Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2006-03-16

00:00:33 Quit petur ("here today, gone tomorrow")
00:02:59sharpei'll be back in a bit...
00:03:45*sharpe prances off to non-rockbox land...
00:04:07linuxstbkkurbjun: Your latest Doom is giving me lots of warnings about LONG_MAX, SHRT_MIN, SHRT_MAX etc being redefined...
00:04:42linuxstbAnd then it dies with errors about functions being declared static after non-static...
00:05:18linuxstbAnd something I don't understand: "p_mobj.c:885: warning: dereferencing type-punned pointer will break strict-aliasing rules"
00:05:26 Quit ender` (" Don't take life too seriously; you won't get out of it alive.")
00:06:47webguest55Sounds like a sentence generated by a clever script
00:07:20 Join midkay [0] (n=midkay@rockbox/developer/midkay)
00:07:43kkurbjunlinuxstb: Where are you getting the static / non-static errors
00:07:45 Join antoine [0] (
00:08:58stripwaxlinuxstb- for latest patch - fixes sprite mangling when it goes off left-hand tunnel; also (hopefully) some more minor optimisations of sprite and char rendering.
00:09:10 Join quobl [0] (
00:09:19stripwaxwant me to upload it to the patch tracker or are you cool w that?
00:09:46kkurbjunlinuxstb: I have that mobj error also, I havn't looked into it, the SHRT stuff is because of some definitions in rockmacros.h, I was including the values from limits.h
00:10:31 Join nnod [0] (
00:10:34linuxstbstripwax: It's OK, I'll try and look at it tonight.
00:10:48stripwaxactually small prob.. one sec..
00:11:05linuxstbkkurbjun: I'm confused... Does it fail to compile for the H300 as well then?
00:11:31kkurbjunlinuxstb: no, not on gcc 3
00:11:38linuxstbAh, so it's a gcc 4 thing.
00:13:26stripwaxlinuxstb - sorry, patch should be fine now.
00:13:32 Quit mikearthur (Remote closed the connection)
00:14:06 Join mikearthur [0] (
00:15:48stripwaxI'm at 14fps on h1x0 so far, I'm just about to build in the patch Paprica referred to and see how it goes!
00:16:07linuxstbstripwax: That patch is in cvs now.
00:16:21linuxstbWith just that patch, it's faster than 20fps.
00:16:24 Quit Rondom (No route to host)
00:16:37linuxstb(i.e. the framerate control is kicking in)
00:17:31***Saving seen data "./dancer.seen"
00:18:12linuxstbkkurbjun: I finally got it compiling for the ipod, and it's too big for IRAM.
00:18:19linuxstbs/IRAM/Plugin RAM/
00:18:43 Quit quobl (Remote closed the connection)
00:18:46stripwaxlinuxstb - great, just trying now
00:18:48linuxstbBut there are a _lot_ of warnings.
00:19:10kkurbjunlinuxstb: yeah, I know there's a ton of warnings on this new version, I was just getting in all the code
00:19:40 Join didj [0] (
00:19:42linuxstbIt looks like we are back at the beginning then... And there was me hoping to commit it tonight :(
00:19:55kkurbjunlinuxstb: there is alot of code in this new version, how much over is it on the plugin ram
00:20:03linuxstbBut from your comments in the patch tracker, it seems worthwhile.
00:20:22kkurbjunlinuxstb: yeah, it's a much better code base
00:20:39linuxstbI think we're going to have problems with unaligned memory accesses again though - if this is the same as the test version you gave me.
00:21:21linuxstbI'll try increasing the plugin ram size and see what happens.
00:21:27kkurbjunlinuxstb: I made one change to the memory allocator that was more or less a bug, I was also having some memory corruption that I fixed as well so I was hoping that wouldn't be an issue
00:22:46kkurbjunI also have some packet attribute flags for some of the wad information that wasn't really being loaded correct before
00:22:58 Quit tinodeleste (Read error: 110 (Connection timed out))
00:23:12kkurbjunif that's an issue though you can remove them by redefining PACKEDATTR in rockmacros.h
00:24:05imphasinglinuxstb: Which file has the entry point for the rockbox bootloader?
00:24:26stripwaxlinuxstb - nice, that z80 patch makes it fly
00:25:04linuxstbimphasing: firmware/crt0.S
00:25:09imphasingah, cool
00:25:16imphasingLooks like pacbox got fast
00:25:28linuxstbYep, and getting faster :)
00:26:12linuxstbkkurbjun: I've just tried it, and it's not finding my wad - I have /games/doom/doom1.wad
00:26:51imphasingWhoa, Linus sure knows his ARM ASM
00:27:14kkurbjunlinuxstb: you need prboom.wad in that directory also
00:27:25kkurbjunone sec, let me give you a link
00:27:34linuxstbimphasing: Why do you say that?
00:28:04imphasinglinuxstb: It looks like he wrote the rockbox bootloader
00:28:06linuxstbkkurbjun: It's OK, I've got it from last time. I just forgot about it.
00:28:14imphasingah, never mind
00:28:22imphasingpreglow knows his ARM ASM
00:28:53 Nick antoine is now known as tinodeleste (
00:29:47linuxstbkkurbjun: "data abort at 01FB0058"
00:29:57linuxstb(i.e. unaligned memory access)
00:30:34 Join quobl [0] (
00:30:41kkurbjunlinuxstb: where was that, r_main.c again? did you try redefining the packedattr?
00:30:44linuxstbThat's the R_RenderBSPNode function
00:31:20linuxstbNo, I haven't touched packedattr.
00:31:59linuxstbBut that's a bad thing for ARM.
00:32:55kkurbjunyeah, I figured, if you redefine it blank in rockmacros.h it might help
00:33:02linuxstbJust trying that now...
00:33:39linuxstbAnother data abort - different location.
00:34:21linuxstbIt's now in R_InitTextures
00:34:34kkurbjunthat's earlier in the code
00:35:00kkurbjunI think the alligned attr is needed for some of the map data as it's packed on disk
00:35:44preglowimphasing: most of the bootloader was stolen from ipl...
00:35:54imphasingpreglow: ah, hehe
00:36:01imphasingIt's still impressive
00:36:58linuxstbstripwax: You could try changing FPS to 30 for the H140 and see how fast it goes.
00:37:47preglowhalf of it is ipl code, looks like
00:38:14*preglow kicks cordic
00:38:22linuxstbstripwax: The blit_display() function could possibly benefit from some optimisation as well.
00:39:17kkurbjunlinuxstb: all the memory allocation is effectively done in z_zone.c now, I thought I fixed the problems with it, but aparantly there are still some more. I would guess that that is what's causeing the allignment issues
00:39:56kkurbjunlinuxstb: did the game at least get to the menu screen with PACKET ATTR?
00:40:28kkurbjunand without it didn't?
00:40:34linuxstbThe data aborts have both been after I selected play game.
00:40:39 Quit matsl ("Leaving")
00:40:41kkurbjunoh, hmm
00:41:05linuxstbI assume you mean the Rockbox menu, not the Doom menu.
00:41:14kkurbjunno, doom menu
00:41:24linuxstbThe menu with "New Game" etc?
00:41:36linuxstbNo - it doesn't get that far.
00:41:42imphasingpreglow: cordic?
00:41:50linuxstbIt crashes during the initialisation.
00:41:52preglowtrig function generator
00:42:00preglowneeded to fix the eq
00:42:27kkurbjunyou looked at the z_zone stuff before though dind't you?
00:43:17 Quit tenzing (Read error: 104 (Connection reset by peer))
00:43:54 Join tenzing [0] (
00:44:12kkurbjunit's strange that it's calling R_RenderBSPNode before showing anything. it should at least get to the doom menu before doing that
00:44:26stripwaxpreglow - what's "up" with the eq?
00:44:48stripwaxlinuxstb - yep, I'll give it a go. was thinking about rewriting the blt to use coldfire asm but don't have time to do that right now
00:45:36linuxstbkkurbjun: Yes, and I don't think I could find the problem.
00:46:05 Nick ScoTTie_ is now known as ScoTTie (n=scott@unaffiliated/scottie)
00:46:32linuxstbI was hoping you would give up on this new code and stick with the original Doom :)
00:46:58stripwaxlinuxstb - @ 50 fps! D-) and still rising .. 51ish
00:47:17linuxstbstripwax: I've just tried your latest patch, and it doesn't seem to make any difference for the 5g.
00:47:28stripwaxit fixes the bug, no?
00:47:31kkurbjun: ), I thought about it a bunch of times, but I really wanted to have a version of doom that could play quite a few of the new mods reliably
00:47:35preglowstripwax: the frequency is inaccurate sometimes
00:47:43linuxstbstripwax: Ah yes, I forgot there was a bugfix in there....
00:47:55stripwaxit's mainly a bugfix! ;-)
00:48:07preglowamiconn: timer.c is one-shot mode only?
00:48:16linuxstbI'm too obsessed with optimisations at the moment.
00:49:21stripwaxpreglow - oh! silly question but - how can you tell / how did you spot that? or is it so inaccurate you can hear that it's wrong?
00:50:48preglowstripwax: some people did measurements, and then i plotted the filter response for the coefficients in octave, and it turned out it's off
00:51:07preglowif you ask for 60hz, you get 140hz, etc
00:51:24preglowturns out my sine function wasn't accurate enough
00:51:47stripwaxpreglow - wow. I mean, that's great, I wouldn't have expected eq freqs to be completely perfect anyway.
00:52:05stripwaxit's great that the rockbox devs care this much about quality, is what i meant, of course
00:52:48kkurbjunlinuxstb: I think another good canidate for some of the alignment problems is in p_setup.c in function P_LoadNodes, it's adding an integer value to *nodes and then assigning that to a pointer, I'm supprised it works at all...
00:53:08preglowwell, it's a pretty major bug, and i also need a very accurate sin() function for the eq plotting
00:53:08stripwaxpreglow - you only need the sine to build the coeffs though, right? would a slow-and-very-accurate trig fn be more appropriate?
00:53:19preglowthat's what i'm doing now
00:53:28preglowit's only used every time you change the settings
00:53:34preglownothing's changed in the static case
00:53:51stripwaxcools, I guess that's what I was asking. i can go back into hiding now ;-)
00:53:56linuxstbkkurbjun: OK. But I'm too tired to look at it any more tonight.
00:54:16linuxstbIf you manage to get it compiling in the sim, that may help debugging.
00:54:42kkurbjunlinuxstb: no problem I'm looking into what I can, I'm working on that now
00:54:48kkurbjunthanks for looking at it though
00:56:09 Join rkostynu [0] (
00:58:32preglowbloody annoying, this is
00:58:42preglowthink i'll just wait for safetydan
00:58:59preglowsleep also sounds sweet
00:59:00linuxstbstripwax: I've just tried making pacbox display the full 60fps on the H140, and it's managing about 51.4
00:59:19stripwaxlinuxstb - yep, almost exactly what I just reported for h120..
00:59:37stripwaxabout 15mins ago in fact!
01:00:46 Quit kernelsensei (Read error: 104 (Connection reset by peer))
01:00:50 Join kernelsensei [0] (n=boris@gentoo/developer/kernelsensei)
01:02:41 Join linuxstb_ [0] (
01:02:50 Quit linuxstb (Read error: 104 (Connection reset by peer))
01:02:59 Join damaki_ [0] (
01:03:02 Nick linuxstb_ is now known as linuxstb (
01:03:17linuxstbstripwax: Were you trying 30fps or 60fps?
01:03:35stripwax60fps. which is how I got 51fps
01:03:48stripwaxI figured 30fps would be too easy.. :-)
01:03:51linuxstbOf course, I forgot that I had enabled the framerate control...
01:04:12linuxstbIt seems a bit pointless though...
01:04:22linuxstbMaybe we should try the greyscale lib.
01:04:36stripwaxmight be better to try 30fps and adjust cpu boost?
01:04:51stripwaxat 50fps the h1x0 lcd is pretty darn blurry
01:05:53*stripwax seems to remember that dynamic cpu boost would play havoc with framerate control tho..
01:06:12preglowtime for bed, y'all
01:06:17stripwaxgnight you
01:08:19 Quit Kohlrabi (Nick collision from services.)
01:08:26 Join Kohlrabi [0] (
01:12:04 Nick scf is now known as Zzzcf (
01:13:55stripwaxdo we know how fast the H1x0 can actually do continuous full screen lcd updates?
01:13:59linuxstbstripwax: We've also got sound to implement.
01:14:13stripwaxlinuxstb - heh.. yeah..
01:14:30linuxstbWould you be interested in attempting that?
01:15:31stripwaxcertainly, although no promises on when I'd be able to finish it!
01:16:09stripwaxare there any other examples of plugins that play sounds? i guess metronome would be a starting point, and then I'd need to generate the samples somehow..
01:16:27linuxstbDoom does it.
01:16:47stripwaxI'd say that's a complicated starting point.
01:16:52linuxstbThe original source is here:
01:17:00linuxstbBeware - C++ inside.
01:17:13stripwaxyeah, got that already, just need some pointers on how to do the rockbox audio buffer thang
01:17:35stripwaxhey, you're on f2s also. how you finding the network speeds since march? :-D
01:17:51stripwaxyeah. .
01:17:57linuxstbBut it seems to have got better recently.
01:18:23stripwaxyeah, last two weeks or so in my case. can't believe I'm paying for 8Mb and getting <1Mb most of the time
01:18:42 Quit damaki (Read error: 110 (Connection timed out))
01:20:29linuxstbYou could test the audio code by just writing the PCM data the emulator generates to a file.
01:21:34stripwaxsure could. I used to hack Buzz plugins so that should be pretty straightforward. just not familiar yet with how to make it actually play. but it seems that all the audio data is in the rom anyway so it should be a piece of cake
01:23:49linuxstbI've just remembered that the audio frequency is 96KHz. So we'll have to resample to 44.1KHz.
01:25:07linuxstbstripwax: Just committed your patch.
01:25:12stripwaxneat! cheers
01:25:33 Join lolo-laptop [0] (
01:25:44 Quit lolo-laptop (Client Quit)
01:25:49stripwaxgotta go. hopefully i'll get time to try out audio tomorrow; if not then soonest..
01:26:03 Part stripwax
01:26:32 Quit tinodeleste ("Client exiting")
01:36:24 Quit amiconn (Read error: 104 (Connection reset by peer))
01:42:39 Quit Rick (Read error: 104 (Connection reset by peer))
01:44:31 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
01:45:06 Quit sharpe (Read error: 110 (Connection timed out))
01:46:27 Quit herz42 ("Und wech")
01:50:37 Quit rkostynu ("Chatzilla 0.9.72 [Firefox]")
01:50:51 Join Rick [0] (
01:51:15 Nick BHSPitLappy2 is now known as BHSPitLappy (
01:52:03 Join savagenator [0] (
01:52:24savagenatorhello people. anyone can help with the activation email to get into the rockbox forums?
