| 00:02:00 | prpplague | GoRK: eewww, bowels doesn't sound pretty |
| 00:02:09 | prpplague | howdy Tangent |
| 00:02:15 | prpplague | hows the UK today? |
| 00:02:26 | Tangent | prpplague : Dark and cold.. It's night :) |
| 00:02:45 | prpplague | hmm time for a yard of guinness then |
| 00:02:49 | Tangent | prpplague : Weather was actually pretty good, but I spent today in my windowless offfice playing computer games with a mate |
| 00:03:12 | prpplague | ahh |
| 00:19:03 | prpplague | hey! whats the deal with the wiki changes! |
| 00:19:25 | prpplague | "handheld people" ?? |
| 00:21:48 | Tangent | prpplague : Hrm.. yes... that one could do with being deleted I think |
| 00:27:47 | prpplague | Tangent: i drop bzflag an email |
| 00:38:39 | prpplague | later folks, off to the house |
| 00:51:01 | Russ | GoRK: wheaties/serialflash need some fixing |
| 00:51:19 | Russ | GoRK: they need to use asm/arch/shannon.h, and they need scripts/Lindent |
| 01:29:42 | as_ | Evnin all. |
| 01:30:00 | jacques | I just got 2 129MB CF cards |
| 01:30:05 | jacques | oops 128MB |
| 01:30:16 | jacques | heh, actually, 129MB is kinda accurate: |
| 01:30:31 | jacques | Nov 25 18:26:33 meccano kernel: hdg: 251904 sectors (129 MB) w/1KiB Cache, CHS=984/8/32 |
| 01:31:56 | jacques | hi Russ, any new kernel patches yet :-D |
| 01:37:33 | jacques | I just realized how stupid it is to use reiserfs on my CF cards |
| 01:37:53 | jacques | considering the fixed size 32MB JOURNAL FILE |
| 01:37:56 | jacques | arg!!! |
| 01:39:10 | jacques | i would rather use jffs2 :-( but I get that was never finished |
| 01:41:41 | jacques | anyone have any favorite ext2 settings for CF? |
| 01:45:15 | spse__ | jacques, cant you use ext3? |
| 01:45:36 | jacques | spse__: hey, that's a good point - I never even considered it |
| 01:45:57 | Tangent | jacques : Yeah.. use ext3 |
| 01:45:57 | spse__ | i think the journal size is dependant on the partition size |
| 01:46:22 | jacques | spse__: you mean in ext3? that's good. I cant understand the stupid fixed journal size in reiserfs |
| 01:46:24 | spse__ | my 260meg / partition has a journal size of 8meg |
| 01:46:44 | spse__ | probably easier to code |
| 01:47:05 | spse__ | i use ext3 on my desktop machine on the paritions i had to leave mounted whilst i converted them from ext2 |
| 01:48:20 | jacques | hmm, I notice ext3 is not supported in my tuxscreen kernels |
| 01:48:30 | jacques | i guess I will have to build it in |
| 01:48:41 | jacques | along with adding support to my desktop |
| 01:48:53 | Tangent | jacques : And emailing the new kernel to me |
| 01:48:56 | Tangent | :D |
| 01:49:14 | jacques | Tangent: hmmm, are you implying that it will be a difficult task? |
| 01:49:26 | spse__ | no, Tangent is implying that he is very lazy |
| 01:49:37 | jacques | ah, that's all right then |
| 01:49:51 | Tangent | I'm not implying it... I thought it was a well accepted fact |
| 01:50:47 | jacques | Tangent: which kernel do you want? I have a working 2.4.15pre6-etc but they changed the ts driver and microwindows doesn't work anymore |
| 01:50:59 | jacques | (well working minus the ext3 stuff) |
| 01:51:06 | Tangent | jacques : Don't worry about it for now.... |
| 01:51:17 | spse__ | you could use that tscal drive SpaceCoaster wrote |
| 01:51:21 | Tangent | jacques : It'll probably be a couple of weeks before I get my CF card plugged into the tux |
| 01:51:43 | Tangent | jacques : I'm wearing 'python monkey' hat today |
| 01:52:01 | jacques | Tangent: :-) |
| 01:52:09 | EGg0 | jacques: Change the reiserfs journal size by recompiling the module... http://www.namesys.com/faq.html#small-device |
