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

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

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

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

#rockbox log for 2005-03-04

00:00:03kergothi loooove the h320 i just bought. so nifty
00:00:05preglowthere are other factors involved
00:00:10DMJCihp-340s cost $690.00
00:00:16DMJCi've seen them in stores
00:00:32DMJCthere are different tax rates between countries
00:00:33preglowi'll have to give around 700$ for a h340 here
00:00:41ScheeI can't find anything else than H320 or H340 here... (for ~400 and ~500 EUR)
00:00:50preglowh1x0 has been discontinued
00:00:52preglowhard to find
00:01:04preglowa pity
00:01:25DMJCif I had the money I'd buy the design for the h1X0
00:01:40preglowit's a great player
00:01:42DMJCimho it's a better starting point for an mp3 player than the h3x0
00:01:45preglowthey could have made a wiser choice of cpu
00:01:55DMJChow so?
00:02:08preglowi hear the arms are great
00:02:32preglowhad a look at arm assembly, looks pretty damn powerful
00:02:48DMJCwhat do the pmps use?
00:02:54preglowno idea
00:02:56kergotharms are indeed great
00:03:27kergoth"arm assembly reads like poetry ;)" −−erikm
00:03:38preglowhe's got it backwards, PERL reads like poetry
00:03:39 Quit ashridah (Read error: 104 (Connection reset by peer))
00:04:10preglowno assembly is particularily pretty, imho
00:04:13DMJCno... poetry reads like poetry
00:04:15preglowbut arm seemed pretty powerful
00:04:15kergothperl reads like vomit spewed onto a disk
00:04:22DMJCprogram code reads like ugliness
00:04:27kergotharm's multiple register load/store are quite nice
00:04:30Bagder !
00:05:16preglowi like the 'list of possible words'
00:05:30preglowleaves no doubt that there actually went some work into this
00:05:58BagderI did a grep
00:07:52BagderI personally like the download binary option
00:08:00Bagderveeery useful
00:08:43DMJCMan strangled by escalator
00:08:52DMJCnow there's a headline you don't expect
00:10:06preglowbut now i'll have to face the problem of how to commit this code without apologizing loudly every time someone mentions it
00:10:27preglowit really is quite ugly
00:10:50BagderI'll give you permission to blame me
00:11:11preglowthe imdct_l replacement is a HUGE blob of asm
00:11:15*Bagder - the skapegoat
00:11:32DMJCAnti-gun crusader busted for illegal handgun, drugs
00:11:46 Join amiconn_ [0] (
00:11:55preglow500 lines
00:11:58preglowa bit above, actually
00:12:12Bagderthat's a big munch of asm
00:12:55 Quit amiconn (Nick collision from services.)
00:12:55 Nick amiconn_ is now known as amiconn (
00:20:16 Quit methangas (" HydraIRC -> <- 100,000+ downloads can't be wrong")
00:21:37preglowimdct_l around line 1770
00:22:03preglowdon't tell me that's pretty
00:23:27 Quit Renko ("Leaving")
00:24:04preglowif you're 68k asm wiz, feel free to assemblify the imdct routine that's #if 0'ed out
00:24:26Bagderbeen too long since I did m68k asm
00:25:12preglowthere's alot more plain ordinary 68k asm work to be done in that than the one i did
00:25:21preglowand i guess it'll be faster with data in iram
00:26:49preglowand it'll have a smaller footprint
00:28:21 Quit Chamois ("CGI:IRC")
00:28:37 Join ashridah [0] (
00:28:40Rickcan the iriver 'joystick' toggle two directions at once? (eg diagonal)
00:28:56Rickwonder how that'll work out with rockboy
00:28:59preglowit wasn't designed for stuff like that
00:29:05preglowwell, you'll quite simply just have to get used to it
00:29:16preglowyou can't even press a button and a direction at once
00:29:22preglowwith the notable exception of the play button
00:29:28DMJCwhy is that?
00:29:43preglowit's designed that way
00:29:49preglowthere isn't one wire per button
00:30:02preglowthere is one wire for all buttons, and an adc that checks which has been pressed
00:30:15preglowit's logical for an mp3 player, but not a handheld gaming device, which the h1x0 was never intended to be
00:30:15Rickno way around it?
00:30:28preglowno, no easy way apart from a really extensive hardware mod
00:30:30Bonkershow about a hardware mod? ;)
00:30:38Rickwhat about the remote plug?
00:30:41Rickis that a standard plug?
00:30:43preglowBonkers: it would have to be a really sofisticated mod
00:30:47*Rick has never seen a plug like that before
00:30:48DMJChmm.... idea
00:30:57DMJCwould it be possible to create a usb hand controller?
00:31:03preglowRick: the remote is connected the same way
00:31:04Bonkerspreglow: the buttons form a voltage divider or what?
00:31:05Rickwell, methinks data can travel both ways for the remote thing
00:31:09preglowBonkers: exactly
00:31:09DMJCthat plugs into the port on the bottom...
00:31:10Rickso I dunno
00:31:15ashridahDMJC: there's no access to the usb port
00:31:16DMJCthat'd be cool
00:31:21RickDMJC: from what I understand, it's slave
00:31:23Bonkerspreglow: so wouldn't two buttons also have a unique voltage?
00:31:27DMJCI thought it had usb host
00:31:33Rickdunno then
00:31:35ashridahDMJC: since the usb port connects directly into the usb-hdd bridge
00:31:37preglowRick: if you made a button controller that transmitted via spi, you'd have a thing that could work
00:31:39Bagderthe 3x0 has usb host
00:31:44ashridahDMJC: the 3xx can act as a host
00:31:46Bagderthe 1x0 does not
00:31:58Rickis that remote control plug a standard thing?
00:32:07preglowi have no idea, why does it matter?
00:32:10DMJCiirc h1x did have usb host
00:32:15DMJCbut it's disabled
00:32:16preglowyou'll still have the same limitations
00:32:18BagderRick: I very much doubt that
00:32:23DMJCthe hardware actually supports it
00:32:26DMJCask linus
00:32:38RickBagder: that's why i'm asking :)
00:32:38RickI've never seen a plug like that
00:32:42DMJCI remember ppl talking about that
00:32:46preglowRick: it's not so much the connector standard, it's how the cpu is wired to the connector
00:32:52preglowRick: you can't change that easily
00:33:07*Rick isn't much of a hardware guy :(
00:33:38DMJCok, so how would we go about making an spi remote?
00:33:43RickI was just thinking about that today because I havn't been able to use my remote (it's "broken")
00:34:10Rickhoping the iriver site still offers them
00:34:36preglowDMJC: well, a set of buttons and an spi interface
00:34:45Rickspi = ?
00:34:49preglowDMJC: for all i know, the spi might be unidirectional and might not work
00:34:56preglowRick: just a way of transmitting serial data
00:35:01DMJCwhat if you have an old nes remote hehehe
00:35:10Rickspi to me is a plugin system for susie (image viewer :p)
00:35:16DMJCI mean controller damnit
00:35:20Bonkerspreglow: say there's 5 buttons, that's only 2^5 = 32 combinations, why can't an ADC reconginze 32 different voltages? or is it like a 3-bit ADC?
00:35:21Camilospi is like a bus, you can have masters and slaves on it
00:35:33preglowBonkers: it's not like that, the button that you press override all the others
00:35:45preglowBonkers: combinations are out of the question from an elecrical point of view
00:36:17Bonkerspreglow: it's not 1 resistor per button and then a common resistor for the other half of the divider? I figured the button resistors would just be in parallel when multiple ones are pressed
00:36:41Rickpreglow: does the remote also override any buttons pressed on the main device?
00:36:50preglowBonkers: if you press a button, the button acts as a short circuit through to a resistor
00:36:54preglowBonkers: i THINK
00:37:26preglowBonkers: in any way, i'm pretty sure it won't work, there are schematics available, so feel free to investigate
00:37:30preglowRick: no
00:37:30Bonkerspreglow: so it's like 5V−−−−-/\/\/\−−−−-button−−−−-GND
00:37:47Rickpreglow: hmm. :p
00:38:02Bonkerserr, 5V−−−−/\/\/\−−−−button−−−−−−**−−−−-\/\/\/\−−-GND and the ** is where all the other buttons connect too
00:38:12DMJCit's wired in series
00:38:17preglowRick: haha, well, you try, one finger on the directional stick, one on the play button of the unit, and the last non-existing hand on the remote
00:38:18DMJCnot gonna work
00:38:21DMJCnuff saif
00:38:30Bonkerspreglow: what device is this exactly? or do you have a URL for the schematic
00:38:30Rickpreglow: that's the problem
00:38:32Rickmy remote is broken
00:38:34Rickso I can't test
00:38:37BonkersDMJC: I'm just curious, it's not like I actually have one of these things
00:38:38preglowBonkers: it's on
00:39:13 Part Bluechip
00:39:28amiconnpreglow: I think multi-button would work in principle. The archos button hookup is similar, different resistors per button and a common resistor connected to ground.
00:40:05amiconnThe problem is that the levels are rather close together, and the tolerances between 2 units might be larger than the difference between 2 levels
00:40:08preglowamiconn: depends on resistor choice and adc resolution, really
00:40:20preglowand adc is probably eight bit
00:40:21amiconnADC is 10 bit on archos
00:40:40amiconn(it's the SH1 integrated adc)
00:40:56preglow10 bit adcs seem to be pretty common on microcontrollers
00:41:08RickWish I knew more about hardware :)
00:41:23preglowRick: read up, it's not that hard
00:41:34amiconnOn the recorders there is another problem though. Any combination that includes the button with the "strongest" resistor leads to an adc overflow :(
00:41:46Camiloit's still a weird design for a remote
00:41:52preglowCamilo: why?
00:42:21amiconnIt saves wires, and that's the point here
00:42:27preglowwell, yes, but you have to do it like that to not have to use a parallel design
00:42:58preglowit's a perfectly sensible design
00:43:06CamiloI suppose if you already have a DAC spare and you want to build the cheapest possible remote...
00:43:27ashridahsaving space on a pcb saves you more than the engineering effort involved in designing a multiplexed system
00:43:58amiconnHowever, I wonder why archos didn't use a proper R-2R network.
00:44:07preglowamiconn: laziness
00:44:21Camilocheapest components?
00:44:39Bonkersamiconn: isn't R-2R going to use n-1 more resistors were n is number of buttons?
00:44:51preglowamiconn: if i'm not completely mistaken, there's more than one thing "odd" about the archos hardware ;)
00:45:00Bonkersor maybe n-2 more
00:45:43amiconnI presume there are integrated resistor arrays for R-2R networks
00:45:54preglowyes, there are, i think
00:46:32amiconnpreglow: Yes. That seems to be a common sense across mp3 player manufacturers though (the hardware oddness I mean)
00:46:48preglowat least iriver endian swapped the ide interface
00:47:36amiconnThinking about iriver... Choosing a CPU capable of running at 140 MHz in principle, but not stable due to thermal problems
00:47:58preglowthey probably never intended to have it running at 140 in the first place
00:48:01preglowbut who knows
00:48:09elinenbeamiconn: what speed does iriver run the chip at?
00:48:19preglowelinenbe: i've heard 96 mhz max
00:48:27amiconnI'd really like to know how Motorola manages the mp3 decoding in realtime with just 20 MHz
00:48:44DMJCwrite to the metal
00:48:48preglowamiconn: well, the source is available if you contant them :)
00:49:06preglowi guess iram is a major component
00:49:11preglowand l337-ass optimizing
00:50:01preglowtheir programmers are probably hung like bulls
00:50:21preglowi dreamt about vim the night after i did the imdct_l asm opt
00:50:25Rickhave any problems come up with the bootloader?
00:50:33preglowRick: none
00:50:41XShocKby the way, did anyone check the demo MP3 decoder on motorolla site?
00:50:54preglowXShocK: it's binary only, isn't it?
00:50:58XShocKno. source
00:51:06RickWanted to try doing some dev but reluctant because I like my H120 :P
00:51:15XShocKi will check now and say for sure
00:51:18Camiloyou have to sign up to get it though
00:51:20BagderRick: now don't be a chicken ;-)
00:51:32preglowi was the second person apart from linus to flash mine
00:51:45preglowcan't remember who the first was
00:51:53RickI've never done any dev outside of a PC
00:52:04Rickwell, aside from TI assembly. ;P
00:52:08Bagderpreglow: I could've been me, but i don't know
00:52:09preglowRick: i've mostly done pc work as well, with the exception of microcontroller stuff
00:52:20preglowBagder: don't think so, some guy just jumped in and flashed it
00:52:26Rickcan plugins cause an unrecoverable crash?
00:52:41Rickor will the RESET button take care of that just fine?
00:52:42preglowRick: everything can, if it writes the flash
00:52:51preglowRick: odds are very small of that happenin, though
00:53:00preglowRick: and yes, reset has handles everything i've ever run across
00:53:06Rickthat's very good news
00:53:13Bagdereverything anyone has come across actually
00:53:26Bagderexcept for the guy whose reset button was broken
00:53:34RickI guess I should test that
00:53:43Rickbefore I try ;)
00:54:20XShocKpreglow: it is really BINs only, libsrc.a library... :(((
00:54:32XShocKand header file for it
00:55:28preglowXShocK: yes, i thought so
00:56:26Rickjust curious, does the iriver firmware keep its settings in flash?
00:57:06preglowRick: eeprom
00:57:26preglowRick: keeping things in flash isn't very wise, it can't take more than a couple of thousand write cycles, at worst
00:58:11 Quit webguest43 ("CGI:IRC")
00:59:16preglowlooks like DMJC was the first to flash it
00:59:53preglowDMJC: madman
00:59:53DMJCnow they tell me
01:01:09DMJCit worked beautifully
01:01:17DMJChaven't flashed it since
01:01:29amiconnBagder: rockboy.ovl ends up in .rockbox/rocks, not in .rockbox/viewers where it should be....
01:01:40preglowi've flashed it a couple since
01:02:03preglownothing more nerve wracking than the first time, when it just blinked its hdd light for about five minutes or so
01:02:23preglowi was more or less preparing to package the thing in and ship it off for sweden, heh
01:02:45DMJCI'd have had to give it to a mate of mine to work on
01:02:49DMJChe's a hardware junkie
01:03:42DMJCit's funny actually... if he saved his cash instead of buying all the crappy players he has... he'd be halfway to an iriver
01:03:46***Saving seen data "./dancer.seen"
01:04:06 Join joe1 [0] (
01:04:27 Quit Aison ("( :: NoNameScript 3.72 :: )")
01:04:46Rickhow are file operations on the iriver?
01:04:56preglowspeedy enough, why?
01:05:02Rickjust curious
01:05:15DMJCyou want to move data around?
01:05:15 Quit joe1 (Client Quit)
01:05:36Ricknah, just random thoughts popping into my head
01:06:00DMJCfile cut/paste would be kinda cool
01:06:08preglowwhy? :P
01:06:31DMJCso after you've leached music from a friend you can spend the bus trip home organising it
01:06:41preglowwill be doable, for sure
01:06:41RickDMJC: the latest firmware added deleting iirc
01:06:45preglowbut i think i'd wait
01:06:57preglowRick: rockbox can rename as well!
01:07:01RickI was thinking of making a text editor actually
01:07:08Rickan efficient vkeyboard or something
01:07:13preglowthere is on already
01:07:16DMJCso I can edit my phonebook..
01:07:20preglowtry renaming a file
01:07:32RickDon't have rockbox on my iriver yet
01:07:43Ricki'm building a gentoo vmware install so I can have access to buildtools :p
01:08:07preglowthere are limits to how efficient the keyboard can be
01:08:09preglowbut it works
01:08:30RickI was thinking about how to do a vkeyboard
01:08:45RickI was thinking different keyboard modes, lowercase/uppercase/macro (user definable)
01:09:20DeadManvkeyboard is pretty easy with joystick
01:09:32DeadManscroll to letter click
01:09:36*Rick nods
01:09:47DeadManhold rec key for shift click
01:09:54Rickyou can't press two keys at once
01:09:58Rickhence mode changing
01:10:04 Quit elinenbe (" Like's GUI? Then try HydraIRC -> <-")
01:10:21DeadManwell double click then
01:10:37DeadMandouble click in joystick for cap
01:11:18 Quit Camilo (Read error: 60 (Operation timed out))
01:11:28DeadManor just use rec key for click on for cap...alsorts of ways
01:11:40preglowyou can hold play and press other things
01:11:53preglowbut that's the only exception
01:12:14DeadMananyhow nothing amazing gapless working yet on iRiver?
