**** BEGIN LOGGING AT Sat May 14 02:59:59 2011 May 14 07:26:57 GNUtoo|laptop: you missed me by some minute :/ May 14 07:51:21 moin May 14 08:02:04 moin JaMa|Off May 14 08:18:15 someone with 2.6.34 running on om-gta? can you show ls /sys/devices/platform/leds_pwm/leds/ ? May 14 08:19:49 JaMa: let me boot May 14 08:21:42 root@om-gta02 ~ # cat /proc/version May 14 08:21:43 Linux version 2.6.34.8 (bitbake@localhost) (gcc version 4.5.3 20110214 (prerelease) (GCC) ) #1 Mon Feb 21 06:53:23 CET 2011 May 14 08:22:00 root@om-gta02 ~ # ls /sys/devices/platform/leds_pwm/leds/ May 14 08:22:00 ls: cannot access /sys/devices/platform/leds_pwm/leds/: No such file or directory May 14 08:22:30 root@om-gta02 ~ # ls /sys/devices/platform/leds_pwm/ May 14 08:22:30 modalias power subsystem uevent May 14 08:22:48 and where does fsodeviced find them? May 14 08:23:17 because someone reported that they are not working with 2.6.37, but fsodeviced find them (maybe with different names again) May 14 08:23:18 it probably doesn't May 14 08:23:29 gta02::vibrator gta02:blue:power gta02:orange:power May 14 08:23:42 2011-05-13T16:21:24.451028Z [INFO] Kernel26Led : ^^^ supports the following triggers: '[none] rfkill0 rfkill1 nand-disk battery-charging-or-full battery-charging battery-full adapter-online usb-online ac-online mmc0 mmc1 timer heartbeat backlight gpio rfkill2' May 14 08:23:48 2011-05-13T16:21:24.451377Z [INFO] Kernel26Led : Created May 14 08:24:00 this is beter: May 14 08:24:01 ls /sys/class/leds/ May 14 08:24:02 gta02::vibrator gta02:blue:power gta02:orange:power gta02:red:aux May 14 08:24:41 2000.01.01 01:01:52.671 oeventsd.fso_triggers INFO Receive PowerStatus, status = charging May 14 08:24:44 2000.01.01 01:01:52.683 oeventsd ERROR DBus call returned an error May 14 08:24:50 no such class here May 14 08:24:59 nothing about leds in fsodeviced.log May 14 08:25:13 weird but we had leds working with 2.6.34 right? May 14 08:25:33 and no leds exported to dbus May 14 08:25:41 no idea May 14 08:26:42 /org/freesmartphone/Device/LED/gta02_blue_power May 14 08:26:42 /org/freesmartphone/Device/LED/gta02_red_aux May 14 08:26:42 /org/freesmartphone/Device/LED/gta02_orange_power May 14 08:26:49 modprobe leds-s3c24xx May 14 08:26:57 gives me an empty /sys/class/leds May 14 08:27:18 strange that it looks right with .37 and wrong with .34 :) May 14 08:27:38 probably another module missing May 14 08:27:39 hmm May 14 08:28:25 * mrmoku rebooting after depmod May 14 08:28:41 did you opkg upgrade from tests/shr-u already? May 14 08:29:20 that could explain it as it removes modules which exists in .37 from /lib/modules/*.34* May 14 08:32:31 hmm May 14 08:32:33 possible May 14 08:32:53 root@om-gta02 ~ # ls /lib/modules May 14 08:32:53 2.6.34.8 2.6.39-rc5 May 14 08:33:01 I also tried out 2.6.39 it seems May 14 08:33:13 so maybe that's just something personally broken :P May 14 08:34:53 JaMa: even worse... May 14 08:34:55 src/gz jama-all http://jama.dyndns-home.com/org.openembedded.shr/all May 14 08:34:56 ah.. that's the image you downloaded from me? May 14 08:34:57 ;) May 14 08:35:03 yeah May 14 08:35:22 so it's completely without autoloads for .34 only modules :) May 14 08:35:29 because you had to install them manually May 14 08:35:52 which image do you want me to test? May 14 08:36:13 latest synced unstable ? May 14 08:36:33 nvm thenmodule_autoload_leds-s3c24xx = "leds-s3c24xx" May 14 08:36:33 module_autoload_leds-pwm = "leds-pwm" May 14 08:36:34 module_autoload_leds-gpio = "leds-gpio" May 14 08:36:34 http://build.shr-project.org/shr-unstable/images/om-gta02/shr-lite-eglibc-ipk--20110510-om-gta02.rootfs.tar.gz May 14 08:36:54 mickey|bbl: hi, i added first anotations for qdbusxml2cpp: https://github.com/radekp/fso-specs/commit/57e587f8f6e64571330559c9a06d0b59c5502834 May 14 08:37:11 mrmoku: for leds please try modprobe module_autoload_leds-s3c24xx = "leds-s3c24xx" May 14 08:37:14 module_autoload_leds-pwm = "leds-pwm" May 14 08:37:14 sorry May 14 08:37:16 module_autoload_leds-gpio = "leds-gpio" May 14 08:37:21 modprobe leds-s3c24xx leds-gpio leds-pwm May 14 08:37:22 root@om-gta02 ~ # ls /sys/class/leds/ May 14 08:37:23 gta02::vibrator gta02:blue:power gta02:orange:power gta02:red:aux May 14 08:38:24 so it's the same as with .37, I'll ask Neil Jerram for fsodeviced.log then May 14 08:38:36 ok May 14 08:38:37 because here it seems to be working (tested with dbus calls) May 14 08:38:53 maybe it's missing some module autoload there too? May 14 08:39:08 and if you want to test something then tests/shr-unstable latest image is for you May 14 08:39:18 ok, will do :) May 14 08:39:20 there is new kernel as well as newer xorg May 14 08:39:25 good May 14 08:39:34 * mrmoku loves experimental stuff May 14 08:39:42 and after index there will be newer qt if you're using something qt-based May 14 08:40:22 no, nothing qt (yet) May 14 08:40:26 btw GPRS is still known broken on gta02, right? May 14 08:40:30 yup May 14 08:40:52 fso and ppp don't play nice together May 14 08:41:53 and can be fixed only by gsm_n? May 14 08:42:07 or is there plan for some other fix? May 14 08:43:18 gsm_n + internal ppp of fso would be the good fix May 14 08:43:33 probably the fast fix is to look into what is happening exactly May 14 08:48:13 ok I'll include this snippet to answer to Neil May 