| 01:52:21 | Sammy | morning |
| 01:52:28 | Tangent | Sammy : hi |
| 01:52:44 | jacques | EGg0: oooh, aaah. never seen that |
| 01:52:58 | Sammy | hello Tangent |
| 01:54:06 | jacques | heh heh since I am lazy too, I may just recompile with different reiserfs params |
| 01:54:12 | jacques | :-D |
| 02:05:26 | EGg0 | is away: Running my Dell diagnostics programs to troubleshoot this serial port failure. |
| 02:05:38 | jacques | WTF?? looking at the ext3 home page |
| 02:06:05 | jacques | it's either way out of date or ext3 is not ready for prime time |
| 02:06:19 | jacques | The ext3-0.0.3b kernel version inncludes ext3 code which is functionally complete except for one detail: it currently implements only "full-data journaling". In other words, all filesystem writes --- including new data --- get written to the journal, so effectively it ends up writing twice as much data as it needs to. This reduces write performance under sustained load by up to 50% relative to ext2. |
| 02:07:18 | jacques | Ext3 has much stricter restrictions on the order in which it writes data to disk than ext2 has. This is necessary to provide the guarantees that on a crash, the contents of the disk always represent a recoverable state. As a result, ext3 cannot always write its modified cached data out to disk immediately if the kernel becomes low on memory. |
| 02:07:18 | jacques | It is advisable to make sure that you have enough physical memory in your system for ext3 to cache all of its journal files in the very worst case. |
| 02:07:32 | jacques | none of this is good |
| 02:07:48 | jacques | latter esp. on a machine with 16MB RAM |
| 02:09:40 | jacques | i guess that page is criminally out of date |
| 02:09:49 | jacques | it's the one linked off of freshmeat |
| 02:10:43 | jacques | GREAT |
| 02:11:01 | jacques | getting permission denied going to the download site |
| 02:11:43 | jacques | where do I find ext3 tools? |
| 02:11:50 | jacques | I need a mkfs.ext3 |
| 02:12:50 | spse__ | you need e2fsprogs |
| 02:13:22 | spse__ | http://e2fsprogs.sf.net |
| 02:13:45 | spse__ | then use mke2fs with the -j option |
| 02:14:20 | jacques | spse__: thanks. |
| 02:18:46 | jacques | hmm, when did ext3 make it into the mainstream kernel? |
| 02:18:55 | jacques | doesn't appear to be in 2.4.10 |
| 02:19:03 | spse__ | 2.4.15 i think |
| 02:20:05 | jacques | heh, well I'm not putting that on my machine |
| 02:20:10 | jacques | maybe 2.4.16pre1 |
| 02:21:53 | spse__ | jacques: i thought ext3 was a bit memory hungry for your tuxscreen |
| 02:23:37 | jacques | spse__: the info I quoteed above appears to be very old - though I don't know if they've fixed those issues |
| 02:23:53 | jacques | ooh, pcmcia-cs 3.1.30 is out |
| 02:24:00 | spse__ | cool |
| 02:24:54 | jacques | looks like some good changes |
| 03:07:17 | jacques | ARG |
| 03:07:36 | jacques | where are the reiserfs tunable parameters defined? |
| 03:07:49 | jacques | I am grepping the code and finding nothing |
| 03:09:17 | jacques | ah, I think I found them in reiserfsprogs |
| 03:10:48 | jacques | but the mentioned JOURNAL_MAX_BATCH is nowhere to be found |
| 03:11:20 | jacques | do any of these damn journalling filesystems actually have uptodate documentation??? |
| 03:20:11 | Keiser_Soze | did you try www.namesys.com? |
| 03:21:04 | EGg0 | jacques: Did you look in ../include/reiser_fs.h and the others? |
| 03:33:41 | GoRK | Russ: i think i can handle those. apparently something else in ucb1x00-audio is messed up too. recording does *not* work really at all.. :( |
