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 2005-02-08

00:01:07 Part GnagelRam
00:03:04preglowamiconn: it's not really important, Thom Johansen is my name
00:03:49 Join Eodun [0] (~Eodun@80.26.195.172)
00:04:15Eodunhi there, lots of ppl today :)
00:05:13 Join xen` [0] (~xen@ADijon-151-1-72-248.w83-196.abo.wanadoo.fr)
00:06:23 Quit webguest67 ("CGI:IRC (EOF)")
00:08:23 Quit ripnetUK ()
00:08:40[IDC]Dragonnite all
00:08:45 Quit [IDC]Dragon (" Like VS.net's GUI? Then try HydraIRC -> http://www.hydrairc.com <-")
00:09:06rashernight
00:14:58 Quit Hadaka (Read error: 110 (Connection timed out))
00:19:39 Quit stripwax (Read error: 110 (Connection timed out))
00:21:43*HCl realized he haven't had time to look at rockbox plugins..
00:21:57HClgee, my bad spelling o.o; realizes* hasn't*
00:22:44Eoduna pity I only have a player... but that nim one is really annoying anyway
00:24:55 Join Cham [0] (~3e234217@labb.contactor.se)
00:26:04Chamthere is a little probleme with teh firmware
00:26:29Chamwhen the butoon hold is on and we pressed play the player iriver power on
00:26:38Chamit will be better if it stayed off
00:26:42Chamno ?
00:27:03HClwhat?
00:27:15lolo-laptoppower on still works with the hold switch on
00:27:30HCloh.
00:27:40HClthats not really a firmware problem, i think o.o;
00:27:47Chamno sorry
00:27:50preglowbootloader problem
00:27:56preglow"problem", really
00:28:01HClwell, is it a bootloader problem?
00:28:04Chamno not a problem
00:28:08preglowdepends on where you want it
00:28:12Chami didn't use the good word
00:28:19preglowif you put it in the firmware, it has to be loaded before it detects that hold is on
00:28:20HClits easy to add code to turn it off when turning it on while hold is used
00:28:23preglowso battery is used
00:28:26HClyea.
00:28:27HClexactly
00:28:27amiconnrasher: dansk.lang voice:-fixed now, about to commit...
00:29:04Chambut it's a problem for my battery when my iriver is in my bag and power on
00:29:24preglowCham: yes, a fix should be easily done as well, so it'll probably be fixed
00:30:08 Quit CrunchyWhiteMeat ("Download Gaim: http://gaim.sourceforge.net/")
00:30:12Cham do not doubt it
00:30:21Cham do not doubt it
00:30:23rasheramiconn: great, I'll get to translating them tomorrow probably
00:30:31EodunWell, that's a point: the need to push the on button several seconds to turn it on removes that problem
00:30:56preglowit'd be more practical with a hold check
00:31:02Eodunbut if you can turn itonwith a little push, you'll need to make a new security check
00:31:32Chamthe HOLD button is here for that
00:31:43HCldid the original firmware not load if hold was on?
00:31:49preglowno
00:31:52preglowjust try
00:31:56HClno what?
00:31:56preglowit shows a image saying HOLD
00:31:58Eodunyeah, but the player hasn't got a hardware HOLD button...
00:32:00HCloh.
00:32:00HClok.
00:32:02preglowthen quits
00:32:18HClwell, i can't try since i have the rockbox bootloader now
00:32:23preglowbest would be if it just didn't do anything
00:32:25Chambut the hard disk didnt load
00:32:41HClpreglow: should be almost trivial to implement..
00:32:57preglowyes, but i'll leave it to linus, i do not want to brick my player
00:32:59ChamHCl :you push the record button and it loads the orginan firmware
00:33:01HClmhm
00:33:06HClCham: yes.
00:33:16HClCham: does it detect hold then?
00:33:24Chamyes it detect HOLD
00:33:26HClokay.
00:33:28ChamI test it
00:33:38HCli guess that makes sense.
00:34:35Eodunamiconn: is english.lang correct enough by now to consider the other updated languages as correct as it?
00:34:40*HCl sighs a bit and is a bit shocked by being confronted with iriver not being able to do up and left at the same time..
00:34:45Eodunawful language...
00:34:59amiconn???
00:35:00HClsuggestions on how to work around it are welcome :X
00:35:08preglowHCl: that's a real bummer
00:35:13HClpreglow: yea, well...
00:35:13lolo-laptopsigh, the 'd pad' on the iRiver is a bit depressing overall.
00:35:16EodunI mean if it doesn't have big bugs...
00:35:16HClstuff like pokemon
00:35:18HClwill be playable
00:35:26HCland final fantasy
00:35:33HClsince they don't require live action
00:35:38HClbut have a turn based system
00:35:38preglowquite lukcily i love rpgs ;P
00:35:49HClwell, i was hoping to play zelda.. but...
00:35:58preglowwhy can't you play zelda?
00:36:07Eodunlol, buy a psp...
00:36:10HClzelda has a realtime fighting thing stuff. what do you call it.
00:36:19preglowi know what you mean
00:36:19HClyou need to be able to move up and left
00:36:23HCland preferably press A
00:36:24HCltoo
00:36:25HClat the same time
00:36:30 Quit Marder ()
00:36:30preglowyou dont NEED to, but it's a plus :P
00:36:35HClno, seriously.
00:36:38HClyou can't beat its endbosses
00:36:40HClwithout it
00:36:44preglowok
00:36:48HCl :/
00:36:54amiconnEodun: english.lang is the master, so it is correct per defintion. Or do you mean something else?
00:37:06Chamnite
00:37:11*HCl is rather disappointed by all that.
00:37:18preglowwell, only thing i can think of is to make a separate control unit to plug in the remote port
00:37:22preglowbut that's not very practical
00:37:24HCli've been thinking for semi-solutions like a button on the remote
00:37:26HClto go up and left
00:37:30HClbut thats far from ideal
00:37:45HClbetter than needing extra hardware.. i suppose..
00:37:51 Quit Cham ("CGI:IRC (EOF)")
00:37:56Eodunamiconn: yeah, I was trying to know if espanol.lang, for example, is correct by now too, as I saw several updates of them these days
00:38:27HCli guess pokemon is still coolish.
00:38:35Eodunamiconn: sorry for my terrible english this late... my exams don't leave room for more
00:38:35amiconnquelsaruk overlooked some changes (only did the additions) in the first commit.
00:38:41HClmario also needs up and left movement at the same time....
00:39:37EodunHCl: what about a cards game? it should be easier
00:40:13 Quit jyp ("poof!")
00:40:56amiconnrasher: I have a problem with your realname in the commit message. bash doesn't like the 'ä', and I'm not sure whether cvs interprets the \0xx octal escape.
00:41:05 Join ripnetUK [0] (~mirc@82-70-100-230.dsl.in-addr.zen.co.uk)
00:41:11rasheramiconn: just put it as an a then
00:41:36amiconndone.
00:41:37rasher(please not 'ae' as the wiki did :|)
00:42:22amiconn'ae' would be the way to substitute 'ä' in german, that's why I was asking.
00:42:56amiconn2 .lang files to do...
00:43:00rasheryeah, but ä doesn't really exist in Danish, so usually I just put a when I can't use ä
00:46:53preglowahh, charset blues
00:46:58preglowi wish everyone went utf-8 :)
00:47:24*Eodun loves his spanish keyboard. (sure you don't have a ñ)
00:47:40EodunX-D
00:47:41amiconnpreglow: Somebody has utf-8 support for rockbox in the works
00:47:58preglowamiconn: who's making the font? :P
00:48:33amiconnUsing a conversion to, as now
00:48:56amiconnThe hardest part for unicode support is the necessity of a font cache
00:49:24xen`I dunno if I'll get the time before exams, but I'll be greatly interested in contributing for mod/xm/s3m etc... codecs
00:49:31amiconnYou cannot load the whole font in memory, at least not on the jukeboxes
00:51:18***Saving seen data "./dancer.seen"
00:51:35preglowxen`: and i'd be greatly interested in using those codecs
00:52:30preglowthat's also a consideration that has to be made when designing the codec api for rockbox: not all codecs are streaming
00:56:03linuxstbamiconn: strange you say that - I'm currently looking at FLAC metadata, which is stored in utf-8
00:57:27amiconnlinuxstb: Nothing strange. 'markun' already tested utf-8 support, and showed some screenshots. Must be somewherer in last week's logs
00:57:57 Quit ripnetUK (Read error: 60 (Operation timed out))
00:58:56linuxstbamiconn: I just mean it's a co-incidence, I'm working on a flac decoder for rockbox as we speak.
00:59:10amiconnAh ok.
00:59:50amiconnBtw, rockbox does already have to deal with unicode a bit (utf-16 in this case), when reading/ writing fat32 file names.
01:00
01:00:21amiconnVery limited implementation so far
01:00:34linuxstbI'm assuming OGG metadata is in UTF-8 as well. Anyone know about ID3v1 and ID3v2?
01:01:06xen`preglow: cool to see some ppl interested :)
01:01:07amiconnI don't know, but ID3 support is already there, so you can compare
01:01:15 Join bagawk [0] (~Lee@bagawk.user)
01:01:26rasherI think id3v2 defines encoding in some frame
01:02:00amiconnpreglow: You left most of the ### comments in (no big deal though, since I'm going through the whole file anyway)
01:02:13preglowxen`: you do now of the can of worms you're opening on yourself in getting proper xm playback, yes? ;)
01:02:25preglowamiconn: ahh, yes, i forgot about those
01:02:42preglowamiconn: i actually deleted most of the, but forgot about those that are on a line of their own
01:03:17amiconnThe "Partiton?" thing says "Batt. tomt?" in norwegian?
01:03:26xen`preglow: hmm I gonna do my best to not deceive music sceners ;)
01:04:03DMJC-Lyay
01:04:20preglowamiconn: hell no
01:04:27*Eodun is going to study for his exams...
01:04:29preglowamiconn: it should say "Partisjon?"
01:04:29Eodunc ya
01:04:30rasherxen`: just please don't base it on mikmod :|
01:04:33preglowamiconn: i swore i'd fixed that
01:04:44 Quit Eodun ("good luck")
01:04:47amiconnWhen? Maybe I didn't notice you updated...
01:05:06preglowamiconn: no, i haven't, i just thought i had
01:05:11*amiconn grabs latest version and starts over
01:05:14preglowamiconn: and while you're at it, i spotted another bug just beneath
01:05:23pregloweng: "No partition?"
01:05:36preglownorwegian should be "Ingen partisjon?"
01:05:54preglowamiconn: that is the latest versionm
01:15:38xen`ok
01:17:01preglowrasher: well, what can he base it on that does good playback?
01:17:05lolo-laptopin rockbox, can the player be used while recieving power from USB? On the stock firmware, it can either charge from USB, OR be in connected mode, but it cannot recieve power from USB and be in play mode.
01:17:13preglowrasher: there are precious few modplayers that play correctly
01:17:14lolo-laptop(in an iRiver)
01:17:19preglowrasher: especially opensource onews
01:17:30preglowlolo-laptop: not possible because of hardware
01:17:39preglowlolo-laptop: the player has to give up control of the harddisk
01:17:48HCllinus said mod would be impossible cause it needs multiple channels
01:17:58HClbut i dunno, timidity renders midi into a single audiostream too.
01:18:05HClso it should be possible to do it in software..
01:18:06lolo-laptoppreglow: no, I don't want to be able to have it in UMS mode and play mode same time, I want it to be taking power from USB and in play mode at the same time
01:18:08preglowHCl: software mixing
01:18:13lolo-laptoppreglow: taking power does not require giving up the HDD
01:18:14HClpreglow: mhm
01:18:27preglowHCl: it is very possible, most soundblasters also just had two channels
01:18:31HClyea.
01:18:40HCllinus was prolly thinking audio
01:18:42HClerrm
01:18:43HClhardware
01:18:44HClsorry
01:18:45HCli'm tired
01:18:47HClx.x
01:18:50*HCl goes to sleep :X
01:18:57preglowlolo-laptop: ahh, i misunderstood, i don't know then
01:19:14rasherpreglow: modplug
01:19:17preglowHCl: yes, but no players do it hardware any longer anyway
01:19:42preglowrasher: haven't tried that in a long time
01:20:02rasherpreglow: nor have I, but I've been told by people that should know, that modplug is way better than mikmod
01:20:19preglowahh, most things are better than mikmod, i think
01:20:24preglowit used to be decent
01:20:38preglowwe'll find out any way
01:20:45preglowmod support would rock, that's for sure
01:20:49preglowsid support :PP
01:21:46rasherI'm so hoping some poor soul would take that up
01:22:13preglowresid engine should port just fine, cycle accurate 6510 emulator might just be a bit intensive, though
01:22:52amiconnmodplug is inaccurate, at least the versions I tried.
01:22:53DMJC-Lcommodore 64 emulator...
01:23:06DMJC-Lget the old games going..
01:23:07 Join ashridah [0] (ashridah@220-253-121-151.VIC.netspace.net.au)
01:24:26xen`I'll take a look at these libs and check if its feasable
01:24:41rasherhttp://prdownloads.sourceforge.net/modplug-xmms/libmodplug-0.7.tar.gz?download <- there's libmodplug
01:24:44xen`anyways, xm support seems not that hard, I saw this mini fmod implementation in few kbytes
01:24:46preglowamiconn: everything is inaccurate
01:24:51xen`thx
01:24:54preglowxmplay is quite nice, though
01:24:57preglowbut not open source
01:25:17rasherI forget what it was people were bitching about in mikmod
01:25:41amiconnImho DeliPlayer does the best mod playback on PC
01:26:15amiconnIt even plays modules with embedded m68k code
01:26:25preglowdeliplayer plays everything, heh
01:29:27hubbleargh.. I've tried following the CrossCompiler wiki, compiled gcc-3.4.2 and installed, compiled binutils 2.15 and installed.. but binutils-cvs does not compile
01:31:14hubbleUnable to create ../bdf/bdf.h needed for size.o
01:31:31preglowcvs might be unstable just now
01:32:14hubbleany known good date for binutils cvs?
01:32:29ashridahhubble: you're running the wrong configure
01:32:43 Join Naked [0] (naked@naked.iki.fi)
01:33:15ashridahand make sure you do it from an empty directory the same way you do gcc and binutils originally as per IriverBoot.
01:33:47ashridahhm. i missed a bit there
01:34:07ashridahhubble: run the configure script in the 'src' directory you would have checked out via cvs, not the subdirectory 'binutils'
01:34:21 Nick Naked is now known as Hadaka (naked@naked.iki.fi)
01:34:24hubbleashridah: aha.. thanks.. going to try that
01:34:36ashridahheh. that one bit me too :)
01:35:51 Quit xen` ()
01:36:16hubbleashridah: yes, and when i'm successful in compiling it i'll try to do a writeup in the wiki =)
01:37:32cYmenwhat actually is the "voice" option in the menu good for?