14 08:48:22 fine May 14 08:49:26 * mrmoku booting May 14 08:50:34 * JaMa rebooting May 14 08:53:52 root@om-gta02 ~ # ls /sys/class/leds/ May 14 08:53:52 gta02::vibrator gta02:blue:power gta02:orange:power gta02:red:aux May 14 08:54:21 root@om-gta02 ~ # cat /proc/version May 14 08:54:21 Linux version 2.6.37.6 (bitbake@opmbuild) (gcc version 4.5.3 20110311 (prerelease) (GCC) ) #1 Fri May 13 08:06:01 CEST 2011 May 14 08:54:26 this is that new image with .. :) May 14 08:54:27 so indeed that is unchanged May 14 08:54:30 good May 14 08:54:54 * JaMa plans to send himself sms to see red aux blinking :) May 14 08:56:59 weird http://paste.pocoo.org/show/388522/ May 14 08:59:22 * JaMa surprised that UnreadMessages gets 0 so quick May 14 08:59:32 but it does wibrate and blink for few secs May 14 09:00:13 ah.. those messages are imediatelly shown on screen and probably then marked as read May 14 09:00:29 even with screen locked :) May 14 09:12:30 time to go out for a while... bbl May 14 09:38:05 hi mrmoku May 14 10:00:24 hi GNUtoo|laptop May 14 10:01:09 I'll check in some acked oe patches and ask you vala question on pointers arithmetics May 14 10:01:56 GNUtoo|laptop: I need half an hour for gardening work first though May 14 10:02:11 ok May 14 10:02:24 bb after that May 14 10:02:25 so you'll be back in half an hour? May 14 10:02:41 the question was: May 14 10:02:49 yeah, read it in the backlog May 14 10:02:52 void * pointer = foo; May 14 10:02:56 no pointer arithmetics May 14 10:03:00 foo += something; May 14 10:03:15 so how do I advance in the alsa buffer? May 14 10:03:34 hmm May 14 10:03:45 maybe having an index and doing foo[index] works? May 14 10:03:50 not sure though May 14 10:03:54 have to find out myself May 14 10:03:54 hmmm May 14 10:03:57 ok May 14 10:04:07 so think about it while doing gardening May 14 10:04:11 and come back with the idea May 14 10:04:48 #vala gave me some answer May 14 10:04:53 with cast to pointer May 14 10:04:55 for instance May 14 10:05:08 (uint8 *) foo = foo + something; May 14 10:05:18 but that made the program segfault May 14 10:06:53 ok May 14 10:07:21 maybe playya_ or mickey|bbl know the correct solution May 14 10:07:31 otherwise I will try to find out after the garden May 14 10:07:32 bbiab May 14 10:27:26 GNUtoo|laptop: can you pastebin your diff? May 14 10:27:44 maybe do the gardenning now May 14 10:27:53 since I need to eat and checking oe patches May 14 10:28:05 I've sent something like 12 patches yesterday May 14 10:29:56 GNUtoo|laptop: ok, gardening is delayed after siesta though May 14 10:30:26 ah ok May 14 10:51:39 * mrmoku lunch May 14 11:05:37 GNUtoo|laptop, could you paste youe source? May 14 11:08:32 yes May 14 11:08:33 ok May 14 11:08:41 I'll do it now that I've pushed the oe patches May 14 11:08:47 I can modify patchwork later May 14 11:09:59 http://www.pastie.org/private/ohh4eidu2v5kdalbadm2g May 14 11:11:03 sorry for not publishing before May 14 11:17:23 hmm. try to use ptr = &ptr[160]; May 14 11:18:05 GNUtoo|laptop, ^ May 14 11:18:34 ok May 14 11:21:45 playya_, you mean : May 14 11:22:02 buffer = &ptr[160] May 14 11:22:08 or May 14 11:22:11 better: May 14 11:22:22 buffer = &ptr[frames] May 14 11:22:41 but that won't work May 14 11:22:44 better doing that: May 14 11:22:54 buffer = &ptr[buffersize] May 14 11:22:56 and then May 14 11:23:09 buffer += frames May 14 11:23:14 is that ok? May 14 11:27:59 alsa.vala:514.3-514.15: error: Pointer arithmetic not supported for `void*' May 14 11:28:07 playya_, ^^^ May 14 11:28:18 ahhh May 14 11:28:25 I understand what you told me now May 14 11:40:48 playya_, it segfault May 14 11:41:55 ah not this time May 14 11:42:02 this time it buffer underuns May 14 11:43:33 #0 0x405b0844 in memcpy () from /lib/libc.so.6 May 14 11:43:33 #1 0x406ef9a0 in snd_pcm_area_copy () from /usr/lib/libasound.so.2 May 14 11:43:33 #2 0x00000000 in ?? () May 14 11:44:51 playya_, are you there? May 14 11:45:32 GNUtoo|laptop: be damn sure that &ptr[160] is still valid May 14 11:45:47 that is to say? May 14 11:46:14 I'll try something May 14 11:46:23 you mean that it doesn't cross the boundaries May 14 11:47:14 I think it does May 14 11:47:36 yup May 14 11:48:02 I'll try the hardcoded version May 14 11:48:47 ok it works better but still xruns May 14 11:48:58 but but.... May 14 11:49:09 maybe the xruns are not alsa xruns May 14 11:49:20 CMTSPEECH: nokiamodem_backend: DL xrun, reset hw/appl at 644 May 14 11:49:20 CMTSPEECH: nokiamodem_backend: DL buffer overrun (mmaphw 649, hw 646, appl 645, slot 1, count 3, hwdelay 4). May 14 11:49:20 CMTSPEECH: nokiamodem_backend: late appl wakeup (hw 646, appl 645, delay 4, count 3, boundary 768). May 14 11:49:52 maybe I'm not taking into account the cmtspeech requirements May 14 11:50:48 I did that btw: May 14 11:50:50 device.writei(buffer,160); May 14 11:50:51 buffer = &buffer[160]; May 14 11:50:51 device.writei(buffer,160); May 14 11:50:51 return 320; May 14 11:54:47 moo May 14 11:54:55 hi May 14 11:56:17 are you running cmtspeech on realtime scheduling? May 14 11:56:28 what's that? May 14 11:56:43 I'm not running alsa in realtime scheduling tough May 14 11:56:49 since I used the higher level API May 14 11:56:57 I didn't have to choose the scheduling May 14 11:57:03 mhm May 14 11:58:02 I have to admit I'm not entirely sure