**** BEGIN LOGGING AT Mon Sep 24 02:59:57 2007 Sep 24 03:02:49 roh: it seems i can cat together our kernel and jffs2 images and write the resultant file to the rootfs partition. nand read.e will correctly load and boot it, and the jffs2 fs scans past the kernel and happily finds the real fs to mount once it comes up Sep 24 03:03:02 thanks for the ideas :) Sep 24 03:07:20 cheriff jikes... interresting it works.. Sep 24 03:07:37 hopefully jffs never tries to use the blocks where the kernel lives Sep 24 03:09:32 it doesn't, it whinges quite noisily, but then mounts ok. (I have disabled the write routines in the driver, so we're safe on that front) Sep 24 03:10:16 next up i'll put in something to lie about the partition layout so it'll skip the kernel completely Sep 24 03:35:57 good morning Sep 24 04:15:27 (script) wiki RecentChanges: [[PyFeedRead]] Sep 24 04:16:41 does anyone have a sample GTK+ .bb file that runs from moko/local/packages/? Sep 24 04:48:48 <|R> http://www.xpcgear.com/bta3210.html <- does anybody know if this bluetooth dongle works with linux? Sep 24 04:49:25 <|R> (I'm really unfamiliar with anything bluetooth on linux and i'm trying to prepare :P) Sep 24 04:52:40 So far I've tried 3 usb bluetooth adapters and they've all worked right out of the box Sep 24 04:52:58 But I don't know what chip is in that one Sep 24 04:53:17 <|R> ok, i was wondering if there was a bluetooth USB class or something... Sep 24 04:53:46 Class 2 is the normal one, 10m range Sep 24 04:54:29 <|R> Class 2 has to do with RF output power / law right? I meant USB Class like the Video class which unifies all video device using those specs under a single driver :) Sep 24 04:54:50 Ah, sure :) Sep 24 04:54:57 * CM rubs eyes Sep 24 04:55:02 <|R> :) Sep 24 04:55:05 It's 06:54 here Sep 24 04:55:17 <|R> oh haha, waiting for the day to start then ;) Sep 24 04:55:30 <|R> europe or africa? :) Sep 24 04:55:37 Sweden Sep 24 04:55:48 Just about to go to work :) Sep 24 04:56:06 <|R> hehe, don't! they'll make you... dare i say... work! Sep 24 04:56:15 <|R> haha (yeah 00:53 here, i should go to bed instead :P) Sep 24 04:56:25 For sure ;) Sep 24 04:56:37 * CM gets on the bicycle Sep 24 04:56:45 <|R> hehe, have a good day :) Sep 24 05:53:26 /tet Sep 24 06:49:03 I'm still not having any luck booting from SD :( Sep 24 06:49:15 I followed the directions on the page Sep 24 06:49:32 It flashes some text, and resets Sep 24 06:50:01 any ideas on what I can try? Sep 24 06:51:07 my doesnt' even boot at all... after laying around 2 days (switched off of course) it's battery is more than empty... :( Sep 24 06:53:50 emdete: http://wiki.openmoko.org/wiki/Neo1973_Battery_Charger#from_deep_discharged_battery Sep 24 06:54:12 you probably already have it, but just in case Sep 24 06:54:57 gcb77: thanks... prob is that i don't underdstand why its empty at all.... it was switched off. Sep 24 06:56:13 emdete: heh, there's been a lot of discussion here on that... some of the components are connected directly to the battery, and continue drawing power even when off Sep 24 06:56:15 emdete: you must removed the battery, then it's switched off Sep 24 06:56:31 hi there. Sep 24 06:56:31 morning all Sep 24 06:56:34 fgau: :D Sep 24 06:56:37 is there any dbus guy here? Sep 24 06:56:47 what's an dbug guy? Sep 24 06:56:52 dbus :) Sep 24 06:56:53 oups... Sep 24 06:57:01 s/bug/bus/ Sep 24 06:57:29 i would need to know if dbus is sending a signal when someone is trying to own a name Sep 24 06:57:42 bedboi: don't think so Sep 24 07:00:53 any ideas on how I can troubleshoot booting from card? Sep 24 07:02:55 gcb77: with the factory delivered uboot version, i have no probs Sep 24 07:04:17 fgau: hmm, I updated my uBoot when 2007.2 came out Sep 24 07:06:37 gcb77: is that correct? Sep 24 07:07:02 gcb77: other reported the same issue Sep 24 07:07:37 I curently have qtopia on my neo, when I shut it down it powers off, had been off for a couple of days before I powered it on this morning Sep 24 07:08:03 ScaredyCat: and it startet? battery not empty? Sep 24 07:08:10 yes Sep 24 07:08:14 booted fine Sep 24 07:08:22 The other strange thing is when I ran: mmcinit and ext2ls mmc 0 the phone reset itself Sep 24 07:08:24 puh, i had mine off also and battery was dead Sep 24 07:09:10 gcb77: lokk here: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=799 Sep 24 07:09:20 look^^ Sep 24 07:09:35 ScaredyCat: what does wtopia different when powering down? Sep 24 07:09:56 emdete: note 'using qtopia' - maybe it takes power from the devices properly? Sep 24 07:10:28 echo >0 blah Sep 24 07:10:32 iyswim Sep 24 07:10:32 fgau: thank you, has it been resolved or still onging? Sep 24 07:10:43 I'm thinking of flashing the latest uBoot Sep 24 07:10:55 ScaredyCat: no prob i would do all the scripts for that - is there a llist what echo 0 i had to do? Sep 24 07:12:14 for a in gsm bluetooth display vibration ; do echo "look at wiki : Manually using $a"; done Sep 24 07:12:38 the thing is... Sep 24 07:12:54 so gsm bluetooth display vibration are the devices to look at? what about gps? Sep 24 07:13:12 is that ever powered up ? Sep 24 07:13:15 atm Sep 24 07:13:26 * ScaredyCat revises Sep 24 07:13:28 ScaredyCat: who knows what the default is after power on Sep 24 07:13:49 for a in gsm bluetooth display vibration gps ; do echo "look at wiki : Manually using $a"; done Sep 24 07:14:19 you'd hope stuff got powered u when asked to and not before Sep 24 07:14:39 are there more devices? i2c? Sep 24 07:14:56 yeah Sep 24 07:15:53 power-managed? can i switch off power to any device in the neo using /sys/...? someone said some are connected directly to battery some minutes ago.. Sep 24 07:16:11 directly to the battery? Sep 24 07:16:18 why? Sep 24 07:16:51 gcb77 said "connected directly to the battery, and continue drawing power" - isn't that true? Sep 24 07:18:37 I dunno, do we have shematics? Sep 24 07:18:41 schematics Sep 24 07:18:58 at the moment i can't look at it, but i could to `echo 0` to all files named 'power' in /sys? i saw some called power_on, as i remember Sep 24 07:19:08 schematics? i can't read those :/ Sep 24 07:19:45 the suggestion that my neo with qtopia lasted longer than when running om2007.2 would point to something.... Sep 24 07:20:11 * ScaredyCat decides to poke about Sep 24 07:21:14 emdete: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=770 Sep 24 07:24:32 gcb77: thnx. but what they say is that components are not switched off at shutdown. qtopia seems to shutdown the most important part: the gsm with AT@POFF Sep 24 07:25:35 what is the difference between AT@POFF and echo 0 > /sys....gsm....power? Sep 24 07:29:43 oups... my neo disapeared. :( Sep 24 07:32:15 emdete AT@POFF followed by echo 0 > /sys....gsm....power switches it off. Sep 24 07:32:38 echo 0 > /sys....gsm....power ;; echo 1 > /sys....gsm....power resets it Sep 24 07:33:05 from off to on is just echo 1 > /sys....gsm....power Sep 24 07:33:49 if somebody have GTA01 up & connected now, can you guys run ts_test for a while? Sep 24 07:35:05 I get strange behavior in qemu & want to know if it hardware specific or bug in touchscreen driver Sep 24 07:38:44 what behaviour? Sep 24 07:39:25 roh, random reads, out of screen positioning Sep 24 07:40:09 DOH!!! I just flashed u-boot-gta01bv4-1.2.0+svnnow-r14_a89cbbd27a60e6740772000fd0688ffba1c2576a_0_2938.bin Sep 24 07:40:40 Is the power button supposed to still turn on the NEO? Sep 24 07:41:58 of course Sep 24 07:42:14 wow, a `find /sys/ -name power|wc` gives 81 - are there 81 devices to switch off? Sep 24 07:42:43 roh: I take it its not a good sign that the phone won't turn on at all after flashing uBoot Sep 24 07:42:54 not really. Sep 24 07:43:28 * gcb77 will be getting very intimate with the JTag board and the "Embedded Linux Primer" book this week Sep 24 07:46:32 oh well, I guess I've contributed all I can for today, I'm out. Thanks for the help fgau Sep 24 07:47:03 gcb77: np :) Sep 24 07:50:16 hm.. build through... nice Sep 24 07:50:21 http://people.openmoko.org/roh/untested/ Sep 24 07:58:20 roh: what's new? Sep 24 08:14:37 ScaredyCat: the real challenge is if I can get the pm working better... Sep 24 08:15:44 but I have to wait till a v4 device returns to my desk Sep 24 08:16:31 any news on commercial release date? Sep 24 08:16:56 Dec earliest Sep 24 08:17:52 rwhitby: do you know what the release requirements are? are they documented? Sep 24 08:18:14 no. I expect working hardware is near the top. Sep 24 08:19:03 I've been saying (with no inside information), Nov for GTA02 P0, Jan for P1, who knows for general market release (cause that depends on the software) Sep 24 08:19:34 last we've heard (from mickey|portugal at a conference) is Dec, with no information on which phase that is. Sep 24 08:20:06 are you saying