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 2003-12-29

00:00:03diddystar5(or atleast one needle!)
00:00:17***Saving seen data "./dancer.seen"
00:00:25diddystar5i just set up a table of values
00:00:28Dr_Slumpbut the must would be to have real NFO viewer ( and reak cue file viewer ( :)
00:00:35Dr_Slumpbut i am dreaming then ;)
00:01:05BC|codelet your dreams become a reality :)
00:02:15BC|codeIt is the dreamers who create the future
00:02:58diddystar5the logarithmic scale dosent look as good for realt loud music
00:03:12diddystar5i guess i set a button to switch betwene linear and log
00:03:22BC|codeI hope so
00:03:52BC|codeI wanna BPM counter from you next :)
00:03:54diddystar5i need to figure out a way to make it delay, it jumps to much
00:04:08BC|codesample less often?
00:04:20diddystar5well then its not so smooth
00:04:34diddystar5i an go from pixle 0 to 30 in an instant
00:04:42diddystar5and thats not smooth
00:05:15BC|codetake a sample once every second (second is just and easy example)
00:05:35BC|codethen remember THIS sample and LAST sample values
00:05:59BC|codethen 10 times during the second ...move from one to the other
00:06:02diddystar5im not 100% sure it was the best forula to get the value for log, by my formula was log(x)*30.09, where x would be the original x position
00:06:23BC|codenot got a log button on your calculator?
00:06:30diddystar5i do
00:06:38diddystar5i have ti 83+ :)
00:07:05BC|codei've got a casio fx-451 ...and original with the giant solar panel
00:07:06diddystar5i acctually wrote a program on it to figure up all 56 calculations so it was easy to do
00:07:18BC|codebest choice
00:07:30BC|codeprogram on the ti83 of course :)
00:07:33diddystar5i dont know anything about the casio calcs
00:07:38diddystar5of cource
00:07:56diddystar5that basic built in language is very usefull
00:08:21BC|codestill hunting for the ultimate programmers calculator
00:08:31BC|codeso what about my decay idea?
00:08:40diddystar5i found a program someone write on that solves quadratics, you just write down what it says, it shows all the work, it made my homework very easy :)
00:08:48diddystar5the decay idea may work
00:08:59diddystar5ill try it when i get the log scale all nice
00:09:14BC|codeall in a big array?
00:09:55diddystar5just enter in the AB and C and write down what it says
00:10:06diddystar5cant get much easier than that
00:10:24BC|codeyou can take that into an exam?
00:10:39diddystar5i used it on the state tests
00:11:17diddystar5well see ya BC|code
00:11:20diddystar5takl to you later
00:11:33 Quit diddystar5 ("Leaving")
00:18:21 Join Electronic [0] (
00:36:04 Quit Electronic ("Leaving")
00:48:53 Join LinusN [200] (
00:52:55 Join c0utta [0] (
00:53:22c0uttau there linus ?
00:54:20 Nick edx{visited} is now known as edx{code} (
00:55:59BC|codec0utta: he signed on 4mins before you he is likely to be around soon :)
00:56:10LinusNi'm here
00:56:25c0uttano worries. i've isolated the source of rockbox hanging.
00:56:48 Quit Dr_Slump ()
00:57:13c0uttathey are a specific set of files that have worked previously on Rockbox
00:57:36c0utta192kbps - nothing special about them. You try to play them and they hang. Work OK on WinAmp
00:58:13LinusNhow big are the files? (i have only a v.90 modem right now)
01:00:09LinusNcan you put it up for download somewhere?
01:00:48c0uttai'll try. i think i have some webspace somewhere. leave it with me
01:00:50BC|codedoes the file still exhibit the same behaviour if you cut some of it out of the middle?
01:01:27LinusNif it is the id3 tag, i would only need the beginning of the file
01:01:43c0uttawhat's the best way to cut it ?
01:02:09LinusN(linux program)
01:02:21c0uttausing xp
01:04:02BC|codedownload a dos compile of dd?
01:08:00BC|codebookmarked :)
01:09:33BC|codewonder why he didn't use DJ to compile - then he would have NO file dependancies !?
01:13:36c0uttawhich of the unxutils does the cutting ?
01:14:54LinusNdd if=original.mp3 of=newfile.mp3 bs=1024 count=128
01:15:15LinusNwill copy the first 128kbytes from the file to newfile.mp3
01:20:17c0uttabc: FYI - doesn't work
01:20:40BC|codec0utta: okay - cheers, sorry for the misdirection
01:20:58c0uttano problem.
01:22:35c0uttai'm on 56K so i'll be about 4 hours downloading the unix utils
01:23:15BC|codehang on dude - i'll get it on cable and send over the one you need.
01:24:28BC|codecan't see the d/l link...
01:24:57BC|codegot it
01:27:03 Quit AciD (Read error: 54 (Connection reset by peer))
01:28:32BC|codec0utta ...comin at ya'
01:32:09 Join AciD [0] (
01:33:25c0uttasorry BC - i was parenting
01:33:37c0uttawhat type of file are you sending ? dcc ignored the type
01:33:49BC|codeunderstood ...we stumbled over the term NMI the other night ;)
01:33:52LinusN.exe i guess
01:33:57BC|codesending dd to save you time
01:35:06c0uttaok BC - try again
01:35:41BC|codethat's from the zip from the sourceforge link from linus
01:36:17c0uttagot it. i'll still download the unix utils tho.. :)
01:36:28BC|codefer sure
01:36:40c0uttaLinusN: i also understand your issue about recursive keypresses
01:37:33c0uttai made a change to the F3 screen last night to load the playlist viewer
01:37:47c0uttacaused all sorts of problems - had to turn off and on again.
01:39:42c0uttai have worked out the problem too..i think
01:39:57 Join Electronic [0] (
01:40:17 Quit Electronic (Client Quit)
01:40:21c0uttarockbox is designed to be very modular in that you can press almost any key at any time and it will perform that function
01:43:20c0uttaok linus - i've dd'd
01:43:53LinusNc0utta: nice
01:44:06LinusNwhere can i get the file?
01:48:17BC|codeLinus, do you know a way to make diff "ignore TRAILING spaces"?
01:48:51LinusN"trailing spaces"?
01:49:25BC|codemy editor strips trailing spaces and there are load of lines of code in Rockbox which flag up as "changed"
01:49:51BC|codeI hav a number of fiddly work arounds
01:50:25LinusNregular gnu diff has a -b flag
01:50:39LinusN-b means "ignore whitespace changes"
01:50:45BC|codeI need "trailling" spaces only
01:51:08BC|codeignore "start" spaces means you cannot correct indenting
01:51:58c0uttaLinusN: i can't dcc it to you ?
01:54:41 Quit AciD (Read error: 60 (Operation timed out))
01:55:48 Quit methangas (" Try HydraIRC -> <-")
01:57:35c0uttaLinusN: incoming...
02:00:19***Saving seen data "./dancer.seen"
02:02:05BC|codeLinus, is there a font editor/creator fro windows?
02:02:49LinusNthere are lots, but most of them are expensive
02:03:00BC|codei want to generate a font for rockbox
02:04:06LinusNc0utta: got it, analyzing
02:04:07c0uttaLinusN: did you get the file ?
02:05:45LinusNc0utta: got it, analyzing
02:11:18 Quit uski ("Fermeture du client")
02:13:24 Join midknight2k3 [0] (
02:13:49midknight2k3hi all
02:14:30midknight2k3hi linus
02:20:35 Join kurzhaarrocker [0] (
02:21:01kurzhaarrockerLinusN: you seem to be an owl.
02:23:14LinusNor vampire
02:23:35BC|codeno Shrek, moves by day ;)
02:24:49BC|codeKurz - do you have any idea how to make a rockbox font using windows?
02:25:03kurzhaarrockerI never bothered with that
02:25:12BC|codeelse I gotta dupe a whole big fat bunch of code to deal with my own format
02:25:31 Join Gazel [0] (
02:25:35BC|codeok thanx
02:30:53kurzhaarrockerJust for my interest: Is it right that the mas can provide timing information - but only while recording?
02:31:30LinusNit can include a time stamp in the frames
02:32:00kurzhaarrockerSo - theoretically they could be kept persistent?
02:32:27kurzhaarrockerstored in the file and read back while playing.
02:33:17LinusNnot very useful if you edit the file, though
02:33:41kurzhaarrockerhm. And it would only concern files recorded with the mas I assume.