01:12:24preglowwe have no codec api yet, so no
01:12:40RickI was actually thinking of using a/b key for modes
01:12:51*Rick shrugs
01:12:56DeadManoh so that libmad at 104% ain't worth anything unless it has some place to output to?
01:13:13preglowwhich reminds me i should have commited libmad code
01:13:21Rickpreglow: are there any plans on "cleaning up" the plugin api?
01:13:25preglowbut i should have done that before i had two beers and grew tired
01:13:30preglowRick: in what way?
01:13:50Rickrenaming things to better suit the fact that multiple players are supported
01:14:00Rickmp3_* -> audio_* or something :P
01:14:11preglowahh, like that
01:14:16preglowthat's got nothing to do with plugings
01:14:20preglowand we'll get around to that
01:14:23Bagderfeels pretty secondary to me
01:14:28Rickyeah, just curious
01:14:31preglowit's just semantic
01:14:34*Rick nods
01:15:03*Rick watches his gentoo stage tarball extract
01:15:20preglowwhat we really need is people working actively on the codec api
01:15:23XShocKso the current problem is only in the architecture of the sound framework, since all parts: decoding and playing are done.
01:15:33preglowand perhaps another person on codecs optmisations
01:16:38preglowbut it doesn't matter, it's just been one month since the first person flashed his player, we've really gotten quite far
01:17:01RickI had been checking routinely for updates and figured I should check it out in more depth
01:18:08preglowi've just been involved for little under a month myself
01:18:16DeadManyeah lot of progress...makes me optimistic it will be nearly there in the next month ;)
01:18:44RickDeadMan: nah... not that quick ;P
01:19:12DeadMandepends on what's happening and who works on what
01:19:13XShocKi am sure that all features of a normal player will be done in next month. :)
01:19:39RickI thought it was a matter of simply 'fixing' the features and retaining the features that rockbox already has?
01:19:49Rick(and then, implementing features afterwards)
01:20:01Ricknew features that is
01:20:04preglowRick: there's more to it, h1x0 series brings along a lot of possibilites rockbox never had before, like software codecs
01:20:18RickI just thought there would be more focus on getting what is already there to work first
01:20:20preglowand that alone is a very big part
01:20:25preglowwell, it does
01:20:32preglowrockbox does work
01:20:45preglowthe only things it lacks are stuff that was handled by hardware before
01:21:17DeadManmore software coding involved on the codec end
01:21:30BagderRick: you really should install rockbox and try it out
01:22:13preglowRick: hell, if it does nothing but bore you to death, you can still show your friends how cool you are by loading super mario land on your mp3 player!
01:22:20amiconnBagder: Got my remark concerning rockboy.ovl ?
01:22:26DeadManI'm not installing it until I can get MP3 playback with gapless. it's the only reason for otf playlist I guess but that's about it
01:22:33Bagderamiconn: I did, fixing it tomorrow
01:22:34Rickwhat is gapless?
01:22:35 Quit matsl (Remote closed the connection)
01:22:40preglowRick: no gaps between songs
01:22:40amiconnBagder: Okay.
01:22:46Rickdoesn't the latest firmware have that?
01:22:50XShocKRick: do it.. you will not regret it at least. :)
01:22:51preglowRick: no
01:22:54preglowRick: it never will
01:22:56Rickwhat's the option for?
01:23:08preglowRick: god knows, it doesn't do what it should
01:23:13DeadManseamless playback of DJ mixes and live recordings
01:23:25RickI hadn't tried it yet
01:23:26preglowyes, and electonic music in general
01:23:30RickI guess it's like that language menu
01:23:33preglowmore than half of the cd's i own have no gaps between songs
01:23:43Rickyeah i've noticed that myself
01:23:52Rickthat's one feature I wished winamp had
01:23:55Rickplaylists in playlists
01:24:11preglowdump winamp, use foobar2000
01:24:24Rickbasically, a playlist in the playalist would play that playlist with no shuffling in order
01:25:17XShocKare there people good at electronics?
01:25:24preglowXShocK: perhaps
01:25:52Rickdefine electronics
01:26:06XShocKi want to add RTC to iriver.
01:26:15preglowreal time clock
01:26:31XShocKthrough the i2c line, for example fm-radio bus line.
01:26:32preglowXShocK: you'd have to find a free interrupt pin
01:26:51preglowor some way of having the rtc switch the cpu on
01:27:20*Rick watches portage extract
01:27:36XShocKfirst i wanted to implement the timing only, no auto-starting on alarm.
01:27:37RickI think I should have picked a lighter flavor of linux for the vmware buildtools :P
01:28:03preglowRick: gentoo is for men
01:28:17XShocKi have a DS1340 chip, it has i2c bus.
01:28:20RickI like gentoo, don't get me wrong
01:28:31Rickjust taking forever to install atm ;)
01:28:42preglowXShocK: well, if i don't remember incorectly, you can just stick things strait onto the i2c bus and have it work
01:28:47 Quit lolo-laptop ("Client exiting")
01:28:54preglowXShocK: provided you've got a driver, of course
01:29:07preglowRick: yes, i know, i had to install xwindows on it today
01:29:23XShocKfrom what i understood i don't need to change any component on the player itself. i only attach the chip through resistor on bus lines.
01:29:28preglowXShocK: yes
01:29:30Rickluckikly i'm just using a stage3 for the vmware install ;P
01:29:31preglowXShocK: that's correct
01:30:17XShocKso, what i am worring about is electrical current while the player is off, but the rtc is working.
01:30:36XShocKshould i cut the rtc from i2c bus when the player is off?
01:31:08preglowit shouldn't try to sense the bus when power is off, no
01:31:28preglowbut i don't know how you should handle that
01:32:26XShocKso, if i let rtc chip be connected to bus while all other chips do not have power, then i will get all sorts of troubles, yes?
01:32:54preglowdoes anyone know if sim buils vomit on seeing something declared belonging to section .idata og .icode?
01:33:15preglowXShocK: you might, i'm not sure
01:33:45Bagderpreglow: I believe it defines the attribute() to nothing for the sims
01:33:53preglowBagder: that was what i was hoping
01:34:13amiconnpreglow: Iirc these sections don't matter for the sim builds. They are included just like normal .text / .data sections
01:34:16preglowi believe linus agreed on having a #define to do the attribute iram stuff, but hasn't been done yet
01:34:41amiconnI believe I've seen this while searching for the cause of the non-working lseek(()
01:34:58 Join elinenbe [0] (
01:35:10preglowbtw, anyone know of a free noncygwin x server?
01:35:11amiconnpreglow: HCl does that in rockboy
01:35:17preglowtesting the sim myself would be good
01:35:30preglowamiconn: ok, i'll drop ifdefing it then
01:35:38XShocKok. i am pretty afraid i need to isolate it, but it will make everything pain in the ass since i will need additional components to put...
01:35:45amiconnIf you're on windows, why don't you use cygwin?
01:36:08preglowXShocK: depends on how the bus is wired
01:36:09Rickbecause cygwin is bizaarely slow at compiling
01:36:29preglowXShocK: ask linus, he should know this
01:36:31amiconnI use it here; can't say that it is excessively slow
01:36:46preglowamiconn: i've never been able to make cygwin work very well
01:36:46Bagderamiconn: compared to a linux build, it is
01:36:47RickI havn't touched cygwin in years
01:36:50Rickso it could be different
01:36:52preglowamiconn: so i've grown to dislike it ;)
01:37:18XShocKyep, i will ask him. he addmitted that i don't need a pull-up resistor, but i forgot to ask him that problem..
01:37:20preglowi've got a linux box anyway, i prefer to use that
01:37:46Rickiriver doesn't have floating point right?
01:37:54preglowRick: correct
01:38:59amiconnBagder: When rebuilding my whole assortment of simulator & target builds, it doesn't take significantly longer than the official build sequence. That's on a laptop btw
01:39:40preglowamiconn: depends on what you're doing. configure is insanely slow on cygwin
01:40:12amiconnThe only slow thing in configure is the language selection - takes a few seconds
01:40:21preglowahh, didn't mean your configure
01:40:24preglowautoconf configure
01:40:51amiconnrockbox doesn't use autoconf, so I can't tell much about that
01:41:03XShocKi think getting clocks and probably alarm on the cost of 4 el. components. will be good. i actually was pretty delighted that DS1340 has such small size(2 by 2 millimeters. :) )
01:41:23preglowi tried to use cygwin to build csound, failed completely in making it work, ended up installing linux on this box as well
01:42:34elinenbeto all the developers... awesome work!
01:42:38amiconnHowever, I one tried building gcc, and I built gdb
01:42:40 Join DrRick [0] (
01:42:52elinenbeHCl: I used to joke about a gamboy emulator on the archos... now that's a reality!
01:43:15amiconnelinenbe: yup.
01:43:52XShocKgoing to sleep. see you all
01:45:25 Quit XShocK (" Try HydraIRC -> <-")
01:45:28preglowyes, i'll do that myself
01:45:29preglowgood night
01:45:44 Quit preglow ("off")
01:47:08 Join XShocK [0] (
01:48:03XShocKin file firmware/mp3_playback.c on line 1175, shouldn't it be "#else /* CONFIG_HWCODEC == MASNONE */"
01:49:18XShocKjust a very small thing of course, but still
01:51:25XShocKoops. ignore it... i just messed up with the way the code was commented
01:51:35XShocKi really need to go to sleep.
01:51:37 Quit XShocK (Client Quit)
01:55:59 Join tHefaInt [0] (
01:57:00 Part tHefaInt
02:01:11BagderI remodeled how the cvs builds are made
02:01:40Bagderthey now check for new changes every minute
02:01:49RickBagder: you're an official rockbox dev? :)
02:02:01*Bagder puts on his hat
02:02:10Bagderthat is Mr Bagder please ;-)
02:02:13*Rick doesn't know who anybody is yet
02:02:20 Join lostlogic [0] (
02:02:48BagderI'm one of those in-it-from-the-start
02:10:17Bagdertime to sleep
02:10:28Rickwhat timezone?
02:10:39Bagdercentral european
02:10:44Bagder02:10 now
02:10:51RickPST here :
02:19:16rickst131CST here...
02:22:32 Join lostlogic_ [0] (
02:30:00 Quit DMJC ("Leaving")
02:30:55 Join DMJC [0] (
02:36:11 Nick midk_ is now known as midk (
02:36:32 Join skav [0] (
02:39:44 Join geoff_o [0] (
02:40:07 Quit lostlogic (Read error: 110 (Connection timed out))
02:49:02 Part geoff_o ("Kopete 0.9.1 :")
03:03:50***Saving seen data "./dancer.seen"
03:21:19Rickla la la
03:21:26*Rick frollicks
03:23:36 Quit skav ()
03:37:55 Quit DrRick ()
03:39:09 Quit midk ("Leaving")
03:39:23 Join midk [0] (
03:42:43 Quit midk (Read error: 104 (Connection reset by peer))
03:48:24 Join midk [0] (
04:06:11 Join QT_ [0] (as@area51.users.madwifi)
04:14:43 Quit QT (Read error: 60 (Operation timed out))
04:20:16 Join ze__ [0] (
04:27:36 Quit Stryke` (Read error: 110 (Connection timed out))
04:32:21 Quit ze (Read error: 110 (Connection timed out))
04:32:21 Nick ze__ is now known as ze (
04:36:35 Quit ze (Read error: 104 (Connection reset by peer))
04:40:40 Join ze [0] (
05:03:52***Saving seen data "./dancer.seen"
05:11:55 Join ze__ [0] (
05:14:53 Join wacky_ [0] (
05:15:01wacky_any big guys in here ? :)
05:15:21Rickdefine big guy
05:15:49wacky_this afternoon I came and asked for my plugin to be added to CVS.. and preglow say I should wait for the big guys that take decisions to be there :)
05:16:14Rickthey're probably all sleeping atm
05:16:23wacky_yeah.. just saw the current.txt log..
05:16:30wacky_I'll be back!!
05:16:34 Quit wacky_ (Client Quit)
05:23:19 Quit ze (Read error: 110 (Connection timed out))
05:23:19 Nick ze__ is now known as ze (
05:28:48Rickyay ;)
05:32:45 Quit cYmen ("zZz")
05:44:25Rickgot my gentoo install up and going
05:45:28Rickgoing to setup my build enviornment now
06:12:34Rickmaking binutils ;D
06:21:14Rickmaking gcc....
06:42:57pabsHCl: adds load and save state (both to quicksave slots and to arbitrary files, although the latter needs a litle refinement)
06:43:03pabsHCl: i'm sending it to the mailing list also
06:44:51RickError: unrecognized architecture specification '5249'
06:46:16ashridahyou didn't build gcc with the newly built binutils added to the path afaik
06:46:28pabsit's binutils that's wrong
06:46:36pabsi had the same problem a couple days ago
06:46:46RickI had it in path
06:47:20pabsRick: m68k-elf-ar −−version
06:47:28pabsRick: it'll probably give you th e wrong thing
06:47:40RickGNU ar 2.15
06:47:50pabsRick: yeah, you need something newer than 2.15
06:48:07Rickso what needs newer, binutils or gcc?
06:48:20pabsRick: you need to redo binutils
06:48:21Ricknevermind, stupid question
06:49:11RickCrossCompiler should be updated to say that somehow ;P
06:52:52ashridahit probably didn't need to to begin with. feel free to update it with current info
06:54:55Rickit mentioned getting latest cvs snapshot
06:54:59Rickbut it said it was an alternate
06:55:22ashridahyeah, it isn't really.
06:59:18ashridahthe problem is, it probably still is for the SH platform archos uses
07:00:12Rickfirst it lists (what I assumed were the 'default')
07:00:20Rickthen under Coldfire it says the cvs binutils is optional/alternate
07:00:28Rickso I assumed I could use binutils-2.15
07:02:58Rickhence confusion
07:03:53***Saving seen data "./dancer.seen"
07:05:58 Join ze__ [0] (
07:06:46ashridahas i say, feel free to correct it
07:17:23 Quit ze (Read error: 110 (Connection timed out))
07:17:24 Nick ze__ is now known as ze (
07:20:22Rickrebuilding now ;P
07:21:53 Join StrathAFK [0] (
07:26:13 Join LinusN [0] (
07:27:46Rickgreetings LinusN
07:33:46LinusNRick: the CrossCompiler page doesn't say that CVS binutils for coldfire is alternate
07:34:25LinusNalternative 1: get directly from CVS
07:34:41LinusNalternative 2: get a CVS snapshot
07:35:30Rickunclear definition of alternate then :p
07:35:46Rickno biggy
07:36:08LinusNyes it is, if it's unclear we need to change it
07:36:44Rickwell, instead of saying alternate you can say 'OR' in front of the get a CVS snapshot ;)
07:37:20LinusNare you editing it?
07:37:37LinusN"There is a possibility that RickLaCharite is editing this topic."
07:37:47RickI forgot to release
07:37:56Ricktry again :)
07:38:49 Quit Strath (Read error: 110 (Connection timed out))
07:40:45LinusNtake a look at it now
07:40:55Rickyay :) building rockbox
07:41:02*Rick looks
07:41:13Rickmuch more clear, :)
07:41:27LinusNwiki is a blesing
07:41:34*Rick nods
07:42:29Ricknice, all built with no errors ;)
07:44:31LinusNhave you installed the boot loader?
07:44:35Ricknope, not yet
07:44:54Rickwas about to do that now
07:45:16LinusNword of advice: use the bootloader.bin on the wiki page
07:45:33Rickalright, will do
07:48:13pabsLinusN: i fired off another rockboy patch to the mailing list
07:48:52pabsHCl doesn't have email, and i can't dcc at all, and it's inconvenient for me to access ftp (because of my firewall)
07:48:56pabskind of amusing
07:49:10LinusNdoesn't have email?????
