**** BEGIN LOGGING AT Wed Nov 14 03:00:00 2012 Nov 14 03:02:02 Drathir: as I understand it, your N900 boots up fine? Nov 14 03:03:38 DocScrutinizer05: i think he is in the process Nov 14 03:03:42 then you want to flash RX-51_2009SE_10.2010.13-2.VANILLA_PR_EMMC_MR0_ARM.bin first. Then remove battery, and start again, this time with RX-51_2009SE_20.2010.36-2_PR_COMBINED_MR0_ARM.bin Nov 14 03:04:42 * DocScrutinizer05 ramps up PC highlight volume a bit Nov 14 03:04:52 ping me when sth new Nov 14 03:04:55 afk Nov 14 03:05:34 DocScrutinizer05: i do it without removing the bettery(fully charged) Nov 14 03:06:04 if that works for you... Nov 14 03:10:59 DocScrutinizer05: yes because i must waiting when fw download... not have to fast connection.. Nov 14 03:13:30 i write when im start to flash but this pproppably take a few hours when downloads end... Nov 14 03:16:21 hours? poor fellow Nov 14 03:16:53 Drathir: if you dont mind my asking...in which country are you? Nov 14 03:25:32 well, see you guys later Nov 14 03:50:00 kerio 50mA variation isn't unusual, especially if you don't pay special care to starting learning state at same state of charge each time Nov 14 03:50:46 50mAh* Nov 14 03:51:46 I usually make sure I disconnect charger when it has terminated once and is now discharging with charger off Nov 14 04:06:42 maybeHere: poland Nov 14 04:06:47 fail Nov 14 04:07:09 to mad_dev may be... Nov 14 04:13:17 DocScrutinizer05: latest from nokia is RX-51_2009SE_3.2010.02-8.002_PR_COMBINED_002_ARM.bin and RX-51_2009SE_10.2010.13-2.VANILLA_PR_EMMC_MR0_ARM.bin Nov 14 04:24:28 any script to download files at risk of being lost? (Nokia proprietary) Nov 14 05:13:42 how to use pidgin for irc? Nov 14 08:20:29 Luke-Jr: I've mirrored everything Nov 14 08:22:03 teotwaki: sure, but you can't legally share that mirror Nov 14 08:22:06 so I need to make my own Nov 14 08:22:15 I can't legally share that mirror? Nov 14 08:22:19 http://wedrop.it Nov 14 08:22:22 well Nov 14 08:22:23 not if it has Nokia proprietary crap Nov 14 08:22:27 http://maemo-archive.wedrop.it Nov 14 08:22:32 there, shared. Nov 14 08:22:51 teotwaki: that has nokia proprietary crap? Nov 14 08:23:40 Luke-Jr: I'm willing to say "check it out yerself?" Nov 14 08:24:05 I probably won't know what I'm looking for until next year Nov 14 08:24:06 <.< Nov 14 08:24:34 Luke-Jr: http://talk.maemo.org/showpost.php?p=1293065&postcount=16 Nov 14 08:27:43 does anyone recommend a better n900 camera app, the one i have now crashes sometimes and takes ages to load? Nov 14 08:30:36 Psi, camera-ui? Nov 14 08:41:14 Hurrian: yeah Nov 14 08:41:31 uhh, camera-ui usually doesn't hang up Nov 14 08:41:59 BlessN900 and FCam being slow is expected due to their non hw-accelerated nature Nov 14 08:42:01 sometimes i go to click the button to take a pic and the app just closes Nov 14 09:01:33 Psi, uhh, you may want to check your system's health Nov 14 09:01:48 running processes, background processes, used memory, filesystem corruption, etc. Nov 14 09:01:58 because here on my clean N900 it just werks Nov 14 09:02:46 it works most of the time Nov 14 09:03:12 just takes 5 seconds or so to load and crashes about 1 out of every 50 times i go to take a photo Nov 14 09:04:01 didn't CSSU bring some fixes to camera-ui? Nov 14 09:04:33 yeah, it didnt used to crash Nov 14 09:55:21 phew, discharged a battery far enough to get EDV1 in ~30 minutes. Nov 14 09:55:42 WiFi + BT + FMTX + GPS with the screen