modules-2.6.22.5-moko11+svnr3238-r10-neo1973.tgz Dec 17 15:36:13 abraxa__: yes it is a custom module so I will check in the openembedded wiki as recommended by CM Dec 17 15:36:23 jmichel: Ah okay Dec 17 15:37:38 jmichel: are you wan't to build loadable modules outside of the tree? Dec 17 15:39:22 Writchie: I want to build a module for a usb device that is not in modules.gz Dec 17 15:40:19 ATM there no "kernel headers" package - so you need the full sources for your kernel from the build system. Dec 17 15:41:05 once you have these you can set the CROSS_COMPILE and KERNELDIR for your system and build in any directory. Dec 17 15:41:16 I'll post more detailed instructions on the wiki later. Dec 17 15:41:58 Writchie: ok I just understood what you meant by "outside the tree"... I already have the complete build environment so I guess I can add a folder in my local overlay for my module Dec 17 15:42:09 There is already a USB device built into the kernel and I'm not sure how you get it go away. Dec 17 15:43:37 jmichel: if you use a conventional kernel make file you can build with make ARCH=arm CROSS_COMPILE=OMDIR/build/tmp/cross/arm-angstrom-linux-gnueabi/bin/ Dec 17 15:43:38 KERNELDIR=OMDIR/build/tmp/work/fic-gta01-angstrom-linux-gnueabi/liknux-openmoko-2.6.22.5-moko11+svnr2937-re/linux-2.6.22.5 Dec 17 15:43:38 where OMDIR is the path to the top of your openmoko directory. Dec 17 15:44:09 openmoko: 03mickey * r3662 10/trunk/src/target/OM-2007.2/libraries/libmokogsmd2/libmokogsmd/moko-gsmd-connection.c: libmokogsmd2: silence DEBUG output -- closes #1035 Dec 17 15:44:24 Writchie: ok I will try this it would be great if it is that simple! Dec 17 15:44:56 like most things with linux its simple once you know how ;-) Dec 17 15:45:12 and near impossible until you do ;) Dec 17 15:46:32 oh nice Dec 17 15:46:35 * balrog-kun didn't know CROSS_COMPILE could contain full path Dec 17 15:46:54 jmichel: as I said, i don't know how you "unload" the exiting monolithic driver that controls the resources - its running the usb net. Dec 17 15:47:18 but for sure we will need to be able to load lots of different kinds of usb device drivers eventually. Dec 17 15:48:08 you can probably use BT to get a remote shell. Dec 17 15:49:44 Debug board with a serial console is the best bet for that sort of thing; note that you'll have to prevent gsmd from starting in order to ensure that the serial console remains functional. Dec 17 15:49:59 balrog-kun: it worked fine for me. We just need to find a script to strip a full kernel tree to headers only so can be 7MB instead of 783MB Dec 17 15:50:38 mwester: if you have a serial board ;) Dec 17 15:50:51 Writchie: i don't think that many people want to compile out-of-tree kernel drivers Dec 17 15:50:54 Writchie: there is a package Dec 17 15:50:55 oh Dec 17 15:51:02 maybe there isn't a package Dec 17 15:52:51 balrog-kun: probably not - but its much easier to play with modules. Dec 17 15:53:18 yes Dec 17 15:53:38 i think the script would consist of just three "cp" commands Dec 17 15:53:58 Writchie: 2G microSDs are really cheap. Dec 17 15:54:32 but you can only have one :) Dec 17 15:54:41 Well, you could get a 6G Dec 17 15:55:10 too bad its not externally accessible. Dec 17 15:55:16 Why isn't it? Dec 17 15:55:20 really need two of them - one internal one external Dec 17 15:55:28 nfs over bluetooth Dec 17 15:55:43 that will work Dec 17 15:55:49 Or unless you're actually doing kernel module hacking on the road, over usb Dec 17 15:55:49 should Dec 17 15:56:15 nfs alone could replace the microSD Dec 17 15:56:17 usb-net that is Dec 17 15:56:41 I was surprised that a kernel compile only took an hour Dec 17 15:56:44 BTW anybody know how the low-level drivers get away spin-locking for 1 second Dec 17 15:56:49 (local) Dec 17 15:57:20 Writchie: which driver