02:34:02kurzhaarrocker-> pretty much effort for little effect to use these infos
02:34:22LinusNpretty useless, if you ask me
02:40:26BC|codeLinus, is it a simple task to retarget the .rocks to ./$DEVICE/.rockbox/rocks
02:40:41BC|code...for sim
02:41:18LinusNdon't know for windows, but it should be fairly simple
02:41:52BC|codeits more of a makefile problem than a windows problem
02:43:02BC|codeif I do everythng else for the sim, would you help with that bit?
02:44:37LinusNi'll try
02:44:49LinusNc0utta: i have found the cause of the hang
02:44:57BC|codeok thanks, i'll let you know when I get there
02:47:20*kurzhaarrocker is reading tech info
02:47:44kurzhaarrockerwho's dma controller is it that transfers the data? Of the MAS or the cpu?
02:48:50LinusNthere is no dma
02:49:00LinusNwhen recording
02:49:19kurzhaarrockerI was still reading about playback
02:49:36LinusNthe cpu sends serial data to the mas
02:50:01LinusNthe serial controller on the cpu can fetch the data to send with dma
02:50:03 Quit k4waii (Read error: 54 (Connection reset by peer))
02:50:24LinusNrockbox uses the dma capability for this
02:51:06kurzhaarrockerfor playback and recording?
02:52:55LinusNrecording uses the 8-bit parallel connection to the MAS
02:53:39LinusNhowever, we can't use DMA for this, so the CPU has to handle it "manually"
02:54:07LinusNpretty tricky
02:55:11kurzhaarrockerAnd I still have no clue where I could start to understand the beast. There are so many interrupts I have no clue about, what they might be for.
02:55:27LinusNkurzhaarrocker: what are you trying to do?
02:56:13kurzhaarrockerFirst I'd like to understand a _little_ of what happens. Long time goal is mp3 frame precise looping.
02:56:47LinusNfor playback:
02:57:05LinusNthe MAS has a DEMAND signal that is active high
02:57:07kurzhaarrockerlooping recordings is a job for musicians
02:58:14kurzhaarrockerIs that what feeds (on negative edges) the DEI3?
02:58:24LinusNthe DEMAND signal is connected to two pins, PA15 and PB14, PA15 is inverted
02:59:27LinusNso, PA15 can be configured as IRQ3
02:59:33BC|codei have (a=b=c=d=0xff) and I get "warning operation on rr may be undefined" ...any ideas?
02:59:40LinusNand PB14 as IRQ6
03:00:02BC|code "warning operation on d may be undefined"
03:00:08kurzhaarrockerSo we have irqs for both: mas needs data and mas is fed up?
03:00:08LinusNBC|code: show me the entire expression
03:00:24LinusNyes (on recorder only)
03:00:24BC|codell = rr = rl = rr = 0xC000;
03:00:39LinusNBC|code: you have rr twice
03:01:27LinusNso, IRQ3 enables the serial DMA and IRQ6 disables it
03:02:09c0uttaLinusN: parenting. what's the cause of the hang ?
03:02:30LinusNthe id3 tag is bad, basically
03:02:56LinusNolder versions of rockbox handled that more or less by mistake
03:03:35kurzhaarrockerDo those dma chunks _have_ to be 64 kb or may they be smaller?
03:03:37LinusNkurzhaarrocker: the DEI3 interrupt is triggered by the DMA controller when the DMA block has been sent
03:03:46LinusNkurzhaarrocker: they are 8k iirc
03:04:05c0uttayou were correct then - id3 tag was the issue
03:04:29LinusNi don't know how to fix it yet
03:04:50c0uttaand no-one else reported it ?
03:05:46GazelHi all geniuses ! I just got my FMR back from repair (warranty), all fine, and so i come back to rockbox to see what's new :) Can u briefly tell me what is currently incorporated in the daily builds that is not present in 2.1 ? i know about playlist viewer and video player... anything else ?
03:05:49LinusNc0utta: the frame size field of the TYER tag is wrong
03:05:51kurzhaarrockerSo there's a short period of undefined length between DMA is complete (DEI) and Mas wants food (irq3)?
03:06:22LinusNGazel: LCD upside down flip, bug fixes, new rocks
03:06:35kurzhaarrockerdon't forget the new clock plugin :)
03:06:37LinusNkurzhaarrocker: yes
03:06:41Gazel(sorry to appear at that moment :o) thanx !
03:06:44LinusNkurzhaarrocker: "new rocks"
03:06:55kurzhaarrockererm. yes.
03:07:04LinusNthe clock is cool
03:07:22Gazelwhat does it do ? show time ?
03:07:23LinusN(especially the digital one :-)
03:07:29LinusNGazel: yes
03:07:37kurzhaarrockerBut you can have the help (f1) only once. And I don't know how to change the date format...
03:07:55LinusNthe date format is fixed, methinks
03:08:01kurzhaarrockerme too
03:08:47LinusNthe dma block size was lowered to 8k some time ago, to increase the resolution of the resume info
03:08:58midknight2k3oh my clock?
03:09:23kurzhaarrockerWas my assumption right that the dma chunk always has to be the complete size - be it 8 or 64 kb?
03:09:40 Quit Gazel ("Leaving")
03:09:40midknight2k3Gazel: my clock
03:09:44midknight2k3too late.
03:09:53LinusNit will be shorter in two cases:
03:10:03midknight2k3i should make IRC beep everytime someone says "clock"
03:10:10LinusNthree cases:
03:10:17LinusN1) end-of-song
03:10:22LinusN2) track change
03:10:28LinusN3) buffer wrap
03:10:55 Quit edx{code} (Read error: 110 (Connection timed out))
03:11:18kurzhaarrocker? buffer wrap? Sounds delicious. New kebab. What's that?
03:11:34LinusNthe mp3 ram buffer is circular
03:17:43kurzhaarrockerIn the source comments I repeatedly read interrupt on level 5. Does the sh1 have priority handling for interrupts?
03:17:56LinusN15 levels
03:18:15kurzhaarrockerHave there been problems with them?
03:18:26LinusNi believe so, yes
03:18:42LinusNwe had to adjust a few levels once, iirc
03:19:04kurzhaarrockerSounds like a hell to debug.
03:25:28LinusNgotta sleep
03:25:36LinusNkurzhaarrocker: read docs/TECH
03:27:29LinusNnite all
03:27:31 Part LinusN
03:28:05*kurzhaarrocker is reading exactly that :)
03:34:20 Join earHurts [0] (
03:34:48 Part scott666
03:46:46 Join AciD [0] (
03:52:33 Quit kurzhaarrocker (Read error: 110 (Connection timed out))
03:56:15 Join Nomad [0] (
03:56:18Nomad'lo all
03:56:21Nomadhey, any chance of getting ext3 support into rockbox?
03:57:25BC|codecan you code in C?
03:58:26NomadI wish
03:58:33Nomadthats why I "ask" instead of tell :>
03:58:49Nomadalso why I ask if it's possible
03:59:03BC|code:( in which case I can only suggest that you put in a feature request and see if it gets any interest
03:59:30Nomadit'd would be great if we could use ext3 so I could rsync from my linux filesyerver instead of having to copy the whole frigging archive every time
03:59:39Nomadfigured I'd ask if it's possible before bothering anyone
03:59:52BC|codethat's fair :)
04:00:21***Saving seen data "./dancer.seen"
04:01:20Nomadfeature request is down it seems
04:01:32NomadI'll try again in a bit
04:01:43BC|codesorry - coding in hex
04:02:48 Quit Nomad (Remote closed the connection)
04:03:21 Join Nomad [0] (
04:03:33Nomadthat was neat, irc client hung up
04:09:05Nomadbounced it to the mailing list
04:44:38BC|codeNomad: just to advise, your email was posted as an attachment and will likely be ignored by many
04:57:09 Quit earHurts (Read error: 110 (Connection timed out))
05:19:47 Join SirZoltan [0] (
05:19:53SirZoltanhello :)
05:21:20 Part SirZoltan
05:26:57BC|codehow are you gonna use the toilet now?
05:27:22BC|codei dread to think - LOL
05:28:03BC|codeI have "rtoflmfao" reservered for VERY special occasions :)
05:29:28BC|codemy code looks like that at the moment
05:32:37 Join rhak [0] (
05:38:02 Quit rhak ("[BX] Beavis & Butthead use BitchX. Huh-huh huh-huh huh-huh he said "Bitch".")