on maximum brightness. Nov 14 09:56:06 I wonder how much faster I can drain the N900's battery if I had 3.5G and CPU+GPU+DSP load was 100% ;) Nov 14 10:03:29 it would probably melt Nov 14 10:42:11 Ive got a debian chroot into n900, can I make chrooted pulseaudio connect to main pulseaudio through esd? Nov 14 10:42:29 or is there any other suggested method to do that? Nov 14 10:44:40 I guess pulseaudio only allows esd as server, but can not connect to it as client Nov 14 10:46:15 perhaps through env variables you could tell chrooted pulse clients to connect to main pulse Nov 14 10:46:36 PULSE_SERVER etc, from man pulseaudio Nov 14 10:47:04 the problem is that new pulseaudio versions doesnt work with old pulseaudio on n900 Nov 14 10:47:43 wow, I see that the pulseaudio I have access to on debian chroot is 2.0, that is even worst Nov 14 10:48:09 client connecting to server should work regardless of version difference Nov 14 10:48:29 should :) Nov 14 10:48:34 mpd also has this problem Nov 14 10:48:41 (or so I think) Nov 14 10:49:26 Something to note, "The dbus machine ID in the chroot needs to be the same as that of the main host, which it wasn't for me. That caused audio apps to report "connection refused" when I tested them." Nov 14 10:51:10 Skry: that is new information to me Nov 14 10:53:45 amospalla: you might want to take a look at these: https://wiki.archlinux.org/index.php/PulseAudio/Examples#Pulseaudio_from_within_a_chroot_.28ex._32-bit_chroot_in_64-bit_install.29 and http://blog.einval.com/2010/11/12 Nov 14 10:54:18 adapt to maemo if necessary Nov 14 11:00:07 new version of battery status plugin with huge debug output in syslog Nov 14 11:00:11 on same url Nov 14 11:00:44 kerio, Hurrinan ^^^^ Nov 14 11:03:59 Pali, got the message, I'll install it later once my N900 finishes charging Nov 14 11:04:10 ok Nov 14 11:07:09 Why the hell do people send emails to the board/council in the "name of the community", without including the community mailing list? Nov 14 11:07:36 Fuckin' i-wanna-feel-special-and-be-a-saviour buffoons. Nov 14 11:10:05 who did? Nov 14 11:10:35 Pali, just curious, why don't you increase version numbers of updates? Wouldn't it be easier to spot if people have an outdated version installed? Nov 14 11:11:03 kolp, this is testing version Nov 14 11:11:12 kolp, iirc mp-fremantle-pr depends on specific versions Nov 14 11:11:22 Hurrian, no Nov 14 11:11:37 Pali: Yepp, I'm aware of that :) Nov 14 11:12:00 Pali, understood. Why not tag testing versions with "0.9" or something to make sure people upgrade, though? Nov 14 11:12:09 The debs aren't even tagged with build date. Nov 14 11:12:26 Hurrian, so library is tagged Nov 14 11:12:44 ah Nov 14 11:22:28 Pali: we should move KCSSU assessment from wiki, it could take too long for wiki to be fixed Nov 14 11:22:38 merlin1991, DocScrutinizer05 ^^^ Nov 14 11:37:36 hmm, battery 104% charging. Nov 14 11:37:50 I assume it just took last full capacity. Nov 14 11:44:50 Hurrian, Pali: well, a limit of 99% until "charged" signal received will look better :D Nov 14 11:46:21 freemangordon: If $current_now/$current_full > 1 then battery status should be "Calibrating", as iirc under these conditions VDQ is recalculated Nov 14 11:47:30 Hurrian: Pali: either battery meter uses % from bq27200, then there CAN NOT be 104%, or bq27200 has CI=1 then the cell voltage should get used to calculate %capacity via a LUT ->again no way to get 104$ Nov 14 11:47:35 