#qi-hardware IRC log for Monday, 2011-01-10

kyakxiangfu: hi! do you remember the issue with gmenu2x started before loading of sound modules and those disturbing messages from alsa?05:16
xiangfukyak: yes.05:17
kyaki don't know if this is a good solution, but i load these modules from gmenu2x wrapper: http://downloads.qi-hardware.com/people/kyak/tmp/gmenu2x05:17
kyakit adds a small overhead when starting gmenu2x, but it is neglectable05:18
kyakit's /usr/bin/gmenu2x05:18
xiangfukyak: ok. I am testing the gmenu2x time. need to know how much time the 'load_modules' need :)05:23
kyaki already tested, but your results may be different05:24
xiangfuI remove the "./gmenu2x" from your version and old version files.05:27
xiangfuyour version: 0.50s05:27
xiangfuold verison: 0.06s05:27
kyakwould be better to leave load_modules only05:28
kyakoh, ok05:28
kyaki misunderstood05:28
kyakok, the overhead is 0.42s05:28
kyakit is only needed when first starting gmenu2x05:29
kyakmaybe we can use different wrappers ?05:29
xiangfuhow about add something like:05:36
xiangfulsmod | grep snd_soc_qi_lb6005:36
xiangfuif [ "$?" = "1" ]; then05:36
xiangfu        sleep 105:36
xiangfuin fact, I still don't understand why Lars build the sound as module. why we can just build the sound into kernel.05:38
xiangfumaybe I need ask him in mailing list :)05:39
kyakoh yeah05:39
kyakmakes sense to have monolithic kernel05:39
kyakthis would speed up boot, too05:40
kyaki think it's better than wrappers :)05:40
xiangfuif someone want start 'gmu' instead of 'gmenu2x' , he need the sound modules :)05:48
kyakor, i just want a fancy sound right after boot :)05:49
kyakthe one from Windows XP05:49
wpwrakkyak: forget splash screens and such. when booting, show a little video clip, with a girl getting naked to some music.05:50
wpwrakkyak: people will beg for the boot time to be increased ! ;-)05:51
kyakawesome idea!: )05:51
wpwrakhmm, soldered an atusb in 135 minutes. not bad. i'm a bit reluctant to destroy the perfection and test it ...08:02
rohwpwrak: :)08:19
qi-bot[commit] kyak: fbterm now builds against both tiny and gnu iconv http://qi-hw.com/p/openwrt-packages/8de229d09:04
qi-bot[commit] kyak: gcc-mips: copy missing files manually http://qi-hw.com/p/openwrt-packages/5282f8110:31
qi-bot[commit] Werner Almesberger: f32x/: renamed OpenWRT target from "ben" to "ben_openwrt" and added "ben_jlime" http://qi-hw.com/p/f32xbase/901998510:43
qi-bot[commit] Werner Almesberger: tools/atrf-txrx/atrf-txrx.c (enter_test_mode_231): clean up and exit on error http://qi-hw.com/p/ben-wpan/110d5ab10:43
qi-bot[commit] Xiangfu Liu: flite: include in upstream http://qi-hw.com/p/openwrt-packages/fe66b5410:51
qi-bot[commit] Xiangfu Liu: gmenu2x: add mediatomb icon http://qi-hw.com/p/openwrt-packages/75a5e5e10:51
qi-bot[commit] Xiangfu Liu: hnb: add default .hnbrc, thanks to mark adrian bell http://qi-hw.com/p/openwrt-packages/842102010:51
kyakxiangfu: having fun with flite :) i noticed that playing the resulting file with aplay produces much better result than playing the output immediately11:12
xiangfukyak: don't notice that.11:17
kyakhmm, ok11:20
xiangfukyak: what is the text, you using in flite?11:25
kyakxiangfu: compare flite -t "test string" play.wav && aplay play.wav and flite -t "test string" play11:26
kyakthe first one is much cleaner11:26
kyakno "jittering"11:26
xiangfukyak: seems same here :), I can not find out the difference.11:33
kyakflite -t "this is just a test string" play && flite -t "this is just a test string" play.wav && aplay play.wav11:35
kyakmaybe this way you can hear the difference?11:35
xiangfukyak: sorry. still same here. with kernel  image 12-14.11:43
kyakxiangfu: ok, don't bother! still sounds not bad with aplay11:46
kyakxiangfu: btw, it seems that "dialog" is also upstream11:49
xiangfukyak: yes.11:51
qi-bot[commit] Xiangfu Liu: dialog: remove it since it included by upstream feeds packages http://qi-hw.com/p/openwrt-packages/c269f7611:53
qi-bot[commit] Werner Almesberger: tools/lib/misctxrx.c (wait_for_interrupt): restore signal also after timeout http://qi-hw.com/p/ben-wpan/6ff235e11:54
qi-bot[commit] Joachim Steiger:  removed old files http://qi-hw.com/p/m1/17ab0e112:45
qi-bot[commit] Joachim Steiger:  added 27c3 release of case http://qi-hw.com/p/m1/3b3095f12:47
Action: Jay7 have found his nanonote at home15:41
Jay7yeah.. it really need player control buttons at some side15:41
Jay7or at least wheel+ok/cancel buttons like zauruses15:41
Jay7or remote control like some players have (via 4th connector on audio jack)15:43
Jay74th ring even15:43
wpwrakthese wheels are surprisingly expensive16:08
Jay7buttons are cheaper :)16:08
wpwrakthe ben already has a ton of these :)16:14
Jay7no one is accessible when ben is closed :)16:15
Jay7at least 3 button needed: play/pause, stop, ff16:16
Jay7s/ff/next track/16:16
wpwrakyeha, it's not really designed as a media player16:20
wpwraki.e., the dictionary shows :)16:20
wpwrakild love to16:21
wpwrakget rid of the function keys, though16:21
kristianpaultouchpad !16:21
wpwrakwrong century :)16:21
wpwrak(the fn keys)16:21