01:53:59 Quit obo (Read error: 110 (Connection timed out))
01:55:22 Quit nnod (Remote closed the connection)
01:56:26BHSPitLappy:O NES on iPod!
01:56:41*BHSPitLappy runs away from the potential can of worms now opened
02:00:42 Join sharpe [0] (
02:00:48*sharpe prances back.
02:02:33XavierGramiconn, Bger: I added the 'zoom more while playback thingy' on the tracker.
02:03:19XavierGrBHSPitLappy: I think this will be very nice and doable if we find a good emulator to port.
02:04:11savagenatorso anyone?
02:04:18BHSPitLappyXavierGr: it wasn't a suggestion, it was a "FYI"
02:04:24savagenatorrockbox activation email?
02:04:36BHSPitLappysavagenator: umm, it's not complicated?
02:05:01savagenatori just dont recieve it
02:05:12BHSPitLappyfor your information
02:05:13savagenatori tried at least once every day for the last week
02:05:25BHSPitLappysavagenator: what domain is the address?
02:05:47savagenatorwhat do you mean?
02:07:50 Quit webguest55 ("CGI:IRC (EOF)")
02:08:13 Quit Moos ("Glory to Rockbox !!!")
02:08:43*savagenator finds there is no answer...
02:10:35 Quit mirak (Remote closed the connection)
02:12:19BHSPitLappysavagenator: it's possible it doesn't accept yahoo... create a gmail account
02:12:20 Quit savagenator ("CGI:IRC (EOF)")
02:12:26BHSPitLappylol, even
02:16:59 Join kernel_sensei [0] (n=boris@gentoo/developer/kernelsensei)
02:17:29 Join actionshrimp [0] (
02:17:32***Saving seen data "./dancer.seen"
02:19:20 Quit kernelsensei (Connection timed out)
02:19:27 Nick kernel_sensei is now known as kernelsensei (n=boris@gentoo/developer/kernelsensei)
02:21:15 Quit tenzing (Read error: 110 (Connection timed out))
02:22:12 Join damaki [0] (
02:26:56 Join tenzing [0] (
02:27:42 Join damaki__ [0] (
02:30:32XavierGrBHSPitLappy: From a little search FCE Ultra is your answer
02:30:44XavierGrOpen Source Emulator under GNU in C language
02:30:50BHSPitLappyXavierGr: I wasn't asking :P
02:30:50XavierGrSeems a right candidate to me
02:30:52BHSPitLappyI was telling
02:31:03BHSPitLappyinfoNES has been ported to the iPod
02:31:20XavierGrthen this should be more easy
02:33:47 Join nnod [0] (
02:34:26XavierGrIt actually runs, but really slow.
02:34:29XavierGrthis is bad
02:34:39XavierGris it for a long time now?
02:35:57XavierGrthat emulator port, are they fiddling with it for a long time, or it just started?
02:36:28XavierGrbecause if it is like that for a long time.... (and active) I can't see how much more we can do.
02:36:33 Quit damaki_ (Read error: 110 (Connection timed out))
02:38:55BHSPitLappyI wasn't saying anything about rockbox... I was just mentioning it because it's interesting
02:41:32 Quit damaki (Read error: 110 (Connection timed out))
02:47:45 Join JdGordon [0] (
02:48:09 Quit JdGordon (Client Quit)
02:48:19 Join JdGordon [0] (
02:49:41sharpehope i figured out the bits correctly...
03:18:04 Join Spida_ [0] (
03:18:42 Join tvelocity [0] (
03:22:35sharpewonderful, so far it successfully maps each bit to a pixel... and i know no one is paying attention to me.
03:28:00 Quit Spida (Read error: 110 (Connection timed out))
03:33:36 Quit actionshrimp ("( :: NoNameScript 4.01 :: )")
03:38:48 Quit damaki__ (Read error: 104 (Connection reset by peer))
03:44:39midkaysharpe, ey!
03:45:17 Quit Daishi ("Client exiting...")
03:49:02sharpehey midkay
03:49:20midkaywhat're you up to?
03:49:32sharpeworking on the pixel maze :)
03:51:26sharpei've got the map storage done
03:51:35sharpei just need to figure out an algorithm that works
03:52:25midkaycool :)
03:52:46sharpeyeah, the coordinate system is like, flipped
03:53:27sharpefor the ipod, the maze takes up 40*240 bytes...
03:54:08sharpeand each byte corresponds to 8 horizontally consecutive pixels
03:54:58sharpeand... i've got nothing so far for making the maze
03:56:43imphasingsharpe: What are you working on now?
03:56:54sharpethe pixel maze :P
03:57:26JdGordonpixel maze?
03:57:31sharpejust need to find an algorithm for making the maze
03:57:33imphasingI just renewed my hatred of C++
03:57:47 Join damaki [0] (
04:01:44sharpeheh, lets see if this goes out of the bounds of the map and starts overwriting parts of memory... :)
04:10:39sharpeis it just me, or does 'make' pause for a while in the plugins/bitmaps/mono or plugins/bitmaps/native?
04:11:39midkayit's not just you!
04:12:42sharpedoes it go over plugins/bitmaps/native, mono, mono, native, native, mono, mono, native for you too?
04:15:56midkaysomething like that..
04:16:02midkaya bunch of dirs..
04:16:09midkayi think they may be different, though..
04:16:11sharpethe same ones?
04:16:20midkayie apps/bitmaps/native; apps/plugins/bitmaps/native?
04:17:33***Saving seen data "./dancer.seen"
04:18:03sharpeyeh... mine are plugins/bitmaps/native and /mono that are repeated...
04:19:23sharpejust wondering if something like that happened for anyone else...
04:19:33midkayi'll check in a sce
04:27:26sharpeno wonder that switch statement didn't run
04:30:14midkaysharpe, it looks like it does the multiple bitmap build thing for me too.
04:30:16midkayand. why? :)
04:31:09sharpeheh, i was writing a if statement to check if x and y were out of bounds when it chose a direction to go in, but i put in the the while loop, but forgot to remove the one right before the switch statement... :)
04:31:38midkayhaha, alright........ :)
04:32:02sharpei'm just making up an idea for a map generator
04:33:42sharpewhich, doesn't seem to be working too well
04:33:49sharpefor some reason, it won't draw the map anymore.
04:35:20sharpeoh well... i'll work on it tomorrow.
04:35:47sharpegoodnight everyone who's conscious.
04:36:58Aditya_nighty night
04:37:11midkaynight Aditya_
04:37:16midkaynight sharpe
04:52:49JdGordondoes any1 know a winamp plugin that plays the first ouple seconds of each song? or alternativly, a program that can download the lyrics for every song in your collectino?
05:01:44 Join Rob2222_ [0] (
05:07:44 Quit jaebird ("Leaving")
05:10:12 Quit Shadowarrior13 ("OMG <3 <3 <3 lolwtfbbq?")
05:15:15 Join virtualball2 [0] (
05:16:19 Quit sharpe (Read error: 110 (Connection timed out))
05:20:00 Quit Rob2222 (Read error: 113 (No route to host))
05:30:37 Join Iceman [0] (n=icechat5@
05:31:21 Part Iceman
05:56:31 Quit midkay ("Leaving")
05:57:21 Nick TiMiD[FD1 is now known as TiMiD (n=TiMiD[
06:01:16 Quit virtualball2 ("leavin")
06:11:51 Quit DJ_Dooms_Day (Read error: 110 (Connection timed out))
06:17:36***Saving seen data "./dancer.seen"
06:21:07 Quit [TCK] (Read error: 104 (Connection reset by peer))
06:21:47 Quit imphasing (Read error: 110 (Connection timed out))
06:21:59 Join nave7693 [0] (
06:22:33 Part nave7693
06:30:49 Quit nnod ("Leaving")
06:32:44 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
06:45:02 Quit tvelocity ("Ex-Chat")
06:47:21amiconnpreglow: timer.c implements an interval timer, with a priority system and an unregister callback
06:49:17amiconnThe callback is necessary on H1x0 because the second timer is shared between backlight fading and plugins. Plugins requesting the timer can preempt backlight fading.
07:11:43 Quit SereRokR ("XChat Aqua")
07:22:33 Quit quobl_ (SendQ exceeded)
07:25:19 Join needleboy [0] (
07:25:33 Quit needleboy (Client Quit)
07:25:38 Join needleboy [0] (
07:30:58 Quit needleboy ()
08:01:34 Quit darkless (Read error: 110 (Connection timed out))
08:07:33 Join midkay [0] (n=midkay@rockbox/developer/midkay)
08:17:38***Saving seen data "./dancer.seen"
08:31:03 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
08:40:53 Join needleboy [0] (
08:40:54 Quit needleboy (Client Quit)
08:41:31 Join needleboy [0] (
08:41:32 Join ender` [0] (i=ychat@
08:42:59 Join damaki_ [0] (
08:47:05 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:47:10 Part LinusN
08:47:26 Join LinusN [0] (
08:48:10 Join B4gder [0] (
08:48:10amiconnmorning to the bouncing bouncer ;)
08:48:30B4gderxchat crap
08:49:36amiconnThat was directed @ LinusN
08:49:57LinusNmoo to you too
08:49:58B4gderwell, I've been fiddling with my IRC
08:50:10B4gderit turned out xchat had blanked my nick name field
08:50:22B4gderso all irc servers disconnected me immediately
08:50:30LinusNB4gder: machines don't have a mind of their own
08:50:40B4gderwhat? since when?
08:51:07B4gdershould I stop talking to my machines then?
08:52:09B4gderI think I better leave that to the voices in my head
08:52:18amiconnOh, they don't?
08:52:42*amiconn got the impression they sometimes have
08:58:49 Quit damaki (Read error: 110 (Connection timed out))
09:00:07 Join Acksaw [0] (
09:07:51 Join petur [0] (
09:12:07 Quit Acksaw ("( :: NoNameScript 4.01 :: )")
09:12:36 Join ashridah [0] (
09:14:27midkaybtw.. i'm planning to commit an updated quick screen soon.
09:19:22 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
09:27:56 Join adiamas [0] (
09:28:08 Join darkless [0] (
09:43:09LinusNmidkay: how nice for you ;-)
09:43:44midkayLinusN, oh, so you don't want it. fine! *deletes quickscreen.c* :)
09:50:07JdGordonmidkay: whats changed?
09:51:03midkayJdGordon, i moved the bitmaps+text to appropriate positions on the LCD (left aligned, right aligned, center aligned) to give a better idea of "this direction does this".. while retaining the ability to scroll the lines
09:51:22JdGordonfun (y)\
09:53:02JdGordoncan that screen be made customizable? so it actually can be used for user shortcuts?
09:53:34JdGordonthe best i can tihnk of doing it is having a list of available functions the keys can call.. but that sux
10:04:31 Join Paul_The_Nerd [0] (
10:06:18B4gderMale to Female Ratio: 85:1
10:06:25B4gderin the rockbox forums
10:06:47JdGordonthats better than i would have expected
10:07:56 Join adiamas_ [0] (
10:09:59 Quit adiamas (Read error: 113 (No route to host))
10:10:04 Nick adiamas_ is now known as adiamas (
10:12:25safetydanpreglow, what's up with cordic?
10:17:42***Saving seen data "./dancer.seen"
10:17:58Paul_The_NerdHey, another "Hey look, Paul is still a linux n00b" question: How would I go about putting error/warning messages I get while compiling into a .txt?
10:18:16JdGordonmake > blaa.txt
10:18:29JdGordonor make 2> blaa.txt
10:18:48JdGordon2> gets stderr
10:18:53Paul_The_NerdJdGordon: When I tried make > blaa.txt, everything *but* the warnings/errors were caught
10:18:56safetydanmake > out 2>&1
10:19:08safetydandumps everything to out
10:21:19Paul_The_NerdAnd apparently the new Doom source doesn't compile.
10:21:32JdGordonkill him...
10:21:58linuxstbPaul_The_Nerd: No, I tried it last night. After I got it compiling on the ipod, it just crashed with data aborts.
10:23:12Paul_The_Nerdlinuxstb: The new source he posted?
10:25:04 Quit _ironi_ (Read error: 110 (Connection timed out))
10:27:21 Join tucoz [0] (n=martin@rockbox/developer/tucoz)
10:29:49tucozlinuxstb, those macros we talked about a few days ago. Do you think it's possible to define something like is_bitmap which includes all the bitmap targets?
10:30:03tucozfor the manual that is
10:30:11 Quit darkless (
10:31:46NJoindarkless [0] (
10:32:18 Quit ohrn (Read error: 104 (Connection reset by peer))
10:33:56 Join banan_ [0] (
10:35:35Paul_The_Nerdlinuxstb: I wonder if he left out something... It doesn't seem to compile for H300 for me either.
10:35:45tucozAs some sections include all targets _but_ the player, it's easier to just define something like that in a central place and not spreadout through the entire document.
10:37:46 Join skaffen-amitskaw [0] (
10:40:00skaffen-amitskawHi, anyone awake?
10:40:13Paul_The_Nerdkkurbjun: Just so you get a highlight, as far as I can tell your new doom source zip doesn't compile for H300 or iPod Nano (and probably others)
10:40:18Paul_The_Nerdskaffen-amitskaw: A few of us, yes.
10:40:32*skaffen-amitskaw sits sharpening knife missiles..
10:40:50skaffen-amitskawOh, Hi Paul.
10:41:23skaffen-amitskawI assume you've chosen Paul_The_Nerd to be descriptive?
10:41:33*skaffen-amitskaw hopes so
10:41:44Paul_The_NerdI chose it to be descriptive a long time ago. I kept it for consistency's sake.
10:41:56Paul_The_NerdIt's accurate enough.
10:41:57linuxstbPaul_The_Nerd: Which version of m68k-elf-gcc are you using?
10:42:25Paul_The_Nerdlinuxstb: 3.4.5
10:42:42linuxstbtucoz: The Rockbox source uses HAVE_LCD_BITMAP for that.
10:43:32linuxstbPaul_The_Nerd: What problems do you have on the H300? static/non-static declarations?
10:43:34tucozlinuxstb, I know. But, I am not really sure if it's possible to define that by using the optional package in latex
10:44:01skaffen-amitskawWell, I think I checked it out, flame me if I missed it, but ne1 know why my rockbox H340 comes up in iriver firmware when I plug external power in? It comes up rockbox USB bootloader fine if I use USB, but mains?
10:44:15Paul_The_Nerdlinuxstb: Yeah, same stuff as for the iPod
10:44:26linuxstbJust add the line \def\UseOption{have_lcd_bitmap} to all the platform.tex files for bitmapped targets.
10:44:52tucozlinuxstb, oh. I'll try that right away
10:45:16linuxstbOr maybe we should standardise on upper-case names - HAVE_LCD_BITMAP.
10:45:19Paul_The_Nerdskaffen-amitskaw: It's just the way it works right now. USB/Power detection is still somewhat of a work in progress, I believe
10:45:42 Quit merbanan (Read error: 110 (Connection timed out))
10:45:50tucozlinuxstb, that's ok. I use the rockbox-code naming convention
10:46:23skaffen-amitskawThx, Paul. I'll just keep picking up the new one and wait for the code gods to shine. It's a ridiculous minor point anyway.