104% Nov 14 11:48:31 DocScrutinizer05: "VDQ unset, charge your battery" should tell the story ;) Nov 14 11:49:08 please provide output of "grep battery_status_plugin /var/log/syslog" Nov 14 11:49:31 + write if using new or old hald-addon-bme Nov 14 11:49:41 Pali, I'm using the previous version for now, I'll use it when VDQ is calibrated Nov 14 11:50:28 Also, which hald-addon-bme is that? Mine logs to /tmp and does shutdown correctly. Nov 14 11:51:42 old = nokia, new = replacement Nov 14 11:52:29 Yes, it's the replacement one. Nov 14 11:55:55 VDQ is a flag of bq27200 (valid discharge qualified), it means that a calibration cycle would result from continuing to discharge battery. If something happens to battery that spoils the calibration, VDQ is unset Nov 14 11:56:31 you then need to charge battery completely to set it again, aka to start new learning cycle Nov 14 11:58:27 main reason for VDQ unset: short charging of battery. and current flowing *into* battery during discharge cyclke will spoil the calibration abd thus unset VDQ Nov 14 11:58:57 short charging? Nov 14 11:59:05 like 5 seconds connection? Nov 14 11:59:11 5s hooking up to charger Nov 14 11:59:19 yes Nov 14 11:59:35 does this affect the stock phone? Nov 14 11:59:50 any charging unsets VDQ until battery at 100% Nov 14 12:00:20 so just leave it on charge until it's green again to fix it? Nov 14 12:00:21 this is a effect of bq27200 chip hw Nov 14 12:01:06 in stock fremantle basically nobody cares about bq27200 calibration Nov 14 12:01:37 I see, so this won't actually affect the nokia stock version Nov 14 12:02:40 nope, it won't. The string "VDQ unset, charge your battery" is invented by me, in my bq27k-calibrate.sh Nov 14 12:03:24 iirc Nov 14 12:05:03 Well, once I get this battery all nice and calibrated, my NokiaBME-free system will run nicely. Nov 14 12:05:29 hmm, no. kerio invented that string Nov 14 12:05:42 Anyone up for tackling DSME, the most deeply-rooted binary blob in the system? Nov 14 12:06:06 Also, it seems WMO is back up. Nov 14 12:06:07 maybe Nokia will start suing for circumventing their stuff :) Nov 14 12:06:19 and spend more on legal fees than marketing Nov 14 12:06:42 and get bought out by MS in the process Nov 14 12:10:58 jon_y: the latter will happen anyway Nov 14 12:11:26 it's a good story :) Nov 14 12:12:15 Nokia is still to expensive, it needs to be really deep down and then MS will buy it for cheap $1b Nov 14 12:13:29 DocScrutinizer05: btw, what is the thumb erratas that the kernel actually workaround? Nov 14 12:14:36 oh right, something about flushing the insn pipe or cache on a jump Nov 14 12:19:06 yep Nov 14 12:19:29 branch prediction cache Nov 14 12:19:35 DocScrutinizer05: is it supposed to be done in hw? Nov 14 12:19:57 well, in proper thumb cpus Nov 14 12:20:31 hi chem|st Nov 14 12:21:45 DocScrutinizer05: in effect, any kernel code will need to do the errata workaround themselves? Nov 14 12:21:52 like in uboot Nov 14 12:22:17 no, since it only has to happen on context switches Nov 14 12:22:38 jumping between tasks? Nov 14 12:23:01 is an example of a context switch Nov 14 12:23:25 ok, user->kernel calls too? Nov 14 12:23:38 DocScrutinizer05: i think it was yours actually Nov 14 12:23:44 or your script said something *really* similar Nov 14 12:24:47 does arm have a concept of a callgate like x86? Nov 14 12:25:01 * jon_y stupid to relay his x86 experience Nov 14 12:25:25 *relate Nov 14 12:26:14 you also