| 00:32.31 | *** join/#gpe andreas_gi (~andreas@L1262P17.dipool.highway.telekom.at) |
| 00:39.42 | *** join/#gpe TheMasterMind1 (~aman@ip68-100-127-11.nv.nv.cox.net) |
| 02:43.52 | *** join/#gpe semtex (~Speuler@mnch-d9ba4de6.pool.mediaWays.net) |
| 04:07.14 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 04:07.14 | *** join/#gpe paul (~paulj@itg-gw.cr008.cwt.esat.net) |
| 04:07.14 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) |
| 04:07.14 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 05:19.15 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) |
| 06:22.24 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) |
| 06:22.24 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 06:22.24 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) |
| 06:22.24 | *** join/#gpe paul (~paulj@itg-gw.cr008.cwt.esat.net) |
| 06:22.25 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 06:29.37 | *** join/#gpe florian (~fuchs@proton.hrz.uni-siegen.de) |
| 06:29.44 | florian | morning |
| 06:30.53 | pigeon | morning florian! |
| 06:32.07 | florian | hi pigeon |
| 06:44.11 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) |
| 06:44.11 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 06:44.11 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) |
| 06:44.11 | *** join/#gpe paul (~paulj@itg-gw.cr008.cwt.esat.net) |
| 06:44.11 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 07:04.40 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) |
| 07:04.40 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 07:04.40 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) |
| 07:04.40 | *** join/#gpe paul (~paulj@itg-gw.cr008.cwt.esat.net) |
| 07:04.40 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 07:18.28 | semtex | 'morning |
| 07:23.06 | florian | morning Speuler |
| 07:23.33 | Speuler | ah, the_florian |
| 07:23.57 | Speuler | grias di :) |
| 07:26.16 | florian | eh? :- |
| 07:26.20 | florian | ) |
| 07:27.28 | Speuler | that's how it sounds in the south |
| 07:27.49 | florian | :-) |
| 07:30.43 | florian | Ahh... |
| 07:33.31 | Speuler | maar ik kan ook plat snakken |
| 07:36.48 | florian | Maybe we should agree to keep using English?! ;-) |
| 07:40.53 | tigert | nej, english är den sukka |
| 07:43.05 | florian | tigert: Finnish?! |
| 07:43.10 | Speuler | looks danish |
| 07:45.17 | *** join/#gpe paul ([CTLjy16E9@s0.alphyra.bbnplanet.net) |
| 08:02.58 | tigert | that is something nordic |
| 08:03.06 | tigert | not finnish |
| 08:03.16 | tigert | probably wrong wiht any language :) |
| 08:03.25 | florian | :-) |
| 08:12.28 | *** join/#gpe Speuler (~Speuler@mnch-d9ba4bd4.pool.mediaWays.net) |
| 09:12.22 | florian | bbl |
| 09:37.03 | *** join/#gpe florian (~fuchs@kc.celox-online.de) |
| 09:45.02 | *** join/#gpe pb_ (~pb@dsl-62-3-66-201.zen.co.uk) |
| 09:45.51 | pb_ | g'day |
| 09:47.22 | florian | hi pb_ |
| 09:47.56 | pb_ | hi florian |
| 10:16.04 | *** join/#gpe nilsf (~nils@kc.celox-online.de) |
| 10:16.10 | nilsf | Moin! |
| 10:16.33 | pb_ | hi nilsf |
| 10:36.51 | *** join/#gpe Speuler (~Speuler@mnch-d9ba4bd4.pool.mediaWays.net) |
| 10:36.51 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) |
| 10:36.51 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 10:36.51 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) |
| 10:36.51 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 10:56.20 | *** join/#gpe Soopaman (~Soopaman@h24-66-54-32.wp.shawcable.net) |
| 11:09.25 | *** join/#gpe andreas_gi (~andreas@L1262P06.dipool.highway.telekom.at) |
| 11:11.05 | florian | nilsf: cool |
| 11:14.34 | nilsf | florian: Partially... |
| 11:14.38 | *** join/#gpe pigeon (~pigeon@203.206.217.82) |
| 11:16.43 | florian | nilsf: What did you expect? |
| 11:19.48 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 11:19.48 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) [NETSPLIT VICTIM] |
| 11:19.48 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 11:19.48 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) [NETSPLIT VICTIM] |
| 11:19.48 | *** join/#gpe Speuler (~Speuler@mnch-d9ba4bd4.pool.mediaWays.net) [NETSPLIT VICTIM] |
| 12:32.23 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 12:32.23 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) [NETSPLIT VICTIM] |