wpwrakanalog joystick !16:21
kristianpaulyeah fn are not good16:21
Jay7wpwrak: how to switch VT instead? :)16:21
Jay7openpandora have lot of joysticks :)16:22
kristianpaultoo many16:22
wpwrakJay7: (vt) have the fn keys as some key + modifier16:22
kristianpaulhow old is qwerty16:26
wpwrak(according to wikipedia)16:35
wolfspra1lwpwrak: lekernel and I had an excellent day at the reichl-emv.de testing lab!17:25
wpwrakwolfspra1l: so it is as good as it looks ? :)17:26
wolfspra1lfor milkymist one, the tests went great, everything passed except for one test related to an ESD induced magnetic field17:26
wpwrakwolfspra1l: excellent. congratulations !17:26
Action: kristianpaul quack17:26
wolfspra1land for that one, we already found a solution in the form of an additional shielding layer on the bottom of the case (under the pcb)17:26
wolfspra1land, amazing, Berlin is like little China, we sourced a 0.2mm 99.5% aluminum foil, had it cut to 149x124mm, got 100 pieces for 25 EUR, all in 2 hours :-)17:27
kristianpaulThat explaing why lekernel had troubles in 27C3 in some particular place?17:27
wolfspra1lonly the holes were missing but we drilled them by hand17:27
wpwrakwow ;-)17:28
wolfspra1lroh will polish this improvement a little, but basically it's under control17:28
wolfspra1land except for that all other tests passed17:28
wolfspra1lalso, this must be the cheapest ce/fcc testing - total is only 450 EUR including report! :-)17:28
wolfspra1lit gets better17:28
wpwrakthat's a very good result.17:28
wolfspra1lthe guy has a whole daniel duesentrieb style lab, and sebastien will get access for 'personal research'17:29
wpwrakhey ! when's the next plane to berlin ? ;-)17:29
wolfspra1lI mentioned the ben-wpan project and he said sure he can help.17:29
wolfspra1lso as of right now, that's definitely the place I would first try to go to.17:29
wolfspra1lif roh is involved in the case, either roh or Sebastien can go there, with a little irc session on your end to back it up17:30
wpwrakperfect. ben-wpan will probably be more of a challenge, since it's designed to radiate stuff.17:30
wolfspra1lSebastien also offered to go there with the ben-wpan, since he is interested to go to the lab anyway (some radar stuff he wants to do...)17:30
wolfspra1labsolutely no problem17:30
wolfspra1las I had suspected, this guy is a really good source and if he and us stays around, I'm sure we will work with him more...17:31
wolfspra1ltotally fits our style and culture as well17:31
wpwraksounds wonderful17:32
wolfspra1lkristianpaul: I don't think the esd/field related bug we found is the same as the video-in issues Sebastien saw once (they went away the next day btw).17:33
kristianpauloh, SAR research nice :-)17:34
kristianpaulgreats news then :-)17:35
wolfspra1lyes I think so.17:36
wolfspra1lit was a long and hard day, but it was good.17:36
wolfspra1lwe still need to improve the shielding solution a little, but more important is that all the other tests passed (quite a few...)17:37
wolfspra1land we have a fully understood and verified solution for that one test that failed17:37
wpwrakexcellent. now i have to figure our why atusd produces such an ugly spectrum. i was hoping it was me getting test mode wrong with the new chip, but atusb (with the 231) looks quite sane.17:42
wolfspra1lwpwrak: btw, roh pushed the latest case dxf file, but he wanted to call the subdir 'cad' not 'cam'. oh well :-)17:42
wolfspra1lso it's cad now...17:42
wpwrak(cad/cam) as long as it's unique .. ;-)17:43
wolfspra1lI still want to add a little more, for example a little description/documentation about this production process, as well as details for some of the parts, like screws, spacers, etc.17:43
wolfspra1lbut most importantly the root dxf he uses for production is now published, and a little .svg assembly manual (ikea style) as well17:43
wpwrak(assembly manual) oh, very good !17:44
wolfspra1lit's great17:44
wolfspra1lsome more info about specs of screws, spacers, etc. missing - will follow...17:44
wpwrak** (inkscape:8207): WARNING **: Unable to open linked file: /home/labor/Desktop/Kunden/Milkymist-Howto/Mm1_rc1_parts_on_pcb_farb.png17:45
wolfspra1lhe :-)17:45
wolfspra1lroh: you there?17:45
wpwrak(in response to inkscape milkymist-assemble-howto.svg)17:46
wpwrakdo the rubber feet go on top of the screws ? or are they located on the side of them ?17:48
wpwrakin the latter case, it would be clearer if they had their own dashed line17:48
kristianpaulIf you remove the linked image is more clear17:59
kristianpaulyes rubber position is confusing18:00
kristianpaulguess is on top18:01
rohwpwrak: the rubberfeet are mounted diagonally to the center from the screws20:03
rohjust like 1 or 1.5cm offset from the hole20:03
wpwrakroh: ah, good. makes them simpler. the drawing out be clearer if you added a dashed line and maybe even a dashed ellipse for their final resting place for them too.20:53
rohwpwrak: will talk about that with my inkscape-operator20:54
wpwrakhehe :)20:56
kristianpaul"Automake: Escaping makefile hell"21:29
--- Tue Jan 11 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!