relayed it to us :) Nov 14 12:27:16 I was wonder how it was hooking on the context change Nov 14 12:27:21 nfc what's a callgate, and the conditions for BPcache invalidation are tricky, you need to load new program text of the other 'gender' to a location that has valid branch prediction but not for that gender. Then the jump will go to obsolete destination Nov 14 12:27:24 ShadowJK: (re: changes in bq27k calibration) that's fair - i didn't use the script to *charge*, this time - only to discharge Nov 14 12:27:54 s/you need/you must not/ Nov 14 12:27:55 DocScrutinizer05 meant: nfc what's a callgate, and the conditions for BPcache invalidation are tricky, you must not to load new program text of the other 'gender' to a location that has valid branch prediction but not for that gender. Then the jump will go to obsolete destinatio... Nov 14 12:28:12 so it didn't immediately stop charging once VDQ was set Nov 14 12:29:15 so the workaround is for some memory layout? Nov 14 12:29:41 the workaround is branch prediction buffer flush on context switches Nov 14 12:30:02 Hurrian: "tackling" dsme is kinda hard Nov 14 12:30:06 everything talks to it Nov 14 12:30:08 what happens without the workaround? Nov 14 12:30:17 see above Nov 14 12:30:33 the jump will go to obsolete destination Nov 14 12:30:55 ok I see Nov 14 12:31:34 maybe the phrasing threw me off Nov 14 12:31:51 it's complicated Nov 14 12:32:52 usually BP queue entries get invalidated by some hw in instruction decoder Nov 14 12:33:21 the cpu does something extra that needs to be workaround? Nov 14 12:33:26 but that fails when you swap arm program text with thumb program text, and both ahve a jump instruction at same locatzion Nov 14 12:34:22 so your thumb code jumps to the destination that been valid in the arm code context Nov 14 12:34:31 aiui Nov 14 12:34:36 ok, may jump to the wrong area due to insn size difference? Nov 14 12:35:15 no, because of instruction decoder doesn't invalidate the branch prediction entry Nov 14 12:35:53 yup, still over my head :) Nov 14 12:36:03 so basically, thumb->thumb jump is fine Nov 14 12:36:14 it *seems* to me the instruction decoder doesn't detect there's a arm branch prediction when it decodes a thumb branch instruction, or sth like that Nov 14 12:36:26 but thumb->arm jumps and vice versa needs flushing? Nov 14 12:36:49 it's program text swaps that need the flushing Nov 14 12:37:29 it's really not so simple to explain Nov 14 12:37:40 I'm not even sure I completely understood it Nov 14 12:38:31 you have memory with instructions, and for each instruction there's possibly an allocated branch prediction in a separate buffer Nov 14 12:38:52 definitely a RISC thing Nov 14 12:39:20 now if the jump instruction on a particular memory addr changes from a arm to a thumb instruction or vice versa, the branch prediction in buffer doesn't get cleared Nov 14 12:40:05 ok, so far so good, I understand that at least Nov 14 12:40:21 while *usually* the CPU hardware is supposed to clear those invalid branch predictions Nov 14 12:42:14 it has a 50% chance to take the wrong path and execute stuff that is still in the pipeline? Nov 14 12:43:37 since context switches alter the memory addr translation table, those changes of program text at a certain memory addr are most likely to happen during those context switches, aiui Nov 14 12:44:42 ok, its due to the tlb change Nov 14 12:44:46 but yes, self modifying code as well would need special