10:48:12 Join darkless_ [0] (
10:48:25Paul_The_Nerdskaffen-amitskaw: It probably won't change until the whole charging thing is pretty well known, and it's possible it may end up being a bootloader update, as I'm really not that familiar with where it needs to be handled.
10:49:16Paul_The_Nerdlinuxstb: Actually, no, I was wrong
10:49:29Paul_The_Nerdlinuxstb: r_bsp.c:84: error: structure has no member named `memchr'
10:49:45linuxstbOK. You just need to add memchr to the plugin API.
10:49:47 Nick banan_ is now known as merbanan (
10:50:13linuxstbvoid *memchr(const void *s, int c, size_t n);
10:50:31linuxstbOr, in plugin.h - void* (*memchr)(const void *s, int c, size_t n);
10:51:05LinusNplenty of warnings though
10:52:35linuxstbYes, it seems to be a major rewrite compared to the previous version - which was almost ready to be committed.
10:56:56tucozlinuxstb, it looks like that has to be defined like this \def\UseOption{h1xx,HAVE_LCD_BITMAP}
10:58:04linuxstbI was just going by the definitions already in the platform.tex file - e.g. \def\UseOption{ipodcolor}
10:58:27linuxstbWhich is then checked with lines like \opt{ipodcolor,ipodnano}
10:58:50 Quit darkless (No route to host)
10:59:34tucozlinuxstb, yes. But if I add another define like \def\UseOption{HAVE_LCD_BITMAP} then all the defines for h1xx is lost. Like \genericimg etc.
11:00:26tucozBut, it works fine if the HAVE_LCD_BITMAP is defined within the same curly braces as the platform.
11:01:34linuxstbOK, that makes sense - you can only have one \def\UseOption line...
11:02:26tucozYes. But it is far easier to define that within the platform.tex files, than to use \opt{h1xx,h300,ipodcolor...} etc.
11:03:20linuxstbCan you split that \def\UseOption onto multiple lines? e.g.
11:03:47 Quit quobl (Remote closed the connection)
11:03:54tucozI would assume yes, as LaTeX is very forgiving on the whitespace
11:04:29LinusNi wonder why it generates so many spaces after the "rockbox.iriver" string
11:04:39tucozit does? where?
11:05:05LinusNat least for me, chapter 1.7
11:05:09LinusNpage 8
11:05:32linuxstb(apologies for the tiny flood):
11:05:34linuxstb ipodvideo,
11:05:34linuxstb HAVE_LCD_BITMAP
11:05:34DBUGEnqueued KICK linuxstb
11:05:39linuxstbThat works fine.l
11:06:05LinusN1.4 is the same
11:08:35 Join imphasing [0] (
11:10:04tucozlinuxstb, aren't the plugins lost if you do that?
11:11:54tucozLinusN, I think that is because there are newlines around the \opt statement
11:13:14LinusNtucoz: i tried it on one line, but it didn't help
11:13:19tucozIt helped for me
11:13:33amiconnWhat's up with the distributed builds?
11:13:48linuxstbtucoz: Yes...
11:13:52amiconnLast round took 25 minutes...
11:14:13 Join Kohlrabi [0] (
11:14:32B4gderI've mailed t0mas about it, we need more detailed logs
11:14:42tucozlinuxstb, if I do like your example I loose the other macros defined in the platform.tex file. But, if I do \def\UseOption{h1xx,HAVE_LCD_BITMAP} it works fine
11:14:44B4gderto better be able to analyze problems like this
11:15:12CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
11:15:12*amiconn would also like to see archos bootbox builds...
11:17:35B4gderI'd prefer to see the build "under control" first
11:17:51B4gderit seems to fragile atm
11:18:02B4gdertoo even
11:18:52 Join hshah [0] (
11:18:59 Quit hshah (Remote closed the connection)
11:19:39tucozlinuxstb, should I just commit that for now?
11:20:32 Join BoD[] [0] (
11:21:06BoD[]hey does anyone have a uisimulator for the ipod video, already compiled for win32?
11:23:29linuxstbtucoz: Can you test this? \def\UseOption{ipodcolor}
11:23:29linuxstb (on two lines)
11:25:18linuxstbI just think it's going to get messy if we can't declare the options individually.
11:25:39tucozthat works fine
11:25:47tucozI'll try to add another option
11:26:23 Quit BoD[] ("leaving")
11:26:39linuxstb\def and \edef are TeX commands - \def defines a macro, and \edef expands the arguments before defining the macro - so it basically adds one more to the list.
11:26:50 Part skaffen-amitskaw ("ttfn")
11:28:33tucozlinuxstb, it works fine with adding two \edef's as well
11:32:59tucozlinuxstb, I commit this. If we want to define it in a different way later on, that is ok. For now, this works fine and we could use \opt{HAVE_LCD_BITMAP} throughout the text. I could add HAVE_RECORDING as well.
11:33:48linuxstbYes - and the radio.
11:34:35linuxstbNo, it's "CONFIG_TUNER"....
11:34:40tucozhehe, ok
11:37:15tucozgrr, I have to add another ondio platform file then
11:38:55tucozIs it ok if the ondio sp is called ondio, and ondiofm for the other one?
11:43:35 Nick kclaf is now known as AwAyhOmO (
11:46:19tucozlinuxstb, how are RTC units defined?
11:48:39 Join nnod [0] (
11:49:16linuxstbtucoz: If you wanted to keep one Ondio manual, you could just put \opt{CONFIG_TUNER,ondio} around the radio section.
11:49:17 Join Moos [0] (
11:50:22tucozlinuxstb, that is true. I don't know why there is only one ondio manual. The fm have both recording and radio, which the sp lacks.
11:51:18JdGordonis the manuall split into each model like the actual code?? wouldnt it be nicer to have one manual and just mention which targets each section relates to?
11:51:41linuxstbPersonally, I would go with one manual for each target - the same as the code. But I think Cassandra wanted to limit the number of manuals as much as she could.
11:53:24linuxstbJdGordon: Yes, one of the reasons for switching to LaTeX was to have target-specific manuals.
11:54:01tucozwell, let's say the ondios are old and only make one manual for those. That is, include recording and radio.
11:54:12tucozAnd be more selective when adding newer targets
11:55:04linuxstbI don't think it reallly matters - it will be easy enough to change in the future.
11:58:25Papricalinuxstb, pacbox is awesome now, it runs on 20 fps on h300
11:58:51 Quit petur ("lunch")
11:59:01 Nick Zzzcf is now known as scf (
12:06:59 Join quobl [0] (n=quobl@tor/session/x-34180ca4ae9965dd)
12:08:52 Part tucoz ("Leaving")
12:14:34preglowsafetydan: i was trying to make it work like the old fsin i have
12:14:37preglowsafetydan: but i can't make it
12:14:59preglowsafetydan: does the cordic algo only work from -PI/2 to PI/2 ?
12:15:27Paul_The_Nerdpreglow: He said that when he gave it to yah. ;-)
12:15:32amiconnpreglow: Check today's log 06:47...
12:17:20safetydanpreglow, yes, I thought I'd mentioned that
12:17:45safetydanIt's possible to change that, you just need to look at the incoming values and set the initial x and y differntly
12:17:46***Saving seen data "./dancer.seen"
12:19:25preglowsafetydan: i know this implementation only did that, i was wondering if it was intrinsic to cordic itself
12:19:37preglowamiconn: read it
12:20:05 Part Paul_The_Nerd
12:20:24preglowamiconn: was just wondering why set_timer_period is said to be called every time the timer trigs
12:23:16 Nick AwAyhOmO is now known as kclaf (
12:25:29safetydanpreglow, no it's not intrinsic to cordic, just my lazy implementation
12:28:58LinusNgah, the new doom doesn't work
12:29:04LinusNat least not on the h300
12:29:18LinusNfirst it only looks for "prdoom.wad"
12:29:40LinusNi fixed that, but then it hangs when starting the game
12:29:56LinusNsomething about a missing SINTABLE or something
12:30:31LinusNmaybe i just don't understand how to use it...
12:32:58preglowsafetydan: i can't make it work anyway :/
12:32:59JdGordonok, annoying q.. is there any reason the wps and langs and stuff need to be included in the cvs builds zip?? could it just have the firmware/codecs/plugins and have a seperate zip for langs and wps?
12:33:15preglowsafetydan: i'd very much like to take the parameter format fsin currently does
12:33:40linuxstbLinusN: I think it needs both "prboom.wad" and whatever wad you are playing - e.g. doom1.wad
12:33:58linuxstbIf you search the IRC logs from last night or this morning, kkurbjun gave a link to prboom.wad
12:34:36 Join Zoide7777 [0] (
12:34:47Zoide7777hi guys
12:35:01Zoide7777I think my hard drive is starting to make periodic clicking noises.... (ipod 4g gray)
12:35:04preglowsafetydan: if i have to keep the current format in fsincos, i have to throw away another couple of bits just to include the full range of the angle parameter
12:35:08safetydanpreglow, okay, I'll work in it some more. Did you make any changes or improvements?
12:35:20Zoide7777i'm in the us. does anyone know if I could just take it to an apple store and ask for a new ipod?
12:35:25preglowsafetydan: nothing i can't redo
12:35:36Zoide7777(after switching the bootloader to original and deleting all traces of rockbox that is...)
12:36:33JdGordonZoide7777: prob not, but theyd know where to send it
12:37:47Zoide7777When I try to change themes, the hard drive starts making noise and takes forever to load. I tried switching to the iAmp theme and timed it: 1 MINUTES 40 SECONDS!!!!!
12:38:24Zoide7777JdGordon: oh, I had hoped someone from the "Genius" bar would just swap it right there for a new 5G :D
12:44:53markunQuake3 arena on a PDA..
12:45:45 Join asdf [0] (
12:47:04 Join ara-fat32 [0] (
12:47:46ara-fat32i saw some kind of ipod firmware available, is it like the iriver firmware , playable for mpc and flac?
12:47:52JdGordonawesome :D just played pacman.. ~19fps :D
12:48:52Mikachuara-fat32: i don't know about mpc, but rockbox plays at least mp3, flac and ogg
12:49:17JdGordonrb plays mpc
12:49:23Mikachuara-fat32: yes
12:49:26 Quit Sinbios (Nick collision from services.)
12:49:26 Nick asdf is now known as Sinbios (
12:50:41 Join sharpe [0] (
12:50:59 Join DrMoos [0] (
12:51:05JdGordondoes any1 know if the rockbox downloader can update all files excpet a few when it downloads an update?
12:53:01 Nick scf is now known as scf[aw] (
12:53:18ara-fat32thats awesome
12:54:02 Part linuxstb ("Leaving")
12:54:43amiconnpreglow: Why do you think it does? Btw, there is no set_timer_period(), only timer_set_period()
12:55:01amiconn...and it's supposed to be called whenever you want to change the period
12:55:18amiconnBacklight fading has to do so, of course
12:57:48 Join petur [0] (
12:58:50preglowamiconn: forget it, i misinterpreted a comment
13:04:34 Join Jungti1234 [0] (n=jungti12@
13:05:25 Join fantomas [0] (i=foobar@
13:06:56fantomasHi. Musepack question - I can't find that article on with comparsion of different encoders, would someone help?
13:07:03Jungti1234Isn't there a not busy person?
13:07:23JdGordonim a not busy person
13:07:36JdGordoni am :D
13:08:11LinusNgah, i installed prboom.wad and now doom starts, and even runs the demo
13:08:31peturJdGordon: did you find a way to ping names from colinux?
13:08:32LinusNi can even start the game, but when i press UP it hangs :-(
13:08:58JdGordonpetur: i just fixed my /etc/resolve.conf file and its working
13:09:42Mikachuwrite the ip of your nameserver in it
13:09:48JdGordonchanged one of the lines to
13:09:51Mikachuwith the word 'nameserver' in front
13:10:19peturMikachu: I know, but adding the DNS servers of my ips in it didn't help
13:11:05peturJdGordon: will try that tonight, thanks
13:11:19 Quit Moos (Read error: 110 (Connection timed out))
13:12:20 Nick DrMoos is now known as Moos (
13:24:40peturwiki spam :(
13:26:24preglowfantomas: isn't there just one mpc encoder?
13:27:58JdGordonooh.. devcon is this w/e... hope u all have fun
13:31:24 Quit Rick ("I… don't need to be here.")
13:31:52B4gderdevcon fun!
13:34:08 Join Rick [0] (
13:35:11 Quit JdGordon (Read error: 104 (Connection reset by peer))
13:40:43Jungti1234I can't although I want to participate..
13:43:28XavierGrLinus: Any news on the corrupt mp3 timings?
13:44:06B4gderArchos AV700 announced =>
13:45:10 Part Zoide7777
13:48:11 Join ohrn [0] (
13:50:54ashridahrockbox on a tv pvr eh?
13:51:21B4gderyay! ;-)
13:52:31Jungti1234It's different from Archos that was ignorant.
13:53:13Jungti1234 hahaha..
13:54:29fantomaspreglow: yes, the one integrated in rockbox
13:55:10XavierGrwow! 76 replies on gl's "Signing off" mail in the ml!
13:55:22XavierGrIt is more like a foroum now.
13:55:51B4gdermore like a forum?
13:56:05XavierGryes sorry for the spelling
13:56:22XavierGrI mean that there are so many replies there that it reminds me long discussion on MR
13:56:54Moosvery long closed discussion :-)
13:57:10preglowfantomas: there's no musepack encoder in rockbox
13:57:56fantomaspreglow: you understood what mean, did you?
13:58:41 Join Paul_The_Nerd [0] (
13:58:49Moosfantomas: there is one musepack *decoder* in Rockbox, and one for encode from your PC
13:58:53XavierGrMoos: At least we learned which is the conspiracy behind rockbox, and all sort of malevolent acts :p <-kidding
13:59:24B4gderLinusN is actually Darth Vader
13:59:29B4gderbut don't tell anyone
13:59:39Mooshehe :-)
14:00:08MoosB4gder: and you, who are you? :)
14:00:22B4gder"SYNTAX ERROR, input malformed"
14:01:03XavierGrnice! :D
14:01:04markunfantomas: what are you trying to ask?
14:01:13fantomasMoos: thank you very much, it was very useful information. hehe
14:01:34*B4gder is a humanbot 0.3, written by logbot
14:02:07 Quit B4gder ("time to say moo")
14:02:18fantomasmarkun: nothing, it was preglow who pointed me out on difference between encoder and decoder :)
14:02:26markunok :)
14:02:32XavierGrcan logbot interact with anything, anyone? Is there commands for it or scripts (except logging)?
