**** BEGIN LOGGING AT Thu Dec 18 02:59:56 2008 Dec 18 03:00:11 thats just dumb :) Dec 18 03:00:20 hehe :) Dec 18 03:00:31 gonna go sleep Dec 18 03:00:36 good night man Dec 18 03:00:39 yeah me too, good night Dec 18 03:00:40 nice talkin to ya :) Dec 18 03:00:44 you too Dec 18 03:00:50 bye! Dec 18 03:00:53 bye Dec 18 03:54:56 So... There HAS to be a better wifi gui I can use than Mofi... Mofi isn't working. At all. Dec 18 03:55:03 It starts, but never connects, just hangs. Dec 18 04:00:43 Windows has a good one. Dec 18 04:06:17 THE TERMINAL WORKS BEST Dec 18 04:06:23 SORRY Dec 18 04:08:08 * mwester covers his ears Dec 18 04:08:31 H e l ll lll OOOOO Dec 18 04:11:05 mwester: Actually, every wifi gui I've ever used on windows is utter crap. Dec 18 04:11:20 slaxxin: Yes, but not when you have no real keyboard. :) Dec 18 04:11:46 And not when you're working with WPA2. :P Dec 18 04:12:29 They work better than any of the ones we have on the FR. Dec 18 04:27:32 its allabout the width of the screen Dec 18 04:27:55 which is perfect for my hand Dec 18 04:28:33 an a good pair of dental glasses Dec 18 04:57:10 hey! Dec 18 04:57:47 is there a tool to strip jffs2 images downloaded (uploaded in dfu-util lang) from the device? Dec 18 04:58:05 flashing 250MB is very long.... Dec 18 04:58:26 s/is very long/takes a very long time/ Dec 18 04:58:26 drjnut meant: flashing 250MB takes a very long time.... Dec 18 04:59:53 drjnut: strip in what way? Dec 18 05:00:50 strip in a way that it only takes 50MB or so, cause the end (200Mo or so) is filled with 0xFF anyway Dec 18 05:24:35 i'm shopping for a bluetooth keyboard Dec 18 05:24:51 recommendations? Dec 18 05:24:53 heh Dec 18 05:28:25 do bluetooth keyboards all speak the same protocol? Dec 18 05:28:43 orzo: ThinkGeek has a really interesting one. Dec 18 05:28:46 when a userspace program access a tty device is the driver that is used s3c2410.c? Dec 18 05:28:48 the keyboards i see list phones they are compatible with Dec 18 05:28:54 i saw thinkgeek's laser one Dec 18 05:28:59 but i'm a touch typest Dec 18 05:29:17 Oh, does it not work for touch typers? Dec 18 05:29:18 i think it'd be better to have one i can feel Dec 18 05:29:25 Oh.' Dec 18 05:29:29 Heheh Dec 18 06:19:10 shr settings worked partway 1 time now all say unknown Dec 18 06:39:02 So what's this? SCUMMVM on Freerunner, but no package? :\ Dec 18 06:39:13 Has anyone found any games to work well? Dec 18 06:39:53 freesmartphone.org: 03mickey 07specs * rbc025e7d1fe9 10/ (3 files in 3 dirs): odeviced: add missing description of org.freesmartphone.Device.Audio.Scenario signal. Dec 18 06:40:19 man i wish i had a new fdom image to play with Dec 18 06:40:37 are they updateing fdom at all? Dec 18 06:53:50 good morning Dec 18 07:21:58 Digital_Pioneer: debian has had scummvm package for armel for ages :) Dec 18 07:35:09 man Dec 18 07:35:13 i fucked up my pass through cable Dec 18 07:35:17 it doesnt work at all anymore Dec 18 07:35:27 anybody want to just sell me one Dec 18 07:39:49 which cable? Dec 18 07:52:12 gecco: usb cable that provides power from an external source Dec 18 07:52:31 ah Dec 18 07:52:33 any usb cable that is usb to mini-usb will do Dec 18 07:52:38 you can buy in a shop Dec 18 07:52:45 orzo, I have one of those. Dec 18 07:52:52 i have pay 5 euros Dec 18 07:52:54 orzo, you need a link? Dec 18 07:54:19 moin, raster Dec 18 07:54:28 DocScrutinizer2: ooon! Dec 18 07:54:47 :-) Dec 18 07:55:09 OOON Dec 18 07:55:34 freesmartphone.org: 03mickey 07framework * r99209eb98249 10/framework/config.py: do not error out, if /etc directory can't be found Dec 18 07:55:35 freesmartphone.org: 03mickey 07framework * r7192c7d23bfe 10/framework/subsystems/ousaged/ (helpers.py lowlevel.py): ousaged: gathering resume reason now works for Openmoko devices Dec 18 07:56:15 it's time.... Dec 18 07:56:18 for... my next widget Dec 18 07:56:20 MUHAHAHAHHAHA Dec 18 07:57:53 moin Dec 18 07:58:05 and the europeans awake Dec 18 07:58:27 Oh, I need to write up my widget proposal. Dec 18 07:59:04 raster: s/awake/begin to work/ Dec 18 07:59:22 Sargun: you need no steeenkin widgets Dec 18 07:59:25 actually Dec 18 07:59:26 you do Dec 18 07:59:27 :) Dec 18 07:59:36 all your widgets are belong to us Dec 18 07:59:45 raster, I'm thinking of adapting: http://www.w3.org/TR/widgets/ Dec 18 07:59:53 and proposing it as a standard for all FSO dialers. Dec 18 07:59:59 ^_^ Dec 18 08:00:17 Err, maybe I should go for a devil face >:-) Dec 18 08:00:21 (horns) Dec 18 08:01:25 MUHA-widget? nice :-) Dec 18 08:01:31 henk: yeah, know that. working but still asleep Dec 18 08:01:31 oooh Dec 18 08:01:35 lol Dec 18 08:01:36 them is different widgets Dec 18 08:01:45 my last day TPE