| 12:32.23 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 12:32.23 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) [NETSPLIT VICTIM] |
| 12:32.23 | *** join/#gpe Speuler (~Speuler@mnch-d9ba4bd4.pool.mediaWays.net) [NETSPLIT VICTIM] |
| 12:37.08 | andreas_gi | pb_: hi, ive read yesterday it would be usefull if you would have a siemens simpad sl4 for gpe development? |
| 12:38.20 | pb_ | andreas_gi: yeah, I was trying to find a place I could buy one. |
| 12:38.42 | andreas_gi | pb_: ive checked today, and i can manage that you get a new on for free |
| 12:39.16 | andreas_gi | pb_: its some kind of give back to the community from a company i know |
| 12:39.51 | andreas_gi | pb_: they prefer to give back code and bug reports, but they do that too ... |
| 12:42.57 | andreas_gi | pb_: if its ok for you - there are no expectations - you can do what you want with it |
| 12:43.10 | pb_ | andreas_gi: ah, that's very generous of them. |
| 12:43.13 | pb_ | andreas_gi: thanks! |
| 12:43.28 | andreas_gi | pb_ n.p. - its not my money :) |
| 12:43.35 | pb_ | :-) |
| 12:43.47 | andreas_gi | pb_: if you email me your address - ill send it on monday |
| 12:44.09 | andreas_gi | oh |
| 12:44.59 | nilsf | andreas_gi: Puh ... perfect... also thanks from me! That's a very good initiative! |
| 12:45.04 | nilsf | pb_: Congratulations :) |
| 12:45.27 | pb_ | nilsf: :-) |
| 12:48.06 | florian | woh... |
| 12:49.18 | florian | ipaqsleep didn't power down my ipaq :-/ |
| 12:55.07 | bobf | morning florian, andreas, pb :) |
| 12:55.15 | florian | hi bobf |
| 12:55.28 | andreas_gi | hi |
| 12:56.39 | pb_ | florian: :-/ |
| 12:56.41 | pb_ | hi bobf |
| 12:57.16 | florian | hmm... i upated yesterday.... gpe-dm doesn't come up |
| 12:58.03 | pb_ | ah, yes, tigert mentioned that too. |
| 12:58.06 | pb_ | can you tell what's wrong? |
| 12:58.14 | pb_ | I guess I probably screwed up these Xinit/Xsession changes. |
| 12:58.23 | florian | not yet, but i'll try to find out |
| 12:58.32 | pb_ | cool |
| 13:00.38 | florian | Starting Xinit manually it tells me about an exec format error of the ts-calibrate script... |
| 13:01.19 | florian | ok, it has no exec info inside :-/ |
| 13:02.28 | pb_ | Ah, whoops. |
| 13:02.47 | florian | ok, the same with the screensaver script.... |
| 13:03.23 | pb_ | sounds like I should make a new gpe-session-scripts. |
| 13:04.00 | florian | but this is not all... |
| 13:07.42 | florian | ok, xdefaults script |
| 13:07.57 | florian | and norepeat |
| 13:09.14 | pb_ | ok |
| 13:09.16 | pb_ | any others? |
| 13:10.36 | florian | i think that's it - gpe-dm starts now |
| 13:13.35 | pb_ | right |
| 13:14.34 | florian | wait.... only gpe-dm starts :-/ |
| 13:17.40 | florian | yeah: Xsession.d/10unsetgtkrc |
| 13:19.38 | florian | the same with 20, 30, 50 |
| 13:20.19 | pb_ | can you take a look in CVS and see if those are still broken there? |
| 13:20.28 | pb_ | I think I fixed at least some of them a couple of days ago. |
| 13:20.36 | florian | ok |
| 13:21.37 | pb_ | thanks |
| 13:23.42 | florian | where in cvs are these` |
| 13:23.50 | florian | s/'/? |
| 13:23.56 | pb_ | gpe/base/gpe-login/ |
| 13:24.48 | *** join/#gpe jamin2 (~trillian@zool-logisense.sentex.ca) |
| 13:25.07 | florian | broken: 30xTs_Calibrate |
| 13:25.40 | florian | 55xScreenSaver |
| 13:25.49 | florian | 60xXDefaults |
| 13:26.11 | florian | Only xmodmap is ok.... |
| 13:26.36 | florian | where is the stuff in Xsession.d? |
| 13:34.21 | pb_ | Xsession.d contains old cruft that isn't used. |
| 13:34.28 | pb_ | Only Xsession-gpe.d is important. |
| 13:35.22 | florian | intersting... my xsession-errors tell me about /etc/X11/Xsession.d |
| 13:36.17 | pb_ | Oh, hm. |
| 13:36.22 | pb_ | Maybe Xsession is malfunctioning. |
| 13:44.20 | florian | yeah |
| 13:56.29 | pb_ | oh, it looks like I wrote "-f" where I should have written "-d". doh. |
| 13:57.04 | florian | ok, i see why... much of the stuff in Xsession-gpe.d is linked to Xsession.d |
| 13:59.02 | florian | pb_: are you creating a new package? |
| 13:59.26 | pb_ | florian: yep, one minute |
| 13:59.34 | florian | pb_: great :-) |
| 14:00.20 | pb_ | florian: gpe-dm 0.25 is uploaded, which should fix the -f / -d confusion. |
| 14:00.35 | pb_ | I can't create a new gpe-session-scripts until the xserver package is changed, though. |