14:02:36fantomasmarkun: and Moos kindly helped him )
14:02:54XavierGr^are there
14:03:09preglowbut yeah, there's only one musepack encoder, so there's not much to be gained in talking about that, and the musepack decoder in rockbox is very good
14:03:21 Join KN|stiff [0] (
14:03:26preglowit gives almost the exact same result as on your pc
14:03:42Moosfantomas: I have myself plenty of MPC albums, I like his quality −− litle bad point: no seeking in Rockbox for those
14:04:04Moospreglow: great work made on MPC front
14:05:35preglowmore to be done
14:05:46 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
14:05:53 Join BL|phoffmann [0] (
14:06:16Moospreglow: lets hope so :-)
14:06:42linuxstbHow is musepack on ARM?
14:06:54preglowpretty fast
14:06:57preglowbeats mp3, i think
14:07:00fantomaspreglow, Moos: cool. I have all my music in falc and need to decide what format to use for my iPod with Sennheizer px-200. I'm choosing among ogg and musepack actually.
14:07:06linuxstbAnd no assembler opts yet?
14:07:17preglowi did a synth_filter assembler opt that helped quite a bit
14:08:11Mikachuheh, first newsitem on is rockbox
14:09:13 Quit mikearthur (Read error: 104 (Connection reset by peer))
14:12:28 Join mikearthur [0] (
14:15:14didjis there any way to control a H340 w/o having to butcher a remote? (finding the connector seperately seems...unlikely)
14:16:37Paul_The_NerdButcher a remote?
14:16:52Paul_The_NerdYou mean "control an h340 by means other than a remote?"
14:17:14petur"... and your fingers"
14:17:50***Saving seen data "./dancer.seen"
14:17:56XavierGrmaybe mind control could help you....!
14:18:09Paul_The_NerdI've had some luck with Jedi mind tricks.
14:18:13Paul_The_NerdBut results may vary
14:18:57XavierGrReally share it with us, I have been trying such tricks, but I can't get it to work. :D
14:19:00preglowi'm surprised they haven't updated the news item with ipod support
14:19:05preglowthey seemed quite happy when that happeneed
14:19:07didjas in read digital inputs in rockbox that are available using easily available connectors external on the main unit
14:19:11didja bit more precise ;) ?
14:19:31 Join Hobbit_HK [0] (
14:20:22Paul_The_Nerddidj: I suppose if you got the USB Host feature working...
14:20:40peturwas thinking the same
14:20:49Hobbit_HKhey all, is there any progress with the ifp port?
14:21:11preglowHobbit_HK: i don't think there's been much
14:21:21preglowHobbit_HK: tomal is in some need of help, he's the only developer
14:21:22didjcan you read line-in at anytime (ie, during media playback)?
14:21:27preglowthat gets quite taxing after a while
14:21:35 Quit KN|stiff (Read error: 110 (Connection timed out))
14:21:51Hobbit_HKhmm too bad... too bad i can't code much too
14:22:18 Join KN|stiff [0] (
14:22:52Paul_The_Nerdpreglow: I actually have a 799, but I'm a little afraid of experimenting at all until there's a bootloader that doesn't feel quite so hackish.
14:24:19 Quit BL|phoffmann (Read error: 110 (Connection timed out))
14:26:33Jungti1234Windows is miserable.
14:27:43 Quit mikearthur (Read error: 104 (Connection reset by peer))
14:28:27 Join mikearthur [0] (
14:28:57didjanyone know bout H/W limitations of reading line-in adc, while playing audio ?
14:29:30 Quit ashridah ("Leaving")
14:30:30 Quit mikearthur (Remote closed the connection)
14:31:54 Join SereR0KR [0] (
14:32:25 Join mikearthur [0] (
14:32:36*amiconn spots talk about logbot interactivity and wonders why logbot still has no clue about rockbox things
14:32:50amiconn...i.e. why 'explain' still doesn't werk
14:32:56preglowwhy does it matter?
14:33:22preglowas long as it's logging, i'm happy
14:34:16amiconnIt's not high priority, but it worked... once... back in time
14:35:21LinusNyes, i have tried to make it work, but i have so far failed
14:35:43preglowi could whip up an explain bot in perl in a jiffy, but i don't think it's very important...
14:40:36 Quit SereR0KR ("XChat Aqua")
14:41:08 Quit Jungti1234 ()
14:42:58didjwho should i hassle for info about H300 hardware/low level stuff? :P
14:43:25*petur points to Linus
14:44:23didjso linus, any ideas on prev q (reading line-in during playback) ;) ?
14:44:23 Quit scf[aw] (Read error: 104 (Connection reset by peer))
14:45:36 Join scf[aw] [0] (
14:46:31Papricacould someone test tetrox on archos target?
14:49:33linuxstbPaprica: Tetrox works well on my 5g, but seems slow.
14:50:32Papricais it look nice on the 5g?
14:50:53linuxstbYes - everything looks nice on the 5g :)
14:51:09preglowand closing of mouths began
14:51:13preglowlest preglow should buy a 5g
14:51:14LinusNdidj: read line-in, as in record?
14:51:23LinusNdidj: or just mix?
14:51:32Papricaabout the speed, just need to play with the speed
14:51:36didjjust read the ADC level
14:51:45didjtoying with the idea of using it as a simple bus
14:51:48Papricawith the numbers*
14:51:54linuxstbpreglow: The gigabeat has the same sized screen, and is in need of developers...
14:51:56Papricathat control on the speed
14:52:31preglowi think i'll wait with buying yet another target until i have more time for rockbox again
14:52:54linuxstbPaprica: Do all the targets have the same playing area?
14:53:07LinusNdidj: it might be possible, yes
14:53:56markunpreglow: and has a ARM CPU which I know is a criteruim for you :)
14:54:22markunok :)
14:54:24safetydancriteria is plural
14:54:35linuxstb300MHz makes it too easy though....
14:54:38preglowi'd like something new now, i've learnt arm now
14:54:50linuxstbHow about an ipod shuffle?
14:54:55preglowi now 56k asm as well
14:54:58preglowknow, even
14:55:03Mikachuthe singular is kriterium en swedish :)
14:55:05preglowthere are no tools for it
14:55:08markunlinuxstb: we shouldn't clock it at 300MHz
14:55:36preglowmarkun: know how to clock it yet?
14:55:40Mikachuwhat are they doing with 300 mhz?
14:55:56linuxstbNothing at all in the original firmware by the sound of it.
14:56:16markunpreglow: yes, it's in the manual. Toshiba clocks it at 296 MHz
14:56:21preglowholy shit
14:56:33preglowthat's a bucket of waste if i ever saw one
14:56:34linuxstbThere must be a lot of eye candy then...
14:56:36safetydanblimey... I can see a Quake port for that :)
14:56:59Mikachuthe sliding menus on ipods lag when you play music, maybe they wanted to avoid that
14:57:06linuxstbmarkun: Any idea if it is constantly at 296MHz?
14:57:29amiconnpreglow: Isn't 56k used in the shuffle?
14:57:33preglowamiconn: it is
14:57:35Paul_The_Nerdsafetydan: Quake ran fine on my 486/66 if I recall...
14:57:50markunI don't think it is. It's easy to change the clock to 16 MHz
14:57:50safetydanPaul_The_Nerd, righto then... Quake II :)
14:57:57preglowamiconn: and that pretty much implies you 1) need to code everything in asm, 2) need to code a free assembler...
14:58:04Paul_The_Nerdsafetydan: I had Quake 3 running on a 200mhz AMD...
14:58:05Mikachuyou'd need an aimbot or something for it to be playable
14:58:08 Join _FireFly_ [0] (
14:58:30safetydanPaul_The_Nerd, righto then... Doom 3... nah that's just not worth playing
14:58:32MikachuPaul_The_Nerd: it was okay as long as you didn't have rocket trails on my p2 350mhz with a riva tnt
14:58:51markunI looked at the quake2 code briefly but it's C++ and full of floats
14:59:08Paul_The_NerdYeah, I heard that id code isn't fun to read in general, either.
14:59:11linuxstbA "challenge" then...
14:59:19safetydanWhy start with the C++ version, didn't someone write it in Java at some point? Much easier to port right? Just write a JVM...
14:59:30didjheh @ java overhead
14:59:40Mikachuwhy don't music players have fp in general?
14:59:48preglowpower and space
14:59:52preglowand price
14:59:58markunlet me check if it was c++ or if I'm just making that up now..
15:00:08preglowit was
15:00:14safetydanand Real Men (TM) don't use floating point anyway
15:00:27preglowsafetydan: Easygoing Men do, however
15:00:49preglowafter trying to force the bloody eq plotter to behave, i'm pretty sure i'll never look lightly upon fixed point again
15:00:52Mikachuwould it really take that much space if it was integrated in the same core?
15:00:59preglowMikachu: no, not much
15:01:03Mikachui mean, an athlon core is pretty small
15:01:19preglowyeah, but those are in a completely different league
15:01:34linuxstbBut why would you need an FPU?
15:01:40preglowordinary embedded cpu makers use process technology that is waaaay older
15:01:43safetydanan athlon core is also a very small process (60 nm now?) and hugely expensive per unit
15:02:19preglowbut yeah
15:02:42preglowdoing it in fixed point is cheaper and usually faster to boot
15:02:42safetydanI imagine the mp3 player manufacturers are looking at chips in the $5 per unit range rather than the $100 per unit range of an athlon
15:02:46safetydan(figures made up)
15:02:47preglowunless you've got a really posh fpu
15:03:40markunhm, quake 2 is not c++, but it is full of mallocs and floats.
15:03:56preglowmarkun: it isn't?
15:04:04preglowmarkun: so quake3 isn't either, then
15:04:19linuxstbmallocs aren't that much of a problem (we just implement malloc) - but obviously floats are.
15:04:30markunpreglow: did you see this?
15:04:44Mikachui would suspect quake3 has one or two places with x86 asm, but maybe the icculus port has taken care of that already
15:05:07preglowi still say we implement haxx malloc in the plugin library
15:05:18preglowit works, no more bloody free() hack problems
15:05:24preglowwe've already had them and we'll have more
15:05:24Paul_The_Nerd"Software Optimization - We have converted the entire codebase over to fixed point arithmetic for faster execution on handhelds."
15:05:27linuxstbpreglow: I'm agreeing with you now.
15:05:37preglowmarkun: nicety
15:05:41MikachuPaul_The_Nerd: what game?
15:05:44Paul_The_NerdQuake 3
15:05:55Mikachuah, missed that line
15:06:31Paul_The_NerdMan, even at completely unplayable speeds, being able to show someone Quake 3 on an MP3 player would just be something else.
15:06:57Mikachumake a sequence of jpg files :)
15:07:00linuxstbHow much RAM does it need?
15:07:21 Join DrMoos [0] (
15:07:40Mikachuthink of the largest number you can think of, then multiply it by 8
15:08:06Mikachu(three dead trolls in a baggie - how to buy a computer)
15:08:16preglowquake on gigabeat might be entirely realistic
15:08:17 Quit adiamas (Connection timed out)
15:08:22preglowmarkun: how much ram's it got?
15:08:29safetydan48 MB on PC was the minimum requirement for Q3
15:08:29linuxstbAh - q3ce requires "40MB of available working memory"
15:08:33Mikachuhm, with the position sensitive cross, maybe you could even aim
15:08:33markun32MB only
15:08:48linuxstbLooks like the 60GB 5g then...
15:09:02Paul_The_NerdWell, of Q3 could be reworked, you just have to find a fixed point port of Q1 or Q2
15:09:08Mikachuif it's threaded maybe it would run will with all four cpus
15:09:51 Join adiamas_ [0] (n=adiamas@
15:10:08 Nick adiamas_ is now known as adiamas (n=adiamas@
15:10:43Paul_The_NerdARM and SH3 binary versions of Quake and Quake 2, as well as source
15:15:43 Join Febs [0] (
15:16:26markunPaul_The_Nerd: the source is also full of floats
15:16:47markunof quake2 at least
15:17:54Paul_The_NerdThe Quake 1 one is supposed to have "Fixed point optimizations" haven't looked at the source
15:18:29Bg3rPaul_The_Nerd planning to port Q1 to rb ? ;)
15:19:53Paul_The_NerdBg3r: If I manage to find an implementation where someone's already done the fixed point gruntwork, sure. ;-)
15:20:19Paul_The_NerdApparently though, this is not it
15:21:15Bg3ri can't imagine this working on irivers/ipods ...
15:21:55Paul_The_NerdNot in a playable way.
15:22:34Bg3rthat's what i mean
15:22:54safetydanActually where's the Wolfenstein port?
15:23:20safetydanor is there a doom wad to simulate that?
15:23:56didjsimpsons doom is pretty cool :)
15:24:00Paul_The_Nerdsafetydan: There could be, as there was a Wolf3D level in Doom 2, but you'd really have to support Dehacked patches to get the full Wolf3D experience.
15:24:32Mikachuthere was a wolf3d port to snes
15:24:38Mikachuso the cpu should be fast enough :)
15:24:42Mikachui think snes had 3 mhz
15:25:28 Quit Moos (Read error: 110 (Connection timed out))
15:25:39Mikachuoh, it's actually 1024 hz :)
15:25:54Mikachu"#$#" that was the sound cpu, the main cpu runs at about 3
15:31:55midkayminimum requirements aren't necessarily _the_ minimum requirements! :)
15:32:19Paul_The_Nerdmidkay: *especially* with Quake3 and earlier games from id.
15:32:27amiconnwolf3d ran fine on 286
15:32:55*Paul_The_Nerd is *pretty* sure the AMD he ran Q3 on had 32mb of ram, but couldn't guarantee that.
15:32:58midkayPaul_The_Nerd, iirc doom 3's requirement was a 1.3 or 1.4ghz PC.. iirc even out of the box someone sub-1ghz had it running (of course quite slow, though)..
15:33:10midkayq3 would be so unbelievably cool :)
15:33:17Mikachuis wolf3d opensource?
15:34:49midkayand you could probably modify it to use less memory (e.g. lower the texture res a bit)..
15:34:55midkayanyways, got to run.. laters
15:35:06 Join Matze [0] (
15:35:26Paul_The_NerdMikachu: I see various sites saying they've got the source code. I don't know what licenses though.
15:35:57Mikachui found one in gpl, but it's opengl ported :)
15:36:07Mikachuit's very likely that this one is gpl too,
15:36:48Mikachuhahaha, wtf
15:36:48 Quit adiamas (Read error: 104 (Connection reset by peer))
15:37:02Mikachuwho would pay $20 for wolfenstein 3d now?
15:37:11Paul_The_NerdFrom what I've read, Wolf3D was released under the same license Doom was later released under. Then doom was re-released as GPL.
15:37:21Paul_The_NerdMikachu: Does it come with Spear of Destiny?
15:37:30Paul_The_NerdYes it does. So I would
15:37:47MikachuSoD is listed separately at $15
15:38:41Paul_The_NerdMikachu: SoD is moderately hard to find, and has many more levels if I recall
15:38:50Mikachui've actually never heard of it before
15:38:57 Join tvelocity [0] (
15:39:19 Quit sharpe (Read error: 110 (Connection timed out))
15:39:20Paul_The_NerdIt was the expansion. If I recall it had twice as many episodes.