for this time Dec 18 08:01:55 DocScrutinizer2: back to wamr and wonderful .de Dec 18 08:01:56 :) Dec 18 08:01:59 warm Dec 18 08:02:14 DocScrutinizer2: hehe, ok, that's true... Dec 18 08:03:32 brrrr *shudder* Dec 18 08:04:03 raster: warm? compared to what? north pole? it's fscking cold in germany! Dec 18 08:04:12 i guess noone can fabricate a battery dummy for me now? Dec 18 08:04:18 (in the office) Dec 18 08:04:38 will take me til new year to get used to this shock Dec 18 08:06:15 henk: its warm relative to 0 kelivn :) Dec 18 08:06:36 hi Dec 18 08:07:07 will teh palm 3245ww keyboard work? Dec 18 08:07:30 are bluetooth keyboards interchangable? Dec 18 08:07:32 like usb Dec 18 08:07:37 raster: not fair! ;) Dec 18 08:08:11 orzo: mostly, yes, there is a standard. but I'd usually check first, since there aren't that many of them and they're very expensive. Dec 18 08:08:51 well check how Dec 18 08:13:16 mick_laptop: could ask... Dec 18 08:13:16 what kinda dummy would you need? Dec 18 08:13:17 sorry, -> mickey Dec 18 08:13:43 i want to attach my neo to my lab PSU Dec 18 08:13:51 duh, what a delay when IRCing via WiFi Dec 18 08:14:06 to be safe though (not a hardware guy), I'd like someone to make me a battery dummy Dec 18 08:14:34 ooh, I'd buy one. Dec 18 08:15:10 so something that fits into the Neo Dec 18 08:15:20 but has 2 cables instead of the actual cells Dec 18 08:15:46 mickey|ICE874: my reply delayed 9minutes, like your train Dec 18 08:16:49 ok, I will ask Candy, just entered my room Dec 18 08:16:52 cya Dec 18 08:16:57 thanx Dec 18 08:16:58 cu Dec 18 08:27:58 can you use bluetooth networking while using a bluetooth keyboard? Dec 18 08:40:30 ok, I'm back Dec 18 08:41:57 mickey: everybody seems a little busy right now. I will try to find your handset, and take a few bat with me, so I could do this for you during xmess Dec 18 08:44:59 freesmartphone.org: 03mickey 07framework * rbcc477049b4a 10/framework/subsystems/ogsmd/modems/ti_calypso/channel.py: ogsmd: [TI CALYPSO] remove %CGREG unsolicited code, we're not using it Dec 18 08:45:01 freesmartphone.org: 03mickey 07framework * r5218a43ed072 10/framework/subsystems/ogsmd/modems/abstract/unsolicited.py: ogsmd: [ABSTRACT] start handling +CGEV (neccessary to detect gprs connection teardown) Dec 18 08:45:06 freesmartphone.org: 03mickey 07framework * rf6e9c6ce88b8 10/ (docs/TODO framework/subsystems/ogsmd/modems/abstract/pdp.py): ogsmd: pdp: do not use ATZ in connect script, do not query for CPIN Dec 18 09:04:04 * beni praises ssh Dec 18 09:09:01 :) Dec 18 10:18:54 btw is anyone interested in buying an gta02 ? Dec 18 10:19:02 i might be selling mine - pm me Dec 18 10:21:53 d-b: if it is extra it would be useful to be able to run some tests on it remotely :) Dec 18 10:22:06 ah? Dec 18 10:22:28 depends. Dec 18 10:22:44 d-b: like measure consumption with different scenarios Dec 18 10:22:53 mmm? Dec 18 10:22:58 i need to use my phone daily so i can't run very long tests Dec 18 10:23:20 like #1158 seems to need Dec 18 10:23:32 which is? Dec 18 10:25:43 a bug Dec 18 10:25:55 d-b: in the battery charging logic Dec 18 10:26:00 ah. Dec 18 10:26:30 when leave my phone to charger it charges it to 100% and then starts to very slowly discharge Dec 18 10:26:56 how odd Dec 18 10:27:10 is this usb / wall Dec 18 10:27:48 i don't think i have noticied it really Dec 18 10:27:54 bar on om... Dec 18 10:28:02 happens with both Dec 18 10:28:13 d-b: PMU logic was totally rewritten in andy-tracking kernel afaik Dec 18 10:28:30 d-b: check the bug report Dec 18 10:28:46 freesmartphone.org: 03mickey 07framework * r083dbfe01549 10/framework/subsystems/ogsmd/modems/abstract/ (mediator.py pdp.py): ogsmd: pdp: make singleton handling more consistent with CallHandler Dec 18 10:29:12 ok. Dec 18 10:29:39 on debian i dno't notice it tho Dec 18 10:30:13 don't* Dec 18 10:31:03 d-b: it's all about the kernel Dec 18 10:31:15 sure Dec 18 10:31:25 i use debian with andy-tracking kernel and see it Dec 18 10:31:33 i'm on the original debian kernel. Dec 18 10:31:43 but i think i see it with the semi-official debian kernel too Dec 18 10:31:49 k Dec 18 10:32:25 d-b: yes, that's what i say in a comment there Dec 18 10:37:59 lindi-: what about the bug where the sdcard kills itself ? Dec 18 10:38:25 d-b: i have not hit that Dec 18 10:38:35 but i have daily backups :) Dec 18 10:38:40 so it dares not hit me Dec 18 10:38:47 rofl. Dec 18 10:40:10 yeah issue is i 'm not always some where i can fix it Dec 18 10:41:10 lindi-: how do you suggest i make a good back up ? Dec 18 10:41:40 d-b: i use rsync just like on my other systems Dec 18 10:42:05 k... i'm going to tar it Dec 18 10:42:21 shouldn't hit any problems with