what's realtime schedule, it's a special "niceness" you set your process to May 14 11:58:10 I think so May 14 11:58:15 DocScrutinizer: that might cause the robot voice problem, right? May 14 11:58:20 yes May 14 11:58:24 not like nice -20 May 14 11:58:29 yes May 14 11:58:50 the more I split the buffer, the more robot the voice is May 14 11:59:01 sure May 14 11:59:31 as your app needs even more wakeups to service the buffer, but it will get those only when a timeslice exires May 14 12:00:46 I think realtime prio makes the process grab cpu as soon as there's a signal, no matter what's the timeslice frames May 14 12:01:36 NB that's mere handwaving May 14 12:01:56 actually I have NFC May 14 12:03:19 anyway, your buffer should be large enough to work for longer than several timeslices of your scheduler - usually that are 5..100ms/slice May 14 12:03:39 depending on compile time config May 14 12:03:47 ok May 14 12:04:06 though some systems have /proc or /sys nodes to tweak the scheduler May 14 12:05:56 it's a tradeoff between response delay to signals (a task is in state running, but waiting for another task to finish the timeslice it owns) and overhead for frequent taskswitching which tends to impact on overall system performance May 14 12:06:56 realtime prio is a way to ship around that aiui May 14 12:09:39 I think your robot voice is a precise projection of the scheduler's timeslices May 14 12:10:52 try using a buffer consisting of 4 segments each worth 50ms of audio May 14 12:11:04 ok May 14 12:11:09 to start with May 14 12:11:47 how many bytes is 50ms? May 14 12:11:53 SQ May 14 12:12:05 duh, depends on your samplerate and format May 14 12:12:12 S16_LE mono May 14 12:12:18 8000hz May 14 12:12:30 s/SQ// May 14 12:12:42 so that's 2 bytes * 8000 / second May 14 12:12:48 * 50ms May 14 12:12:59 ok May 14 12:13:12 right May 14 12:13:29 16bit * 8000 in one second May 14 12:13:55 though afaik alsa always uses unit frame for buffer calculation which is depending on format and 1 frame = 2 bytes for S16_LE May 14 12:14:32 mono May 14 12:17:25 ok May 14 12:17:40 your buffer also needs to sync the both rates (cmt and audiocard) as they both introduce their own timebase to the system. So if your CMT is a lil bit slow, then your audio buffer that you propably started when it got 50% filled, will slowly deplete May 14 12:19:36 if you detect the audio buffer depleting, you may counteract that by injecting a frame doublette every 1000 frames, then going down to 1 / every 999 and so on, until the doublettes injected compensate for the cmt clock being a bit slow May 14 12:20:44 same for cmt being fast (aka audio being slow): you drop one frame every 1000, then increasing till you adjusted to sync and audiobuffer is again at an average of 50% filling May 14 12:21:24 it's a slightly complex process and not really easy to explain verbally in a few sentences, but I think you got the picture May 14 12:22:21 basically it's like buffer underruns handling? May 14 12:22:30 but you don't need that for first PoC May 14 12:22:31 you drop or inject to keep up May 14 12:22:39 yes May 14 12:23:31 but you don't drop or replay a whole buffer segment of maybe 50ms, but single samples aka frames May 14 12:23:51 I've a 320 frame buffer by cmt May 14 12:23:54 *frames May 14 12:24:00 I need 400 buffers for 50ms May 14 12:24:01 mhm May 14 12:24:13 so I create a new buffer May 14 12:24:17 and memcpy it May 14 12:24:20 inside May 14 12:24:21 sure May 14 12:24:54 apt!!! resurrect! May 14 12:25:05 grrrr May 14 12:26:16 err there's sth wrong with your math May 14 12:26:36 8000 / 320 = 25 buffers/s May 14 12:28:16 le me look May 14 12:28:33 so it's probably absolutely sane to have an audio buffer with 4 segments a 320 frames, for a total of 160ms May 14 12:28:49 maybe use 5 segments ;-) May 14 12:29:04 ok May 14 12:30:09 start playback when 3/5 of buffer are filled with data May 14 12:30:16 I'm lost between buffers and bytes May 14 12:30:27 let me re-look what is buffer and what is bytes May 14 12:30:43 s/buffers/frames/ May 14 12:30:53 1 frame is one sample May 14 12:31:19 which is 2 bytes May 14 12:31:26 but I meant May 14 12:31:26 one sample is sizeof(datatype) * channels May 14 12:31:58 one sec May 14 12:32:01 I'm doing that: May 14 12:32:14 return device.writei(buffer,buffersize); May 14 12:32:19 that's plain wrong May 14 12:32:23 let me verify May 14 12:32:51 it should be buffersize / 2 May 14 12:33:05 anyway I'll use your advises May 14 12:33:07 yeah May 14 12:33:25 what I meant is : May 14 12:33:35 I've lost track of what is frames and what is bytes in the code May 14 12:33:50 as alsa (writei) always uses a unit of frames, this should be buffersize/sizeof(frame) May 14 12:34:36 btw this pretty much explains your robit voice :-D May 14 12:34:45 robot* May 14 12:37:15 some hints: May 14 12:37:24 http://www.alsa-project.org/alsa-doc/alsa-lib/group___p_c_m___dump.html#a0 May 14 12:37:46 for the logging printout ala aplay -v May 14 12:38:05 thanks a lot May 14 12:39:04 freesmartphone.org: 03mickey 07cornucopia * r701348c1c88e 10/fsoaudiod/ (src/lib/alsa.vala vapi/alsa-ext.vapi): fsoaudio: alsa vapi: add abstraction for write and read, use uint8 instead of void* to ease usage May 14 12:40:21 some example code: http://www.alsa-project.org/alsa-doc/alsa-lib/_2test_2pcm_8c-example.html#a89 