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).

#rockbox log for 2006-09-04

00:00:11 Quit Lear ("Chatzilla 0.9.74 [Firefox]")
00:00:28dionoeahum ... you can choose USB charge Amperage. Maybe that's why the ipod is slower when charging in rockbox compared to original firmware
00:00:36dionoea(maybe this is completely unrelated)
00:00:39 Join MarcoPolo [0] (
00:01:43Angryman_h10It has to do something with the disabling of the cpu frequency settings... Used my own rockbox.h10 and it works again
00:03:51barrywardellso maybe when we get cpu frequency working properly it will just work again
00:04:27Angryman_h10i think so... so no bug report ^^
00:04:32barrywardellfile a bug report anyway so that i can come back to it later
00:04:44barrywardellmention that it may be related to cpu freq
00:06:26 Part TrueJournals
00:06:32 Quit _FireFly_ ("Leaving")
00:06:55Angryman_h10In the jpeg.c it also looks which cpu is defined so it's nearby that this is the source of the issue
00:07:08Angryman_h10#ifdef HAVE_ADJUSTABLE_CPU_FREQ
00:13:49 Quit markun ("leaving")
00:13:58 Join markun [0] (
00:14:13Angryman_h10The H10 is not yet in the flyspray
00:14:43 Join pyromancer [0] (
00:15:09 Join jhMikeS [0] (
00:15:14Bagdernow it is
00:16:07Angryman_h10thx ^^
00:18:21 Quit Rondom ("Ex-Chat")
00:23:20stripwaxdionoea - hm, maybe Apple Disk Mode is also slower with 100ma draw than with 500ma (somehow?). wild guess, based on no evidence
00:23:20 Quit pyromancer ("Ex-Chat")
00:23:59 Join Pyromancer [0] (
00:24:35 Quit Febs (Read error: 110 (Connection timed out))
00:25:17 Join XavierGr [0] (
00:27:58 Quit tvelocity ("Αποχώρησε")
00:28:19 Quit Pyromancer (Client Quit)
00:29:18 Quit daurn|laptop (Read error: 110 (Connection timed out))
00:29:37 Join Pyromancer [0] (
00:33:14 Quit Pyromancer (Remote closed the connection)
00:34:16 Join Pyromancer [0] (
00:34:34 Join tvelocity [0] (n=tony@
00:36:18 Quit Pyromancer (Remote closed the connection)
00:38:45 Join Pyromancer [0] (
00:38:55 Quit Pyromancer (Remote closed the connection)
00:41:08 Join Pyromancer [0] (
00:53:25 Quit bluebrother ("Leaving")
00:55:25 Quit Angryman_h10 ("CGI:IRC (EOF)")
00:57:48 Quit rr3tzbach (Client Quit)
01:05:58***Saving seen data "./dancer.seen"
01:16:17 Join JdGordon [0] (
01:18:20 Quit ender` (" Programming is like sex becuse: 22. Other people just get in the way.")
01:19:13 Join nave7693 [0] (
01:22:45 Quit TeaSea (Remote closed the connection)
01:24:45 Quit Pyromancer ("Ex-Chat")
01:26:04 Join Pyromancer [0] (
01:31:49 Join uski [0] (
01:37:13BigMacwere has PaulTheNerd been recently
01:37:19BigMacand why is irc so dead
01:37:58 Join aliasone [0] (n=aliasone@
01:38:43 Quit Kohlriba ("Quit")
01:39:13markunBigMac: because it's quite late on this side of the world
01:40:10BigMacwhat time is it
01:40:20BigMacand i believe paul is from the us
01:40:34BigMacand i have been here all day and it has been rather dead
01:40:36markun1:40 here (netherlands)
01:40:50BigMacstill not too late
01:41:05markunlate enough for me. Good night.
01:41:12 Quit aliasone_ (Read error: 104 (Connection reset by peer))
01:41:25 Join aliasone_ [0] (
01:42:40 Quit tvelocity ("Αποχώρησε")
01:44:46 Join webguest38 [0] (i=5438880b@gateway/web/cgi-irc/
01:45:04stripwaxBigMac - did you get your Ratings stuff working in the end?
01:45:57Soapfrom what was said in the referenced thread the %rp and %rr WPS tags are not currently "attached" to anything.
01:47:09 Quit webguest38 (Client Quit)
01:48:01SoapIt wouldn't be hard to test the %rp one, maybe I'll get off my butt and do so.
01:49:38 Quit Pyromancer ("Ex-Chat")
01:50:25 Join Pyromancer [0] (
01:55:27 Join TrueJournals [0] (
01:56:03 Part stripwax
01:58:53 Quit Pyromancer ("Ex-Chat")
01:59:26 Quit aliasone (Read error: 113 (No route to host))
02:00:50JdGordonmp3z dont compress to zip very welll do they?
02:01:15 Part TrueJournals
02:01:57Soapwouldn't say much about the format if they did.
02:02:15Jasonmp3 is something like a zip
02:02:22belZenot really
02:02:31Jasonlike its compressed allready
02:02:47belZeyes, it uses a huffman-compression
02:03:09Jasonso its something like a zip :P
02:03:09belZeso the data is already "zip-compressed"
02:03:32 Join Pyromancer [0] (
02:03:37 Join Angry [0] (
02:03:46belZeit uses the same compressio
02:04:34 Quit Angry (Client Quit)
02:05:50 Join Angry [0] (
02:06:23 Join Febs [0] (
02:07:16 Quit MarcoPolo (Remote closed the connection)
02:10:44 Join webguest81 [0] (i=5438880b@gateway/web/cgi-irc/
02:11:14 Quit Angry ("Miranda IM! Smaller, Faster, Easier.")
02:15:21 Quit Jason (" HydraIRC -> <- Try something fresh")
02:15:21 Quit webguest81 (Client Quit)
02:15:21 Join nudelyn [0] (
02:16:41 Part nave7693
02:29:39 Part pixelma
02:42:59 Quit barrywardell ()
02:48:50 Quit spiorf (Remote closed the connection)
03:05:59***Saving seen data "./dancer.seen"
03:24:11 Quit Arathis ("[rl_bot quit]")
03:25:11 Join _jhMikeS_ [0] (
03:26:36 Join scott666_ [0] (
03:41:12 Quit jhMikeS (Read error: 110 (Connection timed out))
03:43:44 Quit _jhMikeS_ ("User pushed the X - Because it's X-tra")
03:49:41BigMacstripwax doesn't have a quit line
03:49:45BigMacbut is not here
03:50:03BigMacand no i never got rating working
03:50:04 Part scott666_
04:00:19Soap[2006-09-03 19:54:11] * Parts: stripwax
04:06:28 Join BigMac2 [0] (
04:06:42BigMac2can someone kick my ghost?
04:09:25Soapyou can
04:10:26BigMac2i forgot the ghost command
04:10:49BigMac2"/msg nickserv ghost #rockbox bigmac?
04:18:02 Quit BigMac (Read error: 110 (Connection timed out))
04:21:39 Nick BigMac2 is now known as BigmAC (
04:21:51 Nick BigmAC is now known as BigMac (
04:26:42billytwowillywell kinda.
04:26:52billytwowillyBigMac: if you forget again I think you can type /msg nickserv help
04:27:17billytwowillyI think the command is /msg nickserv ghost ghostnic password
04:27:23billytwowillyor something like that..
04:28:20BigMacthanks man
04:28:50BigMacI wish i was decent at this wps stuff
04:29:04BigMacso i wouldn't have to wait for people on the forum
04:31:43billytwowillyyou'll figure it out.
04:32:13billytwowillyit's cool that you're trying. I can't do anything with the wps stuff, so you're ahead of me;)
04:32:59BigMacbut its not that hard
04:33:18BigMacits just stupid things that keep snagging me
04:34:43billytwowillywhen they get it working on the sansa e200 or the meizu miniplayer I'll try my hand at it;)
04:35:10BigMaci was originally going to get a sans
04:35:20BigMacbut no rockbox support turned me away
04:35:54billytwowillyyah. I think dan_a and possibly bagder are working on it.
04:36:02billytwowillyhopefully soon.
04:36:11billytwowillyI haven't bought one yet. I have an ipod mini right now.
04:36:28billytwowillyI tried rockbox on it and liked it.
04:36:48billytwowillyI used to use rockbox on my archos jukebox 6000 back in the day;)
04:38:19billytwowillyI would probably stick with the original firmware on either of the previously mentioned players if they had bookmarking.
04:38:23BigMacanyone know a way to record an mp3/wav from whats playing on your comp
04:38:26BigMaclike a vid
04:38:32BigMacjust record the audio
04:39:51billytwowillywhat do you mean?
04:40:00billytwowillylike you have an avi file and you want just the audio?
04:40:20 Quit rotator ("zzzzzzzzz")
04:40:32BigMaclike i have a video playing and i just want the audio of it
04:40:36billytwowillyare you in windows or linux?
04:41:25billytwowillyI'd check out for some apps to use to do that.
04:41:38BigMacwhy are you on linux?
04:41:52billytwowillybecause linux is awesome.
04:41:57billytwowillyno viruses.
04:42:00BigMacno not why
04:42:11BigMaci was asking if you are
04:42:16billytwowillyyou forgot a comma.
04:42:26billytwowillyyah, I figured it out.
04:42:36BigMacis it easy? on linux
04:42:37billytwowillyyou meant to say why, are you in linux?
04:43:03billytwowillyI dunno.
04:43:08billytwowillyI haven't had to do that in a long time.
04:43:13billytwowillythe last time I did it was in windows.
04:43:13billytwowillyIt might be.
04:43:22BigMacah nvm
04:43:50billytwowillyand I think I followed a howto on
04:44:29BigMacnot a site anymore
04:45:49billytwowillyit's a site.
04:45:55billytwowillyit's loading for me right now...
04:47:21billytwowillyI'm lagging out like crazy.
04:48:09BigMaccan't find it:-(
04:55:35billytwowillythis was like 5 years ago
04:55:35 Quit solarflare (Read error: 104 (Connection reset by peer))
04:55:47 Join solarflare [0] (
04:56:02 Quit uski (Read error: 104 (Connection reset by peer))
04:56:28 Join uski [0] (
04:58:25BigMacah i see
04:58:41BigMacwell i guess ill learn html again
05:05:01 Join EV|Toshiba [0] (
05:05:32EV|ToshibaOk, im using Xubuntu, and I have the ipod plugged in
05:05:42EV|ToshibaI downloaded the source code (ipod_fw.c)
05:06:01***Saving seen data "./dancer.seen"
05:06:06EV|ToshibaNow it says to type gcc -o ipod_fw ipod_fw.c
05:06:16EV|ToshibaDo I type that in terminal or what?
05:08:22Galoisin the terminal
05:09:35EV|Toshibaok...that isnt working
05:09:39EV|Toshiba is it o or zero?
05:11:13Galoisletter o
05:11:16EV|ToshibaGalois, if I do a install under windows, will I be able to load files on linux?
05:13:05Galoisrockbox, unlike apple, is not evil. You can upload files using standard file copy.
05:13:47Galoisit would help if you describe why it's not working instead of just stating that it isn't working
05:21:01EV|ToshibaI have a 30gb ipod using ubuntu linux to install it (I think i'll use my sindows machine to install rockbox then upload my media files from my linux machine)
05:25:35billytwowillyit's ridiculously easy to do in linux.. just copy/paste all the commands... you might have to change the /dev/sdXXXXXXX only.
05:26:00billytwowillyatleast, the tutorial was really good.
05:26:26billytwowillythe only thing semi confusing was when it linked to the windows howto to get the bootloader
05:26:33ShadowdogMUI need a good idea for a website, I am desperate and asking everywhere :P
05:26:45EV|ServerShadowdogMU, like what?
05:26:56ShadowdogMUAnything at this point, except porn
05:27:03EV|ServerYOu just want a website?
05:27:08EV|Serverjust to do it?
05:27:15ShadowdogMUI have one currently
05:27:23ShadowdogMUbut when I switched WebHosts I got a free domain
05:27:28billytwowilly1) take a picture of a midget doing complex math on a big blackboard. 2) Register 3) profit
05:27:28ShadowdogMUand need to use it up
05:27:37EV|Serverbillytwowilly, xubuntu isnt letting me copy ANY commands..
05:27:52EV|Serverwell it lets me copy them, but terminal wont let me paste them
05:28:10billytwowillyEV|Server: what the gui for x in xubuntu? And why are you using that? Is your computer old or something?
05:28:27EV|ServerNot really.
05:28:31EV|ServerIts the only disk I had around.
05:28:45EV|ServerAnd I kinda like playing with ubuntu. I'll be chaning it to something else soon. Probley zenwalk.
05:28:54billytwowillyyou should probably install gnome or kde (my choice)
05:29:09EV|Serverkde prolly
05:29:17billytwowillybecause you don't sound like a really advanced user, which is who xubuntu is targeted at I think..
05:29:31billytwowillyyou don't necessarily have to reinstall anything though eh?
05:29:44billytwowillyyou can just go sudo apt-get install kubuntu-desktop in the command line
05:29:54billytwowillythen it'll install all the kubuntu stuff.
05:30:19EV|ServerIm not really advanced user yet. I've used DSL for about 2 years..thats about it. lol
05:31:05billytwowillyyah... I am almost certain you don't want xubuntu... that's aimed at advanced users who don't want any bloat
05:31:19billytwowillyyou want something really user friendly like kubuntu or ubuntu..
05:31:27billytwowillyboth are nice. I like kde more myself, so I use kubuntu
05:32:05billytwowillyyah, go sudo apt-get install kubuntu-desktop kubuntu-default-settings
05:32:33billytwowillythen you should have a new kde install... you'll just have to change it so that kdm boots as default (i think that should happen when you install it) then tell kdm to load kde
05:32:41billytwowillyvery user friendly.
05:33:16billytwowillyno reinstall required;)
05:34:26billytwowillyI take no responsibility if it doesn't work though..
05:34:36billytwowillyback up your crap first, yada yada yada
05:35:53EV|Serverthere aint much crap on ere. lol
05:35:57EV|Serverits pretty much a fresh install
05:36:03EV|Serverwith xchat, and a few docs on it only
05:36:09billytwowillythen no problem;)
05:36:26billytwowillyif you are using kde you might want to check out konversation.
05:36:35billytwowillyonce it's up and running. kde xchat look alike
05:40:32EV|ServerUsing windows I cant get the ipodpather 0 command to work either...WTF!
05:40:40EV|ServerI downloaded the three files.
05:40:50EV|Serveripodpather, ipod_fw, and that zipped file
05:42:26EV|Serverit says `ipodpatcher 0` is not a reconized command..
05:42:46billytwowillyit worked flawlessly in linux. are you in linux now?
05:42:47EV|ServerThink someone fucked up on their typing when they did the manuel..
05:42:50billytwowillyI'll help you if I can.
05:42:53EV|Serveryea, I'll do it on this machine then.
05:43:09GaloisI wrote the linux manual. Many many people have used it without any problems.
05:43:11billytwowillyone sec. lemme get to the linux howto I used.
05:43:18 Quit |AhIoRoS| ("Abandonando, see you")
05:43:20billytwowillythen we'lll go through it together.
05:43:45EV|ServerGalois, I used the windows manuel. lol
05:44:13billytwowillythat's what I used exactly.
05:44:22billytwowillylets go through it.
05:45:03EV|ToshibaAlright, got the ipod plugged into the linux machine
05:45:09 Nick EV|Toshiba is now known as EV|Sindows (
05:45:16 Nick EV|Server is now known as EV|xubuntu (
05:45:26billytwowillyk, compile ipod fw.c first...
05:45:35billytwowillyjust go through that howto till you get stuck.
05:45:51EV|SindowsI wanna save the source code right? not open it with mousepad?
05:46:10billytwowillyyah, save it.
05:46:59EV|Sindowsbash: gcc : command not found
05:47:19EV|SindowsI must have missed somethin..
05:47:26Galoisubuntu doesn't install gcc by default
05:47:32EV|Sindows>_< shieat
05:47:35billytwowillyyou don't have the devel stuff install
05:47:48EV|Sindowsinstalling that now
05:47:49billytwowillyone sec. I'm trying to dig up the meta package that will install it all for you.
05:47:57EV|Sindowsoh ok
05:48:04EV|Sindowsusing kdm defalt display manager right?
05:48:20EV|Sindows(for that file you had me download for linux, billytwowilly )
05:49:08billytwowillyyou're installing all those packages while we do this?
05:50:00EV|Sindowsshould be about done
05:50:15EV|Sindowsits running downthe lists very fast...
05:50:25billytwowillysudo apt-get install build-essential
05:50:35billytwowillyk, well we're kinda stalled till that's done...
05:50:44billytwowillydid it download all the packages already and is installing?
05:50:48billytwowillyor still donwloadingt?
05:51:06EV|Sindowsits unpacking
05:51:22EV|SindowsThis is a weird error....
05:51:24billytwowillywell, hopefully it'll be fast.
05:51:37EV|xubuntuE: Could not get lock /var/lib/dpkg/lock - open (11 Resource temporarily unavailable)
05:51:37EV|xubuntuE: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it?
05:51:37EV|xubuntujustin@justin-xubuntu:~$ sudo apt-get install build-essential
05:51:37DBUGEnqueued KICK EV|xubuntu
05:51:37EV|xubuntuE: Could not get lock /var/lib/dpkg/lock - open (11 Resource temporarily unavailable)
05:51:37EV|xubuntuE: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it?
05:52:02billytwowillyyou can't do that while apt is installing the other stuff.
05:52:06billytwowillyrun that command when you're done.
05:52:27billytwowillywhen apt is running it locks the package database so you can't install multiple things at once and break your system.
05:52:33EV|SindowsWell, Im gonna go throw a bag of popcorn in the microwave then while that works. lol brb
05:52:56billytwowillyk, I'll be here listening to the fatboy slim discography I downloaded.. I have about 15 hours left;)
05:53:44 Join TheBeast [0] (n=whoami@unaffiliated/thebeast)
05:54:03billytwowillysay my name in here and konversation will blink and let me know you need my help.
