**** BEGIN LOGGING AT Sat Mar 14 02:59:57 2009 Mar 14 08:45:04 PaulFert`: diplomatic carreer ambitions? ;D Mar 14 08:46:08 PaulFert`: sure you'd perform exceptionally good there Mar 14 08:48:32 PaulFert`: OTOH a commonly accepted mantra of politics seems to be "everybody probably is evil". Doesn't fit your recent POV Mar 14 08:53:02 DocScrutinizer: are you about my recent comment at community ML about "fair" treatment? Mar 14 08:53:13 sure Mar 14 08:53:47 nothing wrong with it Mar 14 08:54:10 I just wondered (or more honestly, I was quite impressed) Mar 14 08:55:43 DocScrutinizer: I really think the way i stated, no filthy diplomacy involved. I don't like his POV and his manner of expressing it but he has every right to do it that way. At least he answers the questions if asked directly. That's how i see. Probably the fact i've just woken up somehow contributed to the tone of my reply. Mar 14 08:56:15 :) Mar 14 08:56:27 * DocScrutinizer remembers missing his coffe Mar 14 08:57:48 DocScrutinizer: and to be honest to me it seems that Mirko has good (read not-money driven) intents, and that is quite important for me. Mar 14 08:58:11 that's correct for sure Mar 14 08:59:20 And i highly doubt the same is true for the majority of TPE devs (though any direct evidence to that or to the contrary is missing). Mar 14 09:00:25 community is rather harsh every once in a while. And maybe you're not the only one pestering him. Everybody has different means to protect against that Mar 14 09:01:57 He said he received like 40 letters of complaints. Mar 14 09:06:21 so his reply wasn't that unfriendly I'd say ;D Mar 14 09:06:36 also keep in mind he's just quite new to OM Mar 14 09:07:36 And I like his absolute dedication towards creating a useable phone Mar 14 09:08:46 PaulFert`: I tought about mic calib. (caly-AGC) Mar 14 09:09:38 PaulFert`: it's probably best to adjust mic-sensitivity to a subjective point-of-best-performance Mar 14 09:10:04 i.e. exactly what they're trying to do in TPE ;) Mar 14 09:10:26 DocScrutinizer: why is better than going "scientifically correct" way? Mar 14 09:10:33 as there are too many contradicting parameters/effects included Mar 14 09:11:01 Inside the calypso? Mar 14 09:11:55 PaulFert`: nope, whole system. "scientific" way would be: 1. approach:level up until you get distortion Mar 14 09:12:31 PaulFert`: 2.approach: oops we see issues with echo, so maybe we don't need *THAT* much sensitivity Mar 14 09:12:34 I thought that the "voice profile" (or how do they call it) was prepared by TI engineers and it's a reasonable expectation to have an information on the recommended signal levels to feed it. Mar 14 09:12:45 scientific schmific Mar 14 09:13:05 raster: Hey :) Mar 14 09:13:12 lol, HI RASTER!! :))) Mar 14 09:13:30 PaulFert`: boo Mar 14 09:13:52 PaulFert`: which "voice profile"? Mar 14 09:13:53 DocScrutinizer: jööörg! Mar 14 09:15:24 PaulFert`: I now of efforts the modem-manufs do, like meddling equalizer settings inside their module, etc Mar 14 09:15:26 DocScrutinizer: http://wiki.openmoko.org/wiki/GTA01_gsm_modem#AT.40AUL Mar 14 09:17:11 yup, exactly these audio tables. I'm not sure to which extent TI did some tuning of the standard/default values of table0 Mar 14 09:17:58 and I'm quite sure nobody defined the mixer patch overall gain at which those tunings (if any) had been done Mar 14 09:18:03 DocScrutinizer: and i'm sure that an alsa state profile should be created which doesn't care for echo at all, aimed at the best AGC performance, as echo can be easily avoided by using an external headphones (with internal mic). Mar 14 09:18:12 s/patch/path/ Mar 14 09:18:12 DocScrutinizer meant: and I'm quite sure nobody defined the mixer path overall gain at which those tunings (if any) had been done Mar 14 09:19:52 PaulFert`: yup, we needed different audiotables, at%n settings, and mixer setup(obviously;) for different scenarios like handset vs headset vs speakerphone Mar 14 09:20:48 DocScrutinizer: and i guess we don't really need any tuning of those table0 values, we just need to find out what gain we should get in the mixer. Can examing the reference design (for calypso) help to determine that? Mar 14 09:21:16 shr-unstable built this morning (~0000utc) - e dies when trying to change size of top shelf Mar 14 09:21:24 alrm doesnt wake from suspend Mar 14 09:21:31 know problems ?? Mar 14 09:21:43 PaulFert`: up til now nobody was able to tell me what is calypso's input sensitivity :-( Mar 14 09:21:44 s/known/known/ Mar 14 09:22:04 s/know/known/ Mar 14 09:22:05 billk meant: known problems ?? Mar 14 09:22:06 billk: hmm... (??) Mar 14 09:22:42 DocScrutinizer: i see, but why can't that be deduced from a reference design schematics? Mar 14 09:23:12 mwester: kick bzzbot please. Seems it's sleepy Mar 14 09:23:53 it probably can, and I did to find it is sth like 30mVss Mar 14 09:24:30 but out settings in table0 and generally in FW could completely void this Mar 14 09:26:26 And i assume there's no way to find somebody who does know if any settings were changed in FW. Too bad again. Then probably you're correct, a subjective way is probably the best. OTOH i'm sure there're ways to make it less subjective (than simple meddling with alsa controls until it "feels" best). Mar 14 09:27:14 yup, of course there are clues of the limits Mar 14 09:28:52 like e.g "volume shouldn't drop if you talk to the phone from 80cm distance, so level up sensitivity to meet this demand" Mar 14 09:29:20 And i'm yet to imagine any reasonable way to measure the resulting output volume (as we can't just hook up another GSM cell to a PC and read the raw audio data). Mar 14 09:29:35 ...or "no distortion when shouting at the phone, so don't level up too much" Mar 14 09:30:01 PaulFert`: that's the main problem all the time Mar 14 09:30:48 PaulFert`: you'd need a CMU200 or similar equipment to do such tests in a decent way Mar 14 09:32:13 taking in account all these parameters it's probably best to just empirically find a "good enough" setting Mar 14 09:32:37 s/parameter/aspect/ Mar 14 09:32:38 DocScrutinizer meant: taking in account all these aspects it's probably best to just empirically find a "good enough" setting Mar 14 09:34:02 PaulFert`: FWIW I don't know of any other phone user can adjust mic sensitivity Mar 14 09:34:21 DocScrutinizer: though we can measure it in a subjective way as well: put the sim in a cell phone considered to adhere to de facto industry standard (like Nokia 3210 or smth), call the fixed line (the fixed line is hooked up to a soundcard of PC), Sync by playing some loud sound through the speakers aimed to the mic, play the test data (sine waves with different volumes) through the same speaker. Collect the results and use as a reference. Mar 14 09:35:31 s/fixed line/voip TAM (e.g. joerg's at +49-911-3083854x)/ Mar 14 09:36:12 * PaulFert` thinks about the fixed line all the time because it's the cheapest way to test voice transfer here :) Mar 14 09:37:10 voip voice box comes in handy by recording wav and sending them via mail to me ;) Mar 14 09:38:56 PaulFert`: still you have GSM-provider's audio processing circuitry in between. This might comprise of an additional AGC as well as it surely has some AEC Mar 14 09:39:37 billk: for me alarm wake up from suspend but even rings if alarm is turned off -.-" Mar 14 09:40:32 DocScrutinizer: TPE devs have CMU2000 and all the equipment and time to get this right. :( Mar 14 09:40:45 sure Mar 14 09:44:32 * PaulFert` wonders what is their plan wrt gta03's modem... Our conversation is getting depressing again, Joerg: too many pessimists (realists) in one openmoko channel ;) Mar 14 09:47:43 I was trying to claim these aspects of mci75 to be my domain, and sync with cinterion to make sure they get well defined starting conditions to do the "meddling". Alas it's not been perceived to be an issue at all :-/ Mar 14 09:53:57 DocScrutinizer: btw, i just realised that i subconsciously percieve TPE stuff as robots and treat them accordingly. To me it looks like all their communications (except Tsai Harry's, but he's not OM anymore) completely lack "personal touch". You can't on earth figure what kind of human (tough/kind/frank/joyful) you're communicating with. Mar 14 09:55:09 duh, *which* communications?? ;D Mar 14 09:55:16 And _that_ too Mar 14 09:56:11 bzzbot: are you taiwanese? Mar 14 09:56:12 Like Candy's letter. The only one. Too little reference points to have any judgment. Mar 14 09:56:47 PaulFert`: QED XD Mar 14 10:04:51 DocScrutinizer: otoh a "normal" person can (and usually do) supply some reference points even in one letter. The letters i saw from TPE engineers are mostly purely "mechanical". Mar 14 10:05:28 lost in translation Mar 14 10:05:53 Most of us are not native too Mar 14 10:06:46 I don't say they're robots, i just subconsciously percieve them as such. Mar 14 10:11:50 DocScrutinizer: u actually have stopped working for om.. right? Mar 14 10:12:07 yup, kinda Mar 14 10:12:13 "kinda" Mar 14 10:12:14 ? Mar 14 10:12:37 well, I'm on a "on demand" agreement Mar 14 10:12:47 aaah ok Mar 14 10:12:53 actually there's no demand Mar 14 10:12:58 hahahaha Mar 14 10:13:27 except ALSA for mickey, which isn't exactly OM Mar 14 10:15:34 raster: can i make a label clickable