tar cvvf Dec 18 10:42:28 (its in via usb card reader) Dec 18 10:43:00 freesmartphone.org: 03mickey 07gsm0710muxd * rbbf01a25713a 10/src/gsm0710muxd.c: improve usage output and show version. patch courtesy Luca Capello Dec 18 10:44:01 d-b: tar would transfer every file again and again? Dec 18 10:44:08 i only transfer modifications Dec 18 10:44:17 mmm i want a quick restore Dec 18 10:44:55 i have quick restore :) Dec 18 10:48:12 lindi-: are you using ext3 or 2 ? Dec 18 10:48:49 ext2 Dec 18 10:49:13 k Dec 18 10:49:18 i have an 8gb scandisk Dec 18 10:49:32 which gets killed after it halts inncorrectly Dec 18 10:49:37 a few times Dec 18 10:50:06 otherwise i would like the phone and use it -> on debian Dec 18 10:50:22 i have 2G Dec 18 10:53:41 ok backed it up i guess i can use debian for a bit and see how that works. Dec 18 11:05:30 openmoko: 03charlie 07org.openmoko.dev * rbad4ab8693c8 10/ (2 files in 2 dirs): paroli: initial import of the bb file Dec 18 11:11:38 openmoko: 03charlie * r4877 10/developers/charlie/Paroli/ (2 files): add paroli package Dec 18 13:50:03 * Weiss turns off all of Illume's animations, and discovers how much more responsive it makes everything feel Dec 18 14:37:23 hi, I'd like to install FSO on the uSD of my gta02. I built the FSO so I have fso-testing/tmp/deploy/glibc/images/neo1973/fso-image-om-gta02.jffs2 Dec 18 14:37:40 should I just dd it to the second partition of the uSD? Dec 18 14:44:10 Impatient guy.. Dec 18 14:44:25 * CM_work points at mujo Dec 18 14:59:41 hello everybody =) Dec 18 16:32:34 well i'm at a standstill on this s3c2410 gps Dec 18 16:32:41 all because I can't get the USB host to work correctly =( Dec 18 16:51:23 anybody knows how usable is android on the freerunner? Dec 18 17:25:07 Whenever I make or receive a call using SHR, I cannot release the call, and I have to reboot the phone to make or receive calls again. Is there any way to avoid this? Dec 18 17:25:59 How recent is your SHR? Dec 18 17:26:06 yesterday Dec 18 17:26:16 Ah. Don't know then. Dec 18 17:26:32 was there an existing issue on earlier images? Dec 18 17:26:45 I'm using SHR and it mostly works Dec 18 17:26:58 ie.. only few times it exploded on me like you say Dec 18 17:27:41 dent: other then those times the phone functionality works pretty well? Dec 18 17:28:38 there's the echo... apart from that it's ok Dec 18 17:29:07 I think if somebody calls me and hangs up right away, then this triggers the "endless ringing" state, but I didn't investigate this yet Dec 18 17:29:38 once it happened to me that I couldn't hang up a call Dec 18 17:29:42 hrm... you don't recall any particular setting you changed? Dec 18 17:29:59 I'm mostly running with the defaults Dec 18 17:30:01 dent: that's what I'm havning. I press release and nothing happens Dec 18 17:30:05 ok, me too Dec 18 17:30:09 but I'm not sure when did U update for the last time Dec 18 17:30:21 ehm, I mean when did I update Dec 18 17:30:36 might have been ~ 3-4 days, as I'm quite busy ATM Dec 18 17:30:45 I know that feeling Dec 18 17:32:26 I may have switch back to 2008-testing. Less polish, but th ephone works Dec 18 17:33:48 if you gfind some time, you might investigate this issue Dec 18 17:33:54 it's not that difficult... Dec 18 17:34:44 Toaster`: http://wiki.openmoko.org/wiki/FSO#How_to_debug Dec 18 17:34:55 anyway I'm off now, bye Dec 18 17:35:11 Thanks dent Dec 18 18:32:06 how usable is android on the freerunner? Dec 18 18:33:10 tomer: as usable as it can be without suspend, I believe Dec 18 18:33:47 android has no suspend support ? Dec 18 18:33:53 nope Dec 18 18:34:09 is it just because of its kernel ? Dec 18 18:34:12 it goes white when it shall wake up Dec 18 18:34:20 jc, I have no idea Dec 18 18:34:23 oh Dec 18 18:34:32 suspend in latest git kernel seems to work Dec 18 18:34:37 why not try that with android ? Dec 18 18:34:51 try if you like Dec 18 18:35:15 well, I neither have android image nor the time to compile it Dec 18 18:36:29 * Tm_T is currently using SHR && Qtextended Dec 18 18:36:33 I read something on the list that they are working on some menu button emulation for android or whatever it was, maybe thats what is stopping it right now Dec 18 19:23:46 how do i install other languages for the openmoko? Dec 18 19:23:53 shr i mean Dec 18 19:24:41 freesmartphone.org: 03stefan 07framework * r1c32b9fb63f7 10/framework/subsystems/ogpsd/ (eten.py factory.py gpschannel.py): Dec 18 19:24:41 freesmartphone.org: ogpsd: Add EtenDevice for the glofiish devices. Dec 18 19:24:41 freesmartphone.org: This patch also introduces a baudrate config option in ogpsd.serialchannel. Dec 18 19:24:42 the older images had other languages i think Dec 18 19:42:17 