07:49:16LinusNwhat a caveman
07:49:18pabshe said he doesn't have easy access to it
07:49:24pabsor something
07:49:29pabsor not a good one, i can't remember exactly
07:49:42pabsbut i was basicall like "where do i send this patch?", and eh said dcc or ftp
07:50:08LinusNin fact, the patch tracker is the official way of submitting patches
07:50:29*pabs searches ye ole wiki
07:50:56 Join Rick-Gentoo [0] (
07:51:59 Quit Rick-Gentoo (Client Quit)
07:52:54LinusNpabs: i like the looks of your menu, but it seems redundant to have two menu systems
07:53:14LinusNwould have been better to export the core menu code to the plugin api
07:53:14pabsLinusN: where's the other one?
07:53:29pabsLinusN: i looked for a menu api and didn't see one handy
07:53:38pabs(in the plugin api documentation, etc)
07:53:42pabsso i wrote that one
07:53:51LinusNin fact, i like it
07:54:11LinusNit has a few drawbacks though
07:54:27pabsprobably, i just whipped it up
07:54:28Rickoh that's evil
07:54:28LinusNone obvious is that it can't handle long strings
07:54:31ashridahthe one in rockbox? it does look decent. a smaller font might be good, plus a scrollbar...
07:54:32*Rick heads to recompile is kernel
07:54:46pabsLinusN: nope, and it won't handle huge fonts gracefully either
07:54:51pabsLinusN: that's easy enough to fix though
07:55:18pabsLinusN: like i said, i ddin't see a menu api, so i tossed that one together
07:55:22pabsLinusN: i'm still refining it
07:55:39pabsLinusN: mainly i was mad because rockboy was quitting on me, and i wanted to be able to save tsate
07:55:45pabsLinusN: (which the second patch adds)
07:59:32Rickthere, kernel rebuilt with vfat support :)
08:01:00pabsLinusN: i can either crop the strings to fit them in the menu, or add a thread that scrolls them (since i kind of like having the game showing a bit in the background, it gives a nice modal feel to the menu)
08:01:32LinusNpabs: the rockbox lcd driver already supports scrolling
08:01:37pabsLinusN: if you're going to expose the menu stuffto the plugin api it'd be nice to have a standard open and save file call too
08:02:15LinusNpabs: "a standard open and save file call"?
08:02:38pabsyeah, somtehing that brings up a menu that allows users to scroll around and select a file
08:02:58LinusNa file requester?
08:03:07LinusNwe don't have one
08:03:19LinusNfeel free to write one
08:04:36pabsLinusN: i know the lcd driver supports scrolling, but i didn't think it'd support scrolling for part of the screen (like, a chunk of text that wasn't full screen)
08:05:07Rickpabs: sorta like buffer areas within the main lcd?
08:05:16LinusNthat's true, it needs support for cutting off the right part
08:07:11pabsRick: yeah
08:07:23pabsthat's what i'd want to do text scrolling in teh rockboy menu i did
08:07:42pabsmaybe i'm just picky :o
08:07:45Rickoh great, forgot to unzip
08:07:51*Rick feels silly
08:09:04Rickwhat's the modified firmware do?
08:09:16RickI mean, other than load up rockbox
08:09:41pabsRick: well, it makes you popular with the ladies!
08:10:23Rickjust curious because I got a nice blank screen for ~5 seconds before it did anything
08:12:46Rickother than that it seems i'm all good to go
08:14:22LinusNit doesn't do anything else but load rockbox, but the FAT driver might recalculate the free space the first time you boot
08:15:04LinusNand it might take a few disk operations to find out that the rockbox files are missing before it boots the original firmware
08:15:14Rickah, okay
08:16:18LinusNpabs: is see that windowing and menus is your thing (profanity) :-)
08:17:14pabsLinusN: damn i just posted that!
08:17:22pabslike 10 minutes ago =-O
08:17:42pabsi was just posting something about rockbox too :D
08:17:43Rickposted what?
08:18:14pabsRick: i just posted an entry to about a library i'm working on called profanity
08:22:38Rickit's working ;)
08:22:43pabsRick: cool!
08:24:00pabsLinusN: the profanity stuff is really just for raggle :)
08:24:04Rickthe md5 hash list on the IriverBoot page −− is that of an encoded final merged firmware?
08:24:09Rickor unencoded?
08:24:28pabsLinusN: but i'm planning on using it for other stuff too (eventually)
08:25:16Rickthere, added
08:26:00pabsLinusN: it really scares me that you saw the post on my page that fast
08:26:05*pabs looks around for cameras
08:26:18LinusNRick: the page says "No need to add md5 duplicates"
08:26:24RickOh, sorry
08:26:42LinusNpabs: pure luck
08:26:44Rickdidn't realize there was a duplicate ;)
08:26:57dwihnoyay! friday!
08:27:25dwihno7 hours until state of euphoria ;)
08:28:20RickI like the upside down setting
08:29:15LinusNquite handy when you have it in your belt
08:29:27Rickyeah :)
08:29:53LinusNespecially for the archos that doesn't have an lcd remote
08:30:45Rickgraphical glitches with pong
08:31:13LinusNlots of plugins needs work on the iriver
08:31:22LinusNthey need work overall
08:31:36LinusNlike a common menu system :-)
08:32:10LinusNpabs: you're up to a fairly big task?
08:32:52Rickdoes the current iriver rockbox function fine with file transfers over usb?
08:33:29LinusNinsert the cable and see for yourself
08:33:35Rickokay ;P
08:34:42Rickthat's nice and speedy
08:37:43Bagderwhat no commits yet today?
08:39:24*Bagder wants the new script tested
08:42:08Bagderdaily ondio fm build works again now
08:44:01Rickdoes snake2 fit the entire screen of the archos?
08:55:30pabsLinusN: i am?
08:55:51pabsLinusN: what would you awnt in a common menu system?
08:56:29pabs(sorry, i was off taking pictures of rockboy for pablotron)
08:56:32LinusNif you're up to it, i'd like you to add the right-side clipping for the lcd scrolling code
08:56:57LinusNso we can use the standard menu system for popups too
08:59:05 Join ze__ [0] (
08:59:32pabsseems like it'd be as straightforward as adding a call with a width parameter
08:59:34pabsfor that case
08:59:51pabsbut what you really want is subwindows and with clipping regions, and htings like that
09:00:01pabswhich is more work
09:00:28LinusNthat may not be what we really want
09:01:15pabsLinusN: i'm still getting up to speed on the guts of rockbox, i don't mind writing that code for you but i'm not sure how you'd want it done
09:01:18 Join einhirn [0] (
09:01:48LinusNfirst thing is to add a width parameter to lcd_putsxyofs()
09:02:06LinusNand then make the scrolling thread use it
09:02:40LinusNperhaps adding width/height (or right/bottom) to lcd_setmargins()
09:03:14LinusNa poor man's clipping window
09:03:15pabsah i see wher eyou're going with that
09:03:34Rickis there any way to obtain the width of a string rendered in a font?
09:03:55***Saving seen data "./dancer.seen"
09:04:00LinusNRick: int lcd_getstringsize(const unsigned char *str, int *w, int *h)
09:04:14Rickgreat :)
09:04:35*Rick is working on a basic scale hack for snake2
09:08:35LinusNRick: perfect
09:09:56 Quit ze (Read error: 110 (Connection timed out))
09:09:56 Nick ze__ is now known as ze (
09:10:58Rickworks, just some graphical glitches
09:11:05Rickand I need to figure out how to handle the fact that the Y scale is warped
09:11:11Rick(it's scaled more than X)
09:14:10ashridahRick: it's because the lcd display won't be square
09:15:26*Rick nods
09:15:27RickI know
09:18:18Ricknearly fixed the graphics for scaling :)
09:33:30 Join Sando [0] (
09:34:03 Join ghode|afk [0] (
09:35:38ghode|afkLinusN: wacky_ is looking for you i think, concerning the m3u converter plugin, would be nice to get that in cvs ;p
09:41:02pabsLinusN: i finished converting the api calls everywhere for setmargins and setxyofs, but it'll take more time to make the margni calls do what they're supposed to
09:41:29LinusNcan i see the patch?
09:42:10pabsi'll have to strip out teh rockboy stuff though
09:42:16pabs(i'm just doing the changes in the same tree)
09:42:17HClugh, morning.
09:42:37pabsHCl: did another patch for you with load/save stuff
09:43:03HCli'll take a look at it.
09:43:16HCli'm still rather cranky at my toothache
09:43:28pabsHCl: it's got 5 slots for each and a save to/load from file option (although the latter could use some work)
09:44:29LinusNpabs: have you tried your patch on a real target?
09:44:50LinusNthe save/load patch
09:45:03pabsi'm using it here
09:45:10pabson my iriver
09:45:39LinusNi'm puzzled by the lack of plugin api handles
09:45:51LinusNopen() instead of rb->open() etc
09:45:58ghode|afkHCl: do you have any plans to add .gbc viewer support to rockboy?
09:45:59pabsLinusN: oh that's from the rockboy defines
09:46:05pabsLinusN: presumably HCl did those
09:46:18Rickghode|afk: how would that work without color support?
09:46:27pabsLinusN: he's got #define open rb->sys_open and stuff
09:46:34ghode|afkdid not mean it like that ;p
09:46:43pabsLinusN: hence the lack of api calls
09:46:54ghode|afkyou can play gbc roms atm, but you can't open them from the dir view, only .gb roms
09:46:55pabsLinusN: i was just going with his system
09:47:16pabsghode|afk: if you add an entry to your viewers file they'll show up
09:47:50ghode|afktotal novice here ;p
09:48:04pabshey, me too! :)
09:48:10pabsLinusN: i'm editing your patch at the moment
09:48:16Bagder.gbc? they're named .cgb in the viewers.config file
09:48:38ghode|afkhmm all the gbc roms i've vound have had the ext .gbc
09:48:39pabsBagder: all teh ones i've ever seen are .gbc i think
09:48:42BagderI fixed the problem with multiple extensions for a single plugin yday
09:49:01Bagderthen "someone" should fix the viewers.config
09:49:39Bagderthe problem was only in how the viewers.config is made when make zip is used
09:52:02HClpabs: are you writing the filename to the savestate too? o.o;
09:52:32HClah... description, ok.
09:55:25pabsHCl: description for the quicksave slots
09:56:13pabsLinusN: it's a start anyway
09:56:31 Nick kergoth is now known as kergoth`zzz (
09:58:24pabsLinusN: i'd need to grep through the code and find all the places that check LCD_WIDTH, and swap that to LCD_WIDTH - rmargin, in order for the margin stuff to be functional
09:58:50pabsLinusN: but that change compiles clean here
09:59:03LinusNmay not be necessary
09:59:46LinusNignore me
10:00:22pabsobviously that's only in lcd-h100.c and lcd-recorder.c, so it's not too bad
10:00:40LinusNpabs: while we're at it, why not move the common code to a new file, lcd-common.c
10:00:48Bagderwe should unite more things between lcd-h100 and lcd-recorder
10:00:56Bagderpossibly in a new lcd-bitmap.c
10:01:12pabsi noticed a lot of stuff was duplicated in there
10:01:16LinusNlcd-bitmap.c might be better than lcd-common.c
10:01:16 Join vrm [0] (
10:01:50pabsLinusN: is that roughly what you're looking for though?
10:02:08vrmit's possible to remote control the rockbox using USB, for example start playing a song ?
10:02:36LinusNi'm not sure if i want left/right/top/bottom or x/y/width/height
10:02:53LinusNvrm: no
10:03:01Bagderthe new build script seems to work
10:03:21pabsLinusN: seems like margins are usually relative to their respective edge
10:03:25pabsLinusN: but it's all the same to me
10:03:40pabsi was just going on how people seem to do them normally
10:03:56pabsi guess i'd call it a clipping region if it's x/y/w/h, and a margin if it's relative ot the edge
10:04:01pabsthat's just semantics though :D
10:04:08LinusNyes, l/r/t/b is appropriate for a setmargins function, but maybe it would be better with a setwindow() function, i don't really know
10:04:49pabsLinusN: if you prefer w/h that's doable
10:04:49vrmLinusN, thanks
10:04:54pabsi should sleep befoer my girlfriend kills me
10:05:02pabs(it's 4am here)
10:05:28LinusNi think i'll apply your patch in the meantime
10:05:34pabsLinusN: i'll catch up with everyone tomorrow at some point
10:05:41LinusNsleep tight
10:05:50pabsnight all!
10:05:53pabsLinusN: thnks
10:06:36pabs(hopefully id idn't mangle anythign when i stripped out the rockboy changes)
10:06:43pabsguess we'll find out
10:06:57vrmLinusN, so the USB communication is enclosed in some mass storage chipset I suppose ?
10:07:41vrmannoying :/
10:10:10 Join Patr3ck [0] (
10:11:19DeadManHow far away would you say you guys are from MP3 playback to the headphone out? :)
10:11:54Bagderabout this: −−−−> <−−−−
10:12:18DeadManJust insert ___ then ;)
10:12:24 Join markun [0] (
10:14:04markunI just did a fresh build this morning. booting rockbox: "FileType string buffer empty". Also my settings were all gone, there were duplicate entries in the "open with" menu and vorbis2wav was not there..
10:14:17Rickyes, I get the FileType string buffer empty too
10:16:09LinusNhow does your viewers.config file look like?
10:16:59*LinusN points at Bagder
10:17:14*Bagder ducks
10:18:19 Join preglow [0] (
10:19:54Bagderactually I bet it is a true bug
10:20:03Bagderthat my fix reveals
10:20:34Bagdersince viewers.config never contained more than one plugin <=> extension before
10:20:48Bagderand now a single plugin suddenly is listed for multiple extensions
10:21:35Bagderat least I can't see any problem with my generated viewers.config
10:24:37LinusNviewers.config was never designed to contain more than one entry per plugin, iirc
10:24:45ghode|afkBagder: the cgb ext should be "gbc" not "cgb"
10:26:15BagderLinusN: did we really limit it to that?
10:26:28Bagderseems very restrictive
10:26:35 Part vrm ("Leaving")
10:27:30Bagderand it would even more explain the problem :-)
10:27:37LinusNmaybe i am wrong, but iirc, the code has that limit
10:27:53LinusNthough it shouldn't, imho
10:27:59BagderI agree
10:28:04LinusNBagder: go fix
10:28:13Bagderthe buildzip flaw has hidden this limitation all this time
10:28:24LinusNpreglow: how's the emac today? :-)
10:29:04DeadManIt's snowing (And settling) like crazy outside :P
10:33:29DeadManHow much % of Coldfire CPU does iRiver's MP3 decoder take?
10:34:27DeadManOnly I made what may be some rather bogus statements in the misticriver forum :)
10:34:37preglowLinusN: ugly
10:34:42LinusNDeadMan: i have no idea
10:34:44preglowLinusN: but i'll fix it up now
10:35:32markunLinusN: viewers.config looks just like before. No duplicate entries and vorbis2wav.rock is also there.
10:36:16LinusNon the actual target?
10:36:39markunWhere are the rockbox settings stored on the iriver?
10:37:10LinusNin an unused sector after the partition table
10:37:25markunhow strange then that they got wiped..
10:37:25DeadManoh so it loads it at boot time?
10:37:30pregloware we planning to use the eeprom for anything?
10:37:39LinusNnot sure
10:37:52DeadManand after a format it still retains those settings?
10:38:16 Join El_Barto2 [0] (
10:38:44DeadManmust be something to tell even when you are formatting via usb not to go near that portion of the disk
10:39:10BagderDeadMan: as long as you have a partition, you won't risk that
10:39:17DeadManI'd hate to see a bad sector appear there
10:39:19Bagderand even if you did, it is only your settings
10:39:27DeadManthat would annoy
10:39:36DeadMankeep losing settings at power off that is
10:39:36Bagderexpect to be annoyed
10:39:42Bagderwe frequently change format of it
10:40:20DeadManI would have thought they would hold settings in memory for as long as the battery was charged
10:40:23Bagderkeeping up with devel means blowing your settings every once in a while
10:40:45BagderDeadMan: they? rockbox surely does
10:41:01DeadManI would have thought so
10:41:07LinusNBagder: no we don't
10:41:17Bagderyes we do
10:41:18DeadManthe more I look at it the more I think iRiver devs are lazy
10:41:20LinusNthe iriver doesn't have an rtc
10:41:20Bagderbut we save them
10:42:02Bagderoh, maybe I misunderstood
10:42:21 Nick StrathAFK is now known as Strath (
10:42:24BagderI thought you meant while the unit is powered
10:42:45ashridahDeadMan: most modern hard drives perform sector remapping if they develop a bad one anyway
10:43:07LinusNsome day, we will use a file instead of a sector
10:43:30DeadManYou could have profiles
10:43:47DeadManboot to profile a,b,c
10:43:54DeadMandifferent settings for each
10:43:56Bagderwe already have that
10:44:00DeadManoh ok
10:44:02Bagderthey're called config files
10:44:09DeadManI've not tried Rockbox yet
10:44:24DeadManI don't see any point for me until I get MP3 playback ;)
10:44:32Rickwhy don't you use a file already?