gta02 needs another year... do i understand you correctly? Sep 24 08:20:28 would it be worthwhile for a new hacker to buy a GTA01 instead of waiting for the GTA02? Sep 24 08:20:56 that might depend on your requirements :) Sep 24 08:21:41 these phones are so needed, the available alternatives are simply horrible (in u.s.) Sep 24 08:22:29 nevertheless it's good to know the company 'gets it' so i'm not waiting in vain Sep 24 08:23:15 khaije1: with 'phone' do you mean the hardware only or hardware+software working as everyday phone? Sep 24 08:25:03 emdete: for me a phone to hack would be nice, but what i really _need_ is a phone to use every day Sep 24 08:25:15 what is the preferred method of getting an openmoko emulator going on the PC? I see references to Xoo,Xephr,Qemu etc. Sep 24 08:25:53 qemu is the only emulator of those three... (and should produce the results closest to the real thing) Sep 24 08:25:57 (I think) Sep 24 08:28:21 khaije1: ic is afraid giving a developer-only phone away but i would really like to have a gta02 soon for deveopment. my everydayphone is a different one anyway... Sep 24 08:28:28 s/ic/fic/ Sep 24 08:28:28 emdete meant: khaije1: fic is afraid giving a developer-only phone away but i would really like to have a gta02 soon for deveopment. my everydayphone is a different one anyway... Sep 24 08:31:21 emdete: i just think the potential is an order of magnitude better with a neo than anything out now Sep 24 08:36:54 emdete: Considering gta02s don't even boot without attached debug board I don't know how useful that could be Sep 24 08:37:34 abraxa_: shure, given it boots and acc. graphics works in x i am fine ;) Sep 24 08:38:00 The driver hasn't been written yet either ;x Sep 24 08:38:14 abraxa_: yes, i know :/ Sep 24 08:38:23 roh, I've run ts_harvest and have all X(Y)_measured out of screen size, or wrong. Sep 24 08:38:28 rwhitby: do i understand your estimate to correctly be nov 08 for the gta02? Sep 24 08:39:31 no, I said nov 07 for P0 (select developers), jan 08 for P1 (sale to developers, software not ready) and no estimate for P2 (general market release) Sep 24 08:39:45 those are not official dates, just my guesses. Sep 24 08:40:52 unfortunately I can't upload the ts_harvest.out (can't get qemu connected via ssh) Sep 24 08:50:27 goodday all Sep 24 08:51:18 g'day Sep 24 08:51:25 hi Sep 24 08:51:52 hi Sep 24 08:52:04 hm.. official dates... well.. when we're lucky the remaining gta01 arrive in the us end of the week Sep 24 08:52:23 there will be a mail about that soon Sep 24 08:54:17 * CM is still hoping for a eabi gllin version Sep 24 08:54:45 CM well.. we have one that segfaults (still) Sep 24 08:55:04 Broadcom sucks.. :/ Sep 24 08:55:15 how far is development of a open driver? Sep 24 08:55:30 roh: is there going to be another batch of GTA01's made after this batch due to arrive end of the week? Sep 24 08:55:40 CM well.. they are much more cooperative than gl before Sep 24 08:55:44 (i.e. has supply met demand yet?) Sep 24 08:56:22 rwhitby yes. propably.. but i cannot give you specifics yet (dunno) Sep 24 08:56:51 rwhitby i think we need to... you guys just don't stop ordering ;) Sep 24 08:57:01 so if I would order a GTA01 right now, is there any indication how long it would tkat for it to leave .us? Sep 24 08:57:13 so how many gta01 are sold right now? Sep 24 09:00:57 Mek could be 2 weeks or so when you order 'now' or some weeks when you wait for too long. Sep 24 09:01:09 emdete sorry.. cannot tell you absolute numbers Sep 24 09:01:15 okay... Sep 24 09:02:34 roh: sad... would be nice to hear. even how many fic plans to sell... Sep 24 09:02:50 Mek but i can only guess.. its statistics im looking at, not 100% certain values Sep 24 09:03:28 :) Sep 24 09:03:33 emdete well.. the whole response was much hotter and bigger than we expected. thats also why we basically build every device we can Sep 24 09:04:26 lead times can be long when you don't know how many devices you need first Sep 24 09:07:36 roh: i really, really hope that fic keeps up with the neo. Sep 24 09:07:55 emdete how do you mean keep up? Sep 24 09:08:13 roh: my english sucks, sorry. i mean they do it further... Sep 24 09:08:19 we will not just drop it like palm the foleo if you mean that Sep 24 09:08:29 roh: for how long? Sep 24 09:09:42 emdete well.. thats not decided yet. if there is real demand after gta01 is out we could do more gta01 as long as the parts are available. Sep 24 09:09:52 i think it costs alot all the developers, don't think those several thousands of neos pay much Sep 24 09:10:56 and: ambitious phones get cheaper all the time. what if a OE dev platform is ported to the iphone... that costs 50$ less than a neo 2 Sep 24 09:11:14 well.. its a investment info the future for fic, so i don't worry there ;) Sep 24 09:12:06 i don't think you sell one GTA02 for more that the price of the smallest iphone in massmarket Sep 24 09:12:22 emdete even if it would run on a iphone that would have no gps and no removable battery or upgradeable storage Sep 24 09:13:23 roh: shure, but massmarket seems not to care - and iphone's case is so much nicer all the people say that see the neo Sep 24 09:13:26 and it would be 100% reverse-engineering which means no specs, no docs and really buggy drivers for a long time.. not really competition Sep 24 09:13:58 roh: i hope so - don't misunderstand me, i want fic to have success with their model Sep 24 09:19:36 emdete well.. so far we do not worry about that Sep 24 09:20:04 roh: i do a little because i like the project alot ;) Sep 24 09:20:35 well.. were just at the beginning Sep 24 09:26:19 roh: Are you familiar with the plans regarding the UI concept? Sep 24 09:41:52 WOOOT Sep 24 09:42:02 Finally made gstreamer and PulseAudio behave on the Neo :D Sep 24 09:42:09 roh: this safe uboot image that fixes need for nand erase rootfs command?: http://buildhost.openmoko.org/snapshots/2007.08/images/u-boot-gta01bv4-1.2.0+git8993e54b6f397973794f3d6f47d3b3c0c98dd4f6+svn20070819-r0.bin Sep 24 09:42:25 strange, libgsm-tool ask for pin twice. looking in the log i see it switches off the modem with 'AT+CFUN=0' after the first pinentry: http://rafb.net/p/1dT1d719.html - someone also seen that? Sep 24 09:42:45 at least I remember at least 2 persons updating to it safely and also I remember some people accidentally doing nand erase without argument Sep 24 09:42:52 abraxa_: congrats, how did you resolve it? Sep 24 09:43:07 so... it sounds like updating to that uboot might be more safe than doing nand erase something ;-) Sep 24 09:43:38 You can - IIRC use dd to wipe the partition. Sep 24 09:43:45 aloril: I'm using that uboot Sep 24 09:43:59 CM: and it fixes need for nand erase? Sep 24 09:44:00 But i have a broken boot_env Sep 24 09:44:07 hmm.. Sep 24 09:44:08 aloril: yes, so it seems Sep 24 09:44:31 I mean, flashing a 32mb qtopia over a 45mb openmoko rootfs worked fine Sep 24 09:44:38 zecke: pulsesink's buffer and latency settings are optimized for low latency, so by increasing them CPU usage drops from 30->5% (PulseAudio) and 45%->30% (gstreamer) Sep 24 09:45:24 CM: is proken boot_env result of updating to uboot or was it broken before? Sep 24 09:46:04 (you are referring to this I assume: "2007-08-28 10:32:32 hhf423_w: But it's broken. Every time I flash, my uboot_env (mtdblock1) is overwritten with parts of the kernel") Sep 24 09:51:52 anybody else updated uboot? any problems? Sep 24 09:59:27 aloril: yes, and yes Sep 24 10:00:16 * LarstiQ has updated his uboot a couple of times after getting his P1, no problems Sep 24 10:00:44 abraxa_ depends on whats your question Sep 24 10:01:24 aloril dunno, i have a tested one i built myself on http://people.openmoko.org/roh/u-boot-gta01bv4-1.2.0+git20070909+svnr2943-r0.bin Sep 24 10:01:35 roh: I would like to know how the termination of applications will be handled in the future - being able to close them only through openmoko-today is... hmm, counter-intuitive in my opinion Sep 24 10:02:39 abraxa_ they are closed via short press on powerbutton currently Sep 24 10:03:24 roh: Ohh, I did not know that. Okay, thanks :) Sep 24 10:09:06 moin Sep 24 10:15:36 XorA: what kind of problems with uboot update? Sep 24 10:15:47 (script) wiki RecentChanges: [[P1_Owners]] Sep 24 10:16:51 CM: boot_env problem however didn't brick your GTA01B_v04 I assume? Sep 24 10:17:31 aloril: I updated to one from buildhost and bricked bv4 Sep 24 10:17:44 aloril: needed to use JTAG to recover Sep 24 10:18:51 XorA: was it from here or was it earlier/later?: http://buildhost.openmoko.org/snapshots/2007.08/images/ Sep 24 10:19:48 aloril: earlier I think Sep 24 10:20:43 http://wiki.openmoko.org/wiki/Flashing_openmoko doesn't list any known working v4 uboot as far as I can see Sep 24 10:21:25 aloril: got debug