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 2006-08-25

00:00:18killa909what does ogg stand for
00:00:21juxtapi got myself a patched build... album art works great...
00:00:22killa909i mean, i have an idea
00:00:34juxtapSUCCESS STORY
00:00:34 Join Kohlrabi [0] (
00:00:35stripwaxkilla909 - pretty sure it doesn't stand for anything.
00:00:44stripwaxjuxtap - yeah, it's neat
00:00:56killa909im going to refrain from telling you
00:01:07killa909juxtap - thats cool
00:01:28killa909is it possible that might be included in future builds of rockbox itself
00:02:07stripwaxkilla909 - it's certainly possible
00:02:18stripwaxalmost certain, I would expect - just not quite yet
00:02:58killa909im noticing this ipod getting mighty warm
00:03:55*barrywardell still can't figure it out
00:04:00 Quit Siku (Read error: 145 (Connection timed out))
00:04:04stripwaxkilla909 - rockbox is not yet optimised for speed, so it has to run the processor faster than the Apple firmware to keep up. Should improve in the future.
00:04:17barrywardellapps/main.c and bootloader/h10.c are identical
00:04:35barrywardellcrt0.S always uses the bootloader code
00:05:22midkaykilla909: to clarify a bit - faster, but certainly within specifications (at rated maximum that even Apple firmware runs it at sometimes).
00:05:26barrywardellcompile bootloader build and main() gets called
00:05:34barrywardellcompile normal build and it doesn't???
00:06:11preglowany fast way to change repository location for an already checked out source code tree? subversion
00:06:30barrywardellis there some other difference between bootloader and normal builds?
00:06:39killa909okay got "data abort" what the hell does that mean
00:07:01 Join oxygen77 [0] (
00:07:01killa909at 004dd40
00:07:11preglowbarrywardell: have you tried disassembling the file to see that your stuff is actually linked in properly?P
00:07:22oxygen77hello all
00:07:33killa909had to restart
00:07:36killa909wasnt even doing anything
00:07:38killa909thats sad
00:07:42barrywardelli'll try that preglow
00:07:47oxygen77amiconn, are you here ?
00:09:10stripwaxkilla909 - is it reproducible? sounds like you've found (another?) bug
00:09:37killa909i will try to
00:09:59 Part juxtap
00:10:09killa909listening to same file as before
00:10:39killa909i think the rockbox software may not like rap music
00:12:15preglowmine does
00:12:31killa909mr me too
00:12:37 Quit oxygen77 ("Leaving")
00:12:43preglowor at least it obviously tolerates it
00:12:44linuxstbbarrywardell: How are you testing that main() is called?
00:13:06barrywardelli'm writing a line of text to the lcd
00:13:19Mikachuhello world!
00:13:27preglowgood bye, cruel world!
00:13:40linuxstbbarrywardell: Where? Before the call to app_main() in main() ?
00:14:09barrywardellit's the only thing main() does
00:14:26barrywardellmy main.c is identical to bootloader/h10.c
00:14:38barrywardellthe one that's in cvs
00:14:46linuxstbWhy not just run the standard Rockbox main() function?
00:15:01linuxstbAnd then see how far through init() it gets.
00:15:39linuxstbThe bootloader has a lot of differences to main Rockbox - see all the #ifndef BOOTLOADER defines in firmware/*
00:15:40barrywardelli tried that first
00:16:15barrywardelli put the lcd update at the very start of main and still nothing
00:16:34barrywardellso I think it doesn't even get that far
00:16:42linuxstb? You can't put lcd_update() in main().
00:17:11linuxstbIt needs to come after lcd_init() is called in init() - and you can't write text until after font_init() is called.
00:17:40barrywardelli also tried putting it straight after font_init()
00:18:35linuxstbSo what happens when you try and boot Rockbox? The bootloader text stays on the screen, and your H10 freezes?
00:18:37*jhMikeS thinks it should say "pluto is no longer a planet"
00:18:53barrywardellyes, exactly
00:19:30 Part killa909
00:22:07barrywardelli'll go through firmware/* and find #ifdef BOOTLOADER. maybe I'll spot something
00:22:51stripwaxsee ya
00:22:53 Part stripwax
00:23:40mirakpreglow: what does a lea ?
00:24:25mirakI am not sure to understand <ea>y notation
00:25:29 Join San||Away [0] (
00:25:29 Quit San (Read error: 104 (Connection reset by peer))
00:31:01jhMikeS<ea> == effective address?
00:35:51*rotator learns the joys of debugging using debugf on the sim
00:36:07mirakjhMikeS: yes
00:36:09 Join juxtap [0] (
00:36:24jhMikeSmirak: do you mean what does lea do?
00:36:55mirakyes I am not sure to understand the interest compared to a move
00:37:37jhMikeSlea.l (offset,Ay),Ax for instance moves offset+Ay into Ax
00:38:16jhMikeSso if Ay = 16, lea.l (10, Ay), Ax puts 26 into Ax
00:38:36jhMikeSThe register must be address registers
00:40:16jhMikeSTo sum up: You create an address from another address whereas move would load data at the address <ea>y
00:41:04preglowmirak: lea just calculates the address specified by the second operand
00:41:26preglowmirak: lea.l %a0, (%d0, %a0) -> %a0 = %d0 + %a0
00:41:55mirakbut if I do
00:42:02jhMikeSpreglow: I know, I'm crappy at explaining stuff.
00:42:06mirakmove (%a0),%a1
00:42:19jhMikeSIt loads the data at address %a0
00:42:46mirakinto %a1
00:43:27jhMikeSDon't forget the size
00:43:28mirakok I see I needed that in a first draw I did
00:43:41mirakI used add #8,%a0
00:44:04mirakwith a move of a base adress into %a0 just before,
00:44:04jhMikeSUse lea.l (8,%a0),%a1
00:44:10mirakyep got it
00:46:07preglowjhMikeS: i didn't notice you tried to explain it, heh
00:46:22preglowjhMikeS: i tend to just read the highlight message, then answer without seeing if anyone else did
00:46:27blacksI'm using rockbox for the ipod mini gen1, is it normal for it to freeze up, requiring me to hold menu+select? It did it about 5 times randomly at work today
00:47:07 Quit idnar (Nick collision from services.)
00:47:07jhMikeSpreglow: it's ok
00:47:10 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
00:49:14jhMikeSWeird how not long ago I didn't squat about coldfire assembly but I guess it reads much like x86 but the destination reversal seemed weird.
00:51:25 Quit Kohlrabi ("Quit")
00:52:07*jhMikeS still needs to learn some (e)mac and lots of other stuff too.
00:52:42preglowit's pretty simple stuff
00:52:52preglownot too much new
00:52:53 Quit ender` (" Cluster bombing from B-52s is very, very accurate -- the bombs always hit the ground.")
00:53:24mirakman it's horrible
00:53:27jhMikeSyeah, those (e)mac instructions look a bit there any pure assembly in RB that uses it (.S)
00:53:59mirakjhMikeS: yes, look imdct_mcf5249.S
00:54:25 Quit matsl (Remote closed the connection)
00:54:34preglowjhMikeS: yeah, and eq_cf.S and dsp_cf.S
00:54:40preglowand coldfire.S in flac
00:55:02preglowprobably all the asm stuff i've done
00:55:25jhMikeSIt should be clearer with the gcc inline stuff out of the way
00:55:52pregloweq_cf.S is a pretty clear example
00:56:02preglowseen together with the generic routine in eq.c
00:57:17mirakI have an int_16* that defines the start of the block I want to run an idct row on just 8 coef that I exepect to be 16 bits, so I do "move.l (%[block]), %d0\n\t" exepecting
00:57:23miraknow I am not sure about order
00:58:07mirakblock[0] should be in %d0u and block[1] in %dl
00:58:11mirakright ?
00:59:00 Quit TeaSea (Remote closed the connection)
00:59:06 Join Drkepilogue [0] (
00:59:11mirakho you understood what I said, I was about to rewrite it
00:59:15Drkepiloguehey dood/doodettes
00:59:34Drkepiloguedo anyone know if Rockbox uses C or C++ to code?
00:59:39mirakpreglow: so I load the 8 coef in d0 d1 d2 d3
00:59:49scorcheDrkepilogue: C
00:59:56mirakpreglow: you might want to give a look ?
01:00:07Drkepiloguehmm, so im guessing taking a C++ class not gonna do much good ey?
01:00:26preglowDrkepilogue: sure it would
01:00:29preglowDrkepilogue: they're not very different
01:00:37scorcheDrkepilogue: well, it will teach your programming basics as well
01:01:00preglowmirak: a quick one
01:01:05Drkepiloguebut say if i want to help out coding rockbox... it will still be some benifit?
01:01:30Drkepiloguealright cool, thanks
01:02:29mirakam looking for a pastebin
01:02:44scorcheDrkepilogue: and in the future, do not ask your question in both the forums and IRC
01:02:47linuxstbmirak: normally works for me
01:03:23Drkepiloguecuz i decided to ask forum first
01:03:32Drkepiloguebut i forgot about irc,
01:03:36Drkepiloguetil like a few min ago
01:06:13mirakpreglow: what I did was using linuxstb paper mpeg2 optimisation. in it there is a idct row with matrix multiplication plus the formula in algebra. So I verified the formula and matrix multiplication in matlab, so I am sure the equation is correct to perform an idct row. Now I am really not sure how the data formats, and not sure also the idct_row in idct.c really performs a 1 dimention idct on a row. The output I got is
01:06:13mirak garbeled and tainted in green.
01:06:53mirakI mean not sure about the data formats, the byte ordering etcetera
01:08:06 Join webguest69 [0] (
01:08:11webguest69hey guys
01:08:28webguest69I have an error question if anyone is up for answering
01:08:33 Join lukaswayne9 [0] (
01:09:00mirakby the way I don't really exept this thing to be faster than C version, since it's optimised a lot more, but I would like to have thing work at least first. The other option would be to just transcribe in asm the idct_row, I will do that next.
01:09:17 Quit lightyear (Remote closed the connection)
01:09:40preglowmirak: what's up with the macsr move?
01:09:40 Quit webguest69 (Client Quit)
01:09:44preglowmirak: where'd you get that value from?
01:10:08preglowunless i'm completely mistanken, you just want to do move.l #0, %macsr
01:10:09mirakpreglow: it's to enable the signed mode
01:10:14Mikachuhow do you like this shell command to convert webirc hex to decimal ip? (zsh only syntax)
01:10:20Mikachu% for b c (`for a ({1..8}) {i=44d9358b;echo $i[a]}`) {hextodec $b$c}|xargs|sed s/' '/./g
01:10:47mirakpreglow: signed mode plus overflow limitation.
01:12:30preglowwhat's with all the 0xffffff ?
01:12:41preglowthe docs clearly state bits 31-12 are reserved and should be cleared, you've set them all...
01:13:41mirakpreglow: I have taken that from an how to on the web, but will try without.
01:13:52mirakindeed the BUTTERFLY function works with this setting
01:14:42preglowprobably has nothing to say, no
01:14:47preglowthe low byte is correct
01:14:51jhMikeSthere's that problem of when you hold the power key not long enough to cause RB to go into shutdown but you wait a few seconds and it shuts down that due to the ONKEY1s interrupt not being cleared in the PCF50606?
01:14:58preglowi don't believe you need saturation, thougfh
01:16:20mirakpreglow: yes probably unecessary
01:16:35preglowbut yeah
01:16:39preglowotherwise it looks good
01:17:14jhMikeSDatasheet says pcf50606 will go into standby in 8s if not serviced
01:17:33 Join OLD45 [0] (
01:18:39 Quit OLD45 (Client Quit)
01:19:16mirakpreglow: hum probably it should be 40 instead of 80 ?
01:19:47preglow40 is unsigned mode
01:20:47 Join webguest45 [0] (
01:21:53preglowi'm having a really hard time seeing how idct_row can benefit much from using the emac
01:22:37mirakwritten as it is hardly believe
01:22:43preglowanyway, what the hell are you doing in idct_row?
01:22:46preglowit's clearly not based on the c code
01:22:47mirakwritten as it is, hardly I believe
01:23:04preglowit's a bit hard to verify if it's correct if i don't know what it's supposed to do
01:23:06mirakpreglow: no it's not
01:23:20 Quit Rondom ("Ex-Chat")
01:23:31mirakpreglow: well just spotting some incoherent asssembly operand using and such
01:23:44mirakbut I can tell you
01:24:12mirakyou have a 8x8 block of coefs
01:24:15preglowthere's lots of potential for optimising here, that's for sure
01:24:29preglowthe first four move.l instructions should be one movem.l instead
01:24:46preglowthe accumulator clearing is not needed, since you should always use movclr.l to get answers anyway
01:25:04mirakpreglow: yes, but in doubt ...
01:25:28mirakok for the movem
01:26:13preglowthe lea.l (1, %a0), %a0 should be written addq.l #1, %a0, since that's shorter and faster
01:26:20 Quit barrywardell ()
01:27:04mirak"movem.l (%[block]),%d0-%d7\n\t" ??
01:27:26preglowdo you really want to add just one byte to %a0?
01:27:38mirakone word
01:27:39preglowno, movem.l (%[block], %d0-%d3
01:27:47preglowthen you need to do addq.l #4, %a0
01:27:47mirakone long word
01:27:53preglow1 is just one byte
01:28:57mirakwhat I do with mac
01:29:15preglowmacs look fine as far as i can see
01:30:03preglowbut i would rather use a movem.l instead of doing the parallel load with mac.w
01:30:05preglowbut that can be fixed later
01:30:23mirakI think the mac is wrong
01:30:30mirakit should be d0u
01:30:47mirakoh no it's ok, sorry, it's block[1]
01:31:03mirakI know what's wrong
01:31:28mirakI concatenated the two matrix, but inverted the order for computation so it's wrong.
01:31:39mirakI need to swap them in mat_coef array
01:31:40 Join fatherfork [0] (
01:32:00webguest45I am looking for a tutorial or guide for creating FLAC files. I am new to rockbox and using an H 120. I want to create a directory of folders, subfolders and files using FLAC as the compression theme. The files would be bird songs with the bird name displayed. The folders and subfolders would help me organize the bird songs to family.
01:32:41mirakpreglow: if you see what I mean ... :)
01:32:44linuxstbwebguest45: What format are the tracks now, and how are they stored?
01:32:59webguest45wav and they are on my PC
01:33:13webguest45no golders yet
01:35:56 Nick JoeBorn is now known as JB_away (
01:36:25mirakpreglow: I am wondering what the mac is for, if it can't be used to perform idct or imdct
01:38:19preglowmirak: it's great for filtering
01:38:34preglowmirak: and it _can_ be used for those tasks as well, it just depends on how you do it
01:38:37jhMikeSThe iPods don't have recording? Is that because RB doesn't support it yet for that?
01:38:40preglowmirak: like you've just done it
01:39:00webguest45linuxstb> are you looking for someting that I may be able to use
01:39:04preglowmirak: for mp3, i use almost 90% mac instructions in the imdct routine
01:39:23jhMikeSwhat the diff between idct and imdct?
01:39:42preglowimdct is a lapped one-dimensional transform
01:40:27linuxstbwebguest45: Sorry, no - I've got no ideas...
01:40:36 Quit lukaswayne9 ("Ex-Chat")
01:40:42preglowother than that, they're not that different, both decompose a signal into cosine components of different frequencies
01:40:51webguest45OK I will do more searching
01:41:10*rotator nearly has sample-accurate flac seeking implemented
01:41:32rotatorjust a few more bugs to work out
01:41:32 Quit MarcoPolo (Read error: 104 (Connection reset by peer))
01:41:43jhMikeSHow does it get the phase? It's easy to see how a fourier series get the phase by combing differnt amount of sin and cos for a frequency
01:41:49linuxstbjhMikeS: I've started work on ipod recording, but it's not in CVS yet. And it's probably badly broken by all the recent changes - I dare not run cvs update in that tree...
01:41:53mirakpreglow: what we lose indeed is we can't use immediate value
01:42:10preglowrotator: w0000t
01:42:20mirakI am not sure it wouldn't be faster be add #Wsomething
01:42:28preglowjhMikeS: it doesn't have phase
01:42:37jhMikeSlinuxstb: I'll have my changes up soon in FS
01:42:48rotatorpreglow: hehe, works about half the time right now
01:42:49preglowjhMikeS: there's just cosines, which means they've all got a fixed phase
01:42:54jhMikeSpreglow: It didn't seem like it could.
01:43:03 Quit webguest45 ("CGI:IRC")
01:43:27jhMikeSpreglow: I guess that mean I do have grasp on it. It's much like a fourier series with the sin part taken out
01:43:39preglowjhMikeS: it's more or less exactly like that
01:43:56preglowjhMikeS: cosine transforms have good energy compaction properties, which is why they're so often used
01:44:14jhMikeSbecause the it start at 1 right?
01:44:31***Saving seen data "./dancer.seen"
01:44:34jhMikeSthe dc coeficient is at the left then
01:44:48preglowjhMikeS: the dc coefficient is unimportant for sound
01:44:59preglowjhMikeS: so no, that's not just why
01:45:02jhMikeSnot for video
01:45:11preglowfor video the dc component is very important
01:45:17preglowand pictures
01:45:34*linuxstb goes to bed
01:46:22jhMikeSvideo obviously has a lot of dc
01:46:33preglowvery good example of dct energy compaction
01:46:44preglowalmost all the energy are in the lower frequency coefs
01:47:19jhMikeSA 2d dct is a dct of a 1d dct?
01:47:20preglowit seems i suck at english again
01:47:25mirakpreglow: I found one explanation of DCT decomposition like in idct.c but I haven't found one for th inverse dct
01:47:33preglowjhMikeS: yeah, i guess so, similar to a 2d dft
01:47:37 Quit fatherfork ()
01:48:12mirakgood night
01:49:44jhMikeSI think I can just do it up in a spreadsheet and compare the results with the fs waveform coeffs. I did the fs in excel real quick.
01:51:08jhMikeSIf it can compact energy so well it must rely less on high harmonics to reproduce the function
01:53:07preglowsleepy time for me too
02:05:42 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
02:06:51 Join webguest82 [0] (
02:08:07webguest82linuxstb> I found this for FLAC users, thought you may want to know.
02:09:12webguest82was webguest 45
02:14:33 Join Bloody_Sorcerer [0] (
02:14:44Bloody_Sorcererfor the rockbox VMWare dev platform, whats the login
02:15:05scorcheBloody_Sorcerer: user:rockbox
02:15:38scorchesorry...that was it
02:15:48scorcheuser: user pass: rockbox
02:15:58Bloody_Sorcererah, there we go :)
02:17:09 Quit webguest82 ("CGI:IRC")
02:18:44 Join JdGordon [0] (
02:19:54 Join lodesi [0] (
02:20:11 Join andrewy [0] (n=boo@
02:20:44andrewydoes anyone have use an iriver h320 with rockbox? i'm wondering how well it works, as in, if it's useable
02:22:48JdGordonits 10000x more usable than the iriver software
02:23:27andrewyjust out of curiousity, does it show the remaining battery life like the h120?
02:23:32andrewyor is that not implemented
02:23:34Genre9mp3add some more zeros please....
02:23:55JdGordoneverything that works on the 120 works on the 320
02:24:44andrewyah, cool
02:25:06Bloody_Sorcerereverything +1
02:25:35Bloody_Sorcerermeaning even better
02:25:39andrewywhat's the 1?
02:26:06andrewydon't think all my music will fit on a 320 though :/
02:26:15Bloody_Sorcererisnt there a 340?
02:26:29andrewy3x as expensive
02:27:51 Quit t0mas (Read error: 104 (Connection reset by peer))
02:28:18 Join t0mas [0] (n=Tomas@rockbox/developer/t0mas)
02:31:05Bloody_Sorcerermake zip is puking size constraints errors
02:31:07andrewyhave you tried any of the things with it? i think there are patches to write the songs listened to to a file for submitting to them
02:31:09Bloody_Sorcererwtf does that mean :(
02:31:29 Quit lodesi_ (Read error: 110 (Connection timed out))
02:31:53*Bloody_Sorcerer pretends it doesnt matter
02:35:11Bloody_Sorcerermpegplayer plugin = sick nasty
02:39:23 Quit lee-qid (Read error: 110 (Connection timed out))
02:48:58MikachuBloody_Sorcerer: those are not errors
02:52:58jhMikeSMikachu: I meant to ask about those. Why are the rwps's not being included?