10:45:38LinusNhistorical reasons
10:45:45DeadManwill the gapless portion of the code have to be rewritten for the iRiver?
10:46:07LinusNnot for pure mp3 gapless playback
10:46:43DeadManso how will audio be loaded into memory?
10:47:01DeadManiRiver's firmware buffers the entire memory up
10:47:17BagderDeadMan: that's the only way to do it
10:47:46Bagderneed to keep the disk spinning as little as possible
10:47:47DeadManthat's how it saves on power too I guess
10:48:21DeadManso for gapless it would need to buffer in the next song as well or at least the first portion?
10:49:16DeadManI mean. If the buffer is already full...
10:49:41Bagderthis is nothing new for us
10:49:48Bagderrockbox already plays on Archos, you know
10:49:52DeadManjust curious :)
10:50:03Bagderif the buffer is full, there's no need to load more data
10:50:22Bagderthen you play music until there's a need to load more data again
10:50:45DeadManyeah ic
10:51:09DeadManI forget playback and read into memory are seperate events. linear thinking me ;)
10:51:43preglowthe iriver firmware doesn't even buffer a song if it can't fit it in the buffer
10:51:52preglowthat's plain stupid, if you ask me
10:52:05DeadManso if the song is bigger than 32mb it does what?
10:52:31 Quit ghode|afk (Read error: 110 (Connection timed out))
10:52:41DeadManjust spins the disk continually?
10:53:27DeadManif that's a stupid question I apologise :)
10:55:21*DeadMan wonders how much CPU iRiver's firmware uses for MP3 playback in order to get 16 hours battery life
10:55:31preglowDeadMan: it handles a single song bigger than the buffer
10:56:32LinusNpreglow: so how's the libmad opts going?
10:56:33DeadManI know. Just wondering what you meant by doesn't even buffer if the song can't fit in the buffer
10:56:41LinusNwhat do we have left to do?
10:58:22DeadManSomeone gonna do some testing on battery life, decoding an MP3 in Rockbox? :)
10:58:49LinusNsomeday we will
10:59:19HClmy rockbox crashed while shutting down with usb attached
10:59:33LinusNcrashed how?
11:00:20HClhanged on shutting down
11:00:35LinusNi use to differ between crash and hang
11:01:10 Nick Strath is now known as StrathAFK (
11:01:17LinusNhehe, i must admit i've never tried to shut down with usb attached
11:02:30 Nick QT_ is now known as QT (as@area51.users.madwifi)
11:02:44preglowi like the libmad guys selective use of tabs
11:03:13preglowshall i rectify it or let it be?
11:03:27LinusNi dislike all use of tabs :-)
11:03:40LinusNi think we should change as little as possible
11:03:57preglowi think it's neat if done correctly, but partial use is silly
11:03:57LinusNmakes it easier to keep in sync with the original
11:03:59***Saving seen data "./dancer.seen"
11:08:41preglowbut ok, i'm testing to see if synth.c opt breaks anything
11:08:50preglowif not, i'll stuff it in
11:09:03amiconnBagder et al: The "file type string buffer empty" isn't really a bug, it's just a bit misleading message. It should rather say "filetype string buffer full"
11:09:33amiconnThis is because of the supported file types increased so much that the buffer storing all the plugin names and icons is too small
11:10:14amiconnapps/filetypes.c line 49: #define STRING_BUFFER_SIZE 256
11:10:53preglowLinusN: should i let all .idata attributes be in there as well?
11:11:22Rickam I to assume correctly that to add a new plugin for building, just add it to the 'SOURCES' file?
11:11:22HClmarkun / amiconn.. whats the current idea about grayscale?
11:11:28LinusNpreglow: do so
11:11:41LinusNRick: yup
11:12:18amiconnBagder: That's probably also the cause for .cgb not working...
11:12:26 Quit ashridah ("Leaving")
11:13:05 Join ashridah [0] (
11:14:42BagderI renamed it now
11:14:46 Join Aison [0] (
11:15:53amiconnmarkun: The settings didn't get wiped for a strange reason; they were reset because the settings structure changed, and the settings version bumped because of that.
11:16:12markunamiconn: ok
11:18:40markunHCl: I will not be able to work on rockbox for the weekend. You can tell me if there were any ideas on grayscale on monday.
11:18:50preglowcvs diff insists that a couple of identical lines have changed
11:19:38Bagderwhitespace perhaps
11:19:48preglowsure as hell can't find it
11:20:14preglowand this is a part of the file i'm pretty sure i haven't touched
11:20:51preglowcan you see anything different in the first lines?
11:22:17markunI'm off, have a nice weekend guys!
11:23:22 Quit markun ()
11:23:30Rickgood night
11:23:43preglowgood morning
11:26:41LinusNpreglow: cr/lf?
11:26:46 Quit Aison (Read error: 104 (Connection reset by peer))
11:26:51amiconnLinusN: Do you have an explanation why your added .S support in the codec Makefiles "magically" fixed the heapload of inline warnings for libFLAC?
11:26:55preglowLinusN: can't find a cr
11:27:13LinusNpreglow: not in the diff
11:27:28 Join Aison [0] (
11:27:43LinusNperhaps the diff output doesn't contain the cr/lf differences
11:28:04amiconnpreglow: That sometimes happens here too. Just open a source file, save it, then diff. The diff shows some lines as changed that actually arent.
11:28:41preglowLinusN: no dice, i reinserted lf's, and still shows diff, i'll just ignore it
11:29:58LinusNamiconn: i fixed those warnings long before the .S fix
11:30:27amiconnNot all of them, and only for gcc 3.4+
11:30:44amiconnI still got loads of them when building the iriver sims on cygwin
11:30:49amiconn(gcc 3.3.3)
11:31:03amiconnThey disappeared the same moment you added .S support
11:31:19 Join Camilo [0] (
11:31:25 Quit Camilo (Client Quit)
11:31:39amiconnThis was also observable in the build table
11:34:19 Join ze__ [0] (
11:34:48*HCl stretches
11:35:10HClamiconn: would it be a bad thing if i added grayscale support to rockboy in cvs, disabled like how dynarec is disabled?
11:37:32preglowdid i commit
11:37:37preglowshould nok break sims this time
11:37:45preglows/did i/i did/
11:41:03*HCl stares at the void of response and guesses he'll just go and add it..
11:41:37LinusNadd it
11:41:46LinusNor is there a problem adding it?
11:43:28HClnot really
11:43:35HClaside from that markuns grayscale isn't in cvs
11:43:39HClso it can't work with cvs
11:43:42preglowlibmad is actually 103% realtime with only the changes i added now
11:43:48HClbut it will when markun's patches are applied
11:43:58 Join stripwax [0] (
11:44:01HClbut it'll be deactivated by a define in cvs
11:45:21LinusNHCl: go ahead, you're the rockboy boss
11:45:31LinusNstripwax: yo
11:45:53LinusNwhat are markuns patches?
11:45:59stripwaxhow's iriver audio looking (sounding)?
11:46:08 Quit ze (Read error: 110 (Connection timed out))
11:46:08 Nick ze__ is now known as ze (
11:46:13LinusNstripwax: i haven't had time to work on it
11:47:56LinusNpreglow: so what's our next move on libmad?
11:48:16preglowLinusN: santize the imdct opt i made
11:48:33preglowLinusN: and then i plan to make version of synth_full that doesn't use OPT_SSO
11:49:08stripwaxis it possible to trace/profile the code using bdm?
11:49:31LinusNstripwax: yes, but not with gdb
11:49:48LinusNpreglow: 95% with my test file
11:50:59preglowLinusN: what bitrate is it?
11:51:07LinusNflac is 104% with my test file
11:52:38preglowdoes ata driver use pio mode?
11:53:07preglowlibflac is 170% realtime with my file
11:53:24preglowat 120mhz
11:53:34LinusN120MHz here too
11:53:35preglowdrops to about 150
11:53:40preglowbut stays there
11:53:43preglowLinusN: how did you encode it?
11:53:59LinusNi got it from someone in this channel
11:54:20HClLinusN: um, i have them.. shall i put them on ftp? as far as i know, they pretty much enable native 2bit grayscale for the iriver
11:54:33preglowweird, even with a −−best encode, i end up at around 145%
11:54:35HCland just switch rockbox's format to 4 pixels / byte
11:55:02HClbut its prolly best to wait for markun, he's cleaned them up and stuff
11:55:03LinusNHCl: let me see
11:55:03preglowi need some food
11:55:31HClhold on.. can you accept dcc..?
11:55:48LinusNpreglow: what's opt_sso?
11:55:53LinusNHCl: no
11:56:12HClhold on
11:58:42preglowLinusN: optimisation that lets you do subband synthesis with a 32 bit accumulator, i used that at the time, since i wasn't quite certain how emac fractional mode would work
11:59:19preglowLinusN: it's enabled if you have no architecture specific MAC macros defined, which we do not have
11:59:31LinusNand how would non-sso perform?
11:59:40preglowgood, once i code support for it
11:59:55preglowequally good
11:59:56LinusNcompared to the sso version?
11:59:58preglowno difference for us
12:00:06preglowjust get more precision
12:00:26LinusNwe will probably have to find ways to trade quality for performance
12:00:42preglowi might actually want to postpone makeing it use no SSO_OPT, it's not going to make a lot of difference
12:00:54preglowand spend time doing more vital stuff
12:01:44*HCl commits grayscale
12:01:47HClfor rockboy
12:01:50HClnot for rockbox
12:04:16preglowbut i'll go generate some food
12:05:42LinusNpreglow: btw, my test file is 320kbit/s
12:06:15LinusNiRiver, Catch the digital flow!.mp3
12:06:46preglowLinusN: that's bound to do some more damage than my 192 avg. vbr
12:07:56preglowhmm, my sim build has invalid plugins
12:08:04preglowfailed to load /.rockbox/rocks/cube.rock
12:08:05preglowdlopen(archos/.rockbox/rocks/cube.rock): archos/.rockbox/rocks/cube.rock: invalid ELF header
12:08:08preglowSegmentation fault
12:08:50 Join jyp [0] (
12:10:02*HCl builds a grayscale rockbox/rockboy from cvs...
12:11:16LinusNjyp: found your next project: mt-500/index-en.html">
12:11:58jyphehe ;)
12:12:06jypwe'll see in due time ;)
12:12:10HCl :P
12:14:50HClfor anyone interested; -> grayscale rockbox/rockboy
12:16:39HClits nice that grayscale doesn't seem to slow it down much
12:18:02 Quit [FlaT]Heidel (Read error: 110 (Connection timed out))
12:19:12LinusNHCl: look cool
12:19:30LinusNi can't even make it past the first brick wall in marioland :-(
12:19:38HClaw, why not? o.o
12:19:39*LinusN is a gameboy loser
12:19:51HCl :p
12:19:59LinusNi can't make him jump high/far enough
12:20:01HClthere are savestates now
12:20:05HClyou can save right before the jump
12:20:09HCland try again and again and again
12:20:10HCltill you got it :P
12:20:14preglowyou need to run there, don't you?
12:20:21HClyou need to hold the play button
12:20:29HClthe longer you hold it the higher you jump
12:20:38ashridahyeah, holding the jump button jumps higher
12:21:37LinusNholding! aha!
12:21:51LinusNi was stuck in my giana sisters mode
12:22:05LinusNwhere you ran a few more steps to jump higher
12:22:37LinusNgrayscale lowers the contrast, doesn't it?
12:23:05HCli dunno
12:23:31HClthis is clean cvs + grayscale.tgz from my ftp
12:23:33LinusNmy screen got darker with your version
12:23:50HClmarkun added the background in the normal rockbox
12:23:54HClto be one of the grayscales
12:24:23HCli'm not sure, to test, i guess.
12:24:27HCli prefer it to be white
12:24:34HClbut i'm not exactly sure where in the code its located
12:25:25LinusNlunch time
12:26:52HClthe save state stuff of pabs has some bugs.. but i'll fix that later.. i really want to get the lcd modes to work..
12:29:08El_Barto2anyone know why i have a cable with videoutput with my gmini 120?
12:29:20ashridahHCl: need to be able to toggle certain buttons, to get 'run'
12:29:31HClashridah: mm?
12:29:37HClwhat do you mean?
12:30:06HCli know there's button work to be done.. but i really want to get lcd to work first..
12:30:41ashridahHCl: i'm just saying, pressing a button constantly is a bit of a bitch. need to have it 'on + press == off for a tic then on again)
12:30:50 Join ripnetuk [0] (
12:31:01HClhuh? i don't understand what you mean :/
12:31:02ripnetukDid someone say giana sisters :) one of the best games of all time
12:31:54ashridahHCl: i'm basically asking for a way to set a button to be held on without actually holding it :)
12:32:12HClyea, shouldn't be much of a problem to add that
12:32:19HCli'm planning to use remote buttons for functions like that
12:32:49HClfor now, i want to add more ways to deal with the 16 dropped lines
12:32:53HClsince they're really annoying me with mario
12:32:59HCl - i keep dropping in holes i can't see :3
12:34:18 Join [FlaT]Heidel [0] (
12:35:06ripnetukhcl - could we maybe rotate the screen 90 degrees, then you lose left and right margin, not top or bottom? or is the aspect ratio on the iriver not what it looks like?
12:35:52ripnetukmight make mario and tetris playable
12:35:57HClthen you'd have too many vertical lines
12:36:17ripnetukyeah, its a compromise
12:36:27HCland you'd lose not 16, but 32 lines
12:36:29HClon the side
12:36:38*ripnetuk tries newest cvs
12:36:48HClget from my ftp
12:36:56HClits pretty much state of the art as it is at the moment
12:37:13ripnetukgrayrockbox - i thought the grayscale stuff was in cvs now...
12:37:29HClonly for rockboy
12:37:32HClnot for rockbox
12:37:39HCland rockboy can't do grayscale while rockbox doesn't support it
12:37:59ripnetuki see... can u remind me the ftp addy (yet) again please ;)
12:38:13HClyou can also get grayscale.tgz with current cvs
12:38:25HCland uncomment the grayscale define in rockboy's makefile
12:38:39ripnetuki think i will just grab your binaries for now - thanks HCl
12:39:20DeadManis any work being done on LCD remote?
12:39:31*ripnetuk loves scp - my work has blocked ftp, so I have to ftp to home server, and scp to work :)
12:39:41HClhehe :)
12:40:02DeadManThe remote LCD could be interesting
12:40:11DeadManfitting in info
12:40:13ripnetukDeadMAn - until we have a proper driver for it, I reckon we should at least scroll the rockbox logo across it :)
12:40:32DeadManas long as I can see what I am ding :)
12:40:33ripnetukhcl - does grayrockbox need you to manually boost the cpu?
12:40:46HClit automatically boosts to 120mhz
12:40:56HClyea, linus increased it
12:41:06Lynx_awyis the temperature issue solved?
12:41:07ripnetukwoah! i missed that... what happened?
12:41:18HClhe didn't seem to have any trouble at 120mhz
12:41:20BagderLynx_awy: no
12:41:31ripnetukhow did he solve the crashing, presumed overheating?
12:41:39DeadManmight be worth making sure apps only use that as a max
12:41:50HClwell, that happens at 140mhz, not 120mhz, as far as i know
12:41:52Bagderripnetuk: he didn't, that was at 140
12:42:17ripnetukoh yeah <−−−− me->memory != good
12:42:17DeadManI think iRiver use 96 as a cautious measure and to cut down on battery use?