| 14:02.19 | florian | :-/ |
| 14:03.06 | *** join/#gpe _fishy (fishy@adsl-81-167-155.asm.bellsouth.net) |
| 14:08.33 | pb_ | florian: want to make a new xserver-tiny-h3600? I think this is just a case of removing /etc/X11/Xsession.d from that package. |
| 14:09.27 | florian | pb_: ok |
| 14:09.33 | pb_ | cool |
| 14:19.02 | florian | pb_: try this: http://handhelds.org/~florian/xserver-tiny-h3600_4.2.20030126-14_arm.ipk |
| 14:20.12 | florian | pb_: if it's ok, i'll upload it... |
| 14:20.30 | pb_ | florian: thanks. I don't have my ipaq on hand right now (it's in my car) but I'll try it within the next half-hour or so. |
| 14:20.48 | florian | pb_: ok |
| 14:29.15 | *** join/#gpe jamin2 (~trillian@zool-logisense.sentex.ca) [NETSPLIT VICTIM] |
| 14:29.15 | *** join/#gpe tigert (~tigert@mopo.tigert.net) [NETSPLIT VICTIM] |
| 14:29.15 | *** join/#gpe moray (moray@mma29.dar.cam.ac.uk) [NETSPLIT VICTIM] |
| 14:29.15 | *** join/#gpe LunohoD (alex@rz-alex.rz.uni-karlsruhe.de) [NETSPLIT VICTIM] |
| 14:29.15 | *** join/#gpe PaxAnima (~petter@b097c.studby.ntnu.no) [NETSPLIT VICTIM] |
| 14:29.15 | *** join/#gpe Speuler (~Speuler@mnch-d9ba4bd4.pool.mediaWays.net) [NETSPLIT VICTIM] |
| 14:36.59 | florian | pb_: Do you know if it's possible to compile memprof for arm? |
| 14:40.22 | bobf | florian: got my basic app to cross compile and run! :) |
| 14:40.55 | bobf | (my contrast adjusting app.. which means my brightness app will also be easy. The setup app might be slightly harder, it's more complicated) |
| 14:41.06 | pb_ | <PROTECTED> |
| 14:41.15 | pb_ | seems that older versions, at least, would build on arm. |
| 14:41.20 | florian | bobf: cool... it starts to become easier if you get used to crosscompiling problems... |
| 14:41.22 | pb_ | I don't know whether they actually worked or not. |
| 14:42.19 | florian | pb_: I tried to compile 0.5.1 on the cats... stack-frame.c:152: error: `stack_frame' undeclared (first use in this function) |
| 14:42.36 | florian | (works with debian/x86) |
| 14:43.24 | *** join/#gpe paul (~paul@hibernia.jakma.org) |
| 14:43.36 | florian | hi paul |
| 14:44.03 | paul | wie gehst florian? |
| 14:44.41 | florian | paul: kann nicht klagen :-) |
| 14:47.15 | pb_ | florian: ah. |
| 14:47.16 | paul | gut :) |
| 14:47.29 | pb_ | <PROTECTED> |
| 14:47.36 | pb_ | yeah, it would appear 0.5.x is i386 only. |
| 14:47.48 | *** part/#gpe jamin2 (~trillian@zool-logisense.sentex.ca) |
| 14:48.39 | florian | These guys rom SCO must be braindead...! |
| 14:50.20 | paul | florian: one possible outcome is that SysVx ends up GPLed! |
| 14:51.16 | florian | paul: yeah, would be cool :-) |
| 14:52.14 | paul | florian: its a possibility. (remote though) |
| 14:53.43 | bobf | florian: Could you apply a patch to GPE's busybox and make a new bootgpe jffs2 image? |
| 14:55.42 | florian | bobf: boot images would be a a job for pb if he is idle... concerning busybox i'd need to check if i'm allowed to write to this part of cvs. |
| 14:56.24 | bobf | Hrm.. I think I'm confused - it's nilsf's /skiff that I would need patched, I guess. |
| 14:56.55 | bobf | I may be able to fix it without that though.. investigating |
| 14:58.11 | florian | bobf: Where is the problem? |
| 14:58.17 | bobf | silly me, fixed it. Just had to define IF_NAMESIZE in my file, because the libc used by the skiff toolchain doesn't have it defined in sockios.h |
| 14:58.23 | florian | s/Where/What |
| 14:58.35 | florian | ok |
| 15:00.29 | bobf | pb_: Could you tar up the contents of a h3600 bootgpe.jffs2 for me? |
| 15:16.26 | pb_ | bobf: I think the 0.7.1 versions are available for download on th familiar site. If you need unstable, I can do that for you later today. |
| 15:17.34 | *** join/#gpe cantona (~cantona@vp165147.hk.uac1.hknet.com) |
| 15:19.03 | bobf | pb_: available as a tar? |
| 15:19.41 | bobf | I know you can get a .tar of the jffs2 image, but I'd like the *contents* of the image, if that's possible? |
| 15:21.21 | bobf | Otherwise I'm stuck making changes on my system, then tar'ing everything up and making my own jffs2 image from that |
| 15:21.51 | bobf | Just figure it'd be easier if I had a tar of the filesystem to begin with, make changes to it, then create my own jffs2 image |
| 15:27.44 | pb_ | bobf: yes, there are .tar.bz2 files of the contents |