15:40:46Paul_The_NerdI may be wrong though
15:41:15Paul_The_NerdOh, I am wrong
15:41:32Paul_The_NerdWolfenstein3D had 30 levels. "The Nocturnal Missions" were another 30. Then Spear of Destiny was 21 on its own
15:43:01Mikachuah again
15:45:09 Join gracchus [0] (
15:46:59Paul_The_NerdMan, 640k RAM required. Wolf3D was a beast.
15:48:32*preglow doesn't miss qemm
15:48:59 Quit Sinbios (Nick collision from services.)
15:49:23 Join Sinbios [0] (n=Sinbios@
15:49:27Mikachui used to love pressing that optimize button and playing with the load order
15:49:46 Quit Matze ("Miranda IM! Smaller, Faster, Easier.")
15:49:48 Join mirak [0] (
15:53:08 Quit mirak (Remote closed the connection)
15:58:55 Quit needleboy ()
15:59:40 Quit XavierGr (Read error: 110 (Connection timed out))
16:01:54BHSPitLappyhello, IRC.
16:06:11t0mas[11:15:08] <B4gder> -I've mailed t0mas about it, we need more detailed logs
16:06:15t0masJust writing my reply
16:06:16 Nick CoCoLUS is now known as CoCo^uni (
16:06:27 Join quobl_ [0] (n=quobl@tor/session/x-92587ec08ca9f140)
16:08:13 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
16:17:42 Join Genre9mp3 [0] (
16:17:52***Saving seen data "./dancer.seen"
16:17:57 Quit Genre9mp3 (Client Quit)
16:22:08 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
16:22:32 Join DreamTactix291 [0] (
16:25:51 Quit quobl (Remote closed the connection)
16:27:01 Join webguest14 [0] (
16:27:35webguest14How about this for wolf3d: - a port to S60 cellphones
16:28:24webguest14okay, upon further inspection, that's based on the icculus port
16:28:41webguest14someone should just port sdl to rockbox and be done with it
16:28:55webguest14that way we could run the rockbox sim on rockbox and make the world implode
16:29:40 Part LinusN
16:39:52preglowamiconn: i'm not sure how to make the int_prio parameter work on arm, is it important?
16:42:06amiconnint_prio isn't used on coldfire either
16:42:45amiconnLines 173/174:
16:42:46amiconn#elif defined CPU_COLDFIRE
16:42:46amiconn (void)int_prio;
16:43:07preglowreg_prio is what?
16:43:56amiconnIt's the 'registration priority', used for sharing the timer
16:44:21amiconnThis should work the same way on all targets
16:44:33preglowso something can hijack it
16:47:06 Quit Ashex ()
16:47:29amiconnYes, and in that case the unregister callback is called if it is non-NULL
16:47:47 Quit fantomas (Read error: 110 (Connection timed out))
16:48:29*preglow wants datasheets....
16:50:12 Quit Febs (Read error: 104 (Connection reset by peer))
16:51:15 Join Matze [0] (
16:52:09 Join xmixahlx [0] (n=xmixahlx@
16:57:04 Quit darkless_ (Read error: 113 (No route to host))
16:59:34Papricahow can i take the 96 from the number 1996?
16:59:52Papricathe type of the variable is integer
17:00:35preglowvalue % 100
17:00:42Papricaoh right
17:02:50webguest14Haha, so simple. My head was already turning and spinning looking at the question.
17:05:48preglowheh, it helps having used fixed point
17:06:26 Join Nico_P [0] (
17:06:48 Quit tipi^ (Read error: 104 (Connection reset by peer))
17:06:49Mikachuwould % 0x100 and & 0xff be the same thing?
17:07:07preglowthe latter being way faster
17:07:23preglowand the reason people prefer to use power of two based fixed point
17:07:26Mikachumaybe the compiler will optimise it if you have that constant hardcoded
17:07:46Mikachuit would make no sense to use fixed point in a nonnative base i guess
17:07:54preglowpeople do it all the time
17:07:59preglowour old cube.c did it, for instance
17:08:03preglowusing base ten fixed point
17:08:11preglowi believe amiconn bettered iot
17:08:46preglowit's a bit simpler to understand if you're not fully familiar with anything but base 10 numbers
17:09:27Mikachumaybe no sense is a bit strong :)
17:09:40pregloware you a coder?
17:10:14Mikachuyeah, but lowlevel things are a bit confusing still
17:10:27 Quit Zagor ("Client exiting")
17:10:30Mikachui'm in charge of the openbox wm
17:10:46*amiconn considers lowlevel code most of the fun
17:11:09 Join Rondom [0] (
17:11:24Mikachumostly by accident
17:11:37linuxstbMikachu: So are you only interested in projects with box in their name?
17:12:07HCl :o
17:12:55Mikachulinuxstb: that seems to be the trend :)
17:13:12Mikachui usually join the irc channel for new programs i use
17:13:34Mikachuand try to help and/or submit changes i make that i think make it better
17:14:04Mikachuand also to try and affect the development in directions i want :)
17:14:24linuxstbNot interested in combining openbox and rockbox?
17:14:45Mikachunot sure how that would work :)
17:16:26preglowfor some reason i didn't know that
17:16:31Rondomopenbox is windows manager?
17:16:57preglowcan't say i have a preference for what code i like to do
17:16:57 Join ][milwaukee][ [0] (
17:17:08][milwaukee][what's up
17:17:25preglowdsp code and 3d graphics are my favourites, i believe
17:17:44][milwaukee][through rockbox, isn't it theoretically possible to write in the ability to play video on non video ipods?
17:17:53Mikachucan you make the cube plugins speed depend on peak levels?
17:17:55RondomMikachu: isn't openbox the successor of blackbox, a window manager?
17:18:12Paul_The_Nerd][milwaukee][: "Theoretically" yes. Likely to happen soon, not so much, since that's not a real focus of the project.
17:18:17MikachuRondom: openbox2 was a fork of blackbox 0.60something, but openbox3 is a rewrite in c
17:18:32Mikachuand then the guy who wrote it didn't feel like doing it anymore so i took over
17:18:33][milwaukee][Paul_The_Nerd: what goes into it?
17:18:44Rondomdo you want to run a window manager on a portable player?
17:18:50Rondomthe makes no sense to me
17:18:58Paul_The_Nerd][milwaukee][: Someone has to create a fixed point decoder for their video format of choice that runs fast enough on the ipod processor to be usable?
17:19:15MikachuRondom: no, it was just in reference to the question about me being a coder :)
17:19:43Rondomsorry Mikachu linuxstb asked
17:19:47Mikachuand not doing so much lowlevel things
17:19:48][milwaukee][Paul_The_Nerd: Apple did it, how much faster are the videos than nanos or minis?
17:19:57Rondomall questions were actually directed to him, but nvermind
17:20:36Paul_The_Nerd][milwaukee][: The video has a special chip in it that's designed more for that purpose.
17:21:00MikachuRondom: that makes no sense since you adressed me and then continued talking without adressing anyone else
17:21:26][milwaukee][Paul_The_Nerd: ah, emulating that would be ridiculous, nevermind
17:21:56][milwaukee][next question, i just picked up a nano, previously i've owned a slew of mp3 players but never an ipod
17:22:22][milwaukee][is it possible to throw a little program onto the nano that keeps together my workout schedule, allowing entry of reps and what not?
17:22:27Paul_The_Nerd][milwaukee][: Emulating it would be pointless anyway. It'd be slower than just writing a proper decoder to run in software. It's been done with ipodlinux, to an extent.
17:22:46][milwaukee][or do i have to run a non-apple firmware to consider it?
17:22:53RondomMikachu: all questions were actually meant to be directed to linuxstb
17:23:12Paul_The_Nerd][milwaukee][: This channel's basically dedicated to non-apple firmware.
17:23:43Paul_The_NerdI don't even know how you install third-party apps like the various 'games' that Apple includes
17:25:19][milwaukee][so no way even with non-apple firmware?
17:25:24Paul_The_Nerdlinuxstb: Apparently pacbox doesn't get included with "make zip" for 5g with the source tarball...
17:25:40Paul_The_Nerd][milwaukee][: With non-apple firmware it's a cake walk if you can program.
17:25:44linuxstbPaul_The_Nerd: It has done for me ever since I started developing it...
17:26:21Paul_The_Nerdlinuxstb: I just tried it. Downloaded a source tarball, extracted, compiled, make zip, extracted that zip somewhere else, no pacbox.rock in the zip. Works fine from CVS though
17:26:21linuxstbI can't see how the source tarball would differ in that respect to the CVS.
17:26:52Mikachumaybes omething is missing from MANIFEST if you sue that to make the tarball
17:27:06*Mikachu makes a random guess
17:27:29Mikachumaybe there isn't such a file.. i thought i saw it though
17:27:31preglowlinuxstb: you know if ipl kernel is still in sf cvs?
17:27:47preglowi need to find the irq number for timer2 on pp5002
17:29:37 Quit damaki_ (Read error: 110 (Connection timed out))
17:30:36linuxstbpreglow: A quick grep doesn't appear to show any references to TIMER2 for pp5002 - only pp5020.
17:31:14preglowlinuxstb: lucky us, then
17:31:27preglowthen someone with a pp5002 needs to just test until it works
17:35:01preglowi don't even know if timer2 works at all, for that sake
17:35:03preglowtime to test
17:35:59preglowsf cvs is so bloody unusable
17:36:11preglowwhy do they bother maintaining a service that never works?
17:36:28webguest14I submit that they don't.
17:37:44 Quit ][milwaukee][ ("CGI:IRC (EOF)")
17:40:33linuxstbPaul_The_Nerd: How big is your pacbox.rock compiled from the bleeding edge source?
17:40:44linuxstb(I mean the source tarball)
17:41:36 Quit ara-fat32 ("Fnord!")
17:41:41Paul_The_Nerdlinuxstb: Aha, 0 bytes. That might be the problem
17:41:56linuxstbYes - that's the problem...
17:41:57Mikachuit works most of the time..
17:42:58linuxstbPaul_The_Nerd: Is pacbox_arm.S in your source file?
17:43:11Mikachuand by it i mean sourceforge
17:43:40Paul_The_Nerdlinuxstb: Yes.
17:44:09linuxstbI'm just downloading the bleeding edge source now. Are any other plugins 0 bytes?
17:45:03Paul_The_Nerdlinuxstb: firmware_flash and rockbox_flash, but I supposed those are probably supposed to be.
17:46:00linuxstbAh, I found the error - pacbox/SOURCES is missing
17:46:22linuxstbThe build system clearly displays it.... (before it scrolls off the screen)
17:46:29Mikachui was close then
17:47:38linuxstbFix committed.
17:49:14preglowwebguest14: yeah, perhaps that's it...
17:50:38linuxstbDo Sourceforge have an SVN service now?
17:51:26linuxstblinuxstb: Yes, they do.
17:51:46Mikachuthe command they list isn't good to run though
17:52:08Mikachuit'll download all branches and tags (i noticed after 232M of gphoto2)
17:52:53linuxstbDon't you just do "svn co http://....." ?
17:53:17Mikachufirstly they use https so i had to upgrade svn, then the command they list is like http://.../projectname projectname
17:53:29Mikachuinstead of /projectname/trunk/module projectname
17:54:49Mikachufor cvs they list this
17:54:52Mikachucvs -z3 co -P modulename
17:54:57Mikachuso it's obvious what part to change
17:55:14Mikachubut for svn they say
17:55:15Mikachusvn co gphoto
17:55:31Mikachuand what i wanted was
17:55:31pregloware the paranthesis in #if (CONFIG_BACKLIGHT == BL_IRIVER_H100) && !defined(SIMULATOR) really necessary?
17:55:33Mikachusvn co gphoto2
17:56:13Mikachunot sure if == takes precedence over && in the preprocessor but it should
17:56:26preglowwell, i've never used the parantesis, and it's always worked
17:56:26 Quit webguest14 ("CGI:IRC (EOF)")
17:56:34peturI find it easier to read
17:56:35Mikachuyou need the ones around SIMULATOR though
17:56:41preglowyeah, sure
17:57:54linuxstbI've seen lines like that both with and without the parenthesis all over Rockbox. Personally I prefer to add them.
17:57:54preglowahh, hooray, every ipod generation has its own BL_IPODxxxx
17:58:19linuxstbApart from the 5g, which I think has the Nano's backlight.
17:58:25preglowi don't add them in ordinary c code, so i don't add them in preproc directives either
17:58:52linuxstbI do add them in ordinary C code, so I add them in the preproc directives...
17:59:18peturI do too
18:00:00linuxstbI'm sure there's some language I use that requires them, so I'm in the habit. But I've long forgotten which language that was...
18:00:37Paul_The_NerdI make far too much use of parenthesis because I have friends who can't work out order of ops properly in their heads for many things I end up writing for them.
18:01:20Mikachuif you write code and you think "hm, are these needed?", then they are needed :)
18:01:31preglowlinuxstb: you know what's up with the dual outls in ipod nanos backlight switch code?
18:02:21preglowi would have thought flipping one bit would be enough
18:02:37Mikachui tried commenting either out one of them
18:02:41Mikachuand they are needed :)
18:02:45linuxstbNo idea... It's just the usual copy/paste...
18:02:46Mikachuout either*
18:03:09Mikachuin one case it almost broke the backlight in apple os, at least it wasn't on at startup but when i touched the pad
18:04:05 Quit petur ("work day eof")
18:04:27Mikachuwhat i wonder is how to control the intensity, but that's probably a luxury feature
18:05:53 Quit linuxstb ("Client Exiting")
18:09:36 Quit tvelocity ("Ex-Chat")
18:10:59 Join Aditya [0] (
18:10:59 Quit Aditya_ (Read error: 104 (Connection reset by peer))
18:12:59preglowyeah, perhaps
18:14:08 Quit quobl_ (Remote closed the connection)
18:16:08preglowa bit weird that intensity is set by flipping a bit
18:17:03Mikachuapples firmware can do it in steps...
18:17:07preglowi need a drink for the cold
18:17:14Mikachuso i was thinking one might want to set an intermediate intensity to save battery
18:17:33lostlogicwhee, about to do proper charger input and charge state detection for ipod video!
18:17:56***Saving seen data "./dancer.seen"
18:18:17preglowyou can set intensity?
18:18:20preglowwhy'd i never see that?
18:18:30Mikachuno, that is what i want
18:18:43preglowoh, it sounded like you knew it was possible
18:18:43Mikachui think it might be possible since apple os fades in and out
18:18:50preglowbut that can be done by pwm
18:19:03preglowpulse width modulation
18:19:08preglowlike we do backlight fading on iriver
18:19:15Mikachuhm, you think that's what they do?
18:19:18preglowjust turn the backlight on and off fast enough and it looks like fading
18:19:24preglowyeah, i think that's what they do
18:19:36Mikachuit would use some cpu power to set a lower intensity then
18:19:43preglownot necessarily
18:19:55Mikachudoesn't someone need to flip the bit for it to be... flipped?