01:38:05preglowverbal menus
01:38:07preglowrockbox talks to you
01:38:18cYmendoes it??
01:38:24preglowwell, if you want it to
01:38:34preglowsomeone has to make voice files first, though
01:38:43hubblenice, bootloader and rockbox works on my ihp-120 now =)
01:38:51bagawkhubble, :)
01:38:56cYmenah ic - i had hoped there d be a feature to voice controll it
01:38:59hubblebrowser is a little unstable though.. random crashes =)
01:39:05bagawkhubble, What all can you do?
01:39:45preglowamiconn: the translation giving you trouble?
01:40:51*rasher attempts to compile pcf2bdf
01:41:22hubblebagawk: file browsing, rockbox settings, sokoban..
01:41:26bagawkSilly microsoft platfirm SDK setup: does "Size: 15513" mean 15.513mb, 155,13 mb or somethign else?
01:41:42bagawkhubble, Fun :) any other plugins work?
01:41:59rasherHello world :)
01:42:50 Quit mecraw ()
01:42:53bagawkI thin it mean 155.13 mb :( progress bar does not appear to be going anywhere
01:43:02cYmenwhere is that stupid wakeup function hidden?
01:43:24bagawkcYmen, depends
01:43:42cYmenbagawk: give me a hint :)
01:44:01bagawk1. It may not be there unless 1. You have modded your erecorder, and compiled it with alarm or 2. you own a v2/fm
01:44:15bagawkcYmen, I beleive the system setup menu
01:44:57rasherAny ideas on how to compile pcf2bdf (from http://www.tsg.ne.jp/GANA/S/pcf2bdf/)..? I get this error:
01:45:03rasherpcf2bdf.o(.text+0xccb): In function `main':
01:45:03rasher: undefined reference to `operator new[](unsigned)'
01:45:09cYmenbagawk: erh..no :)
01:45:14bagawkrasher, linux app or windows?
01:45:23rasherbagawk: linux
01:45:48rasheror actually, I think it's almost ANSI C, because it has a makefile to compile with VC
01:46:11bagawkI would have to look at the source
01:46:20bagawkyou should do that :)
01:46:26ashridahuh. 'new' is a c++ keyword....
01:46:43rasherI don't know what I'm looking for though :\
01:47:11rasherashridah: sounds weird
01:47:13mrmagsyeah change the makefile CC from gcc to g++
01:47:16ashridahwell, if you're compiling c++ on linux, use g++ not gcc
01:47:28mrmagscrazy japanese
01:47:40rasherHah
01:47:52rasherwell that sure worked
01:48:01rashersilly silly japanese
01:48:03bagawkDid you use a makefile with the thing?
01:48:17rasherYes.. very simple though
01:48:24rasherI guess I might as well have used g++ directly
01:48:27mrmagshehe
01:48:41rasherIt's just a single sourcefile
01:48:48bagawkahh .cc .cxx or .cpp is generaly c++ :)
01:48:59bagawkYour case: .cc
01:51:01rasherCrazy japanese
01:52:02 Part hubble
01:56:07 Quit DMJC-L ("Leaving")
01:56:43 Quit lolo-laptop ("Client exiting")
01:57:19 Part Patr3ck
01:58:32 Quit preglow ("offz0rs")
01:59:37 Quit bagawk ("Leaving")
02:00
02:00:05rasherwell that sure did work
02:01:03amiconnArgl, preglow went away...
02:04:39amiconnAnyway, norsk.lang looks good, only 2 minor issues
02:05:08rasherMmm, Atadore font looks nice
02:05:24rasherand now that I can convert pcf to bdf, sky's the limit!
02:06:25rasherdammit, typo in dansk.lang
02:06:58rasherLANG_BOOKMARK_SETTINGS_MAINTAIN_RECENT_BOOKMARKS
02:07:12rasher"Vedligehold liste a seneste bogmærker?" -> "Vedligehold liste af seneste bogmærker?"
02:09:47 Join knoppix_ [0] (~knoppix@cpe-24-94-54-216.stny.res.rr.com)
02:09:55knoppix_hey okay so.. im about to flash this fw...
02:10:24knoppix_i get an md5sum of the edited hex file to be like the EU version shown...
02:10:41knoppix_is it okay to flash it now?
02:10:56knoppix_because, its not the same as in the table of everything else..
02:11:04rasherThere have been no bricks so far
02:11:13rasherI think there were modifications to the bootloader after that
02:11:16knoppix_okay so i should be set?
02:11:24rasherBut don't blame me if it breaks
02:11:29knoppix_haha
02:11:30rasherAlthough it shouldn't :)
02:11:36knoppix_well...
02:11:41rasherh120 or h140?
02:11:51knoppix_where does everyone get the firmware md5 numbers they put in the table?
02:11:52knoppix_h120
02:12:36knoppix_rasher: you got be0bf0e1df367e5cffd9e27eff988ef1
02:12:41knoppix_where from though?
02:12:57rasherthat's md5sum ihp_120.hex (the edited hex file)
02:13:32rasherWith a h120, you may have to wait for 4-5 minutes on the first boot, as it recalculates free space on the drive
02:13:40knoppix_yeah
02:13:42knoppix_alright
02:13:43knoppix_hmm
02:13:44rasherthat's happened to every h120 so far
02:13:48knoppix_yeah
02:14:14knoppix_did you compile everything in your own environment? or just use the bootloader.bin?
02:14:15amiconnrasher: In case you find more mistakes/ typos, collect them, fix .lang file (based on cvs version), and tell me tomorrow, or submit to the patch tracker on sf
02:14:30rasheramiconn: alright, I will
02:14:47rasherknoppix_: actually, someone sent his to me, because I don't have a build environment set up
02:15:01knoppix_i see...
02:15:12knoppix_do you have any idea of when linus will be on?
02:15:30rashernope, not at all
02:15:30knoppix_*does anyone
02:15:32knoppix_k
02:16:25knoppix_rasher: would you happen to have that bootloader.bin file that you received from someone else?
02:16:38rasherYeah, I can send it to you.. email?
02:16:44knoppix_sure...
02:16:47knoppix_newhouseb@gmail.com
02:18:23rasherdone
02:18:44rasherweee, a bunch of the plugins work
02:20:47amiconnXShocK: Could you please check LANG_SETTINGS_BATTERY_PLAYER again? It should say the equivalent of "Partition?", but seems to say the equivalent of "Battery low". It's also too long, player string is limited to 11 chars (not that it matters much; the player is unable to display cyrillic anyway).
02:21:21XShocKI have a question about rtc for h120. since there is no rtc for it, a question to dev's, how difficult would be making a circuit yourself and put it in the box?
02:21:29amiconnSame goes for LANG_SETTINGS_BATTERY_RECORDER ("Battery low?" -> "Partition?"). This one can be displayed
02:21:45XShocKok. i will check now
02:23:16XShocKwhat does it ask? "what is the partition?" or "should i partition it?"
02:23:42amiconnNo. It's a message which is displayed when saving the settings failed.
02:24:27amiconnIt was first thought the most common would be low bat, but then found it's that the partition is missing.
02:24:32amiconnHence the change...
02:24:50XShocKÐàçäåë
02:25:38rasherall of the plugins appear to work (except screen/keyboard problems)
02:25:50amiconnXShocK: What?
02:26:34XShocKdo you want me to send the equivalent of "Partition?" here to irc? it is "Ðàçäåë?"
02:26:49XShocKso i just copy-paste it there. :)
02:27:00amiconnArgl, diffenrent code table here. I'll copy verbatim, then check
02:27:03 Join lostlogic [0] (~lostlogic@node-4024215a.mdw.onnet.us.uu.net)
02:27:53XShocKi think it will work out.
02:28:17amiconnXShocK: Could you give "No partition?" as well? (The longer variant for graphic displays)
02:28:37XShocK"Íåò ðàçäåëà?"
02:29:06ashridahknoppix_: that md5sum is from an earlier build of the modded firmware
02:29:12XShocKactually the player is capable to display cyrilic, but you need to change font in the player to win_crox1c
02:29:35amiconnNo, the player has no font support
02:30:03amiconnCharcell display only, with merely 4 or 8 user-definable characters
02:30:19XShocKHmm... I probably don't understand you. I use this Russian menus already, and it works perfectly.
02:30:39amiconnWhich device?
02:30:47XShocKiriver h120
02:30:47rasherhttp://www.rockbox.org/twiki/bin/view/Main/PluginIndex <- I guess we should start adding h100 to that
02:31:20amiconnXShocK: There it'll of course work (graphic display, like the archos recorder etc.)
02:31:31amiconnI'm speaking of the Archos player/Studio units
02:31:49XShocKAah. my bad. :)
02:40:57amiconnXShocK: You didn't use uplang, did you?
02:41:21XShocKI used it
02:41:32XShocKI deleted all ### stuff later
02:42:25amiconnNo, that's not what I mean. There are some strings in it that are marked as deprecated in english.lang
02:42:55amiconnHowever, that was a recent change by me
02:43:38XShocKIf I remember right those are with eng: "" ?
02:43:56amiconnNo, those have desc: DEPRECATED
02:44:00XShocKok
02:44:05amiconneng: "" may also mean voice only
02:44:17XShocKok
02:44:21amiconn(buit these have usually id: VOICE_*)
02:44:27amiconn*but
02:44:55amiconnI'll run uplang, if there's no other change that should do
02:45:41XShocKok. i will be here 2 hours more, so if something needed i am here
02:46:07knoppix_rasher...
02:46:09knoppix_sorry, i left
02:46:14knoppix_but thanks a lot for the file
02:46:17rasherknoppix_: did it work?
02:47:37cYmendoes the alarm clock mod allow setting the wakeup time from within rockbox?
02:47:54knoppix_i havent tried it yet...
02:49:47XShocKso, will it be hard to implement rtc for the iriver?
02:50:27XShocKI mean in hardware.
02:50:40amiconnXShocK: Could "Ñâîáîäíî:" be shortened to "Ñâîá.:" ? (It's a bit too long)
02:51:19***Saving seen data "./dancer.seen"
02:51:24XShocK6 first letters would be better actually. but fine
02:51:46 Quit linuxstb ("Leaving")
02:52:48knoppix_*prays* as he flashes his fw
02:54:23knoppix_umm
02:54:46knoppix_it should hang with a blank screen for a bit right? (on an h120)
02:54:59knoppix_recalculating something...
02:55:06XShocKyes
02:55:12knoppix_k good...
02:55:20XShocK3-5 minutes
02:57:24knoppix_alrighty, well i have to go gentlemen...
02:57:28knoppix_it looks like it's working
02:57:31knoppix_i'll be back
02:57:38knoppix_thanks for all your help
02:57:48 Quit knoppix_ (Remote closed the connection)
03:00
03:01:11 Quit cYmen ("blub")
03:03:31rasherHurray
03:06:39 Join CrunchyWhiteMeat [0] (~CrunchyWh@12.145.185.82)
03:06:49XShocKhehe... under Ubuntu gcc 3.4.2 compiler perfectly.
03:07:23*rasher starts using dansk.lng
03:07:30rasherthis really is horrible
03:08:34 Quit midk (Read error: 104 (Connection reset by peer))
03:08:43 Join midk_ [0] (~midk@c66-235-14-120.sea2.cablespeed.com)
03:09:16lostlogichas anyone gotten this md5 on bootloader.bin? 75a81a96b7814abb4e932f0809945875
03:11:06XShocKi got 1dbd7e2b2494cf54639117dccdc68599 bootloader.bin
03:11:16mrmagsI built from cvs a few hours ago with gcc 3.4.2 and got: 6b9075eec74ea42b437684af764367de
03:12:14lostlogicI'll take those for nos... I built with gcc-3.4.3 and binutils 2.15.94 *shrug*
03:15:23XShocKthe easiest way to check if anyone had that md5sum is to search in irc logs
03:17:16amiconnXShocK: Committed russian.lang
03:17:57rasheramiconn: I just fixed a couple of lower/uppercase problems
03:18:12rasherwhat is "Caption backlight" ?
03:18:25mrmagsooh I was wondering that too
03:18:33rashersong lyrics?
03:18:34XShocK:)
03:18:56amiconnSwitch on the backlight shortly before a track change, switch off shortly afterwards
03:19:20rasheraha
03:19:32rasherI'll change it to "Light on songchange"
03:19:38rasheror the Danish equivalent
03:21:57rasherwhoa the diff on dansk 1.12->1.13 is large
03:22:09amiconnyup.
03:22:11*rasher diffs his small fix
03:22:36amiconnsame goes for norsk 1.1->1.2
03:22:53amiconnThese weren't updated for 21 months
03:24:15rasherherm
03:24:24rasherI think it's better if you apply these fixes manually
03:24:32rashermy lang file hasn't had the voice fixes
03:25:09XShocKwhat happened with 07 Feb logs in irc?
03:25:50mrmagsamiconn: damn, that is a nice feature.
03:26:07amiconnXShocK: They're there, just reload the http://www.rockbox.org/irc/ page
03:26:46XShocKno, i mean 7 Feb log is messed up
03:26:59XShocKhttp://www.rockbox.org/irc/rockbox-20050207.txt
03:27:11rasherhow so?
03:27:14rasherlooks fine here
03:27:42XShocKhmm
03:28:42 Join jpburton [0] (knoppix@cpe-24-94-54-216.stny.res.rr.com)
03:28:47XShocKcompared to 06 Feb which is fine, the 07 one does not have line breaks, maybe i did something
03:28:53jpburtonokay, so... i installed the fw...
03:29:02rasherXShocK: it has linebreaks here :\
03:29:26jpburtonand, it brings up the error "Loading firmware / Result: -1"
03:29:30jpburtoni could just be stupid
03:29:41rasherjpburton: did you put a .rockbox on your drive
03:29:45jpburtonnope
03:29:50jpburtonthats what i thought...
03:29:51rasheralright
03:29:54jpburtonthanks
03:30:03rasherdownload this: http://www.rockbox.org/daily/h100/rockbox-h100-20050207.zip
03:30:08rasherand unzip in your drive root
03:30:26mrmagsooh the dailys started showing up? nice
03:30:32rasherApparently :)
03:30:54rasherthis is the first one
03:31:52rasherAnyone know about the bleeding edge builds?
03:32:14amiconnTime to sleep now.
03:32:17amiconnNite.
03:32:54rasheroh, nevermind, they're being (attempted) built
03:33:45jpburtonk thanks
03:33:48jpburtonwooohooo it works!
03:33:55jpburtonthanks rasher...