| 15:28.16 | bobf | ftp.handhelds.org? |
| 15:29.01 | pb_ | familiar.handhelds.org/releases/v0.7.1/install/files, I think |
| 15:29.06 | bobf | k |
| 15:29.44 | bobf | Awesome, thanks. |
| 15:30.17 | pb_ | Looks like only the 3900 image is there for some reason, but the only difference between the two is the kernel. |
| 15:30.30 | bobf | ah |
| 15:30.49 | bobf | okay, thanks. |
| 16:28.02 | florian | pb_: is the xserver package ok? |
| 16:36.40 | nilsf | Apropos xserver... what about switching the ipaq xserver to tslib? Or is this done already? |
| 16:38.54 | florian | Nope... i just repackaged the standard Xipaq to get rid of some old Xinit.d cruft. |
| 16:39.18 | nilsf | Sure, I just pointed that out as a general maybe good idea :) |
| 16:39.41 | nilsf | The server Matthew built is a good example for it. Makes life a lot easier. |
| 16:39.59 | florian | Yep, this would make at least our life muach easier :-) |
| 17:04.06 | florian | bbl |
| 17:16.46 | *** join/#gpe ade|desk (~adrian@modem-4004.gorilla.dialup.pol.co.uk) |
| 17:31.45 | pb_ | nilsf: I would rather not switch the ipaq xserver to tslib until we have a good way to do touchscreen calibration with tslib. |
| 17:32.29 | bobf | Doesn't tslib do ts_calibrate? |
| 17:32.29 | pb_ | Keith and I spoke about this a few weeks ago, and I have a plan for how to make it work; it's just a case of finding time to write the code. |
| 17:32.46 | pb_ | bobf: Yes, but ts_calibrate doesn't necessarily play well with the X server. |
| 17:32.53 | bobf | Ah. |
| 17:33.00 | nilsf | pb_: Hmm... the calibration program from the tslib worked quite fine. Waht's the problem you see? |
| 17:33.37 | nilsf | *someone changed the OSS sound drivers or L3 ... treble and bass mixer don't work anymore, grmpf* |
| 17:33.57 | pb_ | nilsf: The main problems are the necessarily to switch VCs, the fact that it doesn't inherit the framebuffer rotation from X, and the fact that it looks rather ugly. |
| 17:34.05 | pb_ | s/necessarily/necessity/ |
| 17:34.13 | nilsf | :) |
| 17:34.33 | nilsf | Oh, I see. I just saw it on the SIMpad and I suppose Matthew on his C700... |
| 17:34.55 | pb_ | Yeah, indeed. It certainly works well enough to be better than nothing, but I think it would be a retrograde step on the ipaq. |
| 17:35.32 | nilsf | Yes, the rotation is indeed a problem. The VT switch could be done... but without rotation it's bad. |
| 17:35.56 | pb_ | Most of the same technology that is required to make calibration work would be useful for other things as well, so it seems like something that would be worth implementing. |
| 17:38.16 | bobf | I'm going to have to get some calibration stuff working, probably starting tomorrow. Finished porting my gtk stuff to arm, trying to get my ts working now.. after that is calibration |
| 17:38.33 | pb_ | As an interim (albeit ugly) solution, one could even consider implementing the h3600_ts interface in the kernel for other platforms, which would allow you to use xcalibrate/xtscal with no changes. |
| 17:41.06 | bobf | A guy at ADS says he has a ts working under GPE on an AGX.. so I should be able to get one working too |
| 17:41.15 | pb_ | Cool. |
| 17:43.00 | bobf | Any idea if the ts stuff changed between 0.6 and 0.7? |
| 17:43.46 | pb_ | What ts stuff in particular? |
| 17:44.29 | bobf | The guy at ADS might be using 0.6, he had some problems with 0.7 - I'm trying to get it under 0.7, just wondering it that could be the difference. |
| 17:44.58 | pb_ | Ah, I see. I don't think so, no. |
| 17:47.35 | bobf | k |
| 17:51.23 | *** join/#gpe andreas_gi (~andreas@L1244P19.dipool.highway.telekom.at) |
| 18:11.36 | *** join/#gpe florian (~fuchs@p509296B8.dip.t-dialin.net) |
| 18:11.39 | florian | re |
| 18:12.06 | bobf | wb florian |
| 18:12.49 | *** join/#gpe BBrox (~ulmer@bbrox.org) |
| 18:24.55 | nilsf | I don'tg get it... setting volume works, but setting bass/treble doesn't anymore :(( |
| 18:35.45 | florian | nilsf: Tried "cvs diff" to find out what changed? |
| 18:36.31 | nilsf | florian: If I only knew when it changed :( |
| 18:37.10 | nilsf | Oh perfect, alsa drivers do not cross-compile, as usual :( |
| 18:37.51 | florian | nilsf: yep, looks like jamey is the only one who knows how to build.. |