board? Sep 24 10:21:41 nope.. thats why I would like to be sure Sep 24 10:21:53 aloril: Id wait then :-) Sep 24 10:22:01 I know this is another option: http://wiki.openmoko.org/wiki/Nand_erase Sep 24 10:22:17 but I have heard people accidentally bricking neo with that command Sep 24 10:22:43 well, I'm swure there's someone with a debug board in your area. Sep 24 10:22:44 aloril: yeah if you arease u-boot, no way to fix without debug board Sep 24 10:23:03 == nand erase without arguments Sep 24 10:24:31 XorA: ahhh. nope. IIRC uboot should be all loaded into memory. so you can actually erase the nand and then flash a new uboot into the device before rebooting. Sep 24 10:24:46 that _should_ work if I remember the bootloader semantics correctly. Sep 24 10:25:10 ixs: didn't know that, but: yes, if you notice error before reboot ;-) Sep 24 10:25:53 aloril: of course. In general though you are much safer fiddling with the bootloader with a debug board as your safety net. Sep 24 10:26:27 aloril in about 20 minutes or so i can test some binary on my device and then tell you if it works on my bv4 and a md5sum Sep 24 10:26:37 yeah... just trying to decide which one is more likely: problem with uboot update or accidental too wide nand erase Sep 24 10:27:19 roh: OK, and I'll update http://wiki.openmoko.org/wiki/Flashing_openmoko page then Sep 24 10:27:34 aloril i did not have any problems with updates since bv2 where the build was ok, so if it works on one moko, it works on all Sep 24 10:27:41 if it matters anything the procedure described at http://wiki.openmoko.org/wiki/OpenMoko_under_QEMU doesn't work for me, fails at the openmoko/flash.sh step (it worked before that august update) Sep 24 10:28:00 hhf423 has tested http://buildhost.openmoko.org/snapshots/2007.08/images/u-boot-gta01bv4-1.2.0+git8993e54b6f397973794f3d6f47d3b3c0c98dd4f6+svn20070819-r0.bin Sep 24 10:28:00 and it worked ok for him(or her) Sep 24 10:28:12 has anyone looked into vala Sep 24 10:28:35 http://live.gnome.org/Vala this one...? Sep 24 10:28:50 also Ghiottone reports it being good, but .. still is it still ok after kernel/image flashing? (ie.. is there boot_env problem with it) Sep 24 10:31:05 viq: Removing the "nand createbbt" line from flash.sh should resolve that issue Sep 24 10:33:05 testing Sep 24 10:34:36 abraxa_: nope Sep 24 10:34:57 viq: Aww =( Sep 24 10:35:43 why is it that the future consumer version will lose the stereo speakers? Sep 24 10:35:59 to make room for the wifi I believe.. Sep 24 10:36:15 so does that mean there isnt support for stereo headphones? Sep 24 10:36:30 Do stereo speakers in such a small container really make a difference? Sep 24 10:36:35 the headphones still be stereo Sep 24 10:36:40 King_InuYasha: stereo headphones, mono speaker. Sep 24 10:36:46 good... Sep 24 10:37:27 there seems to be a discrepency in when the Consumer Version of the OpenMoko phone will come out? Sep 24 10:37:32 btw.. I guess later produced GTA01B_v04 have never uboot without that "needs nand erase" problem? Sep 24 10:37:41 and i found a typo on the website Sep 24 10:38:00 openmoko: 03abraxa * r3034 10/trunk/src/target/OM-2007.2/applications/openmoko-mediaplayer2/src/ (main.c persistent.c persistent.h playback.c): Sep 24 10:38:00 openmoko: Force playbin to use a specific sink again - pulsesink this time Sep 24 10:38:00 openmoko: Make gstreamer/PulseAudio's CPU usage drop to sane values by forcing bigger buffers and allowing higher latency Sep 24 10:38:00 openmoko: Remove trailing slash from default file browser path for the SD card Sep 24 10:38:12 King_InuYasha: you can just fix typo Sep 24 10:38:24 not on a static web page Sep 24 10:38:34 King_InuYasha: about when consumer version comes out: *nobody* knows Sep 24 10:38:48 King_InuYasha: every date you see is a guess Sep 24 10:38:52 aww darn Sep 24 10:39:05 i wanted to get the phone in October like it said on the static page Sep 24 10:39:18 but the wiki says December... Sep 24 10:39:34 October does sound unlikely Sep 24 10:39:41 why is that? Sep 24 10:40:16 and even if it would be released in october, it would be far from a usefull phone, as the software is far from complete... Sep 24 10:40:17 P0 phase hasn't happened for GTA02 for example Sep 24 10:40:34 P0 phase? Sep 24 10:40:37 no distribution system for mass market and what mek said Sep 24 10:40:53 free phones to selected developers to test hardware further Sep 24 10:40:57 Mek, what is left for the software? Sep 24 10:41:58 stability, usability... just minor stuff :P Sep 24 10:42:10 functionality? :P Sep 24 10:42:21 Yes, that too ;) Sep 24 10:42:24 >.< Sep 24 10:48:45 some other stuff: they need to set up mass production line for GTA02, produce them, ship them to outside China Sep 24 10:49:06 each of these... week or few? Sep 24 10:49:39 so even if they decided: lets mass produce them it sounds unlikely for you get one in hand at October Sep 24 10:49:43 aloril: are they already certain what will be inside the final one and have exact plans/maps/whatever for it? Sep 24 10:50:20 viq: they have 2 different configurations and latest info is that they are testing Sep 24 10:50:25 aloril, well, there is a vid up on youtube of gta02's being produced Sep 24 10:50:44 those are probably the p0 ones Sep 24 10:50:47 well, i would like to be able to order one in October Sep 24 10:50:48 cb22: at Taiwan or at China? Sep 24 10:50:58 and maybe hopefully have it by December Sep 24 10:51:12 so that in January, my new budget year starts and I can replace my crappy laptop Sep 24 10:51:26 dunno, http://www.youtube.com/watch?v=Mpt6R4ZNl1w Sep 24 10:52:17 that factory may not be running Windows Sep 24 10:52:29 though it would be sad if it did Sep 24 10:52:36 true Sep 24 10:53:06 my reason for thinking it may not is that Win95-style Window Managers do exist for Linux Sep 24 10:53:18 and a lot of them are low-memory consumers Sep 24 10:53:40 there was a lot time between Taiwan P1 GTA01B_v04 phones for P0 devs and P1 phones sold for devs (>3 months) Sep 24 10:53:53 don't expect this time to be that big delay Sep 24 10:55:38 cb22: that looks like P0 batch being produced (just guess) Sep 24 10:55:46 well.. i now have 2 tested and working uboot on gta01bv04.. Sep 24 10:56:54 roh: cool, which uboot's? (and have you tested flashing kernel and rootfs after uboot flashing?) Sep 24 10:59:01 aloril on people.openmoko.org/roh/ are 2 uboots, in md5sums are the checksums Sep 24 10:59:19 both work fine here Sep 24 10:59:32 one is from the 10th, one is from 'just now' Sep 24 11:01:11 roh, any inside knowledge on that vid :) Sep 24 11:01:41 cb22 of the smd robot? eh.. thats gta02bv2 afaik.. got mone seconds ago (finally) Sep 24 11:02:58 nice, is that the one with all the features, but still the 2410? Sep 24 11:03:03 next time i need to hit the dhl-guy with something sharp and heavy Sep 24 11:03:56 railway spike? Sep 24 11:05:23 hmm.. would do... they forced me getting counted money due to vat Sep 24 11:06:19 huh, now card is visible! I saw some crc errors and repeated boots didn't work, but switching to 512MB card (which didn't get visible) and back to 2GB: now 2GB is visible Sep 24 11:06:39 sounds like a contact problem Sep 24 11:07:04 I thought there was some crc errors in flash that somehow prevented card working and thus thought about reflashing rootfs/kernel ;-) Sep 24 11:07:55 thought it stopped working without opening case, but I guess its kind of 'close' Sep 24 11:07:59 ljp: well it works somewhat at the mo :) Sep 24 11:08:14 anyway, I guess I want to update uboot anyway..., but less hurry now Sep 24 11:11:22 the Dude Where's My Car? wishlist item sounds quite nice actually Sep 24 11:13:07 King_InuYasha, link? Sep 24 11:17:56 what if new bad blocks happen? will reflashing automatically fix it or is something else needed? Sep 24 11:18:10 or is there then need to manually mark it from uboot? Sep 24 11:19:17 aloril well you should be able to update your uboot now without problems. Sep 24 11:19:52 c22, its on the software wishlist Sep 24 11:20:03 errm cb22, its on the software wishlist Sep 24 11:20:37 on the wiki Sep 24 11:21:11 roh: ok, I'll do that tomorrow and then if all seems ok, update wiki (I have moko9_r8 uboot from May 19 installed) Sep 24 11:21:58 King_InuYasha, does sound cool Sep 24 11:22:10 although not too sure how it would be implimented using one neo only Sep 24 11:24:05 well, if a car has GPS and reports its location, it is possible for a Neo to grab this location and track it i suppose Sep 24 11:24:49 roh: I guess uboot update is not useful for anybody with GTA01B_v04 more recent than X (X being some serial number or day when it was shipped) Sep 24 11:24:50 true, but how would it report its location? Sep 24 11:25:03 i dunno Sep 24 11:25:08 im not an expert on GPS Sep 24 