05:55:08 Part TheBeast ("Leaving")
06:02:46 Quit uski (Connection timed out)
06:06:28 Join perplexity [0] (
06:13:16 Join uski [0] (
06:14:56 Join Jazz00006 [0] (
06:15:11 Quit Jazz00006 (Client Quit)
06:16:54EV|Sindowsbillytwowilly, finishing up the install for build essnetial
06:18:00billytwowillynow follow the howto and let me know if you run intoa problem
06:18:56EV|SindowsMan my hands hurt today..
06:19:09EV|SindowsThen again.. I was pratcing stuff for the marching band for abuot 4 hours. lol
06:19:53billytwowillythe moral of the story is don't practice stuff for marching band for 4 hours;)
06:20:12EV| I kind halfto. :p
06:20:15billytwowillyor install hands 2.0
06:20:20billytwowillythe new and improved version;)
06:20:34EV|xubuntujustin@justin-xubuntu:~$ gcc -o ipod_fw ipod_fw.c
06:20:34EV|xubuntugcc: ipod_fw.c: No such file or directory
06:20:34EV|xubuntugcc: no input files
06:20:44EV|SindowsI lasted: 1.3 seconds! lol
06:20:47billytwowillycollege or high school level?
06:20:57EV|SindowsHigh school... 3rd bass drum.
06:20:57billytwowillyrun the command in the same dir you saved the file to..
06:21:04EV|Sindowswhich would be home..
06:21:07EV|Sindowserr desktop
06:21:16billytwowillythat file is the source code that that you are compiling...
06:21:20billytwowillyyah.. go gthere.
06:22:05EV|Sindowsso it'll be gcc -o ipod_fw justin/desktop/ipod_fw.c
06:22:20billytwowillyjust go to the dir.
06:22:26billytwowillyit's easier
06:22:37billytwowillycd /home/justin/desktop/
06:23:08billytwowillywherever you put the file you download cd into that dir.
06:23:14billytwowillythen follow the howto from there.
06:24:10EV|Sindowsstill getting errors of no such file or directory.
06:24:19EV|SindowsIts on my desktop... im looking at the frickin icon. :P
06:24:22billytwowillydude.. if you can't find the file just download it again. put it in /home/yourusername
06:24:52billytwowillyopen up a terminal
06:24:58billytwowillytype cd desktop
06:25:38EV|Sindowsok, its in my user name
06:26:03billytwowillythen type gcc -o ipod_fw then type ip then tab. it should fill in ipod_fw.c if it doesn't it isn't there.
06:26:36EV|Sindowsok, now it made an ew command line....errrr
06:27:08CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
06:27:08*billytwowilly doesn't understand what you just said
06:27:20EV|xubuntujustin@justin-xubuntu:~$ gcc -o ipod_fw ipod_fw.c
06:27:21EV|xubuntugcc: ipod_fw.c: No such file or directory
06:27:21EV|xubuntugcc: no input files
06:27:26EV|Sindowserr hold on
06:27:32EV|xubuntujustin@justin-xubuntu:~$ gcc -o ipod_fw ipod_fw.c
06:27:45Galoisthat means it worked
06:28:03billytwowillyso now you have an ipod_fw there
06:28:09billytwowillyso keep following the faq...
06:29:05*billytwowilly is lagging at 3 seconds.. hurray!
06:29:18 Quit webguest67 ("CGI:IRC 0.5.7 (2005/06/19)")
06:29:33EV|xubuntuNow i gotta figure out what the hell the device is assinet to...yey.
06:29:47billytwowillydude.. it's hella easy...
06:29:53billytwowillyare you in kde?
06:30:01billytwowillydoes the ipod show up on the desktop?
06:30:06EV|xubuntu( i think )
06:30:08billytwowillyif so, right click and hit mount
06:30:15 Join secleinteer [0] (
06:30:19EV|xubuntuIts already mounted
06:30:26billytwowillythen right click and you should see where it's mounted.
06:30:28billytwowillyI think
06:30:32*billytwowilly goes to get his ipod to test it out.
06:31:19billytwowillyright click properties
06:31:39 Quit uski (Read error: 110 (Connection timed out))
06:31:49billytwowillymeta info
06:31:53EV|xubuntudont have that..
06:31:57billytwowillydevice node: mine says /dev/sdf2
06:32:29billytwowillyk, open up another command line and go ls /dev/sd then hit tab until it shows you all the entries for /dev/sd
06:32:31EV|xubuntuSetting up kaffeine (0.7.1-1.3ubuntu10) ...
06:32:32EV|xubuntuSetting up kubuntu-desktop (0.86) ...
06:32:35EV|xubuntuThat did something earlier..
06:32:45billytwowillydon't worry about it.
06:33:06billytwowillyyou're running breezy aren't you?
06:33:18EV|xubuntuI dont know wtf im running
06:33:25EV|xubuntusda, sda 1, sda 2
06:33:35Galoisdude, just type mount in the terminal
06:33:36billytwowillyremind me later and we'll find out what you're running
06:33:43Galoisit'll say something like /dev/sdb2 3904176 2216364 1687812 57% /media/IPOD
06:33:55EV|xubuntuI got a 40gb main drive, 80gb slave and the ipod
06:33:55billytwowillyyah, you're text book
06:34:01billytwowillyfollow the howto exactly
06:34:02EV|xubuntuNow to figure out what one the fucker is...
06:34:07billytwowillybut those are IDE
06:34:23EV|xubuntuWell..the ipod thought it was an IDE one last night...
06:34:30billytwowillythose will show up in /dev/hd probably
06:34:37billytwowillydude. right now it's textbook.
06:34:41billytwowillyjust follow the faq exactly.
06:34:42Galoisif it shows up in the desktop, it will show up when you type mount
06:35:17billytwowillytype fdisk -l /dev/sda and I can almost gaurentee you'll see what is in the howto I pointed you to (and I think galois wrote)
06:35:28billytwowillythen just follow the howto and you're set
06:36:25EV|xubuntujustin@justin-xubuntu:~$ fdisk -1 /dev/sda
06:36:25EV|xubuntufdisk: invalid option −− 1
06:36:25EV|xubuntuUsage: fdisk [-b SSZ] [-u] DISK Change partition table
06:36:25EV|xubuntu fdisk -l [-b SSZ] [-u] DISK List partition table(s)
06:36:25EV|xubuntu fdisk -s PARTITION Give partition size(s) in bloc ks
06:36:26***Alert Mode level 1
06:36:26EV|xubuntu fdisk -v Give fdisk version
06:36:28EV|xubuntuHere DISK is something like /dev/hdb or /dev/sda
06:36:30EV|xubuntuand PARTITION is something like /dev/hda7
06:36:32EV|xubuntu-u: give Start and End in sector (instead of cylinder) units
06:36:34EV|xubuntu-b 2048: (for certain MO di
06:36:38EV|xubuntuUhh huh..sure.
06:36:42Galoisyou screwed up
06:36:47EV|xubuntuYes i am..
06:36:52Galoisfdisk -l as in the letter l, not the number 1
06:36:58billytwowillythat is a -l not -1
06:37:21billytwowillyjust highlight the commands in the howto, go to the terminal and hit the middle mouse button. should paste in
06:37:30EV|xubuntujustin@justin-xubuntu:~$ fdisk -l /dev/sda
06:37:30EV|xubuntuDisk /dev/sda: 30.0 GB, 30005820928 bytes
06:37:30EV|xubuntu255 heads, 63 sectors/track, 3647 cylinders
06:37:30***Alert Mode level 2
06:37:30EV|xubuntuUnits = cylinders of 16065 * 512 = 8225280 bytes
06:37:30***Alert Mode level 3
06:37:30EV|xubuntu Device Boot Start End Blocks Id System
06:37:31***Alert Mode level 4
06:37:31EV|xubuntu/dev/sda1 1 5 40131 0 Empty
06:37:33EV|xubuntu/dev/sda2 6 3648 29262397+ b W95 FAT32
06:37:39EV|xubuntuAnd we have a winner.
06:38:15billytwowillyyah. you are text book. just keep following the howto and only message stuff in here if you need help.. prefereably not big stuff.. use pastebin if you get anything much bigger than above...
06:39:22billytwowillyit should work no problem. you're textbook...
06:39:41billytwowillyheh. I think that should be my new catch phrase;) "you're textbook!"
06:39:55EV|xubuntu * ./ipod_fw -o apple_sw_5g_rcsc.bin -e 1 bootpartition.bin
06:40:53 Join uski [0] (
06:41:23EV|Sindowsit says that command bootpartition.bin is not valid or some shit like that
06:41:33billytwowillyyah... you screwed up something using dd if that didn't work..
06:41:49EV|Sindowscopying and pasting. :)
06:41:56billytwowillyhighlight the dd commands in the faq and hit th emiddle mouse button in the terminal window...
06:42:03billytwowillypost the exact error.
06:42:14EV|xubuntujustin@justin-xubuntu:~$ /ipod_fw -o apple_sw_5g_rcsc.bin -e 1 bootpartition.bin
06:42:14EV|xubuntubash: /ipod_fw: No such file or directory
06:42:29billytwowillyyah.. you messed something up.
06:42:34billytwowillyuse ./ipod_fw
06:42:36EV|Sindowsimagine that
06:42:53billytwowillyor you didn't compile something right.
06:43:22billytwowillydo you see a file called ipod_fw in that dir?
06:43:48Galoisthe howto clearly says ./ipod_fw instead of /ipod_fw
06:44:12EV|SindowsI did dd if=/dev/sda1 of=bootpartition.bin
06:44:18Galoisthat dot in the beginning is not optional. You should not expect things to work when you leave out characters.
06:44:19EV|Sindowsdd if=/dev/sda1 of=bootpartition.bin
06:44:32billytwowillyyou didn't screw that up.
06:44:33EV|Sindows./ipod_fw -o apple_os.bin -e 0 bootpartition.bin
06:44:46billytwowillyyou either didn't compile ipod_fw or you didn't use the ./
06:44:53billytwowillythats period forward slash...
06:45:30EV|SindowsI got it
06:45:32DBUGEnqueued KICK EV|Sindows
06:45:36arpthe dot is optional for people who know how unix works.
06:45:48EV|Sindowsit finally decided to work. :)
06:45:56billytwowillyteach us oh awesome arp!
06:46:04CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
06:46:04*billytwowilly bows down before arp and begs for knowledge
06:46:11arpyes. by explaining how dots work, this makes me awesome
06:46:19arpin fact, I've put that on my resume
06:46:26arp"I know how dots work. this makes me awesome"
06:46:34*billytwowilly cries tears of joy because arp is here.
06:47:09billytwowillyseriously though. how do I make the . optional? If you say put it in /usr/bin or some such thing I'm going to smack you with a large trout....
06:47:23Galoisput ipod_fw in /
06:47:27arp... or you could type the full path or relative path to the binary
06:47:32***Alert Mode OFF
06:47:41arp./ being relative to your current directory
06:47:55arpor... you could put "." in your path
06:48:00billytwowillylame. I knew that already.
06:48:08*billytwowilly gets off his knees and smacks arp with a large trout
06:48:10arpnot lame, awesome!
06:48:38arpyou said it yourself!
06:49:27billytwowillybut I already knew that. I thought you were going to say something new and cool
06:50:35billytwowillyI wonder what's going on with my isp.. uploading one torrent and having 6 seconds lag..
06:50:36arpnope, I was only remarking about whether "a dot" was "optional" or "mandaatory".
06:51:43arpnothing new in open() or execvp() in a long time.
06:51:56billytwowillyEV|xubuntu: how's it going?
06:52:17 Join fatherfork [0] (
06:52:21EV|xubuntuuhhh...I fucked something up again
06:52:28EV|Sindowsim on setup 3
06:52:55 Join RoC_MM [0] (
06:53:00billytwowillyyou have a video right? you copied the second thing right? apple_sw_5g_rcsc.bin?
06:53:11EV|Sindowsno I dont have a video..just a color
06:53:29billytwowillyk, you're set then.
06:53:54EV|Sindowsim just now chaning the nano to color in this:
06:54:02EV|Sindows./ipod_fw -g nano -o rockboot.bin -i apple_os.bin bootloader-nano.bin
06:54:15EV|Sindowschange those 2 nanos to color and thats it right?
06:54:34EV|xubuntujustin@justin-xubuntu:~$ ./ipod_fw -g color -o rockboot.bin -i apple_os.bin bootloader-color.bin
06:54:34EV|xubuntuGenerating firmware image compatible with iPod mini, 4g and iPod photo/color...
06:54:34EV|xubuntuCannot open loader image file bootloader-color.bin
06:54:42EV|xubuntuDamn it. This is starting to piss me off..
06:55:00billytwowillydude. you didn't download a precompile bootloader
06:55:12EV|SindowsWait, I dont need to do the boatloader?
06:55:20Galoisbootloader-color.bin is something you need to download. The page tells you where to download it.
06:55:33EV|Sindowsi did download that
06:55:36billytwowillyyah.. you do.. that's what you've een working towards.... getting a bootloader that can load rockbox...
06:56:00 Join midkay_ [0] (
06:56:08arpls -l bootloader-color.bin
06:56:14arpdo you see that in your dir?
06:56:19arpyou might need to add a dot!
06:56:49billytwowillyman, macy grey + fatboy slim = awesome
06:57:27EV|SindowsI remeber why its doing that. I had to move the file into my home dir.
06:57:40arpgood to know!
06:58:05EV|Sindowsthere we go
06:58:39billytwowillyyou're new to linux right?
06:58:45EV|Sindowspretty much, yea.
06:59:04billytwowillythis is usually easier.
06:59:14billytwowillyyou'll look back in three months and laugh at how much trouble you had;)
06:59:20EV|Sindowslol, yup
06:59:26billytwowillyanyhoo, enjoy your rockbox
06:59:28EV|Sindowsalright, setup 4
07:00:11billytwowillyso now you have a bootloader that can load the rockbox firmware or the apple firmware. so now you put a rockbox build on there and you're golden.
07:01:22arpI was very impressed with how fucking easy it was to install rockbox.
07:01:34billytwowillyme too.
07:01:37EV|Sindows:P Just kidding
07:01:45billytwowillythe only thing that confused me is the downloading the boot loader.
07:01:50billytwowillybecause it linked to the windows howto.
07:01:54billytwowillyand I was like "WTF?"
07:01:56EV| setup 4 is all I duno know..
07:02:00billytwowillythen I figured it out...
07:02:02arpanyway, I'm just annoyed that it doesnt have great support for the 3G ipod
07:02:25billytwowillyclick "daily builds" then download the one pertaining to you.
07:02:42billytwowillythen punzip it to the root dir of your ipod..
07:02:55arpso I dont actually run rockbox yet.
07:03:20EV|SindowsIt says i dont have to unplug it..
07:03:22billytwowillyyou haven't installed rockbox yet...
07:03:27billytwowillyyou did the hard part.
07:03:29EV|Sindowsbut then the first thing in setup 4 is to plug it back in
07:03:49arpI believe you have to extract rockbox onto the device now
07:03:49billytwowillywhen your ipod boots up there is a program that loads the bigger program that gives you the mp3 features.
07:04:08arpall you did was put the boot loader program which can load rockbox
07:04:13billytwowillyyou changed the bootloader from the ipod one which only loads the apple firmware to the one that can load rockbox or ipod
07:04:55billytwowillyso download the daily build that pertains to your player then type unzip /media/sda2
07:04:59billytwowillythat should probably work
07:05:14billytwowillyor just use ark (the unzip/unrar etc gui)
07:05:17EV|Sindowsdidnt work
07:05:24 Quit fatherfork ()
07:05:27EV|Sindowsusing xarchiver
07:05:28billytwowillyand point it to your ipod directory
07:05:39billytwowillyk, fine.
07:05:53Galoistry unzip -d /media/sda2
07:05:56EV|Sindows /media/sda2
07:06:03billytwowillythe important bit is to extract it to the root dir of your ipod, which should be /media/sda2
07:06:06***Saving seen data "./dancer.seen"
07:06:08billytwowillyyep. almost certainly.
07:06:18billytwowillyyou can ls /media/sda2 and see what it is
07:06:34EV|SindowsI got the zipped file save to the desktop
07:06:37billytwowillyif you see Notes, iPod_Control, Contacts and Calendar that's where it goes
07:06:58billytwowillythat's fine.
07:07:10billytwowillyextract the zip file to the root dir of the ipod....
07:07:19EV|Sindowsi cant get it to find the fucker
07:07:39billytwowillydefine "fucker"
07:07:50EV|Sindowsipod = fucker
07:07:56Galoisyou said it showed up on the desktop
07:08:01EV|Sindowsit does
07:08:15EV|Sindowsbut I have a zipped file open in xarchiver
07:08:19billytwowillymake sure it is mounted...
07:08:28Galoisif worst comes to worst open the ipod in the desktop
07:08:35Galoisdrag the files from xarchiver into the ipod
07:08:49billytwowilly heh.
07:08:55billytwowillyI forgot about that...
07:09:21*billytwowilly doesn't use guis for that stuff. Too slow
07:10:24 Quit uski (Read error: 113 (No route to host))
07:11:13billytwowillyso now reboot the ipod and let us know if it worked.
07:11:57 Quit midkay (Read error: 110 (Connection timed out))
07:15:12EV|xubuntuOk, lets reboot ipod..
07:16:39EV|xubuntuah no bastard!!!
07:17:17billytwowillyapparently grammar school sucks in america....
07:17:22billytwowillywhat happened?
07:17:34EV|xubuntuIpod software is still there!!
07:17:45billytwowillythat's fine...
07:17:50billytwowillyyou want it still there.
07:17:53billytwowillytry rebooting it again.
07:17:56billytwowillyhold the middle and top buttons down.
07:18:13billytwowillyif you see some text instead of an apple you have the bootloader installed
07:18:17EV|xubuntui rebooted it
07:18:24Galoisturning it off and on is not the same thing as rebooting it
07:18:36EV|xubuntupressing menu and select
07:18:39billytwowillyif, after the second reboot, it doesn't show up rockbox then you farked up installing the rockbox firmware.
07:18:41EV|xubuntuthen when it reboots let go?
07:19:00billytwowillyyah, do that alteast twice
07:19:15EV|xubuntuapple logo comes up no matter what I do
07:19:18EV|xubuntufucking apple
07:19:38billytwowillyyou don't see any text scrolling?
07:19:41Galoisthe apple logo always comes up. The rockbox text displays right after the logo.
07:19:45EV|xubuntuthat damn apple logo comes up
07:19:47billytwowillyyou screwed up installing the firmware.