in elementary? Mar 14 10:15:41 is there an attempt to oysh back bits of the moko kernel in mainstrearm ? Mar 14 10:15:48 push Mar 14 10:15:55 evas_object_smart_callback_add(lb,"clicked",lb_clicked,NULL); doesn't work Mar 14 10:16:16 sakkara: no Mar 14 10:16:30 =( Mar 14 10:16:31 a label is just that - a label. its not an interactive clickable element Mar 14 10:16:42 an entry is Mar 14 10:17:11 morning! Mar 14 10:17:27 raster: hm ok, thanks Mar 14 10:17:46 I needed bits of moko's kernel for my 2440 board, so I "back"ported the touchscreen support to mainstream tree, but I've been wondering if I ought to try to push them to the arm tree, or if the moko people do it, or care... seems to me they don't care.. Mar 14 10:17:48 u can have an entry that is not "editable" Mar 14 10:17:48 DocScrutinizer: I want my pingpong champion cup! Mar 14 10:17:53 :) Mar 14 10:17:55 ie u can select from it Mar 14 10:17:57 and click links Mar 14 10:17:59 etc. Mar 14 10:18:02 (anchors) Mar 14 10:19:00 mrmoku: you here ? Mar 14 10:19:33 raster: hm yes, i just wanted a label to display smthg and clear it on tap, but i will make a button for that Mar 14 10:20:15 the thing is - u want to be careful Mar 14 10:20:33 elementary adjust sizes of things u can "tap" or press/drag to make sure its at least a "finger size" in size Mar 14 10:20:36 to make it easy to interact Mar 14 10:20:43 labels are not adjusted as they are for reading only Mar 14 10:20:48 buttons, sliders, lists etc. are Mar 14 10:20:57 so there is a fundamental difference in a label and a button Mar 14 10:21:01 labels are "Read only" Mar 14 10:21:07 buttons are "read/write" Mar 14 10:21:25 buttons have a different look for usability reasons Mar 14 10:21:42 in elementary's default thteme things u can press/click are black Mar 14 10:21:51 to let u know its an interactive element u can press Mar 14 10:22:01 a label that just looks like one is a bad ui choice Mar 14 10:22:05 as a user has no idea they can press it Mar 14 10:22:27 ah ok, that sounds clear to me Mar 14 10:23:22 thus - stick to the right widget Mar 14 10:23:29 if u want a button that looks different Mar 14 10:23:32 less "butotn-like" Mar 14 10:23:35 thats what styles are for Mar 14 10:23:39 not all widgets support it Mar 14 10:23:41 but a few do Mar 14 10:23:48 likely i will add more aas i go on Mar 14 10:23:53 some widgets it makes no sense for Mar 14 10:24:12 yeah Mar 14 10:24:17 others it does Mar 14 10:24:23 and some i just havent "done yet" Mar 14 10:24:30 like all things - only so much time in a day Mar 14 10:24:41 and i do the things that are most important to those who are paying me first Mar 14 10:24:41 :) Mar 14 10:24:52 :D Mar 14 10:25:11 (thats the whole point of paying.. you get your stuff to the top of the priority queue first) Mar 14 10:25:11 what are you working on at the moment? Mar 14 10:25:22 elementary+evas+ a bit of edje Mar 14 10:25:36 speedups in evas, edje features to support elementary where needed Mar 14 10:25:39 and elementary manily Mar 14 10:25:57 may i ask who pays you for that? Mar 14 10:26:09 clients Mar 14 10:26:09 :) Mar 14 10:26:14 :D Mar 14 10:26:32 (who are not openmoko) Mar 14 10:27:03 hm Mar 14 10:28:18 is there any other doc for elementary than the wiki/elementary on enlightenment.org? Mar 14 10:28:25 thats it Mar 14 10:28:28 plus example code Mar 14 10:28:32 test.c Mar 14 10:28:35 and elementary.h Mar 14 10:28:56 ok... Mar 14 10:29:53 i like elementary, but i'm used to qt assistant so its a bit harder for me, always browsing example code... Mar 14 10:30:05 yeah Mar 14 10:30:13 that is somehting i tried to help with with the wiki page Mar 14 10:30:19 at least the begining "getting started" bits Mar 14 10:30:26 but again - only so many hours a day Mar 14 10:30:30 yeah, that helped a lot Mar 14 10:30:35 so either add mroe widgets/features or write more docs Mar 14 10:30:40 its always a tossup Mar 14 10:30:48 i almost always lean to "more code" Mar 14 10:30:58 yes, i understand Mar 14 10:31:14 but i hoped the wiki will be enough to get yopur feet wet - the rest should be easy enough with just a read of header files for api calls then browsing exmaple code Mar 14 10:31:16 i hope Mar 14 10:31:29 on a list of things i'd like to do is a ui builder Mar 14 10:31:36 that can save a ui as a datafile Mar 14 10:31:44 or generate code for you to produce the ui Mar 14 10:32:04 like the qt designer Mar 14 10:32:04 so u just drag and drop widgets from a palette into your "window" Mar 14 10:32:09 play with their properties Mar 14 10:32:10 etc. Mar 14 10:32:19 and u get an instant preview that "Works" Mar 14 10:32:22 yup Mar 14 10:32:23 or glade Mar 14 10:32:24 etc. Mar 14 10:33:21 that sounds good =) Mar 14 10:33:28 but it may be a while Mar 14 10:33:41 right now i have clients who are now well familiar with efl and like elementary a lot Mar 14 10:33:49 they know it and really just need more power/features/widgets Mar 14 10:33:59 a designer would be great Mar 14 10:34:06 but less important than "features" Mar 14 10:34:22 a designer combined with edje_editor would be pretty good Mar 14 10:34:29 well use the tools together Mar 14 10:34:47 design themes or layouts or "extension themes" (to add new custom styles) Mar 14 10:34:53 and then select these that you designed Mar 14 10:34:53 oh yeah Mar 14 10:35:00 load/import them into the designer Mar 14 10:35:11 but even so a designer can only do so much as a tool Mar 14 10:35:13 wow, you really have good ideas Mar 14 10:35:18 but maybe it can do 90% of the boring footwork for you Mar 14 10:35:42 right now i am wondering about code generation vs putitng the ui design in a data file and loading at runtime Mar 14 10:36:02 but i'm just letting the idea lurk in the back of my mind for now Mar 14 10:36:09 elementary needs a lot lot lot more widgets Mar 14 10:36:14 it has a long todo list Mar 14 10:36:22 chances are over time it will get more and more and more Mar 14 10:36:31 =D Mar 14 10:36:46 the idea being that if u want to do an app that somewhere there will be some pre-canned widget/object that will have just what u want Mar 14 10:36:52 in a standardised theme setup Mar 14 10:37:08 in the end you only write logic/glue code for your app Mar 14 10:37:25 or you are busy using cusotm edje layouts or custom smart objects u created yourself Mar 14 10:37:30 for very special purposes Mar 14 10:37:40 i think one of the most important things is speed, so i am for code generation, ui loading at runtime may slow down a lot Mar 14 10:37:59 think about it this way Mar 14 10:38:04 it already loads lots at runtime Mar 14 10:38:06 edje Mar 14 10:38:18 but... no one is screaming to get rid of that! Mar 14 10:38:19 :) Mar 14 10:38:27 if i do it.. chances are it'll be fast anyway Mar 14 10:39:00 btw - if u do your elementayr apps right - u can use elementayr's quicklauncher Mar 14 10:39:06 if u have slow flash/disk io Mar 14 10:39:10 like the freerunner Mar 14 10:39:21 in my example i got app start time down from 4.7 to 1.0 seconds Mar 14 10:39:22 i think you'll do it right =) Mar 14 10:39:26 if u have really fast flash/io Mar 14 10:39:36 it still went from 1.7 to 1.0 secs Mar 14 10:39:42 (this was a s3c6410 system) Mar 14 10:39:47 wow Mar 14 10:39:55 elementary_test builds a versuion of it for the quicklauncher Mar 14 10:40:06 also elementary_sms (in svn) builds for quicklaunch too Mar 14 10:40:13 basically u build both a binary and a .so Mar 14 10:40:19 the .so goes in the lib dir Mar 14 10:40:27 by default elementary just runs the "long way" Mar 14 10:40:34 and u get no speedup - but u split your app into 2 bits Mar 14 10:40:49 a very small exe that loads the .so (elementary takes care of all of this for you) Mar 14 10:41:00 (the macros and startup setup all handles it) Mar 14 10:41:08 u just need the #ifdef Mar 14 10:41:19 around the code for the meat of your app Mar 14 10:41:24 and use ELM_MAIN() macro Mar 14 10:41:27 thats all Mar 14 10:41:38 then.. u get a small exe that loads the .so for your app and runs as usual Mar 14 10:41:46 it is the same speed as normal - so no speedup Mar 14 10:42:01 hm Mar 14 10:42:12 now Mar 14 10:42:17 delete that small exe Mar 14 10:42:25 and ln -s the exename to elementary_run Mar 14 10:42:27 and bingo Mar 14 10:42:29 well ok Mar 14 10:42:40 u need tyo run elementary_quicklaunch as part of your ui session Mar 14 10:42:44 it needs to be floating about Mar 14 10:42:47 but.. if it is Mar 14 10:42:50 and u do that symlink Mar 14 10:42:54 your app will start.. faster Mar 14 10:42:58 try it Mar 14 10:43:00 oh nice Mar 14 10:43:06 a quick test is Mar 14 10:43:10 yeah, i'll try that Mar 14 10:43:15 delete the elementary_testql binary it installs Mar 14 10:43:42 and ln -s elementary_run elementary_testql Mar 14 10:43:49 then run elementary_quicklaunch Mar 14 10:43:59 somehwhere in your gui session Mar 14 10:44:06 (automatically or by hand Mar 14 10:44:08 doesnt matter Mar 14 10:44:16 then just run your app as before Mar 14 10:44:19 elementary_testql Mar 14 10:44:22 it will start faster now Mar 14 10:44:26 sit and time it Mar 14 10:44:34 time from run to window on screen will be less Mar 14 10:44:38 hopefulyl a lot less Mar 14 10:44:41 =D Mar 14 10:44:54 (and that entirely depends on your os/flash/io/caches etc.) Mar 