03:34:05 Part amiconn
03:34:39mrmagsso, earlier today someone was asking me if I knew any player that supports midi. and I said "no but cowon/jetaudio/iaudio might do if you ask nicely. also, rockbox-supported players could do." and then I came here and ppl were talking about implementing it. heh. rockbox, well, rocks.
03:34:51rasherjpburton: you're very much welcome
03:49:05XShocKI finally built it myself, and flashed, everything works.
03:49:37rasherwhere "everything" doesn't include "sound output, remote, battery status" :)
03:49:59XShocKyea. :))
03:51:01ashridahXShocK: as in both the bootloader and the firmware from the latest cvs?
03:51:47XShocKaaam.. no. daily build. :)
03:51:54ashridahright
03:52:30ashridahrasher: lets not forget usb, sane button assignments, and screen aspect ratio adjustments
03:52:47rasherashridah: heh
03:52:48XShocKI would really like to write something for it. :)
03:54:30XShocKvery pity that I am 100% dumb in the embedded systems.
03:54:56 Part CrunchyWhiteMeat
03:56:38HClmrf
04:00
04:11:17 Quit QT (Read error: 60 (Operation timed out))
04:30:13lostlogicWhee, just added my own board images to the H3X0 hardware page, wonder if they are helpful
04:34:51lostlogicOh, and a side note: The US version of H3x0 has a clock, but the intl version doesn't...
04:35:11lImbusa clock ? you mean a clockchip ?
04:35:12lImbuswoot
04:35:19lImbusthat'd be nice
04:35:32lostlogicwell I don't know if the intl has the chip or not, but I know that the US displays a clock on the screen and the intl doesn't.
04:36:08lostlogicit is probably a hardware difference, because the intl doesn't have DRM and that's probably what the clock chip is used for.
04:36:14XShocKi have an unanswered question, how hard would it be to implant clock into ihp1x0 ?
04:36:47ashridahXShocK: i don't think it has the hardware to keep the clock running when powered off.
04:37:07ashridahthat said, lots of things are possible with messy hardware mods
04:37:09XShocKbut clock can be attached directly to battery
04:37:37lImbuslostlogic: if you own a scanner, it's always a good idea to scan your board.
04:38:08lImbustho your pictures are of very high quality already
04:39:05lostlogiclImbus: yeah, I don't have a scanner... but my digicam is pretty good ;-)
04:39:25lImbusic
04:40:25XShocKyea, you are right, messy. but still, would it be hard to make that? i mean i think making a counter of time is not very hard, but how to connect it to player? does it have any possible extensions?
04:42:13ashridahwell, supposedly there's a headerless serial port. that'd work in a pinch, i guess.
04:43:11ashridahi just don't see why you'd want a clock that badly
04:43:33XShocKyou are right. :)
04:48:24 Quit jpburton (Read error: 110 (Connection timed out))
04:49:13XShocKtime to sleep, see you all. thanks for all your help
04:49:26 Quit XShocK (" HydraIRC -> http://www.hydrairc.com <- Try something fresh")
04:49:38lImbuspoof
04:49:49lImbustoo fast to say g'bye
04:50:23lImbusit's almost 5 am here. gotta go to bed too
04:50:27lImbusgood night all
04:51:05 Quit lImbus ("KVIrc 3.0.1.99 'Realia'")
04:51:23***Saving seen data "./dancer.seen"
05:00
05:06:44 Join CrunchyWhiteMeat [0] (~CrunchyWh@12.145.185.82)
05:12:21 Join elinenbe_ [0] (trilluser@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
05:14:19 Part CrunchyWhiteMeat
05:14:35 Join HCl_ [0] (hcl@titania.student.utwente.nl)
05:16:06 Quit HCl (Read error: 113 (No route to host))
05:36:42 Join cYmen [0] (~cymen@nat-ph3-wh.rz.uni-karlsruhe.de)
05:56:31 Join fuk3d [0] (~dev-@adsl-209-30-158-50.dsl.rcsntx.swbell.net)
05:56:55 Part fuk3d
06:00
06:46:01 Join courtc [0] (~court@adsl-158-3-127.asm.bellsouth.net)
06:51:25***Saving seen data "./dancer.seen"
06:52:46 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
06:57:21 Join LinusN [0] (~linus@labb.contactor.se)
06:58:47 Join BigCrock [0] (~email@r42h108.res.gatech.edu)
07:00
07:03:59 Quit MooMaunder (Read error: 110 (Connection timed out))
07:04:01 Quit elinenbe_ (Read error: 110 (Connection timed out))
07:04:14 Quit BigCrock ("—I-n-v-i-s-i-o-n— 2.0 Build 3515")
08:00
08:04:05 Nick HCl_ is now known as HCl (hcl@titania.student.utwente.nl)
08:51:26***Saving seen data "./dancer.seen"
09:00
09:03:08 Join lImbus [0] (lImbus@237-44.244.81.adsl.skynet.be)
09:03:18lImbusg'mornin
09:03:41LinusNmorning
09:05:13dwihno'allå 'allå
09:05:45 Join Zagor [242] (~bjst@labb.contactor.se)
09:06:35 Nick Lynx_awy is now known as Lynx_ (HydraIRC@134.95.189.59)
09:06:38Lynx_hiho
09:24:55Bagderhappy birthday LinusN!
09:25:18rashermorning
09:25:21Bagderyou oldie
09:25:27Zagor:)
09:25:38 Join bobTHC [0] (~foo@l05v-4-93.d1.club-internet.fr)
09:25:41LinusNthanks
09:25:49bobTHChi folks!
09:25:58LinusNsalut
09:26:07bobTHC:)
09:31:59dwihnohaiduk!
09:37:38lImbusurg. not that one
09:38:25lImbusoh! happy birthday Linus !
09:39:04LinusNthx
09:55:01dwihnooho! happy b-day!
10:00
10:04:35HClhappy bday :p
10:05:53HClwhat'dcha get? :p
10:06:06LinusNa year older :-)
10:09:55lImbus:D
10:10:49Lynx_happy birthday linus :)
10:13:33 Quit bobTHC (Read error: 110 (Connection timed out))
10:14:18 Join bobTHC [0] (~foo@l05v-4-93.d1.club-internet.fr)
10:19:25 Join midk [0] (~midk@c66-235-14-120.sea2.cablespeed.com)
10:27:12 Join Patr3ck [0] (~patr3ck@pD9ECEAE1.dip.t-dialin.net)
10:28:11 Quit midk_ (Read error: 110 (Connection timed out))
10:30:44 Quit lostlogic (Read error: 110 (Connection timed out))
10:47:53CassandraHappy birthday, Linus.
10:51:29***Saving seen data "./dancer.seen"
10:55:25 Join lostlogic [0] (~lostlogic@node-4024215a.mdw.onnet.us.uu.net)
11:00
11:11:07 Join linuxstb [0] (~linuxstb@dsl-212-23-31-215.zen.co.uk)
11:13:07linuxstbHappy Birthday.
11:13:31linuxstbAm I right in saying that there is no malloc in Rockbox?
11:13:38Zagorthat's right
11:13:53Zagorwe'll need to add some for the codecs, which use a dedicated buffer
11:14:08linuxstbI was just about to mention the codecs.
11:14:29Zagor:)
11:14:48linuxstbI'm working on FLAC, and even though it has it's own internal wrapper functions to malloc which we can rewrite, they are not used everywhere.
11:15:16Zagorsince we don't have a malloc, it's ok to use that function name
11:15:57 Join QT [0] (as@area51.users.madwifi)
11:16:41linuxstbSo does that mean you plan to implement malloc/calloc/realloc in Rockbox?
11:16:52Zagoronly for the codecs
11:16:56 Join Cassandra_ [0] (~christi@213.78.125.106)
11:17:10 Join oNe`Neo [0] (~Info@pD9EAB6EA.dip.t-dialin.net)
11:17:28Zagoreach codec will get a dedicated memory buffer to use, which it can spend any way it likes. such as using a malloc.
11:18:11Zagorthe "real" way is of course to use static worst-case buffers, but that may make it harder to merge updates later
11:18:21linuxstbRunning my programs under Linux, is there an easy way for me to intercept calls to the libc malloc/calloc/realloc functions so I can see how much memory libFLAC is using?
11:18:42Zagoryou need to use wrappers
11:21:44oNe`Neoany motherfucke tke my counterstrike server
11:22:02oNe`Neosomeone knows schoki?
11:22:21ZagoroNe`Neo: i think you're in the wrong channel
11:22:29oNe`Neono i dont think so
11:23:05oNe`Neoi know his ip adress
11:23:29oNe`Neoand yesterday someone was here with the name schoki
11:23:38oNe`Neowith the same ip adress
11:24:44#>>"seen" used by Zagor (~bjst@labb.contactor.se) [snoop prevented]
11:25:06Zagorwe don't care about counterstrike here. this is a development channel.
11:25:25 Part oNe`Neo
11:26:50Zagorsushi time
11:27:32linuxstbAnother Linux/C question, is there a way I can see which C library functions a particular executable is using? i.e. similar to ldd but more detailed?
11:27:36sneakumsglibc has some debug stuff you can enable with environment variables
11:27:45sneakumsfor malloc
11:27:53sneakumsre ldd, what extra information do you need?
11:28:05sneakumsoh, functinos.
11:28:14sneakumsyou can use nm to see which symbols a program uses
11:28:22sneakumsthe ones with a 't' are defined externally, iirc
11:28:28sneakumser, functions that are defined externally
11:28:52sneakumsexcept it doesn't work on stripped executables, i guess
11:29:27linuxstbsneamums: Thanks, that's perfect. Looks like "U" identifies external functions.
11:29:39sneakumsobjdump -T looks like it works on stripped executables
11:29:45sneakumsobjdump has a bunch of useful stuff
11:29:57linuxstbGives lines of the form lseek@@GLIBC_2.0
11:30:08sneakumsi assume when objdump says *UND* it means undefined and not underpants
11:30:25 Quit Cassandra (Read error: 110 (Connection timed out))
11:30:31sneakumsobjdump -T will demangle that for you
11:30:41sneakumsthe @@ things are the symbol versioning stuff
11:33:12linuxstbThat shows that apart from the memory allocation functions, libFLAC isn't needing anything else.
11:33:30 Join B4gder [0] (~chatzilla@neptunus.contactor.se)
11:34:06B4gderis "Dorian Gray" here?
11:34:29lImbusnot me, gotta go
11:34:33 Quit lImbus ("bbl")
11:34:35dwihnoI got his painting here, you can use it for identification ;)
11:34:47B4gderI'm not fond of the latest CrossCompiler edits
11:35:04B4gderthey're incorrect and adds duplicate info
11:35:30dwihno:/
11:37:00B4gderI guess I'll just edit it
11:38:33B4gdera lot
11:44:30linuxstbsneakums: Yes, objdump -T is even better than nm. Thanks.
11:48:22B4gderdone
11:50:51B4gderit gets tricky documenting things when Gmini enters, as that is an Archos too, just different
11:51:10B4gdercalling them "SH-based Archos" is not very intuitive for users
11:53:41 Join walnut [0] (~garryh@frink.nuigalway.ie)
11:55:22 Join webguest37 [0] (~8ccb1035@labb.contactor.se)
11:56:17 Quit webguest37 (Client Quit)
11:59:55 Quit walnut ("[BX] If idiots could fly, IRC would be an airport")
12:00
12:00:21sneakumsor an aviary, or a beehive.
12:15:33 Join lImbus [0] (~manuel@kernel.cycos.net)
12:15:39lImbusre
12:21:39 Join elinenbe_ [0] (trilluser@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
12:26:44 Quit JJ-Demon ("—I-n-v-i-s-i-o-n— 2.0 Build 3515")
12:27:32dwihnoDamn I'm good at packaging stuffs!
12:29:01 Join elinenbe__ [0] (trilluser@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
12:29:14 Quit elinenbe_ (Read error: 104 (Connection reset by peer))
12:29:14 Nick elinenbe__ is now known as elinenbe_ (trilluser@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
12:33:10linuxstbIf anyone is interested, it seems that libFLAC needs about 150K of memory when decoding 44.1KHz files. I haven't tested other files yet. My stripped-down (decoding only) libFLAC.a itself is anout 65K when stripped.
12:34:57linuxstbSorry, that 150K just refers to malloc'd memory, not stack usage etc.
12:35:45B4gderdoes it malloc a lot?
12:36:34B4gdercurrently there's nothing in Rockbox that uses malloc, so we can adjust the malloc to the codecs' needs
12:38:06HClyea
12:39:05B4gderI would assume that it mostly does a few mallocs at start and then frees them when done
12:41:05linuxstbIt does some large mallocs (4 of about 18K and 1 of 64K) at the start during initialisation, and then some reallocs to vary some pointers from around 200-500 bytes (they grow and shrink), and then some small mallocs when it parses the metadata.
12:41:42B4gderok
12:42:46linuxstbNothing changes during actual playback - just at the first initialisation (creating a decoder instance), and then each time you reset the decoder to start reading a new file.
12:44:04B4gderI think I should bring back the malloc code we once used
12:44:48dwihnoBring in the beast!
12:44:59B4gder:-)
12:45:18dwihno:-)
12:45:47linuxstbThe alternative is to hack each codec so it doesn't use them. Which probably isn't a good idea.
12:46:14B4gderI think we should do as little modifications of the codecs as possible
12:46:42linuxstbB4gder: I agree. My FLAC implementation doesn't change anyrthing from the standard code.
12:47:22HCl :P
12:47:22sneakumswill gapless payback pose much difficulty?
12:47:43sneakumsi assume you could buffer and initialise multiple files
12:48:33B4gdermy guess is that gapless is not our biggest problem
12:48:53linuxstbWe will have two buffers - a "compressed data" buffer and a "PCM data" buffer. As long as the PCM data buffer is never empty, we will get gapless playback.
12:48:59sneakumssweet
12:49:07sneakumsyou'd think iriver could have managed that
12:49:24*HCl personally prefers a pause between songs...
12:49:34sneakumseven for songs that are mixed together?
12:49:40HCleither pause or crossfade
12:49:40sneakumsand live albums with audience noise?
12:49:47linuxstbI think gapless playback is something that needs to be designed for at the very start - it's probably difficult to add if the whole audio playback system isn't designed around it.
12:49:57HClwell, i never listen to the first, but i guess you're right with the second one
12:50:01HCli have a single album thats like that
12:50:31sneakumsheh
12:50:34sneakumsi have a bunch of both
12:50:49sneakumsalthough a crossfade mode would be nice for use when shuffle is on
12:50:55linuxstbHCl: most tracks have silence encoded in them at the start and end (which is why there is a gap on your CD player), and Rockbox will still play that silence. "gapless" just means there is no extra silence.
12:51:04HCllinuxstb: i know.