07:19:57EV|xubuntuthen it roolls into its hsitty ass software
07:20:07billytwowillyso you didn't install the bootloader.
07:20:15*billytwowilly likes the apple firmware
07:20:44billytwowillygo through the steps of the howto again. you missed something.
07:20:45*EV|xubuntu wants to shove it up apples ass..
07:21:04EV|xubuntualright, lets do it again. Hoorrayy.
07:21:14*billytwowilly thinks EV|xubuntu shouldn't have bought it if he didn't like the firmware
07:21:26EV|xubuntuWell i did for awhile
07:21:35EV|xubuntuuntil it started deleteing songs, locking up, and not working
07:21:42billytwowillynot all again. recopy the boot partition back to the ipod for starters. the new boot partition I mean.
07:22:39GaloisI bought the ipod for the express purpose of running rockbox, but then again I also know what I'm doing in linux, so ...
07:22:57EV|xubuntujustin@justin-xubuntu:~$ ./ipod_fw -g color -o rockboot.bin -i apple_os.bin bootloader-color.bin
07:22:57EV|xubuntuGenerating firmware image compatible with iPod mini, 4g and iPod photo/color...
07:23:00EV|xubuntuok, so that works
07:23:20billytwowillystep 3 specifically
07:23:31billytwowillydd if=rockboot.bin of=/dev/sda1
07:23:36billytwowillythat part particularly
07:23:46EV|xubuntuok, i did that and now term isnt doing anything
07:23:55Galoisit takes a while
07:23:55EV|xubuntunow it did
07:24:00billytwowillyjust let it sit.
07:24:04billytwowillyit takes a bit.
07:24:12EV|xubuntunow what?
07:24:16Galoisis it done?
07:24:22Galoisdisconnect the usb cable and reboot the thing
07:24:23billytwowillyif it's done, right click the ipod in kde and select safely remove
07:24:33billytwowillythen reboot it
07:25:29EV|xubuntuGOt a toshiba hard drive in it.
07:25:32billytwowilly you just missed that last command
07:25:45billytwowillythey al do...
07:25:57billytwowillyit'll boot into the ipod firmware the first time I think.
07:26:00EV|xubuntuThis...cant be good
07:26:03EV|xubuntuoh ok
07:26:16billytwowillyreboot it again.
07:26:27billytwowillyif it still boots into ipod then you farked up step 4 somehow
07:26:36EV|xubunturockboc error: -1
07:26:40EV|xubuntuloading orginal firmware:
07:26:45*billytwowilly bets he knows how too.
07:26:57billytwowillyyou farked up step 4
07:27:05billytwowillythere is a hidden dir called .rockbox in the zip archive.
07:27:07Galoisstep 3 is done for sure
07:27:13billytwowillyyou just copied over the rockbox.ipod file
07:27:27billytwowillyI'd put money on it.
07:27:38EV|xubuntuok, its in there.
07:27:50EV|xubuntui dont know about the hidden one
07:28:03EV|xubuntuah, there the little bastard is
07:28:25billytwowillyif the hidden dir is in the root dir of the ipod reboot it again and rockbox should work.
07:28:55EV|xubuntuwell, its still copying stuff
07:29:20billytwowillylemme know once it's done and you've rebooted it.
07:31:19 Quit daurnimator ("I wonder where that fish could be...")
07:32:53 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
07:32:58billytwowillyOMFG! steve irwin is dead!
07:33:55EV|xubuntuIt works!! Hahaha.
07:33:57crashdaye, the crocodile hunter got nailed by a stingray in Queensland
07:34:15*billytwowilly loved that guy
07:34:16EV|xubuntuooohhhh gotta hurt
07:34:27billytwowillyI gues it's true only the nice die young.
07:34:40Galoisin the future, to put songs on the ipod, just plug it in, open the ipod in the desktop, drag and drop your mp3s or whatever into the ipod folder, then do the "safely remove" thing
07:35:09billytwowillyGalois: have you had your ipod delete mp3s on you?
07:35:13Galoisit works the same way in windows, just drag and drop your songs into the ipod (anywhere _except_ the iPod_control folder)
07:35:22billytwowillyie if they were put directly in the root dir?
07:35:32Galoistechnically, I don't have any mp3s on the ipod
07:35:35EV|xubuntuthere all in there now
07:35:40Galoisbut I've never had it delete files on me.
07:35:55billytwowillymy ipod mini deletes files on me.
07:36:01billytwowillyit's pissing me off.
07:36:08EV|xubuntuok, now to go abuot the graphical enviorment
07:36:48billytwowillyEV|xubuntu: you should look at the manual and the parts about the WPS... maybe download some of them. you can make your ipod look a lot nicer.
07:37:06EV|xubuntuyeah, what im trying to find now
07:37:35EV|xubuntuerrr..where is it?
07:38:11EV|xubuntuit had a list of things for the desktop
07:38:15EV|xubuntuerr, menu.
07:40:22billytwowillyon the website.. I can't remember exactly where.
07:42:59EV|xubuntuDo you know where the skins and stuff are?
07:43:01EV|xubuntuI cant find them.
07:43:14Galoissome of them come with rockbox already
07:43:19scorcheright on the wiki..
07:43:27scorchefront page of it actually
07:43:31Galoishit the Menu button and scroll down to Browse Themes
07:43:36scorchethere is a link to the WPSs
07:44:32EV|xubuntuthere it is
07:44:49Galoisyou can download more themes from but you have to put them in the right place, something like .rockbox/wps
07:46:53 Nick dongs is now known as umwhat (
07:49:15 Nick umwhat is now known as dongs (
07:56:16EV|xubuntuI cant get it to apply its theme..
07:58:32EV|xubuntuI dont care anymore. I'll deal with it tomrrow. im going to bed...
08:07:13 Join theli_ua [0] (
08:07:47 Join ender` [0] (i=null@
08:09:46 Quit XavierGr (Read error: 113 (No route to host))
08:09:52[HO]vo|tanyone need parts for a v2 recorder? i have two batteries (one stock and 2yr old/1yr use, one 2800mah and a 1yr use) and a case/circuitboard (minus headphone/line out jacks)
08:10:17EV|xubuntuheh, i dont even know what that is. lol
08:11:07[HO]vo|twhat what is ?
08:11:11*billytwowilly doesn't think too many people use the original archos models anymore
08:11:33*billytwowilly stopped using his quite a while ago.
08:12:02[HO]vo|tagreed, but i can't see tossing something that could help someone else from buying a used ebay model
08:12:28billytwowillytrue dat.
08:12:48billytwowillywe should have a forum section for posting players for free/ just shipping cost
08:12:59[HO]vo|tyeah that would be good
08:13:10[HO]vo|ti would have used my archos v2 recorder for another few years if my headphone jack self-repair would have went through
08:13:24billytwowillyheh. I think I actually still have my archos jukebox 6000
08:13:31billytwowillydid you buy a new mp3 player yet?
08:13:44billytwowillyWill the parts from a working archos jukebox 6000 help you?
08:14:00[HO]vo|ti've already decided to move away from older models actually
08:14:17billytwowillyk, because if it would have helped you I would have sent you it for just shipping cost
08:14:27 Nick EV|Sindows is now known as EV|Toshiba (
08:14:27DBUGEnqueued KICK EV|Toshiba
08:14:30 Nick EV|xubuntu is now known as EV|Server (
08:14:30DBUGEnqueued KICK EV|Server
08:14:43[HO]vo|ti bought an iaudio x5 60gb a few weeks ago from amazon, but there's a deal i found online that is making me return the one i have for another new one
08:14:54billytwowillysweet! The gospel of the flying spaghetti monster has been released!
08:15:13[HO]vo|tis that the one that there's porn/etc with?
08:15:13 Join uski [0] (
08:15:13billytwowillythat is morally reprehensible
08:15:45billytwowillyIt's the one where the basic tenent is that there is a flying spaghetti monster that created everything and watches over everything.
08:15:54[HO]vo|tyeah that's right.
08:16:01billytwowillyand he uses his noodly appendage to influence scientific findings.
08:16:12billytwowillyand heaven has a beer volcano and a stripper factory
08:17:37billytwowillyWAY better than christianity.
08:17:49billytwowillyand it explains why the scientists keep finding things that contradict its teaching.
08:17:54billytwowillyerr. teachings.
08:17:57billytwowillyit's all about the noodly appendage.
08:18:12billytwowillyand the bookstore by my house is showing one in stock.
08:19:07billytwowillyso in 8 hours I shall be purchasing a copy
08:20:35[HO]vo|tthat sounds rather blasphemous and hilarious
08:23:09billytwowillyit is
08:23:11CtcpIgnored 2 channel CTCP requests in 3 seconds at the last flood
08:23:11*billytwowilly is an atheist
08:23:16*[HO]vo|t too
08:23:21*billytwowilly is a member of the least trusted group of people in america
08:23:31billytwowillypeople will elect gays and muslims before atheists;)
08:23:38[HO]vo|tive read something like that
08:24:24[HO]vo|twasnt there a video about that? like it namedropped all the athiests in popular culture
08:25:12billytwowillyyah, on digg.
08:25:36billytwowillybut I knew that before that video came out. it was in a study done a while ago.
08:25:50[HO]vo|ti see
08:26:10[HO]vo|tbillytwowilly what do you run rockbox on?
08:26:19billytwowillysome times on my ipod
08:26:22billytwowillynothing at the moment.
08:26:35billytwowillyif it comes out for the sansa e280 I will run it on that when it comes out.
08:27:09[HO]vo|twell, i can get an ipod 60gb or an iaudio x5 60gb for $10 less. would you say ipod or no
08:30:00 Join MarcoPolo [0] (
08:30:46billytwowillyI'd go iaudio. ipod is kind of a scam.
08:31:13billytwowillyiuadio has a nice fm tuner built in.
08:31:20billytwowillyand can record and has usb to go built in too.
08:31:43crwland much better documented hardware :)
08:32:28billytwowillyI wouldn't buy a hard drive based player though...
08:32:37*billytwowilly is moving to flash soon.
08:32:43Galoisflash for the win
08:32:44*billytwowilly is crazy hard on his electronics
08:33:09billytwowillyI'm honestly surprised how long my ipod mini has lasted.
08:33:26billytwowillyI've dropped it off my bike numerous times onto concrete and even out of my car onto concrete a couple times.
08:36:56[HO]vo|ti have a 1gb ipod nano coming in a couple months from KeyBank (explanation needed?) so ill have that for physical activity... my mass media will mostly stay in the car
08:37:48[HO]vo|tare you USA or eur or otherwise?
08:39:15[HO]vo|tkeybank in the states has an offer where if you are new to keybank and you open a checking acct and then either do direct deposit or get a credit card with them, you get a free 1gb ipod nano
08:40:01Galoisif I were buying I would just get the largest capacity flash player on the market that runs rockbox
08:40:02crwli think i might upgrade from my h120 when at least, say, 32 GB flash-based players are available and competetively priced :P
08:40:45[HO]vo|tGalois: what is it currently? ipod nano 4gb? (they'll have 8gb out on sept 12 iirc)
08:40:52billytwowillywe have a similar thing here in canada for td canadatrust
08:41:09billytwowillyGalois: which will hopefully soon be the 8 gb sansa;)
08:43:12*EV|Server is away: zZz
08:43:35 Quit EV|Toshiba (Read error: 104 (Connection reset by peer))
08:48:48 Quit [HO]vo|t ("( :: NoNameScript 4.02 :: )")
08:49:30 Join [HO]vo|t [0] (
08:50:47 Quit uski (Read error: 104 (Connection reset by peer))
08:51:33 Quit tuwe (Remote closed the connection)
08:53:42 Join webguest35 [0] (i=c0647cda@gateway/web/cgi-irc/
09:00:10 Join Kohlrabi [0] (
09:00:51 Join webguest12 [0] (i=d4d10305@gateway/web/cgi-irc/
09:00:53 Quit Kohlrabi (Client Quit)
09:01:17 Quit webguest12 (Client Quit)
09:06:10***Saving seen data "./dancer.seen"
09:16:19 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
09:19:20 Nick midkay_ is now known as midkay (
09:23:09amiconnSlasheri: Your scheduling patch doesn't build for archos
09:23:25amiconndebug_menu.c: In function 'dbg_os':
09:23:25amiconndebug_menu.c:129: error: 'struct thread_entry' has no member named 'priority'
09:23:47*Bagder don't want no priority
09:24:37 Join the_fuchs [0] (n=the_fuch@
09:25:04Bagderand I don't see how we can introduce a priority to the scheduler without bending backwards
09:28:21amiconnWell, I'm not fond of introducing priorities myself, but Slasheri's patch seems to work according to him, and I just wanted to test it
09:28:42amiconnPriorities are optional with his patch, and hence the archos scheduler would run without
09:28:42Bagderso how does the priorities work?
09:29:10amiconnThey are dynamic priorities, not static
09:29:11BagderI mean, how can you tell when the low-prio ones should get scheduled?
09:29:18Bagderand how are they raised?
09:29:20amiconnSo no thread gets starved
09:29:43billytwowillyyou guys must be in europe
09:30:04Bagderthere's something outside of Europe?
09:30:06amiconnYou could just check out the patch. It's not on the tracker, but available here:
09:30:22billytwowillybecause I couoldn't talk about what you're tlaking about at 2 am;)
09:30:32amiconn9:30 here
09:30:49amiconnApart from that, 2 am is a good time for coding...
09:31:30billytwowillyI'm vegged out by then. can't think, don't want to.
09:31:44Bagderwell, that's too not easy to tell by reading the patch I'd say...
09:32:26Bagderthere's not a single comment added to explain how it works
09:32:45amiconnIt looks like the priorities regulate the ratio with which the threads are scheduled
09:32:59amiconnHigher priority threads are just scheduled more often
09:33:31amiconnThe only dynamics outside the scheduler is the handling of the playback thread. If it runs low, it raises its own priority
09:34:03*amiconn checked out a 'test' working copy and applied the patch there
09:34:05 Quit RoC_MM ("Leaving")
09:34:13*LinusN mumbles something about worms in a can
09:34:28BagderI don't see what good this will bring
09:34:33Bagdercompared to the problems
09:35:38amiconnWell, according to Slasheri it's a working approach against playback skipping
09:35:58amiconnI just wanted to test in order to have some facts for argumentation
09:36:10BagderI can understand that, but the question is: is it the proper fix?
09:36:36Bagderamiconn: of course, that's a good approach
09:37:01Bagderbut I rather start by examining the concept and theory behind it
09:37:07LinusNmy uninitiated guess is that the UI will lag behind when the codec needs cpu
09:37:20*midkay perks his ears at "approach against playback skipping"...
09:37:35amiconnLinusN: UI gets slower when the codec runs low, yes
09:37:54LinusNso what happens when the codec is really hungry, i.e non-realtime
09:38:26Bagdermunch munch
09:38:39 Join sandoaway [0] (i=lolsteam@
09:40:50amiconn 12:20 thru 12:55 and 13:59 and later
09:41:37amiconnIt seems like the patch I linked to is not the latest
09:41:41*amiconn ings Slasheri
09:41:50 Nick the_fuchs is now known as tuwe (n=the_fuch@
09:57:05 Join Jason [0] (
09:59:08 Quit sando (Read error: 113 (No route to host))
09:59:39markunhi Jason
10:00:38Slasheriamiconn: it's not the newest but should work also
10:00:48Slasherii will create a new patch at evening
10:01:19Slasheriin fact UI should be even more responsive with the new patc
10:01:37LinusNSlasheri: do you implement a max-time to prevent starvation?
10:05:32SlasheriLinusN: yes, all threads get cpu time
10:06:02Slashericurrently that is (current_tick - last_run) > current_priority * 10
10:06:12SlasheriLinusN: if you can, please try the patch :)
10:06:17LinusNi must say that i'm *very* reluctant to introducing priority
10:06:28Slasherifor example running cube with the max boost mode, does not cause even q8 vorbis to skip
10:06:34LinusNi'd like to test the patch, if you can bring me the latest one
10:06:44Slasheriand check the improved debug menu for process stack also
10:06:53Slasheriit separates running and sleeping processes
10:07:08SlasheriLinusN: the one on the web should be working
10:07:50LinusNso what's new in your newer, unreleased patch?
10:07:50billytwowillyI wish companies developed their software like this. It's so cool watching the devs duke it out;)
10:08:36SlasheriLinusN: it has only some bug fixes and compiles better on different targets
10:08:40Slasheribut no new features
10:09:04Slasheribut i need to go now, be back soon and check the logs later :) -|
10:12:46 Quit perplexity ("Check out my rear window")
10:13:52 Join perplexity [0] (
10:19:10 Join webguest00 [0] (i=52678186@gateway/web/cgi-irc/
10:19:52 Join _FireFly_ [0] (
10:24:11 Quit webguest00 ("CGI:IRC (Ping timeout)")
10:27:04 Join Rondom [0] (
10:30:00 Join webguest07 [0] (i=50edeb09@gateway/web/cgi-irc/
10:31:35 Quit webguest07 (Client Quit)
10:32:52 Join tucoz [0] (n=martin@rockbox/developer/tucoz)
10:33:27tucozI wonder if i should file a bug-report to the (as far as i know) lacking A-B support on the ipods
10:35:32tucozas the A-B repeat option is possible to select. However, no-one has answered my thread in the forums how to set the A-B marks.
10:39:05 Join sando [0] (i=lolsteam@
10:39:26 Quit tucoz (
10:39:26 Quit Jason (
10:39:26 Quit sandoaway (
10:39:26 Quit webguest35 (
10:39:26 Quit ender` (
10:39:26 Quit Febs (
10:39:26 Quit SanAlmighty (
10:39:26 Quit actionshrimp (
10:39:26 Quit gromit` (
10:39:26 Quit ScoTTie (
10:39:26 Quit excitatory__ (
10:39:26 Quit gopp (
10:39:26 Quit dpassen (
10:39:26 Quit fah (
10:39:26 Quit Strogg (
10:39:26 Quit twisted (
10:39:26 Quit direkt (
10:39:26 Quit Spida (
10:39:26 Quit Hadaka (
10:39:26 Quit crwl (
10:39:27 Join ScoTTie_ [0] (
10:39:28NJoinactionshrimp [0] (
10:39:42NJoinSanAlmighty [0] (
10:41:44NJoindirekt [0] (n=direkt@
10:41:46LinusNSlasheri: your scheduler doesn't guarantee that the thread with the highest priority is run
10:41:56LinusNis that intentional?