11:25:12 i dont even own a GPS Sep 24 11:25:50 aloril depends.. this one has fixed acm console and modem passthrough as well as no nand erase issues i'm aware of Sep 24 11:26:27 morning Sep 24 11:26:31 ah, ok, no nand erase is older fix, acm console and modem passtrough is new thing Sep 24 11:27:26 nand erase though is "problem that makes stuff not work" + you could accidentally brick neo with it Sep 24 11:28:40 is there any known case of bricking neo with updating to "known good working uboot"? Sep 24 11:29:13 (versus bricking by updating to wrong version and bricking by nand erase without arguments) Sep 24 11:30:00 I mean.. which one would be more recommended: update uboot or use nand erase with right arguments? (asking because I wonder what wiki should recommend) Sep 24 11:33:24 aloril, uboot protects itself agains being overwritten by a version for different model Sep 24 11:34:17 yeah, I think this holds for any uboot May 19 or more recent Sep 24 11:35:00 however, not true for older uboots -> quite many v4 to v3 and v3 to v4 uboot writes in past Sep 24 11:35:27 yes Sep 24 11:35:42 but if you are careful it should be no proble,m Sep 24 11:39:32 roh: thanks for testing Sep 24 12:16:50 (script) wiki RecentChanges: [[User:ChrisLord]] Sep 24 12:26:31 hi all Sep 24 12:30:51 hi florian Sep 24 13:18:08 Listen around 52:20 http://www.lugradio.org/episodes/84 Sep 24 13:29:18 cm: interesting... Sep 24 13:31:06 Well.. Guess we have to prove Jono wrong ;) Sep 24 13:36:25 ooh..ooh..what does it day...what does it day... Sep 24 13:37:25 Greetings and salutations Sep 24 13:41:12 stefan_schmidt: are you there? Sep 24 13:57:45 ok i tried booting openmoko from sd, but am getting a bad partition and bad magic number stuff...i couldnt see the error fully since i am brought to the main boot screen rather soon Sep 24 13:59:32 ...anyone getting the same stuff Sep 24 13:59:49 no, we're all sitting on our hands waiting for v2 to come out. Sep 24 14:00:05 ewon: :) Sep 24 14:00:09 bah Sep 24 14:00:28 sudharsh: I always get the bad magic number problem Sep 24 14:00:45 But booting using Factory Reset works Sep 24 14:01:09 But that's on a regular boot, not from sd Sep 24 14:01:38 CM, ah...wait i have a hunch i did something stupid..i forgot to change CONFIG_MMC_S3C=m to y Sep 24 14:01:42 someone spank me Sep 24 14:04:00 sudharsh: Ask ScaredyCat, he seems to like doing that. :P Sep 24 14:04:15 spanking part? Sep 24 14:04:44 oook then Sep 24 14:04:55 Hehe Sep 24 14:12:02 openmoko: 03chris * r3035 10/trunk/src/target/OM-2007.2/artwork/themes/openmoko-standard-2/gtk-2.0/ (6 files in 2 dirs): Fix header style and add customised button styling for openmoko-dates Sep 24 14:15:30 is the date known yet for the final version of the Neo? Sep 24 14:15:31 MetroDelt: 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 24 14:15:47 MetroDelt: You mean the public release? Sep 24 14:15:58 coutner Sep 24 14:16:00 CM: yep Sep 24 14:16:02 counter Sep 24 14:16:02 http://londerings.novalis.org/wlog/index.php?title=Neo1973_counter Got it, get yours from openmoko.com (and read topic) (1811;252) Sep 24 14:16:24 aloril: Might be time to get a GTA02 counter going? :) Sep 24 14:16:34 MetroDelt: We guess in half a year or so.. Sep 24 14:16:37 hehe, yeah I should update that page Sep 24 14:16:52 [[Neo1973]] page has some estimates though Sep 24 14:16:52 Or I do at least, but I'm just a channel lurker ;) Sep 24 14:16:58 HXD8 Sep 24 14:19:08 CM: is the developerversion kinda usable? Sep 24 14:19:21 Not really Sep 24 14:19:56 So it isnt replacing your cells yet? Sep 24 14:20:03 No Sep 24 14:20:20 Okay, too bad Sep 24 14:20:22 heh, so the oct release date for gta02 has slipped a bit then? Sep 24 14:20:22 The battery time is not even half a day, and I can't call or send/recieve sms Sep 24 14:20:24 keep up the good work! Sep 24 14:20:49 ewon: Hopefully P0 for GTA02 will be November Sep 24 14:20:58 Think mickey|portugal said something like that Sep 24 14:22:06 But so far, no official statements. Only rumours and estimates Sep 24 14:24:39 CM: thanks Sep 24 14:25:41 has there been any further news on the 3d accelerator in gta02? Sep 24 14:25:56 (sorry, I've been flat out and haven't been on channel / reading the lists) Sep 24 14:27:04 somewhat Sep 24 14:27:09 I don't know how current you are Sep 24 14:27:20 basically, all docs will be translated and released by FIC Sep 24 14:30:02 Is it an issue of someone with an NDA cleaning up the docs, or is it just translation? Sep 24 14:30:45 SpeedEvil: They never said *all* docs. Sep 24 14:31:08 Hmm.. Seems like webkit-gtk compile is stuck in an endless loop, anyone else seeing that? Sep 24 14:31:13 they said that FIC would look at the vendor docs under NDA, write some new docs, have those new docs approved for release by the vendor, then release the new docs. Sep 24 14:31:47 (I guess you might have meant "all" in the "any" sense of the word ..) Sep 24 14:32:22 rwhitby: pretty much. And yes, and I was meaning 'translation' in the 'rewrite so it's not covered under the NDA' Sep 24 14:32:40 nod Sep 24 14:33:14 Lynet: It takes a long time to compile, over an hour on my comp Sep 24 14:35:40 CM: I'm coming up on 10+ hours here. I have the build env running in a VM, so I don't expect super fast but I'm kinda starting to get worried. Sep 24 14:42:05 Lynet: Then something must be wrong Sep 24 14:42:18 I build it a few days ago with no problem Sep 24 14:58:51 Lynet: i just builded current webkit some hours ago - when did you start? Sep 24 15:00:19 hey, any idea why the build instructions for qemu don't work anymore? Sep 24 15:01:13 this wouldn't happen if we were a Microsoft project. Sep 24 15:01:21 I'll get my coat. Sep 24 15:03:39 anyone using qemu anymore? Sep 24 15:06:19 webjames: Some people are, but not at all as many as before Sep 24 15:07:27 it used to work, not i'm just getting nothing Sep 24 15:08:21 i have tried the moko make file and the manual setup Sep 24 15:16:26 * mwester believes he has discovered the "needs to be restarted after boot" problem with gsmd. Sep 24 15:16:47 mwester: please do tell :) Sep 24 15:28:19 * mwester completed another test Sep 24 15:29:03 Ok, as with most things, there are multiple problems, perhaps the biggest being that the gsmd as currently written has a very weak algorithm for syncing up command and responses from the GSM modem. Sep 24 15:30:21 What prevents it from doing so almost each and every boot is that at bootup, the tty is set to the normal settings -- which echos every received character back to the terminal. Sep 24 15:31:12 So in the interval from power-up to gsmd actually reading data, the modem has been "talking with itself" for several seconds, with several thousand characters exchanged. Sep 24 15:31:45 So even the "flush" operation done by gsmd seldom actually clears out the backlog of trash coming from the GSM modem. Sep 24 15:32:17 is it possible to just issue a hard reset to the modem? Sep 24 15:32:39 Note with the P1's that were purchased, those had the reset line removed Sep 24 15:32:56 The easiest solution is to simply turn off echo before powering up the modem. Sep 24 15:33:02 It syncs up nicely, then. Sep 24 15:33:06 cool. Sep 24 15:33:37 You can even remove the horrible "dances with sleeps" in the gsmd startup file, and it still works nicely. Sep 24 15:48:13 mwester: Great detective work! :D Sep 24 15:50:45 ~seen zdanek Sep 24 15:50:48 zdanek was last seen on IRC in channel #openmoko, 2d 17h 49m 11s ago, saying: 'cu'. Sep 24 16:17:30 (script) wiki RecentChanges: [[User:Online_Shopping]] [[User:Xqus]] [[P1_Owners]] Sep 24 16:21:50 hi Sep 24 16:28:08 Hi Sep 24 16:36:26 hi there. Sep 24 16:37:36 Hi Sep 24 16:37:47 does anyone have de openmoko cellphone???? Sep 24 16:38:35 no, we're all just here because we think they're pretty Sep 24 16:38:43 heh Sep 24 16:39:04 spontaneous self-organization from chaos Sep 24 16:39:19 Inluxc: some people have received a developer version, but the retail version is out at Christmas Sep 24 16:39:42 ckuethe: self disorganization Sep 24 16:39:52 Inluxc: http://wiki.openmoko.org/wiki/P1_Owners Sep 24 16:39:53 i don't want to be called organized Sep 24 16:40:05 * LarstiQ has one, does that make him organized? Sep 24 16:40:23 Thanks Fradeve Sep 24 16:41:06 * mwester has his third from the left, in the row of computers all alphabetically arranged. Does that make him organized? (or insane?) Sep 24 16:41:29 mwester: yes. Sep 24 16:41:45 * cjb_ie raises his eyebrows at ewon Sep 24 16:41:59 * mwester is unsure of where to file ewon's answer... :p Sep 24 16:43:16 mwester: between X and Z Sep 24 17:03:48 hmmm Sep 24 17:04:00 guys i'm porting some applications i wrote Sep 24 17:04:18 one of them needs a library (provided by another .bb i wrote) Sep 24 17:04:31 i put this