12:51:19HCllinuxstb: most of mine don't have gaps at the end, and its just really annoying if they seamlessly start afterwards.
12:51:22HClwinamp does that...
12:51:30HCli couldn't even find the option to add a gap between songs
12:51:31***Saving seen data "./dancer.seen"
12:51:31HCl>.<
12:51:38sneakumsi thought it had one
12:51:47ashridahlinuxstb: need to make sure you handle the fact that mp3s can't have a short frame too.
12:51:49HCli found a plugin that could do it.. but it was malfunctioning...
12:52:07ashridahthat's what probably bites hardware players that don't think ahead during design
12:52:35ashridah(well, part of it, iriver's firmware tends to start reading early)
12:53:08linuxstbI think there was talk about an option in Rockbox to strip out any silence at the end of MP3s - thus solving the short frame problem.
12:53:43*ashridah nods
12:54:44sneakumsall i have is vorbis files, so the keep-the-pcm-buffer-filled strategy should suffice :)
12:55:00ashridahheh
12:55:10*HCl still has loads of mp3s, mostly because people he downloads from have mp3s
12:55:25HCl :/
12:55:34ashridahi have a mix of mp3s from a cd collection i had before it got stolen from my house (they left the computer, took stuff they could sell to what was probably their dealer), and a bunch of oggs from cds i've had after that
12:55:40*B4gder only has mp3s since Archoses can't play anything else ;-)
12:55:46ashridahdon't have an flac, not sure i can be bothered reencoding everything
12:55:49HCl :X
12:55:51Lynx_there is a prog that removes silence from start and end of mp3's
12:56:21HClwell, i don't have the original albums, so. i can't really reencode anything of it.. :/
12:56:38sneakumsi re-encoded a bunch of stuff when vorbis went 1.0
12:56:42sneakumsthat was a project and a half
12:56:51sneakumsworth it, though
12:56:54*ashridah notes that there's a coldfire-based player on the market (iaudio or something possibly) that plays flac.
12:56:59sneakums1.0 sounded a lot better, and the files worked out smaller
12:57:04sneakumsiaudio m3, i think
12:57:07B4gdermeeting
12:57:12ashridahsneakums: that's the one
12:57:18sneakumsthey look nice
12:57:34ashridahstill, i'll be happy when rockbox can play stuff, since then i'll have tetris and my mate that bought one of the m3's won't :)
12:58:17HCllol.
12:58:48HClhas anyone looked at the tetris plugin yet to fix size / orientation?
12:58:51HClfor the iriver
12:59:03ashridahnot to my knowledge
12:59:26sneakumsjust write a gb emulator and run gameboy tetris</facetious>
12:59:33sneakumsgb tetris is about the only one i like to play
12:59:43HCllol.
12:59:46HClok ok...
12:59:52HCli'll take a look at it today.
12:59:57HClonce i finished my anime
13:00
13:00:01rashersneakums: just play rockblox!
13:01:44 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
13:02:03ripnetukdoes rolo work on iriver?
13:03:46Zagorno
13:04:37ripnetukok
13:04:51 Join R3nTiL [0] (~zorroz@217.30.249.151)
13:06:37 Quit ripnetuk (Client Quit)
13:07:54*rasher quietly notes that the Boxwave usb+charger cable for iRiver appears to work
13:08:01 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
13:08:12HClwhats rolo?
13:08:26ripnetuklike lilo but loads rockbox not linux
13:08:33ripnetuklets you load other foirmware from rockbox
13:08:37HCloh..
13:08:57HClshouldn't be hard to get it to work
13:08:59ripnetuki was hoping to be able to load my hacked one that just jumps to original; software to provide a kind of menu option for original iriver rom without a reboot
13:09:07HClmhm.
13:09:11ripnetuki guess thats extrememly lazy ;)
13:09:18HCli guess...
13:09:37HClif you're taking that much trouble to not have to press a button.. i'd say flash it with a bootloader that loads iriver by default for now?
13:09:43HClor am i thinking too simple? :P
13:10:55HClum... what does TEST_PLUGIN_API do?
13:11:01LinusNhmmm, seems i forgot something in the boot loader
13:11:11HClLinusN: mm?
13:11:29LinusNit's supposed to start the original if it can't find rockbox.iriver
13:11:36HCloh
13:11:37HClyea.
13:11:42rasherThat it doesn't
13:11:43HCli was thinking about why that wasn't in there.
13:11:55HCli figured it was due to all the settings the bootloader did before loading rockbox
13:12:10 Quit elinenbe_ (Read error: 54 (Connection reset by peer))
13:12:16 Join elinenbe_ [0] (~elinenbe@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
13:12:40LinusNi'll fix that asap
13:13:58ZagorLinusN: how about adding a hold check while you're at it?
13:13:59 Join Patr3ck_ [0] (~patr3ck@pD9ECFF0B.dip.t-dialin.net)
13:14:10ashridahheh.
13:14:38dwihnoHow much of a mpeg audio violation is a truncated frame?
13:14:39HClyea
13:14:49Zagordwihno: truncated?
13:15:54dwihnoZagor: Was just thinking about frame length and such
13:16:29 Join kurzhaarrocker [0] (~Phil@p5487C659.dip0.t-ipconnect.de)
13:17:14LinusNZagor: ok
13:18:18Zagordwihno: the bit reservoir is the problem that is causing gaps in mp3 files
13:18:29HClhrm
13:18:49HClmy rockbox suddenly crashed while going to browse hdd, it just access the hdd and doesn't do anything..
13:18:55HClbrowse plugins*
13:19:24ZagorHCl: yes, several people have reported file browser crashes on iriver.
13:19:53HClZagor: i've had it crash before.. but not like this
13:19:56LinusNZagor: and added zero samples at the end of the song, because the length has to be in even frames
13:20:13LinusNthat's the main reason for gaps
13:20:21HClmaybe i accidentally clicked through onto the battery test..
13:20:44 Quit kurzhaarrocker (Remote closed the connection)
13:21:06LinusNtodays daily fixed one browser bug (was actually a lcd scroll thread bug)
13:21:16HClk
13:22:20dwihnoZagor: ah, okay
13:22:55rasherI guess for gapless to work with mp3s (to take up an old conversation), you'd need something like iRiver's "Gap delete" that skips the empty frames at the end of songs
13:23:10LinusNrasher: yes
13:23:29linuxstbOr encode them "gaplessly" (i.e. put the start of the next track in the end of the previous track)
13:23:56sneakumsare mp3 frames a fixed size?
13:24:14Zagorrasher: the frames aren't empty, they are half-empty. so we need to remove empty samples.
13:24:21linuxstbMP3 frames are a fixed time duration - i.e. a certain number of ms per frame.
13:24:51rasherZagor: oh-ah.. sounds annoying
13:24:54linuxstbThe frame size in bytes depends on the bitrate.
13:25:00 Quit ripnetuk ("Leaving")
13:25:34Zagorrasher: actually it's not so bad. we'll do it in the pcm layer.
13:26:23linuxstbDoes Rockbox have any concept of samplerate at the moment - i.e. is anything special done when giving the MAS a 48Hz MP3 frame after giving it a 44.1KHz frame?
13:27:36LinusNno, nothing special at all
13:28:08linuxstbThe PCM buffer will need to take care of that - i.e. know the samplerate for each block of data in it.
13:29:23 Quit Patr3ck (Read error: 110 (Connection timed out))
13:29:59 Join webguest46 [0] (~5189b144@labb.contactor.se)
13:30:09webguest46hi all
13:30:55LinusNhi
13:31:38linuxstbLinusN: is audio playback next on your iRiver to-do list, or are there other things you want to do first?
13:31:39webguest46Rockbox scramble and descramble tools.
13:31:44webguest46Rockbox scramble and descramble tools.
13:31:46webguest46I was looking for the
13:32:21LinusNlinuxstb: usb handling, bootloader fixes, performance (faster clock)
13:32:29HClwebguest46: they're in the tools directory.
13:32:34webguest46sory I tryed pasting the text in to mesage but it was posted strate up
13:33:26linuxstbLinusN: OK. That gives me more time to work on the codecs then.
13:33:40LinusN:-)
13:36:33linuxstbOut of interest, what codecs are people interested in seeing support for? I'm personally only interested in MPEG and FLAC, but I'm happy to help implement OGG and WAV if no-one else gets there first.
13:36:33dwihnoyay! price drop!
13:38:08rasherMP3, OGG, SID, Tracker file formats is my list
13:38:33dwihnotracker formats is cool and fun
13:38:48dwihnoI think there are non-float open sourced code for that.
13:39:15rasherI saw something for ARM pdas, do they have fpu?
13:39:30linuxstbIt may be useful to add some info on those formats to the audio codecs table in the iriverport Wiki page.
13:39:54rasherOkay, I'll see what I can find
13:40:13webguest46is ther a compiled version of mkboot ?
13:41:04linuxstbAC3 (as implemented in a52dec) is also on my list - you can then play audio rips from DVD or digital TV without needing to re-encode (and lose quality).
13:41:16HClwebguest46: read the development howto on the wiki.
13:41:45Zagorwe can add any number of codecs once the framework is in place
13:42:12Zagorlinuxstb: are you looking more at the codecs themselves or the framework at the moment?
13:42:50linuxstbZagor: I'm looking at the codecs in detail first, so I can then sensibly put forward suggestions for the framework.
13:43:01Zagorsounds good
13:44:05linuxstbI've almost finished with FLAC,, so will probably start with libmad tonight. I've already written a MP2 player using libmad, so I'm already quite familiar with it.
13:46:33dwihnolinuxstb: great!
13:47:30rasherdwihno: do you know where to find the non-float tracker library?
13:49:33dwihnorasher: have you checked libmikmod?
13:49:59Zagor<rasher> xen`: just please don't base it on mikmod :|
13:49:59dwihnoI think it is, or some spinoff
13:50:02rasherheh
13:50:14 Quit elinenbe_ (Read error: 60 (Operation timed out))
13:50:20dwihnowhat's wrong with the mikmod library?
13:50:48rasherIt lacks support for many features (or so I'm told.. I don't really use tracker formats much) in many file formats
13:50:59rasherI think modplug is better
13:51:04rasherbut if mikmod is fixed-point
13:51:12dwihnohave you checked modplug?
13:51:18rasherI'm looking
13:56:41rasherhrm
13:56:45rashercan't find it now
13:58:02Lynx_http://modplug-xmms.sourceforge.net/, or did you mean you can't find some info on it?
13:59:02rasherYeah, I stumbled over a version of either mikmod of modplug ported to some pdas, but I can't find that now
13:59:54rasherand dwihno said he thought there were non-float libraries around
14:00
14:00:05rasherI don't know if either modplug or mikmod is non-float
14:02:05 Join elinenbe_ [0] (~elinenbe@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
14:04:01webguest46thanks for the ponters bye
14:04:31 Quit webguest46 ("CGI:IRC (EOF)")
14:07:00 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
14:09:03Lynx_rasher: the only thing i find is a libmikmod based player for some korean ARM9 based handheld called mikplay32
14:11:07ripnetukIf you add a hold check on the iRiver, it needs to still be able to boot the original
14:11:16ripnetukeven if hold is on, as it might not be on on the remote
14:11:43ripnetukand it would require hold to be off on the main unit to tuen on by remote
14:11:55HClwhats the include/library dirs i should use when compiling stuff for rockbox?
14:14:39LinusNripnetuk: so it should still start if it is started with the remote?
14:15:01LinusNHCl: please elaborate
14:15:20ripnetuklinus - yes
14:15:32sneakumsthere are two hold switches, one on the unit, one on the remote.
14:15:33ripnetukmaybe it should refuse to load the iriver.rockbox if hold is on
14:15:38LinusNripnetuk: the so be it, i an read the remote hold switch too
14:15:41LinusNcan
14:15:41HClLinusN: um, i'm trying to compile gnuboy for m68k-elf outside of rockbox, so its not being built by the make stuff of rockbox..
14:16:18LinusNyou need to link with the firmware library
14:16:19HClbut its lacking library functions, obviously
14:16:20ripnetukin an ideal world, it would work like the original, ie hold is specific to main unit or remote, so it should only check one
14:16:27sneakumsHCl: i was kidding when i suggested that, but it'd be cool if it worked
14:16:30ripnetukdepending on if its tuned on wiht remote or main
14:16:37HClsneakums: i wasn't :P
14:16:45sneakumsrock on, then
14:16:47LinusNripnetuk: should be easy
14:17:05ripnetukthe original firmware seems to check the correct hold - if i have hold on the main unit, i have to hold on on the remote for longer than usual (untuil the original firm boots)
14:17:16ripnetukotherwise i just get the hold screen and it turns itself off
14:17:24LinusNi'll fix
14:17:36ripnetukim guessing the boot loader should handle this, otherwise too much time passes before the actual firmware starts.
14:17:45ripnetukso you have access to remote buttons already then? nice
14:18:48ripnetukalso, /IDEALLY/ it would check the record thing a little quicker (maybe before initing ATA stuff), so the on button on the remote is still down when the original boots
14:19:03ripnetukand the main hold is on
14:20:25bobTHCi dont if that can help for the framework but i give the link : http://www.tml.hut.fi/Opinnot/T-111.550/Mobileaudioformats2004-10-26.pdf
14:21:24LinusNripnetuk: it does check the rec button before initializing ata
14:22:06LinusNbut there's a sleep() there for debugging purposes that can be removed
14:23:21HCldo plugins have to be a single .c? :X
14:23:30LinusNcurrently, yex
14:23:32LinusNyes
14:23:35HCl :X
14:23:52*HCl has nightmareish visions of a 2mb .c file that makes up gnuboy
14:24:02LinusNhehe, we can fix that
14:24:11HClplease :X
14:25:23 Quit lostlogic ("Going to the moon")
14:25:25linuxstbHCl: it's not perfect, but you could always #include .c files
14:25:45HClew...
14:25:47HCli guess.
14:25:53HCli guess that might work for now...
14:27:03linuxstbBut if the codecs are eventually going to be plugins, then we need a more flexible build system for them.
14:28:01LinusNthey will not be the same kind of plugins as the current ones
14:28:13LinusNthey will be codec-plugins
14:28:57 Quit elinenbe_ (Read error: 60 (Operation timed out))
14:29:38ripnetuklinus - i see...
14:30:03ripnetukand... happy birtday
14:30:06ripnetukbirthday
14:30:10LinusN:-)
14:31:50 Quit R3nTiL ()
14:33:33 Join elinenbe_ [0] (~elinenbe@207-237-225-9.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
14:37:48HClmrf..
14:37:55HClwell, the include .c doesn't work nicely.
14:38:13linuxstbHCl: I didn't think it would be trivial.
14:38:18HClmhm.
14:38:35HClwell... i can't really work on it without having a semi-nice build environment.