freesmartphone.org: 03daniel 07framework * rdef440027b0a 10/framework/subsystems/ogsmd/ (5 files in 3 dirs): Dec 18 19:42:17 freesmartphone.org: ogsmd: Add support for different SMS message types Dec 18 19:42:17 freesmartphone.org: This patch also (hopefully!) fixes the SMS PDU breakage that has been Dec 18 19:42:17 freesmartphone.org: around in master for some time. Dec 18 19:42:17 freesmartphone.org: SMS types are now sms-{deliver,submit} and sms-{deliver,submit}-report Dec 18 20:48:56 Has anyone been able to upload GPS tracks to OSM successfully? Dec 18 20:57:36 lindi-: i found out why the smses where weird on my phone Dec 18 20:57:42 my inbox was full :D Dec 18 20:59:11 Digital_Pioneer: I saw something on the wiki about it Dec 18 20:59:40 Yeah, there's a page in TangoGPS's wiki on it, but it didn't work. It lost the timestamps. Dec 18 21:00:02 Digital_Pioneer: http://wiki.openmoko.org/wiki/TangoGPS#Using_tangoGPS_to_create_tracks_for_OpenStreetMap.org Dec 18 21:00:39 well, its a perl script and format of tangogps track is right beneath it ;) Dec 18 21:02:50 30 print < 31 Dec 18 21:02:50 32 @arr[2] Dec 18 21:02:50 33 @arr[3] Dec 18 21:02:50 34 @arr[4] Dec 18 21:02:51 35 3d Dec 18 21:02:53 36 @arr[5] Dec 18 21:02:55 37 Dec 18 21:02:57 38 Dec 18 21:02:59 39 Dec 18 21:03:01 guess it should be saving the timestamps Dec 18 21:03:29 oooh Dec 18 21:03:36 maybe they're in different format Dec 18 21:03:42 that could depend on your locale Dec 18 21:03:48 might wanna investigate that Dec 18 21:19:44 Anyone tried using deb packages from Hackable:1 on Debian? Dec 18 21:26:38 So... Can I get a keyboard in ScummVM? Heheheh Dec 18 21:26:59 i tried to play the hit george in the head with a shoe game an got bumped off the internet!!!!! Dec 18 21:27:17 that was funny Dec 18 21:27:19 linuxxr: good Dec 18 21:28:56 OK, the mouse gets REALLY screwed up in ScummVM, and there's no way to quit a game because I have no keyboard.... Dec 18 21:31:28 raster: hi do you know how to link a c program against Xlib? (-l parameter Dec 18 21:31:57 I'm trying to find the right one but theres no Xlib or libX or whatever Dec 18 21:33:38 Digital_Pioneer: Debian? Dec 18 21:33:49 azaghal: SHR Dec 18 21:33:56 Ah Dec 18 21:33:56 omg never mind, totally overlooked that, it was -lX11 Dec 18 21:34:24 jc: -lX11 Dec 18 21:34:32 jc: note - u may need Dec 18 21:34:35 raster: yeah, thanks :) Dec 18 21:34:38 --L/usrX11R6/lib Dec 18 21:34:39 depends Dec 18 21:34:42 u'll have to detect Dec 18 21:34:53 just -lX11 worked.. I have it in my lib path Dec 18 21:35:19 I grepped for X in the libraries, went through the list and overlooked X11 right at the top Dec 18 21:36:19 Anyone who can tell me why i'm getting just a black screen after flashing the andriod uImage and jffs2 rootfs? Dec 18 21:36:34 jc: it HAPPENS to be in your lib path Dec 18 21:36:39 but not in everyone's Dec 18 21:36:43 given infinnite number of people, somebody would know for sure Dec 18 21:36:44 and not in every build system Dec 18 21:37:15 in fact its only in more recent times u can get away without -L/usr/X11R6 (for example) Dec 18 21:37:23 yeah I was just figuring out if my desktop is screwed up right at the bottom Dec 18 21:37:24 (or for that matter X11R5...) Dec 18 21:37:46 compiled just a simple app that draws hello world and it ran ok and took no cpu or memory, something else is wrong here... Dec 18 21:38:29 ? Dec 18 21:38:37 must be alsalib or something Dec 18 21:38:38 did it draw? Dec 18 21:38:41 yeah Dec 18 21:38:47 where? Dec 18 21:39:09 well it worked ok, just many other apps hog 100% cpu from time to time Dec 18 21:39:29 it will use some memory Dec 18 21:39:32 but not much Dec 18 21:39:34 x is not a hog Dec 18 21:39:37 its pretty lean Dec 18 21:39:41 yeah it does Dec 18 21:39:43 160KB or so Dec 18 21:39:44 contray to what everyone likes tyo say Dec 18 21:39:50 on a 2GB machine thats nothing :) Dec 18 21:39:52 most people are pretty ignorant of x Dec 18 21:40:02 they blame x for the faults of rteally bad toolkit libraries or apps Dec 18 21:40:07 Xlib is ok Dec 18 21:40:11 x or X ? Dec 18 21:40:16 a well built toolkit lib/app can be nicely lean Dec 18 21:40:21 I'd see the problem in Glib and GTK for the memory usage Dec 18 21:40:27 and expend a bit of resources and it can be very pretty and smooth Dec 18 21:40:29 stupid gnome panel takes 30MB of memory Dec 18 21:40:39 hehehe Dec 18 21:40:39 as does everything else gnome Dec 18 21:40:42 aaah no Dec 18 21:40:43 it doesnt Dec 18 21:40:48 jc: you see, I one reason more to NOT use GNOME Dec 18 21:40:56 *everything else graphical gnome Dec 18 21:41:04 I use gnome, I like it actually Dec 18 21:41:07 again - a fallacy of users who just read