| 03:35:41 | GoRK | which is odd since the mcp divisor is the same for both playback and recording and playback works fine |
| 03:39:54 | GoRK | Russ: do they need to use shannon.h in lieu of their own .h files? ie should the #define's from wheaties.h and spi.h go into shannon.h since they are sa1100 gpios? |
| 03:40:36 | GoRK | spi needs its own .h somewhere since wheaties.c and serial_flash.c need to call functions from it. where does that belong? |
| 03:41:06 | GoRK | err i guess maybe they can just use extern and assume they are going to exist in the finished product :) |
| 03:43:53 | EGg0 | is away: Going to bed. |
| 06:07:06 | Russ | GoRK: yah, use extern, and the shannon.h |
| 06:07:17 | GoRK | Russ: cool i will.. |
| 06:07:34 | GoRK | Russ: im still having hell with ucb1x00-audio... i think there is something else seriously wrong with it |
| 06:07:56 | GoRK | Russ: i dont suppose you have built a shanip board? :) |
| 06:17:26 | Russ | no time |
| 06:19:03 | Russ | I gotta get the backlight stuff cleaner |
| 06:19:19 | Russ | right now, I can see where both a user and pm would want to change the backlight at the same time |
| 06:19:38 | Russ | I can sleep the user, but I don't think I can sleep pm hooks |
| 06:20:21 | Russ | but since there are only 3 settings, I think I can just queue them, overwriting older settings if newer ones come in before they get set |
| 06:21:07 | GoRK | that sounds like it works fine to me... plus if the client wants to set it at all it's probably overridden all the PM stuff anyway |
| 06:21:26 | Russ | should it though? |
| 06:21:39 | Russ | if pm says off, should it stay that way? |
| 06:21:46 | GoRK | well the backlight anyway |
| 06:22:22 | GoRK | if pm turns the backlight off, then the user app should be able to wake it up... take for instance lighting up the screen if the phone rings |
| 06:22:49 | Russ | once I get done, it would be really clean to make a module that listens for input events, and modifies the backlight |
| 06:22:59 | Russ | (ie, make the screen+/- keys work) |
| 06:23:38 | Russ | away |
| 06:23:46 | GoRK | sure... just so long as there is a way to disable screen blanking and override everything from userspace |
| 07:33:37 | GoRK | ooh a superhero! heh |
| 07:39:19 | Russ | heh |
| 07:39:34 | GoRK | i think this audio junk might have to do with the gain setting on ucb1x00 |
| 07:39:56 | GoRK | the "old" driver sets gain, the new one doesnt. the new drivers record level is horribly weak. |
| 07:40:22 | GoRK | i think that mabye the lucent folks had to use that op-amp because they didnt fiddle with the gain register |
| 07:41:26 | Russ | maybe you need to implement /dev/mixer |
| 07:43:14 | GoRK | villany! |
| 07:45:51 | GoRK | there would be only one setting anyway i think... let me go read the datasheet |
| 07:47:57 | GoRK | ahh 2 settings for audio interface anyway.... input gain and output attenuation |
| 07:49:20 | GoRK | there are more bugs in rmk's driver anyway... like the telecom divisor can only be between 16 and 127 and he allows 6 - 127 |
| 07:49:49 | Russ | you can email the patches directly to the patch system if you want |
| 07:50:21 | GoRK | yeah i think ill do that the next time |
| 07:50:39 | GoRK | now that we have a base to stand on |
| 07:51:49 | GoRK | i started working on the wheaties/serial flash but decided i better wait to talk to spacecoaster... probably ought to get wheaties.c using spi.c for its routines like he intended before we send to rmk on that one |
| 07:57:44 | GoRK | now to crank the gain as high as it goes and see how it works :) |