14 10:45:08 ie my 4.7->1.0 sec was runing off nfs over usb Mar 14 10:45:15 thus slow disk - little/no caching Mar 14 10:45:27 the 1.7->1.0 was on local onenand on a s3c6410 Mar 14 10:45:37 and with proper flash drivers written to be fast Mar 14 10:46:03 so a lot more was cached and io was faster Mar 14 10:46:07 bvut u still got a speedup Mar 14 10:46:30 hm Mar 14 10:46:47 if u choose not to use this feature - no gain, but no loss Mar 14 10:46:54 it all works the "old way" Mar 14 10:47:04 if u build "to be able to use it" Mar 14 10:47:07 also no loss, or gain Mar 14 10:47:18 and everything works via compatibility paths Mar 14 10:48:33 ok, then its allways good to build it like that Mar 14 10:48:38 correct Mar 14 10:48:41 as i siad Mar 14 10:48:41 and then check out the gain/loss Mar 14 10:48:44 elementary-sms Mar 14 10:48:57 yeah, i'll have a look on that Mar 14 10:49:09 builds for it Mar 14 10:49:15 elementary-alarm does not Mar 14 10:49:22 elemtary_testql does Mar 14 10:49:27 elementayr_test does not Mar 14 10:49:40 so u can look at the autofoo/makefile/am's etc. to see how its done Mar 14 10:49:44 and look at the src Mar 14 10:49:53 it literally is a #ifdef Mar 14 10:50:15 ok Mar 14 10:50:20 include Mar 14 10:50:20 #ifndef ELM_LIB_QUICKLAUNCH Mar 14 10:50:34 / ... meat of program that does everything Mar 14 10:50:57 // ... Mar 14 10:51:04 EAPI int Mar 14 10:51:04 elm_main(int argc, char **argv) Mar 14 10:51:04 { Mar 14 10:51:17 // init stuff here as you did before - no change Mar 14 10:51:23 return 0; Mar 14 10:51:23 } Mar 14 10:51:23 #endif Mar 14 10:51:27 ELM_MAIN() Mar 14 10:51:30 thats it Mar 14 10:51:56 the rest is the makefile.am (and autofoo to make sure u can build .so's in configure.in/ac) Mar 14 10:52:02 hi guys Mar 14 10:52:05 really simple, you should write that on the wiki :P Mar 14 10:52:17 and then the magic symlink (and deletion of the bin you build) is the final "we run via quicklaunch only" step Mar 14 10:52:43 hehe, great Mar 14 10:53:15 i should Mar 14 10:53:17 i just havent yet Mar 14 10:53:29 time time time :) Mar 14 10:53:33 but its a nice "hidden" feature that is pretty much almsot 0 effort to slide in Mar 14 10:53:50 it allows u to build the binary withotu quicklaunch mostly for debugging Mar 14 10:53:53 it works in gdb as usual Mar 14 10:54:02 even tho it dlopen()'s your app;'s .so Mar 14 10:54:05 it is debuggable Mar 14 10:54:07 as usual Mar 14 10:54:33 so you pretty much would only rm the binary and do the ln -s as a final stage before "production release" after the bugs are gone Mar 14 10:54:41 and u're squeezing performance Mar 14 10:55:27 also rememebr - you do get separate processes in quicklaunch Mar 14 10:55:34 so if your app crashes - it wont bring everything down Mar 14 10:55:38 it is "safe" Mar 14 10:55:51 but you need to follow the template of doing an app Mar 14 10:55:51 yeah Mar 14 10:55:57 ie ELM_MAIN() macro Mar 14 10:55:59 etc etc. Mar 14 10:56:02 to make it work right Mar 14 10:56:46 raster: any progress on you sms app? Mar 14 10:56:55 yes, this is supposed to allways be... Mar 14 10:57:07 thank you raster =) Mar 14 10:57:19 bumbl: hmm - it should work. just nothing to handle incoming sms's Mar 14 10:57:33 ah ok Mar 14 10:57:46 but you still work for swisscom on the moko? Mar 14 10:59:15 or have you had time to start on your build-my-own-open-cellphone idea? Mar 14 11:00:49 bumbl: still do Mar 14 11:00:54 tho only 1/2 time Mar 14 11:01:02 :) Mar 14 11:01:11 bbl Mar 14 11:01:14 need to grab food Mar 14 12:47:10 hi Mar 14 12:47:25 any news about no apps showing in illume SHR-unstable? Mar 14 12:48:05 admiral0, are there files in /usr/share/applications ? Mar 14 12:48:14 admiral0: afaik that should be fixed Mar 14 12:48:45 tomboy64: i'm not that noob Mar 14 12:48:54 pardon me admiral0 :) Mar 14 12:49:00 bumbl: must see .config/menus Mar 14 12:51:11 tomboy64 ;) Mar 14 12:54:29 ahahaha Mar 14 12:54:54 i don't what i did but i have all icons on illume with latest shr-unstable Mar 14 12:55:07 *i don't know what Mar 14 12:55:13 XD Mar 14 12:55:15 bye Mar 14 12:55:20 going to work Mar 14 13:24:45 heyho Mar 14 13:26:32 hi morphis Mar 14 13:26:49 how is python elementarygoing Mar 14 13:27:14 good :) Mar 14 13:27:30 currently only support for the genlist widget is missing Mar 14 13:32:24 Does anyone know if the 'Devirginator' is up-to-date ? Mar 14 13:32:24 I'm trying to rescue my bricked freerunner and I've already disassembled it and connected it to Debug Board via JTAG-cable. Mar 14 13:34:42 ingi: wtf U did to brick it? Mar 14 13:35:19 Yup, the GSM-firmware update got stuck so it's kinda bricked. Mar 14 13:35:41 what a nonsense. sorry Mar 14 13:36:13 1: we never seen gsm-flashing to get stuck Mar 14 13:36:33 2: this won't brick the whole FR Mar 14 13:37:01 3: you impossibly can fix this non-issue *by debug board* Mar 14 13:37:23 You haven't seen my FR. I'm not blaming anyone either, maybe it's just not supported to do the update on a 64bit Ubuntu... Mar 14 13:38:44 "Bricked" for me means that I have no USB-network connection with it! When booting it complains about not being able to set MAC-address when loading g_ether stuff. Mar 14 13:38:47 * DocScrutinizer *strongly* recommends to read wiki or ask before bombing NOR to fix GSM Mar 14 13:39:10 Wiki ? Whats that ? Mar 14 13:39:27 Of course I've been reading both the Wiki and the mailing-list. Mar 14 13:40:08 so who told ya gsm-flashing could brick nor uboot???? Mar 14 13:40:56 as NOR is the only thing you need a debug board to unbrick Mar 14 13:41:33 Nobody told me. But it's evidently a fact that I cant connect to my FR over USB after having tried to upgrade the GSM-firmware. Mar 14 13:41:52 Do you know of better way to restore the state of my FR ? Mar 14 13:42:11 * anttna hides under desk Mar 14 13:42:15 flash a known good rootfs and kernel Mar 14 13:42:26 I've had the FR since september last year, I hope you dont think I got it yesterday... Mar 14 13:42:31 * anttna comes back out Mar 14 13:43:08 gsm and usb is *completely* unrelated Mar 14 13:44:20 Of course I've tried to do that with both recommended rootfs/kernel and newest likewise. I even prepared my uSD with the 89Mb image that's supposed to upgrade the GSM-firmware automatically, but to no avail. Therefor I've started this Debug Board devirginating process. Mar 14 13:44:51 *sigh* Mar 14 13:44:55 bye Mar 14 13:45:06 *sigh* :-/ Mar 14 13:46:59 "if you lost it over there, why for gods sake are you searching the floor here around the street light??" "well over there it's dark" Mar 14 13:47:55 :) Mar 14 13:49:04 * anttna hopes flashing shr might improve the volume level of the mic (as a quick way to fix the problem) Mar 14 13:57:48 ingi: FR is unbrickable by design. Still you seem to be 2 steps away from actually bricking it. For your convenience: step1: kill NOR for no good reason. step: break FPC or connector of debug board Mar 14 14:01:01 Ok, then it isn't broken, just me imagining it is. Thanks. Mar 14 14:01:18 ingi: i'd do as the very kind DocScrutinizer says!, but i'd count to ten first if your not sure :) Mar 14 14:04:13 Then what would you do ? Mar 14 14:04:49 you *NEVER* need debug board and devirginator to recover from any mishap. NOR can't be destroyed and is completely sufficient to flash a new NAND uboot (if this would be necessary) Mar 14 14:06:00 I replaced the NAND uboot with Qi for 2 months ago. Mar 14 14:06:39 if you can boot to NAND uboot menu (press power, then aux), then all you need is a known good kernel and rootfs Mar 14 14:06:44 And I've been trying to boot from NOR uboot with new rootfs/kernel's and as I understand your meaning that would be sufficient to bring the FR up and running. Mar 14 14:07:30 ingi: NOR is mainly (solely) needed to flash NAND bootloader Mar 14 14:08:20 NOR uboot isn't supposed to be able to boot rootfs with special Qi mods Mar 14 14:09:49 ingi: and for sure NOR needs no fixing to recover from a fail on gsm-flashing Mar 14 14:12:33 neither needs the kernel or rootfs you used before gsm-flashing "failed" (basically) Mar 14 14:16:13 My understanding is that if I screw up the GSM-firmware in the phone, no rootfs/kernel will help, wether I boot it from NOR or NAND. Are you saying that my undstanding is wrong ? Mar 14 14:17:03 ptitjes: here now Mar 14 14:18:10 I posted previously to .user-mailing list, search for the subject: "Re: Help! GSM not working after updating to Moko10", you see a reply from Werner Almesberger there were he suggests two methods to recover from this, one of the is the devirginizing. They are a bit unclear for me so I'm kinda lost, therefor searching were the lights are bright... Mar 14 14:19:34 Either you got im wrong, or Werner told complete crap Mar 14 14:20:11 ingi: can you still get NOR bootloader menu (uboot) ? Mar 14 14:21:21 ingi: i.e. what happens if you press and hold AUX, then press and release power while still holding AUX? Mar 14 14:22:28 ingi: and are you sure you have a Frerunner, not a Neo 1973? Mar 14 14:24:15 hmm, sorry. really late on brunch now. Mar 14 14:24:17 bye Mar 14 14:25:10 mrmoku: hello Mar 14 14:25:45 mrmoku: you wanted to tell/ask me something yesterday ? Mar 14 14:29:30 ptitjes: well... wanted to know about the dinner :-) Mar 14 14:30:54 mrmoku: was good! nice to talk with angus Mar 14 14:31:55 mrmoku: =>query Mar 14 14:32:25 select * from topic where rating='interesting' :P Mar 14 14:33:18 * mrmoku hast to get used to queries :-) Mar 14 14:35:28 LOL Mar 14 14:50:30 who made the timeouts settings in shr-settings? Mar 14 14:50:48 Zorkman: dos1 and spaetz Mar 14 14:50:50 and does anyone know how to use them correctly? Mar 14 14:51:25 mrmoku: i seem to be doing something wrong... Mar 14 14:51:29 it won't suspend... Mar 14 14:51:45 are there sliders that have "priority" above others? Mar 14 14:52:02 like if suspend time < idle prelock (whatever that is....) it won't suspend? Mar 14 14:52:38 Zorkman: no.. it is the timeout from one state to the next Mar 14 14:52:58 i don't understand you i think.... Mar 14 14:53:42 if i want screen to go dark in 60s, the screen to be locked (like when pressing auxw) in 75sec, and suspend in 120sec, how do i set them? Mar 14 14:54:09 60, 15 and 45 Mar 14 14:54:19 (at least that's what I understand :P) Mar 14 14:54:26 ah so it's cummulative Mar 14 14:54:43 and what is idle ? or idle dim? Mar 14 14:56:01 idle is the time the system stays in idle state without the user touching the screen Mar 14 14:56:19 hmm... not very sure though ;) Mar 14 14:59:39 i don't understand is, isn't the everything related to the time without the user touching the screen? Mar 14 15:00:30 (btw: read the fact that someone managed to put the gps into deep sleep?) Mar 14 15:00:45 gps or gsm? Mar 14 15:00:49 gps Mar 14 15:01:04 never heard of gps deep sleep... Mar 14 15:01:16 community mailinglist, topic: Good news: FixNow mode can be wakenup through serial port Mar 14 15:04:01 interesting... Mar 14 15:04:22 indeed Mar 14 15:04:48 btw: still couldn't get tangogps to get a fix with shr-testin, have you heard of anyone else having that problem? Mar 14 15:05:33 mickey|CLT: have you read the community-mailinglist e-mail; topic: "Good news: FixNow mode can be wakenup through serial port" seems like something interesting for the framework :) Mar 14 15:05:59 i have an email backlog of one week Mar 14 15:06:14 but that sounds good, ya Mar 14 15:06:14 my condoleances Mar 14 15:06:17 Zorkman: well... I get a fix. the best method for me is to start tango and just ignore it ;) Mar 14 15:06:22 morphis: ping Mar 14 15:06:44 mickey|CLT: tried to checkout your efl/vala repo... and failed Mar 14 15:07:00 building will fail, but checking out should work Mar 14 15:07:04 how did it fail= Mar 14 15:07:05 ? Mar 14 15:07:05 mrmoku: i never get a fix :( although openmoko-agpsui gets one very fast; i tried deleting ogsmdpickle but nothing changed Mar 14 15:07:40 fatal: The remote end hung up unexpectedly Mar 14 15:07:55 maybe the path is wrong... though I took it from the gitweb Mar 14 15:08:07 git clone git://git.freesmartphone.org/libeflvala.git libeflvala Mar 14 15:08:40 let me check Mar 14 15:09:44 mickey|CLT: (What's a CLT?) at some point, when you have time and are interested, we should discuss bluez4 and libusb issues, and a missing or extraneous patch. No rush, as it builds for Om builds (fails for SlugOS/EABI) Mar 14 15:10:05 mrmoku: the volume (of calls, percieved on my end) also seems a little bit lower that it used to be on shr unstable/previous testing, is it just my perception, or have there been changes? Mar 14 15:10:22 IMO and IIRC, libusb is handled in a broken manner for Om builds. Mar 14 15:10:46 Zorkman: well no changes from SHR side... don't know if something changed in between milestone5 and milestone5.5 Mar 14 15:11:15 mwester: Chemnitzer Linuxtage, some conference. yes, lets do that when I'm back in frankfurt (tuesday), things are quite hectic atm. Mar 14 15:11:29 Ok - enjoy the conference! Mar 14 15:12:02 :) Mar 14 15:12:15 thanks Mar 14 15:12:20 have a presentation tomorrow Mar 14 15:12:25 spreading the word about FSO Mar 14 15:12:26 Bye Mickey - Enjoy yourself :) Mar 14 15:12:31 Ah, so you're actually working, not just attending. Mar 14 15:12:32 Be good! Mar 14 15:12:34 mrmoku: ok i think the repo is broken, i will repush it Mar 14 15:12:41 mwester: yeah Mar 14 15:12:47 mickey|CLT: ok, thanks Mar 14 15:13:23 hi, i updated uboot on my neo1973 to version dec-17 2008 Mar 14 15:13:34 but i have problem with my 8gb minisd Mar 14 15:13:49 from flash minisd is correctly detect Mar 14 15:13:59 but from bootloader i have this error: Mar 14 15:14:05 ===============> MMC CMD Timeout Mar 14 15:14:11 and other message Mar 14 15:14:50 probably is only a problem of the bootloader Mar 14 15:14:58 any suggestions? thanks in advance Mar 14 15:17:23 mickey|CLT, hey how were your vacations? Mar 14 15:18:24 mrmoku: it should work now Mar 14 15:18:59 mickey|CLT: unfortunately, it weren't vacations (only for the others). Mar 14 15:19:04 just one week without internet access Mar 14 15:19:13 ah ok :) Mar 14 15:19:13 morphis: i see you started eflvala as well last night :) Mar 14 15:19:27 jepp, I started Mar 14 15:19:47 wasnt that easy as I thought but it is running :) Mar 14 15:19:51 cool Mar 14 15:20:01 i have 90% of elm bound as well Mar 14 15:20:08 take a look in my repo Mar 14 15:20:18 which repo? Mar 14 15:20:24 it doesn't compile atm. since elementary.h neds a patch Mar 14 15:20:33 Vala doesn't like unnnamed structs Mar 14 15:20:42 mickey|CLT: can you take a look at http://trac.freesmartphone.org/ticket/373 ? It seems AlwaysSolicitedParser is still not handling the case of unsolicited messages between command and reply Mar 14 15:21:10 lindi-: oh Mar 14 15:21:16 AlwaysSolicitedParser is still existing? Mar 14 15:21:20 it should not be Mar 14 15:21:23 it's obsolete Mar 14 15:21:31 mickey|CLT: i have not followed the ogsmd rewrite yet Mar 14 15:22:09 mickey|CLT: should I move to the rewritten parser? Mar 14 15:22:57 mickey|CLT, looks very good Mar 14 15:23:21 lindi: the new parser may have some bugs as well, but it solves the issue once and for all Mar 14 15:24:07 mickey|CLT, it is more feature complete than my work, so I think I can drop my work :) Mar 14 15:24:45 morphis: ok, sorry for that, but i had no internet access during that week Mar 14 15:24:52 no problem :) Mar 14 15:24:58 morphis: send me your ssh key for access Mar 14 15:25:07 juhu :) Mar 14 15:25:58 :P Mar 14 15:26:11 mickey|CLT, which patch elemenetary needs for compiling? Mar 14 15:26:15 mickey|CLT: hmm, because of http://trac.freesmartphone.org/ticket/348 i'm seriously considering to replace ogsmd with my own very simple parser which i can debug more easily Mar 14 15:27:35 lindi-: can't stop you from that Mar 14 15:27:42 lindi-: give the new one a try first Mar 14 15:27:49 morphis: one which i lost Mar 14 15:27:58 hm ok Mar 14 15:27:59 morphis: i need to recreate that Mar 14 15:28:09 bahahaha Mar 14 15:28:10 basically the genlist structure needs callback declarations Mar 14 15:28:18 and a name for the struct in struct Mar 14 15:28:19 mickey|CLT: "do not reinvent the wheel" has been giving me motivation to debug the existing code but now that it is rewritten anyway that point kind of goes away :( Mar 14 15:28:20 I send you my ssh pub key Mar 14 15:28:32 lindi-: rewritten in how far? Mar 14 15:28:41 lindi-: itwas a minor patch Mar 14 15:28:56 thankfully i could do this with very little changes Mar 14 15:29:01 mickey|CLT: sure, but changes the whole idea of the parser Mar 14 15:29:23 but yeah, maybe i'll give it a try Mar 14 15:29:33 mickey|CLT: any tips on http://trac.freesmartphone.org/ticket/348 would still be appreciated though Mar 14 15:30:08 true Mar 14 15:30:18 we need something for inspecting the running process Mar 14 15:30:25 will do that with some interpreter running over a socket Mar 14 15:30:28 i have hit that bug now maybe 7 times Mar 14 15:30:52 mickey|CLT: i have a bad feeling it might be gobject, python-gobject, dbus or python-dbus bug Mar 14 15:31:37 mickey|CLT: $ git pull Mar 14 15:31:37 fatal: Couldn't find remote ref refs/heads/mickey/experimental Mar 14 15:32:15 just pull from msater Mar 14 15:33:16 mickey|CLT: msater? Mar 14 15:33:27 maste Mar 14 15:33:28 r Mar 14 15:33:51 mickey|CLT: 'git branch' shows that i am in master branch Mar 14 15:34:39 strange. works here Mar 14 15:34:50 mickey|CLT: http://paste.nerv.fi/60056559.txt Mar 14 15:35:23 try deleting ths branch Mar 14 15:35:41 uh oh Mar 14 15:35:45 are we having problems here? Mar 14 15:36:35 mickey|CLT: http://paste.nerv.fi/65598673.txt Mar 14 15:37:01 hmm Mar 14 15:37:05 bummer Mar 14 15:37:07 try recloning Mar 14 15:37:18 i'm not much of a git expert Mar 14 15:37:39 i always reclone of something gets in a strange state ;) Mar 14 15:37:43 s/of/if/ Mar 14 15:37:43 mickey|CLT meant: i always reclone if something gets in a strange state ;) Mar 14 15:41:18 mickey|CLT: Autogenerated prefixes for command 'AT+CLIP=1\r\n': set(['+CLIP']) Mar 14 15:41:42 mickey|CLT: won't this mean that if AT+CLIP=1 is replied with +CLIP that tells about incoming call it is not understood as incoming call? Mar 14 15:42:41 correct Mar 14 15:42:51 this is an inherent problem of the GSM spec Mar 14 15:43:43 mickey|CLT: did you read my mailing list question about writing calypso documentation. could i just start collecting various calypso specific quirks to wiki? Mar 14 15:44:08 lindi-: I'd rather prefer enhancing my textfile Mar 