14:38:42HCllike, each module in its own dir
14:38:45Zagoradd it as a second library
14:38:51Zagorwe already have a lib dir in plugins
14:38:57HClhrm.
14:39:22Zagorthat will get you forward, even if it's not an ideal solution in cvs
14:39:30HClyea.
14:39:34HCli'll do that
14:45:03 Join jyp [0] (~jp@64.198-200-80.adsl.skynet.be)
14:45:41lImbussalut
14:46:04jypsalut lImbus ;)
14:46:32dwihnohaiduk
14:50:11 Quit B4gder ("ChatZilla 0.9.35 [Mozilla rv:1.5/20031007]")
14:51:33***Saving seen data "./dancer.seen"
14:53:14HClgeeze. gcc is configured to be strict about warnings.. heh.
14:54:01 Join cYmen_ [0] (~cymen@nat-ph3-wh.rz.uni-karlsruhe.de)
14:54:43HCldoes rockbox provide an implementation for memset/memcpy? or shall i just write those myself?
14:55:05 Quit cYmen (Read error: 104 (Connection reset by peer))
14:55:30ripnetukhow will the clock speed adjustment work? will it be as simple as if there is any idle time, bump it down otherwise bump it up?
14:56:55HClyou'll prolly want to have a little margin...
14:57:06ripnetukyeas to prevent ossicilation
14:57:26HClat least, i know windows has this nasty habit with cpu scaling that it actually assigns too little cpu speed to make the system run smoothly
14:57:35HClbut rather assigns the minimum cpu speed needed
14:58:32ripnetuki have limited experience of windows cpu scaling... i think my laptop (pentium-m) does it, but ive never noticed any slowness on battery
14:58:56HCli remember wanting to watch a movie
14:59:04HCland windows assuming that 300mhz would be enough
14:59:14HCland having to set it to 1.6 ghz manually
14:59:15ripnetuk300mhz is not enought for my watch ;)
14:59:41linuxstbHCl: Yes, memcpy and memset are both used by functions in the firmware directory, so they must be implemented somewhere.
15:00
15:00:08HCllinuxstb: hm, then they're not in the headers somewhere...
15:00:15HClat least, gnuboy can't find them
15:00:17HClbut ok
15:01:03linuxstbThey are in firmware/include/string.h - not sure if they are exported to the plugins though.
15:01:35HClokay
15:01:39HClthanks
15:02:06linuxstbFound them - apps/plugin.h
15:02:58linuxstbYou need to call them using something like rb->memcpy
15:03:17HClok
15:03:18HClthanks
15:05:46HClits a shame gnuboy's code is poorly documented and has lots of warnings...
15:06:47linuxstbHow do the specifications of the gamboy's LCD compare with the iRiver?
15:08:07LinusNthe cpu speed will not be dynamic
15:08:09ashridahlinuxstb: from vague memory, i think they're around the same size, but i suspect the iriver has a higher DPI
15:08:17LinusNat least not to begin with
15:09:06sneakumsi think the gb was 160x160, but i don't know for sure
15:09:20 Quit Nibbler (Remote closed the connection)
15:09:35sneakumsah, 160x140
15:09:39HClyea
15:09:42 Join Nibbler [0] (~sven@port-212-202-78-188.dynamic.qsc.de)
15:09:43HCl160x140
15:09:45linuxstbAnd the iRiver is...
15:10:00HClless.
15:10:14linuxstbHow many greyscales on the Gameboy?
15:10:17HCl4
15:10:20HCland 4 on iriver
15:10:26HClso thats good
15:10:30sneakumsthe iriver is 160x128
15:10:31sneakumsdanm it
15:10:42HClthough rockbox supports only black/white at the moment
15:11:59linuxstbSo we could just crop the 12 pixels.
15:12:05linuxstb(I mean lines)
15:12:20HClyea.
15:12:23HClwe've discussed that alraedy
15:12:27HClalready*
15:12:33sneakumsprobably wouldn't matter for tetris to lose the top 12 lines
15:12:35 Join preglow [0] (thomj@s183a.studby.ntnu.no)
15:12:36HClmy primary concern is getting it running first
15:12:36linuxstbSorry. Or display them on the remote :-)
15:13:43rasherI added some info on the IriverPort page about sid and tracker formats by the way, don't think I mentioned that
15:13:50sneakumsman, i hate "forums" and "boards" and verdammt "lounges"
15:14:55 Join Greenwell [0] (Greenwell@DSL212-235-23-3.bb.netvision.net.il)
15:15:32linuxstbrasher: thanks for that. But I don't understand what you mean in the description of SID by "no notion of playing time".
15:16:00rasherWell, sids don't mark when they end
15:16:11rasherafaik
15:16:13preglowand you can't really predictably seak in it
15:16:15preglowseek
15:16:20rasherthat too
15:16:37preglowyou can "seek" forward by speeding up the 6510 emulator
15:16:42preglowbut backwards, no
15:16:55preglowstuff like this really needs to be taken into consideration when designing codec framework
15:17:00linuxstbBut I assume they have a start and an end? It's just that the playing time is not known without playing the whole track?
15:17:07preglowperhaps have two types of codecs, streaming and non-streaming
15:17:13rasherlinuxstb: no, they literally don't have an end
15:17:18preglowlinuxstb: no, they might very well loop forever
15:17:18rashermost of them just start over at some point
15:17:27rashersome just keep playing silence
15:17:40sneakumsso are they sort of more like a program than anything?
15:17:46rasherYes
15:17:49preglowlinuxstb: and if it has an end, it just stops making sound, you can't tell if it's over unless you start detecting silence in the player
15:17:51rasherin fact they're very much like a program
15:17:58preglowthey _are_ programs
15:18:07rasherThey're *that* much like programs
15:18:12preglowa sid file is an executable 6510 program
15:18:21linuxstbOK, I'm understanding the concepts (and the problems) now.
15:18:57preglowbut yes, i believe separating streaming and non-streaming codecs might be valuable
15:19:13linuxstbMaybe they could be implemented as "viewers" - similar to the text file or jpeg viewers (and games). Outside the main codec system.
15:19:42linuxstbWe would need to give plugins the ability to play PCM audio anyway.
15:19:48preglowthey're going to be plugins, probably
15:20:09preglowbut i'd very much like the ability to choose which ones i'm going to use a lot, so that they can be contained in memory at all times
15:25:17linuxstbThey could still be part of the codec structure, but we would need to introduce the concept of an infinitely long track. You would have to manually press "next" to skip to the next track. We would also have to identify some filetypes as non-seekable.
15:26:14rasherwell as I noted, there exists a database with playing times for pretty much any .sid
15:26:33rasherso if that was used, most songs wouldn't be infinitely long
15:26:43rasheralthough it could still happen in a number of cases
15:26:43linuxstbAre there any standards for saving that information to files to store on your iRiver?
15:26:55rasherit's just
15:27:00rashera flat textfile
15:27:03rasherof the format
15:27:16rashermd5sum playingtime1 playingtime2
15:27:17rasheretc.
15:27:37rasher(each file contains 1 or more songs)
15:27:48 Part LinusN
15:28:02rasherProbably not the most straight-forward format to support
15:28:08rasherto put it mildly
15:29:25linuxstbBut it's useful to be aware of them when thinking about the codec framework.
15:31:06linuxstbAre the "tracker formats" similar, or do they have a defined end-point?
15:31:22rasherI think they have defined end points
15:31:26rasherand no concept of subsongs
15:31:34rasherthat should be mostly straightforward
15:31:42rasherapart from the multitude of slightly different formats
15:32:02rasherbut that'll just be bound to whatever library you end up using I guess
15:32:08sneakumsi think some tracker formats have a way to specify a jump at the end of the pattern list
15:32:17sneakumsbut that's based on vague memories of playing with a tracker years ago
15:32:18preglowyes
15:32:25linuxstbThe subsongs issue is something I already care about - I use cuefiles with long MP2 files to identify tracks. I want to add support for cuefiles to rockbox as well at some point.
15:32:26preglowthere are mods that loop forever
15:32:34preglowmost players let you choose if they should ignore loops
15:32:49preglowlinuxstb: cue files mix poorly with mp3s
15:33:08linuxstbpreglow: why do you say that?
15:33:24 Join lolo-laptop [0] (~lostlogic@68.251.84.226)
15:33:27preglowbecause mp3 doesn't support sample accurate seeking
15:33:49preglowyou have to compromise, or you have to scan the entire mp3 up to the point of the track you want
15:33:57preglowthat's the solution the foobar2000 developers have chose
15:33:57preglown
15:34:11linuxstbBut we're no longer just talking about MP3 files.
15:34:21preglowahh, mp2
15:34:22linuxstbFLAC supports sample-accurate seeking, for example.
15:34:32 Quit lolo-laptop (Client Quit)
15:34:36preglowyes, but i don't get the need for cue files when you have gapless playbackj
15:34:53linuxstbYes, and my MP2 files are always CBR without any padding bytes or ID3 tags. So for my specific case, sample-accurate seeking is easy.
15:34:56rasherwell if you already have a large collection of cue'd files
15:35:13 Join lolo-laptop [0] (~lostlogic@68.251.84.226)
15:35:13rasheror if you use a player on your computer that uses cue files
15:35:16preglowbut yes, might work for mp2, of course
15:35:22preglowbut for mp3, you'd get inaccurate track boundaries
15:36:00preglowworks great for codecs with good seek support, though
15:36:21linuxstbI don't really care about sample-accurate track boundaries either - most of my files are live concert recordings, so most of the time I have a few seconds of cheering between tracks to insert the track mark.
15:36:26preglowhaving cue file support will require pretty flexible codec api
15:37:32linuxstbYes, I'm thinking that the codec stuff should probably have cuefile support from Day 1.
15:39:40linuxstbSo you would have three levels: playlist -> files -> songs. Currently rockbox thinks of tracks and files being the same thing.
15:40:45 Quit ashridah ("sleep")
15:41:20 Join methangas [0] (methangas@0x50a43276.virnxx10.adsl-dhcp.tele.dk)
15:41:21linuxstbWe could always invent a seektable format for MP3 files - maybe using id3v2 tags. Unless someone else has done it.
15:46:40preglowhaha
15:46:44preglowi don't care, i avoid mp3's
15:48:44preglowif you implement it as id3 tags, you're going to have problems, the seektable belongs with the decoder, but is in id3v2, so has to be read by the tag reader
15:49:30 Quit methangas (" HydraIRC -> http://www.hydrairc.com <- The future of IRC")
15:50:53preglowbut we might need some communication between the tag reader and the codec anyway, like with replaygain
15:51:02preglowthere's a lot to consider
15:54:53linuxstbThe tag reader will be different for each codec - so the tag reader will need to call codec functions to read the tags.
15:55:38linuxstbThere's two different types of metadata - total playing time, file samplerate, stereo/mono etc, plus the artist/track/title stuff.
15:59:08preglowwhy should it be different for each codec? apev2, among others, can be used for lots of codecs
16:00
16:00:30 Quit Nibbler ("blubber")
16:01:23rasherwell most codecs include their own metadata format
16:01:26sneakumsflac uses a vorbis comment block, i believe
16:01:30 Join Nibbler [0] (~sven@port-212-202-78-188.dynamic.qsc.de)
16:01:30 Join NibbIer [0] (~sven@port-212-202-78-188.dynamic.qsc.de)
16:01:36sneakumsin a specific frame, or whatever the term is
16:01:46sneakumsbeen a while since i looked at the flac docs
16:01:48 Quit NibbIer (Client Quit)
16:01:51rashera specific thingamajik
16:01:59rasheras the professionals say
16:02:04linuxstbsneakums: yes, FLAC has vorbis comments
16:02:06preglowbut metadata format should not be tied to codec, i think
16:02:38 Quit Nibbler (Client Quit)
16:02:42rasherhm
16:02:44rashermakes sense I guess
16:02:48sneakums"Note that the vorbis comment spec allows for on the order of 2 ^ 64 bytes of data where as the FLAC metadata block is limited to 2 ^ 24 bytes. Given the stated purpose of vorbis comments, i.e. human-readable textual information, this limit is unlikely to be restrictive."
16:02:53sneakumsHeh.
16:02:58rashera "vorbis comment" reader
16:03:08linuxstbno, but how that data is stored in the file could vary depending on the codec (well, specifically the container format used by the codec)
16:03:09rasheran "id3 reader", etc
16:03:14rasheryeah
16:03:22rasherso the codecs should be able to "use" the metadata readers
16:03:31rasherish
16:03:39preglowyes, some way
16:03:45rasheror at least, instruct the metadata readers on where/how to read
16:03:46preglowthey should have access to the info, at least
16:04:02preglowbut metadata will mostly be used by rockbox itself
16:04:02linuxstb"reading" should mean reading from the file (a codec function) and "parsing" could be done by a general tag library.
16:04:11preglowlike show artist, allow you to scroll through lyrics, read track list
16:05:06rasherpreglow: but as you said, things like gain/normalization information *can* live in tags, so the codecs should have access to that
16:06:41preglowrasher: yes, and i believe they're tag format agnostic, so rockbox should do that part, and somehow pass the info to the codec
16:06:47linuxstbThe gain info could go from codec -> tag -> rockbox -> audio playback system -> codec. i.e. the codec doesn't directly "know" it.
16:07:00preglowor the codec can have access to the entire tag struct and just use what it wants to
16:07:03rashermakes sense
16:07:28linuxstbThe user may want to override the gain specified in the tag - so playback gain could be a combination of various things, not just what's in the tag.
16:07:35sneakumsoh, i'd forgotten that flac can store cuesheet info in the file
16:08:19linuxstbsneakums; yes, but I don't think it has very good software support at the moment. But of course of Rockbox had general ".cue" support, it would be trivial to use FLAC-embedded cue files.
16:09:47linuxstbThere is also a standard for embedding cuefiles in id3v2 tags in MP3 files.
16:10:44preglowi don't see why people are creating these standards now, gapless playback being well available
16:12:05linuxstbPersonally, I like having one 2-hour MP2 file I can navigate using a cuefile, rather than about 40-50 seperate ones.
16:12:19rasheron the contrary
16:12:24rasherI like having the individual files
16:12:37rasherso I can pick a move around
16:12:41rasheruh
16:12:47rasherso I can pick a few to move around
16:12:47sneakumsi generally keep a 1:1 cd track to file mapping
16:12:48rashershare
16:12:49rasheretc
16:12:53sneakumsi have some files i'd like to add cues to
16:13:04sneakumsrather than split the files
16:13:18linuxstbMy sources normally aren't CDs - they are radio or TV recordings
16:13:42rasherI only have a few of those
16:13:47preglowand ogg supports having several files pasted together
16:14:09preglowthat way you get one big files containing several songs, and the track titles will change dynamically and all
16:14:50linuxstbSo we at least need the concept of cuefiles to support those kind of OGG files properly.