top/ps wrongly and dont understand real unix memory usage Dec 18 21:41:17 raster: sure Dec 18 21:41:19 yeah I was talking about real usage Dec 18 21:41:21 not virtual Dec 18 21:41:33 "real" memory usage is RESS/RES -SHARE Dec 18 21:41:34 err Dec 18 21:41:38 RSS/REs -SHARE Dec 18 21:41:51 ytou need to take into accoutn SHARE Dec 18 21:41:53 oh thats too much to grasp for me Dec 18 21:42:04 I'm talking stack, bss and stuff, this is too complicated Dec 18 21:42:06 thats share memoyr that it shares for executable + libs + mappd files Dec 18 21:42:17 every gnome/gtk process will load the same libgttk for example Dec 18 21:42:20 they share the same instance Dec 18 21:42:23 of course Dec 18 21:42:28 in fact libc too - EVERY process shares that Dec 18 21:42:33 thats what I'm talking about Dec 18 21:42:40 its REALLY hard sayin "app x uses X ram" Dec 18 21:42:46 its nigh impossible to work out Dec 18 21:42:48 every stupid gnome program hogs memory with instance of Glib, GTK, dbus, alsalib, ... Dec 18 21:43:02 the only way is to measure the system as a whole Dec 18 21:43:10 well Dec 18 21:43:20 a real measurement of an app is - boot your chouse of environment/.system up to a point Dec 18 21:43:23 use "free" Dec 18 21:43:33 I run firefox on fluxbox, about 20% memory usage Dec 18 21:43:40 altogether Dec 18 21:43:44 8:43AM ~ > free -m Dec 18 21:43:44 total used free shared buffers cached Dec 18 21:43:44 Mem: 3285 926 2359 0 261 351 Dec 18 21:43:44 -/+ buffers/cache: 312 2972 Dec 18 21:43:44 Swap: 3820 0 3820 Dec 18 21:43:46 I run it in a gnome session, about 60% Dec 18 21:43:46 for example Dec 18 21:43:49 without cache Dec 18 21:43:58 see the used +- buffers/cache Dec 18 21:43:59 (312) Dec 18 21:44:00 hows that for a number Dec 18 21:44:04 raster: I hate you Dec 18 21:44:04 thats the REAL footprint Dec 18 21:44:21 now run your app Dec 18 21:44:39 -/+ buffers/cache: 370 2914 Dec 18 21:44:40 damn thats not the point, the point is all that stuff are so packed with libraries Dec 18 21:44:44 (i ran inkscape) Dec 18 21:44:46 freesmartphone.org: 03jluebbe 07framework * r78582569e660 10/framework/subsystems/otimed/otimed.py: otimed: accept only valid time from ogpsd Dec 18 21:44:47 freesmartphone.org: 03jluebbe 07framework * r3d274adf9b45 10/ (7 files in 6 dirs): Dec 18 21:44:47 freesmartphone.org: ogsmd: reformat mobile_network_code into networks.tab Dec 18 21:44:47 freesmartphone.org: This adds a new DBus interface "org.freesmartphone.GSM.Data" on the server Dec 18 21:44:48 freesmartphone.org: object, which can be used to query the network database. Dec 18 21:44:50 * Tm_T is crippling with 1 GiB ram Dec 18 21:45:05 every stupid little gnome thing consumes a lot just for existing Dec 18 21:45:08 so 370-312 ius about as close to a "how much memory it uses" number as you can get Dec 18 21:45:23 so.. 58m for inkscape Dec 18 21:45:46 and since all applets are separate proceses in gnome panel for example, that gives you a lot Dec 18 21:46:07 just trying to help people as users tend to just erroneously read numbers for "it uses this much ram" all the time - and read the wrong ones :) Dec 18 21:46:17 but yes - gnome apps do tend to eat a fair bit of ram Dec 18 21:46:19 thats why i use e Dec 18 21:46:20 :) Dec 18 21:46:28 no you use it cause you wrote it ;P Dec 18 21:46:54 on my treo650 the entire memory footprint of the whole os booted, with hal, x e etc. is 12m Dec 18 21:47:07 but I admit the styles look good even without compositing Dec 18 21:47:33 * Tm_T is shameless KDE4 farter Dec 18 21:47:45 $ ps -eorss,cmd|sort -n|tail -n10|awk '{print $1 $2}'|tr '\n' ' '; echo Dec 18 21:47:45 5248xvnc4viewer 5848mplayer 6488/usr/sbin/named 8236irssi 9328/usr/sbin/smokeping 11488SCREEN 17096/usr/sbin/tor 35348/usr/bin/emacs 94980Xvnc4 124132/usr/lib/iceweasel/firefox-bin Dec 18 21:47:45 no, you don't like to know my memory footprint Dec 18 21:47:47 anyway on a system with 2GB ram I don't mind the 10% mem usage just for running gnome Dec 18 21:47:55 its comfortable Dec 18 21:48:11 of course on fr, thats totally another thing Dec 18 21:48:31 you can't hog 100MB on phone just for starting a minimal gnome session Dec 18 21:48:48 jc: well yes. i wrote it - but i pay pretty close attentin to memory usage in pretty much everything i write. i know that there will be a cost for something - i just need to make sure the memory cost is "Reasonable" given the return for it. so design things to use minimal memory and still maximal speed - so spend the time early on Dec 18 21:49:21 good approach Dec 18 21:49:26 indeed Dec 18 21:49:48 I think majority of memory usage in stupid little gnome applets is in the libs its linked to anyway Dec 18 21:49:54 Tm_T: btw - i lost 800m or so of my ram as i wanted shit like flash to work and so chose a 32bit distro. got 4g - but that's like. 3.2g is more than enough. Dec 18 21:50:26 raster: yes, and I would love to have even 2 gigs, would stop this constant swapping for a moment atleast Dec 18 21:50:27 hey flash works on my amd64 Dec 18 21:50:33 and beautifully Dec 18 21:50:36 you need nswrapper Dec 18 21:50:46 nspluginwrapper Dec 18 21:50:50 Tm_T: u can stop using kde... :) Dec 18 21:51:00 jc: thats what i didnt want to bother with Dec 18 21:51:02 raster: oh, it's not usage that uses ram too much Dec 18 21:51:11 along with all the other things that dont come in 64bit binaries Dec 18 21:51:12 or didnt Dec 18 21:51:13 no problem it is Dec 18 21:51:16 raster: compiling KDE4 takes most Dec 18 21:52:00 raster: and if wife left firefox running, that then, in her session Dec 18 21:52:00 I just compiled it against firefox, placed flash lib into ns plugins folder and ran the thing Dec 18 21:52:10 then it automagically works Dec 18 21:52:40 anyway amd64 is now somehow fucked up in the sense that tickless and high resolution timers don't work Dec 18 21:53:16 and for a long time it had a rather serious issue with keyboard layout switching Dec 18 21:53:19 (you couldn't) Dec 18 21:53:56 I'm always using newest git kernel, tickless still isn't fixed Dec 18 21:54:30 oh and nvidia drivers overflow my system with many interrupts from time to time so it totally hangs and reacts only to sysrq Dec 18 21:54:40 hmmmm Dec 18 21:54:42 but other than that its just pleasant :D Dec 18 21:54:50 I have never compiled Linux kernel Dec 18 21:55:10 on the other hand, I compile KDE several times a day Dec 18 21:55:11 root@palmt650:~# free -m Dec 18 21:55:11 total used free shared buffers cached Dec 18 21:55:11 Mem: 28 27 1 0 0 14 Dec 18 21:55:11 -/+ buffers/cache: 12 16 Dec 18 21:55:11 Swap: 0 0 0 Dec 18 21:55:15 no problem with that Dec 18 21:55:15 nice and lean :) Dec 18 21:55:24 * laen  Dec 18 21:55:27 you just cd into it, make menuconfig, set it up and make Dec 18 21:55:45 jc: oh, it's not about if it's hard Dec 18 21:55:48 never had to Dec 18 21:56:01 I take it as a part of installing the system Dec 18 21:56:04 I run gentoo :D Dec 18 21:56:33 it doesn't take long.. and when I did that for a friend with his mandriva, replacing the large stock kernel made a large difference in performance Dec 18 21:56:54 i don't understand how it can affect performance Dec 18 21:57:07 unless there's some buggy things going on Dec 18 21:57:13 well stock kernels tend to be full of functionality and drivers for just about everything Dec 18 21:57:21 we compile just the necessary stuff Dec 18 21:57:21 jc: but they are as modules, on disk Dec 18 21:57:28 sometimes Dec 18 21:57:43 but you get for example all filesystems compiled in, generic performance tweaks and stuff Dec 18 21:57:43 jc: you can just rm the modules if they bother you :) Dec 18 21:57:56 jc: what kind of generic performance tweaks? Dec 18 21:58:06 jc: faster bootup time I can agree Dec 18 21:58:14 well, like compiling bisexually for i686 with generic optimizations Dec 18 21:58:15 but not much else Dec 18 21:58:26 I compile for k8 on my amd64 for example Dec 18 21:58:30 jc: distros usually have i686-specific kernels already? Dec 18 21:58:43 yeah but theyre bisexual Dec 18 21:58:47 both for amd and intel Dec 18 21:58:54 so they have just generic tweaks Dec 18 21:59:26 jc: what benchmark did you use? Dec 18 21:59:33 well, the overall feel Dec 18 21:59:44 it feels good :D Dec 18 21:59:54 sigh Dec 18 21:59:54 mandriva boot time reduced a little Dec 18 22:00:02 and responsiveness increased Dec 18 22:00:08 like no jittery windows and stuff Dec 18 22:00:27 better make a measureable benchmark so that those can be added to stock kernels Dec 18 22:00:28 maybe the stock kernel had low timer freq or wasn't preemptive Dec 18 22:00:38 all those options make difference Dec 18 22:00:51 they can't be added Dec 18 22:00:55 still not very convinced :) Dec 18 22:01:09 when you apply those tweaks the kernel WON'T run on any other CPU family than the one you have Dec 18 22:01:23 but distro could offer kernel for all those Dec 18 22:01:25 and possibly on any other machine at all Dec 18 22:01:38 could, who would maintain all that crap Dec 18 22:01:48 autobuilders Dec 18 22:01:56 yeah they need to be set up Dec 18 22:02:01 of course Dec 18 22:02:18 but still, a benchmark would be useful Dec 18 22:02:25 having 50 .config scripts for every cpu and chipset is just dumb Dec 18 22:02:33 so that you would have numbers to compare against Dec 18 22:02:51 oh why.. Dec 18 22:02:52 and then you could notice regressions caused by individual kernel