May 14 12:42:09 some more: http://www.alsa-project.org/alsa-doc/alsa-lib/_2test_2latency_8c-example.html#a24 May 14 12:44:10 freesmartphone.org: 03mickey 07aurora * r823e305b4a21 10/aurora/aurora/main.qml: aurora: fix typo May 14 12:46:36 quite nice: http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html May 14 12:47:52 writei() def: http://www.alsa-project.org/alsa-doc/alsa-lib/group___p_c_m.html#gf13067c0ebde29118ca05af76e5b17a9 May 14 12:48:24 WOW May 14 12:48:29 there is a small krrr May 14 12:48:31 but that's all May 14 12:48:58 the krrr is not present with aplay May 14 12:49:17 the diff is: May 14 12:49:20 return device.writei(buffer,buffersize); May 14 12:49:45 pcm.open("plug:dmix"); May 14 12:52:06 btw s/segments/periods/ in all my above posts May 14 12:52:39 ok May 14 13:00:19 >> After application transfers the data in the memory areas, then it must be acknowledged the end of transfer via snd_pcm_mmap_commit() function to allow the ALSA library update the pointers to ring buffer. This kind of communication is also called "zero-copy", because the device does not require to copy the samples from application to another place in system memory.<< May 14 13:01:04 that's from http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html regarding using mmap rather than writei May 14 13:01:25 I don't use mmap yet May 14 13:01:30 I use plain writei May 14 13:01:45 and currently I've no more sound May 14 13:01:47 just felt like pointing at it May 14 13:01:49 I don't know why May 14 13:01:51 ok May 14 13:02:46 first you should probably implement alsa_debug May 14 13:02:59 will make your life a lot easier May 14 13:03:06 the aplay -v thing May 14 13:03:24 yes indeed May 14 13:03:31 I'll go to do some sport very soon tough May 14 13:09:09 buffersize / 2 makes it silent May 14 13:09:11 strange May 14 13:17:32 *yawn* May 14 13:18:15 ~seen radekp May 14 13:18:33 hmm, where's apt May 14 13:19:08 mickeyl, I'll go doing some sport and try to reproduce what I did after May 14 13:19:21 ok, cool. if you ahve something to review, i'm all ears May 14 13:19:50 there's no 3g at the sport May 14 13:19:56 so bb in a bit May 14 13:20:00 okies, have fun May 14 13:20:21 it's not very fun May 14 13:20:24 but it's necessary May 14 13:20:29 oh so true May 14 13:20:46 bbl May 14 13:32:39 mickeyl: apt is on holiday May 14 13:32:49 i see. oh well, it was about time May 14 13:33:00 poor little workhorse May 14 13:33:06 [2011-05-14 12:21:27] [Notify] apt ist offline (kornbluth.freenode.net). May 14 13:33:08 [2011-05-14 12:21:41] [Notify] infobot ist offline (kornbluth.freenode.net). May 14 13:33:37 I guess she's on a walk at the beaches with Tim Riker ;-D May 14 13:33:44 hehe, yeha May 14 13:34:49 [2011-05-14 12:46:12] *** ChanServ setzt Modus: +o TimRiker May 14 13:45:22 mickeyl: what's with family? May 14 13:46:33 you're not yet smoking a fat cigar? ;-D May 14 13:50:02 http://docs.openmoko.org/trac/ needs some cleanup there are 3 years old tickets related to Om 2007/8 May 14 14:00:40 Martix: indeed May 14 14:06:33 DocScrutinizer: I managed to login with my trac account, but I cant edit anything, except my account settings May 14 14:07:14 hmm, I have no idea who's managing the permisions there and what are the default ones May 14 14:07:56 probably its related to "Warning: " which I see in head of page May 14 14:08:08 ooh May 14 14:08:50 /query roh May 14 14:09:00 using Firefox 4 May 14 14:09:54 I told roh some 4 weeks ago it's hard to open a trac ticket against trac not working properly May 14 14:12:28 Martix: http://admin-trac.openmoko.org/trac May 14 14:15:26 Martix: anyway suggest a ticket to close - I want to see if I am allowed to do this May 14 14:16:00 also suggest which close reason to use May 14 14:21:14 DocScrutinizer: try this: http://docs.openmoko.org/trac/ticket/69 wontfix or invalid May 14 14:27:18 http://admin-trac.openmoko.org/trac works for me May 14 14:31:31 closing tickets works for me May 14 14:31:52 ping roh to help you sort it out, please May 14 14:32:22 only difference that I see: community is 0.11.2.1, admin trac is Trac 0.11.3 May 14 14:32:49 DocScrutinizer: no news, Sabine still pregant, baby doesn't want to come out yet May 14 14:33:04 patiently waiting :) May 14 14:33:11 oh well, they do what they want :-) May 14 14:33:16 oh yeah May 14 14:33:19 already May 14 14:34:09 mickeyl: are you trac admin empowered? May 14 14:34:45 see ^^^ May 14 14:35:16 I got nfi who's actually taking care about trac nowadays May 14 14:37:35 Martix: anyway feel free to ping me with a list of trac # and suggested close reasons anytime May 14 14:38:11 plus a very short rationale about why this ticket has to be closed May 14 14:38:12 (do what they want) it that will stay that way... the next 18yrs or so ;) May 14 14:38:21 mrmoku: uh oh May 14 14:38:32 :-P May 14 14:38:35 yeah May 14 14:38:47 * mrmoku still waiting for the two bigger ones to come back for lunch :P May 14 14:39:11 oh well... it's cold now anyway May 14 14:39:11 DocScrutinizer: ooh, this problem with Trac is 2 years old https://admin-trac.openmoko.org/trac/ticket/1810 May 14 14:40:11 DocScrutinizer: no, it doesn't look like i still have admin access May 14 14:40:59 pester roh!! May 14 14:41:18 Martix: ^^ May 14 14:41:26 that's actually