02:58:29SoapIf I can make a comment about an earlier conversation:
02:58:31Soap[2006-08-24 15:56:03] <jhMikeS> linuxstb: point is (I think I have one :)...a string of short files will load the codecs and files up to available space...long files won't have a significant relative load from loading codecs.
03:00:24SoapWould a possible solution be (if this possibility (lots of mixed codec short files) is worth being concerned over) to change the code so that a codec is not removed from the buffer until a combination of time/tracks played has elapsed?
03:00:48SoapSo the codec would get removed from the buffer, but only after a wait period.
03:01:10SoapI'm more curious about the idea in theory, than as a pratical need.
03:02:19 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
03:03:04 Quit BHSPitLappy (Read error: 110 (Connection timed out))
03:03:56jhMikeSSoap: I'm not certain there would be much benefit. Like was also said earlier, codecs are small. Things would get messy IMO.
03:04:23 Quit linuxstb (Read error: 110 (Connection timed out))
03:06:57jhMikeSSoap: Most audio files approching the size of codecs are very short indeed and would buffer along with the codecs very quickly.
03:23:26 Quit spiorf (Remote closed the connection)
03:38:11 Join dan_a [0] (n=dan-mirc@
03:40:46 Join sharpe [0] (
03:44:34***Saving seen data "./dancer.seen"
03:47:21 Join webguest82 [0] (
03:48:56 Quit webguest82 (Client Quit)
03:57:22 Quit Genre9mp3 ("I don't suffer from Rockbox psychosis. I enjoy every minute of it.")
03:57:29 Join ni1s [0] (
03:59:22 Quit solarflare_ ("Leaving")
04:02:12 Quit Bloody_Sorcerer (Read error: 113 (No route to host))
04:02:48 Quit dan_a (Read error: 110 (Connection timed out))
04:09:37 Join JoeBorn [0] (
04:10:49 Quit [HO]vo|t ("( :: NoNameScript 4.02 :: )")
04:16:12 Part pixelma
04:27:37 Join XxBigP123xX [0] (
04:28:19XxBigP123xXuh i know what your all going to say but is anyone familiar with ipod linux(good with it)
04:31:06XxBigP123xXgeuss not
04:35:14 Join ryanpg [0] (
04:35:36ryanpgwow, busy channel
04:36:06ryanpgthe website seems a touch out of date... is there any recent progress on the iaudio port? like M5 support :D
04:37:47ryanpgreading the ml now actually... no mention of the M5
04:38:07XxBigP123xXin pacbox can i change the names of the little ghosts?
04:38:38Mikachuyou can choose between two sets of names in the settings iirc
04:41:52Mikachuthis is impressive
04:41:56Mikachu% for a (*) { -b $a}
04:41:58MikachuFile "cover.jpg.ogg" created successfully
04:43:01 Quit ni1s (Remote closed the connection)
04:43:35XxBigP123xXyeah but wen i choose the names it like aaaaaaaaaaaaaaaa, bbbbbbbbbbbbbb, ccccccccccccccccccc,dddddddddddddddddddddd
04:46:21sharpeXxBigP123xX: That's what they're supposed to be.
04:48:14ryanpghrm... different question, different direction, what player is recommended if I want the best rockbox experience?
04:49:28XxBigP123xXo ok sharpe thanks
04:56:36blxMikachu, are you transcoding between lossy formats?
04:57:03 Join Old45 [0] (
04:58:21 Quit Old45 (Client Quit)
04:58:46 Join Old45 [0] (
04:58:47 Quit jaebird (Read error: 104 (Connection reset by peer))
04:59:39 Quit CriamosAndy ("( :: NoNameScript 4.02 :: )")
05:06:11 Quit Old45 ("CGI:IRC (EOF)")
05:14:53rotatordoes anyone have any flac files without seektables that I may borrow?
05:15:52 Join jaebird [0] (
05:16:11XxBigP123xXin podpaint(ipl) do i have to keep tapping the select button to draw a little dot?
05:21:26scorcheXxBigP123xX: what makes you think that we would be able to answer questions about ipl?
05:22:54XxBigP123xXwell, actually idk but i thought id give it a try =)
05:23:54scorchei would think that you would get much better responses in...hmmm...i dont know...maybe ipl's channel rather than the rockbox channel?
05:24:05XxBigP123xXi cant find it
05:24:30scorchehow would i know?...try going to ipl's site and going fromthere
05:39:11sharpeXxBigP123xX: #ipodlinux it's not that hard to find...
05:44:37***Saving seen data "./dancer.seen"
05:45:44XxBigP123xXyeah but im using firefox and wen i press it it says it cannot open it
05:46:57 Join Paul_The_Nerd [0] (
05:47:09XxBigP123xXand frget that now have another question
05:47:44midkayhuh? you press what?
05:47:56XxBigP123xXhow do i load diffrent wads in doom
05:47:57midkaywhy can't you just join #ipodlinux?
05:48:18midkayah, webclient.
05:49:05Paul_The_NerdWebclient isn't an excuse for posting in the wrong channel. It's not like there aren't fifty billion free ways to connect to IRC.
05:49:28*Paul_The_Nerd walks in and talks about something he has no basis to interject in.
05:49:47XxBigP123xX^^^^ i have no idea what you two are talking about
05:50:13midkayPaul_The_Nerd: i was asking why he couldn't join. i agree that he should not ask here.
05:50:36midkayXxBigP123xX: rockdoom? it's in the manual, the wiki, several forum threads..
05:50:44scorchePaul_The_Nerd: we have a have no excuse for this behavior =P
05:50:51XxBigP123xXo ok thanks
05:50:59Paul_The_Nerdscorche: I have an excuse, it's just not a *valid* excuse
05:51:32scorchePaul_The_Nerd: well shame on you!
05:53:24sharpei have excuses for everything...
05:53:48scorchesharpe: ok...what is your excuse for butting into a conversation that had nothing to do with you?
05:55:47sharpewell, it was two minutes after a thing had been said.
05:56:05midkaysharpe: and what's the one for being a total bitch 24/7?
05:56:20sharpeaww. that's not nice.
05:57:13scorchesharpe: why not? ARE a dog =)
05:57:21scorchegender confused one...but still
05:58:16sharpeyeah. that's really funny.
05:59:13 Join RoC_MM [0] (
05:59:13midkayit actually was, but that's beside the point - what was the excuse for it?
06:00:17 Quit Drkepilogue ("CGI:IRC (EOF)")
06:07:20 Join _jhMikeS_ [0] (
06:12:21 Join dongs_ [0] (
06:12:33 Quit dongs (Read error: 104 (Connection reset by peer))
06:22:55 Quit jhMikeS (Read error: 110 (Connection timed out))
06:33:54 Quit dongs_ (Read error: 110 (Connection timed out))
06:37:52 Join matsl [0] (
06:40:03 Join BigMac [0] (
06:40:11 Quit XxBigP123xX ("CGI:IRC (EOF)")
06:40:27 Quit rotator ("zzzzzzzzzz")
06:40:32BigMacPaul_The_Nerd, are you present?
06:41:58BigMacare you a member of
06:42:57BigMacwhat is your forum nickname?
06:43:03BigMacllorean right?
06:43:19BigMacok nvm
06:43:32BigMacthere is a person with lloorean as a nick
06:53:17 Quit daurnimator (No route to host)
06:56:45 Quit merbanan (Remote closed the connection)
07:00:41 Quit juxtap (Read error: 110 (Connection timed out))
07:08:39 Join RoC-MM [0] (
07:12:21 Quit sharpe ("Leaving")
07:23:16 Quit RoC-MM ("Leaving")
07:31:46 Quit t0mas (Read error: 104 (Connection reset by peer))
07:31:49 Join t0mas_ [0] (
07:44:40***Saving seen data "./dancer.seen"
07:46:43Bg3rmorning :P
07:54:47amiconnIt seems we exceed the LCD timing specs on H300 :(
07:58:23Bg3rhuh ?
07:58:30Bg3rmain LCD?
07:59:32amiconnThe datasheet says: write cycle time >= 150ns, low pulse width >= 40ns, high pulse width >= 70ns
08:00:18amiconnWe're running /CS1 with 2 waitstates at 124MHz. Busclk = CPUclk/2, so one bus clock cycle is ~16ns
08:01:08amiconnA non-burst bus cycle is (3+ws) clock cycles, of which (2+ws) clock cycles /WR and /CS is held down
08:02:04amiconnSo the low pulse width is within specs: 4*16ns = 64ns >= 40ns. We could even shave off one waitstate.
08:03:02amiconnBut: There's sometimes only one other instruction between 2 consecutive writes. So our high pulse width is only 1 or 2 bus clocks, i.e. 16 or 32ns
08:03:30amiconn(not sure whether the instruction is shadowed by the write pipeline or not)
08:03:33 Quit dwihno (Read error: 54 (Connection reset by peer))
08:03:44 Join MarcoPolo [0] (
08:03:52 Join dwihno [0] (n=dw@
08:04:30 Quit MarcoPolo (Remote closed the connection)
08:04:40amiconnThat means both the high pulse width (16 or 32ns, need >= 70ns) and the cycle time (80 or 96ns, need >= 150ns) are out of specs
08:05:05 Join MarcoPolo [0] (
08:05:13amiconnThat is, if I don't miss something important...
08:07:55Bg3ramiconn: since when is this time exceeding ?
08:08:17amiconnSince the asm optimised data transfer
08:08:40amiconnSeveral months
08:09:14amiconnOtoh, the overall transfer speed lies in the expected range with obeying the timing.
08:10:10amiconnA cycle time of 150ns means the theoretical limit is 6.67MPixel/s or 13.33MByte/s
08:10:21amiconnWe transfer ~10MByte/s
08:10:42amiconn(of course that includes the DRAM read)
08:11:21*amiconn could need LinusN to solve this mystery :/
08:14:44Bg3rbut i haven't observed any problems ...
08:17:42amiconnI didn't observe any either
08:18:55amiconnI only found this because I was calculating waitstates & stuff for every possible CPU clock, which also involves the LCD waitstates...
08:22:13 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
08:22:15 Join Siku [0] (
08:25:40 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
08:27:28 Join BHSPitLappy [0] (
08:35:11 Quit RoC_MM (Read error: 110 (Connection timed out))
08:39:24 Join Drkepilogue [0] (
08:39:42Drkepiloguefoa question for the coders
08:39:56DrkepiloguecaN U WRITE C PROGRAMS WITH UNIx?
08:40:00Drkepiloguesrt for cap
08:41:52Bg3rDrkepilogue: some people here can ....
08:42:25Drkepiloguei mean, is it possible to use unix to write c programs
08:42:26zei dunno, can you write BASIC programs on a c64?
08:42:26*Bagder doesn't understand the question
08:42:35BagderDrkepilogue: yes of course
08:42:40BagderC was made on and for unix
08:43:19Drkepilogueim not smrt about these things
08:43:25 Join ender` [0] (i=null@
08:45:13 Quit Drkepilogue ("CGI:IRC (EOF)")
08:54:35 Join Rob2222_ [0] (
09:01:34 Quit JdGordon ("using sirc version 2.211+KSIRC/1.3.12")
09:02:07 Join [H5N1] [0] (
09:03:07 Quit Rob2222 (Read error: 60 (Operation timed out))
09:05:00daurnimatorhi al
09:08:26 Nick pabs_ is now known as pabs (
09:11:42 Join webguest84 [0] (
09:13:08 Join _FireFly_ [0] (
09:13:19webguest84Hi. There is a typo in the RB manual. In the section about "A-B" feature, it reads "This option is more complicated to use THAT the others". It should read "...THAN the others"
09:13:51midkaythanks for pointing out one of surely hundreds. :)
09:13:59midkaywe'll consider fixing this. very intensely.
09:14:22 Join daurn|laptop [0] (n=quae@
09:15:40 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:21:05Paul_The_Nerdwebguest84: Typos should be reported on the patch tracker, just like anything else.
09:21:45*amiconn is surprised about midkay's reaction
09:23:43*amiconn had the impression that midkay is usually very picky about spelling & grammar
09:24:12midkayi am, what's your point? it was sarcasm :0
09:25:15amiconnmorning LinusN
09:25:48*LinusN gets some black goo in a cup
09:25:55amiconnLinusN: Could you help solving the LCD timing mystery (H300/X5)? See today's log starting 07:54...
09:26:00webguest84Paul_The_Nerd: yes, I know, that's the official way. But if the authors are reading here it will be quicker and "lighter" way. BTW: IMHO, "A-B" should be written using a dash, not "−−" in TeX.
09:26:19 Quit mirak ("Ex-Chat")
09:27:01 Join linuxstb [0] (
09:27:23linuxstbwebguest84: No need to file a bug report - I've just fixed it. Thanks for reporting.
09:27:35 Join dan_a [0] (n=dan-mirc@
09:30:37LinusNamiconn: you wonder why it still works even if we exceed the timing specs, or what?
09:31:04amiconnYes, and whether we actually exceed the specs or not
09:31:18amiconnSome numbers are simply not matching well...
09:31:20 Join tucoz [0] (
09:31:55 Quit linuxstb ("CGI:IRC")
09:32:27amiconnOn X5 it's even worse. Minimum cycle time is 200ns instead of 150ns (HWM=0), but we're writing at the same speed as on H300
09:32:30LinusNwe all know that the h300 lcd is timing sensitive, but maybe the x5 controller is different
09:32:50amiconnIt is?
09:33:02tucozwebguest84: what did you mean about the dash and A-B?
09:33:04LinusNwe also know that some h300 users have problems
09:33:25LinusNnot many, but some
09:33:26*amiconn didn't know that
09:33:39LinusNit's in the forums
09:34:27tucozAs far as I know, it is written as A-B everywhere. Not A−−B. Or did I misunderstand you?
09:35:33LinusNspeaking of timing, hwm mode is out of the question, unless we can make the coldfire hold CS down during the entire transfer
09:35:42amiconnHmm. Random pixels would indicate exceeded timing
09:36:07amiconnBut according to my cals, we're _significantly_ exceeding timing specs
09:36:30amiconnSomething must be wrong, as I wouldn't expect it to still work if that's _really_ the case
09:36:56LinusNi believe the lcd timing has generous margins
09:37:18amiconnBtw, why is HWM out of question? The controller datasheet doesn't _require_ /CS held low during the 4 consecutive transfers
09:37:45amiconnThe diagram shows it that way, but it's nowhere required in the text or tables
09:38:22Bagdertucoz: the ipod3g manual doesn't build for me...
09:38:31LinusNamiconn: don't you remember that we tried this on the h300, and it only transferred every 4 pixels
09:39:08LinusNa loooong time ago
09:39:28amiconnDidn't know that...
09:39:30tucozBagder: ok. Maybe I forgot something. I´ll check it out later.
09:39:57 Join pondlife [0] (
09:40:58 Quit matsl (Read error: 110 (Connection timed out))
09:40:58 Quit tucoz ("CGI:IRC (EOF)")
09:42:25*LinusN broke 5 wires in his x5 lcd strip the other day
09:43:02LinusNbut i managed to repair them by soldering
09:43:25LinusNthis means that i can't open it up and measure the lcd timing, because then i'll break it for good
09:43:46webguest84 tucoz: no, it's written as "A−−B", i.e. with two dashes (at least at one place, in the A-B section)
09:44:44***Saving seen data "./dancer.seen"
09:46:10 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
09:46:39 Join linuxstb__ [0] (
09:46:43 Nick linuxstb__ is now known as linuxstb (
09:47:01 Join BHSPitLappy [0] (
09:47:02 Quit linuxstb_ (Nick collision from services.)
09:47:18 Part Paul_The_Nerd
09:49:00 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
09:49:49 Join BHSPitLappy [0] (
09:50:23 Part webguest84
09:50:24linuxstbdan_a: Are you around:
09:50:51 Join webguest84 [0] (
09:51:22 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
09:52:13 Join BHSPitLappy [0] (
09:56:14 Join BHSPitLappy_ [0] (
09:56:33 Part BHSPitLappy ("Leaving")
09:56:50 Nick BHSPitLappy_ is now known as BHSPitLappy (
10:01:53 Join tucoz [0] (n=martin@rockbox/developer/tucoz)
10:02:38dan_alinuxstb: Not for long. I've (just) figured out where the stack overflow is coming from
10:02:52tucozBagder, you can try now
10:03:33linuxstbdan_a: If I remove the sleep() in apps/main.c:init(), then it doesn't panic any more.
10:03:34dan_aChange the sleep() in app_main to a udelay or a load of nops
10:03:45linuxstbYes, I've done that...
10:03:55Bagdertucoz: thanks, seems to build fine now
10:04:01webguest84tucoz: have you seen my comment about two dashes in "A-B" section? I don't like to file a bug.
10:04:35dan_aIt needs long enough for the COP to wake up and set the variable to say that it's alive before the CPU decides whether to sleep it again or not
10:04:41linuxstbdan_a: I'm still debugging - I can call create_thread_on_core(CPU,....) from mpegplayer and successfully decode the video in a new thread, but the thread refuses to run if I start it on the COP.
10:04:49tucozwebguest84, I saw your comment, but not a reply to my question. (me checks logs)
10:05:19dan_alinuxstb: Does it show up in the OS Stacks debug screen
10:05:33tucozwebguest84, yes. But I didn't understand what you mean. We write A-B as A-B, not A−−B
10:06:13dan_aYou should have a main thread and your thread on core 1
10:06:31tucozwebguest84, oh. you are talking about the A-B feature, not the button right?
10:06:44webguest84tucoz: yes
10:06:47linuxstbdan_a: Just testing now
10:07:06tucozaha. I am not sure why it is written like that. Let me check that.
10:07:50 Join Paul_The_Nerd [0] (
10:08:08linuxstbdan_a: I don't have a main thread on the COP - just my own thread.
10:08:53dan_alinuxstb: Try commenting out the line that sends the COP back to sleep (after the sleep that you deleted)
10:09:03Bagdertucoz: 3g manuals are online
10:09:19tucozthanks :)
10:11:22linuxstbdan_a: When I first start Rockbox, should the OS Stacks screen show a main thread on the COP?
10:11:45dan_alinuxstb: Yes
10:11:47tucozwebguest84: to be honest, I have no idea why it is written using two hyphens. Maybe the writer simply wanted a longer hyphen?
10:11:56linuxstbdan_a: OK, that's not happening.
10:12:18Paul_The_NerdWhat happens if something tries to un-boost the CPU when it's not boosted?
10:12:33dan_aEven when you delete the line that sends it back to sleep?
10:12:53firenxhmm i got this fm transmitter but my sound sounds all muffled.. how can i reset the rockbox settings?
10:13:13dan_aI have to go now - I'll be back online in about an hour. My guess is that I've got the interrupt addresses wrong
10:13:27linuxstbdan_a: I'll keep investigating - thanks.
10:14:35webguest84tucoz: I think so. But I also thing that the desire was wrong :-)
10:14:36 Quit apo` (Read error: 104 (Connection reset by peer))
10:14:46LinusNfirenx: what makes you think it has anything to do with the rc settings?
10:15:02LinusNfirenx: which player model do you have?
10:15:03linuxstbPaul_The_Nerd: Nothing. The boost counter will get decremented, but there is a check to make sure it never goes below zero.
10:15:07firenxwell it sounds crappy on my headphones too so i think its something i did
10:15:09firenxipod video 30gb
10:15:20tucozdo you think using only one hyphen is the proper British way of writing it?
10:15:44Paul_The_Nerdlinuxstb: Okay. I had assumed that was the case.
10:15:59linuxstbtucoz: Shouldn't "A-B" the feature be something like "A<->B" ?
10:16:29Paul_The_NerdWell, isn't it "A to B" since it doesn't bounce, it loops. So maybe A->B?
10:16:52tucozlinuxstb, maybe. I don't know what it is called in Rockbox. I can not find my charger, and the h120 is without juice atm.
10:18:16tucozbut I agree that it shouldn't be two hyphens now.
10:18:28firenxok so i put a song on the apple os and it sounds good
10:18:33firenxso it must have been something i did with rockbox
10:19:10LinusNfirenx: start it with the hold button on
10:19:26LinusNthat should reset the rc settings to their default values
10:19:35LinusNrb even
10:19:41tucozwebguest84, fixed. thanks for reporting
10:20:27firenxhmm still doesn't sound quite right
10:20:56firenxall my settings look the same tho.. time is right, theme is right
10:21:16webguest84tucoz: yes, I would do so. Wikipedia states something different though. Hmm... I think we should leave it as it is then. Or the other way :-)))
10:22:23tucozIf it is called A-B in Rockbox, then it should be that (Imo). Not what wikipedia says. It is a setting after all.
10:24:10Paul_The_Nerdfirenx: Did it say "Settings Cleared" when you booted?