in the dependencies of course, it gets built Sep 24 17:04:32 but Sep 24 17:05:05 bedboi: but ? Sep 24 17:05:20 for some reasons the application can't find one of the libraries Sep 24 17:05:34 compile time or runtime? Sep 24 17:05:38 compile time Sep 24 17:07:05 bedboi: wrong linker flags? Sep 24 17:09:17 maybe Sep 24 17:09:32 yep it's because i'm stupid Sep 24 17:09:42 i didn't put it in the configure.in Sep 24 17:10:57 bedboi: :-) Sep 24 17:11:12 pebkac Sep 24 17:11:17 as always Sep 24 17:11:22 C is perfect Sep 24 17:11:56 * bedboi hugs K&R Sep 24 17:17:30 * bedboi prays Sep 24 17:42:00 just for curiosity is there any news about the shipment date? Sep 24 17:45:26 the libraries should get installed in tmp/staging/arm-angstrom-linux-gnueabi/ Sep 24 17:45:28 isn't it? Sep 24 17:57:54 bedboi: december ;) Sep 24 17:58:39 and i really hope this time batches will be a bigger Sep 24 17:59:40 shipment of gta02 or 01? Sep 24 18:01:17 gta01 should be shipping now already Sep 24 18:02:10 no no, the one batch gta01 Sep 24 18:02:24 CM: are they shipping, or they are supposed to Sep 24 18:02:24 ? Sep 24 18:04:03 They're supposed to be, but I haven't heard from anyone getting the new batch Sep 24 18:05:05 NOTE: Multiple libraries (libpientityclient.so.0, libpilow.so.0, libpiutils.so.0, libarpiclient.so.0, libarpi.so.0, libpientity.so.0, libpilowclient.so.0) found and LEAD_SONAME not defined Sep 24 18:05:08 what is this? Sep 24 18:06:24 a warning! Sep 24 18:06:29 * anrp continues being unhelpful Sep 24 18:06:46 anrp: ehehe, i know. Sep 24 18:06:57 the problem is that the application cant find that libpiutils Sep 24 18:07:07 so that's why it's failing Sep 24 18:08:04 i mean in that ipk i have the libraries and some sample applications Sep 24 18:08:14 and they work Sep 24 18:08:54 the problem is if i write an application that needs those libraries, for some reason at compile time it doesn't find them Sep 24 18:16:49 (script) wiki RecentChanges: [[Talk:Contacts/2007.2]] Sep 24 18:21:28 hey all Sep 24 18:34:38 hi Ghiottone Sep 24 18:37:53 hey kristian-m. what's up? Sep 24 18:38:09 learning for a test.. :-/ Sep 24 18:39:13 Ghiottone how are you? thanks for the bootup graph btw Sep 24 18:39:33 tests should not take people away from moko development :P Sep 24 18:40:03 I'm fine thanks. Tonight I hope to have a bit of time to dedicate to the moko. Sep 24 18:40:05 Ghiottone tell that to the prof. ;-) Sep 24 18:40:22 Ghiottone great i envy you Sep 24 18:40:26 which bootup graph? the one I put on my user page? Sep 24 18:40:40 chi stato e'? Sep 24 18:40:49 ops wrong window :P Sep 24 18:41:30 Ghiottone yes, i showed it to the author of cinit, maybe we can port that to oe one day Sep 24 18:43:13 great! Sep 24 18:43:30 that bootup showed strange results however. Sep 24 18:44:10 I'm working on a partial replacement of bootchart in C in the hope to lighten the measurement overhead Sep 24 18:45:19 hmmm there's something wrong in the compiling.... Sep 24 18:45:42 all the libraries are referenced with absolute paths Sep 24 18:46:21 but i don't know why my libraries don't get installed in /hdext/moko/build/tmp/staging/arm-angstrom-linux-gnueabi/ Sep 24 18:47:29 in the install/ dir it's ok, i have all i need Sep 24 18:47:59 kristian-m, in which way is cinit different from upstart? Sep 24 18:48:00 what current uboot-version is able to boot from sd? Sep 24 18:49:37 I have 19 aug version and I'm not able Sep 24 18:59:47 Ghiottone i don't know upsatart in detail but cinit is really lightweight - using a directory per process defining start and shutdown scripts and a deamon flag. the boot sequence can be parallelized... - really neat and fast as hell ;-) Sep 24 19:05:05 cinit sounds somewhat djbesque Sep 24 19:05:28 needs more files though ;( Sep 24 19:05:30 ;) Sep 24 19:08:22 buz actually the only really needed file can be a link to the program itself - but it's still the dir ;-) Sep 24 19:10:45 hello Sep 24 19:10:56 hi telmich Sep 24 19:11:23 telmich we were just talking 'bout cinit Sep 24 19:11:24 buz: it's a bit djb-alike, as I think it's a good way to implement configurations Sep 24 19:11:32 telmich is the author Sep 24 19:11:40 i always hated daemontools Sep 24 19:11:49 buz: well, they are a bit unecessary Sep 24 19:12:09 and djb's live philosophy regarding to foss sucks a bit, imho Sep 24 19:12:27 telmich Ghiottone asked "in which way is cinit different from upstart?" Sep 24 19:13:06 Ghiottone: I don't know upstart good, but upstart is event driven, whilst cinit is service based Sep 24 19:14:05 the aim of cinit is really getting everything out of the hardware, without overhead, so after booting is finished events are not handled by cinit Sep 24 19:14:07 upstart is part hal / dbus as well Sep 24 19:14:58 but mostly, upstart was a NIH thing with canonical Sep 24 19:15:09 NIH? Sep 24 19:15:22 not invented here Sep 24 19:18:50 which is interesting, canonical usually doesnt suffer much of NIH Sep 24 19:19:11 (heck the whole distro was invented elsewhere ;) Sep 24 19:20:14 well, it's pimped up debian Sep 24 19:28:01 back Sep 24 19:28:08 telmich, I see Sep 24 19:28:55 telmich, so it basically start all the possible services as soon as possible? Sep 24 19:29:22 I fear that it will not be good for the neo Sep 24 19:29:29 Ghiottone: yes and it works with dependencies Sep 24 19:29:48 because I think that the bottleneck and not the disk Sep 24 19:30:09 Ghiottone: well, normally it's the disk, why do you think it's not in that case? Sep 24 19:30:30 telmich, because the Neo boots from a nand memory. Sep 24 19:30:55 telmich, in the bootchart I got it seemed that I/O usage was always zero Sep 24 19:31:23 telmich, I'm not sure that this result is correct but it's somewhat reasonable Sep 24 19:31:42 Ghiottone well, but still there are blocking operations which could speed up boot if parallelized Sep 24 19:32:03 Ghiottone: well, what do you think is the bottleneck Sep 24 19:32:39 kristian-m, I'm not that sure. the processor context switch has its cost. I don't know what it is with neo's processor but I fear it to be pretty high Sep 24 19:32:49 telmich, I think it is the processor. Sep 24 19:33:23 but to be sure I'm trying to optimize bootchart (which is a bash script) with some C program Sep 24 19:33:23 Ghiottone: what processor is it? Sep 24 19:33:40 just to get rid of a lot of fork and process execution Sep 24 19:33:50 telmich, give me one sec Sep 24 19:34:01 telmich http://wiki.openmoko.org/wiki/Neo1973_Hardware Sep 24 19:34:29 oh yes. that. :) Sep 24 19:35:23 telmich, what does cinit require to run? Sep 24 19:35:45 ok, don't know much about that one. but if you are sure that the CS are really heavier than heaving an idle processor, than it may make sense not to startup in parallel Sep 24 19:35:52 Ghiottone: libc and shm Sep 24 19:36:02 i think portmap is the current bottleneck Sep 24 19:36:44 portmap is used for RPC programs (such as nfs).. why is that there? Sep 24 19:36:48 why is portmap started btw? Sep 24 19:36:56 telmich, shm? sysV shared memory? Sep 24 19:37:28 telmich, just kernel support or external programs ? Sep 24 19:37:59 Ghiottone: yes, shared memory as in message queues Sep 24 19:38:07 Ghiottone: no external programs Sep 24 19:38:17 cb22: for nfs I guess Sep 24 19:39:11 telmich, I'm looking at latest sources... what is that file without a name? (probably some undisplayable char) Sep 24 19:39:47 this is a non existent file :P Sep 24 19:39:57 eheheh Sep 24 19:40:02 telmich, nice easter egg :) Sep 24 19:40:03 Ghiottone: it's a 'hidden' file to attract intelligent developers... ;-) Sep 24 19:40:40 you could have it start with a dot in next version :) Sep 24 19:40:54 hehe Sep 24 19:42:43 telmich, you seem to like files and directory a lot, don't you? :) Sep 24 19:42:56 telmich, I've never seen such a particular configuration system Sep 24 19:42:57 yep, did you have a look at cconfig? Sep 24 19:44:08 hmm, new gitweb.cgi looks good: http://unix.schottelius.org/cgi-bin/gitweb4.cgi?p=cLinux/cinit.git;a=summary Sep 24 19:45:20 telmich, does makefile support a $ROOT variable? Sep 24 19:45:52 $DESTDIR Sep 24 19:46:05 great Sep 24 19:46:07 if you like I can add $ROOT, I don't care Sep 24 19:47:11 hi @ all.. Correct me if I'm wrong, but the GTA02v3 won't have USB 2.0? Sep 24 19:47:11 I mean the mass market version of the phone Sep 24 19:47:57 yeah, getpid() support works now Sep 24 19:48:01 unique, it won't Sep 24 19:48:12 telmich, I think DESTDIR will do Sep 24 19:48:34 hi Sep 24 19:48:35 :) Sep 24 19:49:01 mjr: well I hoped it would have, cause USB 2.0 should be standard today Sep 24 19:49:36 telmich, are the steps: adjust config, make, make install? Sep 24 19:50:00 Ghiottone: yep; take care about conf/, as I don't use autoconf Sep 24 19:50:09 unique, it isn't in many