10:42:06NJointwisted [0] (
10:42:55LinusNthe loop in switch_thread() breaks as soon as it finds a thread with higher priority than the running one
10:42:56NJoinSpida [0] (
10:42:58NJoinexcitatory__ [0] (
10:43:05NJoingromit` [0] (
10:44:11LinusNthat approach probably works fine, just wanted to know if that was what you wanted
10:45:12 Quit dan_a (Read error: 110 (Connection timed out))
10:46:01LinusNheh, the variable "highest_priority" confused me at first, since lower values mean higher priority :-)
10:48:02NJointucoz [0] (n=martin@rockbox/developer/tucoz)
10:48:02NJoinJason [0] (
10:48:02NJoinsandoaway [0] (i=lolsteam@
10:48:02 Join webguest35 [0] (i=c0647cda@gateway/web/cgi-irc/
10:48:03NJoinender` [0] (i=null@
10:48:03NJoinFebs [0] (
10:48:03 Join ScoTTie [0] (n=scott@unaffiliated/scottie)
10:48:03NJoindpassen [0] (n=dpassen1@resnet-236-163.resnet.UMBC.EDU)
10:48:03NJoinStrogg [0] (n=user@
10:48:03NJoinfah [0] (
10:48:03NJoinHadaka [0] (
10:48:03NJoincrwl [0] (
10:48:04***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
10:48:21 Join crwll [0] (n=crawlie@
10:48:23tucoztest. can someone read what I write?
10:48:25 Quit tucoz (Connection reset by peer)
10:48:29 Join karl [0] (n=karl@
10:48:31 Join ender1 [0] (i=null@
10:48:37karlhi all !
10:48:41LinusNhey ho
10:48:48LinusNtucoz: yes
10:49:14karldoes anyone here know where to find info about how to replace the headphone plug on a iriver ihp-140 ?
10:49:25 Join tucoz [0] (n=martin@rockbox/developer/tucoz)
10:49:39 Join Naked [0] (
10:49:42LinusNwell, that connnector is very special, i haven't found a source for it
10:49:46karlit seems mine is slowly breaking, its not making good contact to the plug any more
10:50:21 Join Strogg_ [0] (n=user@
10:50:32tucozLinusN, my connection lagged
10:50:41tucozseveral minutes
10:51:07 Join Jason_ [0] (
10:51:08LinusNkarl: i'm sure you can convince someone to send you a connector from a broken h1xx
10:51:34 Quit Rick ("I don't need to be here.")
10:52:13karlbut it might be quite difficult to solder, doesnt it combine mic and headphone ?
10:52:26tucozLinusN, do you know anything about the A-B mode on ipods?
10:52:50LinusNkarl: no, but it combines the headphones and the remote control connector
10:52:56LinusNtucoz: not a thing
10:53:21LinusNkarl: and yes, it might be tricky to solder if you are inexperienced
10:53:51tucozI think that has to be selected through the context menu, but that item is missing afaik
10:53:55karlbugger.... i wonder if I can maybe just bend the contacts in the plug a bit, so it gets better contact again
10:54:02tucoz(due to the lack of buttons on the ipods)
10:54:39 Quit dpassen (Read error: 113 (No route to host))
10:54:44 Quit crwl (Read error: 110 (Connection timed out))
10:54:53 Join Rick [0] (
10:55:10LinusNkarl: it depends on whether it is the contacts inside the connector or if it is the soldering
10:55:11 Quit solarflare (Read error: 54 (Connection reset by peer))
10:55:30 Quit Hadaka (Read error: 111 (Connection refused))
10:55:30 Nick Naked is now known as Hadaka (
10:55:31 Join solarflare [0] (
10:55:33LinusNtucoz: i guess nobody has cared until now
10:55:59tucozprobably not. well, one person did. He came in here telling me that the manual was wrong ;)
10:56:04 Quit Strogg (Read error: 113 (No route to host))
10:56:21tucozand then I started to care
10:56:30LinusNtucoz: i hope you flamed him real good :-)
10:56:39 Quit ender` (Read error: 110 (Connection timed out))
10:57:02 Quit Jason (Read error: 113 (No route to host))
10:57:02 Nick Jason_ is now known as Jason (
10:57:53 Quit Febs (Read error: 113 (No route to host))
10:58:50karlthanks linux, i'll have a look at the circuit-board pics on the rockbox website, and maybe have a look at ebay :)
10:59:04karls /linux/linus/
11:00:06 Quit ScoTTie (Read error: 110 (Connection timed out))
11:00:36 Quit sandoaway (Read error: 113 (No route to host))
11:03:19 Join uski [0] (
11:04:27 Join bluebrother [0] (i=810d4658@gateway/web/cgi-irc/
11:06:14***Saving seen data "./dancer.seen"
11:10:57LinusNkarl: germany?
11:13:13amiconnSlasheri: The patch on the web doesn't even compile for archos, in case you didn't see that in the logs already
11:14:36LinusNkarl: hmmm, looks like the headphone connector might be a standard connector after all
11:14:48 Quit uski (Read error: 104 (Connection reset by peer))
11:15:42amiconnLinusN: Hmm, do you know whether the headphone connector in the recv1 is available somewhere?
11:15:59Slasheriamiconn: ah, yes i did.. i have already fixed that in the new version but i am not able to test it yet
11:16:20Slasheriamiconn: you could just remove the priority variable from the debug menu, then it should work
11:16:40amiconnMy connector is still working, but a bit of plastic broke off the top. The remote contact is a bit loose, probably because of this
11:18:47LinusNamiconn: i dunno, but i guess jrg could have a spare or two
11:19:14LinusNSlasheri: read my questions earlier?
11:19:21SlasheriLinusN: true, that is just a performance thing so we don't need to go through the entire process list to find the highest priority. And running lower priority threads from time to time shouldn't be an issue because they are run anyway
11:20:03LinusNi see
11:20:54LinusNhehe, i played some AAC files that requires 100% boost and are still not realtime
11:21:07LinusNthe ui lagged somewhat
11:21:43LinusNand it still skipped of course
11:22:24LinusNi'd really like a few comments in the code
11:32:25JdGordonSlasheri: i had problems with your schedular... i tinhk it was that anyway... it seems to freeze up on startup if resume on startup is on... music doesnt play, and i can access the menus but nott he file browser, but if i press stop it crashes
11:32:42 Join uski [0] (
11:32:57JdGordonalso, +-2% on the pitch screen pauses the audio for half a sec when u use it.
11:33:19*JdGordon back in an hour
11:34:40 Join spiorf [0] (
11:38:06preglowanyone got an url for the patch?
11:38:59*Bagder got to use xchat's fancy URL grabber window
11:40:29*amiconn uses this feature quite often
11:40:42amiconnIn HydraIRC it's called URL Catcher
11:41:06billytwowillykonversation has that too.
11:41:15billytwowillythe more you know (tm)
11:41:26billytwowillyveryone picture a star with a rainbow tail;)
11:43:11SlasheriJdGordon: did you try wit the newest version?
11:43:21Slasherii updated the patch twice yesterday
11:45:41 Join Spida_ [0] (
11:45:41LinusNworks fine for me, takes a good while until it starts the playback though
11:46:42LinusNpitch works fine for me
11:49:46*preglow tries out the patch with the struggling 24 bit flac file
11:52:00 Quit Spida (Read error: 110 (Connection timed out))
11:57:52preglowworks nice, ui becomes very sluggish
11:58:28preglowit actually does skip less
11:58:33*linuxstb isn't sure if a sluggish UI is good thing
11:59:00 Quit SanAlmighty (Read error: 110 (Connection timed out))
11:59:07*preglow isn't sure if skipping is a good thing
11:59:21 Join SanAlmighty [0] (
11:59:33linuxstbNeither are a good thing - but if a user is using the UI, that should maybe get priority...
12:00:07amiconnUnfortunately there is no real choice
12:00:39amiconnEither make it skip less and slow down the UI, or make the UI responsive and let it skip more
12:01:03amiconnWe only have a fixed amount of CPU power available
12:01:13 Join Xerion_ [0] (
12:01:30 Quit Xerion (Read error: 104 (Connection reset by peer))
12:01:33 Nick Xerion_ is now known as Xerion (
12:01:38 Quit dwihno ("Reconnecting")
12:01:38 Join dwihno [0] (n=dw@
12:01:41linuxstbor 3) Improve the code so neither skips.
12:04:22amiconnOf course
12:04:23 Quit spiorf (Remote closed the connection)
12:05:29 Quit Xerion (Read error: 104 (Connection reset by peer))
12:05:35 Join Xerion_ [0] (
12:05:41tucozwouldn't it be possible with that approach to tune the priority of the ui-thread and the codec-thread so that they get slight more priority depending on what the users focus is atm?
12:05:51linuxstbHow much of a problem is skipping on the iriver? I can't recall ever experiencing it on my h140, but then I always use FLAC.
12:06:43tucozit never skipped for me. I mainly listen to >=192kbps mp3s.
12:06:48 Join spiorf [0] (
12:07:05 Quit spiorf (Remote closed the connection)
12:07:14preglowlinuxstb: flac is what skips for me right now
12:07:17preglowonly thing i can make skip
12:07:32 Join pixelma [0] (i=pixelma@
12:07:42linuxstbDoes it skip if you do nothing, or only when heavily using the UI?
12:07:52amiconnpreglow: AAC does skip on coldfire, doesn't it?
12:07:55 Join spiorf [0] (
12:07:59preglowamiconn: yes, but i have no aac files
12:08:13preglowi'm not entirely convinced priorities is a bad thing
12:08:37*amiconn read that there is now a free (beer) good quality aac encoder
12:08:48amiconnfrom Ahead iirc
12:09:12preglowi prefer something not patented to pieces
12:09:26 Join spiorf_ [0] (
12:09:28tucozisn't the ffmpeg project working on a fixed point aac decoder?
12:09:30 Quit spiorf_ (Read error: 54 (Connection reset by peer))
12:09:39tucozah, you wrote encoder
12:09:55amiconnI surely won't switch to using aac as standard format, but a free encoder is good for making test files
12:10:12amiconnEven if it's only free as in free beer
12:11:10 Quit spiorf (Remote closed the connection)
12:13:43 Quit Rob2222 ()
12:14:21 Join Kohlrabi [0] (
12:14:57 Join spiorf [0] (
12:15:54 Join webguest83 [0] (i=c27f0812@gateway/web/cgi-irc/
12:20:19preglowamiconn: i can't find this encoder
12:23:52linuxstbHmm.... "Compression Ratios ranging from ultra high (58 CDs fit on one!) to High-End Audio (2.5:1), for absolutely perfect audiophile encodings "
12:24:13preglowbah, windows only
12:24:57linuxstbThere's no free beer under Linux...
12:25:07JdGordonSlasheri: i tryed the veriosn u sent me about 1am my time.. so probably not the latest
12:28:17preglowi wonder how functional the aac decoder is now
12:30:10 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
12:30:50 Join barrywardell [0] (n=barrywar@
12:32:55linuxstbpreglow: I've been thinking that another approach to getting a decent AAC decoder would be to start by stripping all the unused code out of libfaad, seeing what's left, and then trying to optimise/rewrite/simplify it.
12:33:07preglowall the unused code?
12:33:09preglowwhat would that be?
12:33:32preglowunless you mean we're only going to support aac-lc, it's not going to be much
12:33:34linuxstbIt's an #ifdef nightmare last time I looked. All the floating point paths, profiles which haven't been ported to fixed point.
12:33:52preglowi'd rather just wait for the ffmpeg code :>
12:34:44 Quit karl (Remote closed the connection)
12:34:54linuxstbI know, but they don't seem to be working on a fixed-point version. Or do you think it would be easier to convert the ffmpeg decoder to fixed-point instead of working with libfaad?
12:36:37linuxstbWe need an AAC fan to come along and do something...
12:37:47Bagderapple certainly knows how to wreck a good gcc
12:38:42 Quit amiconn (Read error: 60 (Operation timed out))
12:38:42 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
12:39:10 Join tvelocity [0] (n=tony@
12:40:54JdGordonwill gcc compplain if you try initialising a struct that has an array in it like this... blaa = { a,b, {ca,cb,cc}, d}; ?
12:44:51JdGordonhmm... it will let me if the array is the last item.. which may be enough
12:48:56 Quit bluebrother ("CGI:IRC 0.5.7 (2005/06/19)")
12:51:01amiconnBagder: Btw, what happened to the darwin) case in configure during your rework?
12:52:14Bagdervery good question
12:52:16barrywardellthat apple patch is still in flyspray...
12:52:42JdGordon:'( why are you aloud to init a char* in a struct but not a int* ?
12:53:36 Quit MarcoPolo (Remote closed the connection)
12:54:50Bagderamiconn: I think that was in
12:55:09LinusNJdGordon: there shouldn't be a difference
12:55:58preglowlinuxstb: the lpc is definitely the bottleneck for 24 bit flac files. performance problems arise just about at the exact same spot in the file where lpc frames start appearing more commonly than fixed frames
12:56:24JdGordonLinusN: thats what i thought... but i get warning: excess elements in scalar initializer when i try to init the int*... ?
12:56:36amiconnThe same lpc stuff that you wrote an asm version for for 16bit files?
12:56:41preglowamiconn: yes
12:56:43Bagderbarrywardell: anything that speaks against committing that macos patch?
12:57:03barrywardelljust the .map thing
12:57:43barrywardelli don't think it affects any other builds so mac just wouldn't have a .map file created
12:58:02 Quit Kohlrabi ("Quit")
12:58:10amiconnBagder: 20:18
12:58:22 Join dan_a [0] (
12:58:40Bagderamiconn: ?
12:58:55Bagderwhen I changed what?
12:59:04BagderI don't see me ever changed anything darwin in there
12:59:05barrywardellthere are also some linking warnings about multiple definitions of functions (I think it was memset among others)
13:03:41preglowanyone have any ideas on what'll be faster? doing two separate mac chains with parallel loads for both the upper and lower part of the 64 bit result, or just using emac for upper part and mulu for lower part without having to load data again
13:04:44amiconnBagder: Hmm. Maybe that totoetlititi used a patch before, which now doesn't apply anymore because configure changed
13:05:10amiconnHe should have statet this though
13:05:22Bagderyeah, it might be that macos patch barrywardell talks about
13:05:50amiconnpreglow: It probably depends on where the data is read from. If it's SDRAM, loading once should be substantially faster
13:05:56linuxstbbarrywardell: Were you able to reliably crash your H10 when enabling cpu frequency scaling?
13:06:18***Saving seen data "./dancer.seen"
13:06:30barrywardellnot necessarily reliably, but disabling scaling stopped the crashes
13:07:12preglowamiconn: should be iram on most targets
13:08:19barrywardellanother thing was that TrueJournals said he had changed boost to 0 in the debug menu. changing it to 1 stopped the crashes even with scaling enabled
13:09:26linuxstbManually setting the boost count to 1 is the same thing as disabling the scaling - the CPU will just run at a constant 75MHz.
13:09:44barrywardelllinuxstb: kernel.c:397. Should that be |= instead of =? It says it should be in system.c:1273
13:10:12linuxstbI think so. Have you tried it?
13:10:35barrywardellit doesn't cause any new problems
13:10:52barrywardellthe crashes are intermittent so i'm not sure if it helps them
13:12:08linuxstbIMHO you should have left the CPU scaling in CVS - to help debug it.
13:12:45dan_alinuxstb: From the comment in system.c it looks like = instead of |= was intentional in a lot of places - is that just something we've copied from IPL?
13:13:08preglowamiconn: on second thought, the mac chain approach is probably faster, yes. a 32 bit mac has a 1 cycle throughput if you chain lots of them, but a mulu always takes about 3-4 cycles, i think
13:13:36linuxstbdan_a: I'm not sure - I didn't copy those parts of the code. The comments in system.c are mine though.
13:14:45 Quit webguest83 ("CGI:IRC")
13:17:30theli_uacan anyone explain what ICONST_ATTR attribute does?
13:17:46barrywardelli don't see why it should be = instead of |=
13:18:00barrywardelllinuxstb: you make a good point
13:18:15barrywardellregarding scaling
13:18:41JdGordonis there any way to get the english version of any translatable string without doing it manually? (i.e counting the \0's in builtin[]?
13:19:33preglowtheli_ua: places a const variable/array in iram
13:20:32theli_uapreglow, i'm just asking becuse when i define some char array as ICONST_ATTR i get different results then when i declare it without such attribute ...
13:21:05theli_uaand that seems strange to me
13:22:20Bagderwhat "different results" ?
13:23:00theli_uai use unsigned char pixmask[4] /*ICONST_ATTR*/ = {
13:23:00theli_ua 0xC0, 0x30, 0x0C, 0x03
13:23:00theli_ua };
13:23:00DBUGEnqueued KICK theli_ua
13:23:00theli_ua for rendering a frame
13:23:21theli_uawhen it has ICONST_ATTR i get corrupted picture
13:23:24preglowtheli_ua: do you remember to copy iramcopy to the actual iram?
13:23:36preglowthere you go
13:23:51theli_uathanks preglow :)
13:24:43amiconnpreglow: What about that 24bit flac on arm? Does it also struggle?
13:26:22preglowamiconn: checking it out now
13:27:38theli_uapreglow, when specifying IBSS_ATTR i need to copy to?
13:28:36preglowtheli_ua: you should always copy if you use iram at all
13:28:50preglowamiconn: it does way, way better
13:28:52preglownever skips
13:28:59preglowcompletely decent boost ratio
13:29:10amiconnIBSS needs to be cleared, not copied
13:29:42preglowlinuxstb: do you know if the qlevel shift parameter in flac can ever be zero
13:30:00theli_uaamiconn, i got lost ... so what needs to be done when specifying IBSS?
13:30:22amiconnCheck how the other plugins which use IRAM do it, e.g. doom or rockboy
13:31:08preglowi still think the iram procedure should be made a macro
13:31:36amiconnrockboom.c lines 43..49 and 788..794
13:34:03theli_uathnx, i've already implemented this :)
13:34:23theli_uai think
13:38:16 Join MarcoPolo [0] (
13:38:47 Join Hansmaulwurf [0] (
13:38:50 Quit SanAlmighty (Read error: 110 (Connection timed out))
13:39:21 Join aliask [0] (
13:39:31aliaskdionoea: Around?