spin-locks from 1 second? Dec 17 15:57:45 serial for the samsung, waiting for FIFO to clear Dec 17 15:57:55 ah Dec 17 15:58:03 i'm sure it takes less than that - but it looks ugly to see such things Dec 17 15:58:03 yeah, the serial is broken Dec 17 15:58:11 it gets away with it because spinlock is a nop on uniprocessor Dec 17 15:58:31 no - its a disable all interrupts on a unirocessor AFAIK Dec 17 15:58:40 spinlockirq Dec 17 15:58:52 right Dec 17 15:59:07 is anyone else's ipkg update failing with wget --passive-ftp errors? Dec 17 15:59:32 balrog-kun: and why not just disable the one irq: Dec 17 15:59:56 Writchie: you have to talk to ben dooks Dec 17 16:00:03 which is not an easy task in itself Dec 17 16:00:35 well, this driver will need attention in a few weeks Dec 17 16:02:44 i found the no-fifo mode is also broken in serial_s3c2410, it still uses the fifo registers in no-fifo mode Dec 17 16:02:51 probably nobody ever uses it Dec 17 16:03:14 balrog-kun: ack, it looks rather ancient Dec 17 16:09:33 CM: After doing make queu-clobber - it still tries to build qemu. Dec 17 16:09:36 NOTE: Running task 128 of 5289 (ID: 878, /mnt/data/om/openembedded/packages/qemu/qemu-native_20070613.bb, do_configure) Dec 17 16:09:42 (and fails) Dec 17 16:09:49 SpeedEvil: :( Dec 17 16:09:55 thoughts? Dec 17 16:10:07 Have you disabled generating binary locales? Dec 17 16:10:27 Proabbly not Dec 17 16:10:37 ENABLE_BINARY_LOCALE_GENERATION = "0" in local.conf Dec 17 16:10:52 Not sure what else might pull in qemu Dec 17 16:12:28 I did a make clobber;make qemu-local yesterday and it worked just fine.. Dec 17 16:12:57 SpeedEvil: Ahh.. Waid a minute. You're trying to build the old june snapshot. That's not a good one.. Dec 17 16:13:10 I downloaded mmf, and clicked 'go' Dec 17 16:13:36 Hmm.. Should work.. Dec 17 16:15:03 I think it's working so bad because all devs are using update-bleeding-edge and have require conf/distro/include/moko-autorev.inc in their local.conf Dec 17 16:15:30 SpeedEvil: Try putting something like SRCDATE_qemu-native = 20070701 in your local.conf Dec 17 16:15:47 what's the trick for preventing the on-screen keyboard from popping up if a bluetooth keyboard is in use? Dec 17 16:15:49 I rememer I had to do that some time in august for qemu-native to build Dec 17 16:15:57 I just don't want qemu to build Dec 17 16:16:10 Ok Dec 17 16:16:11 as I hacen't got gcc3 installed Dec 17 16:16:20 Ahh.. Dec 17 16:16:21 and I have no use for it Dec 17 16:16:43 Latest 0.9.sometihng should build and work somewhat ok with gcc-4.2.. Dec 17 16:16:57 yes - but it has a check, which is barfing Dec 17 16:17:05 and I can't find where that check is Dec 17 16:17:07 At least according to the bug for in in gentoo bugzilla Dec 17 16:17:45 SpeedEvil: What if you make an ugly hack to get around it and just make a softlink from /usr/bin/gcc-3 to gcc-4.2.2 or so? Dec 17 16:17:53 perhaps, but it should still be possible to just disable qemu in OE Dec 17 16:17:54 Not sure if that would work.. Dec 17 16:19:01 Ah Dec 17 16:19:14 found the qemu source, and run configure --don't bother me Dec 17 16:19:41 hmm Dec 17 16:19:45 nope Dec 17 16:20:55 on 32 bits it likely won't build with a "no register to spill" error Dec 17 16:21:17 it doesn't Dec 17 16:26:00 Ok - that worked. Dec 17 16:26:12 I dramatically simplified the configure and makefiles Dec 17 16:26:21 #!/bin/bash Dec 17 16:26:22 true Dec 17 16:26:25 :) Dec 17 16:27:51 Much simplified. Even I can follow that logic. :) Dec 17 16:28:19 asusim 64 Dec 17 16:28:38 now that's an over-simplification Dec 17 16:28:40 O.o Dec 17 16:29:19 now you saw my router's password Dec 17 16:29:24 /o\ Dec 17 16:37:43 * mwester is burning new firmware into Mononoke's router... Dec 17 16:47:53 ls Dec 17 16:47:55 doh Dec 17 16:48:08 . .. README Dec 17 16:49:55 bash: doh: command not found Dec 17 16:50:24 Dec 17 16:51:16 ipkg install doh Dec 17 16:51:58 ;_; Dec 17 16:53:12 calculating dependencies... 