commits Dec 18 22:03:01 and not trust on feeling Dec 18 22:03:14 I'm pretty sure my kernel can't be more optimized Dec 18 22:03:32 jc: but next release might have something that makes it slower Dec 18 22:03:32 I have only the necessary stuff and its set up for low latency in general Dec 18 22:04:02 usually the .config files are portable, just a handful of stuff change Dec 18 22:04:14 you just go through the menu, see if everythings ok and then compile it Dec 18 22:04:17 matter of 2 minutes or so Dec 18 22:05:40 I don't get it Dec 18 22:05:52 compiling a kernel is a pretty natural thing to do, just like taking out trash and stuff Dec 18 22:06:37 it's quite unnatural to use a computer in any case :) Dec 18 22:06:59 you know how I meant that Dec 18 22:07:26 anyways, an automated testsuite would be very useful for catching regressions Dec 18 22:07:28 you could argue its not natural to wear clothes either Dec 18 22:07:37 jc: yes, that argument does not really work Dec 18 22:07:39 thats overcomplicated... Dec 18 22:08:00 are you arguing that regression testing is not needed? Dec 18 22:08:18 I do my regression testing by booting it and seeing if it works ;) Dec 18 22:08:24 which usually happens for the first timde Dec 18 22:08:25 time Dec 18 22:08:31 jc: but you can't notice 10ms slowdown that way Dec 18 22:08:43 those slowdowns don't happen Dec 18 22:08:47 jc: they do Dec 18 22:09:02 when you like change a major option by mistake, it usually slows it down to a hell freezing point Dec 18 22:09:07 you notice things Dec 18 22:09:10 every time you change the software you usually either make it slightly slower or faster Dec 18 22:09:12 get latency in everything Dec 18 22:09:20 if it feels ok, its ok for me then Dec 18 22:09:32 but it can slowly get worse that way Dec 18 22:09:54 well Dec 18 22:09:59 right now I'm compiling Dec 18 22:10:04 my CPU is at 100% Dec 18 22:10:07 cache at about 20% Dec 18 22:10:12 and memory 100% incl. swap Dec 18 22:10:17 and I'm chatting on irc Dec 18 22:10:20 and playing a video Dec 18 22:10:24 jc: anyways, could you please test the responsiveness of all andy-tracking kernel commits? Dec 18 22:10:36 nothing jitters or anything and moving my composited windows has no lag at all Dec 18 22:10:36 i'd like to know which one is best :) Dec 18 22:10:39 I'm pretty happy Dec 18 22:10:51 jc: i'm not arguing that you are not happy ;) Dec 18 22:10:54 this kind of behavior isnt what I see in default in ubuntu Dec 18 22:11:12 i wouldn't use such systems with non-free bug trackers :P Dec 18 22:11:25 oooh Dec 18 22:11:27 what are we talking about? Dec 18 22:11:30 I'm talking about my desktop kernel :) Dec 18 22:11:33 not the neo one :D Dec 18 22:11:49 jc: why would it be any different? Dec 18 22:11:55 got hopeful when you talked about video... Dec 18 22:12:10 well in freerunner, evely bad option in kernel matters a lot more Dec 18 22:12:16 gets things broken easily Dec 18 22:12:24 jc: aha, so there automated regression testing could be useful? Dec 18 22:12:40 UndrWater: videos work Dec 18 22:12:50 and given proper scheduling, you could watch while compiling Dec 18 22:12:57 just nice the crap of the compile process Dec 18 22:13:16 I don't have to nice anything here :) Dec 18 22:13:40 lindi-: I think they test by trial and error too Dec 18 22:14:02 http://tinderbox.mozilla.org/showbuilds.cgi?tree=Firefox is what mozilla people use to test their every commit on many systems Dec 18 22:14:06 but its pretty hard to do on fr, you'd have to do it via jtag or in a slow emulator Dec 18 22:14:34 jc: why? Dec 18 22:14:34 vga video on fr? or? Dec 18 22:14:36 well, if you feel like you know better than them, you could propose it to them Dec 18 22:14:45 jc: why not do it like on pc? Dec 18 22:14:46 nah, not vga, qvga is fine though Dec 18 22:14:49 vga is slow Dec 18 22:15:08 see the glamo mplayer video output plugin on the wiki Dec 18 22:15:26 will do...used qvga on zaurus...not bad...but not great Dec 18 22:15:43 lindi-: has your pc got an ARM920T with gsm, gps, bluetooth,... in it ? Dec 18 22:15:51 ooh .. what a shame, then it won't run Dec 18 22:15:58 qemu Dec 18 22:16:01 slow Dec 18 22:16:06 but you could Dec 18 22:16:10 jc: i mean, why not test the kernels on fr just like you test your kernels on pc? Dec 18 22:16:48 I do that Dec 18 22:17:06 jc: so, could you please test a few andy-tracking kernels for me? ;) Dec 18 22:17:06 well thats how they do it too, afaik Dec 18 22:17:12 they test stuff they code Dec 18 22:17:27 what for? ;P Dec 18 22:17:42 jc: cause otherwise i need to automate it since i don't have time to do it myself Dec 18 22:17:56 you're writing kernel patches ? Dec 18 22:18:02 testing and reporting