embedded devices, that's the reality, and the reason is cost, space, and power Sep 24 19:50:57 telmich, I'll just try with the defaults Sep 24 19:51:02 sure, it'd be nice, like many other things it won't have yet either Sep 24 19:51:21 mjr: well I guess I could live with it ;) well how fast will wifi be? Sep 24 19:52:16 g standard, unique Sep 24 19:52:30 it's 802.11b/g so presumably 54 Mbit/s (though 27 Mbit per direction) Sep 24 19:52:57 yeah, means that you probably want to use that for bulk data transfer Sep 24 19:53:12 given a choice Sep 24 19:53:14 mjr: yes...that is fast enough for data-transfer. Sep 24 19:53:31 telmich, what do you use to test boot time reduction results? bootchart? Sep 24 19:53:33 unique: why would you need USB 2.0? Sep 24 19:53:49 Ghiottone: a clock Sep 24 19:53:59 Writchie: copying files from computer to the phone for example Sep 24 19:54:09 Writchie, to push new video clips onto the phone quicker ;) Sep 24 19:54:09 Ghiottone: did not try bootchart yet, because I tried to focus on finishing 0.3 Sep 24 19:54:35 unique: you mean you want "high-speed USB 2.0" - not just USB 2.0 Sep 24 19:54:44 a valid nitpick from Writchie Sep 24 19:55:08 i guess even normal USB 2.0 is faster than USB 1.1 .. or am I wrong ?? Sep 24 19:55:18 not -really, device adverstised as USB 2.0 (like the N800) are misleading when they support only full speed Sep 24 19:56:23 yeah :P Sep 24 19:56:24 as to enabling new functionality, high-speed usb 2.0 would allow the use of external display controllers and giving presentations from the neo or somesuch. Fringe stuff, yeah, but there would be _some_ use cases. Sep 24 19:56:37 N800 is full speed 2.0? Hehe. Sep 24 19:56:56 full speed is 12 mbps high speed is 480mbps Sep 24 19:57:41 2.0 != highspeed although all high speed are necessarily USB 2.0 Sep 24 19:57:48 yeah I know Sep 24 19:58:10 this always confused be in the beginning when reading usb-messages Sep 24 20:00:39 * mjr checks the specs and N800 does high speed (in device mode anyway) Sep 24 20:01:00 telmich, where are the files that cinit installs? Sep 24 20:01:07 anyway, back on topic; what was the topic. Ah well :] Sep 24 20:01:23 mjr: that is good new if it does Sep 24 20:01:33 Ghiottone: in src/ Sep 24 20:01:42 * mwester barges into the channel, disturbing everyone. Sep 24 20:02:13 Anyone know if the libmokoui problem has been fixed, i.e. will the latest builds actually work today? Sep 24 20:02:22 no, I need to know where it install them to Sep 24 20:02:37 i'm yet to find where to plug usb in n800 :) Sep 24 20:02:49 Ghiottone: sbin/ Sep 24 20:03:09 Ghiottone: minimal case is to install cinit to /sbin/cinit Sep 24 20:03:32 * telmich is a big fan of git => git branches are pretty cool Sep 24 20:04:17 I've never used git but just for playing a bit with it in its early times Sep 24 20:05:07 it made big progress, in early times I used cogito, but now git is pretty mature Sep 24 20:05:23 * mjr has yet to buy an N800, don't like Nokia's love-hate-openness thing really much Sep 24 20:06:02 * Writchie is contemplated an n800 - the n770 is a pretty terrific buy Sep 24 20:06:26 is still there some frontend? Sep 24 20:06:38 n770 might be better device - all in all, due to the plastic cover, but with more power in the n800... Sep 24 20:08:56 Ghiottone: no Sep 24 20:09:38 ? Sep 24 20:09:43 is it only cmdline now? Sep 24 20:10:03 ahh, you mean like a gui? it's there in git-1.5 Sep 24 20:10:08 I enjoyed the graph of how the sources were branched and merged Sep 24 20:14:35 ERROR: '/home/gua/moko/openembedded/packages/gcc/gcc-cross-initial_4.1.2.bb' failed Sep 24 20:14:45 someone else experiencing the same thing right now? Sep 24 20:15:52 /home/gua/moko/build/tmp/work/armv4t-angstrom-linux-gnueabi/gcc-cross-initial-4.1.2-r8/gcc-4.1.2/gcc/real.h:78: error: size of array 'test_real_width' is negative Sep 24 20:16:10 this, i believe is the correct error Sep 24 20:16:15 or more specific one Sep 24 20:17:10 (script) wiki RecentChanges: [[Neo1973:_GTA01Bv4_versus_GTA02_comparison]] [[Wish_List_-_Hardware]] Sep 24 20:32:23 telmich, it stops with make[2]: *** No rule to make target `w'. Stop. Sep 24 20:32:32 telmich, am I doing something wrong? Sep 24 20:36:31 oeh Sep 24 20:36:34 target w? Sep 24 20:36:54 Ghiottone: which os, which make? Sep 24 20:38:55 mmm Sep 24 20:39:02 my os is linux Sep 24 20:39:04 gnu make Sep 24 20:39:14 but I'm using the OE cross compiler chain Sep 24 20:39:36 what could it be? Sep 24 20:40:29 hard to tell, that description is very generic Sep 24 20:40:50 yeah Sep 24 20:41:08 post at least 3 screenfuls of text to a pastebin, ending with the error line Sep 24 20:41:11 or so D: Sep 24 20:41:11 it's sort of like trying to tell why your computer doesn't show a desktop from you telling me, it doesn't come up Sep 24 20:43:25 telmich, where is this target w defined? Sep 24 20:43:27 Ghiottone: there should not be any w: target, but try do change to src/ and execute make in there Sep 24 20:45:36 I'll post the errors to pastebin Sep 24 20:45:53 that's good Sep 24 20:48:04 http://pastebin.com/d7e9fd206 Sep 24 20:48:09 that's what I get in install Sep 24 20:49:05 and this is for the compile http://pastebin.com/d2a1effb7 Sep 24 20:49:45 ok, the build is fine Sep 24 20:51:28 can you give the output of make -d install to pastebin? Sep 24 20:52:11 what's -d ? Sep 24 20:52:16 damn Sep 24 20:52:17 debug Sep 24 20:52:33 oh. I have to change the recipe. give me a sec Sep 24 20:52:45 why it's not finding the damn library Sep 24 20:53:08 it should be installed in the standard library path Sep 24 20:56:09 ldconfig Sep 24 20:56:26 how do I copy and paste that huge amount of debug info?!?!? Sep 24 20:57:02 Ghiottone: you can also mail it to me: nico-cinit // at // schottelius.org Sep 24 20:58:20 that's better :) Sep 24 20:58:34 mwester: it is compile time Sep 24 20:58:58 there's something wrong either in my toolchain or in the autotools stuff Sep 24 21:00:46 telmich, on its way Sep 24 21:01:17 ok, will take some minutes, as my mailserver verifies each mail before it accepts it Sep 24 21:03:09 ah, already there Sep 24 21:04:35 Ghiottone: could you do a 'grep -r wd *' in the cinit main tree? Sep 24 21:06:31 it's very strange Sep 24 21:07:58 it will take a bit as I had to reconfigure my Openmoko tree Sep 24 21:08:24 so that it does not clean the working dir Sep 24 21:14:03 ~seen zdanek Sep 24 21:14:06 zdanek was last seen on IRC in channel #openmoko, 2d 23h 12m 29s ago, saying: 'cu'. Sep 24 21:14:56 xorg sucks - it just hungup my keyboard again Sep 24 21:15:58 telmich: tell me something that does not suck Sep 24 21:16:05 everything sucks Sep 24 21:16:20 it can suck slightly less, but it will still definitely suck Sep 24 21:16:25 bedboi: ccollect does not suck Sep 24 21:16:59 C does not suck. but that's the only exception Sep 24 21:17:55 telmich: that ccollect seems not to suck that much Sep 24 21:18:15 What's "ccollect"? Sep 24 21:18:25 sh is suck-less, wmii too Sep 24 21:18:31 wmii definitely is suck-less Sep 24 21:18:55 btw, talking seriously. i'm still having problems Sep 24 21:19:26 i don't know why i get this unresolved libraries compile-time Sep 24 21:19:40 bedboi: unix.schottelius.org/ccollect/ Sep 24 21:19:47 yep Sep 24 21:19:49 bedboi: wmii is pretty cool Sep 24 21:19:53 telmich: it is Sep 24 21:20:12 although I miss windows behind windows like in ion3 Sep 24 21:20:49 http://pastebin.com/m38de24f Sep 24 21:20:52 telmich: i don't Sep 24 21:20:59 i like the way it handles windows Sep 24 21:21:03 telmich, http://pastebin.com/d48d76ae0 Sep 24 21:21:39 hehe, all alessandros are italian Sep 24 21:21:55 the libraries should be here -L/hdext/moko/build/tmp/staging/arm-angstrom-linux-gnueabi/lib but for some reasons the are not Sep 24 21:22:00 why i don't know Sep 24 21:23:02 Ghiottone: that's pretty strange Sep 24 21:24:57 Ghiottone: can you come to #cLinux? perhaps we can clearify the problem in there => other devs in there, too Sep 24 21:27:16 telmich, ok Sep 24 21:33:15 nobody has hints? Sep 24 21:33:27 even if manually copy the .so's in the lib path Sep 24 21:33:38 the compilation fails Sep 24 21:42:06 bedboi, are the DEPENDS in the .bb file correct? Sep 24 21:44:57 Ghiottone: it should be Sep 24 21:45:06 lemme check Sep 24 21:45:36 the dependency (libraries), they got compiled without any problem Sep 24 21:45:49 but for some reason they are not installed in the staging/ Sep 24 21:46:28 hello Sep 24 21:46:44 Ghiottone: DEPENDS = "libglade openmoko-libs pinl" Sep 24 21:46:47 yep its ok. Sep 24 21:46:57 pinl is the package that contains those missing libraries Sep 24 21:47:36 in fact it gets built, but then i don't know why the do_compile of this new application fails when linking (it can't find some of those libraries) Sep 24 21:49:04 