41206 packages will be installed, 0 removed, continue? (y/N) Dec 17 16:54:31 Downloading http://packages.microsoft.com/WindowsVista/doh.ipk Dec 17 16:55:47 lol Dec 17 16:56:03 what a bunch of morons we are Dec 17 16:57:07 speak for yourself Dec 17 16:57:11 balrog-kun: wow ;d Dec 17 17:43:52 openmoko: 03mickey * r3663 10/trunk/src/target/OM-2007.2/libraries/libmokogsmd2/ (3 files in 2 dirs): libmokogsmd2: keep track of GSM cipher status Dec 17 17:44:10 openmoko: 03mickey * r3664 10/trunk/src/target/OM-2007.2/panel-plugins/openmoko-panel-gsm/src/openmoko-panel-gsm.c: openmoko-panel-gsm: indicate cipher status in popup Dec 17 17:58:42 are there daily builds of the rootfs images somewhere on the new buildhost? Dec 17 17:58:48 http://buildhost.openmoko.org/daily/tmp/deploy/glibc/images/neo1973/ has only the kernel Dec 17 18:00:24 at this point of time there are no daily builds any more. try asking on distro-devel mailing list wrt. plans about that Dec 17 18:00:50 (i used to do the builds daily, there were not automatic) Dec 17 18:02:28 i wasn't really interested in the plans, just had on my TODO changing qemu to download the daily builds instead of stable since i saw the "daily" directory appear :p Dec 17 18:02:43 and when i got around to do it, i saw there are only kernel daily builds, but that's still good Dec 17 18:04:18 well, it'd be interested to see whether there are plans for that anyways Dec 17 18:04:20 ;) Dec 17 18:07:23 ugh the result is now the kernel, u-boot and rootfs images have to be downloaded from a different directory each :/ Dec 17 18:07:30 agreed Dec 17 18:07:40 might be a reason to mail distro-devel now :D Dec 17 18:08:03 after all, you're part of the tools team, aren't you? :) Dec 17 18:09:48 ~bon appetit Dec 17 18:09:49 rumour has it, bon appetit is smacznego. Guten Appetit. Eet Smakelijk. God Appetitt. Buon Appetito. Buen apetito Bom Apetite. buen apetito Smaklig måltid!. Hyvää ruokahalua. Bo Proveito Dec 17 18:10:28 i thought it meant stuff your face Dec 17 18:10:58 pfft, silly bot doesn't know about utf-8 Dec 17 18:18:28 mickeyl: yeah, or opening a bug Dec 17 18:19:05 for the moment though i'm considering switching to ScaredyCat's private buildhost as the default download source in qemu Dec 17 18:19:18 ScaredyCat: ping Dec 17 18:19:29 sounds good Dec 17 18:19:32 if that's well maintained Dec 17 18:19:35 lets go for it Dec 17 18:19:52 i'm not sure if it's okay with ScaredyCat's bandwidth :) Dec 17 18:20:03 yes, it has regular builds and consistently named images Dec 17 18:20:11 and less downtime so far Dec 17 18:22:40 yeah Dec 17 18:23:52 consistently named images? Dec 17 18:24:05 how do they differ wrt. stock OE builds? Dec 17 18:25:06 i'm not sure about stock OE, but on buildhost.om.o the scheme changes every couple of weeks, so that the script can't decide which file to download Dec 17 18:25:29 hmm Dec 17 18:25:30 interesting Dec 17 18:25:40 since we're using stock OE that sounds odd Dec 17 18:26:04 *shrug* perhaps ScaredyCat is renaming them after the build process manually Dec 17 18:26:46 or if not the file name, then a directory name, the last filename change was "neo1973-" to "fic-" or "gta01-", i don't remember execatly Dec 17 18:27:05 it's this kind of changes, date format etc Dec 17 18:33:30 right Dec 17 18:33:38 well, this depends on the configuration of the people working on buildhost Dec 17 18:33:48 which after all shows that buildhost is not doing automated builds Dec 17 18:34:00 but just a window into the builds the engineers are doing over there Dec 17 18:35:08 yeah, it's still handy for manual downloads i think Dec 17 18:35:13 but for the script not anymore Dec 17 18:46:34 openmoko: 