| 18:39.07 | nilsf | Something is very stupidly broken here.... either me or this strange source... |
| 18:51.05 | *** join/#gpe pb_ (~pb@pc2-cmbg4-3-cust239.cmbg.cable.ntl.com) |
| 18:52.07 | pb_ | re |
| 18:52.09 | pb_ | florian: your xserver package seems to work fine. |
| 18:52.29 | nilsf | pb_: Do you have any idea what has recently changed in kernel that can cause just the treble and bass mixer of the OSS driver to fail? Volume works!? |
| 18:52.40 | pb_ | nilsf: hm, no idea. What version of the kernel are you using? |
| 18:52.57 | nilsf | CVS head I think, hh20. |
| 18:53.06 | nilsf | Yep. |
| 18:53.34 | nilsf | Very strange... I also cannot find anything in any diff that could explain it. |
| 18:53.44 | pb_ | Do you know what the last one that worked was? |
| 18:53.58 | nilsf | Very good question :) |
| 18:54.11 | nilsf | It worked when I checked it in back in April. |
| 18:54.40 | pb_ | Ah, so quite some time. Though I don't think the OSS driver has changed much, if at all, in the last few months. |
| 18:54.58 | nilsf | I suppose it has something to do with the L3 handling- |
| 18:55.04 | pb_ | Most new work is done on ALSA, it seems. |
| 18:55.07 | pb_ | Yes, I guess so. |
| 18:55.19 | nilsf | Also... gmpf. |
| 18:55.26 | nilsf | s/Also/Alsa |
| 18:55.39 | nilsf | I could live with Alsa, but... |
| 18:55.44 | nilsf | I cannot cross-compile it. |
| 18:55.48 | florian | pb_: very nice... finally made everything run manually - only keylaunch is strange, suspending by the power button works, but it doesn't wake up anymore :-/ |
| 18:55.51 | nilsf | It does stutter when playing from MMC. |
| 18:55.54 | pb_ | Well, there are no alsa modules for hh20 yet, so this doesn't really help you. |
| 18:56.29 | pb_ | florian: Hm, that's very weird. Waking up is a kernel function: keylaunch doesn't get involved there. |
| 18:56.56 | pb_ | nilsf: I think the audio player needs to request a larger buffer size, or higher scheduling priority. |
| 18:57.41 | pb_ | It might be that the default ALSA settings are tuned more towards low latency, since the original reason for switching to alsa was to improve performance in telephony applications. |
| 18:57.48 | *** join/#gpe ade|home (~adrian@modem-4004.gorilla.dialup.pol.co.uk) |
| 18:57.52 | florian | pb_: yeah sure - maybe this problem starts while suspending. i'll install the new session scripts whaen apailable and check what happens... |
| 18:58.10 | nilsf | pb_: I would deny higher priority since the OSS does without. Bigger buffer request, would this be an Alsa buffer? |
| 18:58.47 | pb_ | nilsf: Well, the two things are alternatives. With a larger buffer, scheduling priority is not so important, and vice versa. |
| 18:59.17 | pb_ | Yes, it would probably be an Alsa buffer, though you would still request it through the OSS API if you were using the compatibility layer. |
| 19:00.06 | nilsf | pb_: Hmm... I always thought that a whatever big DMA buffer would do the trick and that you can stop the sound engine whenever you want which should allow low latency. |
| 19:00.39 | pb_ | nilsf: Hm, I'm not sure I understand what you're saying. |
| 19:01.37 | pb_ | florian: Okay. I'm building the new package now. |
| 19:02.11 | florian | pb_: cool |
| 19:03.29 | nilsf | pb_: I mean you have different buffers. One in userspace, maybe one in kernel and one DMA buffer the DMA engine cycles through. |
| 19:03.33 | pb_ | florian: done |
| 19:03.44 | nilsf | pb_: Making the DMA buffer larger should not affect latency. |
| 19:05.31 | pb_ | nilsf: How come? Surely if there is a large amount of data in the DMA buffer, this will invariably lead to higher latency. |
| 19:05.55 | nilsf | I'll try chekiing out K2-4-18-rmk6-pxa1-hh19 now :( |
| 19:06.02 | pb_ | Okay. |
| 19:06.25 | nilsf | pb_: Hmm... but audio is a real-time thing. Either you manage to capture / play all data from the buffer or not. |
| 19:07.51 | nilsf | pb_: Hmm... at least for playback it should not matter. For recording you are right I think. |
| 19:09.23 | nilsf | Can CVS also downgrade using date tags? I mean, can it reverse patch a tree? That would make it easier for me to track the bad change back... |
| 19:09.43 | pb_ | nilsf: Sure, just use cvs update with the date or tag you want. |
| 19:09.58 | nilsf | cvs is cool :)) |