10:24:35firenxi dont think so.. anyways i deleted it and put rockbox back on.. sounds proper now :)
10:25:05firenxbut thats cool i had been messing with all the settings to see what they do anyways hehe
10:25:47webguest84tucoz: "...It is a setting after all." OTOH A-B as a feature means "play from A to B". And "X to Y" is written with double dash (en dash). I don't know. We should ask in a typesetting user group :-)
10:27:07tucozwebguest84, hmm. that is probably why it was written like that. maybe i was a bit quick with that commit.
10:27:07 Quit San||Away (Read error: 104 (Connection reset by peer))
10:27:22 Join San [0] (
10:31:27 Join apo` [0] (
10:31:56webguest84tucoz: after having thought about this I still think it should have only one dash. The button has one dash, in the menu we also have one dash. So, to be consistent, we should also have only one dash in the feature name. Even if it's slightly incorrent semantically.
10:32:30tucozIs it 'The build contain a file and a directory ' or 'The build contains a file and a directory' ?
10:32:40webguest84tucoz: one more point: "Repeats between two user defined point within a track" should read "...POINT_S_..."
10:32:43tucozwebguest84, ok. we'll do that for now then.
10:33:03 Quit dan_a (Read error: 110 (Connection timed out))
10:33:20webguest84tucoz: I'd say "the build contains"
10:33:46linuxstbOr "the build zip file contains" - if that's what you're talking about.
10:34:35 Join San||Away [0] (
10:34:35 Quit San (Read error: 104 (Connection reset by peer))
10:34:41 Join dan_a [0] (i=dan_a@
10:38:53 Join San [0] (
10:38:53 Quit San||Away (Read error: 104 (Connection reset by peer))
10:45:37tucozlinuxstb, are the install instructions for the 3g equal to that of the 4g?
10:46:23linuxstbAlthough the hardware key combinations are different on the 4g - see
10:46:41tucozand the file is called bootloader-3g.bin?
10:48:19tucozhmm. sometimes it is 'harddrive' and sometimes 'hard drive'. what should i use?
10:48:45dan_aI think 'hard drive' is correct
10:49:07linuxstbdan_a: It looks like cop_main() is never called...
10:49:54dan_alinuxstb: That's... odd. There's nothing in crt0.S that I've made PP5002 only, is there?
10:50:07 Quit daurnimator ("Changing server...")
10:50:07linuxstbI can't see anything.
10:50:13 Join bluebrother [0] (
10:51:18 Join San||Away [0] (
10:51:18 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
10:51:18 Quit San (Read error: 104 (Connection reset by peer))
10:51:25tucozbluebrother, hi :) did i do something stupid when i changed the A−−B to A-B?
10:51:27dan_aThe way to test that would be to have the first line of cop_main send the cop to sleep, and then check from the CPU what its status is
10:51:30 Quit San||Away (Read error: 104 (Connection reset by peer))
10:52:01bluebrothertucoz, no.
10:52:07tucozok. good
10:52:16webguest84tucoz: have you also changed 'point' to 'points' (see above)?
10:52:41bluebrotherit just produces a longer dash, but maybe that dash is even better in that place.
10:52:48 Join San [0] (
10:53:29tucozwebguest84, ah. right. no. i was going to ask you if you wanted to sum up what you think should be changed, and we can fix several things in one go.
10:54:08dan_alinuxstb: How are you testing if it gets called?
10:54:28webguest84tucoz: then I'd have to file a bug report. And I'm too lazy for that :-)
10:55:06tucozyou could write down what you find, and paste it in a paste bin then :)
10:56:10LinusNamiconn: in the metronome plugin, the timer_register() succeeds, but the timer callback is never called
10:56:26LinusNhowever, if i rolo rockbox, is works
10:56:27 Quit San (Read error: 104 (Connection reset by peer))
10:57:04 Join San [0] (
10:58:10amiconnLinusN: Yes, and it also works if you exit metronome and start it again
10:58:10 Quit San (Read error: 104 (Connection reset by peer))
10:58:16LinusNsaw that too now
10:58:18linuxstbdan_a: I've put a variable (cop_count) in IRAM, and I increment that variable at the start of cop_main() and display it in the debug OS stacks screen.
10:59:03amiconnBut the doom hang must have a different cause. Doom also hangs if you select the tick emulation by current_tick instead of using TIMER1
10:59:27 Join San [0] (
10:59:35LinusNamiconn: yes
10:59:40amiconnBtw, it could be possible to keep /CS low for 4 consecutive LCD writes
10:59:45dan_alinuxstb: You still have COP_CTL = PROC_WAKE in init()?
10:59:49LinusNamiconn: how?
11:00:01amiconn/CS can be switched to GPIO
11:00:02 Quit San (Read error: 104 (Connection reset by peer))
11:00:30amiconnThen all code between this switch has to run in IRAM, must not access data outside IRAM, and disable interrupts
11:00:37LinusNamiconn: but how do you set cs high again without generating additional cycles on the bus?
11:00:43linuxstbdan_a: Yes.
11:00:50dan_aAnd have you tried taking out the 'if(!cop_is_available) COP_CTL = PROC_SLEEP'?
11:01:27amiconnOr does accessing MBAR also generate bus cycles on the external bus?
11:01:27tucozwebguest84, the point_S is fixed
11:01:32 Join spiorf [0] (
11:01:37LinusNamiconn: no, i don't think so
11:01:50linuxstbdan_a: Yes, that's commented out.
11:01:56 Join San [0] (
11:06:22LinusNbut i'm not sure that R/W is silent
11:06:22amiconnDunno whether it's worth the hassle; accessing MBAR is _dea_ slow
11:06:22webguest84tucoz: OK
11:06:22amiconnA write needs 12 cpu cycles, read-modify-write (or/and/eor) needs 25 cpu cycles
11:06:22 Quit _FireFly_ ("Leaving")
11:06:22DBUGEnqueued KICK amiconn
11:06:22amiconnLinusN: Could you perhaps measure timing on H300 then?
11:06:22 Quit San (Read error: 104 (Connection reset by peer))
11:06:22amiconn(re the broken wires on your X5)
11:06:22LinusNamiconn: perhaps
11:06:22 Join San [0] (
11:07:33dan_alinuxstb: What does '@' in front of a line of assembler do? As in '@ldr r4, =PP5020_COP_CTRL'
11:08:46linuxstbdan_a: I think that's a comment...
11:08:53 Quit pike (Read error: 54 (Connection reset by peer))
11:08:56 Join pikester [0] (
11:09:06dan_aAh. I think you need a new bootloader then!
11:10:47 Join San||Away [0] (
11:10:47 Quit San (Read error: 104 (Connection reset by peer))
11:11:37 Join XavierGr [0] (
11:13:38 Join San [0] (
11:13:38 Quit San||Away (Read error: 104 (Connection reset by peer))
11:15:17tucozamiconn, do you think the mpeg player should be ported to the grayscale targets?
11:15:21 Quit San (Read error: 104 (Connection reset by peer))
11:15:51 Join San [0] (
11:15:54Paul_The_NerdIs there really a reason *not* to, assuming anyone wants to bother with doing it?
11:16:17 Join Genre9mp3 [0] (
11:16:41tucozI don't think there is a reason not to. Just wanted to ask.
11:16:41 Quit San (Read error: 104 (Connection reset by peer))
11:17:07tucozand in that case, what needs to be done
11:17:47amiconnOn the grayscale targets, RVF makes more sense
11:17:53 Join San [0] (
11:18:42tucozamiconn, isn't it possible to transcode an mpeg to grayscale?
11:19:54amiconnYes, on a PC
11:20:14CtcpIgnored 2 channel CTCP requests in 0 seconds at the last flood
11:20:14*Genre9mp3 wonders why rvf is only for Archos
11:20:33amiconnGenre9mp3: Because nobody ported video.rock yet?
11:20:45amiconnWett was working on it, then vanished :(
11:20:50Bagderwell someone once made one for h1x0 but never contributed any code
11:21:09Genre9mp3Ah ok....I thought maybe there was a reason for that
11:21:34tucozhehe. the joy of not posting your hard work for others to finish
11:22:17Paul_The_NerdIn my mind, RVF making more sense is just reason to get RVF support working _first_ rather than doing it only. It'd still be nice to have mpeg support on the grayscale targets I think, just because it's a standard format. Couldn't you, in the case of color files, just display the Y and not have it look *too* bad?
11:22:24Genre9mp3BTW the video on rvf uses some special codec or is it just uncompressed 2bpp frames?
11:22:58Bg3rGenre9mp3: in fact it's 1bpp atm
11:23:03Bg3rand mp3 audio
11:23:49linuxstbdan_a: That didn't help - I think r4 was already containing that address anyway.
11:24:52Genre9mp3Bg3r: I understand why it looks like random dots in the screen when its paused
11:27:00dan_alinuxstb: I wonder where the COP is going when you wake it up, then? I'm just reading through crt0.S to make sure there's nothing I've missed.
11:28:34linuxstbPaul_The_Nerd: IMO, rvf will give better quality than MPEG on the greyscale targets - you're avoiding all the artifacts of MPEG encoding. So yes, MPEG support might be nice, but as you need to re-encode everything anyway in order to view it in Rockbox, why not just reencode to RVF?
11:29:23amiconnGenre9mp3: rvf is uncompressed 1bpp frames at the lcd's internal refresh rate
11:29:44amiconnAudio is mp2/mp3 (could also be another codec for swcodec targets)
11:29:44 Quit San (Read error: 104 (Connection reset by peer))
11:30:04 Join San [0] (
11:30:08amiconnBg3r: It will also be 1bpp for the 2bpp targets
11:30:26xorAxAxwhats the bitrate of rvf?
11:30:30Genre9mp3amiconn: I'm impressed about the fact that is 1bpp but looks so good!
11:30:31amiconnPixel flipping and the lcd's internal greyscale generation don't mix well
11:31:03amiconnxorAxAx: Depends on the lcd resolution and internal refresh rate. For archos it's 480kbps + audio bitrate
11:31:13xorAxAxthats quite much
11:31:20Paul_The_Nerdlinuxstb: In the case that it is able to playback color videos in grayscale, because a video can then be shared between users of same-sized screen players. I just think it'd be convenient, for example, if someone with an X5 could share a video with someone with an H120 even in a situation where transcoding software isn't available.
11:31:21xorAxAxfor such a small screen
11:31:33xorAxAxnobody tried using compression?
11:31:38xorAxAx(even lossless)
11:31:51amiconnCompression is simply out of question on archos
11:32:00 Quit BHSPitLappy (Read error: 110 (Connection timed out))
11:32:20Paul_The_Nerdlinuxstb: But remember, I'm not saying I think anyone should work on this. I'm saying that I'd be happy if someone ever did, as I think it'd be nice. I just like the idea of it.
11:32:23amiconnPumping the frames to the lcd already consumes 50% cpu (because the transfer is serial bit-banging - highly optimised btw)
11:32:41amiconnAnd the frames won't compress well
11:32:49 Quit San (Read error: 104 (Connection reset by peer))
11:33:03 Join BHSPitLappy [0] (
11:33:08 Quit XavierGr (Read error: 110 (Connection timed out))
11:33:23amiconnPaul_The_Nerd: Transcoding live and using the grayscale lib will produce worse quality at higher cpu load
11:33:26 Join San [0] (
11:33:28xorAxAxamiconn: does it work on ipods as well?
11:34:05Paul_The_Nerdamiconn: Would that really be transcoding?
11:34:05 Quit San (Read error: 104 (Connection reset by peer))
11:34:17*Paul_The_Nerd is not at all familiar with video.
11:35:08xorAxAxamiconn: so the main archos cpu is constantly feeding serial bitstream data into the lcd controller while displaying menus etc.=
11:35:35 Join San [0] (
11:37:17amiconnNot constantly, only when the lcd is updated
11:38:48 Join San||Away [0] (
11:38:48 Quit San (Read error: 104 (Connection reset by peer))
11:39:03dan_alinuxstb: What happens if you comment out HAVE_ADJUSTABLE_CPU? The COP seems sensitive to that, but it works OK on the 3G now
11:40:54 Join San [0] (
11:40:54 Quit San||Away (Read error: 104 (Connection reset by peer))
11:40:57 Quit San (Remote closed the connection)
11:41:50pondlifeHmm, a Flyspray request here... is there any way to sort a list (e.g. bug reports) in date last edited/updated order?
11:42:19pondlifei.e. see those where new comments have been added recently at the top
11:43:24linuxstbI don't know, but you can subscribe to the rockbox-sf mailing list and get the comments emailed to you.
11:44:00Bagderor read the rss feed
11:44:13pondlifeFrom FlySpray?
11:44:45***Saving seen data "./dancer.seen"
11:45:13Bagderyes, those are sent from flyspray
11:45:24pondlifeThanks, I assumed that was just for the old SourceForge setup.
11:45:35Bagderthe name is _slightly_ misleading ;-)
11:45:46pondlifeJust a tad..!!
11:46:23pondlifeGreat, that's just what I was after
11:46:26bluebrotherpondlife, just click on the table headings like in windows explorer
11:46:38pondlifeYes, but there's no "last updated" heading
11:46:42pondlifeThat was my first thought
11:47:02pondlifeOr can I customize that?
11:48:51 Join shirour [0] (n=mrobefan@
11:49:01tucozBagder, is there a cvs rss feed as well?
11:49:06shirourhi there.
11:49:25tucozBagder, thanks :)
11:49:27shirouri'm a dev. of olympus m:robe 500i....
11:49:41shirourwe've been able to compile C code and run it
11:49:47Bagdershirour: cool!
11:49:55shirourwe wrote remote, lcd and a few hd drivers
11:50:12shirourand now i'm considering porting rockbox to it...
11:50:24shirouryou can check it out at:
11:50:25LinusNshirour: wow!
11:50:25Bagdershirour: we'll love you
11:50:47shirourwho here can help me start this project?
11:50:58Bagderwe're many
11:51:13Bagdershirour: how do you start your own code?
11:51:21shirourfor information about what we've done, check out:
11:51:31shirourbadger - with an exploit
11:51:48Bagderseems to be a common way these days
11:51:54Bagderthe same as the medios guys use?
11:51:59 Join muesli__ [0] (n=muesli_t@
11:52:03 Quit muesli__ (Client Quit)
11:52:07shirournot the same
11:52:09shirourbut very much alike
11:52:16shirourthe exploit is different
11:52:37Bagdershirour: anyway, the way most our ports work is that we have a bootloader for the initial startup, and that bootloader loads rockbox from disk and starts it
11:52:37shirourm:robe 500i is based upon TI's DM320
11:52:44Bagdershirour: the gmini402 is as well
11:52:58Bagderwhich the medios runs code on
11:53:02shirour- i already wrote something that loads a software from the HD and runs it
11:53:19Bagderthen you should be able to get rockbox going pretty easily
11:53:27shirouri've managed to run mediOS
11:53:45shirourwell Badge, i wish... right now i want to know how to start...
11:53:50 Join ni1s [0] (
11:53:53Bagdershirour: 1 get the rockbox code, 2 add your target in tools/configure and a new firmware/export/config-*.h file
11:54:07Bagder3 select to build a bootloader
11:54:10markunshirour: that's great!
11:54:12Bagdermake sure (3) builds
11:54:36Bagderthen there's nothing stopping you from going to a full-fledged rockbox
11:54:53shirouri've downloaded the latest CVS version...
11:54:53markunshirour: did you tell the archopen guys?
11:55:08shirouri've talked to them
11:55:24 Quit BHSPitLappy (Read error: 110 (Connection timed out))
11:55:39shirourbut since we're not completely the same, right now porting from one to the other is still a bit problematic...
11:55:42 Join _FireFly_ [0] (
11:56:01shirourlet me understand - where are all the driver's code located?
11:56:03Bagderyeah, the archopen guys should go rockbox as well... ;-)
11:56:06markunIf they start supporting the m:robe and maybe also the vision:m they should start thinking of a new name :)
11:56:19Bagdershirour: in firmware/ and firmware/drivers
11:56:27Bagdershirour: with target adjustments in firmware/target/
11:56:35shirourso the bootloader takes its includes from there?
11:56:41Bagdercompare with for example how the h10 build is made
11:56:59Bagdersince h10 is pretty new and arm based too
11:57:08shirourmmm... ok... sounds good....
11:57:17tucozI've been thinking of the manual licence. As far as I can tell, the GFDL is subject to the compiled manuals as well as Latex source code. However, it does not include the .svg sources we use for the player images (only the pdf).
11:58:47dan_aBagder: I'm planning to buy myself a present today... a Sansa shaped one
11:59:27shirourwhere can i find docs for the different configuration vars found in tools/configure ?
11:59:27Bagderwelcome to "oh how the heck do we get anything to display on the LCD" project :*)
11:59:49Bagdershirour: only by reading the scripts and code ;-)
12:00:04Bagdershirour: start out cloning the h10 ones and edit those that you notice are wrong
12:00:10dan_aWell, Rockbox is too close to working on the 3G iPod now!
12:00:27tucoz(or maybe the "Transparent" section covers svg's as well.
12:00:56*amiconn considers rockbox still far from release status on all the ipods
12:01:49dan_aamiconn: My comment was a bit tongue-in-cheek - I know there's still a lot to do
12:02:12shirourwhat's the mi4 thing?
12:02:26Bagdershirour: the file format for h10, you don't need that
12:02:43Bagderbut you can instead invoke your own "make firmware" tool there
12:04:04tucozI think I will send an email to the licensing guys at and ask for advice
12:04:19Bagdertucoz: sounds like a good idea
12:05:27shirourmaybe i should use the gigabeatf? it has arm9
12:05:40Bagdershirour: true
12:05:43 Join barrywardell [0] (n=barrywar@
12:05:53Bagderbut you'll notice that they don't differ very much
12:06:32amiconndan_a: Were you able to successfully run test_gray2 in the meantime?
12:06:32Bagdershirour: barrywardell might have some "first stab at rockbox port" hints as he's just walked done the path your walking
12:06:52markunshirour: the interesting ARM9 specific things are not in cvs yet
12:07:06markunlike MMU code
12:07:12shirouroh i see...
12:07:35Bagdershirour: markun is mr Rockbox-on-gigabeat
12:07:35shirouris there a branch for that?
12:07:41*barrywardell goes to check the logs to see what's going on
12:08:10dan_aamiconn: I was never able to get it to be steady, and the buttons didn't seem to exit - so no.
12:08:33amiconnThat's odd, since otoh the grayscale lib works on 3G
12:08:46dan_aI know
12:09:05amiconn...just not properly tuned yet
12:09:54dan_aI'll have a look at the code and give it another go sometime soon.
12:10:05amiconnScrolling back & forth adjusts the cycle. You need to adjust it a bit, then leave the wheel alone
12:10:17amiconnThe timer cycle isn't stable when it's adjusted
12:10:46amiconnWhen you got it stable, Select should end the measuring cycle and display the value
12:10:56amiconnMenu should then quit the plugin
12:11:28dan_aNeither Select nor Menu seemed to work when I tested it
12:11:49shirourwould you recommend to use player:gui first?
12:11:53barrywardellshirour: so you're starting a new Rockboox port
12:13:01shirourbarrywardell: yeah, i want to
12:13:40amiconndan_a: If you make the cycle far too short (requires scrolling forward for quite some time), the plugin will hang in the isr
12:13:43shirouri've been working for a long time on the exploit and drivers, and now i want to have a real software for it
12:13:47barrywardellthere is a page somewhere in the wiki about creating a new build target
12:14:11amiconnBut that's not very likely since I reduced the lcd area used by the plugin, so less data is transferred in the isr
12:16:02bluebrothertucoz, what happened to the license issues mail rasher send to the dev list some time ago?
12:16:34tucozbluebrother, no-one answered him. I discussed it in here with him some time.
12:16:46barrywardellshirour: you'll find it easier to get a bootloader build to work first as it's much simpler
12:16:55 Join Mmmm [0] (
12:16:57shirourbarrywardell: thanks!
12:17:26shirourbarrywardell: so what should i do ? i have a file i've made that loads a binary from the HD, copy it to the RAM and runs it
12:17:53bluebrotherI haven't had time looking on this (but I still want to get around reading and comparing the licenses)
12:18:07dan_aamiconn: It's possible I'd scrolled back too far, but I *think* I remembered to test just entering and leaving the plugin. As I say, I'll retest later.
12:18:25LinusNamiconn: i found the timer bug
12:18:58shirouri've already done most of the things written in the newport...