| 08:06:03 | Russ | run it through scripts/Lindent when you get a chance though |
| 08:06:17 | GoRK | of course |
| 08:09:20 | GoRK | ok do you think one mixer device for both /dev/dsp and /dev/dsp1 (the sblive and others use this) or do you think seperate mixers? |
| 08:12:41 | GoRK | well it is certainly louder but it is also decidedly NOT what is being recorded :) |
| 08:14:54 | GoRK | whoops nevermind... telecom register has no controls anyway |
| 08:21:58 | GoRK | hmm ok the old driver didnt use dma, the new one does... and it doesnt work |
| 08:27:09 | GoRK | it almost works.... like it's reading the bits offset or something |
| 08:27:55 | GoRK | hmm that sounds like a good explanation, GoRK. I'll go check that out |
| 08:28:02 | GoRK | sure no problem |
| 08:28:41 | GoRK | good thing i have ctags and use vi! |
| 08:45:14 | GoRK | Russ: could our DMA buffers for audio be in a different place than on itsy/yopy or something for audio read, telecom read/write ? |
| 09:01:15 | GoRK | hmm i dont suppose our friends at lucent have an off-chip clock driving the MCP do they? |
| 09:30:21 | GoRK | lalalala maybe got it |
| 09:36:23 | GoRK | HAHAHAHA |
| 09:39:40 | GoRK | ok now my application is segfaulting :( |
| 09:47:02 | GoRK | haha now i can record but not play back |
| 10:04:14 | GoRK | yo |
| 10:04:15 | GoRK | what's up |
| 10:06:43 | andersee | just me |
| 10:06:51 | GoRK | heh |
| 10:06:56 | andersee | I fell asleep early. Just woke up |
| 10:07:05 | GoRK | i am trying desperately to hack on the ucb1x00-audio driver |
| 10:07:07 | GoRK | it is braindead |
| 10:07:18 | andersee | just you your a-l-k post |
| 10:07:22 | andersee | saw |
| 10:07:30 | andersee | yawms |
| 10:07:36 | GoRK | i got voip stuff working today... but its useless w/ a broken driver :) |
| 10:07:56 | andersee | cool. |
| 10:07:57 | GoRK | yeah i have a followup now. i set a bit on MCP and now i get recording but broken playback |
| 10:08:02 | andersee | (re working voip) |
| 10:08:41 | andersee | One of the undocumented DSP bits? |
| 10:08:49 | GoRK | thanks. it's pretty cool actually :) i have gotten it more or less half-duplex both ways |
| 10:09:00 | GoRK | nono... MCP mode of serial 4 on SA1100 |
| 10:09:14 | GoRK | this is related to my a-l-k post |
| 10:10:44 | GoRK | i set a bit that causes recieved audio data to be dumped even when the valid bit is not set... im guessing our particular ucb1200 chip doesnt set that properly... but it still breaks playback :-/ |
| 10:12:26 | Sammy | sorry, one silly question , where to get the mafs.jffs2 command ? |
| 10:12:49 | GoRK | mtd cvs or buildroot-tux will download it and build it for you |
| 10:14:11 | GoRK | andersee: i just sent my followup... replying to yourself with helpful information makes people falsely assume there is a discussion happening and so they will jump in :) (that is the hope i guess) |
| 10:14:20 | andersee | heh |
| 10:15:28 | GoRK | some guy from india wrote me that he wrote a DMA using driver that worked on ucb1300 that he's going to mail me to have a look at |
| 10:23:34 | andersee | cool |
| 10:45:26 | GoRK | nite |
| 10:45:44 | andersee | night |
| 15:05:47 | prpplague | morning all |
| 15:05:54 | prpplague | andersee: in a little early today? |
| 15:10:48 | andersee | heh |
| 15:10:54 | andersee | Yeah, an early start.... |
| 15:11:04 | andersee | I got up about 3am actually. |
| 15:11:17 | andersee | my schedule is pretty messed |
| 15:18:33 | pattieja | is away: meeting |
| 15:32:00 | EGg0_in|out | is back (gone 11:47:55) |
| 16:20:14 | EGg0_in|out | is away: school |
| 16:38:45 | BZFlag | crap crap crap crap crap |