05:40:23 Join midk2k3 [0] (
05:42:27 Join earHurts [0] (
05:44:06earHurtswhat's up?
05:44:16midk2k3not much
05:44:19midk2k3coverting vids
05:44:27earHurts(btw I'm typing on my pda)
05:46:35earHurtsand playing gnuz
05:47:06 Join arspy87 [0] (
05:47:39arspy87hey all, i actually have to go in 5 minutes or so, then i'll be back in 30
05:47:57BC|codears dude -yo :)
05:48:12arspy87what's shakin' bc? :)
05:48:40BC|coderattling some audio code - but the user interface is giving me a headache
05:48:48arspy87midk... midknight... interesting, we have a clone here
05:48:55midk2k3we do
05:49:13earHurtsso skip the ui and make an sdk
05:49:35earHurtsor an api
05:49:43arspy87i get headaches and i dont even have to be coding - trying to rip with RealOne player is enough
05:49:56earHurtsrealone sucks
05:50:02earHurtsget a real ripper
05:50:21arspy87i know :) i moved on to musicmatch! lol
05:50:33arspy87WMP was a very long time ago
05:50:46earHurtsmusicmatch is ugly as a 50 year old hooker,s ass
05:50:57earHurtsget cdex or eac
05:52:00arspy87yeah, probably going to get EAC
05:52:20BC|codeEAC is arsey to configure, but worth the effort
05:52:46arspy87why must we pay for winamp5 to rip *rolls eyes*
05:53:14BC|codethink I'll stick with 2.8
05:53:37BC|codedont see the point in 5 - with the classic skin - it is IDENTICAL to 2.8
05:53:51midk2k3is not
05:53:59arspy87every winamp after 2.8 is bloated and leaves a big memory footprint
05:54:07midk2k3sure. sure.
05:55:18arspy87well i'm out for 30, back soon
05:55:18 Quit AciD ("")
05:55:22midk2k3bye as! come back soon!
05:55:29arspy87will do :)
05:55:34 Join scott666 [0] (
05:55:36midk2k3you already told me that
05:55:38midk2k3hi scott
05:55:40 Part scott666
05:56:09arspy87BC - try not to bang your head on the table too much with that audio coding, it will only make your headache worse :)
05:56:37 Part arspy87
05:56:55earHurtswinamp 5 is free; the ripper part is the only part charged for
05:57:07midk2k3hi earHurts
05:57:13earHurtser, the burne
05:57:29earHurtsyou can get free burners that are better
05:57:55earHurtswhatever. I don't have a cd-w
05:58:09midk2k3fine then
05:58:10midk2k3ignore me
05:58:12 Quit midknight2k3 (Connection timed out)
05:58:27 Nick midk2k3 is now known as midknight2k3 (
05:59:19earHurtsmidk2k3 is often known as sdd
05:59:33 Nick midknight2k3 is now known as sdd (
05:59:34sddgood idea
05:59:43sddthats a great idea
05:59:47sddthat oughta be my new nick!
06:00:24***Saving seen data "./dancer.seen"
06:14:48 Join Stevenm [0] (
06:17:37 Join arspy87 [0] (
06:17:45arspy87back in the house
06:17:53arspy87i mean, the 'box
06:18:17BC|codebox is a british euphemism, I'd stick with house if I were you
06:18:46arspy87ah ok, what's it for?
06:19:10BC|codeaeroplane blonde ...blonde on top box below
06:19:27arspy87ah, got it ;)
06:19:33arspy87*writes note to self*
06:20:11arspy87we all know midk = ssd right? :)
06:21:34sddjust casue you didnt see me change nick doesnt mean noone else did
06:23:20 Quit earHurts (Read error: 60 (Operation timed out))
06:24:42 Nick sdd is now known as MidK|GTA3 (
06:25:48arspy87grand theft auto 3, midk?
06:35:22arspy87got it?
06:35:27arspy87sorry iw as making a playlist
06:36:59MidK|GTA3i beat the hard part
06:38:24MidK|GTA3isnt cheating great
06:39:22MidK|GTA3we need a few more video player controls
06:39:31BC|codemk, you missed an entry off that list earlier - do you want to add it, or shall I delete your vote?
06:39:55BC|code<midknight> mono l, mono r, custom, karaoke, stereo wide, stereo, stereo narrow,
06:40:06MidK|GTA3what was missing?
06:40:10MidK|GTA3it was different than that too
06:40:31BC|codei rotated it so mono-l was first is a circle anyway
06:40:52MidK|GTA3i said: mono, mono-l, mono-r, custom, karaoke, stereo wide, stereo, stereo narrow
06:40:56MidK|GTA3or i think i did
06:41:00MidK|GTA3if not just add that
06:41:52BC|codedo you want in 007?
06:42:17BC|code(007 being a famous spy an all)
06:43:01BC|codeclearly not!
06:43:08MidK|GTA3do i want "in" 007?
06:43:19BC|codehey ars pie
06:43:25MidK|GTA3arse pie
06:43:49arspy87how's that coding going?
06:43:56BC|codeI'm looking for opinons on the sequence
06:44:02arspy87ah, cool
06:44:08BC|codewanna vote?
06:47:12BC|codei'll consider that a no
06:48:44arspy87sorry about that BC, what am i voting for?
06:49:02BC|codecompiling - bbs
06:49:20arspy87bbs = ? lol sorry
06:49:29MidK|GTA3be back soon
06:49:34arspy87ah got it
06:49:39 Quit Stevenm ("Client exiting")
06:49:45arspy87what's the voting for midk?
06:50:00MidK|GTA3sequence of channel options
06:57:52 Nick MidK|GTA3 is now known as midknight2k3 (
06:58:51arspy87woo hoo
07:10:22BC|codeneed to write a sprite designer!
07:10:40midknight2k3then make icy tower
07:10:46midknight2k3or boxy jumper
07:11:59midknight2k3nm seeya all later
07:12:26midknight2k3bye BC|code, bye mbr, bye logbot
07:13:25arspy87bye logbot, lol
07:13:33midknight2k3oh, and of course
07:13:43midknight2k3bye adi|home
07:13:56BC|codeg'nite johhn boy
07:14:01midknight2k3ok bye all anyways lol
07:14:03 Part midknight2k3
07:32:02BC|codenite all
07:32:13 Part BC|code
07:37:25 Part arspy87
07:55:10 Join edx{visited} [0] (
07:55:35 Nick edx{visited} is now known as edx (
08:00:27***Saving seen data "./dancer.seen"
08:32:21 Join Kevin [0] (
08:35:59 Quit Kevin (Client Quit)
08:49:24 Join matsl [0] (
09:06:43 Quit webmind (Read error: 110 (Connection timed out))
10:00:28***Saving seen data "./dancer.seen"
10:47:44 Join theTof [0] (
10:51:31 Quit theTof ("ChatZilla 0.8.31 [Mozilla rv:1.4/7]")
12:00:30***Saving seen data "./dancer.seen"
12:16:40 Join methangas [0] (
12:19:42 Quit c0utta (" We ARE the pride of South Australia We are the MIGHTY AD")
12:25:03 Join c0utta [0] (
12:47:11 Join Laurent_ [0] (
13:00:11 Quit mbr (Remote closed the connection)
13:00:44 Join mbr [0] (
13:07:25 Quit c0utta (Read error: 110 (Connection timed out))
13:09:36 Join c0utta [0] (
13:11:34 Quit c0utta (Client Quit)
13:15:21 Join kurzhaarrocker [0] (
13:19:59 Join c0utta [0] (
13:21:07 Quit c0utta (Client Quit)
13:21:55 Join c0utta [0] (
13:22:58 Quit c0utta (Client Quit)
13:23:19 Join c0utta [0] (
13:23:34kurzhaarrockerc0utta: you're blinkin...
13:23:51c0uttayeah, just changing irc clients
13:24:22c0uttatrying to get my profile correct
13:24:43c0uttaone more time....
13:24:46 Quit c0utta (Client Quit)
13:24:57 Join c0utta [0] (
13:25:42c0uttahey, kurz
13:25:53c0uttai heard that you had looked at keyboard mapping
13:26:31kurzhaarrockeroh, yes.
13:26:53kurzhaarrockerI was as far as a working prototype -with some bugs
13:27:08Laurent_hey kurz
13:27:11c0uttai had a long discussion with linus and bluechip about this topic
13:27:22dwihnoso what did you guys find out?