16:15:01preglowno
16:15:03preglowit's part of the ogg spec
16:15:13preglowand the concept is a bit different
16:15:16sneakumsi don't think there's a toc as such
16:15:27sneakumsjust you have multiple ogg streams in the same file
16:15:29preglowsneakums: correct
16:15:32linuxstbI mean that rockbox needs to stop thinking of 1 file == 1 track.
16:15:41lolo-laptopmultiple vorbis or flac streams in one ogg, yes?
16:15:47lolo-laptopnot multiple ogg headers?
16:15:55sneakumsno, multiple ogg headers
16:15:59preglowlolo-laptop: doesn't matter, just paste .ogg files together, and it'll work fine, regardless of content
16:16:02lolo-laptopoh
16:16:03sneakumsyou can cat a bucnh of oggs together, and ogg123 will play them
16:16:05lolo-laptopcrazy.
16:16:12sneakumsalso, does anybody even use ogg flac?
16:16:18preglowdunno, i sure don't
16:16:20sneakumsraw flac seems to be sufficient
16:16:39preglowand having several codecs with just one extension is stupid, the way things work today
16:16:51preglowthat scheme won't work until the day things are no longer file extension based
16:17:08sneakumssame problem with avi and mov, except the software jsut supports all the codecs
16:17:16sneakumsif you're lucky
16:17:17preglowyes, but there you have one player, mostly
16:17:21preglow'cause it's just movies
16:17:22 Join Nibbler [0] (~sven@port-212-202-78-188.dynamic.qsc.de)
16:17:25preglowwith ogg, you can have audio AND movie
16:17:27sneakumsvlc on linux doesn't support qdm2
16:17:31preglowand i have seperate players for that
16:17:46linuxstbThe extension normally refers to the container format, not the codec - most container formats (e.g. AVI, OGG) can contain various codecs.
16:17:58sneakumsoperating systems jsut need to get smarter about that stuff
16:18:06preglowyes
16:18:07sneakumsand they won't without pressure from stuff that exists
16:18:09linuxstbMP3 doesn't have a container format - it's just the raw frames.
16:18:15linuxstbThat's the problem.
16:18:16sneakumsjust out bad luck to be in the mid-stage
16:18:21preglowlinuxstb: well, there are some that are packaged in riff container
16:18:23sneakumss/out/our/
16:18:36linuxstbpreglow: and .rm files.
16:18:37preglowlinuxstb: lame even supports writing mp3s that way
16:18:59linuxstb... and .avi can have MP3 audio in it.
16:20:04preglowsure, loads of movies have mp3 audio
16:23:07linuxstbI don't think we should worry about all the various different container formats yet - just support the de-facto standard for each audio format, whiich is either raw frames (such as mp3), or the most common container format for the other formats.
16:25:16preglowworry, no, but we have to take it into consideration
16:25:31rasherFew people understand that they can play some avis and not others
16:25:37rasherIt's really difficult to explain
16:25:42rasher"But I just played that avi!"
16:25:44preglownot really
16:25:53preglowpeople are quite familiar with the concept of containers
16:25:54preglowheh
16:31:40 Quit bobTHC (Read error: 60 (Operation timed out))
16:32:07 Join amiconn [0] (~jens@pD95D1FF3.dip.t-dialin.net)
16:33:49 Join bobTHC [0] (~foo@l05v-4-93.d1.club-internet.fr)
16:48:26 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
16:50:02 Quit bobTHC ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
16:50:46linuxstbGoing back to reasons for cuefiles, I think they have been adopted by FLAC to enable users to make perfect backups of CDs - i.e. you need a cuefile created from the original CD in order to burn a perfect clone from the extracted PCM data.
16:51:38***Saving seen data "./dancer.seen"
16:51:40linuxstbFor example, EAC supports ripping a CD to one long WAV with cuefile.
16:54:35Zagorthere is no opposition for cue support
16:55:14linuxstbGood :-). My proposal is to make it a core part of the new audio system.
16:55:36linuxstbI don't think it would work well as an after-thought.
16:56:25Lynx_linuxstb: whole rockbox is an afterthought ;)
16:56:25preglowdefinitely not
16:57:20Zagorit's good to keep cue in mind, but still single-song files are the norm by far. we need to handle both types without kludges.
16:59:02linuxstbI'm not sure it will be a kludge to consider every file as potentially containing 1 or more songs, but I've got an open mind about it.
16:59:23lImbushi amiconn
17:00
17:07:23 Part Zagor
17:11:54 Join Aison [0] (~hans@zux166-181.adsl.green.ch)
17:12:00Aisonhello :))
17:12:08lImbushi there
17:12:34lImbusCassandra_: I just noticed the Ondios are missing in the list of supported devices in rockbox-manuel-2.4.pdf on page 3
17:13:08Aisonjust discovered your home page :) and im impressed of the project
17:13:24lImbushehe, we all are :-)
17:16:40 Join Christi-S [0] (~christi@213.78.162.11)
17:16:54 Quit Cassandra_ (Read error: 110 (Connection timed out))
17:20:21 Join mecraw [0] (~mecraw@69.2.235.2)
17:32:15lImbusChristi-S: did you see my comment about the bug I spotted out in the manual ?
17:39:12HClwhats the difference between fopen/fread/fwrite open/read/write etc etc?
17:39:13 Quit ripnetuk (Read error: 104 (Connection reset by peer))
17:39:50 Quit jyp ("poof!")
17:40:14sneakumsHCl: the f* ones are part of the c library, and do buffering
17:40:27sneakumsthey're a bit easier to use also
17:41:19HClno seek function?
17:41:40linuxstblseek works with file descriptors (as used by open/read/write etc)
17:41:40sneakumsthere's an fseek
17:41:51HClah
17:41:52HCllseek
17:41:53HClthanks
17:42:09sneakumsyou can't mix the f* functions and the other ones
17:42:14linuxstbHCl: I had the same discoveries yesterday trying to adapt FLAC to work in the Rockbox environment.
17:42:18sneakumsif you're using fopen and so on you have to use fseek
17:42:20HCllinuxstb: *nods*
17:42:30HCllinuxstb: altering gnuboy to use lseek etc.
17:42:30sneakumsotherwise the buffering will get all confused
17:42:36linuxstbBasically, you can't use any of the f* functions
17:42:42HClyea.
17:43:02linuxstb... and replace FILE* by int
17:43:41HClyup.
17:43:46HCl :P
17:44:47lolo-laptopwhat c library does rockbox use, I assume it is statically linked to the final firmwar?
17:44:58Bagderit uses its own
17:45:23HClits own
17:45:24lolo-laptopis it based on one of the preexisting small C libraries or written from scratch?
17:45:42Bagderit is loosely based on newlib
17:45:53linuxstbSomeone earlier today mentioned the "objdump -T" command - if you run it on the gnuboy executable, it will tell you all the external library functions it uses.
17:46:04HClthanks
17:46:07lolo-laptopcoo'
17:46:18HClgnuboy is written fairly portable
17:46:28HClwith various outputs to sdl, x11, svga, windows, etc.
17:46:48HClso i'm hoping it'll just compile
17:47:24HCli'm mostly fixing warnings etc..
17:48:22Lynx_HCl: does that mean we'll get gameboy on iriver before sound? :))
17:48:42preglowhahaha
17:48:48BagderHCl: I assume it uses malloc() ?
17:49:35HClBagder: so far, i haven't seen any warnings about unknown functions named malloc, but i'll look at it in a bit. and most likely, yes.
17:49:38HClLynx_: nah :P
17:49:47HClunless the porting goes just, scarily smooth
17:50:32Lynx_HCl: ok, i'll check again tomorrow ;)
17:50:55linuxstbBagder: Does rockbox have a qsort implementation?
17:51:23Bagderyes
17:51:55linuxstbCool. In which case, the only missing functions needed for libFLAC are malloc/realloc/calloc and free.
17:52:12BagderI'll bring in the malloc soonish
17:52:24HCllinuxstb: know a replacement for feof?
17:52:56linuxstbHCl: No, I just keep track of the current file position manually and compare it with the size of the file. Anyone else know?
17:53:20HClokay..
17:53:22Bagderno, I don't think we have an equivalent
17:53:30HCli think i'll leave out the gnuboy.rc support then..
17:53:31Bagder(yet ;-)
17:53:31HClfor now.
17:53:42 Quit Nibbler (Remote closed the connection)
17:53:47HClit should fall back on its default settings without one anyways
17:59:18HCldoes rockbox have a time() function?
18:00
18:00:14 Quit elinenbe (Read error: 104 (Connection reset by peer))
18:00:56Bagderthere's a get_time()
18:01:21 Join oNe`Neo [0] (~Info@p54879778.dip.t-dialin.net)
18:01:34oNe`Neosomeone knows Schoki2_
18:02:15 Quit Patr3ck_ ("User pushed the X - because it's Xtra, baby")
18:02:36HClok
18:02:52linuxstbI didn't think the iRiver had a hardware clock
18:03:02oNe`Neosomeone knows Schoki2_
18:03:12Bagderlinuxstb: it doesn't
18:03:14thegeekonce is enough oNe`Neo
18:03:21linuxstbSo what does time() return?
18:03:24oNe`Neokk
18:03:32thegeekif there are no responses
18:03:33BagderI don't know ;-)
18:03:35lImbusoNe`Neo: nicht schon wieder
18:03:36thegeek...
18:03:37thegeek;)
18:03:44HClhrm.
18:03:45Bagderprobably the time since start
18:03:56oNe`Neodoch schon wieder
18:04:00oNe`Neoweisste auch warum?
18:04:00linuxstbThat's going to be a pain for the filesystem timestamps
18:04:09Bagderyes indeed
18:04:13HClnow i need a function to convert that struct tm to unixtime..
18:04:23oNe`Neo10.25.24 Join Schoki2_ [0] (Schoki@DSL01.212.114.237.68.NEFkom.net
18:04:41oNe`Neound der typ mit der ip adresse 212.114.237.68 hat meinen cs server gehackt
18:04:59Bagderwe're english speaking here
18:05:03thegeekthat looks like dynamic ip
18:05:07HCloNe`Neo: we don't know anyone named schoki2, this is an english channel
18:05:14thegeek*a
18:05:20oNe`Neobut he was yesterday in this chanel
18:05:20HCllogbot seen schoki2_
18:05:27HClno access? :x
18:05:32HCllogbot seen schoki
18:05:33Bagderonly privately
18:09:15HClblah.
18:10:23HCli'm getting undefined references to open / read / write etc, with what library should i link
18:10:27HCl?
18:11:30linuxstbIf you're compiling a plugin, I think you need to use rb->open etc
18:12:20HClhmyea..
18:12:21HClok.
18:12:27HClthey'll be part of my interface then...
18:12:37 Part oNe`Neo
18:12:53HCli'm just gonna write an interface .c to provide the rockbox specific stuff..
18:13:15HClhow many milliseconds is a kernel tick?
18:13:25HClwhen talking about sleep() etc
18:14:23linuxstbAnyone have any suggestions for how I can hack libFLAC into Rockbox in order to test decoding speed? Obviously, I need Bagder to implement malloc, and LinusN to increase the clock speed to something sensible first.
18:15:56linuxstbI want to decode a .flac to a .wav on the iRiver to test both speed and accuracy.
18:17:05linuxstb... and then do the same with libmad and Tremor.
18:18:37HClmaybe you can build it into a plugin at first that reads a certain file?
18:19:01linuxstbI was thinking about that, but then I would have all the problems you are having :-)
18:19:24HCllol.
18:19:29HCli'm not having that many problems
18:19:35HCli just made a list of stuff i need to implement
18:20:07HClthe only things that look odd are functions sbrk isatty and fstat, prolly should search where they are used and get them out
18:20:21HClaside from that, its mostly gnuboy interface, really
18:20:23linuxstbThe difference is that the final destination of "rockboy" is a plugin. My code will end up as either part of the core rockbox code, or as a special codec plugin
18:20:34HCltrue
18:20:46HClbut you said you wanted to hack it into rockbox in order to test :P
18:21:25linuxstb"hack" to me means as little work as possible.
18:22:17HCli guess :)
18:22:55linuxstbYou can replace fstat with lstat, but if you just need the filesize (one of the fields returned by lstat), then you can use the Rockbox filesize() function.
18:23:56linuxstbisatty is something to do with terminal I/O, and I haven't a clue about sbrk, even after reading the man page...
18:24:35linuxstb(sorry ignore what I said about lstat - fstat is the right command for file descriptors)
18:26:26 Quit lImbus (Excess Flood)
18:26:41 Join lImbus [0] (~manuel@kernel.cycos.net)
18:29:10 Quit elinenbe_ (" Want to be different? HydraIRC -> http://www.hydrairc.com <-")
18:31:39HClhm...
18:31:53HClwell, i couldn't find any of those 3 functions being used, so i dunno.
18:31:56HCli'll look at it.
18:32:11HClat the moment.. i kind of need to be able to add more .o's while compiling a plugin..
18:35:33linuxstbHCl: Yes, that's why I don't want to attempt to compile my flac code as a plugin.
18:36:42HClmhm
18:36:57HClwell, i'll wait till someone can tell me what the exact commands are that i should execute to get a .c into a plugin file
18:38:23 Join oNe`Neo [0] (~Info@p54879778.dip.t-dialin.net)
18:38:47preglowisn't that pretty well described in the makefile?
18:41:12HClhm.
18:41:15HClnot sure.
18:41:26HCli'm not great at deciphering makefiles
18:42:49HClk, i think i somewhat understand it
18:44:48 Join Trevmar [0] (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
18:45:26TrevmarWhere can I find a list of the Changelog data so that I can determine the latest Ondio version date before the Hotswap was added?
18:51:05lImbusuh. i think the only helping tool is cvs
18:51:37Trevmaryes, I looked through CVS, couldn't find any file of committed changes. Is there one?
18:51:42***Saving seen data "./dancer.seen"
18:52:02lImbuswell, the cvs-comments for each file
18:53:43TrevmarAh yes, thanks. I guess I will have to update my source-tree.
18:53:49lImbusTrevmar: a grep through the irc-logfiles shows that there have been experiments ongoing with hotswap starting 10th of oct
18:55:17Trevmarah - another good idea - thanks. I know that dec 21 seems stable, and that Jan 29 is broken. I would like the audio changes which went in (such as stereo width) before the hotswap code which is breaking my Ondios.
18:55:55TrevmarI guess I will just have to try flashing a few versions in a binary search pattern, or whatever...
18:56:07lImbushehe
18:56:25lImbusget the first with the audio tricks ?