annoying May 14 14:41:39 /query roh May 14 14:41:57 tell him to bloody fix that May 14 14:43:03 mickeyl: please get full access to all OM infra from roh, as he *might* get caught by a bus eventually May 14 14:45:03 I don't want to find wiki.maemo.org et al suddenly orphaned eventually, as we experienced with this very channel May 14 14:46:03 DocScrutinizer: I already told him, sorry for interrupting, I am only suprised that this "blocking" error stayed unresolved for 2 years May 14 14:56:57 wtf?? May 14 14:57:10 aah May 14 14:59:50 mickeyl: are you known (anywhere) as "uur" ? May 14 15:00:13 duh May 14 15:00:16 sorry May 14 15:00:25 not to my knowledge May 14 15:01:32 I gather I'll never get out the mickey|zzZZzz of cs accesslist May 14 15:03:07 how to make it work: May 14 15:03:16 1) make a call with fsoaudiod launched May 14 15:03:21 ctrl + c fsoaudiod May 14 15:03:24 within the call May 14 15:03:29 2) stop the call May 14 15:03:34 3) launch fsoaudiod May 14 15:03:37 [Notice] -NickServ- Information on mickey (account uur): May 14 15:03:38 4) call again May 14 15:03:46 DocScrutinizer: 'mickey' is not me May 14 15:03:49 unfortunately... May 14 15:03:52 yeah May 14 15:04:17 please check accesslist if I've done other crap May 14 15:04:53 GNUtoo|laptop: why would you want to SIGSTOP? May 14 15:05:02 no idea May 14 15:05:05 it works like that May 14 15:05:11 else I've no sound May 14 15:05:19 heh, odd May 14 15:05:24 it can also work like that: May 14 15:05:29 1) make a call with fsoaudiod launched May 14 15:05:38 ctrl + c fsoaudiod within the call May 14 15:05:43 launch fsoaudiod again May 14 15:05:45 stop the call May 14 15:05:49 call again May 14 15:06:00 that's odd May 14 15:06:34 GNUtoo|laptop: anyway I gather you had sound in calls for some time? that's great news now May 14 15:06:47 (on N900 of course) May 14 15:06:49 yes May 14 15:06:59 there is a very little krrr May 14 15:07:07 it was not there during the first test May 14 15:07:15 maybe the order is wrong May 14 15:08:06 next call don't work either May 14 15:08:11 I've to ctrl + c it May 14 15:08:14 and restart it May 14 15:08:22 gathered as much May 14 15:08:23 ctrl+ c *during a call* May 14 15:08:52 krrr is gone May 14 15:08:56 no idea why May 14 15:09:40 anyway I've to check my source code May 14 15:09:56 you might want to create a log of raw audiodata (refer to tee 'user defined' plugin, aka "file" pcm plugin) and analyze this May 14 15:10:05 + buffersize = buffersize / 2; May 14 15:10:06 yes May 14 15:10:08 great May 14 15:10:41 1) I'll look at my logs May 14 15:10:51 2) I'll try to add aplay -v like debugging May 14 15:13:07 mickeyl, I've an oe question on fsogsmd May 14 15:13:27 since I use autorev I've the problem that the nokia 900 fsogsmd plugin is gone May 14 15:13:37 but oe doesn't force autorev May 14 15:13:45 so it should work without May 14 15:13:46 3) create a audio dump by opening "tee dumpfile default" for playback May 14 15:14:05 yes May 14 15:14:06 and upload so I can analyze May 14 15:14:10 GNUtoo|laptop: did anyone disable it in OE? May 14 15:14:13 I've to do 1 and 2 first May 14 15:14:17 not yet May 14 15:14:30 but what if oe has a rev that has the plugin May 14 15:14:31 ? May 14 15:14:35 and in autorev it's gone May 14 15:14:44 we should bump then May 14 15:14:49 how can it be gone? May 14 15:14:57 i don't quite understand May 14 15:15:04 lowlevel_nokia900 is gone May 14 15:15:06 in fsogsmd May 14 15:15:24 ah, you're talking about the lowlevel plugin May 14 15:15:25 because mrmoku reimplemented it in the nokiia_isi backend May 14 15:15:27 yes May 14 15:15:36 can be disabled in the config May 14 15:15:39 or rather should be May 14 15:15:41 it prevent do_rootfs to work May 14 15:15:47 why that? May 14 15:15:55 because do_rootfs looks for it May 14 15:16:00 and it's absent May 14 15:16:23 ah, the manual :/ May 14 15:16:37 that's why i would have preferred do_package_split May 14 15:16:41 http://www.pastie.org/1900509 May 14 15:16:42 hmm cool, gibberish ;-D May 14 15:17:08 ok, so someone should remove it from oe and bump the revision May 14 15:17:19 and add do_package_split rather sooner than later May 14 15:17:19 * DocScrutinizer smiles and feels lucky he doesn't need to uderstand everything May 14 15:17:47 oe is simple compared to alsa.... May 14 15:18:38 well, I tend to agree with you here. It took me like 2 years to somewhat get a grip on ALSA May 14 15:19:00 lol I have bad experience with it too. htcdream..... May 14 15:19:12 but it was mainly htcdream's fault May 14 15:19:13 there are details in the concept that eren't explained anywhere May 14 15:19:22 ok May 14 15:20:19 basic things like "what's a plugin stack and how is it supposed to work" May 14 15:20:50 my alsa learning was somewhat hardcore: May 14 15:21:18 I dived head first in it without any knowledge of it, without serial console(printk) May 14 15:21:25 and with a very badly written driver May 14 15:21:31 the concept of each plugin having two completely unrelated functions, one for playback aka downstream and one for record aka capture aka upstream May 14 15:21:33 on hardware that is not great at all May 14 15:21:44 without reading the leaked documentation May 14 15:21:55 (because it was leaked) May 14 15:22:03 ok May 14 15:22:24 then I suffer of it now May 14 15:22:32 