13:27:33kurzhaarrockerc0utta I've read about that
13:27:46kurzhaarrockerhey Laurent_ :)
13:28:54kurzhaarrockerThe main issue (stack problems) that Linus had are true but can be overcome as long as you impose some restrictions
13:29:05c0uttayeah, i've got very simple requirements. i currently modify f2_screen & f3_screen myself
13:29:43kurzhaarrockerI had that as a by product, too...
13:30:22c0uttait seems that the philosophy with rockbox is that you should be able to use any key at any time
13:30:43c0uttathis causes the stack problems that linus was referring to
13:32:23kurzhaarrockerWell for that you need to 'walkback' the callstack before entering another branch of the user interface.
13:33:04Laurent_do keys generate interruptions ?
13:33:33c0uttalooks like keys are polled to me
13:34:36dwihnoA separate key thread?
13:34:48kurzhaarrocker-> To implement user bindable keys we'd have to rewrite every single switch(button) thingie within all rockbox. That's owfully much.
13:35:17dwihnoIf we want bindable keys, that has to be done.
13:36:10Laurent_it depends, you might just replace switch( button ) by switch ( get_user_button( button ) ) if it's done properly
13:36:54kurzhaarrockerLaurent_: no that doesn't work. It's not just button substitution.
13:37:35Laurent_sorry, I tried ;)
13:37:38c0uttamy simple requirements only involve f2 & f3 so easier to implement. i would not be interested in fully bindable keys personally
13:37:43kurzhaarrockerbuttons behave totally different in different contextes
13:38:26kurzhaarrockerBut within f2, f3 screens you use other keys, too
13:38:49*kurzhaarrocker digs to find his old fn screen patch.
13:39:45dwihnoI think the key bindings are pretty OK
13:40:05dwihnoPerhaps customizable F2/F3 screens might be useful.
13:40:12kurzhaarrockerThat's a matter of taste / use
13:40:18c0uttadwihno: i agree. f2/f3 is very underused
13:40:45kurzhaarrockerThey are just the peak of the mountain
13:41:02c0uttathat's why they're relatively easy :)
13:41:14dwihnokurzhaarrocker: tell me how you would like to change the key bindings...
13:42:27kurzhaarrockerI haven't made up my mind completely and its been a while. But eg I'd prefer skipt to next / previous song to be a button combo. And I'm very lonly with this wish. User assignable keys could solve that.
13:43:10c0uttai'm sooo vanilla
13:43:33kurzhaarrockerLet's make some cake from c0utta
13:43:44dwihnokurzhaarrocker: you apply your own patches to make those changes right now?
13:43:46*Laurent_ is going to be sick from eating too much chocolate
13:44:13c0uttaguys, i have some general development questions..
13:44:29c0uttaif i decide to do a patch for f2/f3 how do i keep synced with the source
13:45:34kurzhaarrockerdwihno: no I've been working key schemes since more than half a year because my unit broke. Now I'm more concerned about split editors and a 'rehearsal plugin' which would suit the need of a musician in a band.
13:45:59kurzhaarrockerc0utta cvs is the only way
13:46:43c0uttai'm concerned i'll trample over someone else's work
13:46:45kurzhaarrockerWhat I did was #ifdef 0 the original code. So not every commit resulted in a conflict.
13:46:56dwihnokurzhaarrocker: split editor?
13:47:44kurzhaarrockerc0utta: Just try to update regularily. I think sometimes there has a price to be paid.
13:48:12c0uttabut if i update incomplete/buggy code does this stuff everyone else ?
13:48:22c0uttado the daily builds get mashed ?
13:48:49dwihnoc0utta: Well, key bindings at some places are things that does not change too often. I don't think you will get much of a problem when checking out new revisions.
13:49:36 Quit kurzhaarrocker (Read error: 104 (Connection reset by peer))
13:49:45c0uttai've never worked in a cvs environment before. i have many, many questions
13:50:07 Join kurzhaarrocker [0] (
13:50:10c0uttawhat if someone else is updating the same code as me ?
13:50:35kurzhaarrockerYou'd know as cvs tries to merge the code
13:50:55 Join orby [0] (
13:51:03orbyhey guys
13:51:07kurzhaarrockerc0utta: Are you used to cvs?
13:51:19Laurent_c0utta: cvs does not replace old code with yours, it integrates the differences between your code and the "old" one
13:52:00orbycan anybody here tell me if the FM recorder 20 is worth $80 more than the recorder 20? :)
13:52:06c0uttanever used cvs. come from a vb programming background. the last c programming i did was talking to manufacturing equipment
13:52:19orby(aside from the fact it'll tune FM stuff)
13:53:10dwihnoorby: it isn't, if you ask me. I think the FM has one major downside that prevents me from ever purchasing one. Custom batteries.
13:53:10c0uttaLaurent_: what if: c0utta->current<-laurent_
13:53:28Laurent_c0utta: this can't happen
13:53:42c0uttawhy ?
13:53:42Laurent_the cvs server merges one first, then the other
13:54:15Laurent_they knew some guy name c0utta would try to mess things :)
13:54:21kurzhaarrockerc0utta: Then I'd recommend to play around with cvs for half a day. Put up a simple demo cvs root on your own machine. After you'll feel better. At least it helped me
13:55:07Laurent_i'll second kurz recommendation
13:55:33c0uttait makes me realise how much i don't know..
14:00:33***Saving seen data "./dancer.seen"
14:01:31Laurent_this will be always true throughout you whole life ;)
14:01:34Laurent_living is learning
14:02:03Laurent_(amen) ;)
14:08:21 Part orby
14:10:09Laurent_meaning : there's always more that we don't know than that we know
14:10:33kurzhaarrockerI don't know
14:20:39 Nick edx is now known as e{eating}dx (
14:21:51 Nick e{eating}dx is now known as e{ating}dx (
14:47:50 Join Pierrot [0] (
14:48:18DBUGEnqueued KICK Pierrot
14:48:18PierrotHi everybody
14:49:08PierrotI'm a beginner, and I need pieces of advice
14:49:44Pierrotkurzhaarrocker => do you know how compil a .c file into a .rock file ?
14:50:06kurzhaarrockeryes. You'll need a complete toolchain for that
14:50:37kurzhaarrockerLook at, there it is explained quite well
14:52:01Pierrotfor .rock files as "snow.rock" or simply "helloworld.rock" , I have to compil all the rockbox ?
14:52:33kurzhaarrockerYou don't have to, but its easier that compiling it separately :)
14:52:43kurzhaarrockerthat = than
14:53:16Pierrotand If I want to compil it separately, how can I do ?
14:53:17kurzhaarrockerActually you need parts of the rest of rockbox for your rock
14:53:40Pierrothave you time to explain me how to compil that ?
14:53:56kurzhaarrockerPierrot: don't try. It would really be the hard way to compile it separately
14:54:19Pierrotis it long to compil rockbox ?
14:54:37kurzhaarrockerIt takes less than a minute once the tool chain is set up
14:54:45Pierrotwhat is this tool chain ?
14:55:02kurzhaarrockerAre you working on linux or win32?
14:55:19kurzhaarrockerThat would be the cygwin way :)
14:55:24Pierrotok :)
14:55:50Pierrotand what do I exactly need ?
14:56:33kurzhaarrockeras said before in you can find all the info including a link to the development environment found at
14:56:57Pierrotthx :):)
14:57:35kurzhaarrockerThe tool chain mainly consists of cygwin - which is something like a little bit linux on windows and the compiler for the sh1 processor
14:58:32kurzhaarrockercygwin is needed for all the tools like make, shells to execute the shell scripts and so on.
14:59:19Pierrot(I've just begun to Download tar.exe and bzip2.exe)
14:59:27 Join AciD [0] (
14:59:52kurzhaarrockerPierrot: do you know cvs?
15:00:27Pierrotjust the name ... but I don't really know how it works
15:01:05Laurent_Pierrot: why do you want to compile only some files ?
15:02:10kurzhaarrockerIn the long run you'll want to find out. But for the first try it would be nothing but an additional hazzle. Thus I'd recommend to use a tarball of the daily builds on
15:02:13PierrotBecause I want to add only .rock files ..
15:02:47Pierrotok. thx :)
15:04:05kurzhaarrockerActually the rocks are based on an api defined by the rockbox. So you'll need something to link your rock against. These things are byproducts of the compile process. They aren't downloadable anywhere so you have to build them yourself. You can't get around building rockbox completely at least once.