bugs Dec 18 22:18:20 guess I could, but I have no toolchain atm Dec 18 22:18:27 I got stuck at glibc Dec 18 22:18:40 have done some small kernel modules too but only for own use (like loop device with CRC checks etc.) Dec 18 22:19:22 i thought it only took 2 minutes ;) Dec 18 22:19:39 the point with the kernel is that a kernel for audio production station is different from kernel for a database server Dec 18 22:19:46 and its not just a few options Dec 18 22:20:02 you tweak the kernel for your purposes, thats the point of making them yourself Dec 18 22:20:25 boot time I accept, as well as adding custom patches Dec 18 22:20:35 but for other stuff I'd like to see some raw benchmark data Dec 18 22:20:44 its in stuff like timer settings, preemption, disk scheduling, filesystem options, I could go on Dec 18 22:21:14 my system wouldn't be very well suited as a web server with many threads Dec 18 22:21:25 it would preemtp the crap of every stupid syscall Dec 18 22:21:35 and neither for safe db transactions Dec 18 22:21:40 it does fair scheduling Dec 18 22:22:16 but I can compile, watch videos, browse web and talk with people without slowdowns :P Dec 18 22:22:28 still, i'd like to see a site that shows examples of different configuration files and how they affect different benchmarks Dec 18 22:22:29 it just generally takes longer to process all stuff Dec 18 22:22:36 but doesn't slow down the gui Dec 18 22:22:39 measuring GUI stuff is just so hard Dec 18 22:23:04 the penalties associated with all the options are obvious and are documentet in the kernel Dec 18 22:23:10 *d Dec 18 22:23:19 help in menuconfig usually helps a lot Dec 18 22:23:35 and when the particular option tells you nothing at all, you just google it and learn what it is Dec 18 22:23:44 or read the source Dec 18 22:23:55 its very time consuming for the first time Dec 18 22:24:03 but every subsequent time, you feel like in control Dec 18 22:24:57 yeah well I meant like if you don't know what is SMP (the concept) you look it up and decide whether its good for you Dec 18 22:25:46 i'm still bit unsure about why automated regression tests are not a good idea Dec 18 22:26:14 they measure just the numbers, not the feel Dec 18 22:26:30 and the feel derivates from maany numbers at once Dec 18 22:26:56 but if all numbers stay the same then the feel is the same Dec 18 22:27:17 and there is no need for a human to check the feel of that commit? Dec 18 22:27:21 may be Dec 18 22:27:50 well they release it and people start shouting at them so they realize the've done something wrong and fix it? ;) Dec 18 22:27:54 and generally if it takes 1.5 seconds longer to open an X window then it should be flagged with red flag and the author warned? Dec 18 22:28:01 as linus says Dec 18 22:28:08 given enough eyballs, all bugs are shallow Dec 18 22:28:18 or was it esr ? Dec 18 22:28:24 i still think that good automated testing can help a lot Dec 18 22:28:29 combined with manual testing of course Dec 18 22:29:16 I think that function testing for all the individual modules is like more to the point Dec 18 22:29:25 but anyways, i better get some sleep now. good night Dec 18 22:29:27 but devs are lazy to do that, I think Dec 18 22:29:39 yeah, good night Dec 18 23:19:57 freesmartphone.org: 03jluebbe 07framework * rc6c9001f33a3 10/framework/subsystems/ (ogsmd/gsm/const.py otimed/otimed.py): otimed: use network database to install a time zone file Dec 19 00:05:26 i keep getting: [ERROR: Please set the 'CACHE' variable.] when trying to bitbake -c clean, anybody know how to solve this? Dec 19 00:15:11 freesmartphone.org: 03daniel 07framework * r898dff053f32 10/framework/subsystems/ogsmd/ (device.py gsm/sms.py modems/abstract/mediator.py): Dec 19 00:15:11 freesmartphone.org: ogsmd: Add support for sms-submit-reports and return timestamps Dec 19 00:15:11 freesmartphone.org: When sending an SMS an sms-submit-report is returned which includes a Dec 19 00:15:11 freesmartphone.org: timestamp. This is the time when the Short Message Service Center Dec 19 00:15:13 freesmartphone.org: received the short message. Dec 19 00:15:15 freesmartphone.org: The DBus calls for sending SMS now also return a timestamp in string Dec 19 00:15:17 freesmartphone.org: format (WARNING, API breakage!) Dec 19 00:15:39 So can anyone tell me why ScummVM has such crappy mouse positioning? I put the stylus on the screen, and the mouse goes half the screen away! Dec 19 00:35:02 freesmartphone.org: 03daniel 07specs * r03286c0102ed 10/ (7 files in 4 dirs): Added timestamp return value for Send{Stored}Message functions Dec 19 00:50:16 does anybody here use usb host mode? **** ENDING LOGGING AT Fri Dec 19 02:59:57 2008