18:56:33lImbusfirst==oldest
18:56:57Trevmarat least it is easy for me to fall back on uart_boot if I get into trouble, all the units are already modded for uart_booting -LOL
18:57:06lImbuslol
18:57:39Trevmaryes, the Dec 21 version doesn't have stereo width (which is real neat), but it can see my MMC cards
19:00
19:08:09amiconnTrevmar: I think I added stereo width *after* hotswap was added, but the proper way would be to find out why hotswap doesn't work for you anayway.
19:08:15amiconn*anyway
19:09:50 Join Eodun [0] (~Eodun@80.28.67.209)
19:11:00amiconnTrevmar: You stated that 20041221 works for you, but 20041229 does not. Could you perhaps investigate further which daily builds is the first broken one? That's only 8 versions to test...
19:12:38 Join edx [0] (edx@p54879303.dip.t-dialin.net)
19:14:35 Part lImbus
19:15:19amiconnTrevmar: The weird thing is that even your 0x0708 masked Ondios don't like hotswap; apart from the different tuner, they should work indentical hardware-wise to the 0x0F08 masked Ondio FM
19:15:37amiconn...and the 0x0708 masked Ondio SP I have.
19:16:26HCldoes anybody know how to get gcc to build 1 .o from several .o's?
19:17:03lolo-laptopprobably just gcc -o 1.o blah.o blah.o ... I'm not sure though
19:18:08HClno, that tries to link them with _start..
19:18:12HCli think.
19:18:24HClor does the extension of the output file matter?
19:18:54HClah. i'm stupid.
19:18:54HClnm
19:20:48 Nick Aison is now known as Aison|unfaehig (~hans@zux166-181.adsl.green.ch)
19:22:15HClhm.
19:22:16HClsorta.
19:23:21 Quit preglow ("off")
19:23:52 Quit Eodun ()
19:25:17thegeekhmm, er det noen begrensninger på ntnu sitt datanettverk?
19:25:18thegeekaltså
19:25:22thegeekhvis man har det på hybel'n
19:25:59thegeekfår man egen ekstern ip
19:26:07thegeekeller er det et internt nettverk ?
19:26:38linuxstbHCl: What about "ld -r"
19:30:17HCllinuxstb: yea, i'm having that...
19:30:20HClthegeek: english.
19:30:22thegeekhmm
19:30:24thegeekoops
19:30:25thegeekhehe
19:30:29thegeekwrong chan;)
19:30:29thegeekhaha
19:30:34thegeekdamn that took long ;)
19:30:34HClhehe..
19:30:49thegeekI was like sitting here wondering why the hell there were no responses;)
19:43:42 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
19:46:31 Join elinenbe [0] (~elinenbe_@65.115.46.225)
19:46:41HClodd
19:46:45HClwe have strrchr, but not strchr?
19:47:02elinenbewhere can one buy the h140 now? I don't see it being sold anywhere!
19:47:13HClBagder: think you can add strchr while you're adding malloc?
19:48:23 Join MooMaunder [0] (~paul@cpc1-stkp4-5-0-cust117.mant.cable.ntl.com)
19:49:08amiconnHCl: strchr() is there
19:49:46amiconnfirmware/common/strchr.c
19:50:04amiconnIn fact, strrchr() uses strchr()
19:50:59linuxstbamazon
19:51:13linuxstbamazon.co.uk I mean
19:52:15HClamiconn: its not in the plugin api then, needs to be updated
19:52:52amiconnAh, yes. Obviously it wasn't needed by any of the exisiting plugins
19:52:58amiconn*existing
19:54:05HClok
19:54:14HClcan someone with cvs access update it..? please?
19:55:33 Join iriver [0] (~Miranda@84.238.16.138)
19:55:43amiconnYou can update it locally. If your plugin gets accepted & committed, it'll be added officially as well
19:58:21 Join Naked [0] (naked@naked.iki.fi)
19:58:26 Quit Naked (Client Quit)
20:00
20:04:59 Join walnut [0] (~Administr@dialup0319.ts551.cwt.esat.net)
20:15:00 Join Aison [0] (~hans@80.254.166.181)
20:20:06 Quit Aison (Read error: 104 (Connection reset by peer))
20:20:49 Join Aison [0] (~hans@zux166-181.adsl.green.ch)
20:31:20 Quit Aison|unfaehig (Connection refused)
20:33:18 Join Marder [0] (~MArder@lsn-boi-catv-c124-p218.vtx.ch)
20:43:24 Join XShocK [0] (~XShocK@pcp09492659pcs.nrockv01.md.comcast.net)
20:44:18 Part walnut
20:48:34MarderXShock : I finally managed to build a working toolchain under Cygwin
20:49:43 Join LinusN [0] (~linus@labb.contactor.se)
20:50:03amiconnhi LinusN
20:50:42LinusNhi
20:51:45***Saving seen data "./dancer.seen"
20:52:23 Join Cham [0] (~52e2b617@labb.contactor.se)
20:52:42 Quit Greenwell ()
20:54:43Chamhi
20:54:50Chamwhats new on the boot loader
20:54:50Cham?
20:55:05Chamhold button is detected ?
20:55:29LinusNhave no time to work on it tiday
20:55:31LinusNtoday
20:55:39Chamoh
20:56:07XShocKhi
20:56:10LinusNin fact, i have to go now
20:56:22XShocKLinus: Happy Birthday :)
20:56:30Chamto go to work on bootloader ?
20:56:34LinusNHCl: just put the .c file in apps/plugins and add it to the apps/SOURCES file
20:56:38Chamsoory happy birhtday in first
20:56:48LinusNCham: no, to celebrate my birthday
20:56:58LinusNXShocK: thx
20:57:02LinusNCham: thx
20:57:11LinusNwell, cu later doods
20:57:18Chambye
20:57:25 Part LinusN
20:58:14 Quit Cham ("CGI:IRC")
20:58:42XShocKMarder: I compiled it in Ubuntu Linux under Cooperative Linux emulation
20:58:53XShocKbut thanks for sharing. :)
21:00
21:00:45MarderOk. The trick was to eliminate the 'm68040' from the list of targets, as it made the linker barf.
21:01:29XShocKhehe.. i wanted to do it, but did do it by some reason. :)
21:02:13MarderOnce 68040 removed, it compiled Ok and I could compile a working version of RockBox
21:03:11XShocK:)
21:03:23MarderI do not trust my compiled bootloader though, as it has a different signature from everbody else :-(
21:05:50linuxstbMarder: you could try checking out older CVS versions of rockbox and see if you can match any bootloader.bin MD5s
21:06:27linuxstbi.e. from yesterday or the day before.
21:06:30iriverJust install LinusN compiled bootloader attached in the wiki.
21:07:48MarderFor the records, the file to edit is 'gcc/config/m68k/t-m68kelf' , remove '/m68040' from the 'MULTILIB_OPTIONS =' line
21:08:33iriverI installed LinusN wiki version and merged it with the EU iriver firmware.
21:08:33iriverit is working fint.
21:08:33iriverIt should be based on the bootloader 1.5 cvs revision.
21:10:53MarderHow do you think I tested my compiled RockBox version ? ;-)
21:12:49iriverI did not know that you tested it on the target. Why do want to compile your own bootloader?
21:16:04MarderJust for the sake of it )
21:50:17 Join courtc_ [0] (~court@adsl-158-11-58.asm.bellsouth.net)
21:50:21 Quit courtc (Read error: 110 (Connection timed out))
21:54:50 Nick courtc_ is now known as courtc (~court@adsl-158-11-58.asm.bellsouth.net)
21:56:21 Part oNe`Neo
22:00
22:03:32 Quit thegeek ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
22:08:05 Quit Trevmar (Read error: 110 (Connection timed out))
22:10:47 Join thegeek [0] (na@ti521110a080-3006.bb.online.no)
22:12:47 Quit Marder (Read error: 110 (Connection timed out))
22:14:04 Join Trevmar [0] (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
22:14:17Trevmaramiconn are you there: I was at the dentist - am back now
22:14:24amiconnI am.
22:14:37TrevmarOK, it was the 20050129 that didn't work
22:15:01amiconnAh, so one of your forum posts is wrong about this.
22:15:30amiconnWould you mind checking which exact day it broke? You could do a binary search...
22:15:49Trevmaroh dear..I will fix that. I am looking at a list of my downloaded daily snapshots now, so I know this is correct
22:16:23amiconnI'd also like to know whether USB access works (both internal flash & MMC) for some/all of your boxes with current cvs.
22:16:29Trevmaryes, I will try and figure that out for you. I really like the audio enhancements recently :)
22:17:12amiconnAnd I'd like to track down such nasty bugs (given that it is one, but it likely is imho)
22:17:55TrevmarI have 5 working sets of hardware today. Only one has a USB which is not working right, and I am pretty sure that the problem is just that whatever is in the memory contenst is stopping the PC from logging in. The USB chip is recognized and sets up the transfer
22:18:14Trevmarcontents
22:18:48Trevmarthat 5th one works OK if I plug in an initiilzed MMC
22:19:33TrevmarI will do a binary search this afternoon to try and track the pronblem down to the nearest day or two. I will track where the MMC card stops being recognized. OK?
22:19:52Trevmarstops being recognized in the 03 08 mask versions
22:19:54 Quit iriver (Read error: 110 (Connection timed out))
22:19:59*HCl yawns
22:20:21amiconnTrevmar: Good :)
22:20:58amiconnDid I get you right that internal & external USB access works for all boxes _with current cvs_ ?
22:21:30TrevmarNo, current CVS panics and I can't get to the stage of enabling USB
22:22:50linuxstbHCl: Have you made any progress with compiling a .rock from multiple .c files?
22:22:58amiconnHmm. I added that out-of-range panic because some cards are formatted in a way that the partition is larger that the physical card capacity, however, this shouldn't hit right at the start...
22:23:28Bagderlinuxstb: add files to the pluginlib to make that happen
22:23:49Bagderall plugins link with that
22:26:05HCllinuxstb: yes, actually.
22:26:05 Quit gromit`` (Read error: 104 (Connection reset by peer))
22:26:38 Join gromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
22:26:41HCllinuxstb: i got gnuboy compiling from both a gnuboy sourcetree and the plugin .c by altering the plugin makefile
22:27:41HClat the moment i'm kind of waiting for Bagder 's malloc implementation
22:27:46HCland for someone to add strchr to the plugin api
22:29:30linuxstbI thought that the malloc was only going to be for the use of the codecs
22:30:54Bagderthat's the plan, yes
22:31:07linuxstbWhich leaves HCl with a problem...
22:31:11Bagderyes
22:31:15*amiconn wonders why we would need malloc() all of a sudden
22:31:23Bagderamiconn: for the codecs
22:31:30amiconnRockbox works well without having malloc()
22:31:48linuxstbBecause the codecs depend on it, and we don't want to hack the standard libFLAC, libmad, Tremor etc code
22:31:54amiconnBagder: Yeah, I read that, however I don't get why the codecs will need it
22:32:23 Join Patr3ck [0] (~patr3ck@pD95483F6.dip.t-dialin.net)
22:32:27Bagderthey're not written for embedded use I guess
22:32:36Bagdernevertheless, they use malloc now
22:32:45Trevmaramiconn- the card which read past the end works fine in the older software version
22:33:14linuxstbBut at least for FLAC, I don't think it would be too hard to try and patch them to avoid malloc. The developer may even accept such patches into the mainstream code.
22:34:09amiconnTrevmar: Well, it appears to work, but if you fill it up 100% (no free space left) I am sure some files can't be read properly
22:35:24HClokay
22:36:29amiconnHCl: I expect you'll run into another problem with gameboy emu: The plugins have a fixed maximum size, which did not yet get adjusted for the bigger iRiver RAM.
22:36:34linuxstbHCl: How much is malloc (and realloc/calloc/free) used in gnuboy?
22:36:46amiconnI don't think you're able to fit a gameboy emu in 32 KB...
22:36:56Bagderwe could extend that for iriver
22:37:06amiconnBagder: Yes of course.
22:37:18amiconnOnly that it's not yet done
22:37:39amiconnIf we increase it proportionally, that would be 512 KB
22:38:22linuxstbWhat's the opinion on malloc - should I investigate if I can FLAC (and libmad and Tremor) can live without it?
22:38:34HCllinuxstb: i'm looking at that now
22:38:56BagderI am in favour for avoiding malloc as far as possible
22:39:11BagderI mean, without going to extremes
22:39:34HClwell
22:39:42HCloddly enough, they use it a lot for strings
22:39:54HCland for the gameboy ram and the rom itself
22:39:57amiconnBagder: Iiuc implementing malloc() and friends would mean setting some memory aside to allocate from?
22:40:04Bagderyes
22:40:06Trevmaramiconn: there is 18 meg free on the 979 meg MMC card
22:40:09HCli'm not sure how to allocate memory without it...
22:40:12HClbut
22:40:22HCli guess i can allocate the max ever needed
22:40:22Bagderwe'd add a special fixed-size pool for malloc
22:40:29HClsince it uses realloc a lot in order to not waste memory
22:41:02amiconnTrevmar: 18 meg is much... compared to the amount the partition might be too large
22:41:28amiconnTrevmar: On my 256 MB card it were just 128 sectors (64 KB)
22:41:30TrevmarLet me track down the initial error and let's see if the rest cascade from there :)
22:43:17amiconnBagder: Imho dynamic memory allocation doesn't give much benefit on an embedded system with little RAM & no MMU, but adds problems instead
22:43:26amiconn(memory fragmentation)
22:43:33Bagderindeed
22:43:37amiconnI know this from the Amiga...
22:43:44Bagderthat's why we don't do it
22:44:02Bagderit would only be added for portability with codecs
22:44:02HClhow am i supposed to allocate memory?
22:44:34Bagdernothing in current Rockbox allocates memory
22:44:40HClyea, i know.. um.
22:44:44HClso what regions can i use?
22:44:45amiconnHCl: The primary rule is: allocate memory static, i.e. use arrays
22:44:49HClah.
22:44:52HClokay.
22:44:59HCli can do that
22:45:01amiconnThen, there are 2 more possibilities for plugins
22:45:30amiconn(1) You can get the remainder of the (fixed size) plugin ram, since only one plugin can run at a time.
22:46:13amiconn(using plugin_get_buffer)
22:46:45amiconn(2) You can grab the main buffer, i.e. the buffer usually holding audio data for playback
22:46:57 Join lImbus [0] (lImbus@237-44.244.81.adsl.skynet.be)
22:47:05amiconnOf course this means your plugin can't run in parallel to playing audio
22:47:07lImbushowdy
22:47:11amiconnhi lImbus
22:47:25lImbuswow. 39 people here in
22:47:30amiconnHCl: For this, you have to use plugin_get_mp3_buffer()
22:47:31lImbushi jens
22:48:40amiconnlImbus: some questions regarding your francais.lang:
22:48:44lImbusga
22:49:14amiconn(1) I found some settings strings that don't start with a capital like the others. Is it okay to change that?