12:42:19 Join methangas [0] (
12:42:38ripnetukim guessing iriver use the lowest they can with their mp3 decoder still working
12:42:50DeadManthat would make more sense
12:42:53ripnetukso is rockbox real time now then?
12:42:53HCli'm not sure what to think of 120mhz either, i'm still gonna work on dynarec to cut it down more.
12:43:02HClyea, pretty much
12:43:03 Nick Lynx_awy is now known as Lynx_ (HydraIRC@
12:43:13*ripnetuk likes
12:43:15HClit didn't speed up as much as i hoped, but the mario seconds seem to roughly be a second
12:43:46DeadMan104% realtime libmad. Could do with a lot lower if possible
12:44:05*HCl hears his belly rumble, goes to search for food
12:44:32DeadManmore realtime that is. what do you think you could ramp it up to?
12:44:36ripnetukHCl - you kudos ! that is fantastic
12:45:09HClhehe :p
12:45:12HClkudos? :P
12:46:04DeadManif iRiver DID use 96Mhz for MP3 playback as the minimum they could clock the cpu down to maybe that is what to aim for with libmad optimization?
12:46:33DeadManI dunno I'm just guessing :)
12:46:58Lynx_could libmad ever be optimized as good as a decoder written specifically for the processor?
12:47:14DeadManPreglow may know
12:47:15Bagder"ever" ? sure
12:47:25Bagdergiven infinite time and effort, everything is
12:47:26stripwaxDeadman - I'd agree, although tremor's likely to use more CPU right? if ogg decode runs realtime at 96Mhz for iriver, libmad should be faster
12:47:39Lynx_Bagder: if it's rewritten 90%? ;)
12:47:42DeadManthat is true
12:47:45stripwaxunless iRiver ramps CPU up for ogg decoding (? would explain shorter battery life)
12:48:04DeadManwould need to decompile the iRiver firmware to know ;)
12:48:25HCli think that eventually we'll wind up either rewriting 90% or looking at iriver firmware's codecs
12:48:27Bagderwhy bother about how they do it?
12:48:38HCleither way, i'm happy that i can at least play gameboy games now :P
12:49:02DeadManit's always handy to look at how others do it for pointers on beating it ;)
12:49:26BagderHCl: there's also the option of using a decoder with less quality and format-support ambitions than libmad
12:49:36DeadManPlaying games is the least concern to me but hey...
12:50:06BagderI think libmad's high ambition in playing with high quality is what gives us performance problems
12:50:11DeadManMP1 and 2 is not really needed
12:50:29DeadManI don't know many people who use such formats
12:51:34DeadManMP2 is only used for PCM MPEG1/2 movies
12:51:47DeadManNo idea what MP1 is used for
12:51:48stripwaxbagder - it should be possible to tune libmad to trade off quality vs performance (in fact LinusN was suggesting exactly that, but only in reference to one small part of libmad).
12:51:49Bagderand some sat radio stuff too I believe
12:52:00LinusNfrom what i have heard, iriver runs at 72MHz for mp3 and 96Mhz for ogg
12:52:21Bagderstripwax: of course it is possible, the question is just how much work it requires
12:52:34Bagderand how much libmad that is left when done ;-)
12:52:44DeadManI say get the API in place forst and optimize the decoder after that
12:52:57Bagdernooooo, let's have more games first! ;-P
12:53:10stripwaxBagder - I'd agree with DeadMan
12:53:12ripnetuki thought that mp3 was very specific about decoding (ie, the decoding is deterministic), but slack in the encoding (ie, the encoders can work how they like). How therefore can you sacrifice quality for speed? is it in the post-processing filtering?
12:53:15stripwaxLinusN - ah, thanks!
12:53:23DMJChow much sound stuff works?
12:53:49ripnetuki like the menus on rockboy
12:54:07stripwaxripnetuk - if you decode an mp3 using internally 16-bit data, the output will not be as good as if you use internally 32-bit data (for example). So it's the precision on the math within the mp3 decoder
12:54:08LinusNDMJC: nothing (in cvs)
12:54:20*DMJC cries
12:54:29DeadManto get the tradeoff in quality vs. decoding right would you not need to analyze the waveform on output?
12:54:48LinusNyou can derive it mathematically
12:54:57stripwaxripnetuk - there's *also* post processing filtering, I'm not sure what libmad does there (e.g. we could turn off dithering if it's built in)
12:54:57ripnetuki see
12:55:15LinusNbut you'd have to listen to observe the effects of course
12:55:26DeadManI can never hear the diff between 16bit,24bit and 32bit anyhow ;)
12:56:01stripwaxDeadMan - I'm not talking about the final result, just the precision of all of the calculations used to *get* to the final result
12:56:08preglowyou can simplify things quite a bit before resolution drops under 16 bits
12:57:53*HCl sighs.
12:58:09*DeadMan size
12:58:09ripnetukmy reference was (the bit under DEciding of MP3 audio)... i guess we are not going to be 'bitstream compliant' !
12:58:16HCli'd appreciate it if anyone would look at rockboy cvs, particularly everything concerning fb.mode, and tell me wth is wrong with it :/
12:58:23preglowripnetuk: well, yes, we probably will be
12:59:18stripwax(cool, they're playing an autechre track as background music on Channel4)
12:59:42preglowripnetuk: the wording is wrong, you don't have to get the EXACT same output, there are tolerance levels
12:59:44ripnetukdid anyone see those programs on c4 about camp xray?
12:59:49ashridahbahaha, the ending of super mario land is silly
12:59:50preglowripnetuk: libmad has an overview on its homepage
13:00:01ashridahthe princess HAD A DAMN SHIP ALREADY
13:00:07stripwaxashridah - chuckle!
13:00:12*ashridah shakes fist at absurdity
13:00:28ashridahi'm guessing she didn't want to break a nail or something
13:00:30ripnetukive never seen a game with a good ending
13:00:41stripwaxripnetuk - Chase HQ had a great ending
13:00:50preglowff8's ending wasn't half bad either
13:00:58ripnetukdid it? i never got very far on that... never played the ff games
13:01:12preglowi've played all of them :P
13:01:15pregloweven the nes ones
13:01:23stripwaxAny, back on topic, I thought 'bitstream compliance' was all about whether *any* decoder could decode the output of *any* encoder; and not that the decoded output would always be identical
13:01:26ripnetuki bought one for the psx, but never got very far
13:01:34preglowstripwax: correct
13:01:46preglowstripwax: but there are also compliance tests, your output has to be reasonably similar
13:02:01ripnetukso wikipedia is wrong? we ought to correct it
13:02:26preglowlibmad has a full overview of compliant codecs on its homepage
13:02:46stripwaxpreglow - oh, neat - i'll check that out
13:03:46stripwaxpreglow - cools
13:04:01***Saving seen data "./dancer.seen"
13:05:46stripwaxInteresting that "MAD(optimised for speed)" is still fully compliant
13:05:59preglowlibmad is fully compliant even in OPT_SPEED mode
13:06:55ripnetuklooks like mad is VERY compliant
13:06:57stripwaxwonder what iRiver's mp3 decoder is ...
13:07:33LinusNvery non-compliant
13:08:07stripwaxmmm, is the output of MAD always 24 bits? sorry if that's a dumb question - that compliance page doesn't mention any 16-bit output variant
13:08:19ripnetukLinus - is that allegeedly very non compliant, or case proved :)
13:08:36ripnetuki think it /sounds/ qquite good
13:08:47LinusNi have a whole bunch of files that it simply won't play
13:08:49preglowit doesn't decode layers 1 and 2
13:08:55preglowso not, it's not compliant
13:09:04 Join Chamois [0] (
13:09:07ripnetuki know it doesnt do 1 and 2 as it wont play streams ripped from freeview on my dvb card
13:09:09LinusNnot even bitstream compliant in that sense
13:09:38ripnetuki think we are being a little harsh on it - it is a mp3 player after all
13:09:43LinusNpreglow: how's synth_full() today?
13:09:50preglowLinusN: commited?
13:10:14preglowhaven't dont much after commiting
13:10:20LinusNi'm blind
13:10:33preglownow you tell me
13:10:36*HCl gets really frustrated >.<
13:10:40 Quit methangas (" HydraIRC -> <- The dawn of a new IRC era")
13:12:09*HCl sighs.
13:12:19*HCl needs help :/ anyone got some time and an iriver? :/
13:12:28stripwaxquote from MAD page- "In one circumstance, MAD is a limited accuracy ISO/IEC 11172-3 audio decoder. This is true when MAD is forced to use an approximation version of its fixed-point multiply routine, and is easily avoided." - is that an option that we could potentially enable to reduce cpu time?
13:12:37stripwaxHCI sure, what's up
13:13:11HClstripwax: i'm not able to get fb.mode to work... if you have a current cvs checkout, do grep -C 3 fb.mode * in the rockboy dir...
13:13:28HClfb.mode needs to be able to be set to either 0 or 1, selecting different drop-line modes
13:13:34stripwaxHCI - I've never yet successfully built rockbox .. this could be a long morning :-()
13:13:38stripwaxoops, was meant to be a :-)
13:13:40HClok :X
13:13:40preglowstripwax: no, we're using emac unit as far as we can, and that's full precision and fast
13:13:49HClif i set fb.mode to 1 instead of 0 in the vid_init section
13:13:50HClit works
13:13:57preglowwell, full precision it's not, but near enough
13:13:59HClbut i seem to be unable to set it to 1 anywhere else
13:14:07HClfor apparently no reason whatsoever
13:15:35HClat least i can see the floor temporarily
13:15:56HClif anyone wants a version of rockboy that drops the lower 16 lines rather than the upper 16, rockboy.rock on my ftp, its a non-grayscale version
13:16:25preglowLinusN: when i want to stuff a fucntion in .icode, do i have to add a separate identical declaration including the __attribute__ statement above the normal declaration?
13:16:44preglowLinusN: it doesn't seem to work if i just add __attribute__ to the one that's already there
13:17:03LinusNpreglow: exactly
13:17:16stripwax(checking out all of rockbox for first time .. woohoo!)
13:18:05preglowLinusN: any way to avoid that? it looks nasty :V
13:18:20 Join Patr3ck_ [0] (
13:18:28amiconnpreglow: Well, the __attribute__ needs to be added to the function *declaration*, not the function definition.
13:19:02preglowahh, yes, and the synth.c boys don't have declarations
13:19:12LinusNHCl: what's the problem with fb.mode?
13:19:27HCli wish i knew! :(
13:19:32HClits refusing to set it to 1
13:19:34HClin my button driver
13:19:48 Quit DMJC ("Leaving")
13:19:58HCland i have no idea why
13:21:01HClits supposed to be really simple code, but for some odd reason, its just not working
13:21:01 Quit Patr3ck (Read error: 60 (Operation timed out))
13:21:10 Join AC [0] (
13:23:30ACHC1: ...can't work without markuns patch -> where can i get the pacht?
13:24:53HCland you need to uncomment the grayscale define in rockboy's makefile
13:25:44LinusNHCl: when is vid_begin() called?
13:26:00HClLinusN: right before the emulation
13:26:06ACok thanks
13:26:31Chamoisi tyhink it's not normal that vorbis2wav or wv2wav have disapperaed in the viewer
13:26:52LinusNHCl: it's called repeatedly during the emulation
13:27:07LinusNsee emu_run()
13:27:17Chamoistwo rockboy, two mpa2wav and two a52towav instead of them
13:27:25HClyou're right, i'm dumb, thanks :)
13:27:35HCllcd modes will be in cvs soon
13:28:43ripnetukdumb question - if I 'play' a mp3 at the moment, I would expect the mpa2wav plugin to convert it to wav, and dump it in the root - 1. do i get any feedback, like progress report, and 2. why isnt is working?
13:29:52LinusNmp3 playback is not handled by the viewer plugins
13:30:03LinusNyou'd have to do "open with"
13:30:19ripnetukah ha...
13:31:16HClthere, thats in cvs
13:31:19HClnow to work on more modes
13:31:19 Join DMJC [0] (
13:31:44Chamoiswhat are the lcd modes ?
13:34:01LinusNHCl: how about losing every 8 pixel line?
13:34:11HClLinusN: planned, planned.
13:34:16HClworking on it :)
13:34:39HClChamois: hold button selects top/bottom row drop
13:34:59preglowwhy does daily build generator report no changes?
13:35:05preglowthere clearly was, or it wouldn't compile
13:36:17Bagderbecause it hasn't kept up with my new way of doing the builds
13:37:31 Quit lostlogic_ (Client Quit)
13:37:32LinusNwhere can i find the iriverify plugin patch?
13:40:48ACLinusN: what should this patch do?
13:41:18LinusNconvert rockbox generated playlists to a format that the iriver firmware understands
13:43:28ACLinusN: what is the current stuff you are working on?
13:45:41LinusNwork :-)
13:45:57HClLinusN: wouldn't losing every 8th line drop 3 lines too many?
13:46:02HCli mean, 2
13:46:12HClshouldn't i drop every 9th line instead?
13:46:17 Join ghode|afk [0] (
13:47:45LinusNhow high is the original screen? 144?
13:47:47DeadManthey are working on games :P
13:47:52HCllet me commit this
13:47:54HCli must say
13:47:56ghode|afkHCl: does incude the hold button thingie?
13:47:58HClthe mode i just added
13:48:01HClworks really well
13:48:03HClghode|afk: no
13:48:37DeadManBlame HCI for pestering the other devs ;)
13:48:44ripnetukHCl - did u say I can get a grayscale rockbox by altering a define in the makefile?
13:48:56HClripnetuk: yes, you need the patch from my ftp, grayscale.tgz
13:49:04HCland uncomment the -DGRAYSCALE
13:49:07HClin the rockboy make file
13:49:11ripnetuki might wait until it makes cvs
13:50:01LinusNHCl: every 9th line sounds reasonable
13:50:21ACgrayscale should get it into cvs..
13:50:54jyp_sound_ should get into cvs ... :P
13:51:12ghode|afkjyp: agree ;p
13:51:18ACwav playback seems to work :)
13:51:24 Quit fuzzie (Read error: 60 (Operation timed out))
13:52:03ripnetukwhats the holdup on grayscale? its seems to work really nicely on that build I got from HCl
13:52:36DeadManBy the power of grayscale! I have the power!
13:52:37HClripnetuk: we're not agreeing yet on the internal rockbox framebuffer format
13:53:41DeadManI just bashed my humourous and I cried instead of laughed
13:54:27DeadManI need to get a file to this desks metal runners. They are too sharp.
13:56:36 Join cYmen [0] (
13:57:34 Quit ghode|afk ()
14:00:54HClxD ok, i think my new mode is rather bugged xD
14:00:59stripwaxDeadMan - chuckles
14:01:09ACHC1: How should look the internal framebuffer look like?
14:01:39 Quit Schee (Read error: 110 (Connection timed out))
14:02:00HClAC: 4 bits per pixel, for rockboy grayscale
14:02:13 Join Schee [0] (
14:02:43LinusN4 bits per pixel? why not 2?
14:03:27HCl4 pixels per byyytte
14:03:35HCli'm obviously not awake yet
14:03:54HCland 2 bits per pixel
14:05:49ACis markun working on it?
14:08:52Chamoiswhy in the "Open with " menu is there two times rockboy, map2wav a52towav ?
14:08:58 Quit mst (Read error: 104 (Connection reset by peer))
14:09:13BagderChamois: we await your bug fix ;-)
14:11:55HClhm, dropping a line every 9 lines proves more diffult than i had hoped
14:13:57amiconnDeadMan: MP2 support *is* needed imho. DAB radio is just one example that uses it.
14:14:54HClannoying scanlines..
14:15:02amiconnAnyway, Stephane Tavenard's decoder supports all 3 layers as well, and it should be faster than libmad...
14:15:18amiconnThe only thing it doesn't support corectly is 8 kHz iirc
14:18:49preglowamiconn: url, please
14:19:50DeadManwhatever works :)
14:20:52DeadManthat motto works for ms
14:20:53Bagderclaims to play fine at 90MHz coldfire
14:21:11preglowunless i remember incorrectly, wasn't that pure asm?