nervousness when something doesn't work May 14 15:22:45 it occur mostly when I'm tired May 14 15:25:28 once you wrap your head around http://www.alsa-project.org/alsa-doc/alsa-lib/group___p_c_m___dump.html and what's the meaning of its output, it helps a lot to get less nervous May 14 15:26:25 ok May 14 15:26:36 I meant in general with the nervousness May 14 15:26:39 since it somewhat work May 14 15:26:43 I'm less nervous May 14 15:26:50 since I know what to do also May 14 15:27:01 I'll rest a bit May 14 15:27:05 since the sport tired me May 14 15:27:49 found this: http://www.alsa-project.org/alsa-doc/alsa-lib/group___p_c_m___helpers.html#g8996218a2d72640723489c5e06f281ea May 14 15:28:44 wow nice May 14 15:28:54 a kinf of frames_to_bytes May 14 15:29:01 *kind May 14 15:29:15 there's also May 14 15:29:16 ssize_t snd_pcm_frames_to_bytes ( snd_pcm_t * pcm, May 14 15:29:17 snd_pcm_sframes_t frames May 14 15:29:19 ) May 14 15:29:28 *could be used to implement a kind of frames to bytes May 14 15:29:29 ok nice May 14 15:29:36 I looked for that May 14 15:29:41 but failed to find it May 14 15:31:36 btw for fso, the vibrator works like a force feedback joystic vibrator May 14 15:31:44 since wordwarvi uses the vibrator May 14 15:31:53 and that it works May 14 15:32:13 and that there is a /dev/input node for it May 14 15:32:24 mrmoku, hi May 14 15:33:36 yo May 14 15:33:53 DocScrutinizer, what's the difference between frames and samples? May 14 15:34:08 mrmoku, any progress? May 14 15:34:17 GNUtoo|laptop: on what part? May 14 15:34:25 any part you're working on May 14 15:34:31 hehe May 14 15:34:35 (for the n900) May 14 15:34:38 the lawn is fine now :P May 14 15:34:46 lawn? May 14 15:34:49 garden May 14 15:34:52 ok May 14 15:35:03 with the messages I started May 14 15:35:03 GNUtoo|laptop: I'd think there's no real difference May 14 15:35:12 still mainly reading docs though May 14 15:35:22 and right now I'm doing a little bit of daywork May 14 15:35:29 on my side I went doing some sports, came back and found a way to make the audio work with some vodoo May 14 15:35:38 GNUtoo|laptop: maybe you can think of a sample to be part of a frame for one channel only, while a frame is holding all channels May 14 15:35:39 ok May 14 15:35:49 ok May 14 15:36:11 GNUtoo|laptop: but if there's something I can do to help you... I'll interrupt May 14 15:37:02 mrmoku, if you want I can give you unrelated tasks for instance: May 14 15:37:18 *change kernel defconfig in oe: add alsa xrun debugging and remove oss May 14 15:37:31 or similar things May 14 15:37:34 but it's not urgent May 14 15:37:43 so you can continue what you're doing May 14 15:37:46 ok May 14 15:37:50 since I'm resting a bit May 14 15:37:55 (tired after sport) May 14 15:38:03 yeah, relax then May 14 15:38:17 * GNUtoo|laptop try to lean vi keys May 14 15:38:24 (playing wordwarvi ) May 14 15:38:45 freesmartphone.org: 03mickey 07cornucopia * r5991f917db7c 10/fsoaudiod/vapi/alsa-ext.vapi: fsoaudio: alsa vapi: bind conversion helpers May 14 15:39:13 frame = samples * channels May 14 15:39:18 afaiui May 14 15:39:49 s/samples/sample// May 14 15:39:57 so, for mono it's the same May 14 15:41:41 ok May 14 15:42:05 fwiw, i just added those conversion helpers May 14 15:42:07 so you can use them May 14 15:42:37 thanks a lot May 14 15:42:51 the code is sooo dirty currently May 14 15:42:56 hmm, one more: May 14 15:43:21 freesmartphone.org: 03mickey 07cornucopia * r4b775f833806 10/fsoaudiod/src/lib/alsa.vala: fsoaudio: lib: set card as public May 14 16:14:36 mickeyl, where is the destructor of cmthandler? May 14 16:14:54 I added a hook in cmthandler() constructor for the setup of my sound card May 14 16:15:04 but I wonder where should I put the .close May 14 16:15:16 and I know the constructor is not the right place May 14 16:15:21 since it consume more power May 14 16:15:26 (DAPM) May 14 16:15:42 DAPM permit to power down the card when it's not used if I understood well May 14 16:15:47 and I open it all the time May 14 16:52:05 GNUtoo|laptop: probably there is none (yet) May 14 16:52:07 just add it May 14 16:52:30 yes but where May 14 16:52:34 where to hook? May 14 16:52:55 fsogsmd/src/lib/modem.vala: ~AbstractModem() May 14 16:52:58 just like this May 14 16:53:00 in your class May 14 16:53:06 ok May 14 16:53:15 after the constructor best May 14 16:53:17 the ~ seem so C++ May 14 16:53:21 yup :) May 14 16:55:03 I'ld rather hook here: May 14 16:55:31 setAudioStatus May 14 16:59:36 public void setAudioStatus( bool enabled ) { connection.state_change_call_status( enabled ); } May 14 16:59:40 basically that's what it does May 14 16:59:47 so I add a hook there May 14 17:00:01 but I wonder what connection.state_change_call_status does May 14 17:00:14 it does somehting in libcmtspeech May 14 17:04:52 GNUtoo|laptop: it tells cmtspeech that a call is active... or not active May 14 17:04:55 I think May 14 17:06:46 yes but what cmtspeech does then May 14 17:08:50 starts to send packets from the modem? May 14 17:09:11 yes but where's that May 14 17:09:14 in the lib? May 14 17:09:17 or in the vala file? May 14 17:11:48 I think that's in the lib May 14 17:11:52 so I should hook there May 14 17:14:08 how do I cache the values of enabled already May 14 17:14:13 there is a right way to do it May 14 17:15:41 