14 15:44:13 it's in docs/ Mar 14 15:44:39 mickey|CLT: ok but wiki already has some docs for AT% commands Mar 14 15:45:04 mickey|CLT: imho all non-standard commands should be documented to best of our knowledge Mar 14 15:45:15 yeah Mar 14 15:45:19 mickey|CLT: (re CLIP) doesn't it always just return OK? Mar 14 15:45:21 i can only do that with code Mar 14 15:45:31 and this text file Mar 14 15:45:37 as i have access to the original file Mar 14 15:45:39 mickey|CLT: and "+CLIP..." is always the unsolicited version and never a reply? Mar 14 15:45:54 mickey|CLT: there are even some missing include errors when compiling libeflvala Mar 14 15:46:00 mickey|CLT: original file? meaning NDA'd stuff? Mar 14 15:46:04 lindi-: yes Mar 14 15:46:09 a simple --pkg dbus-glib-1 for valac fix this Mar 14 15:46:33 lindi-: for those simple commands, we should not use the prefixmap, but rather override to [] Mar 14 15:46:52 mickey|CLT: wouldn't specifying "OK" there be good idea too? Mar 14 15:47:41 no, OK is always a terminal Mar 14 15:47:51 just giving an empty list will be ok Mar 14 15:50:10 mickey|CLT: so should i start collecting calypso information to that file from wiki and other sources which I can read without NDA? Mar 14 15:50:59 yes, feel free to go ahead Mar 14 15:51:36 ok Mar 14 15:52:37 mickey|CLT, without genlist everything is working fine Mar 14 15:55:38 yeah Mar 14 15:56:00 it's just the genlist part that has this unnamed struct and callbacks Mar 14 15:56:03 if you add my ssh-pub-key I can fix the problem with dbus when compiling Mar 14 15:56:04 raster needs to change that upstream Mar 14 15:56:19 I even can do it, if raster allow :) Mar 14 15:56:44 mickey|CLT: hmm, Hardware:AT_Commands.wiki seems to already document quite a few undocumented commands Mar 14 15:57:56 mickey|CLT: and the text file says '* Too many to list here, please see http://wiki.openmoko.org' Mar 14 15:59:02 what problem compiling? Mar 14 15:59:33 raster: no compiling problem with elementary, only with libeflvala :) Mar 14 15:59:45 raster: but mickey has something for you to fix in elementary Mar 14 15:59:51 oooh vala Mar 14 16:00:11 you need to name the structs Mar 14 16:00:18 and declare the callbacks Mar 14 16:00:21 then vala will be happy Mar 14 16:00:26 oh and Mar 14 16:00:36 in one case you have ElementaryFoo where it should be ElmFoo Mar 14 16:00:56 the patch i had was <= 10 lines Mar 14 16:01:00 but i lost it Mar 14 16:01:39 hmm Mar 14 16:01:46 which structs need naming?> Mar 14 16:03:20 a substruct in genlist Mar 14 16:03:34 which contains the callbacks Mar 14 16:04:56 oooh Mar 14 16:05:03 vala doesnt like that? Mar 14 16:05:14 and it insists u typedef callbacks? Mar 14 16:05:20 yep Mar 14 16:06:39 mickey|CLT: I sent 'AT+COPN' with Debug interface and got UnicodeDecodeError: 'ascii' codec can't decode byte 0xf8 in position 17: ordinal not in range(128): http://paste.nerv.fi/21321348.txt Mar 14 16:07:20 yeah Mar 14 16:07:36 that's a bug in the calypso#s table Mar 14 16:07:39 ptoohey Mar 14 16:07:44 so much extra footwork Mar 14 16:07:51 it wont like a lot fo evas and edje too Mar 14 16:08:04 lindi: better use the dbus call where i worked around that problem Mar 14 16:08:07 bbl Mar 14 16:09:07 raster: is it right that an entry is growing when inserting text wich is wider than the entry? Mar 14 16:10:19 errr yes Mar 14 16:10:24 the entry always sizes to its contents Mar 14 16:10:32 if u dont want that - put it in a scroller Mar 14 16:10:35 so u can scroll Mar 14 16:10:36 ah ok Mar 14 16:10:39 bug in the calypso#s table?? Dieter! Mar 14 16:10:52 unlike gtk an entry doesnt include a frame/scroller with it Mar 14 16:10:54 its fairly raw Mar 14 16:11:04 i could add a "scrolledentry" wrapper Mar 14 16:11:33 maybe better Mar 14 16:12:00 another question, can you include some functionality to disable/enable horizontal/vertical scrolling for a scroller? Mar 14 16:13:07 its automatic Mar 14 16:13:24 so why disable it if something needs to scroll in that dir? Mar 14 16:13:28 ie contents are too big? Mar 14 16:14:04 i have code that as part of the list for example forced contents to not exceed the scroller parent size horizontally if u choose that Mar 14 16:14:13 thus the ability to compact labels to hello... Mar 14 16:14:33 ah ok, thats even good Mar 14 16:15:34 but its done withint he widget Mar 14 16:15:42 as it also uses the scroller together with it Mar 14 16:15:49 if u want a specific case... tell me about it Mar 14 16:16:04 eg u want a "limited entry" Mar 14 16:16:09 no, I dont have a specific case, I only remember that in other toolkits it is possible Mar 14 16:16:12 that forces contents to linewrap Mar 14 16:16:18 thats what anchorblock does Mar 14 16:16:20 for example Mar 14 16:16:25 already Mar 14 16:16:45 ok, I thing there sould be a good documentation better than creating some new widgets Mar 14 16:18:51 only have so much time :) Mar 14 16:19:35 :) Mar 14 16:20:46 men, in what unit does sizeof answer ? Mar 14 16:20:50 byte ? Mar 14 16:20:58 in C Mar 14 16:21:03 size_t ? Mar 14 16:21:07 ptitjes: chars Mar 14 16:21:15 chars or bytes ? Mar 14 16:21:43 ptitjes: bytes Mar 14 16:22:37 if i have a void* stuff, and I make stuff+x, will my pointer advance to x bytes ? Mar 14 16:23:10 ptitjes: i don't think that is defined? Mar 14 16:23:19 and also printf("%p", stuff) does print in bytes too ? Mar 14 16:24:24 ptitjes: cast to char * - add +x for bytes u want Mar 14 16:24:27 then do whatever Mar 14 16:24:34 inf act u can happily just use char * Mar 14 16:24:36 and be done with it Mar 14 16:24:41 case later when needed/wanted Mar 14 16:24:53 (if u want to fiddly with your pointer byte by byte) Mar 14 16:25:56 raster: i do printf("%p", stuff) and printf("%p", ((char*) stuff) + 8) Mar 14 16:26:11 but this prints : Mar 14 16:26:12 ** Message: greflect-proxy.vala:51: 0x8051748 Mar 14 16:26:12 ** Message: greflect-proxy.vala:53: 0x8051750 Mar 14 16:26:38 so printf("%p", stuff) does not print in byte, am I wrong ? Mar 14 16:27:06 umm %p shoudl print the pointer as is Mar 14 16:27:08 as a value Mar 14 16:27:10 ie in bytes Mar 14 16:27:13 whats your problem? 0x48+8 is 0x50 Mar 14 16:27:13 no matter what type Mar 14 16:27:26 ok :) Mar 14 16:27:29 I'm bad Mar 14 16:27:34 thank you all Mar 14 16:27:35 you have exactly the result i'd expect there Mar 14 16:27:38 you're bad Mar 14 16:27:42 hahaha Mar 14 16:27:44 :D Mar 14 16:27:49 ;D Mar 14 16:32:18 ok this works Mar 14 16:32:21 thank you all Mar 14 16:36:45 spaetz: ping Mar 14 16:40:15 spaetz: for me that layout of keys in calculator is much more friendly: http://scap.linuxtogo.org/files/857eaa4dbda8edf61ffa78ed0baae8df.png Mar 14 16:40:47 what do you think about that? Mar 14 16:47:40 dos1: I agree with you Mar 14 16:48:08 first I was confused, because the equal button was not in the bottom right corner Mar 14 16:54:19 is there a mean to produce a segfault in C ? Mar 14 16:54:35 I mean with a single instruction ? Mar 14 16:54:45 exit(1) would do ? Mar 14 16:55:13 assert? Mar 14 16:56:19 ok simply set a breakpoint in gdb... Mar 14 16:56:41 but anyway it seems gdb know my symbol names even I don't set -g in my gcc Mar 14 16:57:03 is there a mean to retrieve the symbol name from C just as gdb does ? Mar 14 16:58:51 say I have this : Mar 14 16:58:53 (gdb) bt Mar 14 16:58:53 #0 reflect_int_callback (object=0x8051a00) at greflect-proxy-internals.c:159 Mar 14 16:58:53 #1 0x08048c41 in test_test_interface_test_method () Mar 14 16:58:53 #2 0x08048e9d in test_main () Mar 14 16:58:53 #3 0x08048ee3 in main () Mar 14 16:59:16 is it possible to retrieve the 'test_test_interface_test_method' name from c ? Mar 14 17:01:16 ptitjes: for what purpose? Mar 14 17:01:35 :) Mar 14 17:01:43 lindi-: still the dynamic proxy stuff Mar 14 17:01:55 infrastructure does start to run Mar 14 17:01:55 ptitjes: dynamic proxy? Mar 14 17:02:28 dynamic proxy = implementing a class at runtime Mar 14 17:02:41 s/class/interface/ Mar 14 17:02:42 ptitjes meant: dynamic proxy = implementing a interface at runtime Mar 14 17:03:24 ptitjes: in C? Mar 14 17:03:46 lindi-: yes Mar 14 17:04:07 ptitjes: how would that work? Mar 14 17:05:24 black magic :) Mar 14 17:05:35 ooooh. Mar 14 17:06:16 awesome :) Mar 14 17:06:25 lindi-: in fact, I implement an interface GType and put generic callbacks for each virtual methods Mar 14 17:06:37 ptitjes: virtual methods don't sound like C Mar 14 17:07:22 I create a GType that is said to implement the interface GType Mar 14 17:08:18 For each virtual methods of that interface, I set a generic callback that route the call to an instance of an InvocationHandler interface Mar 14 17:08:24 This works for now Mar 14 17:08:51 ptitjes: that again does not sound very C, is that some vala stuff? Mar 14 17:08:57 however I have to figure out in the callback what virtual mathod has been called Mar 14 17:09:06 lindi-: it is some G stuff Mar 14 17:09:12 not vala specific Mar 14 17:09:14 at all Mar 14 17:09:31 so I have to walk the stack Mar 14 17:09:47 get the second pointer in the call stack Mar 14 17:09:56 retrieve the symbol name for that pointer Mar 14 17:10:00 and then do my stuff Mar 14 17:10:13 how can I do that ? Mar 14 17:10:18 ptitjes: sounds very ugly and non-portable Mar 14 17:10:32 (retrieve the second pointer