14:21:55preglowhaving a portable lib for other ports of rockbox would be better
14:22:00preglowbut performance is important as well
14:22:26LinusNwe can at least get some ideas from it
14:22:34Bagderit is not pure asm in the package at least
14:22:54DeadMan90Mhz might be a bit too slow where battery life is concerned
14:23:11DeadManLinus mentioned 72Mhz for MP3 playback
14:23:19preglowit'll probably be realtime for us, after stuffing it in iram
14:23:20DeadManon iRiver's own decoder
14:23:42Bagderpeople should stop caring so much about what iriver does or doesn't
14:24:01ripnetukits nice to know that it is possible though
14:24:22Bagderyes, but we don't know how much optmizes or asm that take
14:24:23DeadManit's more about worrying about battery life than how iRiver does such and such :)
14:24:41BagderI say we play music first, then worry about batteries
14:24:55DeadManI say you are right
14:25:07DeadManoptimize later. just get it up and running for now
14:25:28ripnetukis the processor usage really significant given that the hdd EATS power like nothing else?
14:25:31preglowthe epitome of elegance
14:25:39preglowit requires a 4 kb lookup table for pows...
14:25:44 Quit ashridah ("Leaving")
14:26:04Bagderripnetuk: well, with 30MB of ram it might be 30 minutes between HD spins
14:26:24Bagderduring 30 mins, I guess 20MHz makes a difference
14:26:27ripnetuki guess...
14:26:47LinusNit doesn't use the EMAC!
14:27:04Bagderthat is good news then
14:27:09ripnetukin fact, is it even worth loading that much in at once? i usually only listen to the thing for 30 mins at a time... there must be a point where its not economical to load in 30megs just in case... also, I very rearely listen through, i am forever swapping tracks
14:27:38preglownot only doesn't it do that, it uses coldfire incompatible assembly
14:27:41Bagderripnetuk: yes, it might be worth doing some statistics to figure that out
14:27:53ripnetukaltough ususally i press next on shuffle, so chances are the track I want is somewhere in the 30 megs
14:28:12preglowit should preload shuffled tracks as well
14:28:21LinusNripnetuk: it will *always* be next
14:28:24ripnetukpreglow - im assuming it will
14:28:39preglowi am assuming it already does
14:28:41LinusNrockbox exists today, remember that
14:28:46ripnetuki wonder how iriver do it (i know, i know!)
14:29:03preglowripnetuk: they don't, their hard disk handling is horrible, don't go there
14:29:09stripwaxthey don't afaict. they just load as much of the current track
14:29:12HCli'm not having much luck dropping every 9 lines...
14:29:14 Join shatting [0] (
14:29:48ripnetukdammit there are a lot of gamebopy roms
14:30:48DeadMando you have to adjust each game to work?
14:31:13ripnetukbut you have to scp them to work to get them onto your iriverv :)
14:31:19LinusNpreglow: incompatible assembly?
14:31:27preglowLinusN: it think it uses 68k assembly
14:31:42LinusNit is supposedly ported to coldfire
14:31:47stripwaxHCI - gameboy is just a Z80?
14:32:07preglowLinusN: translating 68k code to coldfire code is on the todo list in readme
14:32:36Bagderit says it builds for coldfire in the readme
14:32:39LinusNpreglow: but they claim to have run it
14:32:44preglowsure that's just not with c routines?
14:33:07preglowi dunno
14:33:14preglowsomeone should try building it :)
14:33:14LinusNpreglow: that's correct
14:33:16*preglow hides
14:33:19LinusNonly c
14:33:35stripwaxHCI was the core based on another gameboy emu or was it written from scratch?
14:33:42preglowstripwax: gnuboy
14:33:45LinusNleaves a lot of room for assembler opts then
14:33:50stripwaxpreglow cool thx
14:33:55preglowyes, it should be worth a go
14:34:48preglowreading other peoples poorly documented asm code is no fun
14:35:59preglowLinusN: how much iram can we count on having available in a codec? 32kb?
14:36:10LinusN32k seems reasonable
14:36:17LinusNi'd prefer less
14:36:33LinusNwe need iram for the filters too
14:36:45LinusNmixer etc
14:37:06preglowarhh, yes
14:37:33preglowthat's not very much
14:37:50LinusNwe must use the iram wisely
14:38:06LinusNand make good use of the cache instead of running code in iram
14:38:27preglowiram is almost half filled already
14:38:29preglowin mpa2wav
14:38:57preglowand that's just containing subband synthesis data and libmad stuff we probably need
14:41:44LinusNthe mpegdec code looks pretty straightforward
14:42:04preglowiram end is at 0x10016100
14:42:18preglowno code in iram
14:42:54LinusNmpegsub.c is pretty intimidating at first sight :-)
14:43:23preglowlet's play "spot the dct"
14:44:16Bagderthat's my favourite family game
14:44:26LinusNthe kids love it
14:45:22 Join Renko [0] (
14:47:52preglowdon't think we need the entire libmad structs in iram
14:48:01LinusNprobably not
14:48:03preglowbut putting just parts of it in iram is... tricky...
14:48:18 Quit shatting (" Try HydraIRC -> <-")
14:48:24LinusNwe would need some serious profiling to squeeze the last cycles out of it
14:48:44preglowand right now we can't profile
14:48:55 Join shatting [0] (
14:48:55LinusNi think we should give mpegdec a try
14:49:32LinusNi just wish some insane person could add it to the codecs dir and make it build
14:50:41 Quit Ka (
14:50:52*preglow hides
14:51:05NJoinKa [0] (
14:51:28preglowso, any takers?
14:52:01*LinusN is swamped with work
14:52:06ripnetukcant we write a simple profile tool? i wrote one in delphi for work, and it basically keeps a tree of function references, and you have to manually call startprofile and endprofile. It keeps track of the number of times it was called, and the total time in that function. The tree lets you know where the function was called from. Ive also written a graphical displayer for this data. This has helped loads in making our product fas
14:52:27ripnetukthe current 'leaf' of the tree is stored in a lifo stack
14:53:12preglowif you can think of a simple way to make it work in our case and are willing to do some work: go ahead :D
14:53:38ripnetuki might just have a go over the weekend
14:53:55stripwaxso each function would have to log its usage at the entry point and at all exit points? that doesn't sound ideal unless it could be automatic! Also is there a danger that it could significantly slow down any (small, very frequently used) functions?
14:54:06preglowmpegdec doesn't even bloody have compatible makefiles
14:54:15ripnetukyes, its a case of you cannot observe without changing what you observed
14:54:55preglowi know of another way, but will probably require disk writing from an interrupt
14:55:06preglowdon't know if that's possible
14:55:27ripnetukthe main problem is you have to manually make the start/end profile calls... very easy in Delphi with try.. finalyl blocks, but in pure c, every exit path has to be considered and a endprofiel put in
14:56:22ripnetukbut... at the end of the day it works, ive got massive speed savings with it
14:56:39stripwaxpreglow/linus - how fast can the cpu run when using the bdm interface?
14:56:50HClbeeping finally
14:56:51preglowstripwax: don't ask me :)
14:56:58HCljust when i told myself "if it doesn't work now, i'll give up"
14:57:01HClit works.
14:57:01stripwaxpreglow :-)
14:57:15ripnetukcan the gnu tools profiel automatically then? with gdb?
14:57:22stripwaxapparently gdb cannot
14:57:31preglowyou need gprof
14:57:34HCllcd mode 4 - dropping every 9th line
14:57:54ripnetukHCl - i thought you meant you has a audio beep finally :)
14:58:00HClno, heh.
14:58:27preglowbut someone code my idea, please. make a hi res timer that has the interrupt handler check it's own stack frame to see what code was executing when it was called, and log that somehow, then preprocess the data by using the map files to see what functions were used the most
14:58:44 Join muesli- [0] (
14:58:47 Join lolo-laptop [0] (~lostlogic@
14:59:02preglowit doesn't require any mods to current code
14:59:08stripwaxpostprocess you mean? or do you want the interrupt handler to actually do the map file lookup??
14:59:16preglowstripwax: yes, postprocess
14:59:33stripwaxcould we log over serial somehow?
15:00:05preglowstripwax: i have no idea, if interrupt handlers can write to disk without crashing everything, you can just accumulate some data and write it from time to time
15:01:03stripwaxi guess you could have a plugin that writes a buffer to disk - and an interrupt handler which fills that buffer?
15:01:18preglowbetter to just have the interrupt do it
15:02:18 Quit shatting (" Want to be different? HydraIRC -> <-")
15:02:30preglowif you have to load a plugin, you can't profile anything anyway
15:02:38preglowapart from the disk writing plugin, heh
15:03:05preglowand if the interrupt itself does the writing, you'll have no problems with the disk accesses being profiled as well
15:03:22preglowbut i don't know if this is feasible
15:03:43Bagderor, you disable that interrupt while saving
15:04:04***Saving seen data "./dancer.seen"
15:04:08preglowyes, that's a possibility
15:04:14preglowbut this is what i think sounds the simplest to do
15:04:25preglowand it's pretty close to what gprof does already, afaik
15:05:01 Join shatting [0] (
15:06:04stripwaxpreglow - :-s i didn't realise that plugins ran exclusively - oops. is there any chance that the timer interrupt could fire while the disk is already being accessed? would that break stuff?
15:06:10 Join webguest95 [0] (
15:06:23preglowstripwax: wooptidu, indeed, didn't think of that
15:06:23HClmario looks fat xD XD
15:06:23 Part webguest95
15:06:46preglowstripwax: but no, that's something we can fix
15:06:57preglowstripwax: it doesn't make sense to use disk accesses while profiling
15:07:17preglowstripwax: if i were to profile mpa2wav as it is now, the first thing i'd do is commecnt away all disk accesses
15:07:38stripwaxpreglow - mm, good point, i guess you only need to profile decoding in memory, not including loading from disk too
15:07:59preglowstripwax: that's most realistic as well, you'll allways do decode to memory when playing music ordinarily
15:08:02stripwaxi.e. loading the compressed audio, *then* begin profiling
15:08:05preglowalways, even
15:08:16stripwaxmakes sense
15:09:12preglowbut i'm not going to code this, i've got a staggering amount of school stuff to do
15:10:00preglowstaggering because i've neglected it to tinker with rockbox ;)
15:10:39 Join rasher [0] (
15:11:00stripwaxI have a full time job but I'm off sick at the mo :-( I could take a look but it'd be the first rockbox hacking I've done
15:11:20*HCl grumbles about how 0 % 9 is also 0
15:11:46preglowit should be that
15:11:54stripwaxHCI - urhh, what did you expect it to be?
15:12:04HCl0! but, i forgot about that :P
15:12:29preglowstripwax: i don't even know if we've got a high resolution timer yet, i think someone was working on it
15:13:03HClgoddd how i hate this >.<;
15:13:14 Quit edx (Read error: 110 (Connection timed out))
15:13:24HCli'm having trouble making a drop-every-9th-line algorhythm
15:13:56stripwax(y>>3)*9+(y%8 ) right?
15:14:52stripwaxoh right ,sorry.. misunderstood. (y/9)*8 + (y%9) . unless I misunderstood twice now?
15:15:14HClits not that simple.
15:15:14stripwaxgrr, meant %8 not %9
15:15:28stripwaxwhy not
15:15:30HClfirst i need to drop the proper lines using a modulo
15:15:36preglowgod, how i hate coldfire not having data cache
15:15:47*Bagder seconds that
15:15:47preglowwhat fool decided to neglect that
15:16:08Bagderwe should start an online petition for a public punishment of that person
15:16:16preglowoh, i agree
15:16:20HCland why the hellllll
15:16:28HCldo i have elton john - candle in the wind stuck in my head >.<
15:16:39preglowthey've quite obviously got the fast memory
15:16:45preglowthey could have used only 32kb of it as data cache
15:16:56preglowmore than enough, and still 64kb iram for other dubious uses
15:17:32preglowbut i need a breath of fresh air
15:17:49stripwaxHCI - well you drop every line where y%9 == 8, and then use (y/9)*8 + (y% 9) - and i did mean %9 not %8 after all
15:17:52HClsomeone else can have an attempt at making it drop every 9 line
15:18:23 Join geoff_o [0] (
15:18:27HCli hate it having scanlines.
15:18:36 Join edx [0] (
15:18:41stripwaxHCI - what's the actual problem here?
15:19:22geoff_oHey Stripwax.. nice to see you online.
15:19:28HCli don't seem to be able to make it drop every 9th scanline properly.
15:19:52HCli keep getting odd stuff happening
15:20:12Bagderif(y%9) { add line }
15:20:14HCland i'm tired of it, i've been at it for over an hour
15:20:23stripwaxHCI what are you currently doing to drop the scanlines? you need two parts
15:20:30stripwaxadd line needs to map to the line you want
15:20:47HCli'm just letting through a subsection of current scanlines.
15:20:50stripwaxso (if (y%9 != 8 ) { add line where line = (y/9)*8+(y%9) }
15:21:04HClyou can have a go at it if you want.
15:21:10stripwaxright but you also need to map it to where you want or else it'll just use the same value of 'y;', won't it?
15:21:12HCli'm gonna watch some tv and drink my tea ._.
15:21:14stripwaxhey geoff_o, what's up
15:21:21HClit caches every 4/8 scanlines
15:21:35HClits the caching that makes it annoying.
15:22:41geoff_oNot much.. you?
15:23:04HClaltering vid_update to do it would get really annoying.
15:23:19hileHi, I have translated rockbox to finnish, what should I do with the language file? E-mail somewhere?
15:23:42LinusNthe patch tracker
15:23:46*HCl updated his grayrockbox with the new videomodes, by the way
15:24:40LinusNHCl: where are you doing the line skipping?
15:24:52HClLinusN: lcd.c, lcd_refreshline
15:24:57HCland sys_rockbox.c
15:25:03HClfor mode 1 and mode 2
15:25:15HClin vid_update
15:25:26hiletnx - what about updates to the file, do the language files get updated like .po files with intltool-update (getting empty, changed entries there)
15:26:17 Join Quelsaruk [0] (~kvirc@
15:26:19hiletnx again ;)
15:26:42DeadManafternoob I liked better
15:26:42*LinusN is confused by the code in lcd_refreshline()
15:27:11HCli can explain?
15:28:05HClwhat don't you understand? :x
15:28:33HClthe first bit drops the calculation of scanlines that aren't needed for that particular mode
15:29:03LinusNR_LY is the scanline?
15:29:05HClscanline_ind is the position where the current scanline should go in the scanline buffer
15:29:13HClfrom 0-143
15:29:54LinusNso you want to add code that translates the scanline number
15:30:20HClnot only that, you need to make sure the scanline doesn't get into the buffer
15:30:32HCland that scanline 8 ends up in the position of scanline 7
15:30:38LinusNah, of course
15:30:59HClthat last bit is doable - i had a version in which scanline_ind wasn't defined as it is now
15:31:15HClbut it was defined as scanline_ind = (scanline_ind+1) % 8;
15:31:20muesli-l8er mates
15:31:34HClthat works. but. i had problems from that point forward.
15:31:46 Join XavierGr [0] (
15:31:57HClscanline_ind is the position of where the scanline will be put in the scanline buffer
15:32:10HClscanline buffer is either 4 or 8 scanlines big, depending on grayscale mode
15:33:20stripwaxHCI - scanline_ind = (scanlind_ind / 9) * 8 + (scanline_ind % 9 ) ... ?
15:33:43*DeadMan is still pondering afternoob's...or afterboon's....or afterboob's
15:33:45HClum. no. scanline_ind needs to be between 0-3 or 0-7
15:34:02stripwaxoh, that's the buffer? so there's TWO framebuffers, ugh
15:34:09rasheruh.. I managed to turn on my iRiver with the hold button on.. what gives?
15:34:16HClits the scanline cache
15:34:38HClsince rockbox refreshes 4 lines either way, its more efficient to only process them every 4 scanlines
15:34:49XavierGrHello just a quick question about iriver's screen. Why when scrolling in high speeds it gets all blurred? (Either text or graphics ex.Rockboy)
15:35:02stripwaxso you want what I said, mod 8, right?
15:35:04LinusNXavierGr: it's the lcd that is slow
15:35:21XavierGrSo there is nothing that can be done right?
15:35:30LinusNXavierGr: nope
15:35:45stripwaxI had a vague feeling the lcd controller supports scrolling in hardware?