22:49:28lImbusmhmm. I thought about a long time as well
22:49:39lImbusI tried it on the device, and I mainly adapted myself
22:49:54amiconnMost of the strings do, only 5 strings start with a small letter
22:50:12lImbuswhen I added strings, I added them the in the existing style
22:50:47lImbusthe problem is, that mainly french does not know capital letters, this is why I would rather opt for having small letters overall
22:51:37amiconnOIt's only about starting the string with a capital letter...
22:51:48***Saving seen data "./dancer.seen"
22:52:04lImbusmhmm. I suppose it's the id3-db-stuff isn't it ?
22:52:12HClamiconn: how big is the mp3 buffer?
22:53:20amiconnlImbus: (2) There is a mismatch in general settings->system->manage settings. You use "configurations" within "browse .cfg files", but "fichier .cfg" in "write .cfg file"
22:54:24amiconnHCl: mp3 buffer is the remainder of all ram after setting aside the parts needed by the firmware. It's the largest part, I'd expect >25 MB on iRiver
22:55:13lImbusamiconn: I'll check qout first question on the device. It's still in my car, I'd have to get it, will do it later this evening.
22:55:48amiconnlImbus: "au choix" (channel settings), "plus récent", "par type" (file sorting), "banque ID3" (file view), "linéaire(5)" (peak meter settings)
22:55:59lImbusok.
22:56:18amiconnerm, "linéaire(%)"
22:56:29lImbus^^
22:56:33lImbusmhmm
22:56:36lImbushow to say
22:56:45lImbusthose are possible settings for options
22:56:55 Join [IDC]Dragon [0] (~Joerg@pD95123FD.dip.t-dialin.net)
22:57:24amiconnlImbus: (3) The (text) string for indicating free diskspace is way too long. Would "Livre:" be sufficient?
22:57:25lImbuswhile I think starting with capitals is nice for menues, the possible settings should start with small letters
22:57:54lImbuschecking (3)
22:57:57amiconnlImbus: If you prefer it that way, it should be consistent though
22:58:04amiconnhi Jörg
22:58:11lImbushi joerg :-)
22:59:25lImbusabout case sensitivity (is this still (1)?). I don't know if I am in the position of prefering anything. I'm most probably not going to use it. Well, I'll test it for a while. But I don't want to start a poll on the list as well :-)
22:59:43[IDC]DragonHi all
23:00
23:00:02[IDC]DragonHi Trevor
23:00:51lImbusamiconn (3): is that id LANG_DISK_FREE_INFO ?
23:01:01amiconnyup
23:01:08 Join coob [0] (pen0r@82-44-227-205.cable.ubr11.haye.blueyonder.co.uk)
23:02:07coobhiho, anyone who does rockbox.org website backend stuff around? I'm interested in your 'recent cvs activity' bit of the main page.
23:02:31 Join cloudscapes [0] (~d8623c93@labb.contactor.se)
23:02:51BagderI wrote that
23:02:54lImbusamiconn (3): yes. "Libre:" would then mean "free"
23:02:54 Quit cloudscapes (Client Quit)
23:04:13amiconnlImbus: Okay. I really meant "Libre:", only that I couldn't correctly decipher my own handwriting.
23:04:21lImbushehehe
23:05:19lImbusto solve (2), would then be to adapt LANG_CUSTOM_CFG to "Liste des fichiers .cfg" and "Liste des fichier de configuration" for voice
23:06:07amiconnOkay, I'll change that as well, then commit if you don't want to decide on (1) right now.
23:06:50*[IDC]Dragon is puzzled about some people's Ondio problems
23:07:16 Join jyp [0] (~jp@64.198-200-80.adsl.skynet.be)
23:07:18 Quit elinenbe (" The IRC Client of the Gods! -> http://www.hydrairc.com <- HydraIRC")
23:07:23amiconn[IDC]Dragon: If all else fails, maybe we should try to get hold of a 0x0308 unit...
23:07:49[IDC]DragonI was thinking so, too
23:07:54courtchahaha once again #ipodlinux is here to steal your ideas.. ;)
23:07:58 Quit Christi-S (Read error: 60 (Operation timed out))
23:08:07[IDC]Dragonis it only on the 0308?
23:08:09amiconnlImbus: Isn't there an 's' missing for the (2) voice string?
23:08:26[IDC]Dragoncourtc: I don't have any
23:08:45lImbusamiconn: of course, very correct.
23:08:50Bagdercourtc: we're waiting for you to port rockbox to ipod
23:08:55lImbuss/fichier/fichiers
23:08:55Bagder:-)
23:09:00[IDC]Dragoncourtc: how's Helix?
23:09:03courtcheh..
23:09:17Bagdercourtc: rockbox is just now starting to work on two new players
23:09:41courtcits not.. :/ the liscencing is too finiky..
23:09:51courtcnice :)
23:10:01[IDC]Dragonwhat do they "ask"?
23:10:04amiconnlImbus: Probably the same change of voice: applies to the config save "Ecrire un fichier de configuration"
23:10:39lImbusow yeah
23:10:47lImbusi hate those abbreviations
23:11:10lImbusdeciding on (1) would mean another night of changes either way of decision.
23:11:23lImbusso I'll opt for a commit and brace for complaints
23:12:13jypFrench translation?
23:12:39jypI'm a native french speaker; in case you want a reviewer
23:14:13lImbus:-)
23:16:12[IDC]Dragonnot wallonian?
23:16:37lImbusit's virtually the same
23:16:49jypWallonian is only spoken by old people, and a couple specialists
23:17:35 Join Christi-S [0] (~christi@213.78.111.216)
23:17:36lImbuswell, the wallonian dialect certainly
23:17:39jypWallonian (actually walloon iirc) != Belgian French ;)
23:17:43lImbusI meant the belgian french
23:23:40TrevmarJoerg, if we can't identify the problem I can send you onw of my 03 08 units. But let me try to pin down the dat at which it broke, first
23:23:47Trevmardate at
23:24:52Trevmaras you can see, my typing lets you know that I don't code every day any more - haven't for about 25 years :(
23:26:00linuxstbBagder: I've looked at libFLAC a little, and although most of the memory usage is constant, it does depend on the blocksize used to encode the FLAC file - 4 buffers of approximately 4*blocksize are allocated. In my files, blocksize is 4608 bytes (the default I think), but the standards allow for blocksizes up to 32K.
23:27:17linuxstbSo 4608*16 is about 73K, but 32K*16 is 524K - a big difference in "typical" and maximum usage.
23:28:12 Join jpburton5150 [0] (knoppix@cpe-24-94-54-216.stny.res.rr.com)
23:28:24Bagder524K is a lot
23:29:25linuxstbI'm just checking the blocksize values for my collection of FLAC files....
23:29:28amiconnlImbus: A diff -u on francais.lang looks funny (because of your sorting). First all old lines are "-" taken out, then all new lines are "+" added.
23:29:30[IDC]DragonTrevmar, did you read my reply?
23:30:53rasherlinuxstb: maybe a maximum block size could be defined at compile time, with a reasonably large default?
23:31:26linuxstbEvery single FLAC file I have has a blocksize of 4608 bytes, so we could just limit to that blocksize.
23:31:36lImbusamiconn: yes, I'm biting myself in the ass not having made a backup before the last step. would be easier to control
23:31:42linuxstbThese are from various sources, not just ones I've encoded myself.
23:31:51TrevmarJoerg, no not yet, let me read it
23:32:04Bagderlinuxstb: sounds like we could do start for a start at least
23:32:53linuxstbYes, I'll just make 4608 a #define in the code, and we can increase it if anyone complains.
23:33:15rasherlinuxstb: sounds painless
23:33:26rasherunless someone comes up with 32k blocksize
23:35:31TrevmarJoerg - sorry about overwhelming you :) Only the 03 08 units fail to see the cards in the 29 Jan 2005 CVS snapshot, although my 07 08 units switch off with lines across (and up/down) their displays when I hotplug or hot-unplug any MMC or SD card. But the 07 08 units do display the <MMC> directory.
23:36:28linuxstbOK, so for FLAC, my plan is to define a static array of a known size, and then implement my own trivial malloc/calloc functions that allocate memory from there. I don't need to use free, because that's only used when the decoder instance is destroyed - in which case, I just reset everything.
23:36:32[IDC]Dragonok, so we have a 74HC32 problem
23:36:33TrevmarOnly the 03 08 cards are PANICing now with the Feb 2+ releases
23:37:40TrevmarLet me identify the exact date at which the ability to see the MMC cards vanishes and that may help you track it down, else I will send you one of the 03 08 Ondios for testing
23:38:19[IDC]Dragonbetter not, I have't made too good experience with customs
23:38:34TrevmarI can try an HC08 if necessary (would have to drive about 20 miles to the electronics store)
23:39:07[IDC]Dragonno good neither
23:39:17Trevmaryou know, maybe I have a SM HC)* in my toolbox - let me look
23:39:20amiconn[IDC]Dragon: Usually there are no problems with customs if the sender declares the goods as gift
23:39:20jpburton5150Question: is anyone working successfully in cygwin compiling editing the rockbox source?
23:39:21TrevmarHC08
23:39:45BagderI did a very useful iriver fix now
23:39:46[IDC]Dragonamiconn, but since ebay they're more aware
23:39:52Bagderbounce uses full screen! ;-)
23:40:49rasherBagder: hurray! one of the modes were using full width
23:41:05linuxstbCan you fix sokoban next?
23:41:18BagderI started with rockblox, but the amount of fixed sizes made be decide not to ;-)
23:41:35linuxstbI think sokoban is the same
23:41:48rashersnow!
23:41:49rashercube!
23:41:56rasherthose shouldn't be as troublesome
23:42:00TrevmarNope, only an LS138 and S74 and LS240 are the only logic gates in SM in my toolkit. But I am sure I can buy one 'locally' if necessary. But would it work, is the imponderable...
23:42:02amiconnI think since the sokoban levels are standard, that'd need bigger tiles
23:42:57rasherand rockblox could do with being top-down and bigger tiles
23:43:05[IDC]Dragonamiconn, do you have any idea about the '32?
23:43:17Bagderrasher: send in your patches ;-)
23:44:08amiconn[IDC]Dragon: No, not yet, but that isn't what puzzles me most. The 0x0708 units switching off is highly suspicious...
23:44:08rasherBagder: I know.. I have pretty much zero coding experience.. I think this is probably "a little" out of my league
23:44:17TrevmarJoerg- I have a better idea - I can desolder an AC08 off the Ondio I decided to scrap - if necessary
23:45:35TrevmarThe Ondios I have are old units. I ordered some backlight kits and these seem to have the same pinout as the old Bl chip - but I am not swearing on a bible about this until my BL kist arrive :)
23:45:46 Join webguest69 [0] (~52e1440a@labb.contactor.se)
23:45:50TrevmarMy old Ondios seem to have...
23:46:31[IDC]Dragonwhat backlight kit?
23:46:36amiconn[IDC]Dragon: Once I also had to do with customs (I ordered something from Australia), but didn't have to pay anything. The only annoyance was that I had to pick up the packet at the customs office
23:47:16TrevmarJoerg the BL kits from http://www.rockbox.org/twiki/bin/view/Main/OndioBacklight
23:47:56*amiconn wonders why his latest simulator builds no longer scroll (except player sim)
23:47:57 Quit jyp (Read error: 104 (Connection reset by peer))
23:48:02 Join jyp [0] (~jp@64.198-200-80.adsl.skynet.be)
23:48:15Trevmarfrom the JELU web shop
23:48:23[IDC]DragonI wrote that, so you mean the swedish
23:48:24Bagderamiconn: I have no idea...
23:48:27Trevmaryup
23:48:35[IDC]Dragonit's a different chip
23:48:47[IDC]Dragonbut I got chips
23:49:01TrevmarOK, thats why I wanted to wait. I was just looking at the data sheet and it looked awfully similar pinout
23:49:07Bagdercube looks a lot cooler when done bigger
23:49:24[IDC]Dragonbelieve me, it's not
23:49:47TrevmarJoerg: You got chips? No problems with customs here :) You take Paypal?
23:50:03rasherBagder: I can imagine
23:50:07[IDC]DragonI didn't pay for them
23:50:09amiconnBagder: Cube doesn't seem to take perspective into account, so it looks a bit strange to me
23:50:18Bagderyes
23:50:40Bagderbut I think its cool anyway
23:50:50TrevmarOh :) OK, let me try and help track down this 03 08 problem first and then when the BL kits arrive I can worry about the BL then
23:50:54amiconnShouldn't be hard to add...
23:51:12rasheramiconn: you deranged fool :)
23:51:17Bagderthe recorder screen feels so small when you switch back and forth with the iriver one
23:51:29*amiconn would really appreciate a whole week of spare time to dedicate to rockbox coding
23:51:29Bagder:-)
23:51:35amiconnSo many ideas...
23:51:44rasherwrite them down
23:51:48HClyea.
23:51:53HClwe need a public todo/ideas list.
23:51:59HClon the wiki
23:52:04HClwhich also says who's working on what
23:52:09rasherscribble-notes-ideas-page
23:52:10Bagderwe're hardly in a lack of ideas
23:52:21Bagderseen the feature-requests?
23:52:48Bagder"Totally 586 requests (71 marked as fixed, 414 open and 101 bad)"
23:53:06rasheryeah, I noticed a bunch of rejected ones that would totally be feasible on iriver
23:53:11Trevmaramiconn: don't go rushing off on a tanget until the MMC/USB problem is fixed please -LOL
23:53:24Bagderrasher: yes, but that only adds even more ones
23:53:38Bagdermy point being: we don't need new idea for a good while ahead
23:53:43Bagderideas
23:53:51Trevmartangent - sheez my typig is bad today. Went to the dentist this morning...
23:55:59 Join amiconn_ [0] (~jens@pD95D135F.dip.t-dialin.net)
23:56:06 Join Sucka [0] (~NNSCRIPT@host81-156-215-25.range81-156.btcentralplus.com)
23:56:08rasheraw sniff, bleeding edge builds on iriver failing
23:56:16 Quit amiconn (Nick collision from services.)
23:56:16 Nick amiconn_ is now known as amiconn (~jens@pD95D135F.dip.t-dialin.net)
23:56:21amiconnGah!
23:56:33Trevmaramiconn: don't go rushing off on a tangent until the MMC/USB problem is fixed please -LOL
23:56:43Bagderrasher: failing?
23:57:11rasherWarnings, not fails I guess
23:57:14Bagderyou mean missing?
23:57:17rasheryeah
23:58:18Bagderthe bleeding edge builds should be zip files too
23:58:36*Bagder checks it out

Previous day | Next day