hmmm many fsodeviced plugin don't cache May 14 17:15:49 I'll come up with something then May 14 17:18:22 i don't think you want to patch the lib May 14 17:18:35 setAudioStatus() should be the right place May 14 17:42:03 mickeyl, May 14 17:42:07 ERROR: CMTSPEECH: nokiamodem_backend: CS_CONFIG_BUFS returned an error (5): Input/output error May 14 17:42:07 ERROR: CMTSPEECH: nokiamodem_backend: Unable to set up buffer config for call May 14 17:42:26 I'll do something that works first May 14 17:42:31 that works reliabily May 14 17:42:42 then me or someone else will cleanup or help me to cleanup May 14 17:42:45 hmm... if cmtspeech fails, we're lost :) May 14 17:43:04 I think it's more something like open/close in posix term May 14 17:43:13 it works the first time May 14 17:43:15 ok May 14 17:44:13 CMTSPEECH: nokiamodem_backend: Initialized driver buffer: res -1, params size=324. May 14 17:44:25 there is also that ^^^ right before May 14 17:44:28 so I'll look May 14 17:44:32 at the fds May 14 17:47:03 CMTSPEECH: nokiamodem_backend: Initialized driver buffer: res -1, params size=324. May 14 17:47:39 the importnat part according to me is: Initialized driver buffer May 14 17:48:01 it means that it tried to initialize the driver buffer once more(it shouldn't) and that it failed(-1) ? May 14 17:49:23 could be May 14 17:49:52 CMTSPEECH: nokiamodem_backend: Initialized driver buffer: res 0, params size=324. is when it works May 14 17:50:03 and this time it worked the first time May 14 17:50:07 without krrr May 14 17:50:16 nice May 14 17:50:18 and without stopping it in the middle of a call May 14 17:50:39 will look if it can be reproduced May 14 17:51:33 hmmm no I'll see that later May 14 17:57:39 it's not so nice currently May 14 17:57:46 basically when it works I don't know why May 14 17:57:51 heh, fun May 14 17:57:53 and when it doesn't either May 14 17:57:53 timing May 14 17:58:07 not necessarly May 14 17:58:10 but it's a possibility May 14 18:01:16 it could be something in fsogsmd too May 14 18:01:23 since many stuff is not implemented May 14 18:01:27 like list calls May 14 18:02:17 org.freesmartphone.GSM.Call.ReleaseAll not implemented May 14 18:02:39 mickeyl: btw. what order do you prefer inside a class? first publics? or first privates? May 14 18:03:09 GNUtoo|laptop: want me to implement those two? May 14 18:03:10 hmmm strange May 14 18:03:22 mdbus2 -s org.freesmartphone.ogsmd /org/freesmartphone/GSM/Device org.freesmartphone.GSM.Call.Release 1 always return () May 14 18:03:27 even if the call is hanged up May 14 18:03:32 or teorically May 14 18:03:40 since I call 4242 how can I know May 14 18:03:41 mrmoku, yes May 14 18:03:44 ok May 14 18:03:49 more release handling please May 14 18:04:39 mrmoku: i tend to private, protected, public May 14 18:05:05 also with regards to the methods, although i'm sure my code does not always follow that:) May 14 18:05:26 yes that's how java does it too May 14 18:05:32 private, protected, public May 14 18:10:31 mickeyl: ok, but constructor/destructor always first, right? May 14 18:10:55 yeah, was mainly talking about the methods May 14 18:11:14 yes May 14 18:32:03 mickeyl: catch ( Error e ) catches all errors? May 14 18:35:37 sounds like it but I'm not sure May 14 18:35:52 if mickeyl doesn't respond there is also #vala on gnome irc server May 14 18:58:03 (Error) I think so but I'm not sure May 14 18:58:29 looking at morphis code it probably does May 14 19:11:44 hi May 14 19:12:08 I would like to translate /etc/pointercal (from ts_calibrate) into coordinates usable by xinput May 14 19:13:19 I know there is some script, I have done this before May 14 19:13:36 run xinput_calibrate_once May 14 19:13:42 and copy the result in oe May 14 19:14:00 (the result beeing /etc/pointercal.xinput) May 14 19:14:04 let me boot May 14 19:14:26 GNUtoo|laptop: I should explain the background May 14 19:14:55 GNUtoo|laptop: I have two options for touchscreen on kernel level: multi touch boolean value May 14 19:15:38 GNUtoo|laptop: without xorg rotation I have precis pointer - without any calibration May 14 19:15:57 in both, multi touch and non multi touch mode May 14 19:16:37 with screen rotation + multi touch off I need to provide xinput script like following: May 14 19:16:41 ahhh ok May 14 19:16:45 like htcdream May 14 19:17:02 xinput set-prop htcleo-touchscreen "Evdev Axis Calibration" 972 4 569 15 May 14 19:17:05 xinput set-prop htcleo-touchscreen "Evdev Axis Inversion" 1, 0 May 14 19:17:08 xinput set-prop htcleo-touchscreen "Evdev Axes Swap" 1 May 14 19:17:17 I run this automatically and everything is goodi n rotation mdoe May 14 19:17:26 (via Xsession.d) May 14 19:17:35 then add this to /etc/pointercal.xinput May 14 19:17:37 in oe May 14 19:17:39 BUT - with multitouch on, same script doesn't work and I get dodgy pointer May 14 19:17:44 ok May 14 19:18:29 so I think I have to recalibrate May 14 19:18:36 with mt on May 14 19:18:40 yes May 14 19:18:44 but I don't have the converter script :( May 14 19:18:57 why not removing multitouch? May 14 19:19:15 gestures in ubuntu distro May 14 19:19:37 ok May 14 19:19:45 it has a program "jinn" to easily map gestures to keyboard combinations. very nice May 14 19:19:46 and I bet ubuntu people aren't helping a lot May 14 19:19:52 ok May 14 19:20:12 they help with