it' probably the stage part of the pinl package that is broken Sep 24 21:49:20 did you try rebuilding pinl and then trying again? Sep 24 21:51:37 yep Sep 24 21:51:46 bitbake pinl -c clean Sep 24 21:52:12 hello pH5 Sep 24 21:52:45 morning Sep 24 21:53:24 ive not looked into BT on the neo yet, is it possible to use BT to get the neo onto a network? Sep 24 21:53:30 yes Sep 24 21:53:40 look at [[Manually Using Bluetooth]] on the wiki Sep 24 21:53:45 ok, thanks Sep 24 21:53:58 I haven't done this though. Sep 24 21:54:06 anyone know what the range is like? Sep 24 21:54:10 Psi_: I have - it's an easy pand -c .... Sep 24 21:54:50 (assuming you have a pand access point set up, which is easy to do with a consumer router or NAS that supports openwrt or slugos or angstrom. Sep 24 21:57:20 and has USB Sep 24 21:57:32 Psi_: reportedly sucky. Sep 24 21:57:37 Well under 10m Sep 24 21:57:42 oh, ok Sep 24 21:57:56 There are different class devices with different ranges. Sep 24 21:58:11 The one in the Neo is the 10m version. Sep 24 21:58:57 probably not worth me buying a BT card then, id need 50m for what i was thinking about doing Sep 24 21:59:15 You could always have an array of 25 :) Sep 24 21:59:39 * Psi_ waits patiently for the wiki version Sep 24 21:59:39 For point-point, you could use a dish. Sep 24 21:59:42 er, wifi Sep 24 21:59:43 Careful placement of NSLU2's running SlugOS, and you could reach 50M. Sep 24 22:00:43 na, i wanted to use it to update a moving sign while walking around an event that's in a hall about 50m square Sep 24 22:01:15 doesnt matter, it would have been cool but wasnt necessary Sep 24 22:02:03 bedboi, that does only the clean. try with bitbake -f -c rebuild pinl Sep 24 22:02:57 ok Sep 24 22:03:15 i'm quite a newbie so i'm sure that there's something i'm doing wrong Sep 24 22:05:26 rwhitby: can you add kexec-tools to the feed? Sep 24 22:06:07 * mwester is growing exceedingly lazy, too lazy to type "bitbake kexec-tools" on his build machine :-D Sep 24 22:06:54 mwester: you can do that - just add it to task-openmoko-feed.bb - mickey|portugal is happy for anyone to add to that file. Sep 24 22:07:07 Ah - ok, will do. Sep 24 22:07:12 Ghiottone: now the libraries should be installed in staging isn't it? Sep 24 22:07:33 Ghiottone: btw this is not my case Sep 24 22:09:35 Ghiottone: do you have any clue? Sep 24 22:11:05 bedboi, yeah they should be installed in staging. I would thing that the recipe is not correct. Sep 24 22:11:12 s/thing/think/ Sep 24 22:11:13 Ghiottone meant: bedboi, yeah they should be installed in staging. I would think that the recipe is not correct. Sep 24 22:11:23 Ghiottone: i inherited from autotools Sep 24 22:11:31 so it should issue make install Sep 24 22:12:16 Ghiottone: is there something that i need to do to get my libraries in the staging dir? Sep 24 22:12:24 i guess autotools should od Sep 24 22:12:25 do Sep 24 22:13:14 Ghiottone: http://pastebin.com/mede56df Sep 24 22:13:14 bedboi, sorry I don't know Sep 24 22:13:24 it's just basic Sep 24 22:14:14 make install works for that package? Sep 24 22:18:10 bedboi, maybe the --prefix is not passed correctly. what about the configure logs? Sep 24 22:18:27 let's see Sep 24 22:18:50 btw prefix should be set by bitbake in some magic way Sep 24 22:19:35 --prefix=/usr Sep 24 22:19:50 usr? Sep 24 22:20:01 you aren't running it as root, are you? Sep 24 22:20:05 no Sep 24 22:20:06 :) Sep 24 22:20:09 never in my life Sep 24 22:20:25 man i'm not that crazy Sep 24 22:20:34 but see Sep 24 22:20:42 /hdext/moko/build/tmp/staging/i686-linux/bin/install -c -m 644 'libarpi.pc' '/hdext/moko/build/tmp/work/armv4t-angstrom-linux-gnueabi/pinl-0.1-r0/image/usr/lib/pkgconfig/libarpi.pc' Sep 24 22:20:44 this is ok Sep 24 22:21:30 /bin/sh ../../../arm-angstrom-linux-gnueabi-libtool --mode=install /hdext/moko/build/tmp/stagi Sep 24 22:21:30 ng/i686-linux/bin/install -c 'libpiutils.la' '/hdext/moko/build/tmp/work/armv4t-angstrom-linux Sep 24 22:21:30 -gnueabi/pinl-0.1-r0/image/usr/lib/libpiutils.la' Sep 24 22:21:30 arm-angstrom-linux-gnueabi-libtool: install: warning: relinking `libpiutils.la' Sep 24 22:21:34 maybe the problem is that the automatic rules cannot find which files to add to the package Sep 24 22:21:48 have a look at what the final package contains Sep 24 22:21:56 the final package is ok Sep 24 22:22:00 it's working perfectly Sep 24 22:22:33 mmmm Sep 24 22:22:36 do_stage does not contain anything Sep 24 22:22:42 is there a -dev package_ Sep 24 22:22:43 ? Sep 24 22:22:45 yep Sep 24 22:22:59 ok that's the problem Sep 24 22:23:01 the staging Sep 24 22:23:18 i'm sure i have to specify in the receipe Sep 24 22:24:16 should i "oe_libinstall -so something ${STAGING_LIBDIR}/" Sep 24 22:24:17 ? Sep 24 22:25:05 yep it seems that i need it Sep 24 22:26:01 yay, what was the screenshot command ? Sep 24 22:27:17 bedboi, I remember that mickeyl had me try something like that when I had problems with a package Sep 24 22:27:26 bedboi, maybe it is the right way Sep 24 22:28:14 now i'm trying to find the docs :) Sep 24 22:28:57 thseiler: hb9jnx? Sep 24 22:31:35 Hopscotch: no, sorry. Sep 24 22:32:08 thseiler: hehe - but you know him? Sep 24 22:32:10 Hopscotch: sailer vs seiler... Sep 24 22:32:15 right Sep 24 22:32:28 bedboi, http://www.openembedded.info/user-manual/directories_staging.html Sep 24 22:35:09 now it works Sep 24 22:35:21 at least it install 'em in the stagin Sep 24 22:35:32 Hopscotch: not personally, I was reading once a paper with his name on it, then I got curious and googled... ;-) Sep 24 22:35:42 * bedboi crosses his fingers Sep 24 22:35:48 aaaah :) Sep 24 22:36:02 he's kind of the digital radio god :) Sep 24 22:37:57 (power button + screenshot) => http://chmeee.dyndns.org/om/ruby-glade.png Sep 24 22:38:21 Hopscotch: hehe. I always wanted to get my amateur radio licence so I could do packet radio. Problem was: At that time, the EE courses of EPFL were not sufficient to skip the exam, but we are sliding off topic here... ;-) Sep 24 22:38:26 simple app built w/ glade-3 on PC, running entirely on OpenMoko + Ruby + ruby/gtk/glade bindings Sep 24 22:39:20 the work of tonight involved compiling those bindings. Sep 24 22:39:44 Kero: Cool! Sep 24 22:42:02 raises ruby footprint from about 2MB to 2.5 MB :) Sep 24 22:42:54 adding revolution should be easy enough (as ipk, not as gem) Sep 24 22:43:07 Kero: awesome! Sep 24 22:43:47 will take a few days to smooth issues and publish packages... Sep 24 22:44:01 ...mostly since I won't have time until Fri or Sun Sep 24 22:44:53 Kero: 0.5 Mb are a small price to pay for the possibilities ;-) Sep 24 22:45:09 but yeah, I'm happy that the only problem was compiling. Didn't have to edit a single line of the binding, nor of the glade file or ruby app. Sep 24 22:45:37 decent result for the evening :D Sep 24 22:57:23 bye Sep 24 23:12:36 * thseiler searches for tomorrow... Sep 24 23:13:33 mwester: I'm surprised the libmokoui bug didn't get fixed. Makes me wonder if Thomas even knows about it. Do you happen to know if there's a bug report? Sep 24 23:15:49 is fradeve on here? Sep 24 23:18:59 mwester: great work, I'll test on mine as soon as it compiles Sep 24 23:19:31 mwester: if this fixes all the gsmd problems, or if this plus my patch on bug 766 fixes all gsmd problems, it will be very good Sep 24 23:22:35 out of interest, what is the absolute minimum power the neo can / will / does use when in suspend Sep 24 23:22:56 ie, everything off, including gsm, etc, but turns on instantly on power butten press Sep 24 23:23:17 100uA or so Sep 24 23:23:21 under 100uA Sep 24 23:23:27 uA? Sep 24 23:23:28 I was not able to measure it with my setup Sep 24 23:23:32 like nothing Sep 24 23:23:33 microamps Sep 24 23:23:42 and it would turn on instantly? Sep 24 23:23:47 this is not suspend-RAM, but off Sep 24 23:24:03 In principle - suspend-RAM only needs some 1mW or so. Sep 24 23:24:13 hmm, that would kick ass Sep 24 23:24:16 but I've never measured it anywhere close to this. Sep 24 23:24:25 what did you measure around? Sep 24 23:24:32 Hmm. Sep 24 23:24:37 I need to remeasure that. Sep 24 23:24:48 With GSM and audio off Sep 24 23:25:02 just been looking at the iphone's battery life Sep 24 23:25:04 I think it was around 60mA - 250mW Sep 24 23:25:29 i suppose keeping the ram powered was a biatch? Sep 24 23:25:50 i dont suppose suspend-to-microsd / flash will be avalable any time soon? Sep 24 23:25:54 Can't we suspend-to-disk? Or is reading the flash too slow? Sep 24 23:26:10 cesarb, thinking the same thing Sep 24 23:26:16 if it is slow... Sep 24 23:26:25 suspend to microsd sounds supremely iffy ~_~ Sep 24 23:26:25 it is sure as hell a lot faster then starting up Sep 24 23:26:47 RAM should work well - if there are no hardware bugs Sep 24 23:27:02 the battery life