03mickey * r3667 10/trunk/src/target/OM-2007.2/daemons/neod/src/Makefile.am: neod: backout unintentially commited change Dec 17 19:08:06 Writchie: Conserning the build of modules, when I use "make ARCH=arm" I get some errors likes "unrecognized command line option "-mlittle-endian"" Dec 17 19:08:25 Writchie: Have you ever encountered these errors Dec 17 19:08:46 try CROSS_COMPILE=- Dec 17 19:08:47 jmichel: no Dec 17 19:09:31 what does your make file look like Dec 17 19:12:27 jmichel: you need three settings, ARCH= KERNELDIR= and CROSS_COMPILE= Dec 17 19:12:47 you can do it in the environment or on one line to make. Dec 17 19:12:58 make ARCH=arm CROSS_COMPILE=OMDIR/build/tmp/cross/arm-angstrom-linux-gnueabi/bin/ Dec 17 19:12:59 KERNELDIR=OMDIR/build/tmp/work/fic-gta01-angstrom-linux-gnueabi/liknux-openmoko-2.6.22.5-moko11+svnr2937-re/linux-2.6.22.5 Dec 17 19:13:06 the above all on one line Dec 17 19:13:36 replace OMDIR with path to your top level openmoko directory. Dec 17 19:14:05 Writchie: The makefile is the following... I'm not sure it takes all these into account: http://pastebin.ca/820775 Dec 17 19:14:05 for an out of tree build you run the make command from your source directory Dec 17 19:16:01 hmmm: CROSSPREFIX - current kernel build system uses CROSS_COMPILE= Dec 17 19:17:44 Writchie: and the folder you are refering to in KERNELDIR does not exist in my tree Dec 17 19:18:29 did you build with mokomakefile and specify remove tmp directories? Dec 17 19:18:50 rm work or something like that Dec 17 19:19:08 Writchie: yes with mokomakefile and no I didn't remove any tmp directory Dec 17 19:19:09 you need the kernel tree in build/tmp/work Dec 17 19:19:26 the default mokomakefile removes the work directories IIRC Dec 17 19:20:15 it does that by default? then I may have the change this... where can I do this? Dec 17 19:21:42 is it possible that the kernel tree is now in build/tmp/work/neo1973-angstrom-linux-gnueabi/linux-openmoko-2.6.22.5-moko11+svnr3238-r9/ ? Dec 17 19:22:50 jmichel: possible, i last built from svnr2937 Dec 17 19:23:39 jmichel: i have several trees Dec 17 19:24:04 * mwester has a blue spruce, and several pine trees. Dec 17 19:24:08 including an svnr3238-r10 Dec 17 19:24:24 I have a whole forest of kernel trees ;) Dec 17 19:24:39 :p Dec 17 19:25:07 you'll need to match the one you're using in your machine. Dec 17 19:26:45 can't see thru the trees the forest.... (dutch expression) Dec 17 19:26:49 jmichel: i think the neo1973-anstrum*/linux-epen...svnr3238-r10/linux-26.22 etc is correct Dec 17 19:27:02 with correct spelling that is Dec 17 19:27:26 neo1973-angstrom-linux-gnueabi Dec 17 19:28:51 i've been using the fic-gta01-angstrom-linux-gnueabi/linux-openmoko-2.6.22.5-moko11+svnr2937-r3 Dec 17 19:30:15 Writchie: One last question just to make sure my tree is correct, in build/tmp/work/neo1973-angstrom-linux-gnueabi/linux-openmoko-2.6.22.5-moko11+svnr3238-r9/,I have 3 folders: image/ install/ temp/ and 2 files: defconfig patch-2.6.22.5 Dec 17 19:30:25 Writchie: Is anything missing ? Dec 17 19:30:43 mmm Dec 17 19:31:07 anyone know what happened to timezones in oe ... they seem to be missing now. Dec 17 19:31:35 you need linux-2.6.22.5 Dec 17 19:32:10 Writchie: ok so has it been erased by the Mokomakefile after the build? Dec 17 19:32:14 tmp/work/neo1973-angstrom-linux-gnueabi/linux-openmoko-2.6.22.5-moko11+svnr3238-r10/linux-2.6.22 Dec 17 19:32:20 will be the actual build directory. Dec 17 19:32:54 possibly Dec 17 19:33:37 Writchie: I still haven't update to r10 but I will do this right now... thanks a lot for you help Dec 17 19:36:18 i think you have to change the rmwork in you build/conf/local.conf Dec 17 19:37:15 jmichel: better check the wiki - it was a long time ago i did this Dec 17 19:37:30 ok i'll check for that too Dec 17 