13:40:35dionoeaaliask: ?
13:40:53aliaskJust some thoughts about how to make solitaire prettier
13:40:59aliaskOn H300 at least.
13:41:09dionoeai'm listening :)
13:41:25 Join SanAlmighty [0] (
13:41:28aliaskAt the moment the cards are spaced to take up the full height of the screen, and I don't really like it like that.
13:41:34linuxstbpreglow: Looking at the code and the spec (, I would guess the answer is yes.
13:41:45aliaskIs it possible to adjust the spacing?
13:42:03linuxstbpreglow: Although it seems the shift can be negative - but our decoder doesn't allow it.
13:42:31dionoeayeah, i though about adding a max spacing between cards
13:42:57CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:42:57*amiconn likes how solitaire looks on the h300
13:43:15aliaskI much preferred the old spacings before the rework.
13:43:18dionoeai also noticed a bug in 3 cards mode
13:43:27dionoeamust be from one of my commits yesterday
13:44:44dionoeaaliask: i'll do the max spacing thing
13:44:51aliaskThat'd be great :)
13:45:15aliaskI'd say at least 50% of the time actually using the player would be in that plugin.
13:45:33dionoeai use the chopper plugin a lot too :)
13:45:39preglowlinuxstb: negative? sounds weird
13:45:42dionoeabut that's not commited yet
13:45:57pixelmadionoea: I encountered some strange effects using shortcuts in draw three mode (on Ondio)... all effecting the 3 drawn cards (would be a long list...)
13:45:58aliaskNo, though I like the idea of it. Havn't got around to actually compiling it.
13:46:40pixelmaI think that was already the case before your latest commit
13:47:35dionoeai'll fix those tonight
13:48:59preglowlinuxstb: the 64 bit lpc decoder won't allow 0 either, i think
13:49:08 Join webguest11 [0] (i=c27f0812@gateway/web/cgi-irc/
13:49:59preglownot for coldfire, at least
13:50:16preglowhmm, forget that
13:50:27preglowit would work just nice
13:53:36amiconnpreglow: Do you really need 2 full streaks?
13:54:01amiconnI would guess that for a 24bit lpc decoder it would be sufficient to use 32x32->64 bit multiplications
13:54:32 Join Febs [0] (
13:55:06 Quit webguest11 ("CGI:IRC")
13:55:28preglowamiconn: what do you mean? not calculating the lower bits?
13:57:46*JdGordon 's 3rd try with the settings looks promising :D
13:58:34amiconnpreglow: The same way you said that for 16bit it is possible to use 16x16->32 multiplications...
13:59:48preglowi _think_ it's possible. i did a test once, and it sounded just fine
13:59:57preglowi should investigate that further
14:00:19preglowanyway, it all depends on the size of the shift
14:04:14preglowaccording to the same spec page linuxstb pasted, it does seem like it. what gets multiplied is coefs and samples, the samples are 16 bit or lower in that case and coefs can't be any more than fifteen bits, it seems
14:04:28theli_uacvsdo: ERROR: Directory has sticky date revision; cannot add file ... what can i do about this ? :(
14:05:39amiconnpreglow: Then hooray; I should do some experiments with SH1 mac and flac...
14:06:05amiconn24bit flac is very probably out of question, of course
14:06:20preglownot a loss i would mourn anyway
14:07:55preglowhmm, one rotten thing is that both samples and coefs are saved as 32 bits anyway
14:08:06preglowyou should modify that for sh1
14:08:23preglowunless of course that would be suboptimal for some reason i don't know of
14:08:25preglowsh1 is weird :)
14:08:39amiconnsh1 isn't weird
14:09:14amiconnIt's quite straightforward risc
14:09:32amiconn...imho easier to understand than e.g. arm
14:10:15amiconnSH-DSP would be more involved
14:12:56preglowhmm, it wouldn't be completely safe without additional checks, though
14:13:28preglowthe decoder just checks that the sum of the coef precision and sample precision is lower than 32 bits before invoking the "16 bit" lpc decoder
14:20:58 Quit tucoz ("Leaving")
14:26:03 Quit barrywardell ()
14:28:45 Join daurn|laptop [0] (n=quae@
14:29:54 Join bluebrother [0] (
14:36:20JdGordonwhy are there so few good looking wps's for the h300 and so many for ipoo :'(
14:36:41BagderJdGordon: because all the h300 ones are on
14:36:55Bagder... and are made for weird custom versions
14:38:45JdGordonstupid AA wps's :p
14:39:32 Join webguest10 [0] (i=549965f6@gateway/web/cgi-irc/
14:40:09webguest10jdGordon: for H300 wps at mistic
14:40:33JdGordonim there already.. found 1 i migth like... checking what it looks like on the player :p
14:41:09*JdGordon is sick of ajant
14:41:41webguest10some pretty ones and some wacky ones
14:42:48 Quit uski (Read error: 60 (Operation timed out))
14:43:13 Join uski [0] (
14:43:21webguest10StarTrek 2.0 is really wacky, wonder if trekkie fans would like it
14:45:05webguest10iPod H500 looks nice
14:45:13FebsYou know, just before the freeze, there was talk of incorporating the margin patch into CVS.
14:45:38FebsI think that is the bigger issue with using many of those WPSes on a standard build, rather than the AA issue.
14:45:56*Bagder runs for cooooofffeeeeee
14:47:00linuxstbFebs: The problem with doing that IMO is that we will make WPSes work with CVS, but then we will probably break them all again when viewports are implemented - assuming viewports will replace the margins patch.
14:47:29JdGordonviewports isnt really happening any time soon tho is it?
14:47:58 Part webguest10
14:48:11FebsIs it realistic to think that implementing viewports won't break existing WPSes anyway?
14:48:41 Quit tvelocity ("Αποχώρησε")
14:48:55FebsFair enough.
14:49:35 Join Kohlrabi [0] (
14:50:07amiconnViewports will default to the whole screen unless the wps specifies another viewport
14:50:24amiconnSo old wps'es that don't specify viewports will just work like before
14:51:33FebsThat makes sense.
14:51:36amiconnJdGordon: One reason why I still didn't start working on viewports is because I have to catch up with all the archos breakage :/
14:51:41 Join Arathis [0] (
14:52:47FebsStill, it's not like those WPSes work with Rockbox now anyway, so they're going to have to be updated when viewports are implemented one way or the other.
14:52:55amiconnAnd I still use my archoses quite often, one reason being that rockbox still isn't really reliable on swcodec, in modules I don't understand well enough to fix them myself :/
14:52:58JdGordonamiconn: have you tryed the shutdown patch yet?
14:53:51linuxstbFebs: True, but people will also create new WPSs.
14:54:26FebsI don't even use any of those WPSes. I'm just tired of the "Why is my WPS messed up" threads in the forums.
14:54:27*JdGordon gone
14:54:40 Quit aliask ("Chatzilla 0.9.75 [Firefox]")
14:54:51amiconnFebs: I'm not talking about wps'es for inofficially patched rockbox, but about those which are working with current cvs
14:57:32FebsSorry, I wasn't clear. When I said "those WPSes" I was talking about the ones that work with unofficial builds.
14:57:58FebsBut I understood your point.
14:59:06amiconnYes, but adding the margin patch would make them official
14:59:34linuxstbFebs: I'm sure there are always going to be those problems - the WPS seems to be the favourite place for people to hack new features into, and one of many places where no developer with CVS access seems to work...
15:00:57amiconnAnother reason why I didn't start viewports yet is that I deem wps design fairly unimportant. It's just design, which helps nothing if the functionality is broken
15:01:31amiconnBut since viewports will also be useful in other places, I'll start implementing them sometime soonish
15:01:31FebsI agree with you completely on that point.
15:01:46FebsI'm always amazed how much people value the appearance over functionality.
15:02:05amiconnDepends on what other things there are to clean u
15:02:42*amiconn in fact uses the default wps on his swcodec targets
15:02:43FebsThere was a post in the forum the other day by someone who was going to return to the original iriver firmware because he didn't like Rockbox's FM radio screen.
15:03:46bluebrotherJdGordon, have you had a look on There are quite some nice looking for h300
15:04:07amiconnI also want to work on pcm recording for MAS... but hopefully that will be quite simple
15:04:18amiconn(as a techpreview plugin first)
15:04:46bluebrotherI believe all wps from that page work on unpatched versions.
15:04:47dwihnoWhat do I hear? Has there finally been docs made available on the mas?
15:05:14*amiconn wonders how long dwihno was in deep-sleep mode ;)
15:05:21*Bagder puts the blanket back over dwihno, go back to sleep
15:05:25amiconnCheck the datasheet page...
15:05:47dwihnoBagder: Misa wants coffee! No blanketses!
15:05:49amiconnThe wavplay plugin is available & working since early april...
15:06:02dwihnosorry, haven't been too active :)
15:06:22***Saving seen data "./dancer.seen"
15:06:23amiconnPlays canonical 16bit wav files with all standard sample rates, mono & stereo
15:06:29 Quit MarcoPolo (Remote closed the connection)
15:06:36amiconn...on Recorders and Ondios
15:07:19dwihnoWho got the PDF?
15:07:27amiconn[15:05:25] <amiconn> Check the datasheet page...
15:07:36markunI'm wasting my time on glyph drawing again:
15:07:54amiconn[IDC]Dragon finally got the docs, code, and permission for distribution in July 2005
15:08:18dwihnoamiconn: ah, okay... a heapload of thanks to him for his work then \o/
15:08:37amiconnHe also was the one who started working on the wavplay plugin at devcon, but didn't get it working
15:09:01amiconnI fixed it and added some more things like DMA and WAV header evaluation
15:09:13dwihnoneato! great work!
15:09:17dwihnoI need to test it out then.
15:09:56linuxstbmarkun: Can you read Arabic?
15:10:00amiconnWhat's still missing is the integration into the core playback enginwe
15:10:26markunlinuxstb: a bit, but not the meaning of the words :)
15:10:27amiconnSo you can just play a single wav file, then select the next one
15:11:10linuxstbamiconn: Have you considered not unifying the playback engines?
15:11:18 Join tvelocity [0] (n=tony@
15:14:10 Quit daurn|laptop (Read error: 104 (Connection reset by peer))
15:14:21amiconnlinuxstb: I would in fact prefer to unify them, especially if things like metadata on buffer get added, but surely won't start doing that until the engine becomes stable
15:14:38 Join daurn|laptop [0] (n=quae@
15:14:53amiconnThat's why I want to look into pcm recording; this might be easier to integrate in the core than pcm playback
15:15:18amiconnI "just" need to cut out recording from the mpeg thread, and add proper MAS status handling
15:17:26amiconnThis will also help to bring down power consumption in FM radio mode
15:18:31 Join Mmmm [0] (
15:20:01 Nick belZe is now known as belze (
15:20:25 Quit TCK (Read error: 104 (Connection reset by peer))
15:21:02 Quit Pyromancer ("Ex-Chat")
15:21:24 Quit Arathis (Read error: 104 (Connection reset by peer))
15:24:08 Join Pyromancer [0] (
15:26:46 Quit dan_a (Read error: 110 (Connection timed out))
15:31:11 Quit Xerion_ (Read error: 104 (Connection reset by peer))
15:31:12 Join Xerion [0] (
15:32:53 Join Arathis [0] (
15:34:04 Part LinusN
15:40:03bluebrotherwith the FM mentioned some time ago: does somebody want to have a look at
15:41:31FebsIs mpa.codec the codec for MP3?
15:42:25preglowand mp1 and mp2
15:43:26linuxstbpreglow: Do you have any idea what could cause a call to mad_decode_frame() to never return? (I'm trying to use it in mpegplayer...) It's a problem that only seems to happen on my ipod, not in the sim.
15:43:48linuxstb(sorry, mad_frame_decode()
15:44:23preglowhmm, no
15:44:36FebsSomeone on MR claims that there is a sound quality issue with recent builds:
15:47:59 Quit belze (Read error: 110 (Connection timed out))
15:49:48 Quit SanAlmighty (Read error: 110 (Connection timed out))
15:52:07 Quit _FireFly_ ("Leaving")
15:55:59 Join _FireFly_ [0] (
16:03:41 Quit daurn|laptop (Read error: 54 (Connection reset by peer))
16:04:01 Join daurn|laptop [0] (n=quae@
16:14:04 Join TrueJournals [0] (
16:14:57 Join senab [0] (
16:15:13senabjdgordon, you here?
16:20:42 Quit senab (" HydraIRC -> <- Leading Edge IRC")
16:22:04Bagdermore spam than baked beans these days
16:22:21 Join Nibbier [0] (
16:23:51 Part TrueJournals
16:26:56 Join PaulJam [0] (
16:31:48linuxstbFebs: Maybe that person on MR is comparing sizes of the Musepack (mpc.codec) file - which is 40KB in the current build.
16:37:16linuxstbFebs: There have been no changes to the codec for about 4 months. So if there has been a bug introduced, it will probably affect all codecs - it's worth asking the poster to test some other formats (e.g. FLAC) with both his old and new builds.
16:37:32 Join rretzbach [0] (
16:39:01 Quit daurn|laptop (Read error: 104 (Connection reset by peer))
16:39:31 Join daurn|laptop [0] (n=quae@
16:39:34 Quit daurn|laptop (Read error: 104 (Connection reset by peer))
16:41:54preglowwhen did amiconn do his bss magic?
16:43:59 Join XavierGr [0] (
16:44:44linuxstbpreglow: January...
16:45:32linuxstbIf you're talking about revison 1.34:
16:50:53amiconnbss magic?
16:52:40linuxstbRemoving the bss and ibss from the .codec/.rock files I think.
16:53:21preglowi was just thinking about the alleged reduction in file size
16:53:26amiconnAh ok. Dunno if that can be called 'magic' ...
16:53:28preglowthat's the only reason i could think of
16:54:05PaulJamis there a reason why rockbox since a few days sometimes spins up the disk before entering a menu?
16:54:46Mmmmthe settings reset and you forgot to re-enable dircache?
16:54:58Bagdermenus don't use dircache
16:55:25 Quit perplexity (Read error: 113 (No route to host))
16:55:32PaulJambut this happens for example when i enter the playback menu, where the disk doesn't have to be accessed
16:56:44amiconnDo you have voice enabled?
16:57:28PaulJamvoice for menus was enabled, but i have no voice file in the lang folder.
16:57:56 Join mantono [0] (
17:00:15 Quit theli_ua ("Leaving")
17:00:44 Join perplexity [0] (
17:02:23PaulJamalso, with dircache enabled it sometimes spins up the disk when entering a directory
17:04:04 Join dpassen1 [0] (n=dpassen1@resnet-236-163.resnet.UMBC.EDU)
17:06:25***Saving seen data "./dancer.seen"
17:06:57markunPaulJam: which font are you using?
17:07:07PaulJamhmm, this doesn't seem to happen with default settings
17:07:15 Quit JdGordon ("Konversation terminated!")
17:07:15PaulJammarkun: nimbus 12
17:07:28markunMight have something to do with font caching
17:09:13PaulJamyou are propably right, i just changed the font to nimbus 12 and it started to spin up when entering a menu.
17:13:43 Quit Mmmm ("Byeee")
17:15:00 Quit apo ("Leaving")
17:18:25 Join apo [0] (
17:20:15 Quit spiorf (Remote closed the connection)
17:20:21 Nick Strogg_ is now known as Strogg (n=user@
17:20:24 Quit apo (Read error: 104 (Connection reset by peer))
17:20:35 Join apo [0] (
17:20:39 Join barrywardell [0] (i=892b7a6e@gateway/web/cgi-irc/
17:20:48 Join spiorf [0] (
17:21:47 Quit lex (Nick collision from services.)
17:21:48 Join lex [0] (
17:22:01 Quit lex (Nick collision from services.)
17:23:01 Join dan_a [0] (
17:23:53 Quit spiorf (Remote closed the connection)
17:24:13 Join spiorf [0] (
17:26:15 Quit apo (Remote closed the connection)
17:26:47 Join lex_ [0] (
17:28:47 Join apo [0] (
17:30:00 Quit spiorf (Remote closed the connection)
17:30:23 Join spiorf [0] (
17:32:23 Join Criamos [0] (
17:38:36 Join elinenbe_ [0] (
17:38:45 Quit elinenbe (Read error: 104 (Connection reset by peer))
17:38:45 Nick elinenbe_ is now known as elinenbe (
17:39:30 Join bluey- [0] (
17:44:41 Quit tvelocity ("Αποχώρησε")
17:45:54 Quit Febs (Read error: 110 (Connection timed out))
17:47:48 Quit PaulJam (".")
17:49:21 Join rconan [0] (n=richard@
17:49:39 Quit rconan (Client Quit)
17:49:48 Quit CriamosAndy (Read error: 110 (Connection timed out))
17:57:44 Quit bluey- (Remote closed the connection)
17:58:17 Quit webguest35 ("CGI:IRC 0.5.7 (2005/06/19)")
17:58:24 Join tvelocity [0] (n=tony@
18:00:39 Join apo` [0] (
18:02:01 Quit apo (Read error: 110 (Connection timed out))
18:05:58 Join lightyear [0] (
18:07:54 Quit perplexity (Read error: 110 (Connection timed out))
18:08:12 Quit barrywardell ("CGI:IRC")
18:13:06Slasheriamiconn: here is the cleaned up version of the scheduler patch:
18:13:20preglowdoes it have comments? :>
18:13:29Slasheriah, not yet.. =)
18:13:49Slasherihmm, probably i should add those next
18:14:16preglowworth checking out the sentiment to commiting it first anyway
18:14:21preglowi foresee some resistance :>
18:15:10Slasherihehe, true :)
18:16:15Slasheribut at least nobody has yet proven it opens a can of worms ;)
18:16:21dan_aIf I can understand the code and it makes MP3s play at less than 100% boost on 3G iPods, then it'll get my vote.
18:16:35dan_aI'll give it some testing when I get home
18:19:16dan_aSlasheri: Have you had any thoughts about how well / badly your patch integrates with the dual core work I've been doing? Off the top of my head I can't see why it should cause any problems
18:24:30Slasheridan_a: it should work just fine
18:24:51Slashericurrently it operates just on single core as before
18:25:01Slasheribut now it should be easier to move tasks between cores on the fly
18:26:27dan_aExcellent - though I'm not sure we'd want to move tasks between cores (What a big pile of cans of worms you have there!)