care to invalidate the BPB Nov 14 12:45:05 normally done by hw, but needs some help on the n900 Nov 14 12:45:08 kerio, indeed, even friggin tklock talks to it Nov 14 12:47:38 jon_y: the basic design of the workaround been by TI or even ARM (dunno). freemangordon found a way to make those patches work on N900, so we consider them as safe as the original patchset by TI Nov 14 12:48:17 I dunno what TI found out about when to flush BPB Nov 14 12:48:25 TI is good enough :) Nov 14 12:48:37 context switches are the buzzword I recall Nov 14 12:49:02 there might be other occasions that also need that buffer flushing Nov 14 12:56:14 TI is good enough Nov 14 12:56:28 DocScrutinizer05: and that's it >:C Nov 14 12:56:52 hm? Nov 14 12:57:19 we have the kernel patch from TI, thumb works with no unexplicable segfaults Nov 14 12:58:42 seems to Nov 14 12:59:42 we're not doing math, we're doing science! Nov 14 12:59:53 as elaborated above, there might be several occasions where such a patch needs to flush the buffer "manually". Nobody verified that TI dealt with all of them Nov 14 13:01:18 I could figure that e.g. self-modifying code (something *strongly* deprecated, in general) would go "beyond the radar" of such TI patches Nov 14 13:02:08 you mean like... a JIT compiler? Nov 14 13:02:35 "self-modifying code" is a pretty clear term Nov 14 13:03:02 a JIT compiler doesn't modify *own* code Nov 14 13:11:32 ~wiki self-modifying code Nov 14 13:11:50 At http://en.wikipedia.org/wiki/Self-modifying_code (URL), Wikipedia explains: "{{Refimprove|date=April 2009}} In computer science, 'self-modifying code' is code that alters its own instructions while it is executing - usually to reduce the instruction path length and improve performance or simply to reduce otherwise repetitively similar code, thus simplifying maintenance. Self modification is an alternative to the method of 'flag setting' and ... Nov 14 13:19:02 jon_y: in addition to what DocScrutinizer05 said - because os a staled prediction in BTB, the correct code is executed in incorrect CPU state Nov 14 13:19:22 i.e. thumb code is executed as ARM nad vice versa Nov 14 13:19:30 *and Nov 14 13:19:38 ooh, yeah, I forgot Nov 14 13:20:02 so the usual result when errata hits is SIGILL Nov 14 13:21:03 since the branch target from BTB also determines the CPU state, in case of obsolete BTB value getting used, the CPU also switches to wrong state in addition to jumping to wrong addr Nov 14 13:21:29 no, jump is to the correct address, just the state is wring Nov 14 13:21:39 *wrong Nov 14 13:21:39 (pair target addr means ARM mode, impair means thumb mode) Nov 14 13:22:09 yep Nov 14 13:22:54 but instead of taking the mode from the instruction itself, CPU takes the mode from BTB, which is invalid Nov 14 13:22:56 the target addr however is that of BTB and has no relation to the new correct addr as derived from the instruction the stale BTB entry is linked to Nov 14 13:26:54 stage a): 0x50: (arm)jump 0x100; BTB:target100;; stage b): instruction swapped: 0x50: (thumb)jump 0x301; BTB(stale!!):target100;; now in c) the thumb code at 0x50 jumps to 0x100 though it *should jump to 0x301. Target addr pair (0x100) means CPU switching to ARM mode Nov 14 13:27:51 that's how I understand it Nov 14 13:29:04 odds are that 1. the target addr is incorrect (obviously) and 2. that the code at target addr also swapped from ARM to thumb code Nov 14 13:30:16 flushing BTB on swapping code at 0x50 will