19:38:08 you don't want to remove anything so that all of the sources remain Dec 17 19:38:21 unfortunately you need the disk space for that Dec 17 19:39:17 i've been building and test from the fic-gta01-angstrom-linux-gnueabil/linux-openmoko-2.6.22.5-moko11+svnr2937-re/linux-2.6.22.5 tree Dec 17 19:39:44 and the kernel modules work fine so far Dec 17 19:40:30 Do you need to install any cross compiler or it uses the one from the MOKO folder ? Dec 17 19:49:10 jmichel: the build/tmp/cross/arm-angstrom-linux-gnueabi/bin/ tools work fine Dec 17 19:49:21 this is what you are setting with CROSS_COMPILE Dec 17 19:50:10 yeah I just understood this part... thanks :) Dec 17 20:30:33 Does anyone else have a problem with their /tmp/gsm.log filling up at around 300 bytes/second? If I leave my neo on for a few days the tmpfs fills up and I have to reboot. I've seen a few people complain about wget failing and this might be the cause. Dec 17 20:31:41 by wget failures I mean ipkg emitting wget errors Dec 17 20:42:52 apaulsen: no problem here -- the gsm.log I have is 589K (over the course of three days), and as I sit here at my desk, very little is being added to it. Dec 17 20:44:09 Maybe it's worse if you move in and out of coverage. I have pretty strong signal here. Dec 17 20:45:06 quicksand: here's some output http://pastebin.com/m44c5f141 Dec 17 20:52:45 apaulsen: yikes! I get a small burst of activity exactly every five minutes, and each burst adds a couple hundred bytes to the log. Dec 17 20:52:53 apaulsen: om isn't ment to run for long time right now - there is lot of debug output everywhere in the software Dec 17 20:52:57 I get those every second or so Dec 17 20:54:23 that's the only log file that seems to cause any trouble, but I'll just redirect it to /dev/null if it is supposed to behave like this Dec 17 20:55:34 For the record, my results are with libgsmd 0.1+svnr3647-r39 and libmokogsmd2-0 0.1.0+svnr3534-r2 Dec 17 20:55:53 and libmokogsmd0 0.4+svnr2367-r0 Dec 17 20:57:16 I've got libgsmd - 0.1+svnr3657-r39 and libmokogsmd2-0 - 0.1.0+svnr3663-r2 Dec 17 20:57:41 I haven't updated in a couple of days. :) Dec 17 20:58:05 I've been watching this problem for the last week or so Dec 17 20:58:47 It might have always done it and just recently became stable enough that I don't restart it every day :) Dec 17 20:58:53 It looks like /etc/init.d/gsmd is the script that sets up the gsmd output to /tmp/gsm.log Dec 17 20:59:01 yep Dec 17 20:59:06 Should be easy enough to edit that away (temporarily, at least) Dec 17 20:59:37 But can you make calls? ;) Dec 17 20:59:41 I can Dec 17 20:59:44 and receive them Dec 17 21:00:15 Well, there you go. Works fine! Dec 17 21:14:15 /sys/devices/platform/s3c2410-i2c/i2c-adapter/i2c-0/0-0008/battemp << in what units? :) Dec 17 21:52:14 ahven: good question. I get -4 or -5. Dec 17 21:53:05 So I guess that rules out Kelvin. Dec 17 21:55:05 i get 5 or 6 Dec 17 22:11:37 ETA 60 hours to download OE snapshot from downloads.openmoko.org - methinks OM needs better connectivity. Dec 17 22:13:58 rwhitby: elaborate please :) Dec 17 22:21:08 HOW MANY MEGS? Dec 17 22:21:14 oo Dec 17 22:25:07 rwhitby hm.. what are you getting (speed) Dec 17 22:26:13 hm.. its 16msec to paris, abd 100 to washington, 366 to the node wgetting the oe snapshot Dec 17 22:29:36 getting 4.1mbyte/sec from there on my host in ffm (5msec away) Dec 17 22:39:13 can i format an SD card using neo Dec 17 22:39:44 yes Dec 17 22:39:53 just like under linux :P Dec 17 22:40:23 yeah but i am finding it difficult to use terminal in neo Dec 17 22:44:27 log in through ssh? Dec 17 22:46:46 wats the ip ? Dec 17 22:47:51 why are non developers buying the neo? Dec 17 22:48:23 community isn't only about developers Dec 17 22:49:12 xcasex : cause i want to learn and i want to develop Dec 17 22:49:21 