18:19:59preglowthe cpu would need to run at a constant freq for a nice even intensity
18:20:03preglowso it needs to stay boosted...
18:20:08 Join Febs [0] (
18:20:30preglowyou can't set the fade time
18:20:43Mikachuwould you flip it in the timer interrupt?
18:20:54Mikachudo you even have a timer interrupt?
18:21:15preglowwould be a bit hard to do fading without it
18:21:21Mikachuthat's what i was thinking
18:21:22preglowit's basically what i've just codec
18:21:31preglowported to ipod, at least
18:21:34Mikachuin this short time since i asked?
18:22:23preglowna, did it an hour ago
18:23:19Mikachubut just fading in/out then?
18:25:09Mikachuif you enable the backlight while it is fading out, does it switch direction, or wait until it finishes fading out, or jumps to the start of fading in?
18:25:32Mikachunot the world's most consistent grammar
18:25:33 Quit didj (Remote closed the connection)
18:25:44 Join didj [0] (
18:25:45preglowswitch direction
18:25:52Mikachugood :)
18:29:49preglowseems the portalplayer address decoder allows a whole bunch of aliasing to go on
18:31:02Mikachuwhat's aliasing in this context?
18:31:05preglowthe backlight code seems to use 0x6000d824 for writing to port b
18:31:18preglowwhile we seem to use 0x6000d024 other places
18:31:46preglowperhaps the former is some special way of flipping bits instead of using an and
18:31:49preglowwould save some code size
18:35:16lostlogiceventually I'm going to have to make different charge current values for USB power vs. wall power
18:35:26lostlogicthat applies to both H3x0 and IPV
18:36:42 Quit safetydan ("Leaving")
18:41:24lostlogicshould backlight timer control use the plugged timeout when USB powered?
18:41:25lostlogicI think so
18:41:30lostlogicit doesn't now.
18:41:54 Join Genre9mp3 [0] (
18:42:02 Quit Genre9mp3 (Client Quit)
18:46:47Mikachuyou can set different timeouts?
18:56:53amiconnpreglow: You don't necessarily have to keep the cpu boosted for doing backlight pwm.
18:57:10preglowamiconn: then why are we doing that at the moment?
18:57:12amiconnThat depends on the cpu, or precisely whether the timer depends on the cpu clock or not
18:57:34amiconnOn coldfire we need to, since we have to switch prescalers when switching frequency,
18:57:54preglowyeah, nice point
18:57:58amiconnand we don't know *exactly* when the frequency transition happens, as it depends on the pll relock
18:57:58preglowi disabled the prescaler for portalplayer
18:58:02preglowsince i don't know if we've got one
18:58:12preglowit was just dead code
18:58:35amiconnAfaik the pp timer is cpu frequency independent and works in microsecond units
18:58:42amiconn(you told me)
18:58:56Mikachucan't you just adapt the frequency to the boost state?
18:59:23amiconnMikachu: That we do, but there's an uncertainty *during* the switch
18:59:39Mikachuah, and you don't want the display to flicker
19:00:04Mikachuthis would be a bigger issue if you wanted to make a persistent lower intensity
19:00:14preglowit is, yeah
19:00:20preglowtimer tick works just nice when boosting now
19:00:29amiconnThis instability is bearable when the timer period is larger than the pll relock time, so we don't need to boost for the grayscale library
19:00:32preglowwith no adjusting
19:00:37amiconn(timer period 70 Hz)
19:00:38 Join quobl [0] (n=quobl@tor/session/x-4f85c691f632c143)
19:00:58amiconn..but for the backlight pwm, the timer period is much shorter
19:02:13amiconnIt's (1..99) * 50us
19:03:09 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
19:03:17Mikachuhow long is that relock time when switching frequencys? have you tried it to see it really is noticable?
19:03:28amiconnYes it is noticeable
19:03:40amiconnTypical relock time is around 2ms, maximum is 10ms
19:04:02Mikachuthat's quite a while
19:04:25amiconnYes, and until the pll relocks, cpu clock is the base frequency
19:04:52amiconnI.e. when switching 45 -> 124 MHz, we do in fact switch 45 -> 11 -> 124 MHz
19:06:05amiconnlostlogic: I see RED!
19:06:28Mikachuand it always takes different time to switch?
19:06:39Mikachuthat is a bit annoying
19:10:49 Quit TiMiD (Read error: 104 (Connection reset by peer))
19:17:56 Quit Matze ("Miranda IM! Smaller, Faster, Easier.")
19:17:57 Quit Paul_The_Nerd ("Leaving.")
19:20:38preglowthere's some natural variation involved, yeah
19:20:43preglowipod pll seems to use 2ms to relock
19:21:46Mikachuif you guessed on the average value and used three frequencies would it still be noticable?
19:22:03 Join Genre9mp3 [0] (
19:22:05Mikachubut this stuff is not needed on ipods if i understand correctly?
19:22:14 Quit Genre9mp3 (Client Quit)
19:22:28preglowit is
19:22:56preglowall devices capable of adjusting the cpu freq pretty much needs to have a pll
19:22:57Mikachuthe timer interrupt is dependant on the cpu frequency?
19:23:06preglowcan't think of any other worthwhile ways to synthesize a clock
19:23:38 Join herz42 [0] (
19:23:38preglowdependent on external crystal in the case of iriver, and on ipod it's a microsecond timer
19:24:17amiconnOn iriver the timer clock *is* dependent on the cpu frequency
19:24:32Mikachui'm not sure if i understand why the pwm won't work then, but my understanding is maybe not important
19:24:43preglowyou said it's based on the xtal frequency...
19:24:57lostlogicamiconn: fixage pending.
19:25:12amiconnThe value you pass to timer_set_interval() is based on the xtal frequency
19:25:30 Join damaki [0] (
19:25:40amiconnThat's what the prescaler magic is about, and why I changed our boost states to be integer multiples of the base
19:26:33amiconnset_cpu_frequency() calls timers_adjust_prescale()
19:26:42 Quit xmixahlx ("blah blah blah")
19:26:44preglowahh, right
19:26:48preglowi forgot you mentioned the prescaler already
19:27:10herz42lostlogic: seems we are looking at the same codelines recently
19:27:43herz42do you know about the secret port h already? ;)
19:29:34lostlogicherz42: no, I don't
19:29:58lostlogicherz42: does what I've done seem correct for you? Certainly seem to work on my ipod 5g.
19:30:18herz42I've found the bits to detect the various stuff plugged to the ipod
19:30:38lostlogicherz42: gonna wiki it somewhere?
19:30:58herz42registered for the wiki yesterday, was about to add it today :)
19:31:05lostlogicvery cool
19:31:23lostlogicI'm probably goign to start filling in the ipl wiki's gpio page for the 5g with what I've done so far
19:32:48herz42hmm, I was not sure where I should add the info (ipl or rb)
19:32:56 Join control [0] (
19:33:03herz42I didn't too much for the 5g on ipl docs yet
19:34:52herz42you check port a bit 4 for charger_connected. Actually I think this is the detection of a dock-device (don't know how to call it).
19:36:10herz42I'm using port h bit 3==0 for ext_pwr and port h bit 4==1 for usb_pwr (also only tested on my 5g)
19:39:11herz42btw: port h is 'three' behind the "GPIO port E ???" from ipl-wiki pp5020
19:40:26lostlogicherz42: that makes sense −− I guessed the gpioa bit 4 for charger because connecting to my gf's dock set that bit.
19:40:51lostlogicherz42: please if you get a chance submit a patch adding the port h to the debug screen, and/or improving the detection stuff I've been doing.
19:41:00lostlogicherz42: any idea hwo to control the USB current draw yet?
19:41:55herz42no, I didn't risk to try driving gpio pins yet...
19:43:38herz42I can add the patch for debugging soon
19:45:12herz42For charging my estimated charging time from usb is 59hrs atm. Maybe that's true, maybe the characteristics need some finetuning...
19:49:36 Join Aditya_ [0] (
19:49:36 Quit Aditya (Read error: 104 (Connection reset by peer))
19:52:21 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
19:57:29safetydanwell that's interesting... needleboy isn't going to be doing the experimental builds any more
19:57:55safetydanNow who's going to assemble all the dangerous patches for testing? :)
19:59:03Mikachuare all the h300 kids panicking now?
19:59:58safetydannot sure about panicing, but there is consternation certainly
20:05:46herz42lostlogic: debugging patch is added
20:06:11herz42hopefully nothing is missing (the albumart makes diffing quite diffcult atm)
20:17:36herz42lostlogic: I just saw you have started to add g5 on the ipl gpio page. If you don't mind, I'll put my additions there.
20:17:57***Saving seen data "./dancer.seen"
20:18:31lostlogicherz42: go for it :)
20:19:00lostlogicherz42: I probably won't get to adding the debugging patch until tomorrow evening at devcon, just fyi.
20:19:10herz42do you think it makes sense to add something like 'active high' / 'active low' info? And if yes, how?
20:19:32lostlogicherz42: also, I charge via USB while playing on rockbox ipod 5g I get about 6 hour charge time, so I wonder if mine is behaving differently than yours.
20:19:44lostlogicherz42: yes, I htink that should be there, but I don't know how, you'd hafta ask the ipl peeps
20:19:47 Quit Rondom (No route to host)
20:20:39herz42I think charge time is a matter of the inputs to the calculation
20:20:50preglowadd whatever you can
20:20:57preglowthe ipl people don't seem to update it themselves
20:20:59 Quit damaki (Read error: 104 (Connection reset by peer))
20:21:39herz42In my calculation, I'm subtracting the (guessed) 80mA for operation from the 100mA charging current
20:21:40 Join damaki [0] (
20:21:56herz42If backlight is static on, it will never charge :(
20:22:54preglowlostlogic: yo, would you care to test if changing the (0x100 | 1) to just 0x1 in the backlight code for 5g still has it working?
20:23:08preglowit makes no difference on nano
20:23:17preglowand afaik, 5g shares that code with nano
20:25:15herz42preglow: I also didn't understand that code
20:25:53herz42or are there more than 8 significant bits for writes?
20:27:24amiconnoutl means 32bit I'd think
20:27:58herz42yes, but when reading 32bit, only the 8 lsbs are != zero
20:28:12herz42so I'm just not sure if writing the upper bits has some other effect
20:29:06lostlogicherz42: yeah, I did the calculations too, but... it still charges at a reasonable pace :-\
20:29:13preglowamiconn: i think portalplayer port registers are 8 bit wide, even though the register itself allows (and probably requires) 32 bit writes
20:29:16lostlogicherz42: I also didn't measure the input current, so maybe mine has a setting different
20:29:21lostlogicpreglow: will test
20:29:43preglowi the registers are not 8 bit wide, then i want to know how the hell they can have over five registers
20:29:51preglowthe entire pin base would have to be gpio configurable
20:30:05lostlogicpreglow: outl((0x1 << 3), 0x6000d824);
20:30:10lostlogicshould be the new line?
20:30:13preglowlostlogic: yea
20:30:20preglowlostlogic: for both register writes
20:30:32preglowthat is 0x1 << 7 in the other
20:31:15 Quit Nico_P ()
20:32:05lostlogicsame should apply to 4g as well then?
20:32:45preglowprobably, but don't know
20:33:05preglowthe top bits might have some other meaning we don't know of
20:33:12amiconnpreglow: has the timer info for PP5002
20:33:51lostlogicpreglow: no apparent effect here.
20:34:10lostlogicpreglow: wait, I lied
20:34:52lostlogicpreglow: it comes on fine, but then when it times out, it won't come back.
20:36:00Mikachui guess the bootloader sets the bit and then it stays on until it times out and then you don't reenable it?
20:36:33lostlogicah yeah, I didn't install a new bootloader
20:39:17preglowamiconn: only for timer 1
20:39:28preglowlostlogic: right, i didn't either, but it still worked
20:39:52lostlogicpreglow: lemme test something
20:39:56preglowthen i think the top bits probably works as a mask
20:40:02amiconnpreglow: I can read 'timer2 ...' in the "Timer Controller" table
20:40:31preglowamiconn: sure, i'm talking about the irq number
20:40:35preglowamiconn: and that you can't find
20:40:52preglowi just assume it follows timer1
20:41:23preglowbut yeah, the top bits might describe a write mask of sorts, that would explain why you do a (0x101 << x) to set one bit
20:41:55preglowbut that would not explain how you can also just not use the top bits and still have it work
20:41:56crashdevening all
20:41:59preglowcrashd: hi
20:42:12crashdohh, watcha working on ?
20:42:36preglowipod timer.c module
20:42:49*crashd reads up
20:43:07amiconnpreglow: Hmm, not necessarily
20:43:35amiconnser0 and ser1 interrupts use adjacent bits on PP5020, but not on PP5002
20:43:48amiconnSo this might be the case for timer0 and timer1 as well
20:43:53preglowamiconn: exactly, which is why someone with a pp5002 actually needs to test it
20:44:14preglowi'm just going to pretend my assumption stands, if it doesn't, then it's a small thing to fix by trying
20:44:28 Join Rondom [0] (
20:45:18lostlogicpreglow: it's not the top bits
20:45:35preglowlostlogic: hmm?
20:45:36 Quit Sinbios ("If the definition of a klutz is someone who doesn't have eyes on their ass, then yes, I suppose I am a klutz.")
20:45:36lostlogicit works again just by changing the 0x1<<3 to 0x101 << 3 which is still within 8 bits.
20:45:54preglowlostlogic: in what way is 0x808 8 bits?
20:46:05lostlogicoh, wait, I'm a moron.
20:46:14lostlogicI'm looking at it and seeing 3 bits
20:46:40herz42hmm, that BL bit is named L07. So then my port H is actually port L :)
20:46:43Mikachubit operations on hex is confusing, but less confusing than bit operations on decimal :)
20:47:19herz42anybody knows if there is some functionality in the port e−−h range inbetween then?
20:47:21preglowbit operations on hex is really easy, if you ask me
20:47:34lostlogicyeah, it's easy, I'm just a moron sometimes.
20:47:38preglowi find it harder to do bit operations on actual bits thanks to the length of the numbers, heh
20:47:41amiconnpreglow: I'm not sure, but maybe it's possible to partially compensate for the broken cache in PP5002
20:47:47preglowamiconn: how
20:47:56preglowthom's premature enter key dealings
20:47:59 Quit control ("CGI:IRC")
20:48:12amiconnLooking at the product briefs, PP5020 is capable of running up to 80MHz while PP5002 is capable of running up to 90MHz
20:48:51preglowwell yeah, but it's still a far way from being compensated for
20:49:24preglowideally, it should work out ok
20:49:32preglowif we use two cores...
20:49:47amiconnPP5002 is USB1.1 only...