15:04:32CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
15:04:32*Laurent_ doesn't even know what a .rock is ;) this seems to be an object file of some sort ?
15:04:56kurzhaarrockerAfterward the makefiles take care that only the source that has changed (your rock) is compiled.
15:05:03kurzhaarrockerLaurent_: rocks == plugins
15:05:33Laurent_kurzhaarrocker: oh, thanks
15:06:03Laurent_while I am at it : another dumb question -> what's a .ucl ?
15:06:24kurzhaarrockerThat's the file that is used for flashing
15:07:03kurzhaarrockerI think it's a specially compressed .ajz / .mod thingie
15:07:39 Quit c0utta (Read error: 110 (Connection timed out))
15:11:30Pierrotanother dumb question => how does work a patch ?
15:12:17kurzhaarrockera patch is a file that contains _changes_ to the current source. When you apply a patch these changes are inserted into your source.
15:12:45Pierrotand after, I have to compil the source, haven't I ?
15:12:54Laurent_that's the contrary
15:13:02Laurent_huh ?
15:13:09Pierrotlooooooool Laurent_
15:13:16kurzhaarrockererm the yes was for Pierrot
15:13:21Laurent_first changes, then compile, then patch then submit patch ?
15:14:15PierrotLaurent_ => I think it's not a patch as U have the habit to download
15:14:34Pierrotit's not the kind of patch U use for a game or smthng like that
15:14:54Laurent_no, it's a source patch, made with diff
15:15:52kurzhaarrockerso there are two sides of a patch.
15:15:52kurzhaarrocker1.) the programmer who has changed the source creates a patch with his changes.
15:15:52kurzhaarrocker2.) another programmer who wants to try the changes applies the patch from programmer1 to his own source
15:16:27kurzhaarrockerSo there are two processes: creating a patch and applying a patch.
15:16:41kurzhaarrockerI described applying a patch.
15:17:14Laurent_ok, that's why I was wrong
15:17:20Laurent_my apologies ;)
15:17:24kurzhaarrockeryou weren't
15:17:43Laurent_I was wrong in thinking the process was creating a patch ;)
15:18:34Laurent_oops, I must go, if I don't go shopping now, I'll never do it ;)
15:18:51Laurent_back in a few hours
15:18:53PierrotC ya
15:18:55Laurent_bye !
15:18:57kurzhaarrockerbring some beer, please :)
15:19:04Laurent_good idea ! I will !
15:19:09Pierrot(lol Laurent_ ... comment ch'cause ... un vrai americain :p;))
15:19:21Pierrotlool yeah good idea :p
15:19:27 Quit Laurent_ ("shopping")
15:22:54*kurzhaarrocker is hungy
15:22:57kurzhaarrockersee you later
15:23:11PierrotC you :)
15:23:26 Quit Pierrot ()
15:38:07 Join webmind [0] (
15:42:38 Quit kurzhaarrocker (Read error: 110 (Connection timed out))
15:51:42 Join jameson [0] (~reichenb@
15:54:26jamesonHi! I noticed that you guys have a list of fonts available at; are these in the public domain or licensed under some public license?
16:00:36***Saving seen data "./dancer.seen"
16:09:25 Nick e{ating}dx is now known as edx (
17:07:09 Join OrbyX [0] (total2@
17:08:02OrbyXhey peeps
17:08:09OrbyXanybody here with the FM Recorder?
17:09:00OrbyXbetter yet, anybody here? :)
17:10:24adi|homeim here.. but i odn't have the recorder:)
17:11:00OrbyXI'm just trying to figger out if I should upgrade to the FM Recorder to stick with the regular recorder
17:11:11OrbyXI think the FM Recorder has a different design and maybe different battery?!
17:11:22OrbyXwhich has been my big complaint so far withmy recorder :(
17:11:31OrbyXbattery life sux!
17:19:52dwihnoregular recorder rules!
17:19:59dwihnowell, buy a new set of batteries
17:20:05dwihnopowerex 2200! :D
17:21:18OrbyXThey should stop billing the batteries as "10hrs" life if they only provide 1.5 hour
17:21:34OrbyXDoesn't the FM Recorder come with Li-Ion ones though?
17:24:51dwihnothink so
17:25:10OrbyXI wonder what the life of those are per charge?
17:31:41OrbyXhow long do your powerex 2200's last ya?
17:32:46dwihnoI don't have a couple of those yet.
17:33:01dwihnoI'm using a pair of GP 1800
17:33:07dwihnoWorks like a charm
17:33:52OrbyXwhat are thosE?
17:34:01OrbyXI wonder if anybody makes Li-Ion AA batteries?
17:36:24 Join mecraw__ [0] (~mecraw@
17:37:13dwihnohaven't heard of it
17:53:06 Part OrbyX
17:59:07 Join diddystar5 [0] (
18:00:40***Saving seen data "./dancer.seen"
18:01:02 Quit matsl (Remote closed the connection)
18:06:35 Join BC|code [0] (
18:06:44diddystar5hi BC|code
18:06:49BC|codehey diddy
18:07:02BC|codewrote that code you wanted to see
18:07:10diddystar5send it on
18:26:09diddystar5got thelog scale done for both needles!
18:26:23diddystar5not to clean all the code up
18:26:33BC|codeyou do that too huh?
18:26:50diddystar58 i mean TO clean the code up :)
18:27:12BC|codedid the decay ocde too?
18:27:22diddystar5i need to set a button to switch betwene the scales also...
18:27:28diddystar5not the decay yet
18:27:43diddystar5ill work on that when i finish all the log stuff up
18:34:09BC|codeI've got all the basic audio controls running at full steam now :)
18:34:21BC|codeworking on user-definable stereo seperation now
18:34:53BC|code...the worst bit is designing all the graphics and fonts to make it look good
18:47:00diddystar5think i have the button scale switch done
18:47:03diddystar5now to test
18:48:06BC|codeno joke, I am looking forward to trying it ...LOVE the analogue VU meters :)
18:51:18BC|codelol - clearline is the opposite of drawline, but clearrect is NOT the opposite of drawrect !!
18:51:54 Join scott666 [0] (
18:52:05diddystar5humm the linear scale works, but the log scale well.... the needles are all over the place. sometihng is wrong
18:52:22BC|codedo you want me to take a look and make suggestions?
18:52:29BC|codeno promises
18:52:38diddystar5well the left needle on log scael works, just the right needle
18:52:43diddystar5sure one sec
19:12:04 Quit edx (Read error: 60 (Operation timed out))
19:13:32 Join edx [0] (
19:22:05 Join kurzhaarrocker [0] (
19:30:46 Join gsn [0] (
19:33:57 Join Dogger [0] (
19:34:12 Quit scott666 ("i'll be back...eventually...")
19:36:26 Join Laurent_ [0] (
19:37:00BC|codewhere's santa when you need him?
19:37:57kurzhaarrockernorth pole
19:38:42Laurent_probably drunk amongst his reindeer ;)
19:39:57 Join scott666 [0] (
19:40:03BC|codehear about Blitzen the Brown nosed reindeer ...second in line to Rudolf ...just as fast ...but lousy brakes
19:40:39kurzhaarrockerscott666: BC|code wasn't talking about you
19:40:53scott666i know, its still funny
19:43:38kurzhaarrockerinspired by c0utta I just had a glance at my ancient key schemes code. I have hardly any clue what I did there.
19:45:02 Quit gsn ()
19:45:26 Quit kurzhaarrocker ("Trillian (")
19:45:34 Join track [0] (
19:45:56 Quit track (Client Quit)
19:47:50 Quit Dogger ("Client Exiting")
19:51:21diddystar5bye BC|code, and scott666
19:51:27 Quit diddystar5 ("Leaving")
19:56:24 Part Laurent_
19:56:30 Join Laurent_ [0] (
19:57:25Laurent_was dogger present this afternoon ?
19:58:12BC|codeno idea, sorry
19:58:21Laurent_it's ok ;)
20:00:44***Saving seen data "./dancer.seen"
20:08:38 Quit scott666 ("i'll be back...eventually...")
20:31:47 Join hardeep [0] (1098@
20:49:40 Join Lowfiler [0] (
20:55:36Lowfilerwas j÷rg here?
20:58:42hardeepLowfiler: his nick is [IDC]Dragon and he's online quite frequently.