new_bie: good reason :) Dec 17 22:49:30 Mononoke: depends on the project phase. Dec 17 22:51:17 new_bie: by default, over USB networking, the Neo's IP is 192.168.0.202 Dec 17 22:51:46 thax quicksand Dec 17 22:52:23 webbrowsing over bluetooth on Neo is nice, until wifi arrives Dec 17 22:53:23 Indeed. Though with power mgmt being as it is, I rarely disconnect it from its USB umbilical cord. Dec 17 23:13:41 roh: I'm getting 1.3K/s Dec 18 00:35:09 LaF0rge, ping Dec 18 01:00:06 openmoko: 03andrew * r3668 10/trunk/src/host/qemu-neo1973/openmoko/ (6 files): Add commands for controlling qemu from inside scripts. Dec 18 01:11:54 hello! Dec 18 01:32:51 Any news about the Wifi driver ? Dec 18 01:34:10 I can make some up -- what would you like to hear? ;-) :p Dec 18 01:34:29 latest news about the driver Dec 18 01:34:43 (no news in the past days that I've seen here) Dec 18 01:34:48 hmm Dec 18 01:34:52 CVirus: it receives/sends files at 360 KB/s through scp on gta02 Dec 18 01:35:09 ^^^ News! :D Dec 18 01:35:14 :-) Dec 18 01:35:21 balrog-kun: I presume it's CPU-bound at that point? Dec 18 01:35:34 well, i don't know what's been already known and what hasn't Dec 18 01:35:49 Hmm.. no it couldn't be -- it should do a couple of MB/s Dec 18 01:35:53 cpu bound ? Dec 18 01:35:53 mwester-road: according to the driver author nope Dec 18 01:36:15 CVirus: encryption - CPU-intensive Dec 18 01:36:26 (sameo) Dec 18 01:36:49 400 KB/s is the max for the chip so sameo thinks the 40KB/s of difference is the overhead of the sdio interface Dec 18 01:37:09 scp -c blowfish should help lowering cpu usage in encryption (if dropbear supports it) Dec 18 01:37:26 hmm, i don't remember exactly what part's limitation was 400 KB/s, Dec 18 01:37:38 but it was not the CPU, and it was a datasheet limitation Dec 18 01:37:48 * mwester-road thinks we can easily replace dropbear with openssh in openmoko on demand; SlugOS does it today. Dec 18 01:38:13 mwester, yes I did that too Dec 18 01:38:22 I can live with 360KB/second on a phone :) Dec 18 01:38:41 Why is OpenMoko using Dropbear in the first place ? Dec 18 01:38:47 Tradition? Dec 18 01:38:59 who uses dropbear ! Dec 18 01:39:22 Only those embedded systems that are totally starved for flash space. We have plenty. Dec 18 01:39:33 ah Dec 18 01:39:48 my satellite reciever uses dropbear Dec 18 01:40:07 Once we get the basics working, I think there will be room for several base images for the neo -- a small one with dropbear, a more full-featured one with openssh, etc. Dec 18 01:40:22 yeah Dec 18 01:42:22 is the automatic kernel flash on ipkg upgrade working now? Dec 18 01:49:00 openmoko: 03andrew * r3669 10/trunk/src/host/qemu-neo1973/ (hw/neo1973.c openmoko/qemu-auto.sh): Invert the nSD_DETECT input pin polarity on the basis that this way it works and the opposite it doesn't. Dec 18 02:00:06 * mwester-road reads that commit comment, and idly wonders what the real hardware might be doing with the input pin polarity, and if perhaps that might be important... :p Dec 18 02:01:53 mwester-road: it fortunately doesn't matter much on the real hardware because the card is not removable Dec 18 02:02:57 on linux it's only used to trigger the rescanning of the slots, which is done on boot whatever the state of the pin Dec 18 02:03:12 so yes, it may be semantically wrong in the kernel but it doesn't matter Dec 18 02:05:45 balrog-kun == andrew? Dec 18 02:05:57 yesh Dec 18 02:06:34 :) Ok. Thanks for the clarification. I guess we all sorta hope we'll have a removable SD card on a future device. Dec 18 02:08:50 yup but the whole SD slot will be redeisgned three times until that happens :) Dec 18 02:09:10 and new drivers written Dec 18 02:59:22 Run for your life! Raster is here! Dec 18 02:59:41 * raster runs for his life **** ENDING LOGGING AT Tue Dec 18 02:59:56 2007