12:19:06barrywardelldo the things it says in the wiki page. put your drivers in firmware/target/arm/olympus/mrobe or whatever
12:19:11shirouri have my own WIKI which covers it in the URL i gave before
12:19:27barrywardellalso edit firmware/SOURCES
12:19:29LinusNit turned out that the timer was already enabled in TMR, and when that happens, the first write to TMR clears it (because the RST bit does a 1->0 transition)
12:19:36tucozbluebrother, I am writing a mail to ask if our current licence covers our source code, images AND the generated documents.
12:19:54amiconndan_a: Scrolling back increases the cycle, so it shouldn't cause any problem
12:19:57barrywardellshirour: and also bootloader/SOURCES
12:20:25bluebrothertucoz, ok.
12:20:42amiconnLinusN: Huh, interesting. Where does that extra enable happen? It doesn't happen on iriver...
12:20:57LinusNprobably in the cowon bootloader
12:21:11barrywardellshirour: if you look at bootloader/h10.c it just prints a line to the lcd. you could start with that
12:21:22markundaurn|laptop: are you here?
12:21:51linuxstbshirour: What format is the framebuffer for your LCD? RGB565?
12:22:20shirourit has many options... one of them is simple RGB
12:22:42LinusNamiconn: so now i'll see if i can attack the doom hang
12:22:50linuxstbAt least initially, you should use rgb565 in Rockbox - that's the format all the higher-level LCD drawing code works with.
12:23:15shirourno problem
12:23:21 Join dan_a_ [0] (i=dan_a@
12:23:35 Quit dan_a (Read error: 104 (Connection reset by peer))
12:25:04 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
12:25:34 Quit amiconn (Nick collision from services.)
12:25:34 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
12:26:03amiconnLinusN: But why does the RST transition matter? The register setup takes place after that...
12:28:04 Quit Pyromancer (Read error: 110 (Connection timed out))
12:28:12*amiconn is puzzled
12:30:48*Mmmm wonders why amiconn and linuxstb are so against having the samplerate displayed in the statusbar in the WRS
12:31:10preglowwhat clock does that dm320 arm core run at?
12:32:07shirourbut it has a DSP cpu as well
12:32:17preglow200mhz is more than enough to decode audio, though
12:32:20preglowusing the dsp will be hard
12:32:26preglowdue to the apparent lack of tools for ti stuff
12:32:39Bagderthere's an assembler for it
12:32:40shirouryeah... with the dps it can encode video as well...
12:33:05preglowBagder: writing encoders in assembler is really fun, you see
12:33:30BagderDSP details are only available from TI if you sign their NDA
12:33:35*preglow kicks ti
12:33:39preglowbloody idiots
12:34:16preglowall other dsp manufacturers have public specs without it seeming to hurt them noticably
12:34:51shirouryeah... eventough, neuros choose the same SoC
12:35:04preglowthe arm9 things are still armv4t architecture, yes?
12:35:12Bagderlots of companies chose the tms320
12:35:12preglowshirour: yeah, apparently it's the cheapest solution around
12:35:56shirourright... but they'll probably use non open-sourced code for the codecs/encoders
12:36:13Bagdernot for rockbox at least
12:36:37Bagderthe plan is they do the DSP parts "hiddden" and then we do all the rest open and rockbox-style
12:36:39markunshirour: yes, they get the binaries from TI
12:37:00Bagderthey buy their codecs from a 3rd party
12:37:06preglowBagder: so we're still planning to do that neuros thing?
12:37:07Bagderso they're not gpl compatible
12:37:22Bagderpreglow: oh yes, zagor even got a dev board
12:37:30shiroursay - what files do i need to put in target/olympus/mrobe_500i/ ?
12:37:32 Join Arathis [0] (
12:37:35 Quit MarcoPolo (Read error: 104 (Connection reset by peer))
12:37:42preglowwill _all_ our officially endorsed ports work like this?
12:37:48preglowwe get hardware, but no specs
12:38:13barrywardellshirour: you put stuff specific to your hardware there
12:38:18barrywardelllcd driver, etc
12:39:02Bagdershirour: compare how the gigabeat does
12:39:19Bagderyou could even build a gigabeat bootloader and see what it compiles
12:39:45markunIt's not really a bootloader yet though..
12:39:51shirourwhere can i find a file which tells me the file names for the flags i've selected?
12:40:05Bagderfile names for the flags?
12:40:06 Quit webguest84 ("CGI:IRC")
12:41:19shirouryeah, "lcd-{NAME}.h" etc'
12:41:25shirourwhere are the inlcudes?
12:41:49Bagderbootloader/SOURCES specifies what files to build for the bootloader
12:41:55Bagderfirmware/SOURCES for the firmware lib
12:42:57Bagder(as the bootloader links with the firmware lib)
12:43:20shirourbut the symbol of the bootloader is the first to run right?
12:43:32shirouror does it create 2 files ?
12:43:50Bagder the symbol of the bootloader?
12:44:05Bagderthe bootloader is one file
12:44:23Bagderat least all the existing targets only need one file
12:44:58barrywardellmain() in bootloader is what's run first if that's what you mean
12:45:13Bagderwell, strictly speaking crt0.S runs before that
12:46:27Bagderwe need to expand this =>
12:46:50daurnimatorhi badger
12:47:05Bagderdaurnimator: look a fellow tms320-user
12:47:07barrywardellI'll add some extra stuff to it today
12:47:39daurnimatorhe came and talked to me once
12:47:46Bagderits a competition now
12:47:54Bagderwhich of your targets will run rockbox first? ;-)
12:47:57daurnimatorhe got some of our code working on mrobe
12:48:17shirouryeah - hi
12:48:35daurnimatori'm in no rush - i already have a good player - the archos ui is actually good
12:48:52Bagderyeah yeah chicken out already ;-)
12:49:03daurnimatornothing bad i can say about it - cept the diffuculty in scrolling through songs
12:49:09daurnimator(just a dpad)
12:49:16BagderI bet you can't play doom with it
12:49:20daurnimatori can
12:49:37daurnimatorwe ported it
12:49:59Bagderis that a plain linux OS?
12:50:12Bagderto what did you port it then?
12:50:27daurnimator(archopen's project)
12:50:27Bagderyou said archos os was good
12:50:41Bagder"the archos ui is actually good"
12:50:48daurnimatorwell - it is - but i meant "the current state"
12:51:03Bagderand now you mean medios?
12:51:25daurnimatorjust our little exploit
12:51:29 Quit dan_a_ (Read error: 60 (Operation timed out))
12:51:29BagderI don't understand what you're talking about
12:51:34daurnimatorarchos firmware+doom+nes emulator
12:51:38daurnimatorthat is how things are
12:51:48daurnimatorand i'm not unhappy
12:54:48daurnimatoranyway, shirour: did you actually port medios?
12:55:03daurnimatoror just use archopen's code
12:55:18Bagderdaurnimator: what is the point of medios anyway? I mean, is there any benefits compared to Rockbox?
12:56:01daurnimatorits just anther OS is development
12:56:11daurnimator- the people didn't like the rockbox way of doing things
12:56:15daurnimatorso they made their own
12:56:23Bagderbut these guys never even tried rockbox
12:56:28Paul_The_NerdWhich "rockbox way"?
12:56:29daurnimatorthey did
12:56:42linuxstbWhat were their complaints?
12:56:45shirour<daurnimator> yep
12:56:48BagderI think its more "not invented here"
12:56:48daurnimatorjust the way that you manage your targets - not taking full advantage of hardware
12:56:57Bagderdaurnimator: bahahahaha
12:57:00Bagdercrap I say
12:57:07daurnimatorshirour: you got aones going?
12:57:17Paul_The_NerdWhat on Earth does "not taking full advantage of hardware" mean?
12:57:26shirourdaurnimator: haven't tried that.... only the main icons / browser etc'
12:57:41daurnimatori dunno
12:57:43Bagderdaurnimator: in what way does medios "take full avantage" any way better than Rockbox?
12:57:51daurnimatordon't ask me :S
12:57:58Bagderits just a weird accusation
12:58:08daurnimatorask oxygen77 - put him on your notify list
12:58:17daurnimatoror even GliGli - hes online now
12:58:49BagderI have no problems with people doing things on their own
12:59:10Bagderbut saying things like that proves they don't know much about Rockbox
12:59:38daurnimatorjust ask them then :S
12:59:50Bagdernah, I don't see that doing me any good
12:59:51Paul_The_NerdDoes Medios have like... an actual GUI instead of text-based interface? Maybe they mean "Drawing pretty stuff in the menus"
13:00:03daurnimatorit has a gui
13:00:13daurnimatorthough all in dev
13:00:28shirourit has gui... they're currently working on its lib though
13:00:34Paul_The_NerdI mean clearly we aren't taking full advantage of our LCDs since most of the work goes into playback features and battery life and stuff instead of eyecandy. So there's one piece of hardware, at least.
13:00:36daurnimatori just went to them first - cause they already had code going on the gmini400 - i just helped them get it going on the 402
13:01:02daurnimatorPaul_The_Nerd: DON'T complain to me...
13:01:15*Paul_The_Nerd isn't complaining.
13:01:24BagderPaul_The_Nerd: right, but they hardly use their hw to the full either then since they can do full framerate video...
13:01:47Bagdervideo menus!
13:02:15Paul_The_NerdNah. They should clearly be 3D.
13:02:22Bagdertrue, rotating menus!
13:02:23daurnimatorBagder: they/we DO have (eg) a nes emulator running FULL speed - with sound, no frameskip - i call that usig hardware well
13:02:43Bagderbut rockbox would do that too on that hw
13:03:17daurnimatorthen why don't you
13:03:25Paul_The_NerdBecause we're not on that hardware?
13:04:33daurnimatorits running on 5 different lots of hardware
13:04:58daurnimatorWOW - just look at google maps
13:04:58Bagderwe port Rockbox to targets we have
13:05:06daurnimatorthey've REALLY upstepped stuff
13:05:13daurnimatori can see the plot of my house
13:05:24daurnimator- 6 months ago they didn't have nay roads
13:05:29Bagder... 556 rockboxers on the rockbox map now
13:06:17daurnimatorjust added myself
13:06:19Bagdernone in alaska yet
13:06:22daurnimator- you now know where i live
13:06:48daurnimatorbut stil... no name
13:09:47Mmmmamiconn: about the RWPS... You can't have time and filesize on one line because you need both lines for filesplit info
13:10:30pondlifeAnyone understand playlist catalog out there? Nothing in the manual or wiki (that I can find)...
13:11:07linuxstbpondlife: No, but you could try searching flyspray for the original patch - maybe there were some comments there.
13:11:17linuxstb(remember to include closed tasks)
13:12:28Mmmmlinuxstb: why do you think the samplerate shouldnt be in the statusbar in the RWPS?
13:13:51 Quit Kitar|st (Connection timed out)
13:14:25 Join Kitar|st [0] (i=Kitarist@
13:16:12linuxstb(I assume you mean RWRS). It just feels wrong - it doesn't go with the other things in the status bar. You don't put the samplerate there in the WPS for example.
13:16:46shirourtoo bad there isn't a simple api i can wrap and get my port to work
13:17:05Bagdershirour: Rockbox is a full-fledge operating system
13:17:24Mmmmlinuxstb: but samplerate is pointless in WPS, yet important in RWRS
13:17:29Bagdershirour: you need to do the target-specific adjustments, but the upper-layers will work as-is pretty much
13:17:40Mmmmlinuxstb: and it is part of the recording ststus
13:17:55daurnimator3 users in africa
13:18:13daurnimator556 rockbox users.
13:18:14daurnimator * 291 in Europe
13:18:14daurnimator * 193 in North America
13:18:14DBUGEnqueued KICK daurnimator
13:18:14daurnimator * 38 in Australasia
13:18:14daurnimator * 24 in Asia
13:18:15linuxstbMmmm: Why is it so important? I mean, what action does the user take in response to seeing the samplerate?
13:18:15 Quit Mikachu (Read error: 104 (Connection reset by peer))
13:18:16daurnimator * 7 in South America
13:18:18daurnimator * 3 in Africa
13:18:53Mmmmlinuxstb: the user might look at it and think oh, I'd better change that it's not what I want
13:19:18Mmmmits important information
13:19:45linuxstbOK, but I still don't think it should go in the status bar - there must be an alternative...
13:20:07Mmmmbut.. why not the status bar? it is status right?
13:20:26LinusNamiconn: the transition matters because the reset includes the TMR register, so the data written to TMR is immediately cleared if the RST bit does a 1->0 transition
13:20:28linuxstbNo, it's a characteristic of the recording - like bitrate, number of channels, gain etc
13:20:45Paul_The_NerdClearly the proper answer is "Make the Recording Screen use user configurable fonts, and then if users complain about it being too many lines long, tell them to use a smaller font"
13:23:20 Quit Rudy4Pez (Read error: 110 (Connection timed out))
13:25:01LinusNregarding the "doesn't take full advantage of the hardware", it is mostly true
13:25:14Bagderdepending on what you mean, yes
13:25:26LinusNwe don't use all features on all platforms
13:25:35Bagderno, but neither does medios
13:25:36Mmmmlinuxstb: But don't you think that repeat mode and samplerate are both a bit iffy in the "status" status?
13:26:07LinusNmedios will gradually support less and less features as the number of targets increase
13:26:11Bagderbut we do make an effort to use the hw as good as possible when we run on various targets
13:27:08Bagderwhich is what I mean when I say we use the targets to their potential
13:27:10 Join Mikachu [0] (
13:27:41Paul_The_NerdI'm curious now. What would you qualify as things that Rockbox has left out?
13:29:28BagderI assume they simply mean that we don't use icons and colors in our menus
13:29:41Genre9mp3One that I can think of is USBOTG on there something else on any other target?
13:30:03Bagderthere' usbotg on x5 too
13:30:22Paul_The_NerdBut those aren't things that Rockbox doesn't do. They're things it can't do, yet.
13:30:35Bagderyes there's only OSBOTG Paul says not supported YET!
13:31:19Genre9mp3Even video is currently under development
13:31:45Genre9mp3And you CAN use colors in the menus...
13:31:54Paul_The_NerdWell there's various things on the X5, right? Like, lack of full remote support.
13:32:09Bagderand the fact that we don't use every target to its full potential is not a Rockbox idea, it is just that not all targets are properly supported yet
13:32:23Paul_The_NerdThe iPods are lacking a lot of hardware features still (for example, no in-Rockbox USB at all) but none of it qualifies as "not using the hardware to the fullest" since that's the direction it goes in.
13:32:25CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:32:25*Paul_The_Nerd shrugs
13:32:51 Join lee-qid [0] (
13:32:52Paul_The_NerdI'd say stating that someone doesn't use a target to its full potential would imply the thought that they're intentionally leaving something unimplemented.
13:32:58Genre9mp3And there is reason for that most of the time (lack of documentation)
13:33:28Mmmmlinuxstb: I am actually of the opinion that bitrate, no. of channels and samplerate should all go on the statusbar (in recording screen only of course)! :D But it looks like I'm outvoted :(
13:33:32Paul_The_NerdBesides, I think in some instances Rockbox has gone rather far beyond the expected potential. ;)
13:33:35 Join Kohlrabi [0] (
13:33:53linuxstbMmmm: Why does the file-split need two lines on the display? What gets displayed?
13:33:59Genre9mp3Sure...for example: rvf for Archos! ;)
13:34:14Mmmmlinuxstb: the info about when the file will split
13:34:22linuxstbMmmm: If no-one can think of a better solution, then I guess there's no choice. But I would like to try harder to find a better solution...
13:34:47linuxstbMmmm: Again, why is that so important?
13:36:21Genre9mp3Or what about solving even hw malfunctions? (For example the ticking on iriver remotes) My remote ticks with retailOS while not with Rockbox! :)
13:36:49LinusNone example of not using the full potential would be that we only use 16 bits instead of 20 when playing audio
13:36:55LinusNon the iriver/x5
13:37:22 Join dan_a [0] (
13:37:29Genre9mp3Does retailOs support 20bit on those targets?
13:37:38LinusNi doubt it
13:37:57Bagdermost likely not
13:38:11Bagderthe OF people often buy an existing codec
13:38:27Bagderand I don't think 20bit is commonly supported
13:38:48LinusNi'd say that the most lacking part of rockbox is the UI
13:38:55Genre9mp3And how many users actually use 20bit?
13:39:03LinusNit's mostly ugly and not very consistent
13:39:14Paul_The_NerdThere's a patch that helps a lot with consistency...
13:39:20Paul_The_NerdThe play-button returns to WPS from menus, one.
13:39:27LinusNand it does suffer from the common-denominator syndrome
13:39:34Genre9mp3LinusN: why's that? I like text based menus...
13:39:36 Quit midkay (Read error: 54 (Connection reset by peer))
13:39:46Paul_The_Nerd <−−-
13:39:46amiconnLinusN: ouch! Didn't think that TMR would also reset itself...
13:39:54LinusNamiconn: really silly
13:40:10LinusNit sure baffled me when i discovered it
13:40:16 Join midkay [0] (n=midkay@rockbox/developer/midkay)
13:40:37amiconnMmmm: The statusbar is the statusbar. It should be identical for all screens
13:40:47amiconnBesides, how would you fit the samplerate there?
13:40:52amiconnIt's full...
13:40:53Mmmmlinuxstb: You mean the filesplit info? Its really important as a nitifier that it is switched on and a reminder as to when it will split
13:41:15Genre9mp3IMO the text based nature of the menus makes them very friendly and functional
13:41:33Bagderand localizable
13:41:34Mmmmamiconn: I've already done it! :D
13:41:48BagderI doubt medios works fine on 25 languages, including non-latin ones
13:42:24*Bagder stops now and goes poking on code
13:42:31Mmmmamiconn: the thing is that some of the info on the statusbar is totally useless in the recording screen, so why not change it slightly?
13:42:35Genre9mp3And maybe in the future there 'll be a setting for text based or icon based menus...
13:43:02LinusNBagder: i don't think icons need much translations
13:43:10Bagderthey do
13:43:17Bagderthey need to be different I'd guess
13:43:28LinusNi know, just messing with you
13:43:49*Bagder is totally messed up already
13:44:09amiconnLinusN: Your timer fix breaks on-the-fly cycle changes...
13:44:29amiconnThe TMR = 0 should go into the if() block above
13:44:49***Saving seen data "./dancer.seen"
13:45:03amiconn(where the phi &= ~1 is)
13:47:46preglowLinusN: you think it'd be hard to implement 20 bit playback/recording?
13:47:49preglowLinusN: i'd really love it for recording
13:47:53 Quit Mikachu (Read error: 110 (Connection timed out))
13:47:56preglowthat'd really make the h1x0 a nice recording platformn
13:48:08amiconnI'd rather keep dma operation
13:48:17amiconnI wonder what 20bit should be good for
13:48:25preglowamiconn: i'll let you have a guess
13:48:40amiconnThe uda (or the surrounding analog electronics) is so noisy that even 16bit seem like a waste to me
13:48:48LinusNamiconn: s/pdif
13:48:56preglowthe prize goes to linusn!
13:50:06preglowit's nice to exploit the hardware
13:50:10preglowand i think it should be an option
13:51:03LinusNi'm interested in exporting the recording functionality in the plugin api, to allow for dsp effects plugins
13:51:23amiconnI think that who wants o implement 20bit will have a really hard time
13:51:24preglowi'll make a couple asap if that happens :>
13:51:34preglowamiconn: why?
13:51:53Paul_The_NerdLinusN: As this person talked a little about?
13:51:55amiconnThe isr latency needs to be lower than 4 samples
13:52:00LinusNPaul_The_Nerd: yes
13:52:01Genre9mp3LinusN: DSP effect for recording you mean?
13:52:22LinusNGenre9mp3: no, i mean things like guitar effects etc
13:52:27amiconnSo it's practically impossible to disable interrupts
13:52:38amiconn...but we need to do that for several purposes
13:53:12amiconn...and probably longer than 80 s
13:54:01*dan_a meets MTP for the first time... Doesn't like
13:55:14preglowhaving an isr latency of 4 samples sounds bloody outrageous
13:55:21preglowwhere do we disable interrupts for that long?
13:55:35Genre9mp3LinusN: hehe... I imagine my H300 as a guitar pedal! The only problem is that we have to make custom footswitch :)
13:55:54preglowwell, i've been planning to do some kind of effects rig for a long time now
13:56:14barrywardellshould sleep() work in a bootloader build? It hangs when I use it on the H10.
13:56:34dan_abarrywardell: No
13:57:04dan_aYou need to have done init_threads() before you can sleep - as I discovered this morning!