21:00:02Lowfilerthx hardeep
21:33:43 Join [IDC]Dragon [0] (
21:37:21 Join DJBaz [0] (
21:38:36 Quit [IDC]Dragon ()
21:45:33BC|codedoes anyone here have any info on the stereo seperation parameters?
21:45:36 Join electronic [0] (
21:49:53BC|codeI am wondering what the range is for each parameter and whether it is signed or not
22:00:45***Saving seen data "./dancer.seen"
22:02:51Laurent_BC|code: are those params those of the hardware or of rockbox ?
22:03:25 Join Asmotaku-neko [0] (
22:03:32Asmotaku-nekoHi there !
22:04:32Asmotaku-nekoI'm in seach for infos about RockBox history.
22:04:43Laurent_BC|code: isn't this information in the MAS data sheet ?
22:05:06BC|codeno, it just says 4 registers default values are <...>
22:05:20Asmotaku-nekou mean, like : "how we figured how to program the jukebox" ?
22:05:25 Quit DJBaz ("Leaving")
22:05:34Asmotaku-nekoI wonder how.
22:05:58Laurent_Asmotaku-neko: a good start is the mailing list archive
22:06:12Laurent_Asmotaku-neko: i guess at least that it might be so ;)
22:06:25Asmotaku-nekok, I'm off to it.
22:06:44*Asmotaku-neko is reading RockBox mailing list.
22:06:57Asmotaku-neko-archive, that is-
22:09:09Asmotaku-nekok, I get it, it's pure reverse eng.
22:10:23Asmotaku-nekou guy should have put a hit counter on, that could have been a clue to praise Archos for an SDK release.
22:10:35Asmotaku-nekodon't u think ?
22:11:18BC|codearchos SDK - I wish
22:11:50BC|codeMind you, what we've got is sooooo much better it would be irrelevant now
22:12:21Asmotaku-nekowell, first of all is to build a marker study chart, as a proof.
22:12:38Asmotaku-nekothird-party firmware could attact more customers.
22:12:47Asmotaku-nekoattact = attract.
22:12:54BC|codethey are stamping on it big time
22:13:07Asmotaku-nekohuh ? how that ?
22:13:11BC|codethey want ZERO support and are prepared to sue to make it happen
22:14:13Asmotaku-nekoyou mean their are ready to start legal
22:14:21Asmotaku-nekoyou mean their are ready to start legal proceedings ?
22:14:34BC|codeyep, not with us ...with their new kit
22:14:49Laurent_I think it would be a good idea to measure rockbox' popularity
22:15:37DBUGEnqueued KICK Asmotaku-neko
22:15:37Asmotaku-neko(Laurent_): that should be a good mark for alternative firmware dev.
22:15:45Laurent_that might indeed force them to think again
22:15:45BC|codewe already know how many downloads per day
22:16:07Asmotaku-nekoI'm sure Rockbox had taken more than 100K hit,
22:16:15BC|codegotta start thinking like a marketting capitalist ...not a techie ;)
22:18:39 Join [IDC]Dragon [0] (
22:22:19Asmotaku-nekowell, u're right BC|Code... only BIG MONEY will make them move. damn them.
22:22:57BC|codeTime to get out the Rodger Dawson tapes and set up a meeting with their marketting boys
22:23:06Laurent_stupid dd question : i'd like to copy 10Mb of zeros into a file, I'm doing "dd bs=10M if=/dev/zero of=block" but it never stops :(
22:23:29Laurent_there's something in dd that i don't understand correctly
22:23:58Laurent_ok I got it
22:24:02Laurent_dd writes blocks not bytes
22:24:09Laurent_don't bother answering
22:24:53Laurent_yop, that was it
22:27:37BC|codeneko: you truly are the master of the smiley :)
22:27:54Asmotaku-neko░w^ thanx
22:28:17BC|code...are you a dan grade?
22:29:07Asmotaku-nekoAm i a ... what ?
22:29:31BC|codesorry, just being silly and distracting myself ..."dan grade" is a martial arts term
22:29:36Asmotaku-nekor u talkin'bout martial arts ?
22:29:51Asmotaku-nekooops... sluggish answer.
22:31:02Laurent_well, no answer from telechips... but got an answer from the samsung guy
22:31:49Laurent_he basically says : "did you ask telechips first ? if they wont help you, tell me i'll get you in touch with a guy here who can"
22:32:08Laurent_kind of encouraging
22:39:45[IDC]DragonLowfiler: you were looking for me?
22:42:45 Quit electronic ("Leaving")
22:44:04 Join kurzhaarrocker [0] (
22:44:23[IDC]DragonHi Phil!
22:44:44kurzhaarrockerHi Dragon! Startet rewriting mpeg already?
22:45:22[IDC]DragonLinus wasn't exactly jumping on it
22:45:32BC|codewhat are you doing inside mpeg.c ??
22:45:54BC|codeIt' s just that I am working on some new audio prams code ...which will be affected
22:46:09kurzhaarrockerI want to avoid fiddling there
22:46:14BC|code...or maybe even save you some work
22:46:21BC|codeah okay
22:46:39[IDC]Dragonsee my post about splitting it up
22:47:23kurzhaarrockerWhat I _have_ done is to encapsulate some code for calculating the fileposition from time in a function. But that really was a minor change.
22:48:09 Join LinusN [200] (
22:48:18MThello linus
22:48:23[IDC]DragonAh, there's the man
22:48:32[IDC]DragonSo early today?
22:48:33LinusNwhere? :-)
22:48:37kurzhaarrocker[IDC]Dragon: now lets see LinusN jump :)
22:48:47CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
22:48:47*LinusN jumps
22:48:55[IDC]DragonI was afraid Linus and I have no overlap in IRC
22:49:17[IDC]Dragonjumping on the subject of slitting mpeg.c
22:49:32[IDC]Dragon(that's what we were talking about)
22:49:45MTslitting is too much, just split it :)
22:50:38kurzhaarrockerI encourage the idea of going stream oriented, but I will not be able to help with code, just with encouragement ... :)
22:51:26BC|codeMoral Officer Kurz
22:51:34[IDC]Dragon(looks like Linus jumped away)
22:51:47LinusNi think stream oriented mpeg handling will complicate things even more
22:52:08[IDC]Dragonplease explain...
22:52:17kurzhaarrockerDo you mean things like keeping time?
22:52:33LinusNkeeping time, buffer handling, rw/ff
22:52:36[IDC]Dragonit's not simple the way it is...
22:52:42LinusNno it isn't
22:53:11LinusNbut the file i/o is the smallest part of it
22:53:26[IDC]DragonI wouln't call it streaming, but seperating source and sink just has to be easier (gut feeling)
22:53:56LinusNwhen playing mpeg data, what would be the sink other than the MAS?
22:54:11kurzhaarrockerA file while recording
22:54:22LinusN"when playing"
22:54:23[IDC]Dragontrue for sink, but the source may differ
22:54:33LinusNyes it could
22:54:44LinusNi agree that playing from RAM would be nice
22:55:24LinusNi'm just allergic to unnecessary abstractions
22:55:49LinusNbut maybe a source/sink approach would be nice
22:56:24[IDC]DragonI find them (abstractions) helpful to have a mental modularization
22:56:25BC|codenice way forward to new platforms too
22:56:34MTon a side note, ive emailed multichannel asking for possible information sharing about the xclef
22:56:44LinusNBC|code: which platform doesn't have read()?
22:57:13[IDC]Dragonit's more the MAS side to worry about...
22:57:19BC|codeLinusN: My consideration was more to the sink
22:57:21LinusNi also agree that mpeg.c is a mess
22:57:27MTfingers crossed, they may supply some information about their flash procedure
22:57:51kurzhaarrockerBut when we have a source/sink approach - why stop there? Imagine we have different sources - they all could be encapsulted by the same buffer mechanism which is between the source and the real sink
22:57:54BC|codeLinusN: but intuitively if you isolate the sink, then the source becomes isolated by default
22:58:19LinusNkurzhaarrocker: you said the forbidden word - encapsulate!!!!
22:58:38 Join c0utta [0] (
22:58:40kurzhaarrockerI love encapsulating - but I'm a java programmer.
22:58:40[IDC]Dragonwhy forbidden?
22:58:56[IDC]Dragonnot that again
22:58:59MTotherwise it will get quite messy, bad code = the machine stops working :/
22:59:28BC|code(quick change your teminology)
22:59:30kurzhaarrocker(what does Lins mean by K.I.S.S?)