15:36:04LinusNit does
15:36:11LinusNbut that won't help
15:36:31LinusNit's a chemical/electrical issue
15:36:33XavierGrIt's a shame that iriver stuck in there a bad LCD among other good quality hardware....
15:36:34stripwaxoh sorry, so the latency on the lcd itself
15:36:52stripwax(i should read stuff twice)
15:37:31shatting[15:35] <XavierGr> It's a shame that iriver stucker stuck in there a bad LCDd to turn on my iRiver with the hold button on.. what gives?
15:37:31shatting[15:33] <HCl> its the scanline cache
15:37:46shattingaccidential paste
15:38:21stripwaxHCI - so, does ((scanline_ind+1)%9)%8 work for you?
15:38:23Patr3ck_about profiling: i have implemented simple manual profiling, but can't get the hires timer to work properly
15:38:59stripwax(HCI - or %4 as appropriate)
15:39:22HClits not just that.
15:39:32HClits also the passing of the scanline to vid_update
15:39:34HCllike i said
15:39:38HCli'm not gonna work on it.
15:39:51HCltry getting rockbox to compile and feel free to try..
15:39:58LinusNPatr3ck_: what's the problem?
15:40:18Patr3ck_i implemented the timer for TIMER1 as you did for TIMER1
15:40:24Patr3ck_TIMER0 ;-)
15:40:30Patr3ck_but rockbox freezes
15:40:33Patr3ck_at boot
15:40:46LinusNi guess you shouldn't enable the interrupt
15:41:24Patr3ck_will my interrupt function be called anyway?
15:41:26stripwaxHCI - i might... I'm having trouble understanding how the scanline gets 'forgotten' on the way to vid_update - scanline buffer is 4 lines long and we already know what the top and bottom scanlines should be ..?
15:41:29LinusNtell me how you intend to use the timer
15:41:58Patr3ck_exactly the same as the tick_count but with higher resolution
15:42:07LinusNhow high?
15:42:20XavierGrHCl:Oh you may have notice that already but in rockboy while the A button can be pressed simultanesly with other buttons thats not the case for B button. So in mario I cant fire a fireball when running (or run) :P
15:42:42Patr3ck_not sure yet, but I tried using the same values as tick_count uses for testing and rockbox freezes when the splash screen is shown
15:42:50HClXavierGr: yes, i know, thats a hardware limitation.
15:42:59LinusNPatr3ck_: show me the code
15:43:18LinusNwhy do you want interrupts?
15:43:31LinusNjust to increase a counter?
15:43:51Patr3ck_can't show the code I am at work at the moment
15:44:02Patr3ck_but in the evening?
15:44:23LinusNmaybe, i don't know when i'll be back
15:45:08XavierGrHCl:bummer that would made difficult for some games to run. But I guess its an mp3 player right?
15:45:11Patr3ck_ok. next time I see you chatting when I am at home
15:46:26stripwaxwhich version of gcc should I use for rockbox dev? I'm looking at lassauge's page, is 3.3.5 still the 'recommended'?
15:46:50LinusN3.4.x is recommended for coldfire
15:47:43XavierGrAnd a last one, so in the current daily build the the CPU is runnig at 120mhz?
15:47:54stripwaxLinusN - thx. main reason for the recommendation?
15:48:33rasherhurray, I finished first level of Mario! :)
15:48:36LinusN3.3.x is sometimes hard to cross compile
15:49:25 Quit muesli- (Read error: 110 (Connection timed out))
15:49:37stripwaxLinusN - lassauge's instructions for 3.3.x look pretty easy
15:50:24LinusNstripwax: cygwin installation?
15:50:56LinusNhis coldfire compiler is built on binutils-2.15, doesn't work with rockbox
15:51:25stripwaxUGH. I followed his link because he says it does :-(
15:52:02LinusNthat was true in the dawn of iriver rockbox
15:52:11HClmario in a submarine o.o;
15:52:15HClworld 2 level 3
15:52:46stripwaxLinusN - ah, oh. what's the current state-of-play for cygwin builds? can it be done, and if so which binutils do I need?
15:53:12LinusNsee the documentation pages
15:53:46geoff_oLinusN: posted some iFP PCB pictures on the wiki.. enjoy.
15:54:28LinusNthe microcontroller will be the biggest problem, as the data sheets are secret
15:54:59stripwaxgeoff_o/LinusN - neat, that's the funny Phillips do-everything-one?
15:55:27*geoff_o nods
15:55:56stripwaxgeoff_o - any more patterns in the firmware?
15:56:37geoff_oIt's hard to tell what's on it, but there's at least an ARM processor core, a DSP, and maybe a memory controller and USB interface..
15:57:34HClyay, world 3 cleared
15:57:51 Join DrRick [0] (
15:58:16geoff_oNo new ones since the last email (I just woke up)..
15:59:23stripwaxLinusN - hm, I was but the cygwin setup doesn't seem to like lassauge's site .. no matter, I'll keep poking around
15:59:45geoff_oThe two things I want to look at are XOR-style algos that will turn those popular patterns into '0's, and maybe blindly guessing ASCII strings..
16:00:22LinusNgeoff_o: if it's an ARM, we can use a wiggler
16:00:48DeadManmicro controller?
16:00:51geoff_oIf that's a pun, I'm not following.. sorry. What's a wiggler in this context?
16:01:05LinusNa bdm/jtag emulator
16:01:34geoff_o"Please, tell me more" :)
16:02:08DeadManit's a connection you can wiggle about with ;)
16:02:30*preglow decides try to make mpegdec work...
16:02:42HClpokemon gold works fine too..
16:02:46LinusNit's a way of connecting the pc to the cpu, as a "backdoor" for debugging
16:02:46geoff_oDoes bdm/jtag have something to do with iRiver's playlist database format?
16:02:49*preglow also decides to stop leaving out words in his sentences
16:03:14geoff_oOh. Wouldn't that be nice?
16:03:41geoff_oDo you think there's a way of accessing it over USB, or would I need a hardware lab?
16:03:45*LinusN has a wiggler for the ARM
16:03:59LinusNpc parallel port
16:04:02geoff_oA friend of mine pointed out you could also sniff the data bus during a firmware upgrade.
16:04:33LinusNyes, using a logic analyzer
16:04:48geoff_o..but I don't have a logic analyser around the home. (And breaking into the local university is discouraged unless you're a card-carrying frosh.)
16:05:14*LinusN has a logic analyzer
16:05:18DeadManPreglow: it would be nice to have a top notch mp3 decoder working but right now I think just getting something running that is low on resource use is more important. Good luck with mpegdec anyhow :)
16:05:30geoff_oHow do you attach the wiggler to the chip?
16:05:53LinusNsoldering iron, steady hands and a lot of patience
16:05:57preglowDeadMan: well, that's the entire point of using mpegdec, is it not
16:06:09DeadManI like to state the obvious ;)
16:06:17DeadManMakes me feel better lol
16:06:33preglowwhatever floats your boat
16:06:57DeadManor beer
16:07:01DeadManeither will do
16:07:08preglowdon't mention beer in at least six more hours, please
16:07:12geoff_oLinusN: I'm affraid I don't have the skill/confidence/money product to risk such a procedure.
16:07:22DeadMandevs + beer =
16:07:35preglowmy concentration is hard enough to focus as is
16:07:37stripwaxpreglow - hangover?
16:07:44preglowstripwax: no, i just plain want beer ;)
16:08:00DeadManyou need caffiene ;)
16:08:01 Quit Sando (Read error: 54 (Connection reset by peer))
16:08:05preglowDeadMan: good point
16:08:21stripwaxwhat's everyone do when they're not rockboxing? i work (in finance)
16:08:33*DeadMan is not saying
16:08:46LinusNi'm an embedded systems developer :-)
16:08:49*AC studies
16:09:04preglowi'm "studying"
16:09:10DeadManinternational terrorist and psychotic megalomaniac
16:09:12stripwaxLinusN - I'd guessed :-) I used to be too
16:09:21stripwaxDeadMan/preglow - chuckles
16:09:46DeadManbut in my spare time I am a loafer
16:11:25 Quit XavierGr ()
16:11:53LinusNPatr3ck_: i have a running timer now
16:13:06geoff_oLinusN: does that mean you might be able to get a reference manual for the philips microcontroller?
16:13:08 Quit AC ("CGI:IRC (EOF)")
16:13:27LinusNgeoff_o: that means that i will have a hard time getting it
16:13:37 Join AC [0] (
16:13:40geoff_oOh well, it was a thought.
16:13:56Patr3ck_can I download the code somewhere?
16:13:56stripwaxDoes it have a hefty NDA attached, or is it just expensive?
16:13:57geoff_oI'm currently unemployed, but I'm a software developer by trade.
16:14:23DeadManthe micro controller is it on the iRiver?
16:14:33LinusNPatr3ck_: give me your email
16:14:37geoff_oI usually work on server-grade software, but I'm enjoying playing with embedded systems as well..
16:14:44LinusNno, ill put it on my web server
16:15:04stripwaxDeadMan- ?
16:15:14DeadManlost me...what micro controller? :)
16:15:30stripwaxDeadMan - the processor on the iFP-series devices. not iHP
16:15:32 Join R3nTiL [0] (~zorroz@
16:15:34LinusNPatr3ck_: how will you be using the timer?
16:15:49DeadManyou working on iFP too? heh
16:15:56LinusNjust read the counter as a timestamp?
16:16:04Patr3ck_LinusN: yes
16:16:08stripwaxDeadman - well... not a great deal yet!
16:16:15Patr3ck_LinusN: substract for time span
16:16:29 Quit DrRick ()
16:16:43LinusNwhich resolution would you like?
16:17:39Patr3ck_Well for precission as fast as possible, I could change that later, couldn't I
16:17:47*HCl made some more pics of rockboy running pokemon gold
16:18:45ripnetukHCl - is the sound for gnuboy emulated and discarded, or is it not emulated at all?
16:18:46DeadManPoke-A-Bug. Gotta catch em ya'll.
16:18:55LinusNPatr3ck_: of course, i was only concerned about the impact of the interrupt handler
16:18:56HClemulated and discarded
16:19:23Patr3ck_LinusN: Just use something a little faster than tick count, I will change it later
16:19:33DeadManI could have been a programmer. I instead decided to waste my time on the internet. ;)
16:19:59preglowi just became a programmer by default, it seems
16:20:09preglowcouldn't help myself
16:20:24ripnetuki am a programmer, when im not wasting time on the internet... wikipedia random page is a particually good way to waaste time
16:20:28Patr3ck_LinusN: I am very curious what you did to get it running, want to go home now :-)
16:20:33DeadManI used to make basic type apps for my BBS all those years ago. PPL anyone know that? Was basic for PC Board :)
16:21:03preglowi used to frequent bbses
16:21:07LinusNPatr3ck_: the interrupt controller is not the easiest thing to understand
16:21:17DeadManblueboxing was fun
16:21:20LinusNand the timer1 vector had to me moved
16:21:27preglowi'm still in love with ansis because of bsses
16:21:37DeadManI have flying ANSI
16:21:49preglowhaha, i've been coding an ansi viewer in my spare time
16:21:50HCl :)
16:21:52*geoff_o just submitted a support request to philips.. let's see what happens.
16:22:11Patr3ck_LinusN: I will diff with my original try, then I will learn something...
16:22:17DeadManpeople on 14.4's hated my BBS coz of the animated ANSI and huge ANSI intro screen lol
16:22:44DeadManI was l33t with dual 28.8's lol
16:23:49preglowhow can one not love this stuff,!.png
16:23:54preglowbut anyhoo, i need to hack myself a makefile
16:24:04DeadManI'd like to try that view Preglow and find an ANSI art stash someplace
16:24:30ripnetukpreglow - what on EARTH is that png :)
16:24:31preglowDeadMan: it's in, it needs SDL.dll as well
16:24:41preglowripnetuk: ansi art
16:24:47DeadMank ta
16:24:57ripnetuklooks very hires for ansi
16:25:02preglowripnetuk: it's not
16:25:17ripnetukahha its squashed
16:25:27preglowripnetuk: it's 8x16, not 9x16 per char
16:25:39preglowripnetuk: that's the only difference
16:26:21ripnetukthe best thing like that i saw was decss source code formatted to be the dvd logo ;)
16:26:46LinusNi hacked the "view i/o ports" debug menu to show the profile timer, and the timer doesn't start until you enter it
16:26:57LinusNjust to show you how it works
16:27:17Patr3ck_LinusN: Thanks very much!
16:28:14LinusNHCl: have a fresh for me to run on my train trip home from work?
16:28:36Patr3ck_LinusN: I will look at it this evening, integrate it with my profile code and create a patch that will add profiling
16:28:46HClLinusN: yea,
16:28:47DeadMancan that viewer be done in win32 rather than a command prompt Preglow?
16:28:51HClon my ftp
16:28:56LinusNi suggest you use the timer register itself instead of a counter
16:28:59DeadManor would it look crap?
16:29:01HClthe 8-8 mode is really nice to play mario
16:29:20LinusNHCl: where?
16:29:46HClturn hold on and off repeatedly to switch lcd modes
16:29:51Patr3ck_LinusN: The suggestion was for me?
16:29:51ripnetukhcl - does that one include grayscale AND layout options
16:30:12*HCl is gonna try to have another go at mode 3
16:30:17LinusNPatr3ck_: yes
16:30:26*ripnetuk wgets it (felling guilty for not using curl in present company)
16:30:38LinusNripnetuk: :-)
16:31:38*LinusN runs to catch the train
16:32:38LinusNcu guys
16:32:41 Part LinusN
16:32:48CoCoLUSis tetris working now? :)
16:32:54HClpretty much, yes.
16:33:01*ripnetuk suspects sweedish trains work much better than the ones in the uk
16:33:03preglowHCl: the new 8x top and bottom mode is very good
16:33:03HClyou just gotta switch it to mode 1 by turning hold on once
16:33:19preglowripnetuk: most trains work better than the uk ones
16:34:04ripnetuktheyve just taken 300 seats OUT of the trains i use meaning i have to stand up all the way home every day... twats
16:34:22preglowmost trains i've been on in uk have been pretty old and broken down
16:34:27stripwaxright, i'm off to play some HL2.
16:34:50stripwaxpreglow - only the shitty local trains. most mainline trains are pretty new and work fine (although the rails are often broken down :-(
16:35:53ripnetukthe menu on rockboy looks so nice :)
16:36:00ripnetukis it nicked from the mac?
16:36:01preglowright, i've mostly used local trains
16:36:59ripnetuki think they should concrete over the tracks and make a dedicated cycle route
16:37:02stripwaxpreglow - unlucky. they keep saying they're going to get around to replacing them at some time. and then end up pissing all of their budget into a research project which fails. leaving no new train designs, and no money to buy new trains of the previous design.
16:38:14preglowHCl: damn, wario looks strange in grayscale
16:38:42 Quit R3nTiL ()
16:40:36CoCoLUSrockboy looks nice now :)
16:41:08ripnetukthe lcd on the iriver give the exact same motion blur as the real original gameboy
16:41:14preglowdoes it?
16:43:01preglowcolours seem inversed some times
16:43:11 Quit dwihno (Read error: 110 (Connection timed out))
16:43:57 Join muesli- [0] (
16:45:18HClactually, i *might* have the colors slightly off, since i don't feed them through the (2 bit -> 2 bit) palette translation
16:47:15rasherpreglow: I think that's a general grayscale problem
16:47:28rashersometimes after exiting rockboy/jpeg viewer, the display will be messed up
16:49:06preglowHCl: if sprites have individual palettes, that is probably so, one of the people in zelda suddenly has a black face
16:49:30HClthey don't.
16:49:34HClas far as i know
16:49:37preglowdoes the lcd have hard coded palette, btw?
16:49:52preglowHCl: well, most of the game is still right, but some sprites are wrong
16:50:07preglowHCl: load zelda and see for yourself
16:50:19HClpreglow: can you see if you can find an emulator for windows that can do grayscale gameboy to compare?
16:52:01preglowHCl: sure
16:53:50preglowHCl: he's definitely white faced
16:54:09 Join dwihno [0] (~dw@
16:55:26HClpreglow: sure that he's white faced in gameboy grayscale and not just color?