18:28:11Slasherihehe :D
18:28:54dan_aI think if you're changing the way you call create_thread then it shouldn't be a problem to get rid of create_thread_on_core, too. Just add "unsigned int core" to the parameter list, and make anything that calls it at the moment use CURRENT_CORE
18:29:20*EV|Server is back (gone 09:46:07)
18:30:06dan_a(and that goes for *_on_core too)
18:31:47 Join barrywardell [0] (i=892b7a6e@gateway/web/cgi-irc/
18:36:49 Join TrueJournals [0] (
18:39:02 Join webguest30 [0] (i=c3ec4b1b@gateway/web/cgi-irc/
18:50:09 Join rotator [0] (n=e@rockbox/developer/rotator)
18:55:17 Join Lear [0] (
19:01:40 Join TeaSea [0] (
19:01:57markunhi mirak
19:03:22mirakso I did in asm the idct_row and idct_col. I can use them to replace the C version but only separately otherwise there is artifacts. there is also a problem if I don't clear the accumalators each time at the begining of the function, despite clearing them always at the end.
19:03:45mirakso I am wondering if any thread could interfere
19:06:26LearIf using the emac works in other plugins (and the codecs), it ought to be something else, I think...
19:06:29***Saving seen data "./dancer.seen"
19:10:56mirakLear: don't know, maybe there are sound quality loss
19:11:09mirakthat we don't perceive
19:12:10mirakI can show you the idct.c file in a pastebin. I am really sure that my last move from acc are movclr.
19:12:54markunmirak: maybe it doesn't get cleared by some other functions that use the emac?
19:12:55mirakLear: do you know how ordonancing works ?
19:13:06LearWhy not? A fresh pair of eyes might help...
19:13:20LearOrdonancing? What is that?
19:13:26mirakLear: hem threading ?
19:15:26linuxstbmirak: A thread calls yield() in order to tell the kernel to switch threads. So your code won't be interrupted by another thread unless you call yield().
19:15:39Learmirak: you expect me to know something about which google gives no hits? ;)
19:16:23mirakLear: no, I don't know how much you are involved in rockbox, or wich part you know the best
19:16:36mirakLear: unless you tell me to google
19:17:24LearWell, I know a bit, but I don't really have a speciality. Tremor maybe...
19:18:07preglowmirak: do you set the macsr yourself?
19:18:14preglowmacsr is probably set to fractional mode by default
19:19:02 Join ep0ch_ [0] (
19:19:36mirakpreglow: yes I always set it when entering the function for now. I will do a row and columns in one function. in fact I have done it but it doesn't work well yet.
19:19:58preglowand you're absolutely sure you remember to always clear the accumulators?
19:20:31markunmirak: can you show us your code?
19:20:39preglowno thread should be able to interfere if you do that
19:20:45 Part TrueJournals
19:20:49preglowunless you're actually yielding inside the function, which you shouldn't
19:22:10 Join Paul_The_Ner1 [0] (
19:22:40mirakmarkun: yes I am pasting
19:23:23mirakpreglow: there is really not much code
19:23:23 Quit uski (Read error: 104 (Connection reset by peer))
19:23:25Bagderdan_a: your server seems to be compiling fine now
19:23:35preglowpastebin, ples
19:23:49mirakpreglow: there is maybe one thing I don't do it's to clobber the acc
19:23:58mirakbut when I tried gcc didn't liked
19:24:06 Part Paul_The_Ner1
19:24:13mirakand anyway I think gcc doesn't use MAC right ?
19:24:18 Join uski [0] (
19:24:49preglowgcc doesn't even know about it
19:25:33mirakthe idct_row and idct_col wich works are at line 470 and 590
19:25:53mirak407 instead of 470
19:26:20 Join perplexity [0] (
19:28:20mirakactually the file is set to use idct( ). you need to change some ifdefs to have it use the two others
19:29:59 Part pixelma
19:30:01mirakas they are you can't mix the idct_row c and assembly version, some things must be changed, I can tell you
19:31:33mirakidct_row and idct_col must be used both as asm or both as C right now. The C version uses some tricks to have less coefs and multiplications, but it's harder to use mac with that algorithm
19:32:29LearCode and comment doesn't seem to match... (e.g., "load P[1] P[3] P[5] P[7]" and following code)
19:32:42mirakLear: which line ?
19:34:16mirakLear: yes I should comment more, and this point was what gave me a lot of problem. in fact the dots for a row are ranged like that in the block : P[0] P[2] P[4] P[6] P[1] P[3] P[5] P[7]
19:34:34mirakit took me some time to figure this out
19:35:01mirakthey could have explained that in the code.
19:35:15Learand the digits really are the array indexes?
19:36:29mirakhum not sure for you question
19:36:35mirakwhat's your question ?
19:36:39 Join lodesi [0] (
19:37:24mirakLear: if P is the array with the logical representation of the dots.
19:37:32LearE.g., the "1" in "P[1]".
19:37:56mirak[P[0], P[2], P[4], P[6], P[1], P[3], P[5], P[7] ] is the array of the physicall representation
19:39:14 Quit dan_a (Read error: 110 (Connection timed out))
19:39:37mirakLear: look here page 41 . this is the matrix multiplication that I do
19:41:19 Join BHSPitLappy [0] (
19:42:43lex_/dev/sdc2 19G 19G 0 100% /media/ipod-1
19:44:07mirakLear: you gave up ?
19:44:37LearNot really, but you do seem to clear the accumulators properly.
19:44:55 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
19:44:57LearAnd though I think the comment is confusing, the asm code appears to match the C code...
19:45:14 Join Kohlrabi [0] (
19:45:14mirakI didn't knew mac add a fractional mode
19:45:46mirakLear: there is no C code matching this function
19:45:56mirakLear: wich one are you looking at ?
19:46:24Learidct_row. there's a C function in the #else block...
19:47:10mirakLear: yes it's supposed to do the same, at a *sqrt(2) difference
19:47:16 Join Angry [0] (
19:48:21mirakLear: it's by developping back the idct_row c code I realised the dots are not contigous
19:48:29 Quit Arathis (Read error: 110 (Connection timed out))
19:48:40 Quit tvelocity ("Αποχώρησε")
19:49:47amiconnmirak: Do you know whether the data the idct is working on is located in iram or dram?
19:50:37mirakLear: the function at line 204 is an atempt to copy in asm the C fuction
19:50:52amiconnIf it's dram, there is an optimisation possibility in idct_row: If you reverse the order of calculations in the /* mat 1 - mat2*/ part, the results are in the same sequence as they are written to memory. So you can collect the results in 4 32bit registers and write them to memory with a single movem.l
19:50:59mirakamiconn: that's dram, though I know how to put it to iram.
19:51:23mirakamiconn: I did it and got a big boost.
19:51:48 Join Arathis [0] (
19:52:51mirakamiconn: the problem is how this memory block is allocated. It's statically allocated inside the mpeg2_decoder_s struct at line 95 in mpeg2_internal.h
19:53:27Learhm.., in idct_col, the move.w indexes (when reading the block) looks suspicious...
19:53:40 Nick lex_ is now known as lex (
19:55:03LearUnless the matrixes are 8-entry blocks, so the "*16" is just a shorthand of "*2*8"...
19:55:25amiconnThe idct block is 8x8 on shorts
19:55:42mirakamiconn: in slice.c you can declare a int_16 DCTblock [64] in iram, then replace decoder->DCTblock with DCTblock
19:55:43amiconnSo getting one row down means 16 bytes (8 shorts)
19:56:47mirakI think the best thing would to have all mpeg2_decoder_s in iram but I don't think it would fit
19:57:03 Quit Kohlrabi (Nick collision from services.)
19:57:09 Join Kohlriba [0] (
19:59:10amiconnmirak: What was the problem with that code again?
19:59:11mirakquantizer_prescale would be interesting in iram but it's 16k big
19:59:35mirakamiconn: problem is that if I don't clear the accs it's all garbeled
19:59:38amiconnBtw, while I don't understand much about idct, I can see several optimisation possibilities there
19:59:48mirakhum set 0 to them at the begining
20:00:05mirakamiconn: go ahead
20:00:08LearAnd no emac asm in other parts?
20:00:18mirakLear: wich parts ?
20:00:38LearParts of mpegplayer other than the file you pasted.
20:00:38mirakLear: oh in mpegplugin ?
20:01:04miraknot that I modified, but I think the copy to screen is in asm, I will look at it
20:02:17amiconnE.g. I would load all 8 coefficients at once in th erow part
20:02:36 Join webguest78 [0] (i=ccd290a4@gateway/web/cgi-irc/
20:02:45amiconnmirak: The lcd_yuv_blit() uses asm, but no emac instructions
20:02:51 Quit webguest78 (Client Quit)
20:03:05 Join delorean90 [0] (i=ccd290a4@gateway/web/cgi-irc/
20:03:17mirakamiconn: there is not enough registers
20:03:25amiconnIn fact it could perhaps profit from emac, but even if we introduce that, we'd use movclr at the end
20:03:34amiconnmirak: There are enough registers
20:03:45mirakamiconn: well I prefer keep some register to do the block in one asm function
20:03:46 Quit delorean90 (Client Quit)
20:04:15amiconnI am looking at the full idct()
20:04:34amiconnLoading 8 coefficients instead of 4 just means using 2 more registers
20:05:09LearOn the CrossCompiler wiki page, it says that when using gcc 4.0.3 on coldfire targets, "sometimes crashes are observed". Anyone who knows anything about that?
20:05:18amiconnAnd I wonder what you
20:05:32LearI've used such builds a few days, and it seems fine to me...
20:05:33mirakyes, but I need to save the second matrix in d4 d5 d6 d7 , and d3 is used for the loop
20:05:37amiconn're using %a4 for. %[block] itself gets never modified
20:06:07amiconnI never said you'd need d registers. a registers work equally well for emac
20:06:21mirakamiconn: yes, I can do a suba #128,%[block]
20:07:06 Join bbroke [0] (
20:07:07mirakafter doing the 8 rows loop
20:07:16amiconnAh, hmm, in fact you do
20:07:29amiconnBut then you can readjust without wasting a register
20:07:51amiconnBtw, you don't need to specify the i and a variants explicitly
20:08:07amiconnJust write sub.l #128,%[block]
20:08:22amiconnThe assembler knows %[block] is an a register
20:08:26 Join Pyromancer_ [0] (
20:08:30mirakok I wasn't sure for that. seem the compile adjust itself
20:09:01mirakI was thinking it maight be faster
20:09:31amiconnYou just need to remember that some operations aren't possible on a registers (e.g. shifts), or work different (e.g. arithmetic operations don't set flags)
20:09:56 Join niskel [0] (
20:10:04mirakamiconn: yes, discovering that when coding is the hardest part. when you must drop your code
20:10:53amiconnBtw, the accumulator clearing in the middle of the full idct (between rows and columns) should definitely be superfluous
20:11:38 Quit Pyromancer_ (Client Quit)
20:12:09 Quit niskel (Remote closed the connection)
20:12:23 Join Pyromancer_ [0] (
20:12:45mirakamiconn: that was to test why the accs seems to not be cleared
20:13:07mirakthe move #0,acc0 should be superfluous also
20:13:49amiconnThat's what I mean
20:14:03mirakthere is one weird thing, I can't inline idct() because gcc says the row and col labels are used twice.
20:14:23amiconnThat means idct() is used in 2 places
20:14:42amiconnIf you want to inline it, you need to stick to local labels
20:14:47mirakyes, in idctcopy and idctadd
20:15:02mirakamiconn: how do you do that ?
20:15:23Learamiconn: you aware of any reason to not use "-m5206e" instead of "-m5200" when compiling rockbox for coldfire targets?
20:15:48Learit generates slightly smaller/faster code, and works fine in my tests (h140).
20:15:50mirakbecause 5200 is more generic ?
20:16:02mirakah ok
20:16:04amiconn(1) What's the exact difference? (2) I though we use -m5249 ...
20:16:21mirakgreat. Lear : rivers are 5249 right ?
20:16:56Learthe difference I'm aware of, is that with -m5206e, gcc actually uses the div instruction (early 5200 models didn't implement that, so gcc can't use it).
20:17:45 Join dan_a [0] (n=dan-mirc@
20:17:48Lear−−target-help for gcc do list -m5249, but gcc doesn't recognize it. -m5206e works though.
20:18:35mirakthe cores are backward compatible ?
20:18:38amiconnHmm, mine doesn't show -m5249
20:18:42amiconnmirak: Yes
20:19:02amiconnCFPRM.pdf says DIVS, DIVU, REMS and REMU are 5206e and up
20:19:13Learmine does, under 680X0 options.
20:19:33amiconnWe shouldn't run into problems by using it. All irivers we support have 5249, the X5 has 5250
20:19:51 Quit Pyromancer (Read error: 110 (Connection timed out))
20:19:56amiconnAH, indeed
20:20:12LearSo I'll commit the change then. Just wanted to double-check it...
20:21:01 Join bluey- [0] (
20:21:16amiconnLear: Did you check whether it makes gcc also use MAC?
20:21:21amiconn(not EMAC)
20:21:39Learno mac use that I'm aware of. Didn't look much for it though...
20:21:42preglowi'm almost 100% certain it won't
20:21:59 Nick Pyromancer_ is now known as Pyromancer (
20:22:05amiconnChapter 12 lists divide and MAC support beginning with 5206e
20:23:30mirakwoudln't it be faster to make gcc use mac d0,d1,acc0 ; moveclr acc0,d1 instead of mul d0,d1 ?
20:23:53Learonly mac use in rockbox.elf is dsp stuff.
20:24:20mirakmul is 4 cycles and mac is 1
20:24:45Learthere's a stall between the mac and the moveclr, so the first one would take at least 3 cycles...
20:27:52 Quit Angry ("Miranda IM! Smaller, Faster, Easier.")
20:28:06preglowexactly three
20:28:19preglowso it'd be just as slow as an ordinary mul
20:28:29preglowas a matter of fact, it would be slower
20:29:05preglowonly use mac instructions when you need to do actual multiply-and-accumulate operations, or when you need a quick 64 bit mul
20:29:07amiconnMy guess is that the mac/movclr sequence needs the same cycle count as the mul
20:30:18amiconn..because it's very likely that the (e)mac-equipped coldfires actually use the (e)mac multiplier for performing mul as well,
20:30:30preglowit does
20:30:36preglowand yes, the cycle count matches
20:30:51preglowbut thanks to the mac/movclr being bigger than one mul instruction, it'll be slower
20:30:52amiconnand what happens internally is almost the same as the mac/movclr sequence
20:31:08amiconnHello 'risc' ...
20:33:43 Join TrueJournals [0] (
20:34:29 Quit Pyromancer ("Ex-Chat")
20:34:39mirakamiconn: is there any way to load in less moves a column ? I also use a swap just to save two registers, otherwise I could word only on the low part
20:35:15 Quit bluey- (Remote closed the connection)
20:35:54amiconnIf you have enough registers, it would be faster to not swap and only use the low part. But afaics there is no way to speed up the load
20:36:09amiconnDoes the full idct() work as-is?
20:36:16 Quit perplexity (Read error: 110 (Connection timed out))
20:36:32 Join bluey- [0] (
20:36:40mirakamiconn: it doesn't, but it should
20:37:17mirakmaybe I do bad loops or bad incrementations
20:37:44amiconnThe individual _row and _col versions do work, even if both are enabled?
20:38:26mirakamiconn: no
20:38:29mirakI explain
20:38:41 Quit barrywardell ("CGI:IRC")
20:39:32 Join Rob2222 [0] (
20:40:32mirakthe C _row and C_col uses some tricks, like mutiply the theoric coefficient by sqrt(2). this is done because cos(4*pi/16)=sqrt(2). you then have multiplications by 1 wich is the same than not multiplying, therefore you save some multiplications
20:41:29mirakso if you want the asm version to be like the the C version you need to use the commented coefs.
20:41:59mirakamiconn: I do the changing
20:42:02mirakso you can test
20:42:04preglowso, should we switch from 5200, then?
20:42:39Learalready did that... :)
20:42:41amiconnmirak: What purpose/advantage do the other, non-commented coefficients have, then?
20:44:26mirakamiconn: well since mac.l is one cycle I didn't bothered doing just a move in the acc0
20:44:48mirakamiconn: that was also for test purpose
20:45:14amiconnYes, but if the coefficients don't match, the final result will also be different?
20:45:22amiconnOr do I miss something important here?
20:45:37preglowi like the fact that no one noticed until now that gcc never does emits div...
20:45:45 Quit bluey- (Remote closed the connection)
20:46:11mirakthe coef constants Wx are the same than Wa etcetera exept some are *sqrt(2) and the others are not
20:46:38TrueJournalsWhy aren't bootloaders avaliable for download with CVS? The build machine builds them...
20:46:39mirakpreglow: that's a happy surprise indeed
20:46:59 Join pixelma [0] (i=pixelma@
20:47:59amiconnmirak: I understand that the coefficients are different; what I don't understand is how those different coefficients should produce the same final result...
20:49:16TrueJournalsHas anyone ever used to build the tagcache database?
20:49:23mirakamiconn: sorry, the coef are the same exept idct_row in C uses coef*sqrt(2) to save some mults. but this saving as no meaning with the MAC approach exept you could save 1 register and also replace P[0] mutliplications with a move P[0],acc0
20:49:59amiconnYou already said this multiple times, but that doesn't answer my question
20:50:28mirakamiconn: this is because matrix multiplication is linear
20:50:36amiconnIf you have certain values, and multiply all coefficients with sqrt(2), the final result will also be multiplied by sqrt(2)
20:50:51amiconnBut you want the result to be the same, how is that achieved?
20:51:24 Join merbanan [0] (
20:51:31mirakamiconn: because the idct_col also multiply by sqrt(2), and in the end it's like multiplying by 2
20:51:52amiconnAnd then you need to shift one bit more/less?
20:52:18mirakamiconn: yes they do one more right shift.
20:52:31 Quit Caliban (Read error: 110 (Connection timed out))
20:52:44mirakamiconn: but anyway for testing purpose and to ease things for now we should use the same coef than the C version uses.