22:59:30[IDC]DragonI've learned (in practice) that clear interfaces make stuff a lot more maintainable
22:59:36MTim v much looking forward to this machine arriving, 120 MHz is a lot of oomph to play with :)
22:59:40BC|codeKeep It Simple, Stupid
22:59:41LinusNKeep It Simple Stupid
22:59:42MTKeep It Simple Stupid
22:59:59[IDC]Dragonrepeat, everyone
23:00:03LinusNKeep It Simple Stupid
23:00:08BC|codeKeep It Simple, Stupid
23:00:10kurzhaarrocker Keep It Simple Stupid
23:00:11Laurent_also, abstractions leak
23:00:18kurzhaarrockerIs that mantra?
23:00:21BC|codehmm, me too.. brb
23:00:26LinusNLaurent_: Joel Spoilsky?
23:00:46Laurent_ (yup)
23:00:51Asmotaku-neko*sob* *sob* I wish I could find an alternate firmware for my AV140 ! *sob*
23:00:55LinusNK.I.S.S is a Rockbox mantra
23:01:21LinusNAsmotaku-neko: you wanted the history of Rockbox?
23:02:05Laurent_I also think that it would be a nice thing to have on the web site
23:02:19LinusNthe history?
23:02:28Asmotaku-nekoWell, yes, but I wondered how your community managed to dev some firmware on JBM
23:02:29LinusNDaniel is writing it
23:02:52LinusNAsmotaku-neko: that is a completely different beast
23:03:01Laurent_when I discovered Rockbox it has been the first thing I have searched
23:03:05Asmotaku-nekoDaniel is writing some firmware for AV1XX series ? is that it ?
23:03:15LinusNno, the history
23:03:27Asmotaku-nekoOr did I misunderstood ?
23:03:55Asmotaku-nekoduh ! painful answer !
23:04:02Asmotaku-neko*sob* *whine*
23:04:18LinusNAsmotaku-neko: work is being done
23:04:40LinusNbut none of the core rockbox developers have a MM device
23:04:54 Join Pierrot [0] (
23:04:59BC|codebut if you're offering to buy a bunch of them...
23:05:15PierrotHi everybody
23:05:19LinusNPierrot: hi
23:05:23[IDC]DragonAV3xx, please
23:05:30Asmotaku-nekoat least, I've posted a topic about the petition, on the biggest french DivX forum. They reprensent the second community, in terms of population.
23:05:51Asmotaku-nekoreprensent = represent
23:06:04 Join scott666 [0] (
23:06:23Laurent_so the "we wont touch mm devices" motto, is more due to lack of resources than to some other reason ?
23:06:35LinusNalso lack of interest
23:06:47Laurent_I thought it was because there were too much differences on the hardware side
23:06:55BC|codeI guess the interest would grow if we all owned them
23:07:00LinusNif i felt i needed a mm decice i would have bought one already
23:07:03MTAsmotaku-neko: it was a bit like this with the FM Recorder - people were getting there, but quite slowly. Then one kind man bought Linus an FM Recorder, he disected it and within a month the port was stable and complete
23:07:07MTalthough it is VASTLY different hardware
23:07:07Laurent_the MM devices would be interesting PDAs
23:07:25PierrotI have a problem. my girlfriend had plugged her jukebox on an USB port, but she unplugged it without removing it by the icon in the taskbar. Now when she plug the JB on the computer, she get a blue screen ... Have U got an idea of what she could do ??
23:07:34BC|code1 month!? Respect dude !
23:07:50Pierrotshe did
23:07:56Pierrotit's still the same probleem
23:08:22MTwin9x i bet?
23:08:22BC|codePierrot: reboot safe mode - system - device mangler - remove ALL occurences of Jukbox drivers - reboot
23:08:28LinusNPierrot: does the jukebox work without usb?
23:08:36PierrotI don't know
23:08:41Pierrotwhat if it works .. ?
23:08:48BC|codeahhh, check Linus advice first
23:08:52Pierrotlol ok
23:08:55LinusN[IDC]Dragon: about the source/sink idea, what did you have in mind?
23:09:18Pierrotso, I don't know if it works or not ... tell me what to do in both cases please LinusN
23:09:18[IDC]Dragonyou mean, in addition to what I posted?
23:09:40LinusNPierrot: i want to know if you can playt music with it?
23:10:06[IDC]DragonI've worked with DirectShow in the past, and we developed an architecture for mobile phones recently.
23:10:39PierrotI don't know ... I don't have my girlfriend's JB here
23:10:40[IDC]DragonBoth are too big for what we need here, but stripping down is the easier part.
23:10:45*c0utta is AWAY: -[ auto-away ]- away for about 10 minutes
23:10:50Pierrotbut I guess it can play music
23:10:57LinusNPierrot: i can imagine two things:
23:11:31LinusNeither the windows drivers are corrupt (doubt it), or the file system on the JB is corrupt (more likely)
23:11:59PierrotWhat can I do if the file system on the JB is corrupt ?
23:12:02LinusNi suggest you connect it to another computer and scandisk (or reformat) the drive
23:12:52LinusN[IDC]Dragon: i can imagine a simple abstraction of the I/O parts
23:13:11[IDC]Dragonsomething like that
23:13:13LinusNso you can read/seek/write to a RAM buffer
23:13:32[IDC]Dragonwho's RAM buffer?
23:13:38Laurent_does one say zeroes or zeros ?
23:13:43LinusNmy RAm buffer
23:14:00[IDC]Dragonwho is "me"?
23:14:23LinusNLaurent_: i guess it's zeros in USA and zeroes in England
23:14:25[IDC]DragonI suggest the source owns the buffer
23:14:52[IDC]Dragonthe sink consumes it and notifies before it needs more
23:15:17LinusNhow long before?
23:15:42LinusNand how does it notify?
23:15:45LinusNand why?
23:15:49kurzhaarrockerThat's why I'd prefer streaming then the sink can just fetch from the input "queue"
23:15:49[IDC]Dragondragging me onthe ice? ;)
23:16:07LinusNthis is why i prefer a simple read()
23:16:26LinusNno abstract mumbojumbo
23:16:35[IDC]Dragonlet's put it the other way: the source tries to stuff the buffer until it stalls
23:17:11BC|code::: should we code with the intention of our code running on machines which may be supported in the future?
23:17:14kurzhaarrockerIt could stuff its own buffer which the sink reads
23:17:37LinusNBC|code: not if you ask me
23:17:40[IDC]Dragonkurzhaarrocker: that's what I mean
23:17:54BC|codethat makes perfect sense of the opposed views
23:18:21kurzhaarrockerBut that means that all the high / lowwater mechanisms have to be handled by the source.
23:18:30[IDC]DragonI'd like to do the bitswapping late, like just in time, to keep the data intact for seeking
23:18:32kurzhaarrockerBy every source
23:18:39LinusNi see no reason to go lengths to make code portable when you don't even know if it will ever be ported
23:19:12BC|codeI understand your thought, but to many (myself included) it is habitual ...not sure if that is good or bad
23:19:38[IDC]Dragonkurzhaarrocker: the watermarking is because of the source, so should be its business.
23:19:57LinusNwell, i said "go lengths" because there are of course reasons why we stick to the POSIX API for instance
23:20:00[IDC]Dragonthe powered down disk has other latency than a RAM buffer
23:20:21LinusNso some effort is always good
23:20:23Laurent_oops, does anyone know what to type to reset a terminal that switched to some obscure mode ? (off topic I know)
23:20:32LinusNstty sane
23:20:33kurzhaarrockerBut why not put that low highwater mechanism into a pipe that connects source and drain
23:20:57[IDC]Dragonmy vote
23:20:59Laurent_LinusN: hum... not enough unfortunately ;)
23:21:02Pierrotthx for your advices LinusN :)
23:21:09LinusNPierrot: you're welcome
23:21:13Pierrotgood bye everybody
23:21:19kurzhaarrockercu Pierrot
23:21:39 Quit Pierrot ()
23:21:46LinusNkurzhaarrocker, [IDC]Dragon: hold it for a second
23:21:55*kurzhaarrocker holds it
23:22:01*[IDC]Dragon holds "it"
23:22:14LinusNgo back and think of the reasons why you wanted to do this source/sink stuff
23:22:48LinusNwas it to create a nice "architecture" that was really neat and cool, or was it because you wanted to play mpeg data from different sources?