cure the problem Nov 14 13:31:06 usually CPU does that automatically, but not when swapping between ARM and thumb and both have a jump instruction at same memory addr Nov 14 13:32:50 What is the reason headset remote controls do not work? Nov 14 13:33:06 I thought they worked on the N9 and it had the same hardware? Nov 14 13:33:17 hmm? Nov 14 13:33:25 wired multibutton headset? Nov 14 13:33:30 yeah Nov 14 13:33:38 never worked on N900 Nov 14 13:33:46 see my post on tmo regarding that Nov 14 13:34:32 vi_: those aren't standard Nov 14 13:35:02 Pali: sorry, what did you update since yesterday? Nov 14 13:35:23 kerio: I tried to restore an old backup. Nov 14 13:35:27 kerio: bootloop, Nov 14 13:35:30 kerio, what? Nov 14 13:35:36 kerio: Have to start from scratch again. Nov 14 13:38:07 vi_: http://talk.maemo.org/showthread.php?p=799047#post799047 Nov 14 13:40:42 vi_: it's a one-wire serial protocol that's used, with some 2.4kBaud iirc Nov 14 13:41:31 * vi_ reading Nov 14 13:41:50 I.E. both the device and the headset pull mic line to gnd for short times, let's say 0.1ms for a '0' and 0.2ms for a '1' Nov 14 13:43:36 device sends a command, headset answers. OR headset sends an attention code to make device send the command to query what's the event to report. sth along that line Nov 14 13:44:26 it's much similar to old cheapernet ethernet Nov 14 13:44:49 only one sender can use the media channel at any given time Nov 14 13:45:40 Do you think the n900 hardware is capable of this? Nov 14 13:45:49 it might be capable Nov 14 13:46:01 DocScrutinizer05: Do you have a link to the n9 schematics? Nov 14 13:46:13 but all the real time serial protocol needs to get handled in kernel then, in sofzware Nov 14 13:46:35 see hidden/docs on my server Nov 14 13:46:44 iirc it's there Nov 14 13:46:48 I cannot, it is hidden! Nov 14 13:46:56 docs isn't Nov 14 13:47:56 cool! Nov 14 13:48:09 Prepare for server rape! Nov 14 13:48:13 lol Nov 14 13:48:15 yw Nov 14 13:48:59 just never post the URL *anywhere* Nov 14 13:49:05 ofc Nov 14 13:49:10 I will respect your privacy Nov 14 13:50:02 as soon as spiders rape my server, I'll take it down Nov 14 13:50:25 Is there a guide to set up a server like that? Nov 14 13:50:40 pretty simple, change the dir perms Nov 14 13:51:23 I have a lot of stuff for motorcycles I want to host just like this. Nov 14 13:52:19 chmod go-r hidden Nov 14 13:52:33 aka 711 Nov 14 13:53:17 DocScrutinizer05: did Drathir reflash?? Nov 14 13:53:26 eXecute means you may access the content but not read the dir content aka ls Nov 14 13:53:29 you can also change your .htaccess Nov 14 13:53:45 -Indexes Nov 14 13:53:47 mad_dev: I went for a nap, no idea Nov 14 13:54:04 or echo > index.html Nov 14 13:55:09 mad_dev: last I heard was "this will take hours to download" Nov 14 13:55:30 DocScrutinizer05: same here Nov 14 13:55:42 I hope (s)he will come back once download finishes Nov 14 13:56:38 heh, looks like Poland is into reflashing these days Nov 14 13:57:09 yeah, I also thought he's from Poland Nov 14 13:57:30 his hostname suggested that Nov 14 13:58:15 s51.linuxpl.com Nov 14 14:01:48 wtf is reno chip? Nov 14 14:02:17 why they do not have chip numbers on this schematic? Nov 14 14:04:23 because Nov 14 14:04:27 vi_: it's the one with all the casinos Nov 14 14:04:34 the great questions of 21st century Nov 14 14:04:41 reno is one of the phone chip set IIRC Nov 14 14:04:45 «the biggest little chip» Nov 14 14:04:48 etc. Nov 14 