20:50:07preglowyeah, apple probably screams "firewire!" out you if you ask them why
20:50:40*amiconn could actually use firewire at no extra cost
20:50:59amiconnMy sony laptop has an port (as sony calls it)
20:51:06preglowi've got a firewire slot here as well, but no cable
20:51:11preglowand no capability in the nano either way
20:51:16preglowa fact i'm totally at ease with
20:51:37amiconnYeah, the nano has USB2.0
20:52:03preglowyou never get a very grand throughput thanks to the flash anyway
20:52:07*preglow has another drink for the cold
20:52:28amiconnI would like to compare the nano's usb transfer speed with the Ondio
20:52:33amiconnOndio is sloo-ow
20:52:36 Nick scf[aw] is now known as scf (
20:53:39preglowit's not that slow
20:53:41preglowjust slowish
20:53:55Mikachui can transfer a couple of bytes and check
20:54:08preglowi'd be happy to do a test, but linux' usb storage behaviour makes it a bit hard to get consistent results
20:54:11preglowperhaps with a large file
20:55:01amiconnOndio is way slower than archos player, and the player is usb1.1 as well...
20:55:22preglowi think nano beats that, then
20:55:31preglowfor reading as well?
20:56:16Mikachui got 2.9MB/s
20:56:20Mikachuon a 100MB file
20:56:28Mikachunot forgetting to include the time to run sync
20:57:43Mikachu3.6 on the second run
20:58:05Mikachui'll try a larger file
20:58:14linuxstbIs this in the "emergency disk mode" or in the Apple firmware?
20:58:18Mikachuapple firmware
20:59:06 Join erus` [0] (
20:59:12Mikachuin the emergency mode i thought it had crashed while i was just copying rockbox.ipod (or possible the whole dist) over
20:59:19preglowwell, it is the hardware caps we're trying to discover, so use appe fw
20:59:27preglowdiskmode is slow as a dog
20:59:29preglowa really scruffy one
20:59:31Mikachutrying a 800MB file now
20:59:45amiconnOndio read: 780 kbyte/s (76MB folder)
20:59:57preglowamiconn: and i assume writing is no faster
21:00:04preglowamiconn: internals flash or mmc?
21:00:10Mikachui'll remount and try reading it back after
21:00:26preglowamiconn: what about flash? mmcs are pretty notorious at being slow
21:00:35 Join Kohlrabi [0] (
21:00:52Mikachumy gba takes 7 minutes to flash 32MB :)
21:00:56amiconnThe internal flash is just an mmc in a bga package
21:01:13Mikachuthat's 76kB/s
21:01:24Mikachuthat's in windows, with the linux program it takes 45 minutes
21:01:47preglowamiconn: right
21:02:03preglowamiconn: bga with that few pins, or does it pretend to be genuine flash?
21:02:09preglows/pins/balls/ :V
21:02:10amiconnWriting is just a bit slower: 745kbyte/s
21:02:25Mikachui'm at 574MB
21:02:26amiconnbga with only a few more pins than mmc
21:02:44amiconn20 pins iirc, and it has a reset pin which mmcs don't have
21:03:19 Quit erus` (Remote closed the connection)
21:04:08 Quit DrMoos ("Glory to Rockbox !!!")
21:04:21 Join Moos [0] (
21:05:06preglowi think we should consider having a separate games menu, now that we have so many games
21:05:25 Join Shadowarrior13 [0] (
21:05:42Mikachuis there a big reason to not just put all the .rock games in /Games or such?
21:06:16preglowbig, no
21:06:43Mikachui'm not sure which feels more logical to me, via the filesys or the main menu
21:07:03Mikachubut maybe you had that discussion many years ago
21:07:13preglowi've only been here a year :>
21:07:30amiconnInternal flash: 800 kbyte/s read, 650 kbyte/s write
21:08:39Mikachui got only 1.35MB/s now
21:08:52Mikachubut it ran for 10 minutes so i think it should be a good average (write)
21:08:54amiconnMikachu: filesys isn't easily localisable
21:09:12Mikachubut this is the last of my free space, i don't know how much fragmentation affects flash
21:09:25Mikachui haven't removed and added lots of mp3s since i installed rb though, only added
21:09:41amiconnfragmentation doesn't really matter for flash
21:09:46amiconnThere are no seek times
21:09:57Mikachuthat's what i thought
21:10:43 Join michaelc [0] (
21:10:56*Mikachu gets the idea to change rockbox_logo.bmp to a wormeaten apple
21:11:19Mikachulooks like about 4MB/s sustained read
21:11:23linuxstbOne localisable way would be to change "Browse Plugins" into a menu, with each submenu opening up the file browser in a subdirectory of .rockbox/rocks/
21:11:41michaelchi... how's charging coming on the 4G ipods? I've noticed a lot of work on 5G ones in CVS...
21:13:09*linuxstb remembers he's the only dev with a 4G...
21:13:30lostlogiclinuxstb: so every time I make a relatively portable change, you get to relatively port it :-P
21:14:11linuxstbI suppose I do....
21:14:59michaelcguess i'll just keep my eyes opened :)
21:15:22Mikachuokay, to summarize
21:15:42Mikachuon my ipod nano with apple fw, i get 4.15MB/s reading and 1.35MB/s writing
21:15:59Mikachu(on an 800MB test file from /dev/full)
21:17:54Mikachuso it would take 25 minutes to fill a 2GB ipod or 8 minutes to copy it back
21:18:36 Quit michaelc ("CGI:IRC (EOF)")
21:20:09 Quit Moos ("Glory to Rockbox !!!")
21:24:00Mikachuamiconn: happy? :)
21:24:27*preglow learns about the existence of /dev/full
21:24:38Mikachuzero is equivalent for reading
21:24:45Mikachufor writing, zero and null are equivalent
21:24:58linuxstblostlogic: What should GPIOA & 0x10 detect? In power.c you've used it in charger_inserted(), but you refer to the dock on the IPL wiki page.
21:25:03Mikachuwhen reading, null will return EOF immediately, and when writing full will return a disk full error
21:25:09Mikachui think that covers it
21:26:04herz42linuxstb: I have changed that 10sec ago.
21:26:15Mikachuso if you want to waste cpu cycles you can cat /dev/zero > /dev/zero
21:26:35Mikachui get about 15GB/s on that stream according to pv
21:26:56herz42there is another port (L) that can be used to detect usb connection and ext. power connection seperate
21:29:20linuxstbMikachu: You have far too much spare time :)
21:29:27 Join ashridah [0] (
21:30:04 Join SereR0KR [0] (
21:30:31 Join tucoz [0] (n=martin@rockbox/developer/tucoz)
21:32:23lostlogiclinuxstb: actually I believe herz changed it to say dock in the wiki, and he is probably correct, in the interim detecting the dock is a somewhat reasonable way to detect connection to a charger.
21:32:50amiconnMikachu: Nice numbers...
21:32:58linuxstbherz42: I'm just looking at your debugging patch. Is "#if CONFIG_CPU == PP5020" appropriate? Shouldn't it require an iPod?
21:33:42amiconnSo the nano profits from being USB2.0, at least for reading
21:33:56herz42there are other players with pp5020? then of course that should be more specific.
21:34:33linuxstbOK, I'll change that and then commit it.
21:34:34herz42what do you suggest?
21:34:45herz42linuxstb: no, awit
21:35:32herz42I have just changed the port names (and added another four) to pp5020.h to make it consistent with the backlight beeing on port L
21:36:07herz42although I don't know if all those ports are ever used. At least ports e..h do never change on my ipv
21:36:28linuxstbThe same code could be useful on all the PP5020 based devices though.
21:37:08herz42hmm, at least the part with gpios. Detecting the attached devices probably not
21:37:17linuxstbYes, that's what I mean.
21:38:46herz42So what define should I use for that second part? APPLE_IPODVIDEO and APPLE_IPODNANO? or no nano, or is there a more generous define out there?
21:40:46linuxstbWe could just use #ifdef IPOD_ARCH around the whole section - and then adjust the individual checks appropriately as we learn how to do it.
21:46:47 Join [IDC]Dragon [0] (
21:49:11*[IDC]Dragon starts to pack for devcon
21:49:46amiconn'lo' can also be just the last 2 letters of 'hello'
21:51:13[IDC]Dragonseems I haven't missed much serious talk
21:51:27 Quit nnod ("Leaving")
21:51:47 Join sharpe [0] (
21:51:54Mikachumaybe all the serious talk is taken care of and now we do the serious talk
21:52:28herz42linuxstb: I've uploaded the refined patch, so you could have a look...
21:52:28[IDC]DragonBagder: do you read?
21:52:30linuxstbherz42: I think you might be over-optimistic displaying that many lines on a Nano...
21:52:59herz42linuxstb: uups, didn't think about that. So maybe omit the midle part there
21:53:52herz42how many lines does a nano have? on ipv it's just half the screen
21:54:57Mikachui have 9 lines in my wps with 6+12x13
21:54:59Mikachuif that helps
21:55:13linuxstbIt should be OK then. The Nano is just over half the height - 132 pixels high.
21:55:31 Quit ScootScat (Read error: 110 (Connection timed out))
21:56:03herz42btw: is there a way to enlarge that tiny font in the debug screens? :)
21:56:10preglowthat's the system font
21:56:16herz42thought so
21:56:17preglowwe desperately need a new one of those
21:56:27preglowthe one we currently use is for archos screens
21:56:30preglowwhich are tiny^2
21:56:38amiconnThe screens aren't tiny
21:56:43herz42hehe, I remember
21:56:47preglowresolution, that is
21:56:49amiconnThe font is physically larger on archos
21:57:50Mikachulower dpi!
21:57:53tucozFire do look cool on the video.
21:58:08Mikachutucoz: thanks
21:58:11amiconnMikachu: Yes, and additionally non-square pixels
21:58:27tucozcompared to the recorders ;-)
21:59:15 Join Paul_The_Nerd [0] (
21:59:22amiconntucoz: It looks a bit better on archos with the white backlight mod :)
21:59:41tucozamiconn, ah. White backlight is nice.
21:59:44amiconnBtw, are you doing the screenshots in the sim?
21:59:52tucozamiconn, when I do them, yes
22:00:28*amiconn thinks grayscale lib support in the sim was a rather useful addition then
22:00:58preglow'course it was
22:01:00tucozamiconn, certainly.
22:01:02preglowthe more we can use in the sim, the better
22:01:16Mikachuamiconn: is that mod software or hardware?
22:01:33Mikachui hear the nano's backlight isn't white but i haven't noticed :)
22:01:50preglownano has white backlight, yes
22:01:50Mikachui thought maybe it would be possible to compensate the colors displayed
22:02:03Mikachupreglow: that was a confusing answer
22:02:07 Join bluey [0] (
22:02:07amiconnMikachu: ?
22:02:13preglowisn't it white? :>
22:02:19Mikachusomeone said it was bluish
22:02:20tucozI had a look at some new sony players the other day. They do look nice.
22:03:00Paul_The_NerdMikachu: Were they running rockbox at the time?
22:03:10 Join ulli [0] (
22:03:11Paul_The_NerdMikachu: Since y'know... rockbox's default background is a faint blue color
22:03:16Mikachui don't know where i remember it from
22:03:19preglowsure looks white to me
22:03:23Paul_The_NerdIt's definitely white
22:03:26Mikachuyeah, me too, foret i said anything
22:03:27amiconnpreglow: [21:59:55] <preglow> the more we can use in the sim, the better <== playback on amd64?
22:03:39preglowamiconn: i'll fix it
22:03:50preglowamiconn: as far as i can, i'll have a look at it over the weekend
22:03:58preglowif i've got time
22:04:29preglowshould basically just consist of switching all longs to int32_t
22:04:53preglowamiconn: btw, playback works here, it's just libmad that doesn't work
22:05:10preglowahh, no, that's clearly wrong
22:05:13preglowlibmad was the one that crashed
22:05:35amiconnogg plays, but double speed + distorted
22:05:43 Quit ulli (Client Quit)
22:05:45preglowi think it's a long problem
22:05:48preglowit sounds like it, at least
22:06:09tucozIt was one of these:
22:06:25Papricamm someone here with non RTC target?
22:07:18Papricawhich one?
22:07:21tucozPaprica, I have a h120
22:08:00Papricacould you check rockcalendar on your h120?
22:08:01amiconnPaprica: player?
22:08:17Papricahihi no*
22:08:18tucozPaprica, sure. Is the patch on the tracker?
22:08:35Papricathe old version yes
22:08:38Papricathe new one
22:08:55Papricai need to test it on the rtc targets
22:09:06Papricaand then, i hope to commit it =]
22:09:08amiconnPaprica: Hehe, I guessed :)
22:09:21tucozOk, so what should I do?
22:09:23MikachuPaprica: what did you decide about my xobox patch?
22:09:24amiconnBut Ondio would be a useful target to test on?
22:09:47 Join solexx [0] (
22:09:48Papricaamiconn, sure
22:10:13PapricaMikachu, i'll commit it with my changes (i hope tonight)
22:10:19Papricamm sec
22:11:09Papricaamiconn, tucoz,
22:11:11 Join Genre9mp3 [0] (
22:11:12Papricathe bitmaps is there
22:11:26 Quit Genre9mp3 (Client Quit)
22:11:34Papricain a second i will send to you the patch
22:12:25amiconn[IDC]Dragon: ping...
22:13:47Papricaoh, wtf it wants? rockcalendar.c:1778: warning: 'i' might be used uninitialized in this function
22:15:33*amiconn needs to prepare for devcon too
22:15:41 Quit _FireFly_ ("Leaving")
22:16:42preglowlostlogic: you know, it doesn't work for me either if i remove the 1 in 0x100, it does however work if i replace the address with GPIOB_OUTPUT_VAL
22:16:54preglowlostlogic: which suggests that the address used has some special meaning, perhaps being a masked write
22:17:00Papricaamiconn, tucoz,
22:17:59***Saving seen data "./dancer.seen"
22:18:12tucozPaprica, do I apply the bitmaps.diff too?
22:19:39 Join petur [0] (
22:20:02tucozPaprica, that did not apply cleanly
22:20:29Papricai know, i need to make new patch
22:20:48Papricawhen i'll add the new file to FS i'll add a new one
22:21:25linuxstbMikachu: What did your xobox patch do?
22:21:39petur/msg nickserv help
22:21:48Mikachui wrote some comments in the @@ lines,
22:21:59Mikachuie i don't remember offhand
22:22:05 Quit solexx_ (Read error: 110 (Connection timed out))
22:22:28 Part petur
22:22:31Papricacan you download this one?
22:22:46Papricasorry about that, i forgot a 2 lines
22:23:21Mikachulinuxstb: but i have a feeling i forgot to write one thing
22:23:22 Join petur [0] (i=petur@rockbox/developer/petur)
22:26:13linuxstbMikachu: What do you think about the keys in Tetrox? I'm starting to dislike the use of the clickwheel in games, so I am thinking that maybe the iPods could use left/right to move, and MENU/PLAY to rotate.
22:26:39Mikachui generally agree with that, but i think it works well for tetrox
22:27:05nudeli think it'll be horrible either way :) the buttons on the ipod feel horrible to press in games :(
22:27:32Mikachuit plays pretty well
22:27:32nudelmaybe they get less sticky in time, mine's still new
22:27:41linuxstbI much prefer the 4g's button than the 5g.