20:52:53TrueJournalsNo one's used the perl script to build tagcache?
20:53:00amiconnNah, I don't think so
20:53:12amiconnBut the coefficients could be double-checked
20:53:24mirakamiconn: what ?
20:53:35 Join Pyromancer [0] (
20:53:48amiconnThere might be a mistake/typo/whatever in the modified coefficients...
20:54:11mirakamiconn: you mean mine or the original ones ?
20:54:38mirakyou can verify
20:54:49 Quit Landus ("Leaving")
20:54:49mirakamiconn: I used octave/matlab a lot to simulate things
20:55:08 Join lodesi_ [0] (
20:55:17mirakI think the coefs are right I double checked, but you can check also :)
20:57:19mirakamiconn: there is also one thing. They add 2048 to the shifted d0 variable in idct_row and do the same in idct_col by adding 65536. that's maybe to correct some precision loss.
20:57:54mirakyou should see that add commented in the asm code.
20:58:20amiconnIirc the dc component is handled differently
20:58:37amiconnI remember having seen that in the jpeg decoder
20:58:58amiconn...even though I don't understand everything in it
20:59:36amiconnOh, and when you add 65536, you will overflow mac.w for sure
21:00:25preglownice, nice, 3k shaved off rockbox.iriver
21:00:33preglowwhich is still somewhat huge at 330k
21:01:41Slasherii am worried that the bootloader is getting bigger and it will hit the 64 KiB limit soon..
21:02:17Slasherionly 2k left
21:03:29 Quit lodesi (Read error: 110 (Connection timed out))
21:05:31mirakamiconn: yes I probably didn't noticed it.
21:05:36 Quit TeaSea (Remote closed the connection)
21:05:43 Quit _Lucretia_ ("Leaving")
21:05:59 Join TeaSea [0] (
21:06:01mirakamiconn: but that seems obvious
21:06:33***Saving seen data "./dancer.seen"
21:06:42amiconnHmm, but it seems they do the multiplication in a completely different way
21:06:54mirakI think I used it when I used a full register
21:07:06amiconnDo you use the emac in fractional mode?
21:07:12mirakamiconn: I bakward developped the C algorithm
21:07:13preglowSlasheri: i don't think slimming it down further should be impossible
21:07:14 Join rbalboa_ [0] (
21:07:25 Join webguest71 [0] (i=549c2368@gateway/web/cgi-irc/
21:07:44mirakamiconn: it's the same matrix multiplication than described in the pdf I gave a link
21:08:15mirakamiconn: what's fractional mode ? I didn't see that in the manual
21:09:04linuxstbBTW, has anyone noticed if iriver use the EMAC in their firmware?
21:09:04 Quit webguest71 (Client Quit)
21:10:24mirakwouldn't have they if they used asm ?
21:11:29 Quit Arathis (Read error: 104 (Connection reset by peer))
21:11:43linuxstbI guess it depends if any asm code was written specifically for the Coldfire or not.
21:12:22amiconnI guess they do use it, at least in the motorola mpeg audio decoder
21:13:32linuxstbMP3 decoder - it failed to play my MP2s...
21:15:32*amiconn never tried that
21:19:45 Join _Lucretia_ [0] (
21:19:56dan_aSlasheri: Your scheduler changes do bad things to the 3G iPod.
21:20:41Slasheridan_a: hmm, what kind of things?
21:20:44dan_aI didn't notice any major improvement in MP3 playback, but the UI becomes horribly sluggish
21:21:03dan_a(this is with your version 2 patch against a clean CVS tree)
21:21:19Slasheriweird, and you was playing a standard mp3 file ipod should normally handle?
21:21:28mirakamiconn: the value to add is 32 instead of 65536. I do //"add.l#32,%d0\n\t"
21:21:29dan_aI was
21:21:50Slasheridan_a: that was with the scheduler.patch (the newest)?
21:21:57linuxstbSlasheri: The 3g ipods struggle even with MP3s.
21:22:24Slasherii would need to have an ipod :/
21:22:27linuxstbThey have a PP5002 (broken cache), 4g and later ipods have PP502x
21:22:54dan_aMight moving some of the code to iram help?
21:23:14dan_a(I've not worked out the rules of when that does and doesn't help yet)
21:23:41amiconnSlasheri: The patch seems to work on archos (tried on recorder only so far), but it does one thing I don't like
21:24:05linuxstbdan_a: It may help, but most codecs already use all the available IRAM - the Coldfire benefits heavily from it.
21:24:06mirakamiconn: oh no ! I remember why I added 65536. it's because I want to add 1 in the upper part of the register
21:24:08amiconnIt almost maxes out iram usage; only 28 bytes left with the patch...
21:24:28 Join Arathis [0] (
21:24:31amiconnThat will definitely brek debug builds
21:24:57dan_alinuxstb: I thought that half the iram was kept for the core and half for the codecs? Or is that something I've got confused about?
21:25:09linuxstbdan_a: It is. What code are you talking about?
21:25:38amiconnSlasheri: What does the asterisk at the left mean?
21:25:50dan_aI was talking about maybe moving parts of the scheduler code to iram
21:26:25Slasheriamiconn: ah, hmm.. i am not really sure how to fix that.. or maybe we could put some data out of the struct
21:26:37Slasheriamiconn: star indicates that the current process is running (not sleeping)
21:26:52amiconnOkay, seems to work on archos as well
21:27:03Slasherii have had in mind to separate also blocked processes but i think that might become too complex to do
21:27:07Slasherisounds nice :)
21:27:10amiconnI just started a tagcache update, and a second asterisk appeared
21:27:23Slasherihehe :)
21:27:49Slasheriwith iriver, those are flashing when there are many background threads running
21:27:56amiconnmain always shows active. Is that correct?
21:28:08Slasheriyep, because main shows also the debug menu
21:28:15Slasheriso debug menu never sees it sleeping
21:28:16amiconnAH, of course
21:28:45amiconnI don't see any other thread showing activity...
21:29:13amiconn(other than main and tagcache)
21:29:32Slasherithen it sounds correct, because all threads should sleep most of the time
21:30:07Slasherithey probably just wait in a kernel queue loop and will sleep immediately again after woken up
21:30:08amiconnBacklight thread seems to work, backlight goes on & off correcty
21:30:28amiconnThe bad thing is that I cannot see all threads
21:30:37dionoeaanyone her know what BUTTON_ON is in the sym for Hxx0 ?
21:30:38amiconnThat's not a problem introduced by your patch
21:31:42amiconnMaybe we could use the list widget for the thread list? Or does the list widget not allow dynamic entries and programmatic redraw?
21:32:45Slasherihmm, list widget sounds nice if that's possible to do
21:32:49amiconnSlasheri: Do you have a remote?
21:32:57Slasherilist would be nice for logf menu also
21:33:02amiconn(LCD remote that is)
21:33:04Slasherihmm, for iriver? yes
21:33:51amiconnThe debug menu only works on main, correct?
21:33:58 Join rconan [0] (
21:34:06Slasheriyep, it's not displayed on remote at all
21:34:12dionoeapixelma: ?
21:34:14amiconnSo you should be able to see the remote_scroll thread being active if there is a scroling line on the remote
21:34:31pixelmadionoea: yes?
21:34:32Slasheriin fact i think recent changes might have enabled the menu on remote as well
21:34:42Slasheribut debug sub menus don't work
21:34:54amiconnThe menu itself is of course displayed on the remote
21:35:02Slasherii will check that
21:35:05amiconnIt uses the standard menu, i.e. list widget
21:35:19dionoeapixelma: i'll commit a new solitaire bugfix patch in a few seconds. Could you test the shortcuts ? (they seem fine now ... don't know if my code was just all bugged when i first wrote it or if someone added some bugs afterwards)
21:35:26amiconnBut most of the sub-items only display on main
21:35:48pixelmadionoea: sure I will
21:39:06Slasheriamiconn: yep, just tried it and sub items wont work
21:39:07 Quit alberink (Read error: 104 (Connection reset by peer))
21:39:13Slasheribut menu itself is displayed
21:39:55amiconnI guess we should leave it that way, unless using one of the standard widgets does have advantages
21:40:05amiconn(like using th elist widget for thread debug)
21:40:49dionoeaSlasheri: is the threads priority patch working better now ? :)
21:40:59 Join alberink [0] (
21:42:35amiconnSlasheri: Hmm, when browsing by year, the years aren't sorted...
21:43:12dionoeai wonder how my solitaire commit triggered 2 warnings in synth.c in the Gigabeat Sim
21:44:36markundionoea: maybe it just got compiled on a build server with a gcc version which complains more..
21:44:48Learsometime odd build errors pop up, and disappear again a while later...
21:45:53 Quit Rondom (Nick collision from services.)
21:45:59 Join Rondom [0] (
21:48:07pixelmadionoea: the shortcuts work fine now :)
21:48:14dionoeanice :)
21:48:30dionoeasolitaire is now bug free :')
21:50:06TrueJournalsyay :-p
21:50:45Slasheriamiconn: weird, they should be.. probably dir buffer ran out
21:50:49 Quit rbalboa_ (Read error: 110 (Connection timed out))
21:50:58Slasheribut need to go now, cu tomorrow :)
21:51:35Slasheridionoea: hmm, probably not if you tried it last time today
21:52:04dionoeai didn't :)
21:52:06mirakamiconn: I don't know what to do, they work both separately but no together
21:52:11dionoeai only tried the first patch yesterday
22:02:24 Quit rconan ("Leaving")
22:03:08dan_aPutting add_to_list and remove_from_list in IRAM feels like it improves performance by a huge amount with Slasheri's patch.
22:05:22 Quit actionshrimp ("( :: NoNameScript 4.01 :: )")
22:06:00dan_a(on the 3g iPod, at least)
22:09:28 Join mirak_ [0] (
22:09:42 Join Angryman [0] (i=5438962f@gateway/web/cgi-irc/
22:10:49mirak_amiconn: they work independently but not together. I can't determine why
22:12:17 Quit BigMac (Read error: 110 (Connection timed out))
22:13:15Angrymancan someone how to compile a simple plug in and not all again? Try to get some more working on iRiver h10
22:15:32markunAngryman: do you want something like this? "make apps/plugins/logo.rock"
22:15:47Angrymanyes thats what i meant ^^
22:17:34 Quit mirak (Connection timed out)
22:18:59 Join LayDwnAnD [0] (
22:19:14amiconnmarkun, dionoea: The build machine causing the warning is the most advanced native-gcc-wise: (gcc 4.1.2)
22:19:27LayDwnAnDhey has there been any improvement on the video playback for the ipod video ?
22:20:14amiconnHmm, others have that too...
22:20:57Angrymanmarkun: i get an error: "no rule to make target 'apps/plugins/...". stop
22:21:29markunAngryman: are you in you build dir?
22:21:41 Join barrywardell [0] (i=892b7a6e@gateway/web/cgi-irc/
22:22:12markunno, you must make a build dir and run tools/configure in that dir
22:22:30Angryman;-( <−− STUPID ME
22:22:35markundon't worry
22:25:19 Quit lightyear (Remote closed the connection)
22:26:02 Join Nico_P [0] (
22:27:38Nico_Pwow barry nice commit
22:27:48Nico_Pjust seen it
22:29:14TrueJournalsWOW! Scroll pad support?
22:29:16TrueJournalsI have to try this!
22:29:22Nico_Pme too ;-)
22:29:25AngrymanWHAT WHERE
22:29:35TrueJournalsjust got a CVS commit
22:29:42 Join jhMikeS [0] (
22:29:46TrueJournalswaiting for 20 GB build
22:29:54dan_aWell done, barrywardell - I think you've made a few people happy!
22:30:01barrywardellit's still fairly basic, but works non the less
22:30:08 Join newbyx86 [0] (
22:30:31Angryman<−−- Clapping with his hands... happy like a child^^^
22:30:32amiconndionoea: Do you think using the grayscale lib for solitaire on archos would be worth it?
22:30:41dionoeayou could even simulate more buttons if needed i guess ... kind of like what needs to be done on the ipod
22:30:41 Join Davide-NYC [0] (
22:30:56dionoeaamiconn: depends if it has an impact on battery life
22:31:11dionoeacurrent graphics are already quite readable on archos IMO
22:31:18*dionoea fetches his recorder to try
22:31:22amiconnYes they are
22:31:52jhMikeScan you use configfile.c in the core (with a mod, for saving encoder settings)?
22:32:31dionoeaamiconn: is battery life the same when using the grayscale lib ?
22:32:39amiconnThe core has its own incarnation of config file writing
22:33:01amiconndionoea: No, it will go slightly down. More noticeable on Ondios than on recorders
22:33:14dionoeaok, then i guess that it'd be best to keep it as is
22:33:21*pixelma agrees
22:33:28Davide-NYCSorry to interrupt... Can I ask a question about the viability of different video codecs at different resolutions on different targets?
22:33:34amiconnThe grayscale lib loads the cpu with 50% as long as the overlay is running (for fullscreen)
22:33:37jhMikeSamiconn: I've been looking...I don't see...except for the stuff in config.c
22:33:46Davide-NYCEspecially the coldfire CPU in the iRiver H320.
22:34:10jhMikeSof settings.c or wherever
22:34:11dionoeai still need to do the nice animation at the end of the game btw :)
22:34:30amiconnThe jumping cards?
22:34:35dionoeayeah :)
22:34:42amiconnShouldn't be too difficult
22:34:45Davide-NYCWhat are the expected supported codecs? And what (of those codec) will play on an H320?
22:35:28Davide-NYCtheora, xvid, divx, etc, and at what bitrate?
22:35:56dionoeaamiconn: well it's just some downwards constant acceleration, a jump which basically *=-1 the vertical speed and a constant horizontal speed
22:36:04dionoeaso indeed it should be pretty easy
22:36:19amiconnDavide-NYC: We don't even manage to run mpeg-2 realtime... so forget about the mpeg-4 variants for now
22:36:27Davide-NYC[the wife want a video player, but I will try to dissuade from iPod 5G]
22:36:56Davide-NYCso is it reasonable to presume that real-time on the h320 at native res at decent framerate is not likely?
22:37:07Davide-NYCusing xvid as example
22:37:17amiconnPerhaps we'll manage, perhaps not
22:37:23amiconnIt's really hard to say
22:37:31preglowfind some good coders
22:37:31amiconnForget xvid
22:37:36Davide-NYChaha, I'm trying to make a buying decision on a gift for the wife.
22:37:44amiconnIt's more complex than mpeg-2
22:37:54amiconnxvid, divx etc are all mpeg-4 variants
22:37:57preglowamiconn: if iriver could manage xvid, so could we
22:38:04amiconnI don't know about the complexity of theora
22:38:06preglowsomeone really clever just needs to come along
22:38:14preglowamiconn: it's comparable to mpeg4, more or less
22:38:15amiconnpreglow: Yeah, xvid at 10fps, hooray!
22:38:28amiconnI'd rather want mpeg-2 at a decent framerate
22:38:46preglowbut it is doable
22:38:55Davide-NYCwhat about the iPod 5G targets. they have dedicated video proc, right?
22:38:57Nico_Pcool, now we can go up on the H10 :D
22:39:04preglowDavide-NYC: but anyway, it's impossible to say, we need more video coders for any of that to happen very soon
22:39:05TrueJournalsyes :-D
22:39:09Angrymanworks great^^
22:39:11TrueJournalsBarry... you rock :-D
22:39:12preglowDavide-NYC: yeah, which we can't use very easily at all
22:39:14amiconnWell, mirak does have the basic idct in coldfire asm, we just need to find the bugs
22:39:31amiconnyuv->rgb seems to have some room for optimisation as well, with emac
22:39:34Nico_Pbarrywardell: are you close to fixing the screen bugs ?
22:39:37TrueJournalsby "basic" up support, I suppose you mean that it's not like the original firmware where scrolling with your finger would scroll faster?
22:39:37jhMikeSneed to adapt that to gp use...bit entries won't do it
22:39:46 Quit Hansmaulwurf ("( :: NoNameScript 3.81 :: )")
22:40:01Davide-NYCpreglow: I understand, I just wanted to poll some heads in here before I start researching what to get the wife so that she can watch "her shows".
22:40:11TrueJournalsAnd... how do you quit doom?
22:40:19Davide-NYCyou don;t
22:40:20Nico_PTrueJournals: yes, there seems to be an area on the pad where it goes up and one where it goes down
22:40:27Davide-NYC(jusit quits you
22:40:30Nico_PTrueJournals: have you tried pressing power ?
22:40:32barrywardellNico_P: I have a fix for the 6/6GB screen bugs already. it's in the patch tracker
22:40:44markunDavide-NYC: does it really matter which codec we use if you have to reencode anyway?
22:40:49amiconnDavide-NYC: Targets with a smaller screen (pixel-wise) will get realtime video earlier...
22:40:50Nico_Pbarrywardell: cool, i'll try it
22:41:06TrueJournalsNico_P: pressing power takes you to the menu... you scroll to Quit game, press play and it says a message and "press select to quit" I've tried everything, and nothing seems to work :-\
22:41:06 Quit barrywardell ("CGI:IRC (EOF)")
22:41:06 Quit Rondom (Nick collision from services.)
22:41:07Davide-NYCbatch re-encoding is not a problem.
22:41:12jhMikeSpreglow: is the UDA1380 able to provide more sample rates than the coldfire or should it not be in master mode and use the cf audio clock instead?
22:41:16 Join Rondom [0] (
22:41:35amiconnjhMikeS: Afaik the UDA can't be clock master in the iriver
22:41:40AngrymanBarry looks for values greater or lower than a certain one and defines with this up or down.... so it goes up from one value and down
22:41:53amiconnBut Linus probably knows more
22:41:58Angryman<−− as far as a stupid man like me understood it ^^
22:42:08jhMikeSamiconn: The driver file sets it that way when recording is enabled
22:42:20 Join barrywardell_ [0] (i=892b7a6e@gateway/web/cgi-irc/
22:42:32jhMikeSIronically linus says they should always be in slave mode?!?!
22:43:06barrywardell_Angryman: that's pretty much it
22:43:25*amiconn already has too many loose ends in his source tree :(
22:43:30barrywardell_TrueJournals: that's exactly what I mean by basic
22:43:31Nico_Pbarrywardell_: is the battery level read correctly ?