| 19:11.28 | nilsf | pb_: But anyway I think in both cases the size of the DMA buffer should not affect the latency. I you see those effects then your application is too slow, i.e. your buffer fills and you lag back. |
| 19:11.38 | nilsf | s/I/If |
| 19:11.50 | florian | nilsf: Our CAT rulez ;-) |
| 19:12.09 | nilsf | florian: Ah, you fill our DSL line :) |
| 19:12.44 | florian | nilsf: not anymore... :-) |
| 19:12.52 | nilsf | florian: But now me :) |
| 19:14.13 | florian | nilsf: Seems to work fine - my shell is still usable... |
| 19:15.11 | florian | nilsf: We could install the arnold box at university... would be a nice platform :-) |
| 19:16.40 | florian | Last login: Thu Jul 31 10:30:42 2003 from fw64.hrz.uni-siegen.de |
| 19:17.19 | pb_ | florian: Oh dear, now gpe-login is segfaulting. |
| 19:17.56 | florian | pb_: nasty... any idea why? |
| 19:18.07 | pb_ | florian: Yeah, it is my fault. |
| 19:18.32 | nilsf | Is there still some secret button combination to switch an H38xx off completely? Once there was this all-four-button+reset thing... |
| 19:21.23 | *** join/#gpe reenoo (~r@p50867799.dip.t-dialin.net) |
| 19:21.30 | reenoo | hi |
| 19:22.44 | florian | reenoo: hi |
| 19:22.58 | nilsf | Where is u32 u16 and such defined for kernel modules? Alsa complains... |
| 19:23.48 | pb_ | <linux/types.h>, probably. |
| 19:24.07 | nilsf | hmm it includes asm/types already... |
| 19:25.26 | nilsf | Nope, still no go. |
| 19:31.02 | florian | Is there a preferred gpg keyserver? |
| 19:31.16 | pb_ | Nope, any should do. I use the MIT one. |
| 19:34.44 | pb_ | I uploaded another gpe-login which should fix the segfault. |
| 19:35.00 | pb_ | However, I am still not able to log in, owing to missing #!/bin/sh in various scripts provided by other packages. :-/ |
| 19:35.54 | florian | blackhole.pca.dfn.de |
| 19:36.23 | florian | oh, bad... many packages? |
| 19:36.52 | pb_ | not sure. keylaunch, ipaq-sleep, startup-monitor all have this problem. |
| 19:36.57 | pb_ | there might be others too. |
| 19:37.09 | pb_ | I think we need to make run-parts less fussy. I'll try to patch busybox. |
| 19:37.32 | florian | looks like my key reached this keyserver... |
| 19:38.13 | florian | pb_: yep, but we should file bugs against all these packages... bugzilla doesn't forget :-) |
| 19:39.48 | pb_ | florian: right! |
| 19:42.18 | pb_ | ok, new busybox is in unstable. |
| 19:42.30 | pb_ | and now I can log in again! :-) |
| 19:42.59 | bobf | How do I get a touchscreen to calibrate? |
| 19:43.09 | bobf | I run xcalibrate, it works.. gets values, but where do I put them? |
| 19:43.31 | pb_ | bobf: in the h3600 driver model, calibration is done in the kernel. |
| 19:43.43 | bobf | I don't use the h3600 driver ;) |
| 19:43.58 | pb_ | bobf: then I am surprised you can use xcalibrate. |
| 19:44.08 | bobf | It involved some work.. ;) |
| 19:44.38 | pb_ | You could try xtscal, which can be made to emit values that you can put in a tslib /etc/pointercal file. |
| 19:44.41 | bobf | So I can use xcalibrate, it gets appropriate values.. is it supposed to then be automatically calibrated? |
| 19:44.58 | pb_ | Yes. |
| 19:45.05 | *** join/#gpe CAMCAM (~camrein@AMontsouris-108-1-21-151.w80-13.abo.wanadoo.fr) |
| 19:45.16 | bobf | Hrm. |
| 19:45.29 | pb_ | Xcalibrate performs ioctls to set the appropriate calibration in the kernel. |
| 19:46.38 | bobf | My xcalibrate seems like it never sets anything, it's just re-looping through the crosshairs |
| 19:46.54 | pb_ | Ah. In that case, it must be receiving duff information from the kernel. |
| 19:47.27 | bobf | But it does get the values (and they do change, based on where you touch..) |
| 19:47.37 | pb_ | It will loop if the calibration that it calculates doesn't seem reasonable (ie the five points don't delineate the corners and centre of a rectangle). |
| 19:48.06 | bobf | Interesting |
| 19:56.02 | nilsf | Hmm... I tried hh19a, already broken mixers (and broken MMC)... so I'll go more back... |
| 19:58.43 | reenoo | nilsf: just to clarify... I used to think that a user space program can't access a parport LCD via direct I/O calls (inb/outb) if the port is occupied by a kernel driver already. Is that correct? |
| 20:00.37 | nilsf | reenoo: At least it should not be able to, yes. The I/O addresses should be blocked. It might be though that you can still do it since you do not have to allocate them in userspace. |