22:27:58nudelit needs a joypad accessory :)
22:28:07Mikachulinuxstb: but i think with the default delta=4 maybe it should only turn on every three events or so, or it will be too sensitive
22:28:43linuxstbYes, it's definitely too sensitive at the moment.
22:29:55Mikachui got to level 10 here
22:30:18Mikachumaybe it could be an option but maybe that would be complex if games already use all buttons etc
22:30:52 Quit KN|stiff (Read error: 104 (Connection reset by peer))
22:32:15 Join SereRokR [0] (
22:35:02tucozPaprica, what do you want me to try now?
22:35:18Papricago to the main menu
22:35:36Papricaand play with the date
22:35:59 Quit webmind (Read error: 104 (Connection reset by peer))
22:35:59Papricafor example set it to 16/03/2006
22:36:27tucozLooks ok
22:36:54Papricamm quit the calendar
22:36:58Papricaand back to it
22:37:03tucozLooks really good actually. The same date as when I quit
22:37:22Papricanow try to add a note
22:37:42Papricaand lets see if it stays after quit
22:37:53lostlogicpreglow: yes, masked write would make sense, avoid the problems we have with iRiver where we have to or in values.
22:38:09tucozPaprica, it stays
22:38:22*Paprica dancing
22:38:40*Mikachu does the kirby dance <('')> <("<) (>"<) (>")> <('')> with Paprica
22:39:03tucozPaprica, doesn't enter work with the keyboard?
22:39:15Papricanot yet
22:39:26Papricabut you have the down arrow
22:39:53 Join fiftyfour123 [0] (
22:40:03tucozI like this keyboard better than the old one
22:40:41Papricai just need to work on the navigation in keboard mode
22:40:55Mikachudoes it cover the same characters as the old one?
22:40:56Papricaglad you love it =]
22:41:24Papricano, only english(lattin?) letters
22:41:34Mikachulatin yes
22:41:37Mikachuenglish has non-latin letters
22:41:40tucozNo, but for the limited input when writing stuff on a mp3-player it's good enough for me.
22:42:28PapricaMikachu, so non-latin letters and numbers
22:42:42Mikachui think you misunderstood me there :)
22:42:47amiconnlostlogic: It's probablöy even necessary to do more than what we do now on iriver
22:42:53Mikachudo you have letters like é and å?
22:43:07Mikachuokay, there are english words with é
22:43:09lostlogicamiconn: yes, since input and output aren't always the same for the same pin
22:43:13amiconnSome GPIO bits are actually separate GPI and GPO
22:43:21Mikachunot so many
22:43:24Mikachui can't come up with any :) we should have a shadow variable in RAM, AND or OR that, then output
22:44:18linuxstbMikachu: IMO, those aren't really English words... Just words in common usage in England...
22:44:21PapricaMikachu, maybe you can port it to nano?
22:44:28 Quit SereR0KR (Read error: 110 (Connection timed out))
22:44:35MikachuPaprica: maybe..
22:46:09 Join Matze [0] (
22:51:08 Join webmind [0] (
22:53:14 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
22:53:24 Quit amiconn (Nick collision from services.)
22:53:25 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:54:56 Quit Rondom ("I'm leaving on a jetplane, don't know when I'll be back again...")
22:58:11 Quit [IDC]Dragon ("CGI:IRC (Ping timeout)")
22:58:24 Join nnod [0] (
23:02:45*Paprica is waiting for OK to commit rockcalendar
23:05:03lostlogicPaprica: does it now replace calendar?
23:05:15 Join obo [0] (
23:06:16Papricai think it's better, but i want you to say that
23:07:16tucozAnd it also work on non-rtc targets, which is a selling point in my opinion :)
23:07:41 Quit safetydan ("Leaving")
23:08:35Papricaand it works on H3xx,H100,iPod 3/4 G, Archos Ondio/Recorder, iAudio X5
23:08:53 Quit ghode|afk ()
23:08:58Papricawhile calendar works onlt on Archos Recorder
23:09:07 Quit fiftyfour123 ("Chatzilla 0.9.71 [Firefox]")
23:09:10linuxstbWhy doesn't it use the standard Rockbox menu functions?
23:09:48Papricai want a special menu
23:09:55 Join webmind_ [0] (
23:09:56Papricaand nicer...
23:10:13Papricawhat bad with the menu?
23:10:22linuxstbThat's my question to you.
23:10:42 Join [IDC]Dragon [0] (
23:10:50Papricai'm asking you what bad with rockcalendar menu
23:11:45kclafwhat about 5G status ?
23:11:53linuxstbIMO, it's bad to introduce different menus in individual plugins. Yes, the current menus in Rockbox may not look nice today, but they will improve in the future. They also have the potential (with langv2) to handle voice UI and localisation.
23:12:06 Join Genre9mp3 [0] (
23:12:13preglowcompletely agreed
23:12:15 Quit Genre9mp3 (Client Quit)
23:12:26BHSPitMonkeyPaprica: I believe you've just been owned by reason
23:12:56linuxstbAnd they also work on the remote...
23:13:27amiconnSpecial menus in plugins aren't any good, I have to agree
23:13:33preglowthey look good
23:13:41preglowbut that's all
23:13:43preglowno other benefits
23:13:48amiconnSome plugins still have them - that was before putting the menu functions in the api
23:14:15amiconnstar.rock iirc, and rockboy at least. The latter only for H1x0+
23:15:13 Quit webmind (Read error: 104 (Connection reset by peer))
23:15:30Papricathe current rockbox menu will destroy rockcalendar UI
23:16:38*preglow petitions against 'rock' and 'box' in every single plugin name
23:16:54 Join webmind [0] (
23:17:08*Paul_The_Nerd seconds that opinion.
23:17:20Papricacalendar is taken
23:17:29linuxstb:( Someone suggest a better name than pacbox....
23:17:40BHSPitMonkeyPac Man
23:17:42preglowgood honest name
23:17:51preglowthen we can have berofled
23:18:06 Join [TCK] [0] (
23:18:10linuxstbIf we had svn, I would rename it....
23:18:26BHSPitMonkeyand a chopper name called roflcopter
23:18:37crashdsomeone should port the roflcopter 'anim;
23:18:40preglowBHSPitMonkey: now _there's_ a name with a nice ring to it
23:18:40crashdid pay to see that
23:18:47crashdpay good money
23:18:47Mikachuboxman :)
23:18:54crashdchuckle chuckle chuckle chuckle chuckle
23:18:59BHSPitMonkeyhow good
23:19:16BHSPitMonkeybe a nice WPS
23:19:21Mikachui guess Rockman is already taken
23:19:39BHSPitMonkeyhow about WakaBox
23:19:46BHSPitMonkey(for pacman now)
23:22:10 Join damaki_ [0] (
23:22:17preglownot calling jpeg.c pr0nbox.c is an instance of just throwing potential hours of fun out the window
23:23:00Mikachumozillas library for image related things is called libpr0n internally i think
23:23:05preglowexactly :-)
23:23:16Mikachumaybe it's not a wellguarded secret
23:23:18preglowmade me chuckle when i saw it, that did
23:24:09BHSPitMonkeypreglow: hahaha
23:24:44Mikachupreglow: isn't everything in ipl called i* too?
23:25:27*BHSPitMonkey looks... "Cube" "Matrix" "Vortex" "Hunt the Wumpus" "BlueCube"
23:25:49 Join damaki__ [0] (
23:26:03 Quit webmind_ (Read error: 104 (Connection reset by peer))
23:26:52preglowlostlogic: i don't understand the addressing scheme, though, the first write is to what looks like GPIOB_OUTPUT_VAL with an extra 8 in the address, and the next is to what looks like GPIOD_OUTPUT_VAL with an extra 1 in the address
23:27:36 Quit Matze ("Miranda IM! Smaller, Faster, Easier.")
23:27:36MikachuBHSPitMonkey: well, iDoom... and um... yeah
23:27:55Paul_The_NerdAnd uDoom and theyDoom...
23:28:28Mikachuthat picture has an ugly line in it
23:28:30Mikachui think
23:28:46Paul_The_NerdYeah, the plasma plugin doesn't ever do that for me.
23:29:03Mikachumaybe it's the sim's fault
23:29:37linuxstbIf someone gives me a replacement image, I'll commit it.
23:30:28BHSPitMonkeyreplacement for what?
23:30:44crashdplasma image that Mikachu pasted.
23:30:51linuxstbRead back about 8 lines
23:31:00preglowisn't backlight fading enabled on h300 ?
23:31:10Mikachusometimes zsh amazes me, i typed cd build-ui<tab> and it completed to build-sim :)
23:31:25peturpreglow: you can't I think
23:31:39preglowwhy not?
23:31:41herz42preglow: looking at backlight.c? the second one you mentioned is port L07, as in the comment
23:31:52Paul_The_NerdLet me see if I can get a better plasma image then
23:31:58preglowherz42: i hardly think there's a port L
23:32:01peturyou mean toggling on/off?
23:32:08preglowherz42: i just concluded whoever wrote that was on acid
23:32:23BHSPitMonkeyalmost tipped my chair back
23:32:25preglowpetur: well yeah, toggling it fast for fading it, like on h100
23:32:28herz42I have added ports e..l as a patch
23:32:43preglowherz42: then we're starting to talk quite an amount of port pins here
23:32:51herz42and the detection for usb and ext power is also on that port l :)
23:32:56BHSPitMonkeyis there a particular reason sims aren't hosted?
23:32:58peturah maybe - I was thinking of stepping through the pwm values - which isn't usable
23:33:03preglow96 port pins....
23:33:05BHSPitMonkeythe autocompile just discards them?
23:33:15Paul_The_NerdApparently taking them on target doesn't work very well either
23:33:15herz42maybe not all are pins...
23:33:22preglowherz42: then what are they?
23:33:45herz42could be used for internal devices?
23:33:54 Quit webmind (Read error: 104 (Connection reset by peer))
23:33:55 Join webmind_ [0] (
23:33:56preglowherz42: then they shouldn't be called gpio
23:34:01preglowmisleading at best
23:34:07herz42on my ipv teh ports e..h don't change
23:34:10preglowherz42: but yeah, how are they accessed?
23:34:15MikachuBHSPitMonkey: i think they can only build linux sims and people with linux can build themselves
23:34:22preglowherz42: the same (0x100 | 1) trick?
23:34:36herz42some from ports I..L do toggle on hdd accesses
23:34:41preglowMikachu: they can build all sims
23:34:54Mikachufor windows too?
23:34:58preglowMikachu: indeed
23:35:01amiconnThe line is probably caused by the plasma moving during the screenshot
23:35:02Mikachui mean the autobuild system
23:35:04preglowMikachu: at least, they used to
23:35:10preglowMikachu: might be different now with the sdl sim, yes
23:35:46linuxstbpreglow: Yes, I don't think anyone has worked out how to cross-compile the sdl sim for windows yet.
23:35:46Paul_The_Nerdamiconn: Yeah, on-target at least, you can't get a dump without a line (or in some cases two)
23:35:50herz42preglow: Up to now, I'm only reading from those ports, I've only seen that L07 used as write
23:36:22BHSPitMonkeycan't the sim be suspended
23:36:23preglowherz42: but yeah, i doubt those are gpio
23:36:46preglowso no any has any idea why backlight fading is not enabled for h3x0?
23:36:53*amiconn wonders why it's doing that though
23:37:29amiconnThat means write() yields
23:37:40MikachuBHSPitMonkey: ctrl-z
23:37:42amiconnHmm, of course it does...
23:38:04amiconnpreglow: Backlight isn't connected to a port pin on H300
23:38:09 Quit damaki (Read error: 110 (Connection timed out))
23:38:31linuxstbMikachu: Thanks, I'll commit it.
23:39:09 Join ulli [0] (
23:40:09 Quit ulli (Client Quit)
23:40:20 Join ulli-art [0] (
23:40:30markunsmall ARM question: BIQ is used to set bits to 0, right?
23:41:00 Quit damaki_ (Read error: 110 (Connection timed out))
23:41:07preglowmarkun: yes
23:41:11preglowbit clear
23:41:34preglowmarkun: basically just NOTs the argument and ands with the result
23:41:38 Part tucoz ("Leaving")
23:41:40linuxstbRd := Rn AND NOT Operand2 according to my quick reference card
23:41:46preglowamiconn: right, it just looks like it is from the code
23:42:23 Quit [TCK] (Read error: 110 (Connection timed out))
23:42:39Mikachulinuxstb: i think i made a color patch for starfield too, but it doesn't make any sense since stars are white :)
23:42:55linuxstbNo, I like starfield as it is...
23:43:31Mikachuheh, i apparently did star->color=LCD_RGBPACK(rb->rand()%255,rb->rand()%255,rb->rand()%255);
23:43:43Mikachu(this was before all the other stuff)
23:44:08preglowtouch starfield, and you will pay the price for it
23:44:12preglowit is perfect in every way
23:44:30Mikachui just remembered it now and thought it was a bit funny
23:44:43Mikachuit was easier than plasma and fire
23:45:34linuxstbIt should be renamed "starbox" though...
23:45:47Mikachuor rockfield ;)
23:46:04Mikachui have an origami pattern called starbox
23:46:11Mikachui used to be able to fold it in under one minute
23:48:12Mikachuhm, google just found me a fancier version than the one i know
23:48:21Paul_The_Nerdpreglow: I am personally of the opinion that it should have hyperspace like lines at very high speeds... and bitmapped planets randomly inserted... and... and...
23:48:30Mikachuthis is it,
23:48:45preglowbacklight fading is hardwired to h100 everywhere
23:48:56MikachuPaul_The_Nerd: and you should be able to crash into a star
23:49:04peturif the disassembly says bne.w what coldfire instruction would that be in the manual? Or what does bne.w do after btst? branch if bit set or cleared?
23:50:11herz42branch not equal? so branch if the test matched?
23:50:29 Quit ulli-art ("CGI:IRC (EOF)")
23:50:34 Join ulli-art [0] (
23:50:35preglowpetur: bxx
23:50:47preglowpetur: branch on condition
23:51:04preglowpetur: it branches if the compare is not equal
23:51:04peturbcc then
23:51:27 Part ulli-art
23:51:49peturbut it's not comparing, it must be checking some flag set by btst
23:52:05preglowhow do i reset settings on ipod...
23:52:36Mikachuhold button iirc
23:52:45Paul_The_NerdYeah, boot, then turn on hold before rockbox loads
23:52:54herz42petur: usually the test instructions also set the zero flag depending on the result
23:54:16preglowahhj, right
23:54:39*preglow holds breats
23:55:18preglowipod hangs when backlight should fade :/
23:55:53t0maspreglow: weird typo.... "*preglow holds breats"
23:56:22t0masjust missed one s to make it really funny ;)
23:57:16preglowbut this qualifies as full blown annoying
23:58:28preglowi had all my money put on this just magically working

Previous day | Next day