general things a lot. but it is difficult to get devicce specific help May 14 19:20:16 they don't have any qsd or msm May 14 19:20:32 freesmartphone.org: 03mok 07cornucopia * rcd602caf2bd5 10/fsogsmd/src/plugins/modem_nokia_isi/callhandler.vala: May 14 19:20:32 freesmartphone.org: fsogsmd: modem_nokia_isi: throw an error when trying to release an already released call May 14 19:20:32 freesmartphone.org: Signed-off-by: Klaus Kurzmann May 14 19:20:33 freesmartphone.org: 03mok 07cornucopia * r85cd866bd103 10/fsogsmd/src/plugins/modem_nokia_isi/ (callhandler.vala isimediatorhelpers.vala): May 14 19:20:33 freesmartphone.org: fsogsmd: modem_nokia_isi: implement CallListCalls and CallReleaseAll May 14 19:20:33 freesmartphone.org: Signed-off-by: Klaus Kurzmann May 14 19:20:38 ok May 14 19:20:40 GNUtoo|laptop: only compile tested May 14 19:20:44 how does meego handle that? May 14 19:20:48 ok May 14 19:20:56 * GNUtoo|laptop is yawning all the time tough May 14 19:21:02 so maybe tomorrow May 14 19:21:02 I think it's - platfrom independent- interesting to make xorg work with multi touch May 14 19:21:04 :) May 14 19:21:15 yeah, will go to bed early too today May 14 19:21:39 GNUtoo|laptop: so you have any clue how to /etc/pointercal > xinput.sh ? May 14 19:22:17 no May 14 19:22:27 ah wait I know where the script is May 14 19:22:32 on tmzt's personal file server May 14 19:22:36 but look at /etc/pointercal.xinput May 14 19:22:39 it's a script May 14 19:22:51 I'll have to go to sleep, bye May 14 19:23:10 gnu that was quick May 14 19:23:12 good night :) May 14 19:36:35 freesmartphone.org: 03mok 07cornucopia * rfd414133a154 10/fsogsmd/src/plugins/modem_nokia_isi/isicallmediators.vala: May 14 19:36:35 freesmartphone.org: fsogsmd: modem_nokia_isi: actually commit the mediator (list calls) too May 14 19:36:35 freesmartphone.org: Signed-off-by: Klaus Kurzmann May 14 19:55:23 dcordes,what xf86-video do you use btw May 14 19:59:59 GNUoo|n900: fbdev May 14 20:06:58 ok May 14 20:07:18 no xrandr -o 1 then May 14 20:07:42 btw no news of cotulla? May 14 20:07:56 we need him to publish alsa May 14 20:08:18 even if it's incomplete May 14 20:09:41 he sometimes is on irc May 14 20:09:54 but he didn't respond.... May 14 20:12:00 i feel a bit guilty to have forgetten of your requests while waiting for mrmoku's code May 14 20:14:07 GNUoo|n900: guilty? :) you're crazy May 14 20:14:45 GNUoo|n900: you're the single most person working to improve support for 10 devices at the same time in a foss phone project May 14 20:14:53 i try not to waiste time May 14 20:15:02 lol May 14 20:15:46 note that i don't have that much devices May 14 20:16:07 5 phones and 3 bug devices May 14 20:17:06 it's hot here....(no idea of how much degrees) May 14 20:18:23 basically I looked at some openpandora videos and had no new porting game idea May 14 20:18:56 although I ported wordwarvi because I saw it on a pandora video May 14 20:19:06 not so long ago May 14 20:22:56 i'm not saying that I don't do enough.... May 14 20:23:35 i was just saying that i didn't know what to do while waiting for mrmoku..... May 14 20:24:04 and that the time wouls have been better used doing somethung usefull May 14 20:24:36 did you consider some internship at a company working with embedded linux ? May 14 20:24:46 or paid job.. May 14 20:25:03 yes May 14 20:25:10 maybe this summer May 14 20:25:26 there is one near where I go to hollidays May 14 20:25:47 I even reviewed some patches of the founder of that company May 14 20:26:17 and at the end he got oe commit access May 14 20:27:25 when I reviewed his patches I didn't know his company was so close.... May 14 20:27:42 i did it because I was interested in the patches May 14 20:28:17 i even visited that company during last summer hollidays May 14 20:28:43 since i cannot do internships that count yet....i prefer paid jobs May 14 20:29:01 internships has to go with the thesis May 14 20:29:19 else you can do one May 14 20:29:27 but it doesn't count May 14 20:31:22 if I get enough money.....maybe I could buy an usrp May 14 20:33:09 altough I don't know well the law on taxes May 14 20:34:01 maybe student jobs have less or no taxes? May 14 20:36:37 basically usrp aren't cheap.... May 14 20:38:14 btw....how much do you know about msm/qsd on basebbans side? May 14 20:38:24 *baseband May 14 21:41:27 hey guys. I'm just wondering, is shr a version of linux? May 14 21:42:33 RocketSnail: SHR is a distribution May 14 21:42:58 RocketSnail: like ubuntu, fedora, debian, suse May 14 21:43:41 okay thanks. that means that if I add a repository to my shr on my phone, I can basically get any app available for ubuntu for example? May 14 21:43:50 RocketSnail: no May 14 21:43:59 oh. May 14 21:45:49 RocketSnail: just like you can't install ubuntu packages to fedora May 14 21:46:46 okay, thank you a lot lindi :). May 14 21:47:07 RocketSnail: I run debian on my openmoko, I can get any app available for Debian :) May 14 22:41:44 Hi. mickeyl: could you help me to understand how should i use page May 14 22:42:01 s/page/pages on apps?/ May 14 23:51:42 | isicallmediators.vala:76.23-76.52: error: The name `listCalls' does not exist in the context of `FsoGsm.CallHandler' May 14 23:51:44 | calls = yield theModem.callhandler.listCalls(); **** ENDING LOGGING AT Sun May 15 02:59:58 2011