22:43:44Davide-NYCjhMikeS: can I get your quck opinion on bitmaps vs char in the recording statusbar for displaying samplerate?
22:43:48*linuxstb has many loose ends in many source trees...
22:43:55barrywardell_Nico_P: Yes, I think so. on 20GB at least
22:43:57jhMikeSpreglow: Is it needed to detect the SPDIF sample rate if were going to set it explicitly anyway?
22:44:01amiconnMaybe I should stop here, dump all changes, then restart later
22:44:08barrywardell_it takes a few seconds before it's read though
22:44:35amiconnjhMikeS: You can't set s/pdif sample rate, you have to use the measured one
22:44:41Angrymanbut anyway you can define it via system (5g 820mAh or so)
22:44:45Angrymanor not?
22:44:55amiconnYou can only record what's delivered, not magically resample...
22:45:15Nico_PAngryman: i meant the current power level of the battery, not its capacity
22:45:30Angrymanbut for that you need the capacity ^^
22:45:32TrueJournalsso no one knows how to quit Doom?
22:45:37Angrymanneed to know
22:45:45linuxstbTrueJournals: Not on the H10....
22:45:54Nico_Pthe capacity is only needed to know how time you have left
22:45:59TrueJournalsthe capacity of the 20 GB H10 battery is 1550 mAh...
22:46:06Angrymanquit doom<−−- Battery out ^^
22:46:11linuxstbTrueJournals: It's probably just a missing key mapping - so someone will need to fix the source.
22:46:25 Quit LayDwnAnD (Connection timed out)
22:46:29Nico_Plinuxstb: isn't what i just said above true for all targets ?
22:46:31TrueJournalsAngryman: I have a 20 GB :-p
22:46:48Nico_Pwhat's the capacity for the 5 GB ?
22:46:49Angrymanok^^ so i dunno
22:47:03 Quit Angryman ("CGI:IRC (EOF)")
22:47:10barrywardell_TrueJournals: it's probably a bug in the button settings
22:47:14linuxstbNico_P: You've said lots of things...
22:47:27TrueJournalsbarrywardell_: Guess I'll just have to reset :-p
22:47:40 Join aegray_ [0] (n=aegray@
22:47:50 Join Angry [0] (
22:47:53Nico_Pi was referring to what i said just before asking : "the capacity is only needed to know how time you have left"
22:47:54barrywardell_Nico_P: 820mAh
22:48:01linuxstbNico_P: Yes, that's true.
22:48:05TrueJournalslol... tjere
22:48:16TrueJournals*there's one place on the scroll pad where it reads both up and down :-p
22:48:20Nico_Pbarrywardell_: thanks
22:48:48 Join Kohlrabi [0] (
22:49:00Mode"#rockbox +o Bagder " by ChanServ (ChanServ@services.)
22:49:06Nico_Plinuxstb: ok, what you said ("Not on the H10....") made me think maybe it wasn't the same for all targets
22:49:13Topic"" by Bagder (n=daniel@rockbox/developer/bagder)
22:50:54jhMikeSamiconn: ok I don't's set to use the audioclk/4 (44100) then what it really is has to be measured since it won't be that?
22:51:37amiconnThe UDA isn't part of the recording data path when recording from s/pdif
22:52:17amiconnIt's only used for monitoring - and it's impossible to sync the coldfire audioclk output to the s/pdif input signal
22:53:11amiconnSo the UDA will play a ghard-resampled version (no interpolation at all) of the input signal
22:53:22 Quit barrywardell_ ("CGI:IRC (Ping timeout)")
22:53:29Bagderit would be neat if some H10 guys could have a look at adjusting the manual for the H10
22:53:32amiconnStill better than no monitoring at all
22:53:49mirak_amiconn: I manage to make work the two idct in conjonction but there are some glitches
22:55:12jhMikeShere's the comment for SPDIF setup: /* SCLK2 = Audioclk/4 (can't use EBUin clock), TXSRC = EBU1rcv, 64 bclk/wclk */
22:55:12 Quit solarflare (Read error: 104 (Connection reset by peer))
22:55:26 Join solarflare [0] (
22:55:48jhMikeSSo where's the clock coming from?
22:56:02AngryBagder... i will have a look at it
22:56:04amiconnThe UDA clock is coming from the coldfire, as the comment says
22:56:16amiconnBut that doesn't have to do with the recording sample rate
22:56:24Angryat least for the parts i can contribute ^^
22:56:31BagderAngry: great! tucoz and bluebrother are two of the masterminds behind it
22:56:33dionoeaw00t ! jumping cards in solitaire
22:56:39amiconnThe s/pdif input (EBUin in motorola language) is part of the coldfire itself
22:57:19BagderAngry: it should mostly be a matter of describing H10-specific things as the generic things are mostly covered already
22:57:37jhMikeSIsn't the UDA clock from the codec in mm?
22:57:48Angrymaybe you could help me once with compiling a plug in... cant get it out of cygwin antd it seems that there have just to be some lcd size settings adjusted and with my minor programming skills i could solve such issues
22:58:19 Quit Rudy4Pez (Connection timed out)
22:58:23Angrywhere can i find it... in ../ttols/configure ... build manual ???
22:58:28amiconnAgain: the UDA has _nothing_ to do with the recording when recording from s/pdif
22:58:33jhMikeSso the SCLKII follow IIS1
22:58:47BagderAngry: correct
22:58:59amiconnjhMikeS: Not in s/pdif mode (!!!!!!!!!!!!!!!!!!!!!!!)
22:59:16amiconnSCLK2 is free-running, derived from audioclk
22:59:32jhMikeSamiconn: don't get do flustered! :)
22:59:34amiconnTXSRC is passed through from EBU1rcv
22:59:59amiconnThat's the cause of the harsh resampling if the input signal isn't 44.1kHz
23:00:01TrueJournalsHooray being able to turn volume up!
23:00:44bluebrotherAngry, want to help out with the manual?
23:01:23jhMikeSsection 17 some more?
23:01:46bluebrothernice :)
23:01:59*jhMikeS is gonna get a section 8 :)
23:02:23 Part ep0ch_ ("Kopete 0.11.1 :")
23:02:29 Quit Lear ("Chatzilla 0.9.74 [Firefox]")
23:03:16 Quit EV|Server (Read error: 110 (Connection timed out))
23:03:17 Join Lune [0] (i=ce3fe6b5@gateway/web/cgi-irc/
23:04:31Angrycan anyone tell me how to compile brickmania.c to brickmania.rock (saw it now also in the manual^^) and it isnt in the rockbox 5GB build... Needs just some fixes in the LCD-SIZE SETTINGS TO GET IT WORKING
23:04:58BagderAngry: edit SOURCES
23:05:00Bagderin that dir
23:05:19Bagderor possibly edit the source
23:05:48 Quit Kohlriba (Read error: 110 (Connection timed out))
23:06:08Bagderits 128 pixels width isn't it?
23:06:15 Join Paul_The_Nerd [0] (
23:06:37***Saving seen data "./dancer.seen"
23:06:50Angry128x128 5gb
23:07:11bluebrotherdamn. Something is different for on-boot loading wps between the sim and my device.
23:07:15BagderAngry: so see line 41 in apps/plugins/SOURCES
23:07:21Angryah found the source file
23:08:04Angryand its already commented that it needs to be fixed... right place to start ^^ Thx Bagder
23:08:37Bagderthe comment is of course about the ifp7xx, but it has the same LCD width
23:09:13Angry#if LCD_WIDTH != 128
23:09:53Luneis it ok to ask here if anyone could please fix page-up/down (on+down/up) functionality for V1 Recorder ?
23:12:28linuxstbAngry: Someone needs to create smaller colour bitmaps for Brickmania to work on a 128x128 LCD.
23:12:53 Quit rretzbach (Read error: 104 (Connection reset by peer))
23:14:42Angrybluebrother: how shall and can i make suggestions for changes in the manual??
23:15:16bluebrotherAngry, if you can build the manual yourself you could simply generate patches and submit them to the tracker.
23:15:25 Join sharpe [0] (
23:15:56bluebrotherif not, you can still collect your suggestions (sorted by themes if possible) and submit that to the tracker.
23:15:56Angrynever generated a patch before can a tutorial found somewhere on
23:16:01 Join EV|Toshiba [0] (
23:16:02amiconnBagder: Do you know why the plugin bitmaps (native and mono, remote & main) are tried to build 3 times?
23:16:07EV|Toshibabillytwowilly ?
23:16:11bluebrotherdo you have a cvs tree checked out?
23:16:16Bagderamiconn: nope
23:16:35bluebrotherthen it's simply a "cvs diff -u <filename(s)>"
23:16:37 Quit _Veseliq_ (Read error: 113 (No route to host))
23:17:06bluebrotherfor diffing on the complete manual simply use "manual" as filename (assuming your toplevel folder has the manual folder in it)
23:17:07amiconnThis is ugly. It slows down cygwin even more... (other build environments too of course)
23:17:36bluebrotheroh, forgot: cvs diff -u prints it output to stdout, so just redirect it into a file
23:17:52bluebrothercvs diff -u manual > mychangestothemanual.patch
23:18:00Angryuff.. now i gotta make a copy and paste ^^
23:18:17bluebrotherheheh ;-)
23:18:23EV|ToshibaHmm... billy had me install some Kubuntu packages yesterday, and I cant get the machine to connect to the internet..
23:18:29 Quit JoeBorn ("Leaving")
23:18:34bluebrotheralso I'd recommend you read the LatexGuidelines and ManualTodo wiki pages.
23:18:43Angrywill do
23:19:01TrueJournalsDoes rockboy play gameboy and gameboy color roms?
23:19:09bluebrotherwe have quite a bunch of macros for making working on the manual smoother, but you need to know them ;-)
23:19:56linuxstbTrueJournals: Yes.
23:20:16TrueJournalsthat's what I though :-p
23:20:17Angryi already saw them
23:20:49Angry\begin{enumerate} etc
23:20:49preglowTrueJournals: both, afaik
23:20:54preglowTrueJournals: both the colour code has been ripped and
23:20:57preglowripped _out_
23:21:26bluebrotherAngry, I assume you already know LaTeX?
23:21:37amiconnAfaik the colour code was reactivated
23:21:39linuxstbpreglow: I thought it was put back in.
23:22:05preglowlinuxstb: in which case i'm mistaken
23:22:15sharpeyay tinnitus...
23:22:31Angryno... but i thought thats what you meant with makros
23:23:13bluebrotherpartly. LaTeX is (simplified) a bunch of macros for TeX. enumerate for example is part of LaTeX, nothing we added.
23:23:26 Quit Davide-NYC ("Chatzilla 0.9.75 [Firefox]")
23:23:48Angryis there a editor available
23:23:58bluebrotheras a quick start I can recommend the l2short.pdf introduction (pretty easy to find with google, but it should be linked on the wiki)
23:24:19jhMikeSThe IEC958 clock is the SPDIF clock?
23:24:26bluebrotherjust use a text editor of your choice. Syntax highlighting helps.
23:24:34bluebrotherI mostly use kate
23:24:44TrueJournalsoh... no rockboy on H10?
23:25:02 Join Febs [0] (
23:25:06BagderTrueJournals: make it
23:25:32TrueJournalsif I knew anything about programming...
23:25:59TrueJournalsBagder: any idea what would need to be done?
23:26:06amiconnpreglow: Do you know what happens in the mac instruction with parallel load when the destination register of the load is equal to one of the operand registers?
23:26:08bluebrotherapart from that you should have a look that your text editor saves file in unix format (in case you're working on windows) and keeps trailing spaces.
23:26:36linuxstbTrueJournals: The display code in Rockboy doesn't work for colour LCDs smaller than a real Gameboy's LCD - so that includes the H10, ipod Nano and iaudio X5.
23:26:42Angryutf-8 isnt working??
23:27:04TrueJournalslinuxstb: so somebody would need to figure out how to code it so that it scales down the game?
23:27:23linuxstbTrueJournals: Yes. Scaling already happens for greyscale LCDs, but it's not in the colour display code.
23:27:24amiconnTrueJournals: Yes. It's already there for some grayscale targets
23:27:26*jhMikeS is just trying to get some help here cause he can't experiment on an H120 himself :)
23:28:09bluebrotherAngry, the latex sources are still latin1.
23:28:52bluebrotherbut on most files this won't make a difference, there are only 1 or 2 using extended characters.
23:29:09preglowamiconn: i think it works as you think it would
23:29:26AngryI thnink i will download the TexnicCenter Tool linked on the wiki... looks like this is what i need^^
23:29:36preglowamiconn: the values in the multiplier arguments are latched into the multiplier circuit at the start of the first cycle, so it probably works out just fine
23:29:36 Join barrywardell [0] (n=barrywar@
23:29:42 Part Paul_The_Nerd
23:29:44preglowamiconn: as a matter of fact, i think i've done this several times
23:29:50amiconnThat's what I think...
23:30:11*amiconn is looking for ways to optimise the idct thing
23:30:53preglowamiconn: yeah, i do that all the time, it works just fine
23:31:08*jhMikeS is looking for ways of setting sample rates for recording sources on SPDIF
23:31:20 Part Angry
23:31:58preglowi just deleted my 64 bit coldfire lpc routine :///
23:32:06Bagderdescriptive URL even
23:35:44 Join telliott_ [0] (
23:37:12amiconnjhMikeS: You cannot set the sample rate for s/pdif
23:37:37amiconnThe source is digital, and so its clock is fixed
23:37:37 Join tvelocity [0] (
23:37:48amiconnYou have to measure and adapt
23:38:10telliott_Is there a known problem with recent daily builds for Archos V1?
23:38:31amiconnThe measuring is in fact only necessary n order to write the correct value to the wav header
23:38:37telliott_I'm having strange button issues.
23:39:23 Join jborn__ [0] (
23:39:29jhMikeSamcionn: But the codecs don't even adapt to it...they'll all think 44100 anyway
23:39:40preglowthen they need to adapt
23:39:44 Quit jborn__ (Read error: 104 (Connection reset by peer))
23:39:47jhMikeSworking on it.
23:39:57preglowand if an exact sample rate match can't be found, god knows what to do
23:40:00preglowprobably resample
23:40:03jhMikeSWhat is the IEC958 clock?
23:40:06preglowthe spdif clock
23:40:12preglowiec958 == spdif
23:40:22amiconnor aes/ebu
23:40:41 Quit Lune ("CGI:IRC (EOF)")
23:40:46jhMikeSexactly...and that can follow audioclk/(3,4,6,8,12,16), or sclk(1-4)
23:40:49amiconnaes/ebu == professional mode, spdif == consumer mode
23:40:58amiconnYes, for output
23:41:12 Join Lune [0] (i=ce3fe6b5@gateway/web/cgi-irc/
23:41:25amiconnThe input clock is following the input signal of course, and can be measured, but not regenerated for reuse
23:42:20jhMikeSresampling in the context is a biotch because codecs (most importantly mp3) expect a certain number of samples per chunk
23:42:33preglowwhy does that make it a bitch?
23:42:49 Quit Arathis (Read error: 104 (Connection reset by peer))
23:43:37jhMikeSyou have to wait for enough raw pcm to be able to downsample exactly to the # of samples it expects...hope it doesn't cause buffer problems
23:43:53amiconnForget resampling, imho
23:43:57 Part TrueJournals
23:44:06amiconnSupporting the standard sample rates should be enough SPDIF just follows the xmitter
23:44:17preglowwell, at least for now
23:44:20amiconnNon-standard rates aren't supported by iec958 either
23:44:29preglowso what to do if the sample rate doesn't match up?
23:44:33amiconn(but they work)
23:44:41jhMikeSpreglow: chipmunks I guess
23:44:45EV|ToshibaWhere the **** is billytwowilly
23:45:03*amiconn has a very useful tool for testing s/pdif in
23:45:13 Quit gromit` ("Coyote finally caught me")
23:45:18amiconnThe archos recorder can be used as an adjustable s/pdif source
23:45:50 Join x1jmp [0] (
23:45:53amiconnIt can generate everything from 8kHz to 96kHz sample rate (even though it's only specced up to 48kHz)
23:47:19jhMikeSI would support say 32000 by downsampling 44100 or 16000 by downsampling 22050....never UP if possible
23:47:46amiconn32000 is a standard rate
23:48:44jhMikeSBut not creatable by the CF or TLV320 on x5. I don't know if the UDA can be set for those
23:48:59amiconnThe UDA can only do 44.1 for recording
23:49:11 Join jborn__ [0] (
23:49:12amiconnBut s/pdif can deliver any standard rate
23:49:33preglowamiconn: it can do 11025 and 22050 as well, afaik
23:49:39amiconnNot for recording
23:49:43jhMikeSCan't UDA use slave mode for record? Linus thought it should. Which allows 11,22,and 44
23:49:56 Nick jborn__ is now known as JoeBorn (
23:49:58preglowi think it can
23:50:16jhMikeSCould use those funky 3,6,12 clock devides too ? :)
23:51:00jhMikeSEverthing is dsp.c works against 44100 though...not good if I want to support other stuff.
23:51:59preglowshould be pretty quick to fix
23:52:05jhMikeSpreglow: would FIR decimators work well for this app? or too slow? Where's a good design app for that?
23:52:24preglowdepends how good you want them
23:52:32preglowi'd rather use an iir decimator
23:52:38preglowsharper slope at a lower cost
23:52:52x1jmpI have a plugin which works in the simulator, but freezes not on the player
23:52:56***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
23:52:58x1jmphow can I debug it?
23:53:33***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
23:53:52x1jmpoh.. I mean it freezes on the player
23:53:53amiconnWith Lear's change, gcc seems to use a form of div CFPRM doesn't mention...
23:53:58jhMikeSI thought the FIR was preferred for stability (no feedback), less math. I have to read some more about it.
23:54:13 Join Arathis [0] (
23:55:06amiconnIt seems the coldfire allows a 3-register form of div
23:56:58jhMikeSamiconn: ?? you mean quot/rem?
23:57:15newbyx86I love microsoft and their shortcomings.
23:57:20newbyx86 Did you mean bad_pool_heater?
23:57:27amiconnYeah, seems like gnu asm writes the instruction slightly different than motorola
23:57:36telliott_I need to get to NYC to visit Reace.
23:58:30 Part telliott_

Previous day | Next day