13:57:05barrywardellhmm. maybe that's why my normal build hangs then
13:58:48amiconnpreglow: E.g. on iriver H300 during PCF50605 accesses
13:58:52 Join karim [0] (
13:59:58preglowwhy that long?
14:00:01amiconnI didn't measure, but judging from the slowdown the pcf accesses cause, interrupts might be disabled for 200s or more
14:00:41amiconnWe need to protect from concurrent accesses, and we need to access it from an isr
14:00:47amiconn(the button tick)
14:00:57 Join Mikachu [0] (
14:01:27Mmmmamiconn: you really mean that about the status bar ("It should be identical for all screens"). A bit of a shame seeing as some of it is redundant in the recording screen...
14:01:47LinusNamiconn: the audio irq can be on level 7
14:02:00preglowamiconn: why is it so slow? it just sounds stupid
14:02:06LinusNit is on level 7 today
14:02:19LinusNwhich makes it an nmi
14:02:21karimI found this post where the guy talk about is idct asm optimisation for MIPS on mplayer software
14:02:22pregloware anyone here well-versed enough in playback to have a look at snowgoon's seeking patch?
14:02:54karimthat's the same idct than the one in libmpeg2
14:04:27 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:04:27 Quit linuxstb (Read error: 110 (Connection timed out))
14:04:39amiconnpreglow: The bit-banged i2c isn't any faster. It's already pretty optimised
14:06:40 Join Criamos [0] (
14:07:40 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
14:09:47preglowstill think it sounds bloody horrid
14:09:51preglow1 sample is a lot of cycles
14:12:27 Join Rudy4Pez [0] (
14:17:01dan_apreglow: I'm just looking through the logs from yesterday - I don't know if "mov r0,#-1" will be correct, but the compiler didn't complain
14:17:59preglowperhaps it translated it to a mvn r0, #0
14:18:12preglow-1 = 0xffffffff, which you can't set with a mov instruction, afaik
14:20:00dan_aAh - I'm still getting to grips with ARM ASM - I'll work out a correct way to do it
14:23:30preglowmvn r0, #0 = r0 = ~0 = 0xffffffff = -1
14:23:52preglowi can confirm the backwards seeking bug, at least
14:24:03preglowa long ogg file just snaps to the wrong seek point when i try
14:26:24 Join XavierGr [0] (
14:29:13 Quit Mmmm (Remote closed the connection)
14:29:49 Quit ryanpg (Remote closed the connection)
14:30:32markundamn, 608 rockbox users on rasher's page now..
14:30:53Mikachugood thing they don't all have to use irc to get it working
14:33:20 Quit Kohlrabi (Nick collision from services.)
14:33:30 Join Kohlriba [0] (
14:34:07preglowall norwegian rockbox users are in bergen, apparently
14:34:15preglowoh, no, there's a couple of other ones now
14:35:39LinusNmost of the swedes have really odd names :-)
14:35:48 Quit XavierGr (Nick collision from services.)
14:35:52 Join XavierGr [0] (
14:35:58LinusNxlarge, mlind, frause
14:38:25karimHans Gruber
14:38:33karimthat's sweedish ?
14:38:34preglowi take it those are their proper first names? :>
14:38:40preglowkarim: sounds german
14:39:21karimthat's the bad guy in die hard 1
14:39:23bluebrotherHans Gruber? Wasn't that the name of one guy in Die Hard?
14:39:59Bagderwe're actually all terrorists up here
14:40:20pondlifeHmm, I'm trying to fix a hard crash in the ID3 view context menu... select Playlist Catalog and then try to select on the blank menu... kaboom.
14:41:17pondlifeA neat solution would be to enable the View option from the browser contect menus - allowing people to start playback that way.
14:41:50pondlifeThat works but starting playback in this way doesn't enter the WPS.
14:42:12 Quit daurn|laptop (Read error: 104 (Connection reset by peer))
14:42:28pondlifeSo I might just go back and disable the Playlist Catalog option in ID3 view, so people can't get to the empty menu
14:42:42preglowLinusN: would exporting the recording stuff be much bother?
14:42:45pondlifeSeems like a backwards step though it does remove one crash.
14:43:41LinusNpreglow: i don't think so, but the current recording code is tailored for recording to disk
14:43:48preglowis it, now
14:44:01preglowi just want recorded data in a buffer :>
14:44:34LinusNshouldn't be too hard to change
14:44:36preglowi've got an already partially fixed poind-ized echo here i'd love to port
14:44:38pondlifeI suppose that an explicit call to gui_wps_show() would be a bad idea, as the browser may be scrolling stuff...
14:45:00LinusNpreglow: a simple reverb should be possible
14:45:05preglowLinusN: oh yes, easily
14:45:26*preglow might finally have his vocoder
14:45:41 Quit thegeek ("( :: NoNameScript 4.03 :: )")
14:45:46karimare you working on rockbox at your office ?
14:46:03preglowi'm talking about rockbox at my "office"
14:46:28*preglow works from home these days
14:46:35*Bagder does too
14:46:45Mikachuyou mean "works"? :)
14:46:48*pondlife does too
14:46:58preglowi've been quite diligent today, actually
14:47:01pondlifeWell I should get on with that I suppose...
14:48:38preglowLinusN: even a moderately complicated reverb mighty be possible, i guess it's the memory accesses that'll kill us
14:48:44preglowgood reverbs need tons of ring buffers
14:49:02 Join thegeek [0] (
14:49:30dan_apreglow: You're thinking of using an MP3 player as a vocoder?!
14:49:50preglowdan_a: sure, why not
14:49:53*dan_a finds the Microkorg easier
14:49:53pondlifeEffects pedal! Distortion, reverb, digital delay?
14:50:04preglowdan_a: it's also more expensive and's got a crappy keyboard :P
14:50:21karimand you calling video a gimmick ?
14:50:27pondlifeCan we do input and output simultaneously with a low-ish latency?
14:50:33dan_aBut it's got the lovely fake wooden sides!
14:50:37preglowi'm not calling it a gimmick, i'm calling it something i'll almost never use
14:50:47daurnimatorpondlife: latency doesn't matter with reverb
14:51:00pondlifeNot with the wet signal
14:51:01preglownow, a vocoder...
14:51:06karimthe vocoder is to do anonymous phone calls ?
14:51:20pondlifeYou can do that with voice files
14:51:20preglowit's to do scary-ass phone calls
14:52:08karimhow do you wire the player to the phone ?
14:52:32daurnimatorwith wires
14:52:37preglowbut anyway
14:52:45preglowi think the h1x0 will make for a nice little dsp platform
14:52:56preglow120 mhz + fair amount of iram + emac = good
14:53:10preglowyou're actually not _that_ far removed from a proper dsp chip
14:55:02karimargh we can't go on other channels than #rockbox ...
14:55:49preglowfrom the web client? of course not
14:55:52preglowthat would be asking for it
14:56:27daurnimatoryou can have proper dsp with mine - if you want
14:57:20karimpreglow what ?
14:57:36Bagderkarim: use a proper irc client and you can! ;-)
14:58:01 Quit lee-qid ("Trillian (")
14:58:11 Quit XavierGr (Nick collision from services.)
14:58:13karimpreglow I used to use cgiirc as a irc gateway when I am at work, it was installed on my computer. channel and server free
14:58:15 Join XavierGr [0] (
14:58:25amiconnLinusN: You said RaeNye put a patch in the tracker with an optimised pcf50605 driver.
14:58:27preglowkarim: well, this cgiirc is locked to rockbox
14:58:33preglowkarim: and it should be, otherwise it would be abused
14:58:55karimBagder behind a proxy it's not as simple :o)
14:58:57amiconnThis driver seems to attempt to use the coldfire's i2c controller - however, I can't see how this could work??
14:59:36dan_aBagder: I have a new toy :)
15:00:43preglowdan_a: how many channels is the microkorg vocoder anyway?
15:01:08daurnimatorkarim: google for cgi-irc
15:01:16daurnimatoryou'll find one that can connect anywhere
15:01:19dan_apreglow: From memory, 6. Or maybe 4
15:01:30preglowvery improbable
15:01:36preglowi've never seen anything lower than 8
15:01:47*preglow googles
15:02:04dan_aAh, sorry, I was thinking of polyphony
15:02:06LinusNamiconn: wy wouldn't it work?
15:02:17preglowit's sixteen channel
15:02:21preglowahh, yes, it's four voice
15:02:37*dan_a only had 3 hours sleep last night
15:03:26preglowsounds like a blast
15:04:02amiconnLinusN: Afaics, the coldfire i2c controller is not connected to the pins where the pcf is hooked up
15:04:11 Join dongs [0] (
15:04:42 Join lorijho [0] (
15:05:35 Join Pyromancer [0] (
15:08:45LinusNamiconn: gpio10 is scl1 and gpio44 is sda1
15:09:29 Join snowgoon [0] (
15:10:05LinusNbesides, the OF uses the i2c controller
15:10:17 Join lowlight [0] (
15:10:33snowgoonhi preglow
15:11:17amiconnMeh, the coldfire datasheet is clear as mud :/
15:11:25LinusNamen to that
15:11:33preglowsnowgoon: hi, man
15:11:39preglowsnowgoon: good to see you're still on it
15:11:47snowgoonsorry for the delay :)
15:12:23preglowsnowgoon: the newest patch is looking very good, though i haven't had time to look at it yet
15:12:32preglowbut it's on my list of stuff to do very soonish
15:13:02preglowonly hurdle is the seek patch, i don't know nearly enough about the playback system to see if it has any other nasty effects
15:13:14snowgoonhopefully i didn't break gapless :-P
15:13:43snowgoonthere is some strange code in there
15:13:50preglowyes, yes there is
15:13:52snowgoontook me a while...
15:15:06preglowi can imagine
15:15:09lowlightLinusN: Can I be considered "rockbox wannabe coder"?
15:15:13*lowlight refers to yesterday's log
15:15:15amiconnLinusN: If we can use the i2c controller, we could run the button adc conversion in an isr like we do on archos, and just read the values as needed.
15:15:40amiconnThen we won't have a performance problem, and don't need to disable interrupts for so long
15:15:59Bagderdan_a: oooh, congrats to the new toy
15:17:01 Join gucky [0] (
15:17:13dan_aI'm just looking at the list of things that people can do (in the forum thread)
15:17:22preglowamiconn: is that the only place that disables interrupts for that long?
15:17:33LinusNlowlight: nice!
15:18:19LinusNamiconn: we still have the issue with i2c access in background threads
15:18:26 Part gucky
15:19:03amiconnLinusN: Not if the i2c driver serializes the accesses.
15:19:37amiconnThe advantage is that if a thread wants i2c and the controller is busy, we can yield. In the isr we can't
15:19:47LinusNamiconn: i don't see how using the i2c controller would make it any different
15:19:52dan_aBagder: Is there anything that can be added to that list at the moment? (
15:20:17 Join barrywardell_ [0] (
15:20:43Bagderdan_a: not really, nothing much has happened since then. We haven't manged to display anything on the LCD and we haven't even got any proof that our code actually runs on target
15:20:50amiconnLinusN: The i2c controller can fire interrupts. Much less load than bitbanging with interrupts disabled
15:21:16amiconnIt's true that we could do the serialization with the bitbanging driver as well
15:21:26Bagderdan_a: MrH has an idea how to reboot the thing, but when I tried to just delay and reboot to see something happen on the reboot moment, I couldn't
15:21:58Bagderwe need to get the jtag thing going soon to see if we can figure this out
15:21:58 Quit lowlight ("CGI:IRC (EOF)")
15:22:24LinusNamiconn: the coldfire i2c controller is not atomic, i.e it is byte-based, like a simple uart
15:22:28amiconnpreglow: Another place that disables interrupts is the remote lcd driver, unless we're running at >50MHz
15:22:30barrywardell_Bagder: I tried MrH's reboot thing on the H10 and it works
15:22:40Bagderah, nice
15:22:48amiconnLinusN: Yes.
15:22:51Bagderthen perhaps we miss some startup magic for the sansa
15:22:52barrywardell_so it should probably work for the sansa too
15:23:15LinusNso we still need to block the i2c bus for an entire message
15:23:21Bagderbarrywardell: I tried just massive loop doing nothing and then writing to the reboot register but nothing showed
15:23:29 Quit XavierGr (Nick collision from services.)
15:23:33 Join XavierGr [0] (
15:23:34amiconnLinusN: Yes. i2c_begin() and i2c_end()
15:23:35LinusNand protect from being interrupted when doing i2c in the background
15:23:54amiconnWe don't need to protect from being interrupted
15:24:04amiconnThe driver itself should take care of it
15:24:20LinusNeither that, or skip a button interrupt
15:24:24 Join rconan [0] (n=richard@
15:24:30amiconnSkip or delay, yes
15:24:34barrywardell_Bagder: which build did you use it with? I can only get code to run using the bootloader build. other builds just hang and don't seem to run the code
15:24:54Bagderthe sansa only has bootloader build working
15:26:01pondlifeSorry to pester when you're all in low-level mode, but is there any sensible way to force a switch to the WPS. I have a patch up at which fixes a crash, but isn't quite fully baked in the UI department
15:26:01LinusNamiconn: the background thread doing the i2c must not yield in that case
15:26:43LinusNif it does, it might very well be delayed many many milliseconds
15:26:48barrywardell_Bagder: I do a system_init() and kernel_init() and lcd_init() first. maybe they're necessary
15:27:18BagderI don't see why they would but I could try that later on
15:27:26 Quit barrywardell (Read error: 113 (No route to host))
15:28:08 Nick barrywardell_ is now known as barrywardell (
15:28:14amiconnLinusN: I think it can very well yield. It would be delayed a few milliseconds worst-case
15:28:38LinusNnot if a thread misbehaves
15:28:50LinusNlike a jpeg decoding plugin or similar
15:29:10amiconnWell, if a thread never i2c_end()s then all further transfers would fail
15:29:29 Quit XavierGr (Nick collision from services.)
15:29:32 Join XavierGr [0] (
15:30:30amiconnBut that also applies now
15:32:22LinusNyes, but that only happens if the thread doing the i2c misbehaves, if we implement a locking mechanism, any thread can delay the button i2c by not yielding
15:32:49karimjust for the irc log I post that idct decomposition here seems this is how they got the idct.c, along the link I posted before
15:33:14linuxstbkarim: You could add links to the end of the PluginMpegplayer wiki page - so they don't get lost.
15:33:49dan_alinuxstb: Any luck with the COP patch if you disable frequency scaling?
15:33:56*pondlife lunches but will check back later and pester more when not feeling so hungry
15:34:00linuxstbdan_a: No, that made no difference.
15:34:01*amiconn wonders whether that would be a problem
15:34:28amiconnThe only problematic case I can see is button_get(false)
15:34:37linuxstbdan_a: And crt0.S has become incomprehensible now with all the #ifdefs...
15:34:43amiconnAll other methods for getting button input yield themselves
15:35:07LinusNamiconn: that, and lengthy operations without yielding, like calculating mandelbrot, decoding jpeg etc
15:35:40amiconnLinusN: Yeah, but does it matter to delay button input when nobody requests it?
15:35:47amiconnHmm, maybe it does...
15:35:53linuxstbdan_a: BTW, I've been thinking that the cop-related stuff in apps/main.c should go somewhere in firmware/ - it seems to be belong there more IMO.
15:36:04amiconnBtw, mandelbrot yields quite often
15:36:29Mikachuyou can zoom in while it's rendering
15:37:17karimlinuxstb yes good idea
15:37:59dan_alinuxstb: The parts in main.c pretty much mirror what is there for the CPU (setting up the kernel, etc.) so I thought it followed to put them there.
15:39:07 Join MrGreen [0] (
15:39:26amiconnLinusN: Hmm, in fact with my idea a thread not yielding wouldn't lock out the button i2c
15:39:45linuxstbdan_a: But main.c also contains the main Rockbox thread - on the COP, the main thread does nothing, it's just there to let you start new threads.
15:39:46LinusNhow so?
15:39:57MrGreenis ipod ok set to vfat ?
15:40:00amiconnMy idea is that the thread wanting to access i2c would place an "i2c job" in a queue
15:40:30 Quit snowgoon (Read error: 60 (Operation timed out))
15:40:33amiconnIf the controller is available, it could be serviced immediately, if not, then the thread can yield
15:41:03LinusNso it wouldn't yield while waiting for the i2c controller to transmit a byte?
15:41:13MrGreendocs say ipod must be set to fat32 .... is vfat ok ?
15:41:28MikachuMrGreen: vfat is what you want
15:41:40amiconnNo, it would tell the driver "I want 3 bytes from device #12", then yield
15:41:55LinusNand the driver busy waits?
15:42:01amiconnThe driver would perform that access and store the result
15:42:10MrGreen& like can i recover ipod if things go pear shaped?
15:42:13amiconnNo, the driver would run in isr context only
15:42:28MikachuMrGreen: ipods have emergency disk mode that isn't affected by rockbox
15:42:33LinusNso it would busy wait in the driver isr?
15:42:36amiconnIt would be invoked by the timer tick, and check whether there's something to do
15:42:48MrGreenMikachu: I need one of those for my system lol
15:42:49 Join PaulJam [0] (
15:42:56MikachuMrGreen: :)
15:43:01MrGreenthanks again
15:43:02amiconnIf there's a job waiting, it would initiate the first transfer and enable the i2c interrupt
15:43:03 Part MrGreen
15:43:08 Quit dwihno (Read error: 104 (Connection reset by peer))
15:43:34amiconnWhen the first byte is transferred, the i2c interrupt calls the driver again, which starts the next byte etc
15:43:52amiconnWhen the job is done, the last action would be to read the adc
15:43:53LinusNamiconn: and the button adc driver would have to skip a reading if the i2c driver is using the bus?
15:44:18amiconnNo, there would only be _one_ i2c driver which handles both the queue and the button adc
15:44:19dan_alinuxstb: What would you suggest? firmware/cop.c?
15:44:29amiconn...but solely interrupt driven
15:44:52***Saving seen data "./dancer.seen"
15:47:01linuxstbdan_a: I'm not sure if it needs its own file - it's just one small function. I'll think about it once we get it working.
15:47:09tucozAnyone object to remove databox from SUBDIRS, but still keep the sources in CVS?
15:47:24tucozas it is currently not working
15:47:44 Join dwihno [0] (n=dw@
15:48:22linuxstbdan_a: Ah, I see a difference in system.c - you didn't change system_init() for the PP5020. But I've tested without calling system_init() from cop_main() already...
15:48:37 Join CriamosAndy [0] (
15:49:47 Quit XavierGr (Nick collision from services.)
15:49:50 Join XavierGr [0] (
15:54:11markunBagder: I tried to attach a html file to the wiki but it's now redirected to
15:54:43*Bagder blames Zagor's antispam work
15:55:21amiconnLinusN: The only situation where the i2c driver would skip a button reading would be if a transfer job takes longer than one tick period. But that's the same today
15:55:34 Quit Criamos (Read error: 60 (Operation timed out))
15:56:30markunBagder: any idea what I can do about it?
15:56:43Bagdernothing atm, except not having it html
15:56:55markunBagder: It's a leaked partlist for the Gigabeat:
15:58:00ZagorI guess we can remove the html redirect now that we have the manual authorization
15:58:09Bagderyeah I think so too
15:58:14markunZagor: thanks
15:58:23markundo I need to upload it again?
15:58:57 Quit lorijho ("Kopete 0.11.3 :")
16:00:43 Join gorey [0] (
16:01:37Zagormarkun: fixed now
16:02:10markunZagor: firefox shows me the source of the document. Maybe the mimetype is incorrect?
16:02:43Zagorthat's odd. Content-Type: text/html; charset=ISO-8859-1
16:03:02Zagormy firefox doesn't show me the source
16:03:58markunZagor: your URL works fine
16:04:01Zagoraha, that explains the difference :-)
16:04:10Bagderthe 2nd is "text/plain"
16:04:29Zagorright, that's another anti-spam thing. hang on...
16:04:55SlasheriPaul_The_Nerd: now the problem with tagcache should have been fixed :)
16:07:29 Join Febs [0] (
16:10:32 Join mrobefan [0] (
16:10:39dan_alinuxstb: The changes in system.c shouldn't affect whether the COP starts or not - it should only be crt0.S that does that.
16:12:16 Join DarthLappy [0] (n=Angus@pdpc/supporter/student/DarthShrine)
16:16:40 Quit shirour (Read error: 104 (Connection reset by peer))
16:16:41 Join lorijho [0] (
16:17:11 Part Paul_The_Nerd
16:17:44 Quit Kohlriba ("Quit")
16:24:18 Quit barrywardell ()
16:24:19 Quit Rob2222_ (Read error: 104 (Connection reset by peer))
16:24:23coobapple firmware can't see my fat32 partition :(
16:26:55coobyes yes amsuing.