14:04:54 yep, like gaia for bb5 Nov 14 14:05:10 iirc Nov 14 14:05:39 reno for rapuyama Nov 14 14:06:43 well it fulfils the same role as 'stereo audio codec' on the n900. Nov 14 14:07:04 s/yep, like gaia for bb5/yep, like a gaia chip made for bb5/ Nov 14 14:07:39 And it appears to be where the mic bias is connected. Nov 14 14:07:48 or, less lyrical, it's rapuyama bb5's companion chip Nov 14 14:08:21 in N95? Nov 14 14:09:01 n900 Nov 14 14:09:33 N900 micbias connected to APE domain, not cmt domain Nov 14 14:10:04 I.E. probably to gaia Nov 14 14:11:02 yes Nov 14 14:11:09 ~APE Nov 14 14:11:09 APE Portable Environment for C++ Threads, Sockets, etc.. URL: http://www.voxilla.org/projects/projape.html Nov 14 14:11:19 ~cmt Nov 14 14:14:14 Pali: you said you updated the battery applet - did you change anything else? Nov 14 14:14:37 kerio, yes read log :-) added huge debug output to syslog Nov 14 14:14:55 vi_: actually connected to MICBIAS3 of aic34 audio codec Nov 14 14:15:28 use "grep battery_status_plugin /var/log/syslog" for uploading debug log Nov 14 14:16:13 DocScrutinizer05, what is APE? what CMT? and what APE-algo and CMT-algo? Nov 14 14:16:26 vi_: *sensing* however is via ECI_AD of GAIA Nov 14 14:16:37 and what is cmt-mcusw? Nov 14 14:16:52 Application Processor Environment Nov 14 14:17:07 Cell Modem and Telephone (or sth) Nov 14 14:17:19 the algos are for flashing Nov 14 14:17:59 Cell Mode and Telephone Micro Controller SoftWare Nov 14 14:18:07 Modem* Nov 14 14:18:09 so what software deals with the 'audio codec' chip? Is it a kernel module? Nov 14 14:18:23 err, yes Nov 14 14:18:26 basically Nov 14 14:18:33 ALSA I guess Nov 14 14:20:09 An ALSA module? Nov 14 14:20:56 an ALSA soundcard Nov 14 14:21:07 Do you know if it is closed? Nov 14 14:21:16 aaah. who the fuck am I kidding. Nov 14 14:21:18 shouldn't Nov 14 14:24:02 vi_: if you're interested in multibutton serial protocol, check ECI_AD and the two voltage comparators arounf N4007 and N4008 Nov 14 14:24:21 as well as muxer N4022 Nov 14 14:25:00 ok Nov 14 14:26:30 google for ECI, I think that's what the protocol is called Nov 14 14:27:22 how to used dbus command connect gprs? i read the phone_control wiki page......but when execute "gconftool -R /system/osso/connectivity/IAP" ,output:/system/osso/connectivity/IAP/@232@@129@@148@@233@@128@@154@GPRS@232@@191@@158@@230@@142@@165@@228@@186@@146@@232@@129@@148@@231@@189@@145@: Nov 14 14:27:50 https://lkml.org/lkml/2010/12/30/84 Nov 14 14:28:19 DocScrutinizer05: Thankyou Nov 14 14:28:44 what's is my GRPS's IAP Nov 14 14:29:16 DocScrutinizer05: http://mailman.alsa-project.org/pipermail/alsa-devel/2011-February/036544.html Nov 14 14:31:47 vi_: beware! there's much about a special MCU that controls the ECI protocol. It's _not_ about emulating ECI in kernel space via bitbang aiui Nov 14 14:32:19 took me two days to finally recognize it's a dead end Nov 14 14:32:55 dunno to which website and info it applies, I did that investigations like 2 yeras ago, maybe 3 Nov 14 14:33:35 well, feb2011 sounds fairly new Nov 14 14:33:56 DocScrutinizer05: Well those patches are only 1.5ish year old. Nov 14 14:34:27 there's been a Nokia driver for a ECI controller chip, and that was basically moot info, except for some info about protocol itself Nov 14 14:35:05 brb, got to earn a living :(. Nov 14 14:35:16 got to cure my flu Nov 14 14:35:19 o/ Nov 14 14:35:30 DocScrutinizer05: Coffee and painkillers. Nov 14 14:35:54 good