| 16:43:58 | prpplague | BZFlag: when life gives ya crap, just flush it |
| 16:45:22 | BZFlag | the wiki restore script bombed. |
| 16:45:32 | BZFlag | I just reverted a ton of pages by mistake. |
| 16:45:47 | BZFlag | crap crap crap crap crap |
| 16:47:57 | prpplague | darn |
| 16:48:22 | prpplague | this related to that jerk who messed with the pages? |
| 16:48:47 | BZFlag | yep. |
| 16:48:56 | BZFlag | and a broken restore script. |
| 16:49:14 | BZFlag | It picked the wrong backup, and I nuked the current before noticing. |
| 16:49:21 | prpplague | doh |
| 16:49:36 | BZFlag | crap crap crap crap crap |
| 16:54:13 | AlHazWork | someone messed up the wiki? |
| 16:54:56 | boodle | BZFlag: might google have a decent cache of it? |
| 17:09:08 | BZFlag | yes, someone want to suck google's cache? |
| 17:09:17 | BZFlag | I just restored TimRiker from there. |
| 17:10:41 | BZFlag | http://www.google.com/search?hl=en&q=site%3Atuxscreen.sourceforge.net+wiki |
| 17:11:21 | Russ|werk | what happened? |
| 17:11:43 | Russ|werk | http://www.google.com/search?q=site:tuxscreen.sourceforge.net+wiki&hl=en&filter=0 |
| 17:11:52 | BZFlag | Some jerk hacked a page, so I ran a backup, edited the backup, then ran a restore |
| 17:12:08 | BZFlag | unfortunately the restore messed up and pulled an old image. |
| 17:12:16 | Russ|werk | its the risk of running an open system |
| 17:12:35 | Russ|werk | (btw, click the above link and you get everything) |
| 17:12:36 | BZFlag | I did some cleanup, which involved removing the backup I just made (gak) |
| 17:12:46 | BZFlag | and then noticed the restore had messed up. |
| 17:13:09 | AlHazWork | I didn't think we'd offended anyone |
| 17:13:44 | kernelwork | who says someone has to be offended to hack a page? |
| 17:13:48 | BZFlag | so if someone would make a backup from google, that would be much appretiated! |
| 17:13:59 | Russ|werk | some one is ub3r l33t, they defaced a wiki |
| 17:14:02 | BZFlag | we could put it on the site someplace. |
| 17:14:12 | BZFlag | volunteer? |
| 17:14:24 | Russ|werk | BZFlag: like I said, the link I posted has *everything* |
| 17:18:36 | BZFlag | Russ|werk: ok, so are you volunteering? |
| 17:21:25 | Russ|werk | I'm at work |
| 17:23:04 | AlHazWork | kernelwork: I know, the scriptkiddy mind is just a child in need of attention . . . . |
| 17:25:46 | jacques | well well well. 2.4.16 is out |
| 17:37:43 | pattieja | is away: I'm busy |
| 17:37:55 | pattieja | is back (gone 00:00:02) |
| 18:38:46 | pattieja | is away: lunch |
| 19:07:33 | CosmicPenguin | Afternoon all! |
| 19:09:05 | prpplague | howdy CP |
| 19:09:19 | prpplague | CosmicPenguin: break any good code today? |
| 19:10:23 | prpplague | CosmicPenguin: you know you have really have to break at least one line of good code a day so you don't make the m$ programmers feel bad |
| 19:12:09 | kernelwork | I thought you had to mistakenly write one line of bad code to do that |
| 19:12:43 | Niddix | I thought it was to add one major security flaw every day |
| 19:12:58 | CosmicPenguin | prpplague: Haven't touched a line yet.... been digging snow and reading old mail.... |
| 19:13:47 | CosmicPenguin | prpplague: I easily break more than one line a day.... heck, left to my own devices, I'll bring down an entire project.... |
| 19:14:25 | kernelwork | Microsoft has great security, the users often have trouble getting anything done :) |
| 19:15:45 | CosmicPenguin | kernelwork: Funny, my mom says the same thing about Linux.... |
| 19:16:11 | kernelwork | I tried messing around with my brothers win XP and within 20 minutes a profound sense of boredom set in. |