16:27:02coobamusing* heh.
16:27:22coobguess i should go find a win32 machine to abuse
16:27:54pondlifeSlasheri: Don't know if it's your update, but I no longer seem to get any runtime data collected. I'll try a re-init and see if that helps....
16:28:02karimdoes anyone have an idea how the scale factor in MAC instructions can be used for ?
16:31:49Slasheripondlife: hmm, interesting. i will try too
16:32:21goreyare there any presents for the equalizer to download?
16:32:25goreye.g rock / metal?
16:32:54preglowif you like your music, don't use that kind of presets
16:33:48goreymhh why?
16:34:06Slasheripondlife: it still seems to work for me
16:34:57preglowgorey: because they just do totally arbitrary stuff to your music. having a "metal" eq preset implies that all metal sounds the same and that it all needs to have the same stuff done to the sound
16:35:12preglowgorey: anyway, i don't think there are any eq presets like that around
16:35:23Slasheribtw, a good hint for everyone who likes crossfading always to happen when skipping: set anti-skip buffer to 10 minutes
16:35:55goreyhm okay preglow
16:36:30 Join Mmmm [0] (
16:36:32preglowSlasheri: minutes???
16:36:40goreybut i also have the rock preset in my winamp and it sounds a lot better, same with the ipod OS
16:36:56Slasheripreglow: yes, i know. that will eat half of the buffer but it's the only way to make sure crossfading always happens
16:37:08preglowSlasheri: why?
16:37:17Slasheripreglow: well, if user prefers that :)
16:37:21Slasherino other reasons for that
16:38:11Slasherii have tried that a few days now and didn't really notice faster battery consumption at normal use
16:39:31pondlifeDo I have to complete playback of a track for it to be added to the runtime list?
16:39:51Slasheripondlife: no, but you need to wait for the disk to spin up to sync the db
16:40:12pondlifeBut is sync not forced when I do a query?
16:40:14Slasheriit might be necessary to play several tracks before the db will be updated
16:40:18 Quit Zagor ("Client exiting")
16:40:29Slasherinot at the moment
16:40:35pondlifeOK, that's probably what's up
16:40:48pondlifeI'll let you know in a while
16:41:00Slasheriand if you stop playback, the currently playing track will never be added to the db (as that might generate an incorrect score)
16:41:27Slasherihowever, already played tracks are added to the db
16:41:29pondlifeSo it's not added at start of playback
16:41:45pondlifeStopping will force a sync of previous tracks then?
16:41:52Slasherinope, only when stopping playback or filling the buffer
16:42:56 Join Crackerizer [0] (
16:43:14pondlifeHmm, I played one track through, then hit stop on the 2nd track and it doesn't seem to have added anything to the "List played tracks" query...
16:43:47Slasherithat's weird.. do you have the "gather runtime data" option enabled?
16:44:10pondlifeAnd auto update
16:44:15Slasherican you access the logf feature?
16:44:20pondlifeHmm, probably not
16:44:26pondlifeI can do a sim run
16:44:30Slasherithat would tell you what went wrong
16:44:33Slasherisounds good
16:44:34pondlifeBut I don't have a remote
16:44:41Slasheriyou don't need a remote
16:44:45pondlifeWhere does logf data get written on the sim?
16:44:48Slasherilogf can write results to disk
16:44:52CrackerizerHello, how do i get the uisim source?
16:44:54Slasherito the remote :D
16:45:00pondlifeI know DEBUGF -> stderr.txt...
16:45:05pondlifeBut LOGF -> ?
16:45:20linuxstbCrackerizer: From the same places as the rest of the Rockbox source.
16:45:25linuxstb(it's all together)
16:45:26Slasheriwell, if you have a logf enabled build, you can go to debug menu -> logfdump -> ./rockbox/logf.txt
16:45:50CrackerizerOk, thank you.
16:45:57pondlifeDo I need to edit a line to enable logf, or will the sim include it anyway?
16:46:05pondlifeActually IIRC there's no Debug menu on the sim
16:46:17Slasheriyou need to run the ../tools/configure script
16:46:19pondlifeAlthough there should be...!!
16:46:25Slasheriand then enabled the logf option
16:46:31Crackerizerlinuxstb: can i build the full rockbox from rockbox-devel source?
16:46:37Slasherisim doesn't automatically have it but you need to enabled both sim and logf
16:47:00Slasherilogf works only on real targets
16:47:06Slasheriwith sim it's limited
16:47:20Slasherihmm, can't you compile for target?
16:47:48pondlifeYes, but the round trip takes longer. I have USB problems on this PC too.
16:47:55linuxstbCrackerizer: Yes - if you have the required cross-compiler installed.
16:47:55pondlifeLeave it with me.
16:48:02pondlifeSim is easy
16:48:08LinusNSlasheri: with the sim, you just use DEBUGF
16:48:13pondlifeAnd ideal for this sort of debugging I'd think
16:48:25 Join Hansmaulwurf [0] (
16:48:26pondlifeAnyone fancy unifying DEBUGF and logf?
16:48:27SlasheriLinusN: hmm, will that write the logf output on disk too?
16:48:41pondlifeOr is DEBUGF a wrapper for logf ?
16:49:02LinusNDEBUGF writes to stderr
16:49:17LinusNsorry, stdout
16:49:29pondlifeBut it's of no use where the code contains logf, is it?
16:49:41pondlifei.e. I need to modify the logging code..
16:50:07LinusNit should be trivial to make logf() use stdout as well
16:50:23Crackerizerlinuxstb: thank you. I'll check it out from the cvs.
16:51:12pondlifeOK, I'll do that
16:51:25LinusNdebugf and logf have different uses, but i agree that logf output should be visible in the sim
16:51:46LinusNa simple hack in logf.h should suffice, i guess
16:54:21 Part LinusN
16:56:20 Quit DarthLappy (Nick collision from services.)
16:56:24 Quit lorijho ("Lost terminal")
16:56:33 Join rotator [0] (n=e@rockbox/developer/rotator)
16:56:54 Join DarthLappy [0] (n=Angus@pdpc/supporter/student/DarthShrine)
16:57:28goreyhow do i install the Scrolling Margins Patch ?
16:57:42 Quit rconan ("Leaving")
16:58:25tucozgorey, you apply it.
16:58:45 Quit jaebird (Read error: 104 (Connection reset by peer))
16:59:02 Join jaebird [0] (n=jae@
16:59:35tucozgorey, do you know how to compile rockbox?
17:00:00Mikachudoes it still apply?
17:00:26tucozgorey, lots of info on that here
17:00:28goreytucoz im workin on it
17:03:24goreyim trying to "install" the requires for the uni catcher theme
17:03:39goreyi started today with rockbox.. total noob ^^
17:04:44tucoznice, it's fun once you're able to compile your own builds
17:05:33goreyi think so, its even fun to learn those things
17:07:32bluebrothertucoz, I have a new diff for the quickstart section here:
17:07:41bluebrotherdo you think I should commit that state?
17:08:00 Quit Mmmm (Remote closed the connection)
17:11:51tucozbluebrother, just do it. I haven't read what you have written now, but I believe it's good :)
17:12:00 Quit Landus (Read error: 54 (Connection reset by peer))
17:12:15 Join MarcoPolo [0] (
17:13:22tucozbluebrother, do you think we should change the \dap macro for the recorder to be jukebox instead? I think that is what they called it in the past.
17:14:09bluebrotherok. I hope this will make it a bit easier for new users.
17:14:20bluebrotherhmm. Sounds reasonable to me.
17:14:42tucozbut that can wait. I guess we'll have to hear what amiconn and co thinks
17:15:57goreybluebrother funktionieren hier keine queries?
17:16:13bluebrothergorey, queries?
17:16:36Mikachuunregistered users can't send messages to registered users
17:16:38tucozbtw, I could never imagine that we would get .svg images of the players with such a high quality. The ondios look sweet.
17:16:44goreyah oke
17:16:51Mikachuunless the registered user enables them
17:17:53amiconntucoz: The full name would be "Archos Jukebox Recorder NN" with the NN being the capacity in GB (6/10/15/20)
17:18:09pondlifeSlasheri: OK, I have a logf build of the sim now. Shall I just do a rebuild, then restart and let it play awhile?
17:18:29 Quit perplexity (Success)
17:18:31tucozamiconn, ok. But do you think it would be a good idead to use the short term jukebox in the text (instead of player)
17:18:59tucozamiconn, e.g. connect your jukebox to the computer
17:19:08 Join perplexity [0] (
17:19:24amiconnFor the player it's even more funny: "Archos Jukebox NNNN" with NNNN being the (rounded) capacity in MB (5000/6000), and "Archos Jukebox Studio NN" with NN being the capacity in GB (10/20)
17:19:46Slasheripondlife: try to do as similar as on the player to reproduce the problem
17:19:48 Join Landus [0] (
17:20:12tucozaha. so that is what the numbers interpret to
17:20:36amiconntucoz: pixelma will probably work on the images a bit more. With the original device as a comparison, it's possible to achieve even higher quality :)
17:21:17pondlifeDoes "failure to open changelog" mean much?
17:21:17amiconnI think it would be good to use just "Jukebox" for the Player, but keep "Recorder" for the recorders
17:21:45*amiconn has never seen an actual Jukebox 5000
17:21:57tucozhehe. I think a lot of the drawings are of a very high standard. But I assume the term pixel in her nick reveals an interest in CG.
17:22:07amiconnOnly 6000s and Studios... the 5000s seem to be very rare
17:22:23Slasheripondlife: that's not fatal, it just means there was nothing to import (the runtime data)
17:22:53pondlifeAm playing some tracks now, will see what this reveals
17:23:11tucozI would like to clarify the licence situation wrt those images. As it is now, I am not sure what licence they are released under.
17:23:37amiconnI agree, the documentation license situation needs a cleanup
17:23:50tucozthere is no question that the compiled .pdf and html docs are GFDL. But what about the source?
17:24:05amiconnThe manual is releast under GFDL (which isn't necessarily the best choice as at least some people stated)
17:24:23amiconnThe wiki and some other parts have no clear license
17:24:25tucozand the .svg images, that we have to transcode(?) to pdf to use in the manual.
17:25:30tucozAnyway, I have sent an email to the licence team at I hope they have the time to answer my questions.
17:25:31 Join shirour [0] (
17:30:33 Quit mrobefan (Read error: 104 (Connection reset by peer))
17:32:55 Join Sinbios [0] (
17:34:39 Quit Arathis ("[rl_bot quit]")
17:34:42 Quit DarthLappy ("SLEEP! (Is for the weak. So I'm weak...Get over it!)")
17:34:59 Join barrywardell [0] (
17:37:08pondlifeSlasheri: I think I've opened a small can of worms
17:37:28pondlifeAgain, the sim does not list any played tracks
17:37:50pondlifeThe logf output contains some really confused paths too.
17:37:53 Join pixelma [0] (i=pixelma@
17:38:03pondlifee.g. "Checking for /Elvis Costello/This Year's Model/Have You Fed The Fish/"
17:38:03pondlifeWe open the real file 'archos/Elvis Costello/This Year's Model/Have You Fed The Fish/'
17:38:03pondlifeFailed to find: /Elvis Costello/This Year's Model/Have You Fed The Fish/
17:38:37pondlifeConfused between the paths for "Elvis Costello/This Year's Model" and "Badly Drawn Boy/Have You Fed The Fish"!
17:38:47pondlifeI was listening to the latter
17:44:53***Saving seen data "./dancer.seen"
17:46:34 Quit XavierGr (Nick collision from services.)
17:46:37 Join XavierGr [0] (
17:49:19dan_aWould there be any objections if I were to clean up crt0.S a little by doing something like this:
17:50:59 Join sharpe [0] (
17:53:38amiconntucoz: Do you have an idea how the remotes should be handled in the manual?
17:53:54amiconnThey have buttons, which have names and can be used for navigation...
17:55:03tucozNot really. I've thought of it, but haven't done anything more :)
17:55:35tucozBut they should be added sometime soon. Preferably by using action codes
17:56:16 Join apo [0] (
17:56:33tucozHopefully we can extent the "key action" tables to have an extra column for the targets with remotes.
17:56:44bluebrothermaybe one time we could add a chapter in the appendix for the remotes.
17:56:49amiconnI will scan the 3 remotes I have (H100, X5 and Archos)
17:57:28tucozgood. will you add them for download at
17:57:38 Quit apo` (Read error: 110 (Connection timed out))
17:58:08tucozbluebrother, don't you think we should list the remote keys next to the main units? At least for h1xx and h300?
17:58:18Genre9mp3amiconn: Do you wou want me to scan the H300 remote as well?
17:58:45Genre9mp3also the non-LCD remote...
17:58:48bluebrothertucoz, I'm not sure −− it would make it more complex.
17:59:02 Join oxygen77 [0] (
17:59:39 Join lowlight [0] (
17:59:58bluebrotherso maybe it would be better to describe rockbox in the manual itself and reference to some appendix chapter for the remotes.
18:00:45*lowlight sucessfully played some tracks with metadata on the buffer :)
18:01:01tucozGenre9mp3, that would be great.
18:01:20Genre9mp3tucoz: ok
18:01:58tucozGenre9mp3: send them to Bagder, so that he can add them to
18:02:44Crackerizer i got errors while compiling the source from rockbox-devel....any suggestion please.
18:03:03tucozCrackerizer, paste the errors on
18:03:17preglowlowlight: !
18:03:32preglowlowlight: did you reapply lostlogic's patch, or?
18:05:00Crackerizertucoz: I posted it. what's next?
18:05:29coobwow the rockbox ui really doesn't work so well on the ipod
18:05:53CrackerizerI did compile the source from rockbox yesterday, no errors at all...
18:06:16bluebrothertucoz, commited it. I hope it won't confuse too much users ...
18:06:25lowlightpreglow: :) I looked at lostlogic's but it was confused because it was mixed within his playback rewrite.
18:06:51preglowlowlight: right, so you've actually done your own? big patch?
18:07:37 Join MrGreen [0] (
18:07:47MrGreeninstall went ok ... but where has music gone?
18:08:15bluebrotherMrGreen, on ipod? Try TagCache.
18:08:34MrGreenbluebrother: yeah ipod video .... fonts look tiny
18:08:51bluebrotheruse a different theme ... like UniCatcher
18:08:55MrGreenwhat is TagCache *dares to ask
18:09:03tucozCrackerizer, you have to give us the link you got from pastebin
18:09:23tucozbluebrother, nice
18:09:25FebsMrGreen, your questions are all answered in the manual.
18:09:44*MrGreen runs to hide
18:09:48CrackerizerOh, it's
18:10:03*bluebrother points to
18:10:16FebsIt's not a criticism, it's just that we have explained these things in the manual much more thoroughly than we can here on IRC.
18:10:17bluebrotheroh, Febs was faster :o
18:10:18lowlightpreglow: Yes, my own.
18:10:37tucozCrackerizer, have you applied a patch?
18:11:16*bluebrother found the eqf import function on xmms
18:11:37Crackerizertucoz: I checked out the source from cvs 10 minutes ago. do i need applying patch?
18:11:39tucozCrackerizer, or maybe you miss SDL
18:14:15tucozin that case you have to install the libsdl-devel package of your linux distro (in case you run linux)
18:14:36tucozand to answer your question, no.
18:16:18MrGreengot it thanks guys
18:16:39MrGreen..... can I restore pod if things go pear shaped or is that in manual too?
18:16:59 Quit Crackerizer ("CGI:IRC (Ping timeout)")
18:17:00lowlightpreglow: For the daring only. Only vorbis comments are put on the buffer and it will probably crash if the buffer refills :)
18:17:30 Join Crackerizer [0] (
18:17:53FebsMrGreen: yes, you can always uninstall Rockbox, or use Apple's restore tool.
18:18:10*MrGreen wonders where he can find that ?
18:18:17FebsBut since the Rockbox bootloader has a dual-boot capability, you don't usually need to.
18:18:25MrGreenwoah! cool
18:18:35FebsFind what, the restore tool?
18:18:47MrGreendual boot ....
18:18:55MrGreenpress menu & select?
18:19:42CrackerizerOk, i think i dont have the sdl. How do i get that?
18:20:27tucozCrackerizer, what distro?
18:21:04FebsMrGreen: turn the ipod off by pressing and holding play/pause. Then turn it on by pressing select and immediately press and hold menu until you see "loading original firmware." Use the menu+select combination to return to Rockbox.
18:21:41Crackerizertucoz: slack, i donwloaded the source code yesterday. what do i need? lib & include file?
18:22:17tucozCrackerizer, I don't know anything about slack. Isn't there packages available for that?
18:22:34*tucoz googles for Crackerizer :)
18:23:06MikachuDid you mean: crackerized
18:23:41tucozI meant googles at his service. Searching for slackware sdl-devel package
18:24:04Crackerizertucoz: can i build it from source?
18:24:18tucozCrackerizer, sure.
18:24:57tucozwhy not? But it's usually a shorter path to install packages using your distros package managment
18:25:22*MrGreen loves Febs very much
18:25:34MrGreentiming is all
18:25:39CrackerizerWell, i'll try build it first. i dont think slack has sdl-devel package.
18:26:03Crackerizeri tried to install the rpm yesterday. but it didnt work. :(
18:26:15 Quit MrGreen ("Lost terminal")
18:26:15tucozCrackerizer, ok. Read the last post in this page:
18:27:05Crackerizertucoz: thank you very much. you are very helpful. :)
18:27:16Crackerizeri'll try it.
18:27:24tucozYour welcome
18:27:29Genre9mp3tucoz: Are these scans ok?
18:27:59 Join old45 [0] (
18:28:22tucozGenre9mp3, it would be better if you didn't save them as jpg. Use png instead
18:28:40 Join snowgoon [0] (
18:28:49Crackerizer"> >No offense, but... are you sure Slackware is the right distro for you?", what is this supposed to mean?
18:28:51tucozand I think 300dpi is the preferred dpi. Not sure what you used
18:28:51Genre9mp3tucoz: png is lossless, wont affect the quality if I convert them, right?
18:29:11tucozconvert from jpg?
18:29:11Genre9mp3I used 200 dpi...ok...I'll rescan then
18:29:24 Quit PaulJam (".")
18:29:27Mikachuconverting to png from jpg would just be silly
18:29:50 Join jaczehack [0] (
18:30:04jaczehackData −− in accordance with the prophecy
18:31:08 Quit mocker (SendQ exceeded)
18:31:37 Join Mmmm [0] (
18:32:13 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
18:33:14old45Mmmm> I plan to do some recording with my H120 will it providePIP for my electric mic?
18:33:15 Quit MarcoPolo (Read error: 54 (Connection reset by peer))
18:33:37Mmmmyep...2.5V I think...
18:33:57old45Good I asked this before but thought you would be good to ask
18:34:01amiconn_Genre9mp3: Your scanner seems to use the same lighting arrangement as most other flatbed scanners I saw - the light seems to come from below if you scan 3D objects
18:34:29amiconn_I recommend scanning the thing overhead and the flip the image 180 in the software
18:34:40amiconn_s/and the/and then/
18:35:11old45should I have an "eac test" folder in my root with just rockbox installed
18:35:55tucozold45, that is nothing rockbox installed
18:36:16 Quit amiconn (Nick collision from services.)
18:36:16 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
18:39:26 Quit dan_a (Read error: 110 (Connection timed out))
18:40:12Crackerizertucoz: what is your distro?
18:40:44 Quit old45 ("CGI:IRC")
18:41:05 Join TeaSea [0] (
18:41:09 Join old45 [0] (
18:41:13Genre9mp3amiconn: ok... rescaning now
18:41:27Mmmmhello again...
18:41:31Crackerizertucoz: do you have any problem building uisim?
18:41:42tucoztucoz, no
18:41:45 Part jaczehack
18:41:48tucozCrackerizer, no
18:42:06old45MmmmI plan to load some bird songs on the H120 would you use MP3 or Flac? I have them as WAV on my PC now
18:42:19Crackerizer:), i just built and installed sdl from the source.
18:42:45Crackerizereverything is fine. but i still cant build uisim.
18:43:05old45Soory I hit the enter key before I was ready , I am not calling you each time
18:44:06tucozCrackerizer, have you added sdl to your path
18:44:31Mmmmold45: well...depends on the quality you want, Flac is lossless so the quality will be the same as the original but filesize about 1/2 original, MP3 will lose quality but be much smaller.