is around a year in principle with suspend-RAM Sep 24 23:27:16 (though you'd probably want to suspend to SD after a month) Sep 24 23:27:22 suppose self discharge becomes an issue there :) Sep 24 23:27:29 cb22: we could probably do a lot better in startup time Sep 24 23:27:33 err Sep 24 23:27:35 not a year Sep 24 23:27:35 true... Sep 24 23:27:40 40 days Sep 24 23:27:45 like removing portmap Sep 24 23:27:46 cb22: given that, for instance, there's no need to probe things (it's all fixed!) Sep 24 23:28:15 yeah Sep 24 23:28:27 any estimates on ideal startup time? Sep 24 23:29:09 cb22: I'd guesstimate that we can get the current kernel boot time (that really fast scrolling), and double it (to allow for userspace)... that would be my initial target Sep 24 23:29:41 cb22: below that it would start getting hard to reduce boot time, but probably would be possible Sep 24 23:29:49 yeah Sep 24 23:32:04 cb22: well, first we have to make gsmd work... looks like mwester did something to it, I'm going to test right now Sep 24 23:32:10 and just removing unnessacery things will give an quick boost Sep 24 23:32:14 like portmap... Sep 24 23:32:19 all 20 or so seconds of it! Sep 24 23:32:30 cesarb, i heard about that, apparently gsmd is behaving now? Sep 24 23:32:44 or do the parallel thing, and multithread boot Sep 24 23:33:22 i remember something about context switching being quite expensive on the neo's processor Sep 24 23:33:51 cb22: should be behaving, if he found the real cause... Sep 24 23:34:04 hope so Sep 24 23:34:08 cb22: but, as you are just complaining about, it's real slow to boot :D Sep 24 23:34:42 * cb22 gets impatient for gta02 Sep 24 23:34:50 * cb22 thinks of cartman and his nintendo wii Sep 24 23:35:00 * juri_ gets impatient for $20M cash. :) Sep 24 23:35:15 cb22: we've not got to even 10% of GTA01's potential, and you are already impatient for GTA02? Sep 24 23:35:38 cb22: http://portal.acm.org/citation.cfm?id=1281700.1281703&coll=PORTAL&dl=ACM&CFID=15151515&CFTOKEN=6184618 Sep 24 23:35:46 yep ^-^ Sep 24 23:35:53 i dont have a '01 to begin with Sep 24 23:36:03 been holding for the '02, with wifi being the killer Sep 24 23:36:08 there are good reasons for pushing for gta02 Sep 24 23:36:23 you can't realistically expect a device to be developed fully when you don't have the device in question Sep 24 23:36:37 cb22: well, first try worked, now for the real test... cold-starting the neo Sep 24 23:36:51 cb22: heh, for me BT was the killer, but GTA01 has BT, so... :D Sep 24 23:37:00 lol Sep 24 23:37:16 yeah BT was killer enough Sep 24 23:37:29 any specs out on the ar6k? Sep 24 23:37:34 cb22: I can perfectly understand why they delayed everything to add BT. I, for one, wouldn't have gotten one without BT Sep 24 23:37:52 and I, for one, have never actually used bt before Sep 24 23:37:55 .. i should, though Sep 24 23:38:01 i second that anrp Sep 24 23:38:06 cb22, afaik no, just a gpl driver Sep 24 23:38:15 thats almost as good as specs D: Sep 24 23:38:23 so no things like sensitivity, etc? Sep 24 23:38:59 and does / will that driver eventually have Really-Cool-Things(tm) in it, like monitor mode and packet injection? Sep 24 23:43:22 cb22: great, it booted fine... gsmd behaved itself Sep 24 23:43:28 cb22: now, let's get a SIM... Sep 24 23:43:40 cesarb, good so far... Sep 24 23:44:57 * cesarb ponders: my SIM is bigger than the SD card which came with the neo! Sep 24 23:45:28 man, this battery loves leaping out of the neo when I open the back Sep 24 23:45:41 and your sd card is 16384 times bigger or so Sep 24 23:45:47 see, that's why they call it the anti-iPhone Sep 24 23:46:15 mjr: ? Sep 24 23:46:23 removable battery :) Sep 24 23:46:41 indeed :] Sep 24 23:47:12 Anyway, yeah, it's crazy how you can get 4 gigs of flash on a thumbnail. Soon to be 8. Sep 24 23:47:52 (not shipped with Neo, just talking in general availability now) Sep 24 23:48:08 mjr: ...when not long ago it was 1440K on a credit-card-sized (sorta) thing... Sep 24 23:49:22 fd is way bigger than any credit card I've seen... Sep 24 23:49:38 yay, looks like _gmsd_ is working fine now... Sep 24 23:49:48 mwester, congrats! Sep 24 23:49:52 HOWEVER Sep 24 23:49:57 what about gsmd though? Sep 24 23:49:59 Yay! Sep 24 23:50:02 something which should have asked gsmd to register didn't... Sep 24 23:50:44 * cesarb clicks the very very small antenna icon and choses two options in sequence with a few seconds wait Sep 24 23:51:13 it works! (or at least displays a signal strength... which is always full here, since the antenna is like just outside the window) Sep 24 23:52:27 but the dialer doesn't want to load :( Sep 24 23:52:46 it shows the "spinning Qt icon" for a while, and then does nothing Sep 24 23:57:35 so, now we have gsmd working fine, and the two fighting GUI thingys which actually manage it not working that well anymore. nice. Sep 25 00:03:40 cesarb you can always add a third thing to fight too... ;-) Sep 25 00:03:44 night Sep 25 00:03:58 nite kristian-m Sep 25 00:17:10 (script) wiki RecentChanges: [[User:Brainey]] [[User:Fwendt]] [[Buying_Interest_List]] Sep 25 00:30:16 p/qui Sep 25 00:46:05 cesarb: The dialer loading problem requires one to revert libmokoui2 -- that's the problem that prevented mickey|portugal from creating a new snapshot before he left :( Sep 25 00:46:55 But even so, the matchbox panel crashes at startup more often than not. Sep 25 00:47:00 mwester-laptop: ah... so THAT was the libmokoui2 issue everyone was talking about! Sep 25 00:47:21 mwester-laptop: i thought the panel crashed when gsmd died... Sep 25 00:47:34 mwester-laptop: (well, not every time gsmd died, but...) Sep 25 00:47:43 No, it seems that gsmd's problems masked some other issues. Sep 25 00:48:58 * mwester-laptop really doesn't want to debug GUI code; he's much more comfortable with backends and kernels :( Sep 25 00:49:57 Also, I discovered that we need a "hangup" menu item on the GSM antenna icon, so that you can at least end a call when the dialer crashes. Sep 25 00:50:51 nite all Sep 25 00:51:19 g'nite sir. Sep 25 00:52:12 any news on the HXD8 hardware? Sep 25 00:52:40 apart from the information in the wiki i mean Sep 25 00:56:17 mwester-laptop: isn't the "power down antenna" item enough? Sep 25 01:02:19 any news on the HXD8 hardware? Sep 25 01:02:22 anyone? Sep 25 01:04:01 my main interest is on the GSM part Sep 25 01:04:46 and whether there's a touch screen Sep 25 01:05:35 'cause I was planning on finally finishing the GUI work i started back in February Sep 25 01:06:27 the gui work you started back in febuary? Sep 25 01:06:38 well, proposals really Sep 25 01:06:46 ah Sep 25 01:06:52 i'd imagine it has a touch screen Sep 25 01:07:26 cb22: yeah, i'd think so too Sep 25 01:08:22 jannu, it does Sep 25 01:08:26 https://svn.openmoko.org/trunk/src/target/kernel/patches/hxd8-core.patch Sep 25 01:08:33 search for touch screen in that patch Sep 25 01:08:54 cb22: thanks Sep 25 01:09:53 cb22: the main question for me was whether to work on touch screen -based GUIs with a wide-screen aspect ratio Sep 25 01:10:12 hmm Sep 25 01:10:50 cb22: and touch screen -based GUIs for wide-screen and phone capabilities Sep 25 01:11:44 cb22: not that 4:3 and 16:9 aspect ratios make a big difference, but still Sep 25 01:12:45 cb22: actually the real issue is screen size in real-world units Sep 25 01:13:26 the hxd is 480x272 IIRc and the neo 480x640 Sep 25 01:17:29 cb22: I meant in centimetres/inches Sep 25 01:17:58 cb22: to determine the area that you actually see Sep 25 01:18:06 cb22: and touch Sep 25 01:33:56 mwester-laptop: btw, the coolest thing about your finding is that it explains why I was sometimes seeing ATZ and ATE0V1 echoed back TWICE Sep 25 01:34:13 mwester-laptop: gsmd sent ATZ, and the modem (with its echo on) sent back ATZ Sep 25 01:34:31 mwester-laptop: the kernel echoed the ATZ back, and thus the modem (still with echo on) sent back (again) ATZ Sep 25 01:34:42 mwester-laptop: the greatest mistery is why it happened only twice :D Sep 25 02:33:27 cesarb: there was a 1-second delay followed by a flush of the serial port buffers built in - that should have been enough to flush out all the garbage, and start afresh. Sep 25 02:35:06 Except for the fact that HW flow control was turned on -- at some point the buffer filled, flow control turned off the transmit of the modem, leaving garbage buffered in the modem's transmit buffer -- which ended up confoosing gsmd. Sep 25 02:36:08 So the workaround is to disable echo. It would also be wise if the code in gsmd would do a series of short delay-flush operations, rather than one long one. Sep 25 02:38:25 * rwhitby points people to mwester's nomination in http://wiki.openmoko.org/wiki/Community_nominees_for_GTA02_P0 ... **** ENDING LOGGING AT Tue Sep 25 02:59:56 2007