plan, will prepare/take another couple of each Nov 14 14:41:16 Wilson: "and you're feeling like that every day. It sucks to be you" House:"at least I don't have cancer" Nov 14 14:42:13 DocScrutinizer05: old Nov 14 14:43:49 also best season finale ever Nov 14 14:49:53 8/19, seems there are 20..22 still Nov 14 14:50:33 and aiui that's definitive end then Nov 14 14:52:48 yep Nov 14 15:13:09 kerio: What is a good windows IM client with irc support? Nov 14 15:13:16 What is a good windows IM client with irc support? Nov 14 15:13:53 vi_: fucked if i know Nov 14 15:14:15 vi_: kopete? pidgin? Nov 14 15:14:24 I tried the latter Nov 14 15:14:38 the former *should* work Nov 14 15:16:41 vi_: http://irssi.org/download Nov 14 15:18:15 ^ Nov 14 15:19:14 mad_dev: I am using irssi right now. Nov 14 15:19:22 Irssi is king of IRC. Nov 14 15:19:30 The client of the future if you will. Nov 14 15:19:43 vi_: without a doubt Nov 14 15:19:44 But I need other protocols. Nov 14 15:20:05 you can always try this bitlbee thing, or whatsitsname Nov 14 15:20:13 try chatzilla Nov 14 15:20:28 irssi has xmpp plugin also Nov 14 15:20:28 I'm yet to see something that supports more than one protocol and doesn't suck Nov 14 15:20:42 tadzik: n900? Nov 14 15:20:50 hehe Nov 14 15:21:03 it supports absolute basics of xmpp Nov 14 15:21:12 try sending a specific status to specific contact Nov 14 15:21:46 recently I needed to start using skype, and it appears to work quite nice with irssi (via bitlbee and its skype-plugin) Nov 14 15:21:51 tadzik: I am not *that* into im. Nov 14 15:21:53 also, those "user is typing" notifications aren't there. Yes, that's a XEP, but still Nov 14 15:22:01 skype? Nov 14 15:22:26 Skype makes me want to puke a solid shaft of barf vertical so that it may fountain back into my own face. Nov 14 15:22:30 have you changed your email today? ;) Nov 14 15:22:42 lol Nov 14 15:23:12 a standard solution would be preferably by far, but apparently none of them are productized as well as skype Nov 14 15:23:29 flux I just discovered google talk. Nov 14 15:23:41 that's xmpp Nov 14 15:23:49 fortunately Nov 14 15:24:02 I don't really care what it is. I just want it to work. Nov 14 15:24:10 Skype on the n900 is a fucking joke. Nov 14 15:24:12 can you hook your own servers to google talk? Nov 14 15:24:26 the audio is *literally* worthless. Nov 14 15:24:30 only im works. Nov 14 15:24:35 flux: gtalk federates, yes Nov 14 15:25:18 unlike the facebook shit Nov 14 15:25:22 Although a true p2p voice/im would be sweet. Nov 14 15:26:07 what does facetime use? Nov 14 15:26:07 i want a bluetooth PAN module for icd Nov 14 15:26:18 mad_dev: xmpp, just closed to other servers Nov 14 15:26:25 kerio: What a coincidence! So do I! Nov 14 15:26:29 vi_: write it Nov 14 15:26:32 not sure if they introduce anything non-standard Nov 14 15:26:46 kerio: No, u. Nov 14 15:26:51 vi_: luf has a working everything Nov 14 15:26:59 we just need the ICD plugin Nov 14 15:27:06 what is luf? Nov 14 15:27:16 baby don't hur me Nov 14 15:27:34 luf is the dude who built bluez 0.99 for the n900 Nov 14 15:27:53 is his work documented anywhere? Nov 14 15:28:49 vi_: https://gitorious.org/community-ssu/bluez/commits/master Nov 14 15:28:52 So that I may read about it? Nov 14 15:29:15 now doing a NAP server is as easy as configuring it in /etc/bluetooth/network.conf Nov 14 15:29:41 wtf is a nap server? Nov 14 15:29:50 Network Access Point Nov 14 15:29:56 aaah. Nov 14 15:29:59