23:23:16[IDC]Dragonthe latter, but I was thinking about the way to there
23:23:40[IDC]Dragonand other things:
23:23:52kurzhaarrockerfiner control over what is played when
23:24:12[IDC]DragonI'd like the buffer data to stay unswapped for as long as possible, to allow seeking around
23:24:31LinusN[IDC]Dragon: explain
23:24:59[IDC]DragonWe can parse the headers way better when it's plain
23:25:21[IDC]Dragonan old feature whish of mine: have audible FF/FR
23:25:26kurzhaarrocker(apropos swapping: I heared rumors there was someone with access to the MAS specs -> could swapping be delegated to the mas?)
23:25:45LinusNkurzhaarrocker: yes and no
23:25:53[IDC]DragonI was discussing that wit BC yesterday
23:26:06kurzhaarrockerI'll read that later
23:26:14[IDC]Dragonthe avail info is by far not enough
23:26:15LinusNit would probably require quite a lot of code rewriting
23:26:21LinusNin the MAS
23:26:32[IDC]Dragonfor which we have no tools
23:26:42LinusNnor knowledge
23:26:45[IDC]Dragonnor even an instruction set
23:26:54BC|codeLinus ...have you managed to speak to "my contact" yet?
23:26:59*kurzhaarrocker 's dream burst like a soap bubble
23:27:18LinusNyes, we have contacted Micronas once again
23:27:24Laurent_are the recorder/player's archos firmwares compressed ?
23:27:45kurzhaarrockerjust scrambled
23:27:47Laurent_the gmini's is
23:27:51LinusNBC|code: however, the nice guy at Micronas has quit his job, so another person is taking his place
23:27:55BC|codehe can take his time to reply sometimes, (although not as bad as mas) but his replies are always worth waiting for :)
23:28:17LinusNwe contacted him, waiting for an answer
23:29:04kurzhaarrockerLinusN: if you vote for a simple read: would there still any buffering on the side of the mpeg engine?
23:29:11BC|codeLinus: there are more under-handed ways of contacting techies ...if you get stuck, PLEASE ask
23:29:22LinusNBC|code: i'll do that
23:29:40kurzhaarrockereg the dma chunks..
23:29:57LinusNkurzhaarrocker: the mpeg.c code (today) needs a large consecutive RAM buffer
23:30:35LinusNthat can of course change
23:30:56Laurent_255464 Oct 22 13:42 FirmwareGmini100_v120.aos
23:30:56Laurent_253817 Dec 29 23:19 FirmwareGmini100_v120.aos.gz
23:31:48LinusNthe mpeg playing today uses very little CPU (except for the bitswapping)
23:32:13LinusNthanks to the DMA and the large buffer
23:32:17BC|codeLinus: do you know the valid range for the Stereo seperation registers?
23:32:38[IDC]Dragondo you set the DMA for the whole lot, or piecewise?
23:32:44 Join Guest [0] (
23:32:50LinusN8k at a time
23:33:08[IDC]Dragonso there is a granularity, good.
23:33:12Laurent_I read on the mailing list that the hd's cache is invalidated when it stops spinning, is that true ?
23:33:28 Quit Guest (Client Quit)
23:33:48BC|codelinus: 8k !? so where does 7ffff or 80001 come from then?
23:33:48LinusNBC|code: i assume it's -1 to 0
23:33:49[IDC]Dragonthat was short...
23:33:52kurzhaarrockerstrange Guest
23:34:07Laurent_probably very shy
23:34:07LinusNBC|code: where does what?
23:34:26LinusNLaurent_: yes, the cache is invalidated
23:34:39BC|codeLinusN: D0:7F? Stereo/narrow/wide/mono/etc
23:34:56Laurent_LinusN: isn't there any way to prevent this from happening ? that would help buffering a lot
23:35:45 Join chris [0] (
23:35:49LinusNLaurent_: yes, we can use another powersaving mode, which draws a lot more power
23:35:52kurzhaarrockerBut it would need power to keep the hd on
23:36:29Laurent_LinusN: I see, more buffer but less batteries so it's probably better not to do it ?
23:36:45BC|codehi Chris
23:36:51LinusNLaurent_: true, and don't overestimate the importance of the disk cache
23:37:06chrisata -61
23:37:45Laurent_how many bytes is the memory buffer ?
23:38:01chriswhere can Isee it?
23:38:46chrischui franšais
23:38:51LinusNchris: when do you get it?
23:39:11kurzhaarrockerInfo in the main menu
23:39:13LinusNLaurent_: 1.7Mbytes or so
23:39:38chrisI got it when i push on
23:40:01LinusNBC|code: i think the range is -1 to +1
23:40:17LinusNchris: flashed?
23:40:31chrisnot flashed
23:40:45chrisIm formating the drive
23:40:49BC|codeLinus: how does 0x7FFFF fit that rrange?
23:41:11kurzhaarrockerlol! virtual memory!
23:41:18BC|codeLinus: maybe a better example would be 0xC0000
23:41:32LinusNBC|code: page 23 in the 3587F data sheet
23:42:06BC|codeahhaa, maybe I have the wrong PDF open
23:43:29chrisI d formatted the drive but the register hd drive error came back
23:43:55LinusNchris: that is very often a battery problem
23:44:16chrisI hope so
23:44:44kurzhaarrockerDidn't these problems rise somtimes when the battery connecter solder points break?
23:45:36chrismy problems came when i deleted some files
23:45:59kurzhaarrockercould still be a battery problem
23:46:41BC|codeLinus: cheers - wrong data sheet!
23:47:38Laurent_got an idea : let's assume hd's cache is 0.5Mb, proceed so -> read 1.7 Mb into memory in a single run, then read the next 0.5Mb consecutively chunk by chunk into a small buffer (ie don't store it), then power down while keeping the cache valid, as soon as 0.5Mb are played, read them from the cache into memory, then power down completely for 1.7Mb of playing data
23:47:43kurzhaarrockermpeg engine: Would it be reasonable to keep a little queue of maybe 3 or 4 8kb dma chunks in the mpeg engine and the rest of the bufferin has to be done by the source?
23:48:40[IDC]DragonI'd say: the source hands buffers to it, period.
23:48:54[IDC]DragonThe sink eats it up as availabe
23:48:55kurzhaarrockerLaurent_: I have no idea wether there is sucha fine control over the cache of the disk
23:48:59Laurent_so, the hd stays in low power mode during 0.5Mb worth of playing data and is off during 1.7Mb worth of playing data, so general consumption is down by a good factor
23:49:03[IDC]Dragonand passes them back
23:49:09LinusNLaurent_: that may work, but you don't know how and when and what the HD stores in the cache, and i see lots of complications, like if the user pauses the music when the hd is not powered down
23:49:25 Quit chris ("Leaving")
23:50:28scott666and what if it its an 8mb cache?
23:50:41LinusN[IDC]Dragon: that means that the buffer handling would be done at interrupt time?
23:50:43kurzhaarrocker[IDC]Dragon: And the buffer has to be aproppriately granular ?
23:50:48Laurent_I don't know who that user guy is but in every design I ever saw, he's always putting us in trouble ;)
23:50:55scott666youd keep the disk up for 6.3mb of time
23:51:19[IDC]DragonLinusN: not necessarily, it can be a task
23:51:25LinusNLaurent_: "Who is General Failure, and why is he reading my drive?"
23:53:03LinusNLaurent_: i had that as my email signature for a while, and I later received an answer to that:
23:53:37BC|code...He is there to keep Major Problem under control
23:53:38LinusN"He is the commanding officer of Major Disk Crash, who is the commanding officer of Private Files"
23:53:53[IDC]Dragonthe memory doesn't have to be chunked, it could stay one circular buffer as it is today.
23:54:15Laurent_nice one ;)
23:54:16[IDC]Dragonthe sink maintained the read pointer, the source does the write pointer
23:55:44kurzhaarrockerhm. ok. I want functions to control the read pointer, too.
23:56:16[IDC]Dragondepends on if you're a source or sink
23:56:28[IDC]Dragonkurzhaarrocker: for which use case?
23:57:08kurzhaarrockerI want to be able to tell it: when you arrived here, then jump there
23:57:33[IDC]Dragonhmm, then you may better work whith chunks
23:57:44[IDC]Dragonso they can form a linked list
23:58:19kurzhaarrockerIf we want to beep we could tell the read pointer jumpt to the beep and then continue

Previous day | Next day