| 19:17:04 | kernelwork | Though, I must say that I am impressed by the fact it only took 6 years to make a consumer OS that could have it's IP address changed without restarting. |
| 19:17:41 | kernelwork | It doesn't even need the install disk anymore :) |
| 19:18:10 | jacques | HEH |
| 19:18:20 | jacques | i always wondered how long that would take |
| 19:18:30 | jacques | can you change the font size in the GUI without rebooting yet? |
| 19:19:06 | kernelwork | I don't know, but if they figured out how to change 4 bytes and a bitmask without rebooting I'm sure they can figure out anything :) |
| 19:19:14 | jacquesZzZz | heading to work |
| 19:19:19 | jacquesZzZz | be on from there |
| 19:28:35 | pattieja | is back (gone 00:49:50) |
| 19:36:24 | jacques_gone | back |
| 19:43:36 | prpplague | argh, the best looking chic in the build just quit! darn |
| 19:45:19 | jacques | huh? |
| 19:45:42 | prpplague | jacques: just being silly |
| 19:45:48 | jacques | :-) |
| 19:46:18 | kernelwork | People have trouble telling when I'm being wierd and when I'm not. Probably because there is little difference |
| 19:46:37 | prpplague | lol, same here |
| 19:47:54 | kernelwork | likes the snow we got these last 2 days |
| 19:54:37 | jacques | we only have one chick here |
| 20:10:44 | jacques | so, how's the latest kernel patches going? |
| 20:11:46 | GoRK | awful :) |
| 20:11:53 | GoRK | ucb1x00-audio is a mess |
| 20:11:59 | jacques | GoRK: sorry to hear that :-\ |
| 20:12:26 | jacques | GoRK: do you have a link to the tscal driver you mentioned earlier? |
| 20:12:30 | GoRK | i can get playback to work perfectly or i can get recording to work perfectly :) telecom, i cant get to work at all |
| 20:12:42 | GoRK | tscal is in tuxscreen cvs under kernel/ |
| 20:12:50 | jacques | GoRK: tihs is stuff that was working under 2.4.7? |
| 20:13:25 | GoRK | not particularly ... audio under 2.4.7 doesnt use DMA and is quite bloated and bad |
| 20:13:39 | GoRK | both playback and recording worked under 2.4.7, though, yes |
| 20:14:07 | jacques | GoRK: ah, they started using dma |
| 20:14:48 | GoRK | CosmicPenguin: hey if you want some example code for a touchscreen driver that works with the new interface (should be nearly cut and paste for your microwindows) - i wrote one for PicoGUI this weekend... it's in the picogui cvs as input/ucb1x00.c |
| 20:15:12 | jacques | dammit sourcescourge has the cvs files with a strange mime type |
| 20:15:21 | jacques | and galeon wont let me view it in the browser |
| 20:15:34 | jacques | text/x-csrc |
| 20:15:38 | jacques | wtf is that? |
| 20:15:43 | GoRK | c source |
| 20:16:10 | jacques | i suppose it works just fine with IE |
| 20:16:23 | GoRK | mozilla and opera too, even :) |
| 20:16:33 | GoRK | dunno why galeon is bombinb |
| 20:16:36 | GoRK | bombing |
| 20:16:41 | jacques | hm, galeon is mozilla |
| 20:16:51 | jacques | im even using galeon 1.0 with mozila 0.9.6 |
| 20:17:56 | GoRK | haha according to /. my friend's domain name "terahertz.net" might just have become suddenly extremely valuable |
| 20:18:23 | Russ|werk | just click on the version number |
| 20:18:30 | Russ|werk | and you'll get a highlighted version |
| 20:19:27 | jacques | Russ|werk: thanks |
| 20:20:18 | jacques | GoRK: I hope some company doesnt claim he's squatting on it and take it away |
| 20:21:08 | jacques | heh, this highlighted source is purty |
| 20:21:52 | GoRK | no he has a company called terahertz communications (or he did anyway) so everything is cool |
| 20:21:57 | Russ|werk | jacques: apt-get install glimmer |