| 20:00.52 | pb_ | reenoo: the kernel driver can't do anything to stop you using inb and outb, though obviously the results will not be pretty if both drivers try to drive the same hardware. You should probably use ppdev if you want to beat on parallel ports from userspace though. |
| 20:02.03 | nilsf | pb_: But don't you also have to register the I/O addresses you want to use using ioperm() ? |
| 20:02.28 | reenoo | well... LCDd returns bus errors due to the fact that more than one program/kernel driver tries to address the same hardware |
| 20:02.29 | pb_ | nilsf: Yes, but there's no checking that they aren't already in use. |
| 20:02.50 | nilsf | pb_: I would expect ioperm() to fail if a driver has already regsitered them... |
| 20:02.50 | reenoo | and ppdev is not an option really |
| 20:03.01 | pb_ | nilsf: All that ioperm() does is map the port into the application's address space. |
| 20:03.34 | reenoo | ppdev is not availably on non linux systems and also too slow |
| 20:03.54 | pb_ | Applications can also call iopl() to request unrestricted access to all ISA ports. Again there is no checking. |
| 20:04.40 | nilsf | pb_: Hmm... this should be added :) |
| 20:05.22 | nilsf | reenoo: But I had exactly that case today... someone trying lcdproc while having parport_pc in kernel :( |
| 20:05.50 | pb_ | nilsf: It's not really practical to retro-fit the checks, because applications that use iopl() or ioperm(0,0x7ff) would always fail if any ports at all were in use. |
| 20:06.16 | pb_ | I suppose we could have ioperm() return an error if the entire requested range was already reserved. |
| 20:06.16 | nilsf | pb_: Then those apps are ill designed :) |
| 20:06.22 | reenoo | nilsf: heh... /me too |
| 20:06.52 | pb_ | reenoo: you don't have to do all your I/O though ppdev, just use it to lock out other in-kernel use of the port. |
| 20:06.52 | reenoo | nilsf: except it was FreeBSD's ppc |
| 20:07.06 | reenoo | pb_: oh neat |
| 20:08.26 | reenoo | hmm... no inb/outb man page on FreeBSD ?! |
| 20:08.31 | nilsf | Why does scp let me enter my passowrd and only checks for the existance of the file I want to copy (to a remote) afterwards!? |
| 20:09.02 | bobf | That's kind of silly |
| 20:09.05 | reenoo | nilsf: because it's scp ;) |
| 20:09.17 | nilsf | reenoo: Oh thanks, that helps a lot :) |
| 20:09.34 | reenoo | nilsf: hehehe |
| 20:09.37 | florian | pb_: just noticed this strange suspend behavior again: the ipaq wakes up and goes to sleep again... like it does on flat battery... after this it stays off. |
| 20:11.13 | florian | nilsf: right - scp is not really well designed... |
| 20:11.31 | nilsf | At least there is a glitch... |
| 20:12.18 | nilsf | Since when is MMC broken? hh19a was broken and hh19 is so too? |
| 20:12.18 | florian | eeks! |
| 20:13.35 | nilsf | OK, hh19 mixer also broken... going back further :( |
| 20:14.29 | florian | i think since 19a |
| 20:15.28 | pb_ | nilsf: I think MMC was broken around hh19 when the first SD patch was introduced. |
| 20:15.45 | nilsf | pb_: Seems so :) |
| 20:16.08 | nilsf | Now trying hh18... |
| 20:16.15 | pb_ | nilsf: btw, hh19 and hh19a are identical code. |
| 20:18.14 | florian | scap |
| 20:18.58 | pb_ | florian: hm? |
| 20:19.13 | florian | http://handhelds.org/scap/port.7754.png |
| 20:19.32 | pb_ | heh |
| 20:20.05 | pb_ | seems like an interesting choice of font. I guess your Xdefaults has gone missing. |
| 20:20.27 | florian | yeah... but why? |
| 20:20.37 | pb_ | probably some further Xsession.d related lossage. |
| 20:21.12 | pb_ | I did notice some problems that looked like they were caused by /usr/X11R6/bin missing from $PATH when the session scripts ran. |
| 20:21.26 | pb_ | <PROTECTED> |
| 20:21.28 | pb_ | that kind of thing |
| 20:22.08 | pb_ | I suppose this was the reason for having Xinit source /etc/profile. |
| 20:22.14 | florian | heh?! busybox postinst didn't find update-rc.d |
| 20:22.24 | pb_ | crumbs. |
| 20:22.44 | pb_ | oh, make sure /sbin is in your path. |
| 20:22.51 | pb_ | that might be another effect of the same thing. |
| 20:23.37 | florian | profile says /sbin is in my path |
| 20:23.46 | pb_ | check with echo $PATH. |
| 20:23.52 | pb_ | it might be that profile is never being sources now. |