18:45:32Crackerizertucoz: no, but i have to do that?
18:47:01old45Mmmm> That file difference for MP3 is what I would like to have an estimate for, FLAC is half. Is MP3 about 1/4 then?
18:47:19Mmmmdepends on the compression you use
18:47:19tucozCrackerizer, yes. The compiler has to know where to look for libsdl
18:47:40Mmmmold45 128kbs is about 1mb per minute
18:48:04Mmmmwhich is 1/10 original
18:48:17old45Thanks, that helps
18:49:22Crackerizertucoz: sorry, i asked a lot. I'm quite new to linux. :)
18:49:54tucozno problem. but maybe you should ask for linux questions in #linux?
18:50:26old45one more question , does rockbox play the FLAC and MP3 without problems for either
18:51:23tucozCrackerizer, there is also a #slackware channel in freenode
18:51:48 Quit Sinbios ("If the definition of a klutz is someone who doesn't have eyes on their ass, then yes, I suppose I am a klutz.")
18:52:37Mmmmold45: definitely MP3..I don't use flac much though so couldn't say for sure but I think so.
18:53:26Febsold45: It depends on the platform, but MP3 and FLAC should both be pretty much flawless on the H120.
18:53:53old45do you know a FLAC user, I have a lot of files to load and don't want to set it up and have problems
18:54:04Crackerizertucoz: is `UI_LCD_POSY' declared in SDL?
18:55:00 Join mantono [0] (
18:55:46Mmmmold45: I don't think you'll have problems...Just try one file to start with to give yourself some confidence and then do the rest when you know it works
18:57:00tucozCrackerizer, it is defined in uisdl.h. But I have to leave this to someone else as I've got to go.
18:57:03old45Yes that would be good to do, but I have hundreds of files to convert and load, so I don't want to have problems, that is why I asked for a FLAC user
18:57:28CrackerizerThat's ok, thank you tucoz. I'll figure out myself.
19:00:09Mmmmold45: if you can do one, you can do hundreds! :)
19:00:15 Quit XavierGr (Nick collision from services.)
19:00:19 Join XavierGr [0] (
19:00:32old45thats what I need to hear!
19:01:30 Part Crackerizer
19:04:01 Quit XavierGr (Nick collision from services.)
19:04:05 Join XavierGr [0] (
19:04:26Genre9mp3I hope I've done it right this time.....
19:05:28 Join Kohlrabi [0] (
19:07:10 Nick oxygen77 is now known as oxy77 (
19:08:22 Nick oxy77 is now known as oxygen77 (
19:09:00 Quit old45 ("CGI:IRC")
19:11:50 Join bbroke [0] (
19:18:34 Quit BigMac (Read error: 110 (Connection timed out))
19:19:35 Join lorijho [0] (
19:20:53 Quit Kitar|st (Connection timed out)
19:21:54 Quit Siku ()
19:22:00 Quit Hansmaulwurf (Read error: 110 (Connection timed out))
19:22:21 Join Hansmaulwurf [0] (
19:23:24 Join Nico_P [0] (
19:23:52 Join San||Away [0] (
19:27:03 Quit San||Away (Read error: 104 (Connection reset by peer))
19:29:57 Join San||Away [0] (
19:31:16 Join daurn|laptop [0] (n=quae@unaffiliated/daurnimator)
19:33:07 Quit spiorf (Read error: 54 (Connection reset by peer))
19:34:28barrywardelldoes the ipod bootloader use crt0.S?
19:34:48barrywardelllooking at firmware/ it doesn't seem to?
19:35:27barrywardellbut there are loads of #ifdef BOOTLOADER in crt0.S
19:38:30linuxstbbarrywardell: Yes, it does.
19:38:31linuxstbdan_a: Your crt0.S cleanup patch looks sensible to me. Although I think it would make sense to split the portalplayer code into a completely separate crt0-pp.S file, and leave only the generic ARM stuff there.
19:39:22barrywardelllinuxstb: what about line 10 of then?
19:40:01linuxstbIt just means that it's linked automatically, and the linker gives an error if it's mentioned in as well.
19:40:22linuxstbIt's because there is a symbol in crt0.S referenced by bootloader/ipod.c
19:41:16barrywardellah, ok
19:43:08lowlightlinuxstb: I have an preliminary attempt at metadata on the buffer:
19:43:22 Quit XavierGr (Nick collision from services.)
19:43:26 Join XavierGr [0] (
19:44:57***Saving seen data "./dancer.seen"
19:47:57FebsDo people *try* to make it difficult to help them?
19:48:43 Join Rob2222 [0] (
19:49:52 Quit tucoz ("Leaving")
19:50:20 Join dan_a [0] (n=dan-mirc@
19:51:57 Join Rondom [0] (
19:52:43 Join Coldtoast [0] (
19:52:46Coldtoasthi guys
19:53:09Coldtoasthow far along is Rockbox for the Nano? Compared to Rockbox for the h120?
19:53:10 Quit XavierGr (Nick collision from services.)
19:53:14 Join XavierGr [0] (
19:54:06 Join spiorf [0] (
19:58:39lowlightFebs: classic...
20:05:48 Join KlrSpz [0] (n=klrSpz@
20:06:04KlrSpzquestion, what's the 80mb partition on the 4gb nanopod for? is it safe to destroy?
20:07:05sharpeKlrSpz: Do you want to use your iPod?
20:07:30KlrSpzsharpe, do you have a point or are you going to ask assinine rhetoric?
20:07:44sharpeNo, there's a point to it.
20:07:52 Join freqmod [0] (
20:09:19KlrSpzok so how about we stop beating around the bush and get to details?
20:09:27KlrSpzi just want to know how it works
20:09:30KlrSpzram-based storage?
20:09:34KlrSpzfirmware storage?
20:09:35KlrSpzetc etc
20:10:04sharpeFirst off, it'd be 'an asinine rhetoric'
20:10:56sharpeSecond, it is where the firmware/bootloader are stored.
20:11:49KlrSpzthat's what i was looking for
20:12:26amiconnBagder: I have 4 hardware scans I would like to upload...
20:12:45KlrSpzso if i don't care about the original firmware, do i need all that space? can i repartition it smaller just to contain the bootloader?
20:13:22MikachuKlrSpz: yes, you can make it smaller, but no warranty
20:13:24sharpeYou can, but there's not much of a point to it.
20:13:32Mikachubig point on a nano
20:13:47Mikachuit's 4% of the space
20:13:51Mikachu(of 2GB)
20:14:01sharpeYes, but if he has to restore it, it'll be recreated anyway.
20:14:03 Quit mantono ("memtest")
20:14:15 Quit XavierGr (Read error: 113 (No route to host))
20:15:08KlrSpzi'm not concerned personally
20:15:15KlrSpzmy buddy is trying to do some experimental things
20:15:23KlrSpzminimal OS
20:15:30KlrSpzmaximizing the space
20:15:37KlrSpzbut like you said, 80mb is nothing to cry over
20:15:42Mikachui shrunk the bootpartition and created a third partition in the empty space
20:15:43KlrSpzand his is a 4gb drive actually
20:15:51KlrSpzhmm.. would rockbox mount it properly/
20:15:53Mikachuone extra album fits there
20:16:01Mikachunot out of the box
20:16:05sharpeJust be sure you use a dummy apple_os.bin, if you have to install the bootloader again for some reason.
20:16:26 Join lightyear [0] (
20:18:02 Quit Hansmaulwurf ("( :: NoNameScript 3.81 :: )")
20:18:02sharpeFor rockbox, anyway.
20:18:06Febslowlight: I'm beginning to think that ritesh is intentionally trolling.
20:18:14KlrSpzright right
20:18:20KlrSpzsharpe, is there anything else :)
20:18:35MikachuKlrSpz: you can move the start of the second partition too of course
20:19:00KlrSpzwhat did you use exatly? some partition magic or did you go low-level?
20:19:14Mikachui would be the lowlevel type of guy
20:19:23 Join sm007h [0] (
20:19:39KlrSpz[13:19] <Mikachu> i would be the lowlevel type of guy
20:19:42KlrSpzthis is actually for sm007h
20:19:47KlrSpzwhich network is this?
20:19:53sm007hshould have known it was freenode
20:20:06KlrSpzyeah, shoulda
20:20:11KlrSpzprolly coulda looked on the site :)
20:20:23sharpeNot too hard to find there...
20:20:23sm007hthat's too much bs
20:20:35KlrSpzso uhh.. how are you l iking it?
20:21:10sm007hgo get poe2
20:21:28sm007hit's a little over 1.5 gigs
20:21:29KlrSpzi can't, i'm at work.. no axx to my boxes since my server is down thanks to emerge ;)
20:21:37sm007hthat's crapola
20:21:41sm007hI dig rockbox
20:21:45KlrSpzactually i have no idea why it's down
20:21:46KlrSpzhell yeah
20:21:46sm007hworking out pretty dope
20:21:52KlrSpzi told you
20:21:52sm007hthis is shit
20:22:03sm007hyou're not going to play tomorrow?
20:22:14KlrSpzoh shit
20:22:17KlrSpzoh shit oh shit
20:22:19KlrSpzi'm like 22 min late for a meeting
20:22:22sharpeWell, we haven't heard it called pretty dope anytime recent.
20:22:33sharpeYou should get going then, eh?
20:23:26KlrSpzok nm, it was rescheduled.. *whew*
20:23:28sm007hspaz, I can't get partition magic to see my nano
20:23:29KlrSpztalk about a releif
20:23:38KlrSpzMikachu said he had done it low level
20:23:43KlrSpzprobably wrote directly to the mbr
20:23:51Mikachuhaha, not that lowlevel
20:24:03sharpeKlrSpz: Partition Table...
20:24:14Mikachui wouldn't imagine you can do it too easily in windows though
20:24:18sm007hwhat'd you use, fdisk?
20:24:25 Quit Febs ("CGI:IRC (EOF)")
20:24:42sm007hhmm, I'll use my lappy then and see if I can't get in there
20:24:42 Join Febs [0] (
20:24:53sharpeMikachu: Windows doesn't like multiple partitions on removeable devices... :D
20:25:14Mikachui know, it only sees the first partition when i plug it in on windows
20:25:16Mikachupretty awesome
20:26:15sm007hso I have to leave some of that first partiion?
20:26:30sm007h...gentoo, here I come
20:26:51Mikachuwow, i wonder why gentoo users have a bad rep :)
20:27:00ender`eh? i've had no problems with mulitple partitions on removabledisks on windows
20:27:10sm007hwhy would we have a bad rep?
20:27:22linuxstbIf you want to get rid of the Apple firmware completely, you can use the "scramble" tool instead of ipod_fw - that will make an firmware partition image containing only the Rockbox bootloader-xxxxx.bin.
20:28:09KlrSpzbut won't that still leave the partition space used? i think that's also an issue in this case
20:29:38linuxstbsharpe: It's not the multiple partitions Windows doesn't like about ipods - it's the fact that the first partition is marked as type 0 (empty). Neither Windows or Mac OS X will let you access it normally.
20:30:15linuxstb(but Linux is happy, and gives you a normal device node to access it with)
20:30:34sharpelinuxstb: I'm just speaking from experience I've had with usb drives that have multiple partitions, could never get one of them (out of two) to mount as a drive.
20:31:12amiconnlinuxstb: Does the apple loader require the partition to be of type 0?
20:31:37amiconnCould be an interesting experiment
20:31:50linuxstbamiconn: I'm pretty sure it does. I'll test now.
20:34:13linuxstbYes, changing it from type 0 causes a folder with exclamation mark to appear.
20:34:36sm007hso what did you do? just change the size?
20:34:42sm007hI've got a 4gb nano
20:35:06linuxstbI just changed the type of the first partition.
20:35:07sm007hbasically I need to leave some small partition?
20:35:13sharpesm007h: Yes.
20:35:46linuxstbYes - you only need about 100KB, but I'm not sure how small you can make partitions.
20:36:41sm007hI moved everything (.rockbox) off the second partition and thought I bricked my nano. but then I held down menu and select and I'm back to stock
20:37:02linuxstbInteresting message from fdisk when changing the type back to 0: "Type 0 means free space to many systems (but not to Linux). Having partitions of type 0 is probably unwise. You can delete a partition using the `d' command."
20:37:30sm007hthat's what it reports in disk manager in winxp "Free space"
20:37:34Mikachuyou should be aware the man page for fdisk says not to use fdisk
20:37:40sm007hhow much stuff is on it?
20:37:47linuxstbMikachu: Good job I've not read the man page then.
20:38:13Mikachusm007h: the partition is also used to dump ram by the original firmware because they couldn't be bothered to make it boot fast enough
20:38:31KlrSpzso there's your answer
20:38:39sm007hI read that in the forums. what I didn't know was whether rockbox needs it too?
20:38:44Mikachuit doesn't
20:38:55sharpeRockbox doesn't have a sleep mode.
20:38:56Mikachui don't know why they made it 80MB when the nano has 32MB ram though
20:39:05Mikachumaybe they expected the firmware to grow to 30MB one day
20:39:11 Quit JoeBorn ("Leaving")
20:39:19KlrSpzmaybe when they switch to x86 architecture :)
20:39:27Mikachufrying eggs with ipods
20:39:37sm007hI was going to say something along those lines...then I remembered, yeah, they went there ;)
20:39:45sharpeMikachu: Maybe they just wanted to be even and have it start at the 32mb mark, with a little breathing room.
20:40:15sm007hor they knew most people would be using it on windows ;)
20:40:26sm007hso it might have a habit of expanding itself over time
20:41:00sharpeWha? Copying 32mb to and from the disk doesn't make it grow.
20:41:02KlrSpzheh, i'm sure that had to be it
20:41:10sm007hit doesn't?
20:41:14KlrSpzno but the security patches do !!
20:41:19KlrSpzba da dummnnn tssss
20:41:36sharpeNow you're just rambling about windows. :)
20:42:09sm007hspaz, what do you use instead of fdisk. I havne't been on mylaptop in ages
20:42:15sm007hdont even know what's on here
20:43:18 Quit pondlife ("byebye")
20:47:23*preglow fondles octave
20:47:40sharpepreglow: That's beyond acting the fool...
20:48:09KlrSpzbut, i've never done any partition resizing
20:48:15preglowsharpe: i do other roles as well
20:48:24sharpeI see.
20:48:25 Quit damaki (Read error: 104 (Connection reset by peer))
20:48:36sm007hyeah, cfdisk won't open the drive
20:48:59preglowsharpe: among the roles in my reportoire, there is pervert, idiot, drunk, jerk and moron as well
20:49:05KlrSpzis dmesg seeing the drive?
20:49:10preglowalso, illiterate
20:49:13sharpeNow I get it.
20:49:15sm007hI can access it with fdisk
20:49:19KlrSpzhmm, odd
20:49:28KlrSpzyou're specifying the device right?
20:49:37sm007hno, that was retarded. I forgot the path
20:50:16KlrSpzmight help :)
20:50:18sm007hso how do I get all the stuff off it so I don't lose my firmware?
20:50:42KlrSpzdd if=/dev/xxx of=~/ipod_backup
20:50:52KlrSpzi wonder if you need to specify other options?
20:50:57KlrSpzlike block sizes or some crap
20:51:12KlrSpzduno, never done it
20:51:53sm007hwait, do I have it frm that script?
20:52:01sm007hin the ;output; folder?
20:52:17sm007hI have: apple.os.bin, bootpartition.bin, and rockboot.bin
20:53:39sm007hhavne't synced in about 8 monhts
20:53:48sm007hprobably be merging crapola all night now
20:54:11sm007hspaz, those are the original files right?
20:54:18sm007hisn't that what that script did?
20:54:38sm007hlinuxstb, you still around?
20:56:13KlrSpzsorry back
20:56:31KlrSpzwhat script?
20:56:39KlrSpzoh yeah
20:56:40sm007hthe one you gave me
20:56:41KlrSpzit's in output
20:56:57KlrSpzthat bitch still has my ipod
20:56:59KlrSpzi'm so pissed
20:57:00 Quit Mmmm (Remote closed the connection)
20:57:01KlrSpzshe's not in today
20:57:05sm007hI think I'm going to emerge gparted or something so I can try and do this graphicaly
20:57:16sm007hdude, she took it over the weekend
20:57:21 Join ghode|afk [0] (
20:57:36Coldtoasthey guys. What stage is RB for the Nano at? Compared with RB for the h120?
20:57:38sm007hthink mandy needs to show up at her house tonight
20:58:15sharpeColdtoast: The same as most of the other iPods, with the possible exception of the 3G, compared to the h120.
20:58:29Coldtoastand what stage IS that exactly?
20:58:36Coldtoastare they equal?
20:59:19sharpeIt's kind of hard to compare two different players... without and specific things to compare.
20:59:32ColdtoastI bought a nano last night. been using rb on my h120 for a long time now
20:59:49Coldtoastogg playback, mp3 playback, parametric EQ, gapless
20:59:53Mikachuit plays music fine, you will probably have to live without the eq though
21:00:01Coldtoastno EQ? damn
21:00:01sm007huh oh, no eq :D
21:00:08Mikachui haven't tried in a while
21:00:11Mikachuit could be that it works
21:00:22sharpeMikachu: As far as I know it's fine
21:00:28sm007hactually, there's an eq on there. seems better, if you like so sorts of things
21:00:30Mikachumaybe i had the peakmeters on too :)
21:00:53 Join mrobefan [0] (
21:01:05Coldtoasthas the interface translated well to the nano? as in is it a nice marriage between the nano hardware (clickweheel) and rb?
21:01:14ghode|afkdoes anyone know how i can find out what files are causing my tagcache building to fail?
21:01:21sharpeMikachu: Well, the Hardware EQ is fine too, for the 5G anyway...
21:01:27Coldtoasterr.. I should jhust say "ipod" I guess
21:01:31Mikachuyeah but nano doesn't have a hweq
21:01:34sm007hghode, I couldn't get it to work last night until I enabled diskcache
21:02:00sharpeI wasn't too sure if it did or not.
21:02:15linuxstbsm007h: I'm back now...
21:02:21Coldtoaston the h120 I wish you could set a permanent sleep timer actually
21:03:00ColdtoastI have this TERRIBLE habit of dumping my h120 into my bag and accidentally turning it on. then when I go to use it the next day, it's dead
21:03:01sm007hlinuxstb, whatd you do? change the part type, resize it, then change it back to 0?
21:03:24MikachuColdtoast: uh, there is a sleep feature
21:03:33MikachuColdtoast: as long as you don't start playing music too it should kick in
21:03:41Coldtoastyes. But it gets reset when you power off Mikachu
21:03:47Ribsisn't there a 'hold' button?
21:03:54linuxstbsm007h: I use fdisk in Linux. I've no idea how to do it in other OSes.
21:03:55Mikachuthen your settings aren't saving properly
21:04:00sharpeIt's more of a switch really...
21:04:05ghode|afksm007h: funny you say that since diskcache doesnt work either (ipod 5G)
21:04:10Coldtoastribs: of course. But I don't enable it on the remote and the player generally
21:04:13sm007hlinuxstb, I'm in linux
21:04:19RibsColdtoast, maybe you should start :>
21:04:52linuxstbsm007h: Then don't worry about changing the type - just delete the FAT32 partition, resize the first partition as small as you can, then create a new FAT32 partition taking up all the remaining space.
21:05:00KlrSpzspeaking of peakmeters, are those ever going to be conditional?
21:05:02ColdtoastRibs: pfft! I can't be arsed keeping tabs on whether I have both hold switches enabled
21:05:17linuxstbKlrSpz: Conditional based on what?
21:05:21KlrSpzon levels
21:05:22Coldtoasta persistent sleep timer set to, like, 1.5hrs would do it
21:05:36KlrSpzi'd like to do some graphical representations of a peak
21:05:39Coldtoastas I never listen to it for more than that. only takes me 45mins to walk home
21:05:41linuxstbKlrSpz: You mean something like displaying custom bitmaps for the different levels?
21:05:45MikachuColdtoast: just set the idle shutdown to that then...
21:05:52 Join TeaSeaLancs [0] (
21:06:07sm007hlinuxstb, where's the bootloader and misc files that are on that first part?
21:06:25ColdtoastMikachu: I tried today. tested. if I set it, power off then onn, the timer is reset
21:06:28 Quit TeaSea (Nick collision from services.)
21:06:31 Nick TeaSeaLancs is now known as TeaSea (
21:06:41sm007hor are they all sitting at the front of the partition so I don't have to worry about putting them back on?