in the call stack and get the symbol name for it ? Mar 14 17:10:58 lindi-: maybe but this is not my point for now Mar 14 17:11:11 I'll make it portable later Mar 14 17:11:46 i'm not sure if that can be made portable Mar 14 17:11:57 does someone know how does gdb proceed ? Mar 14 17:12:45 ptitjes: GDB uses BFS to parse DWARF debug info of the ELF binary Mar 14 17:17:51 ok let's start simple for now Mar 14 17:17:57 how to walk the call stack ? Mar 14 17:21:14 ptitjes: you're not supposed to, except you're coding assembler Mar 14 17:23:14 DocScrutinizer: I found that glibc has some backtrace() and backtrace_symbols() Mar 14 17:24:13 ptitjes: why you're programming OO when you have to revert to such an ugly machine language level hack? Mar 14 17:24:39 2sec phone Mar 14 17:24:57 :-/ Mar 14 17:26:10 strong indication sth is terribly wrong in system design Mar 14 17:26:44 a botch won't make things better. It never did Mar 14 17:28:16 DocScrutinizer: ? Mar 14 17:28:40 If OO has too much overhead, do it in plain C. If you van't do it in C, do it in assembler. If you can't do that, it's probably not worth being done at all Mar 14 17:28:56 DocScrutinizer: because Vala is based on C, and C no notion of dynamic programming Mar 14 17:30:05 DocScrutinizer: here is an example of what I want to do: http://valadoc.org/paste.php?id=295 Mar 14 17:31:28 DocScrutinizer: this is not the point why I am doing this Mar 14 17:31:41 DocScrutinizer: if you want use cases, then it is easy Mar 14 17:32:28 dynamic proxying enables me to not generate code for generic implementations of some interfaces Mar 14 17:32:34 an example is mocking Mar 14 17:32:47 Well I never heard anybody walking the stack to get to know about details of his callbacks Mar 14 17:33:03 DocScrutinizer: in mocking, all the methods are implemented the same way Mar 14 17:33:16 either I generate code, either I use dynamic proxying Mar 14 17:33:21 either you know implicitly, or by plain value as a parameter Mar 14 17:33:46 DocScrutinizer: except if the callback is the same everywhere Mar 14 17:34:02 DocScrutinizer: also I can't use the parameters as I don't control the way the callback is called Mar 14 17:34:25 that's a clear design flaw Mar 14 17:34:53 DocScrutinizer: GLib has been design to be really close to how C works Mar 14 17:35:13 if you need to distinguish callbacks, the callback has to implement a way to do so Mar 14 17:35:13 this explains why they don't transport the method names along to the other method parameters Mar 14 17:35:50 DocScrutinizer: I don't know a priori the number of use of the callbacks Mar 14 17:36:02 The other way down would be to instanciate my callbacks Mar 14 17:36:16 that's the natural way Mar 14 17:36:23 but lindi-, I think, told me it can't be done because W^X Mar 14 17:36:48 DocScrutinizer: but if it is possible then I want to know how to instanciate a callback Mar 14 17:36:55 in C Mar 14 17:37:31 in fact that is the way I would prefer Mar 14 17:38:14 I have no idea of programming ;) I'm hardware Mar 14 17:38:33 ask a hardcore C hacker Mar 14 17:38:52 err C++ probably Mar 14 17:40:50 cause afail in C it's hard to instantiate anything at all Mar 14 17:41:40 DocScrutinizer: thank you this is what I'm doing here Mar 14 17:41:59 programming in C using glib and gobject is the same thing as C++ overhead wise. Mar 14 17:42:03 any other knows how I can allocate a callback in C ? Mar 14 17:42:32 that question sounds wrong Mar 14 17:42:40 ptitjes: what you mean allocate a callback? storing a function pointer and later calling it? Mar 14 17:43:28 vininim: allocate a callback in memory Mar 14 17:44:14 [2009-03-14 18:40:50] cause afail in C it's hard to instantiate anything at all Mar 14 17:44:40 a callback is a pointer to a function Mar 14 17:44:41 ptitjes: you want to allocate a chunk of memory and store machine code in it? Mar 14 17:44:52 vininim: that is allocating a memory block where I can put some C code to be executed Mar 14 17:44:54 won't work Mar 14 17:45:09 tilman: exactly Mar 14 17:45:21 DocScrutinizer: so JIT compilers don't exist? ;> Mar 14 17:45:46 nah, but no decent precompiled language works like this Mar 14 17:45:59 tha's selfmodifying code basically Mar 14 17:46:05 ptitjes: let me dig some, i wrote something like that some years ago Mar 14 17:46:16 tilman: thank you very much :) Mar 14 17:46:26 ptitjes: the closets you get to that is defining several functions with same signature(and function) and using a function pointer to select them. Mar 14 17:46:28 i think it worksl ike that: mmap w/ PROT_WRITABLE, then write code then mmap w/ PROT_EXEC Mar 14 17:46:49 but i'm not sure if you're really solving the right problem :D Mar 14 17:46:54 tilman: WTF is the use of storing the ever-identical programtext to a plethora of memchunks??? Mar 14 17:47:26 DocScrutinizer: that won't be ever-identical programtext Mar 14 17:47:31 he just wants a callback functionallity if I understands correctly, and that's what is a function pointer for =P Mar 14 17:47:34 I can add a parameter Mar 14 17:47:41 what you need here is sth like a jumptable, plus dynamic allocation of local variables Mar 14 17:48:07 DocScrutinizer: i didn't follow all of the discussion; i'm just answering the question :P Mar 14 17:48:29 men please be correct Mar 14 17:48:46 do you know what dynamic proxying is ? Mar 14 17:48:58 ptitjes said he has all the same callback function all over the place and he want's to know which origin the callback returns from Mar 14 17:49:38 ptitjes: design pattern wise? Mar 14 17:50:02 DocScrutinizer: I have a GLib GType, which enables me to set callbacks in a vtable Mar 14 17:50:42 I have no idea of GLib Gtype, anyway a callback is a callback Mar 14 17:50:52 no matter which language Mar 14 17:51:12 DocScrutinizer: thank you very much for your non-constructivity Mar 14 17:52:02 ! Mar 14 17:52:02 welcome Mar 14 17:52:15 DocScrutinizer: please! you don't even have the slightest idea of what I need to do so please stop Mar 14 17:53:28 a pleasure Mar 14 17:54:20 tilman: where could I find information about mmap ? Mar 14 17:55:23 ptitjes: back then i used mmap(NULL, ..., PROT_READ | PROT_WRITE | PROT_EXEC, ...). not sure _why_ it worked... W^X etc Mar 14 17:55:37 oh Mar 14 17:55:52 finish with mprotect(, PROT_READ | PROT_EXEC) ;) Mar 14 17:56:36 ptitjes: http://9dcc6c445ba58c72.paste.se/ Mar 14 17:58:47 Doctor :) Mar 14 18:02:20 even a fork() doesn't copy programtext Mar 14 18:04:20 and for good reason, either it'd be all the same, or you had it somewhere else in memory already. what's duplicated is variable storage, nada mas Mar 14 18:06:45 DocScrutinizer: programtext? Mar 14 18:07:22 DocScrutinizer: it's probably read-only mapping anyway? Mar 14 18:09:44 yup usually Mar 14 18:10:32 that's why self-modifying code in any form doesn't usually work. It's deprecated Mar 14 18:12:59 lindi-: I always used callbacks by passing the address to triggering_object together with any parameters the callback might expect when called. Triggering_object is responsible to push those on the stack (or whatever) on calling the callback Mar 14 18:18:14 if I can't do this (e.g. triggering_object isn't under my control parameter-def-wise), I have to use an instance of callback address for each triggering_object. Instance means different callback address, though all these can share same code stubs of course, except maybe a oneline 'calledby="triggerobj1"' Mar 14 18:19:52 nice Mar 14 18:20:10 aw, self-modifying code is deprecated, joerg? Mar 14 18:20:21 yup Mar 14 18:20:30 too bad :P Mar 14 18:20:30 huh? nothing new Mar 14 18:20:37 right Mar 14 18:20:48 DocScrutinizer: so you have a fixed number of callbacks anyway Mar 14 18:21:31 by fixed I mean finite Mar 14 18:21:45 yup, except you use a jumptable Mar 14 18:22:12 DocScrutinizer: so you can't do that dynamically Mar 14 18:22:34 but honestly this is very bad style, usually the triggering_obj should return any info the callback needs Mar 14 18:22:52 a jumptable can be dynamic Mar 14 18:22:53 my lib has a priori no knowledge for what interfaces callbacks will be made Mar 14 18:24:06 so I can't say things like: I will ever have at most 10 different callbacks Mar 14 18:24:14 so your scheme does not fit Mar 14 18:24:23 (that would have been too simple...) Mar 14 18:24:42 s/table/list/ Mar 14 18:24:43 DocScrutinizer meant: a jumplist can be dynamic Mar 14 18:25:04 DocScrutinizer: yeah you can change the pointers Mar 14 18:25:09 so great :D Mar 14 18:25:19 if you really think you can't tell any upper limit Mar 14 18:25:45 that is not the problem joerg but you don't want to understand Mar 14 18:26:11 and yes I can't tell any upper limit Mar 14 18:26:29 and anyway, would you see me generate 100 callbacks Mar 14 18:26:43 when sometime a program will only use 10 Mar 14 18:26:50 that is plain stupid Mar 14 18:27:10 no I need either to allocate callbacks at will Mar 14 18:27:27 either I have only one callback and can determine the context of the call Mar 14 18:27:40 and pass an identifier to triggering_obj along with address of callback Mar 14 18:27:56 DocScrutinizer: you mean a closure Mar 14 18:28:30 but then how do I transform the closure to a real C callback without allocating memory with C code ad make it