| 20:22:14 | GoRK | oh whoops i forgot he sold it all to netkraft |
| 20:22:37 | jacques | GoRK: cool. I think the big companies have been using the squatting rules way too much lately |
| 20:23:36 | GoRK | apt-get install vim-gtk exhuberant-ctags vim-rt |
| 20:23:48 | jacques | so, I want to build a tuxscreen kernel that has suppotr for ext3, but I would also like to be able to use the touchscreen :-) |
| 20:23:52 | Niddix | Interesting.. I'm on twice. |
| 20:24:21 | GoRK | jacques: i never actually got the tscal driver to work.. i havent seen spacecoaster yet to ask him about it though |
| 20:24:33 | jacques | I never did find all the params I am supposed to change to get reiserfs working with a smaller journal |
| 20:24:44 | jacques | GoRK: he was on a bit earlier |
| 20:24:53 | | spacecoaster was last seen on #tuxscreen 3 days, 5 hours, 30 seconds ago, saying: thekernel: where are you at? [Fri Nov 23 15:24:23 2001] |
| 20:24:53 | jacques | ibot: seen spacecoaster |
| 20:25:03 | jacques | GoRK: I guess not |
| 20:26:11 | CosmicPenguin | GoRK: Thanks..... |
| 20:26:42 | Niddix | Is there a simple way to configure a program that uses autoconf to use the arm cross compiler |
| 20:27:05 | CosmicPenguin | Niddix: Specify arm-linux as the target |
| 20:27:47 | GoRK | CosmicPenguin: sure. I am going to be fixing the jitter correction up (it's currently pretty braindead) but it at least works (and it doesnt use polling anymore!) - the only tricky thing to look out for is you have to issue whatever you do on pin up (click release or whatnot) when the data stops coming in... so really you should set a timer on something that does get polled to track it.. i think you'll see what im talking about |
| 20:28:01 | GoRK | pin up = pen up |
| 20:28:45 | GoRK | since the new ts driver doesnt output any data that says "the pen has been lifted" or whatnot |
| 20:28:59 | GoRK | it just stops outputting |
| 20:29:00 | jacques | GoRK: you're fixing the jitter in the ts driver?? my hero! |
| 20:29:59 | GoRK | heh well from actually working with the raw data, the old kernel driver did a pretty darn fine job of jitter correction actually |
| 20:30:35 | GoRK | basically what im doing is very cheap -- ignore low pressure inputs and ignore movement if it's < 3 pixels |
| 20:31:07 | jacques | oops - lunch |
| 20:31:09 | jacques | BIAB |
| 20:33:56 | CosmicPenguin | GoRK: Every other driver on the planet can detect a pen up event.... I really think we took a step back here... |
| 20:34:11 | CosmicPenguin | GoRK: How long is your pen up timer? |
| 20:34:56 | GoRK | well currently it's 3 fd_init loops ... so i dont actually have a timer... but i understand that the kernel driver returns data about every 20ms, so setting 30ms or so should be sufficient |
| 20:37:29 | GoRK | well the only thing i can say about the new kernel driver is that 1) it's by far not the only touchscreen driver that requires userspace coordinate correction and 2) the ucb1200 chip costs about $2.. not bad for a touchscreen interface with ADC, 10gpio controls, and two audio devices ... it's cheap. exporting its raw functionality out to userspace is going to give us some cheap audio and touchscreen devices |
| 20:38:17 | GoRK | but without that, rmk's not going to take the driver, so better just do it and move on |
| 20:39:27 | GoRK | btw you can change the code to read the calibration params from the file that fbtuxcal outputs so you dont have to actually *do* the calibration setup in microwindows -- if you have driver messaging or something you could with your old calibrate app i guess :) |