21:06:43Coldtoastwith the daily from yesterday
21:07:01 Quit _FireFly_ ("Leaving")
21:07:21linuxstbsm007h: The first partition is the "firmware partition" - it originally contained the Apple firmware, a normal Rockbox install puts both the Rockbox bootloader and the Apple firmware there, and I'm guessing you just want the Rockbox bootloader there?
21:07:26ghode|afksm007h: also the tagcache build always fails on file 337 so i need to find out which file this is...
21:07:51sm007hlinuxstb, yes
21:08:32 Join lee-qid [0] (
21:08:44sm007hspaz, I have in 'output': apple_os.bin 5mb, rockboot.bin 5mb, bootpartitition.bin 80mb
21:08:44linuxstbsm007h: Do you have a copy of the Rockbox source code?
21:08:59sm007hI didn't grab that yet, no
21:09:08 Join _FireFly_ [0] (
21:09:27KlrSpzhe'd have to set up a cross compiler as well
21:09:44KlrSpzerr wait, which gcc you using sm007h?
21:09:57linuxstbsm007h: You can use the "scramble" tool in the Rockbox source to create a firmware image just containing the Rockbox bootloader.
21:10:01KlrSpzor did you devs get all that squared away
21:10:20linuxstbKlrSpz: I'm not talking about compiling Rockbox - just the "scramble" tool which just uses the native gcc.
21:10:38KlrSpzdoes that come with the source or something?
21:10:45linuxstbsm007h: Or I could run the scramble tool on the Nano bootloader for you if that's easier.
21:11:20sm007hlinuxstb, that'd be cool. thanks
21:11:36 Quit freqmod ("changing xserver")
21:12:20linuxstbAll you need to do is "dd" that directly to the first partition on your ipod - e.g. "dd if=bootloader-nano.img of=/dev/sda1" - assuming your ipod is /dev/sda
21:12:40sm007hyes it is
21:13:00 Join damaki_ [0] (
21:13:11 Join freqmod [0] (
21:13:17sm007hdon't really see any reason to go back to apple_os
21:14:15sharpeAh, nice to hear.
21:15:28 Quit shirour (Read error: 110 (Connection timed out))
21:16:43sm007hbecause that would be nice if you'd get that from me later tonight?
21:17:11sharpeUm. Yes. I love to hear from people that use rockbox...
21:17:14sm007hok, I'll stop seeding, but you can get it from my ftp anonymously
21:17:36sm007hare you at least going to get it from me?
21:18:50sm007hsharpe, lol, sorry about that, I was talking to spaz
21:19:04sharpeI know, it was a random thing, the second message.
21:19:39sm007hspaz, you going to get that from me later tonight then?
21:20:29KlrSpzreally really
21:21:26sm007hwth, are we up to 2006.0 or 2006.1
21:23:04sharpePretty sure we're in the year 2006 either way.
21:23:24sm007hone's likely experimental
21:23:35sm007hI don't want to use the wrong profile
21:23:50sm007hyou can't ssh into your box?
21:24:27KlrSpzme no
21:24:28KlrSpzi can't
21:24:32 Quit Genre9mp3 ("I don't suffer from Rockbox psychosis. I enjoy every minute of it.")
21:24:44sm007hok, you can't. but you're positive it's 2006.1?
21:25:01sm007hok, better not be dev :D
21:25:01KlrSpzor you can ask in #gentoo
21:25:05 Join vcardenas [0] (
21:25:11KlrSpzi asked like a few weeks ago
21:25:17KlrSpzthey said .1 was stable
21:25:31 Part andrewy
21:26:18sm007h2006.0 is stable
21:26:26sm007hI thought .1 == dev?
21:27:02sm007hnothing is every consistent with those guys, though
21:28:04sm007hFieldy> sm007h: not yet. it's just in the tree in advance of the actual 2006.1 release. there could be uknown bugs, or it could be incomplete.
21:29:11KlrSpzwell, wtf, they told me it was stable... oh well
21:29:14KlrSpzmayb ethat's what broke
21:29:19KlrSpzactually i don't recall ever changin mine
21:34:55 Part vcardenas
21:38:12sm007hspaz, are you going to do what I'm doing with your nano? you only have 1gb, too. 80 mgs could really help ;)
21:38:51 Join darkless [0] (n=darkless@
21:41:03sm007hcheck this out
21:41:06sm007hso I got the new kernel
21:41:25sm007hand I'm making it here, and I see that they now have 2100 intel wireless support. so I cnofigure it as amodule
21:41:41sm007hand it asks me if I want to "enable promiscuous mode"
21:41:44sm007hbut of course :D
21:42:10*linuxstb coughs and mentions this is #rockbox
21:43:14 Quit freqmod (Read error: 104 (Connection reset by peer))
21:44:57 Join freqmod [0] (
21:45:01***Saving seen data "./dancer.seen"
21:46:12KlrSpzparty pooper
21:46:39 Quit Coldtoast ()
21:46:47linuxstbIt's just that this channel is logged - so lots of off-topic chat makes it hard for someone to read the logs and find the useful information.
21:46:57 Quit TeaSea (Remote closed the connection)
21:47:50 Join TeaSea [0] (
21:51:13sharpeAnd... you don't want to say something you'll regret if it were logged, do you?
21:56:54 Quit lostnihilist ("Leaving")
21:56:55Rondomthere are also other use cases of promiscuous mode than the one you're thinking of
21:58:02sm007hthan whose thinking of?
21:58:36sm007hI know what it is, it was just a joke ;) goodness
21:59:51 Quit _FireFly_ ("Leaving")
22:16:02lowlightpreglow, linuxstb: had a chance to look at my "metadata on the buffer" patch? I've updated it to work with all formats but mp3 (hopefully I didn't break the metadata handlers).
22:17:55linuxstblowlight: No, not yet. I'm currently trying to get the COP working...
22:25:09 Join JoeBorn [0] (
22:25:35 Join webguest59 [0] (
22:25:37 Quit Pyromancer (Read error: 104 (Connection reset by peer))
22:25:39lowlightlinuxstb: ok. It's here when you get a chance:
22:25:44 Quit webguest59 (Client Quit)
22:25:52lowlightMy premise was to give the metadata handlers a pointer to a buffer to use (instead of assuming the id3v2buf)
22:26:18lowlightand have them return the size of the data written.
22:26:22 Join Icemaann [0] (n=flimpy@
22:26:42linuxstbWhat happens when the track is played and the buffer is then overwritten?
22:27:10 Join Pyromancer [0] (
22:27:16Rondomsm007h: my statement was directed to sharpe
22:27:16IcemaannIve just put rockbox on my ipod nano and Id like to play around with it *while* it is plugged into the usb. Im using linux, and I can eject it, but all the screen shows is "safe to unplug". Id like it to keep charging while I play with rockbox.... is this possible?
22:27:45linuxstbBoot into Rockbox, then hold MENU as you insert the USB cable.
22:28:29_jhMikeS_The updated encoder recording patch is ready for review (please read comments):
22:28:49 Nick _jhMikeS_ is now known as jhMikeS (
22:29:19sm007hoh that's how you do it. that's pretty cool. Are there any plans to make something that will let us use rockbox while it's hooked up, from the computer itself?
22:29:53lowlightlinuxstb: At this stage, there are no such checks.
22:30:53 Quit KlrSpz ()
22:31:09lowlightI guess the plan would be to copy the metadata buffer of the current track before a rebuffer.
22:32:21lowlightBut that requires some trickery since track.id3.artist is a pointer to the buffer.
22:32:47 Join BHSPitLappy [0] (
22:33:22 Quit _Veseliq_ (Remote closed the connection)
22:33:35 Join _Veseliq_ [0] (n=veseliq@
22:34:09linuxstblowlight: Does your patch just put the id3v2buf on the audio buffer, or the whole mp3entry struct?
22:34:58 Join lostnihilist [0] (
22:35:28lowlightWhatever was written to the id3v2buf is now written the the audio buffer.
22:36:24linuxstbOK. I thought the plan was to move the whole mp3entry struct to the audio buffer.
22:37:12lowlightI don't know the plan.:) What would the benefit be?
22:37:47jhMikeSand id3v1 (when preferred or the only thing avail.) converted to id3v2 first or something?
22:37:54linuxstbCurrently there is a static array of 32 mp3entry structs - the benefit would be getting rid of that.
22:38:21 Nick ghode|afk is now known as ghode (
22:38:35linuxstbIt also means the mp3entry struct (or its successor) could be more variable in size, and store things like track art.
22:39:57linuxstbjhMikeS: "id3v2buf" is just another case of a variable being called something which is no longer appropriate - it's used for storing various types of metadata nowadays.
22:40:24lowlightOk, but replacing the id3v1/2 static buffers saves the majority of the space.
22:41:03jhMikeSlinuxstb: then it should be renamed metadatabuf! There's a lot of mp3 centric nomenclature there no?
22:41:07lowlightalbum art would be like the other tags, just a pointer to the data.
22:41:52*jhMikeS thinks JPEGs/PNGs for album art is must have
22:41:56lowlightjhMikeS: until a year or so ago, rockbox only played mp3 :)
22:43:07 Join Bauermanthorn [0] (
22:44:03jhMikeSlowlight: I only found out about it in mid April 2006
22:44:13 Quit Nico_P ()
22:48:14*jhMikeS hopes someone will want to test that patch soon on non-x5 ports so his anxiety level can return to normal.
22:49:04linuxstbjhMikeS: There aren't many devs who record - if you want feedback, you could post some binaries in the Recording forum.
22:49:43linuxstbIt looks like a major change to recording, so it might be useful to get at least some initial feedback before committing to CVS.
22:52:11jhMikeSlinuxstb: There's no way I'm dumping that on CVS yet I would if it only affected x5 but it doesn't.
22:53:08jhMikeSlinuxstb: Mmmm and Petur aren't here or I'm sure they'd do it. Weren't you going to do iPos recording?
22:54:27linuxstbYes, I've got it working, but there is still a lot of tidying up and bug-fixing to do before I can commit. It's also now badly out of sync with CVS.
22:55:47jhMikeSI just synced with CVS before submitting it. What bugs are you getting at? What ports?
22:56:21linuxstbI mean my ipod recording work is buggy and out of sync with CVS - not your patch.
22:58:29 Join San [0] (
22:58:29 Quit San||Away (Read error: 104 (Connection reset by peer))
22:59:16 Quit gorey (Read error: 110 (Connection timed out))
22:59:39jhMikeSlinuxstb: ok...I'm like BUGS?!? Shouldn't be anything major! You scared me for a sec! :)(
23:00:29 Join freqmod_n [0] (
23:00:35preglowlowlight: nope, sorry
23:00:59preglowmy current scheduled items are speex and musepack
23:01:03preglowperhaps when i'm done with those
23:04:37lowlightlinuxstb: shouldn't get_metadata strip id3v2 data from all non-mp3 files? (re your response on the ML).
23:05:35jhMikeSIs it possible to be sure without adding another file section to have an entrypoint always at the end of a codec?
23:06:07preglowlowlight: i don't think it does
23:06:27preglowi think i remember having to do that per metadata format
23:06:46 Join San||Away [0] (
23:07:49lowlightIt doesn't, hence the common complaints about ogg, flac, etc not working when programs unknowingly put id3 tags on them.
23:08:53jhMikeSlowlight: I've been curious: where's the initial file parsing done? Tracking things can be a pain.
23:09:53linuxstblowlight: I thought id3v2 skipping had been implemented for FLAC.
23:10:20linuxstbjhMikeS: The get_metadata() function in apps/metadata.c is called before each track is loaded into the audio buffer.
23:11:02lowlightlinuxstb: flac yes, ogg no.
23:11:48linuxstblowlight: It feels wrong to do it though - the file is no longer a legal ogg stream if there's extra data stuck on the beginning or the end...
23:12:29 Quit freqmod (Read error: 110 (Connection timed out))
23:12:34jhMikeSwhat's actualy put in the audio buffer? Just the file contents after stripping metadata?
23:12:48 Join mantono [0] (
23:13:33linuxstbjhMikeS: I'm not sure, but I think in most cases the whole file is put there. Ideally, we would want format-specific loading functions which only put what is necessary into the audio buffer.
23:13:39lowlightjhMikeS: no, the whole file including metadata (well, id3v1 tags are stripped).
23:14:49Slasherinow tagcache supports browsing by numeric tags (for example year) :) just committed it
23:14:54jhMikeSlowlight: I thought you were just _adding_ that and that now how it is now.
23:15:12linuxstbSlasheri: That reminds me, does tagcache support a "disk number" tag?
23:15:24Slasherilinuxstb: hmm, that is not included yet
23:15:41Slasheriprobably because iirc id3 structure doesn't support it yet
23:15:50jhMikeSlowlight: What if id3v1 tags are the preferred or only tag?
23:15:51linuxstbjhMikeS: lowlight is adding buffering of the parsed metadata to the audio buffer. Currently that's stored in a static array.
23:16:01coobid3v2 supports disc number
23:16:23Slasheriyep, but rockbox doesn't have yet full support for it
23:16:49jhMikeSlinuxstb: that makes sense to me to do. that would/could support extended or custom tags
23:16:59Slasheriwhen support is added, it's very easy to add that tag to tagcache
23:19:51lowlightjhMikeS: in some formats you can't cut out the metadata. get_metadata preprocesses it so that it's readily available for the wps.
23:22:41 Quit San (Read error: 110 (Connection timed out))
23:23:16Slasherilinuxstb: hmm.. ogg files should have "date" instead of "year"?
23:23:28 Join ryanpg [0] (
23:24:09 Join belze [0] (
23:24:10ryanpgI installed rockbox on an ipod nano, it's neat - does this represent what you get on other platforms?
23:25:00jhMikeSlowlight: once it's formatted for wps then it can be overwritten except for what the codec needs? don't you parse it at each song when it's the next track's turn? I suppose I should study it carefully.
23:25:11Slasherilinuxstb: oh.. "Date can be in more any format in a Vorbis tag, so don't try to parse it."
23:25:18dionoearyanpg: pretty much. (some features like recording are device specific)
23:25:31Slasherithen that must be why tagcache is unabled to use the "year" tag for vorbis files
23:25:42ryanpgdionoea, cool thanks
23:26:50ryanpgis there a "play most listened to" option or "play least listened to"? my rio owning friend tells me he loves that feature... I've looked on the wiki
23:27:06ryanpgthe ipod manual doesn't reference it either
23:27:36belzei think i found a bug in the equalizer in rockbox using a iriver h110
23:27:48jhMikeSjust parse common date forms. If they're funky, ignore them. I don't see a problem with requiring certain format(s) if you want it to work with RB.
23:27:49lowlightjhMikeS: yes. Metadata parsing is done just prior to buffering the file (in playback.c).
23:28:05linuxstbryanpg: I think those kinds of features are "in progress"... So maybe soon.
23:28:10belzedepending on the volume you can hear the noise floor going up
23:28:50linuxstbbelze: You need to talk to preglow when he's around.
23:28:59jhMikeSbelze: too much dithering maybe...
23:29:10*blx just ordered a iAudio X5L 30G
23:29:11belzemaybe, it sounds like that
23:29:21jhMikeSCould be a side effect of applying bands low->high
23:29:57belzei have a 2db boost on 60 hz and 4db on the 12khz band
23:30:33 Quit ghode (Read error: 110 (Connection timed out))
23:31:29ryanpgvery good! rockbox is exciting! :)
23:31:29ryanpgthere was one guy on the list that was starting a port to the iaudio m5, now that the x5 is supported maybe it'll happen :D
23:31:29ryanpgerr... kinda supported
23:33:34jhMikeSx5 seems pretty damn supported...what am I missing out on?
23:33:52dionoeareread the sentence :)
23:34:13 Quit mantono ("restarting X")
23:34:17lowlightlater all.
23:34:20 Part lowlight
23:34:32jhMikeSbelze: If you apply bands low to high then noise/distortion (high harmonics) in low bands get boosted by the upper bands.
23:34:58belzei see
23:36:01 Join mantono [0] (
23:36:02jhMikeSbut if the bands above are cut the the noise it cut too...things might work better if sorted depending on cut/boost/frequency
23:38:37ryanpgjhMikeS, maybe nothing... the rockbox web site isn't always current
23:38:37ryanpgjhMikeS, I saw that it was listed as "usable" rather than "supported"
23:39:11Mikachuonly archos and hxx0 are supported i think
23:40:02linuxstbI would say everything with cvs/daily builds is "supported" - as in we think it's good enough for non-developers to use. Only Archos has had official releases, which mean we think it's actually release-quality.
23:41:04linuxstbiriver and iaudio are probably both close to release-quality. Ipods have a long way to go, but are still very usable.
23:41:06jhMikeSbetter order: boost bands->high to low then cut bands->low to high
23:41:20 Quit freqmod_n (Remote closed the connection)
23:41:26 Join illogic-al [0] (n=Your@konversation/user/illogic-al)
23:41:36illogic-alhi. rockbox crashed on me.
23:42:00illogic-alis there anything I can do beyond waiting for the battery to die so I can restart again.
23:42:13dionoeayou can press the magic key combination to reboot
23:42:14linuxstbWhat device are you running Rockbox on?
23:42:20belzejhMikeS: but thats nothing i can change without working on the sourcecode, i guess?
23:42:20jhMikeSme too but that's the price for changing things :)
23:42:29jhMikeSbelze: correct
23:42:53illogic-allinuxstb: h320
23:43:10jhMikeSIt's something I wanted to experiment with however so I'll probably test it out myself
23:43:27linuxstbI don't know about the h320, but the h140 has got a reset button on the bottom - you need a pin or paperclip to press it.
23:43:52 Join BockBilbo [0] (
23:43:56illogic-ali see it. on the side where mic is.
23:44:16jhMikeSjust scream in there and it will reboot
23:45:02***Saving seen data "./dancer.seen"
23:45:59belzeyeah, voice recognition for rockbox :P
23:48:15 Join Genre9mp3 [0] (
23:50:22 Join midgey34 [0] (
23:50:50*jhMikeS needs to make a silent wav file now
23:51:23*linuxstb stays quiet
23:51:57 Quit ender` (" Smoking is one of the leading causes of statistics.")
23:51:59 Part oxygen77 ("Leaving")
23:52:12belzejhMikeS: im not quite sure this is a problem with the equalizer. on ogg i still have that effect when the equalizer is disabled
23:53:03jhMikeSbelze: Just that ogg? Try another.
23:53:06barrywardellI read that the config settings are stored in a special part of the disk. Was it in the mbr?
23:53:10linuxstbbelze: I've generally found the h140 quite noisy, and I think others have commented about it.
23:53:46linuxstbbarrywardell: It's between the MBR (which is the very first sector on the disk) and the first partition (which normally starts around sector 63). Those 62 or so sectors are normally unused.
23:53:52belzelinuxstb: i know its noise, but on certain volume levels the noise floor gets louder
23:53:57belzewhich is kinda weird
23:54:03belzebut ill just test a few files
23:54:22barrywardellso the odds of doing damage to the disk are minimal then?
23:54:41linuxstbbelze: I'm not sure what you mean? I would have expected increasing the volume to increase the noise as well..
23:54:42barrywardelli'm ready to try a full rockbox build but am obviously a little wary
23:55:05 Quit sm007h ()
23:55:08linuxstbbarrywardell: I think your previous experience was either just bad luck, or a result of all those GPIO experiments...
23:55:33jhMikeSThere needs to be a way to make the .rockbox distribution folder without doing the zip thing
23:55:50MikachujhMikeS: you'd think so, wouldn't you? :)
23:55:58belzelinuxstb: no, it gets louder at a certain level and then softens if you increase the volume
23:56:16jhMikeSMikachu: I would indeed...I think I'll try adding that
23:56:28linuxstbbelze: Ah, OK. You're right - that doesn't make sense.
23:56:44barrywardelllinuxstb: you're probably right. I'll just take a chance and try it
23:57:05*barrywardell crosses his fingers
23:57:21linuxstbjhMikeS: I think is the place to look - I think it populates a .rockbox directory in current directory, zips it, then deletes it.
23:57:51linuxstbSo it should be easy to change that behaviour - e.g. to make it build the .rockbox directly onto the device.
23:57:55jhMikeSMikachu: Why are those .rwps files being skipped "due to size constraints"? I asked last night but don't know if you were around or if you don't know.
23:58:12MikachujhMikeS: they wouldn't fit on the lcd
23:58:29linuxstbOr they are too small for the LCD - I think only WPSs which fit exactly are now included.
23:58:34jhMikeSMikachu: it's display issure not bytes
23:58:48Mikachucome again?

Previous day | Next day