executable ? Mar 14 18:29:38 allocate varaible storage space on heap, and pass the pointer to that to the callback Mar 14 18:30:38 and there again I'm not the one who calls the callback Mar 14 18:31:50 I only put the callback pointers in a vtable Mar 14 18:45:57 ptitjes: looks like you can place any code in an mmaped region and it can be executed without issues with NX bit. Mar 14 18:46:21 (provided you mmaped it in a right way) Mar 14 18:48:02 PaulFert`: yeah, this is what I'm reading Mar 14 18:49:04 ptitjes: i still wonder how you decided to deal with calling conventions differences. Are you going to assume it's cdecl and supply the address of the first argument to va_start equivalent with some hack? Mar 14 18:49:38 yeah Mar 14 18:49:59 I have to implement as many base callback as there are possible return types Mar 14 18:50:27 I will allocate memory and copy the content of one of those callbacks depending on the method signature Mar 14 18:50:36 (mmaped correctly) Mar 14 18:50:54 Yes, no way around that it seems. Still looks like a horrible hack, but well, it is cool for sure! :) Mar 14 18:51:31 I have to figure out how to modify the allocated callback to change an literral int value Mar 14 18:51:54 or I will compare the pointer (from the stack) of the callback to all the callbacks of the objects Mar 14 19:04:11 anyone here who can tell me how i get python, sqlite3 and CURRENT_TIMESTAMP to automatically fill a field with the current date on INSERT Mar 14 19:31:49 might it be that shr-testing's gps support is broken? Mar 14 19:31:56 i absolutely cannot get a fix Mar 14 19:35:39 tomboy64: i think shr-testing's gps users are broken :D Mar 14 19:35:47 tomboy64: debug it with Zhone Mar 14 19:35:57 dos1, i tried Mar 14 19:36:00 don't get a fix. Mar 14 19:36:17 but have you looked at signal stripes? Mar 14 19:36:32 that's the most important part of GPS panel in zhone :P Mar 14 19:36:45 the most i get is gps-time Mar 14 19:36:57 dos1, where do i see the signal-stripes in zhone? Mar 14 19:37:06 tomboy64: the middle pane Mar 14 19:37:21 that doesn't work Mar 14 19:37:33 at the bottom right i have the x (close) Mar 14 19:37:35 tomboy64: does it show anything at all? or no sat list even? Mar 14 19:37:49 no sat list Mar 14 19:37:54 ... Mar 14 19:37:55 the only thing i've got is the gps-time Mar 14 19:38:08 and that's the most i managed to get the last week Mar 14 19:38:15 on bottom of screen you have to columns of buttons Mar 14 19:38:23 s/to/two/ Mar 14 19:38:23 dos1 meant: on bottwom of screen you have to columns of buttons Mar 14 19:38:31 ... Mar 14 19:38:40 culumns? 2 rows you mean? Mar 14 19:38:43 on bottom of screen you have two columns of buttons Mar 14 19:38:55 Three columns, two rows. Mar 14 19:38:59 wow! Mar 14 19:38:59 columns; both with 3 rows Mar 14 19:38:59 hmm Mar 14 19:39:00 heh Mar 14 19:39:03 now i found them Mar 14 19:39:10 yikes Mar 14 19:39:13 mine english isn't good :D Mar 14 19:39:14 let's test :-D Mar 14 19:39:21 so i was correct Mar 14 19:39:25 dos1, you're here to help - i appreciate that :) Mar 14 19:39:26 users are broken ;) Mar 14 19:39:32 pffft Mar 14 19:39:36 :D Mar 14 19:39:51 dos1: thanks for the calculator layout suggestions Mar 14 19:39:53 * tomboy64 runs to the window Mar 14 19:39:57 brb Mar 14 19:40:39 implemented. I hadn't given much thought to the layout. I believe I was inspired by the gnome-calc Mar 14 19:40:41 morphis: you should have access now Mar 14 19:41:03 mickey|chemnitz: are you going to openexpo in Switzerland too? Mar 14 19:41:16 spaetz: np ;) Mar 14 19:41:17 100% Mar 14 19:41:22 i have a talk about fso there Mar 14 19:41:58 mmh, cool. I might have to pop in one day. I've heard your talk last year. I was one of the openstreetmap guys next to your openmoko stnd :-) Mar 14 19:42:19 ah, nice Mar 14 19:42:29 this year it's going to be a more interesting talk Mar 14 19:42:31 ;) Mar 14 19:42:37 hehe Mar 14 19:43:18 mrmoku: care to implement the autorev thingie for the claculatur? Mar 14 19:43:29 mickey|chemnitz: any plans on a location based service (which not only uses gps but gsm and wlan too) Mar 14 19:43:34 currently the srcrev is just hardcoded in the .bb file Mar 14 19:45:18 bumbl: yeah. something like olocationd is definitly planned Mar 14 19:45:32 on the community or devel list there were some first sketches iirc Mar 14 19:45:36 oh no. more *d apps :) Mar 14 19:45:46 hehe Mar 14 19:47:04 cheers dos1 Mar 14 19:47:08 you made my day Mar 14 19:47:13 * tomboy64 's got a fix Mar 14 19:49:07 Hello. I have just build a Y USB connector to inject power into my Freerunner (and some other device) while being in host mode, and I have successfully connected and powered a USB keyboard through this method, using the switches in SHS's settings, but I can't get my Freerunner to recieve power. Mar 14 19:50:01 There's some hint on echo:in to /sys/devices/platform/neo1973-pm-host.0/hostmode, but I don't even have that file (or a similar one) Mar 14 19:51:14 Also, if I connect a wall charger to my modded USB cable, the power LED will light up in red if it's in device mode, so I know it can recieve power. Mar 14 19:51:52 LqR, hmmm Mar 14 19:52:07 LqR: find /sys -name hostmode finds nothing? Mar 14 19:52:09 i have latest shr-testing running and experience the same thing with a normal usb-cable Mar 14 19:52:44 i wondered if the cable was broken, but usb-netwerkin werked fine Mar 14 19:55:36 spaetz: argh... was about to do the autorev thingie for calc ;) Mar 14 19:57:21 mrmoku: patches! Mar 14 19:57:23 error: patch failed: packages/tangogps/tangogps.inc:6 Mar 14 19:57:23 error: packages/tangogps/tangogps.inc: patch does not apply Mar 14 19:57:23 error: patch failed: packages/tangogps/tangogps_0.9.5.bb:1 Mar 14 19:57:23 error: packages/tangogps/tangogps_0.9.5.bb: patch does not apply Mar 14 19:57:24 ;) Mar 14 19:57:51 dos1: tangogps 0.9.5 is no more in unstable... remove it Mar 14 19:58:28 ok Mar 14 19:58:36 DOS1 Mar 14 19:59:38 mrmoku: I was quicker :-) But not sure I did it right Mar 14 19:59:39 mrmoku: only 0001-tangogps-add-recipe-for-0.9.5.patch, yep? Mar 14 20:00:21 I added autorev to the include file but I left an SRCREV ?= "..." in the .bb file (in case that someone wants to use it in testing or so Mar 14 20:00:40 Not sure if it will build all right this way. Mar 14 20:00:42 dos1: let me check what the .inc patch does Mar 14 20:01:20 hmm it adds the desktop file... Mar 14 20:01:37 let me check if that is upstream Mar 14 20:01:53 mrmoku: and next: Mar 14 20:01:54 error: patch failed: packages/netbase/netbase_4.21.bb:2 Mar 14 20:01:56 error: packages/netbase/netbase_4.21.bb: patch does not apply Mar 14 20:01:56 error: patch failed: packages/base-files/base-files/om-gta02/fstab:1 Mar 14 20:01:56 error: packages/base-files/base-files/om-gta02/fstab: patch does not apply Mar 14 20:02:00 :( Mar 14 20:02:19 DAVE Mar 14 20:02:30 Yes Mar 14 20:02:36 No Mar 14 20:02:39 dos1: that one is, because part of the fastboot stuff is in milestone5.5 Mar 14 20:04:45 dos1: guess you can remove both tango patches Mar 14 20:04:58 it does not apply and we have a desktop icon... kill it :-) Mar 14 20:05:21 spaetz: should work the way you did it Mar 14 20:06:01 btw... I added it to shr-autorev.inc... because we have different branches it does not influence testing Mar 14 20:11:14 mrmoku: patches removed Mar 14 20:11:48 dos1: did you ever take a look at log_compile for libframeworkd-phonegui-efl? Mar 14 20:12:06 mrmoku: why are you asking? Mar 14 20:12:11 mrmoku: ok, fine with me Mar 14 20:12:38 because I'm fixing those millions of compile warnings - quickdev was a lazy ass :P Mar 14 20:12:46 ;D Mar 14 20:12:52 hahha Mar 14 20:13:00 gogo mrmoku! Mar 14 20:14:10 lindi-: It found /sys/class/i2c-adapter/i2c-0/0-0073/neo1973-pm-host.0/hostmode Mar 14 20:19:17 Hm. There, I set it to 0 as adviced, but seemingly nothing happened. Mar 14 20:19:47 Still discharging. Mar 14 20:22:55 haaaa! libffi's black magic is cool! Mar 14 20:23:20 DocScrutinizer, lindi-, PaulFert`: found a way to make it portable :) Mar 14 20:24:35 ptitjes: Wow :) Are you going to share the details someday? Mar 14 20:24:41 libffi can create real C callbacks from closures (wrapped callback+userdata) Mar 14 20:24:50 PaulFert`: for sure! Mar 14 20:25:01 let me finish it for now :) Mar 14 20:36:19 mrmoku: can you regenerate calc? Mar 14 20:36:35 opkg thinks that old version is newer :x Mar 14 20:38:00 ouch... yeah Mar 14 20:40:06 dos1: tell m eif it doesn't work right Mar 14 20:40:45 I was thinking about a button that slides the basic keyboard away and gives an advanced keyboard (sin, cos, power, mod, etc) Mar 14 20:41:25 but that's low priority. shr-settings comes first Mar 14 20:46:50 Closures? What language are you talking about, ptitjes? Vala? Mar 14 20:47:31 LqR: C Mar 14 20:48:06 There's closures in C? Mar 14 20:48:13 I had no idea. Mar 14 20:48:43 I thought closures were rather lisp/scheme-exclusive Mar 14 20:50:25 dos1: works now? Mar 14 20:54:59 mrmoku: ok ,works Mar 14 21:13:54 hey, how can I disable vibration? Mar 14 21:14:13 I remember seeing something in the ML about someone who managed to add a rule option for that. Mar 14 21:17:10 good.. goood ... gooood VIBRATIONS ;) Mar 14 21:18:18 yay alphaone Mar 14 21:18:38 Hey