| 20:23.57 | pb_ | sourced, that is |
| 20:24.33 | florian | it's definitely there... |
| 20:24.40 | reenoo | JFYI: http://lists.omnipotent.net/pipermail/lcdproc/2003-August/008063.html |
| 20:24.41 | pb_ | weird |
| 20:25.18 | florian | /etc/X11/Xdefaults too, including a rxvt font setting. |
| 20:25.38 | pb_ | check with xrdb to see if Xdefaults was installed. |
| 20:28.11 | florian | i'd like to... but i made a fatal mistake: i logged out... |
| 20:28.46 | bobf | CHECK with Center Coords (320,240): xraw=1865 yraw=1690 |
| 20:28.46 | bobf | CHECK: x(center)=326 y(center)=-3376 |
| 20:28.46 | bobf | off(6,-3616) Badness = 13075492 |
| 20:29.06 | bobf | pb_: So that would be the indication that it isn't working? |
| 20:29.21 | pb_ | bobf: yeah |
| 20:30.17 | bobf | the off value is how much it's off on x,y? |
| 20:31.30 | florian | i used gpe-bluetooth for an ip connection, the serial cradle is at kc and login doesn't work :-( |
| 20:31.46 | pb_ | bobf: yep |
| 20:31.53 | bobf | Hrm.. "center coords" (320,240) might be causing the trouble? It's a 640x480 display |
| 20:31.55 | pb_ | florian: :-( |
| 20:32.22 | pb_ | bobf: (320,240) sounds about right for the center of a 640x480 screen. |
| 20:32.28 | bobf | Oh.. duh. ;p |
| 20:32.49 | pb_ | bobf: but xcalibrate is a crock and it would not surprise me if it is failing in some obscure way due to your geometry. |
| 20:33.21 | florian | ahhh! root login works :-) |
| 20:34.19 | bobf | I see. |
| 20:34.36 | pb_ | florian: phew! |
| 20:34.48 | florian | Isn't jg's staement about warranty worth a story on hh.org? |
| 20:35.04 | florian | s/ae/ate/ |
| 20:35.53 | pb_ | florian: Yes, could be. Maybe leave him a botmail and suggest that he adds it. |
| 20:36.01 | bobf | Well, I'm going to head out.. fiancee wants me home for dinner. See you all tomorrow. |
| 20:36.07 | pb_ | bobf: later |
| 20:36.26 | florian | bobf: yep, cu |
| 20:42.37 | florian | how large is libstdc++? |
| 20:47.22 | florian | pb_: some errors from .xsession-errors: |
| 20:48.13 | florian | /usr/bin/xrandr-panelapp.sh: 9: xrandr: not found |
| 20:48.22 | florian | /usr/bin/xrandr-panelapp.sh: 9: xmodmap: not found |
| 20:48.51 | florian | run-parts: component /etc/X11/Xsession.d/99xCommon |
| 20:49.06 | pb_ | florian: looks like more $PATH problems |
| 20:49.49 | nilsf | pb_: Oh well... hh13 works again... |
| 20:50.40 | nilsf | So somewhere between hh13 and hh19 something bad happened that affects the ols OSS driver... |
| 20:51.25 | pb_ | nilsf: try cvs diff -up -rK2-4-19-rmk6-pxa1-hh13 -rK2-4-19-rmk6-pxa1-hh19 drivers/sound |
| 20:51.54 | nilsf | What is -up ? |
| 20:52.17 | pb_ | shorthand for -u -p -- just the output format specifiers that I happen to prefer. |
| 20:52.26 | nilsf | Ah .) |
| 20:52.29 | pb_ | -u is unidiff; -p makes it prefix each hunk with the function it comes from. |
| 20:53.45 | nilsf | pb_: But I do not think this will really help (I'll try anyway). I think some L3 thing must be bad... |
| 20:54.26 | nilsf | And some additional stuff must have changed. I noticed that the volume is reset after every close->open to some insane high default. |
| 20:55.25 | florian | Is someone from the US east present? |
| 20:56.51 | florian | pb_: Where should PATH be set? |
| 20:57.12 | *** join/#gpe ade|beth2 (~adrian@modem-407.jaguar.dialup.pol.co.uk) |
| 20:59.00 | pb_ | florian: I'm not sure. Maybe in Xinit. |
| 20:59.36 | nilsf | pb_: The changes are marginal and almost do not affect non-PXA machines... strange... |
| 21:00.16 | nilsf | Have to go, CU 2morrow. |
| 21:00.16 | florian | nilsf: sure... cu |
| 21:00.43 | florian | pb_: in a shell everything is ok - but i think this takes it from /etc/profile |
| 21:03.13 | pb_ | florian: right, yeah. |
| 21:18.30 | florian | What does "vcard-normalize"? |
| 21:40.48 | *** part/#gpe herzi (~herzi@kiwi.mediascape.de) |
| 21:42.57 | *** join/#gpe HET2 (~diman@chello080110092229.507.15.vie.surfer.at) |
| 22:12.30 | florian | night all! |
| 22:13.41 | *** join/#gpe pb_ (~pb@pc2-cmbg4-3-cust239.cmbg.cable.ntl.com) |
| 23:51.11 | *** join/#gpe flypiper (~deb-user@12-222-141-210.client.insightBB.com) |
| 23:51.38 | *** join/#gpe _fishy (fishy@adsl-81-160-137.asm.bellsouth.net) |
| 23:52.07 | pigeon | hi fishy, nice nick :) |