dave Mar 14 21:20:50 TAsn: guess via profile Mar 14 21:21:43 mrmoku, what? Mar 14 21:22:00 Greetings Mar 14 21:22:34 TAsn: disable vibration Mar 14 21:22:50 or do you mean manually disable it while vibrating? Mar 14 21:23:04 mrmoku, I mean in general Mar 14 21:23:34 so in profile Mar 14 21:23:40 for rules there is StartVibration and StopVibration Mar 14 21:23:50 I cannot find where it is configured though... Mar 14 21:25:31 Error: Cannot stop vibration Mar 14 21:25:57 Dave: pkill openvibe *duck* Mar 14 21:26:03 :O Mar 14 21:26:10 Dave: tell her to pull the battery :P Mar 14 21:26:13 tremors! Mar 14 21:26:22 mrmoku, pfft, like THAT would ever happen ;) Mar 14 21:27:07 LqR: a closure is a callback and some data to be passed to the closure Mar 14 21:27:34 callback Mar 14 21:40:21 ptitjes: I was going to add "...but I never really understood the concept on the other hand", now you saved me the trouble ;) Mar 14 21:47:41 for some reason adding ppp-generic to the modutils dir Mar 14 21:47:47 doesn't start it automatically Mar 14 21:47:50 I have to do it manually Mar 14 21:47:52 any ideas? Mar 14 21:47:53 ;\ Mar 14 21:49:23 nvm Mar 14 21:49:32 just figured out I forgot running update-modules Mar 14 21:56:12 Tasn! Mar 14 21:56:32 Great, after months and months and months away, I decide to put the FSO m5.1 illume-image to see how things are... and after all this time, it STILL does not display the battery charge correctly. Mar 14 21:57:09 Dave, ! Mar 14 21:57:58 cesarb, they closed your bug report without fixing it? that's odd! Mar 14 21:58:35 TAsn: I do not recall filing any bug report about it Mar 14 21:58:58 cesarb, oh, so you are complaining about them not fixing a bug you haven't reported about? Mar 14 21:59:06 ;\ Mar 14 21:59:12 TAsn: but I will now get the fso makefile, find out how to build things, and fix the damn thing Mar 14 21:59:32 * cesarb did not get the hang of the fso makefile last time, I hope this time I manage to do it Mar 14 21:59:39 cesarb, that's another possibility. enjoy. Mar 14 21:59:44 I prefer to file bug reports with a patch attached ;-) Mar 14 22:00:10 TAsn: but since you already mentioned it, do you know where their bug tracker is? Mar 14 22:00:29 http://trac.freesmartphone.org Mar 14 22:00:30 iirc Mar 14 22:00:36 TAsn: thx Mar 14 22:00:41 np. Mar 14 22:00:46 anyhow, gn ppl. Mar 14 22:02:34 good night Mar 14 22:21:12 Leaving us so soon, mrmoku? Mar 14 22:26:14 Dave: no... was for TAsn :P Mar 14 22:26:36 should have said 'TAsn: good night' ;) Mar 14 22:26:44 Goodnight Tasn Mar 14 22:26:47 don't come back soon :P Mar 14 22:27:05 i'm leaving too Mar 14 22:27:08 good night :x Mar 14 22:28:05 haha Mar 14 22:28:07 fucking dos1 Mar 14 22:28:20 sneaks out when he gets the chance Mar 14 22:28:25 that bastard :P Mar 14 22:38:58 hmm Mar 14 22:39:13 fucking python Mar 14 22:43:25 na na na Mar 14 22:43:58 mickey|chemnitz: i don't understand it Mar 14 22:44:01 hehe Mar 14 22:44:04 what's the problem? Mar 14 22:44:05 it complains it gets Mar 14 22:44:07 """" Mar 14 22:44:19 ( it expects """) Mar 14 22:44:31 and it gets """ Mar 14 22:45:16 mickey|chemnitz: try to run that http://rafb.net/p/xvsjMb35.html Mar 14 22:45:54 the problem are the two execute = lines Mar 14 22:45:59 the first one works Mar 14 22:46:04 the second one doesn't Mar 14 22:46:37 hmm, why do you escape all that? Mar 14 22:46:42 why don't you just use ' Mar 14 22:47:14 mickey|chemnitz: hmm that would be an idea too ;) Mar 14 22:56:20 mickey|chemnitz: still getting a sqlite3.OperationalError: near """": syntax error Mar 14 23:00:23 hmm, then something in your table expression is wrong Mar 14 23:00:37 I would always use values with tuples Mar 14 23:00:44 had less problems that way Mar 14 23:01:02 like Mar 14 23:01:04 SQL_SONG_ADD = "INSERT INTO Songs (fkAuthor, Name, URI, Cached, Duration, PlayedCounter, ProblemsFound) VALUES (?,?,?,'FALSE' Mar 14 23:01:05 ,?,0,'FALSE');" Mar 14 23:01:18 i do that ;) Mar 14 23:01:22 oh Mar 14 23:01:48 but i need a way to change Songs too Mar 14 23:01:49 then i have no idea Mar 14 23:02:16 make sure that there are no " in 'tables Mar 14 23:02:29 and make sure table is a string Mar 14 23:02:35 not something that gets auto repr'd Mar 14 23:04:47 table is a string Mar 14 23:04:59 but i think i need some \" anyway Mar 14 23:05:09 as sql complains about ' Mar 14 23:06:57 raster: ping Mar 14 23:07:12 raster: this is the patch i need to apply to the headers to make everything work Mar 14 23:07:15 http://pastebin.ca/1361027 Mar 14 23:07:43 hmm enough for today Mar 14 23:07:45 freesmartphone.org: 03mickey 07libeflvala * r55cd3a39601a 10/patches/elementary-header.patch: add elementary patch necessary to compile the vapi Mar 14 23:07:58 will debug tomorrow Mar 14 23:17:43 can somebody help me with gta01, 2.6.29 and suspend? Mar 14 23:46:04 Just for the record, Instead of listening to DocScrutinizer (jOERG) I followed Werner Almerberger advise on how to bring back my freerunner to a functional level and it worked! Mar 14 23:46:21 0 points for DocScrutinizer and 12 points to Werner Mar 14 23:46:48 (DocScrutinizer: your assumptions were very rude!) Mar 14 23:47:48 oh yeah, I see. Never recognized NOR uboot will be broken on GSM-flash failing Mar 14 23:47:59 my fault Mar 14 23:48:49 ingi: eat your points Mar 14 23:49:32 DocSy: I quote you "(03:20:16 PM) DocScrutinizer: ingi: and are you sure you have a Frerunner, not a Neo 1973?" Mar 14 23:49:32 Do you think everybody else than you is an idiot ? Mar 14 23:49:53 I know at least *one* Mar 14 23:50:05 he he :-) Mar 14 23:50:13 * tomboy64 raises an eyebrow Mar 14 23:53:29 DocSy: at least thank you for your concern :-) Mar 15 00:09:11 * mwester is confused, and slightly concerned -- there is absolutely no connection whatsoever between the GSM and NOR/NAND flash. Mar 15 00:09:23 ingi: Please provide links to the advice you reference. Mar 15 00:10:12 mwester: yup. :) Mar 15 00:11:17 nevertheless ingi thought I'm kidding when I told it's a bad idea to flash NOR to fix a failed GSM-flashing Mar 15 00:12:02 in fact ingi refused to listen as written above Mar 15 00:13:47 or at least there wasn't any answer, so it's impossible to help. Anyway I don't see what for *I* get (no) points as I did nothing except pointing to this NOR fact and asking questions in vain Mar 15 00:21:01 well, ingi isn't answering my simple question either. Clearly there is a misunderstanding of some sort, and if it is (yet again) bad information in the wiki that needs to be clarified, that should be done before someone else gets confused. Mar 15 00:21:11 1st, I try not to be rude to anyone Mar 15 00:21:12 2nd, I managed (completely on my own) to screw my Openmoko Freerunner (GTA02) when trying to upgrade the GSM firmware to Moko10. Mar 15 00:21:12 I never got any 'usb'-connection to the Freerunner from my Linux PC (the usb-interface never showed up). The phone complained about g_ether.setaddr couldn't be set. Mar 15 00:21:21 For example, the bad page on using u-boot now has everyone using NOR to boot, which is quite wrong. Mar 15 00:21:43 I asked in the 'openmoko.user'-newsgroup, see "Help! GSM not working after updating to Moko10" starting 2009-02-26 Mar 15 00:21:45 ingi: link to the page you reference? Mar 15 00:22:31 I flashed many different rootfs/kernel, but to no avail. Mar 15 00:23:03 If you can flash rootfs/kernels, then your u-boot is fine. Mar 15 00:24:04 Werner's suggestion, in the newsgroup answer, was a bit wague (for me that is), but today I finally understood what Werner was suggesting and reflashed a recent u-boot so the problem was solved. Mar 15 00:24:10 But I don't know any newsgroups for openmoko, so I cannot see the conversation to konw what the issue might be. Mar 15 00:25:59 I had replaced the u-boot with Qi, so maybe I was seeing different stuff than you expected. And I could always get the NOR boot-menu. Mar 15 00:26:39 The newsgroup I'm referring to is: "gmane.comp.handhelds.openmoko.user", you can see the sort thread there. Mar 15 00:27:06 s/sort/short/ Mar 15 00:27:06 ingi meant: The newsgroup I'm referring to is: "gmane.comp.handhelds.openmoko.user", you can see the short thread there. Mar 15 00:27:45 Ah, so the problem almost certainly is the well-known Qi-fails-to-init-the-serial-port issue. Fixed in recent kernels, I understand. Mar 15 00:28:20 I even tried with fairly recent kernel/rootfs combinations, maybe you could define "recent" ! Mar 15 00:28:21 No connection between flashing the GSM firmware; the GSM would fail to work in any case. Mar 15 00:28:47 recent as in you would probably have to pull the kernel from andy's home dir directly. Mar 15 00:31:03 Either way, this is somewhat a "dark" area for me, and I was searching for some explanation/alternatives to solve my problem here, maybe I thought the "light" here would help. Mar 15 00:35:46 At least I learned today that someone here thinks that I'm a complete idiot. Mar 15 00:37:16 if it helps, i can think you're a complete idiot, too ;D Mar 15 00:40:35 I think that everyone should try to stay on topic instead of insulting each other, but then it's only my 0.05 cents... Mar 15 00:52:44 radical. ;) Mar 15 00:56:43 dude Mar 15 01:11:19 uh oh, it's juri **** ENDING LOGGING AT Sun Mar 15 02:59:56 2009