16:55:35preglowHCl: seems like there is something going on with the sprites, i went in to the palette editor in the emu, and it lets me choose colours for exactly the sprites that were wrong in rockboy
16:55:46preglowHCl: well, everything but the sprites is correct
16:55:53HClwell, yes. i took the entire palette stuff out.
16:56:01HClcause i thought it wouldn't be needed for 2bit
17:00:13 Join XShocK [0] (
17:01:40HClanother failed attempt at dropping 9 lines
17:01:49HClpreglow: i'm not gonna look at sprite accuracy for now..
17:02:10 Quit Renko (Remote closed the connection)
17:02:27*HCl is still annoyed at being unable to get dropping every 9 line to work
17:04:01 Quit El_Barto2 ()
17:04:06***Saving seen data "./dancer.seen"
17:05:30 Join fuzzie [0] (
17:07:53preglowthis piece of shit uses fileio directly
17:09:12DeadManzOMG! None!
17:09:56 Quit muesli- (Read error: 110 (Connection timed out))
17:10:03DeadManOHNOES! and stuff. I dunno what it means but it sounds bad Preglow.
17:10:20DeadMancan you code around it? :)
17:12:21*HCl sucks at mario
17:12:24preglowof course, but i was hoping i wouldn't have to touch this filthy stuff
17:12:46preglowi'll do the only sensible thing: comment it out
17:13:11Bagderwell, I don't think we'll expect many updates anyway ;-)
17:13:30preglowit seems to be default access functions
17:13:35preglowi'll just replace them with stubs
17:13:49*HCl finally uses his iriver for what it was meant to do
17:14:36DeadManHCI how>
17:14:49HClhooking it up to usb and using winamp
17:15:08DeadMandon't tease
17:16:01*HCl is tired of the 9 line thing, tried like, 3 different approaches to it, all ending up in a distorted screen, and not distorted in the way it was supposed to be distorted
17:16:32DeadManmight be hardware limitation
17:16:38HCldefinately not.
17:17:56preglowand bam, it compiles
17:18:10HClnow just fill in the stubs.. and benchmark
17:18:15preglowwith more warnings than i can shake a stick at, but i am about to ignore that
17:18:57 Join webguest61 [0] (
17:19:14 Part webguest61
17:25:52 Quit ripnetuk ("Leaving")
17:30:50 Quit Nibbler (Read error: 110 (Connection timed out))
17:31:30 Join Nibbler [0] (
17:36:59 Nick kergoth`zzz is now known as kergoth (
17:37:16 Join Zavatta [0] (
17:37:32 Part geoff_o ("Kopete 0.9.1 :")
17:39:16 Quit XShocK ("CGI:IRC (EOF)")
17:47:50 Join muesli- [0] (
17:48:23 Join [IDC]Dragon [0] (
17:49:10stripwaxpreglow - ooh, nice. we're all waiting to see how it compares vs libmad
17:49:23[IDC]Dragonis there any C compiler option to allow struct reordering in order to minimize padding?
17:51:25 Quit Patr3ck_ ()
17:51:59 Join livesNbox [0] (
17:52:09 Join elinenbe_ [0] (~elinenbe_@
17:52:42livesNboxHey guys... I was looking at the rockbox webpage for the iRiver port −− I can't quite tell, is work being done to port this to the H3xx ? or currently is it only being worked on for the H1xx ?
17:54:48DeadManH1 for now
17:55:17DeadManbig enough job...but H3 should be quicker since most of what is in H1 can be used in H3
17:55:42DeadManI'm guessing
17:56:21muesli-hi DeadMan ;)
18:02:40livesNboxthat's cool −− any way I can help? I'm a programmer..
18:02:42 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
18:02:56 Quit Lynx_awy (" HydraIRC -> <- s0 d4Mn l33t |t'z 5c4rY!")
18:03:27livesNboxbut like most −− not a TON of free time :(
18:04:37DeadMangrab some code and get stuck in
18:05:06DeadMancheck the wiki for what's been done already and what needs doing
18:05:35*HCl sighs.
18:06:35livesNboxis it all in a CVS type of a thing or ?
18:07:42preglowlivesNbox: yes
18:07:56livesNboxok I'll take a look.
18:07:56 Quit Zavatta (Remote closed the connection)
18:08:00preglowlivesNbox: main focus at the moment: making a codec api and optimizing the codecs
18:08:28livesNboxunfortunately I don't have an H1xx −− so not sure how much help I'll be as of yet.
18:08:50preglowwe do have simulators, but depends on what you want to do
18:09:03livesNboxcool.. ok..
18:09:08livesNboxthanks again.
18:09:13livesNboxsee ya
18:10:00DeadManis anyone actually doing anything on the api?
18:13:18preglowi don't know
18:13:24preglowlinuxstb said he'd use time on it
18:13:27preglowbut haven't heard anything
18:13:56 Join ccc [0] (ghj@
18:14:25elinenbe_HCl: you such a pimp... pimping out people irivers with gnuboy.
18:15:18 Part ccc
18:17:34HCl :x
18:18:19 Nick coob is now known as davidc__[3] (
18:21:47 Quit muesli- (Read error: 110 (Connection timed out))
18:27:30HClpreglow: i can't really find the way to fix the palette
18:27:49HCli can't even tell whether its a gnuboy bug or not
18:27:59HClsince gnuboy does not seem to have an original gameboy mode
18:30:23HClit has a fixme that has to do with it, i think..
18:30:25HCl /* FIXME - handle directly without faking cgb */
18:35:07*HCl likes cvs
18:35:11HClrm *;cvs update
18:37:26 Nick davidc__[3] is now known as coob (
18:37:29 Quit crwl (Read error: 110 (Connection timed out))
18:37:58 Quit [IDC]Dragon ("CGI:IRC")
18:46:24HClpreglow: i'm not gonna be able to fix it unless someone manages to explain how the _scan functions work to me
18:47:45 Join Renko [0] (
18:50:16 Part stripwax
18:52:13 Quit shatting (" HydraIRC -> <- Leading Edge IRC")
18:52:40preglowHCl: well, i have no idea
18:55:38elinenbe_preglow: how far away is a codec api?
18:56:17preglowelinenbe_: i don't know, only thing i know has been done is discuss basic functionality
18:56:45elinenbe_ah... don't neglect the cross-fading ;)
18:56:54preglowwe probably wont
18:57:10elinenbe_I know we can't even decode mp3s in realtime, but other codecs that are fast enough can benefit from this.
18:57:26preglowwe should be able to decode realtime
18:57:32elinenbe_also −− just wait until amiconn gets deep into the mp3 codec with his asm magic!
18:57:51preglowamiconn doesn't even have a h1x0
18:58:07preglowi'm the one doing asm hacking on libmad right now
18:58:34preglowbut shop, brb
19:00:07rasherwhoa, I'm decoding in 176% with libmad?
19:00:53rasherah, haha
19:00:55rasherit's a mono file
19:01:13*fuzzie watches it be 22khz too
19:01:24rasherMPEG 1.0 layer III, 160 kbit/s, 44100 Hz mono
19:01:26fuzzieoh, that's quite impressive then
19:04:10***Saving seen data "./dancer.seen"
19:06:50 Join Stryke` [0] (
19:07:00rasherof course, I put the cpu at 120mhz
19:07:44 Join FiZZ [0] (
19:09:47rasheroh.. getting 112% on a 128 kbit/s file
19:10:22 Quit AC ("CGI:IRC (Ping timeout)")
19:10:32FiZZguess thats good news
19:13:26preglowrasher: it's very bitrate dependent, yes
19:14:48 Join crwl [0] (
19:19:41HClpreglow: hows that other decoder?
19:29:31 Join [1]Chamois [0] (
19:37:04preglowHCl: starting to work on it again now
19:42:50HCl :)
19:42:57*HCl made a seperate wiki page for rockboy..
19:44:51FiZZman as soon as someone gets the bootloader going on a H3xx i am so gonna hack into rockboy ;)
19:47:37*preglow cries loudly as he sees the number of errors
19:47:40 Quit Chamois (Read error: 110 (Connection timed out))
19:47:40 Nick [1]Chamois is now known as Chamois (
19:48:53preglowi love it when changes i make just vanish
19:51:34 Quit Renko ("cheerio")
19:54:18HClwhats the resolution of a h3xx?
19:56:38 Join michiel_ [0] (
19:56:49HClheh, the resolution of a h3xx is much better
19:56:50 Join Sucka [0] (
19:56:55FiZZor something
19:57:03*FiZZ checks
19:57:37coobsounds familiar
19:57:43coobipod photo is 220x176
19:58:24FiZZyeah the display is great
19:58:28FiZZits prolly the same *Shrugs*
19:58:35FiZZits still not clear what display it is
19:58:43FiZZits an epson/samsung combo though
20:02:32nozomiyumeit is - the H3xx is 220 x 176
20:02:52FiZZand some 16bits colour depth, supposedly
20:03:13nozomiyumeyeah - it's pretty gorgeous - I was really surprised how well pictures / videos look on it
20:03:18nozomiyumeit must be 16 bit
20:03:28Bagder"up to 260 thousand pixel colors"
20:03:35FiZZthats 18bits in fact
20:03:41nozomiyumehuhn o_O
20:03:45nozomiyumedidn't expect that
20:03:47coob24 bit - one channel for alpha?
20:04:04FiZZi donno
20:04:10FiZZi also donno where the video mem is
20:04:16coobor does that make no sense whatsoever
20:04:19nozomiyumehow far along is the bootloader on h3xx anyways?
20:04:19FiZZin the main sdram ?
20:04:30FiZZas none of the devs have a H3xx :p
20:04:38nozomiyumethat'll do it
20:04:55FiZZwell in the US its ridicolously cheap anyway
20:04:56Bagderor the opposite: no h3xx owner is a dev
20:05:20nozomiyumetrue that
20:05:22FiZZmushroom mushroom!
20:05:36nozomiyumewell, I own one, and I wanna work on it - just need bravery to crack the thing open
20:05:43nozomiyumeoh, and that whole "time" thing =P
20:05:47FiZZsorry, i think I heard that flash one or twohundred times too often
20:06:38FiZZthe h320 is like 250$ in the us now
20:06:44FiZZshouldnt be too hard to get a dev to get one
20:06:52preglowget me one :P
20:06:54FiZZon the other hand, that wiggler they use is hard to get in europe
20:07:17*FiZZ blames his thesis
20:07:21nozomiyumeplus you can only get it online, and I've not seen it for under 270 w/ ship
20:07:22preglowFiZZ: how did linus get one then?
20:07:24BagderFiZZ: we bought the wiggler from the US
20:07:33kergoththere are lots of wiggler compatible jtag cables around, wouldnt be hard to find a place thatd ship to europe. course, keeping the cost low, thats something else
20:07:34FiZZyeah but the taxes and stuff :p
20:07:42FiZZthe thing is
20:07:46FiZZin my uni we got lots of stuff
20:07:52Bagderwe have a bdm wiggler, not jtag
20:07:58Bagder"we" as in Linus ;-)
20:07:59FiZZand we already did some coldfire projects before
20:08:05kergothBagder: uh?
20:08:43FiZZhow much did u guys pay for the wiggler, all in all ?
20:08:54BagderI don't know
20:09:00FiZZshipped to where.... norway ?
20:09:02preglowthere's a link on the wiki
20:09:09FiZZclose enough
20:16:36 Join frank_ [0] (
20:17:00*rasher burns FiZZ slightly
20:17:22FiZZeeeek it burns, it burns
20:17:35FiZZat first I read "bums" haha
20:17:38FiZZwhat a relief
20:19:32*HCl plays mario...
20:22:29 Join Patr3ck [0] (
20:23:05preglowwhat the hell is up with the mem* routines in xxx2wav.h ?
20:23:11preglowthey bloody conflict with rockbox
20:23:40HClohno, game over
20:23:48DMJCman don't drink coke before bed
20:24:00DMJCthat stuff gave me whacked dreams
20:24:09FiZZI dig whack dreams ;)
20:24:30FiZZwow that sounded soo wrong.... -.-
20:30:08*DeadMan ponders the responsibility of giving a home to someone's dog who does not want it anymore
20:42:56preglowand why does everyone have to define their own BOOL?
20:43:47Bagderpreglow: because there was no standard one before C99 (or C++)
20:43:55elinenbe_the h3xx are cheaper than the h1xx!
20:44:35preglowBagder: why just not use a bloody int :///
20:44:53Bagdergood question ;-)
20:45:07Bagderpeople like adding new types
20:45:07preglowand what the hell am i supposed to do about the mem* functions
20:45:11preglowrockbox has them too, and they're clashing
20:45:13Bagderits a programmer's desease
20:45:41Bagderpreglow: I'd say you remove the new ones and use the Rockbox ones
20:45:41 Join XavierGr [0] (
20:46:16preglowBagder: plugins can't do that, can they?
20:46:30Bagderpreglow: they have them in the plugin API
20:46:38preglowi think i'll have remove a header reference to make this work
20:47:06XavierGrhello I've got another question. Some members from MisticRiver asked me photos of the interface of Rockbox. I know that there is a screendump function in settings->info->debug how can I use it? Thanks
20:47:34BagderXavierGr: enable it, then insert the USB cable to make the dumps
20:47:48Bagderthey are them stored as BMP pics on the harddrive
20:49:11 Quit elinenbe_ (" HydraIRC -> <- irc client ownage!")
20:50:15XavierGrbut I guess that I will have to press a button to take the screenshot I need right?
20:51:19 Quit livesNbox ()
20:51:21XavierGrwhich button do I need to press while enabled?
20:52:06Bagderno, you insert the USB cable
20:52:26Bagderduring screenshot mode the ordinary usb mode is disabled
20:55:29XavierGrforgive me but I didnt quite got it. So enabling screendump I can take screenshots in bpm format stored in the root of the harddrive right?
20:56:18XavierGrI have done what you say but I can take one screenshot only of the root folder, what if I want to take a screenshot of the settings or plugins?
20:56:40Bagdergo there, insert cable
20:57:03preglowBagder: is it even correct to have the codecs include standard header files? they can't use any of it anyway
20:57:08XavierGrlets see....
20:57:40Bagderpreglow: right, but strictly speaking the codecs will not be using the plugin API in the future, so it might be...
20:58:28preglowwell, they'll be using a plugin api of some sort, so it'll be just the same
20:59:56XavierGrok it worked thanks, though I cant understand why it is not implemented by pressing one button?
21:00:06preglowoh well, i'll just point the entire bunch to xxx2wav
21:00:24BagderXavierGr: because this way it is much less intrusive
21:00:42BagderXavierGr: the entire UI works like before, all buttons are used
21:00:53XavierGrI wont argue you know better! :p
21:03:48preglow#ifdef AMIGA
21:03:50preglowif love this codec
21:04:14***Saving seen data "./dancer.seen"
21:04:54 Nick FiZZ is now known as fizze`AwAy (
21:09:14 Join prethom [0] (
21:09:14 Quit preglow (Read error: 104 (Connection reset by peer))
21:09:22 Nick prethom is now known as preglow (
21:09:26preglowi love crashing servers
21:09:47 Join Camilo [0] (
21:12:00HClbah, i don't understand this. according to my test-program i made to test scanline-dropping algorhythms, my algorhythm should be absolutely fine
21:12:26*HCl tries to add it one last time
21:15:46 Quit XavierGr ()
21:16:14 Quit Ka (Read error: 110 (Connection timed out))
21:16:56Bonkersis rockboy correctly included in the zip now?
21:17:37 Join Ka [0] ( for loading onto the recorder
21:18:06BagderBonkers: I believe it is
21:21:02 Quit Patr3ck ()
21:21:04HClit should be
21:21:14HClhaven't tested it
21:21:15HClplease do
21:21:34HClthat reminds me, archos needs a button for the new lcd modes
21:23:23preglowwhat about stuffing it in the menu?
21:23:40Bonkersyep, zip all set
21:29:09 Quit elinenbe (" Like's GUI? Then try HydraIRC -> <-")
21:29:46*preglow summons linuxstb
21:31:15HClpreglow: archos does not have a menu
21:32:04preglowHCl: but it will, yes?
21:32:16HClwhen someone fixes it.
21:34:32 Quit Camilo ("Chatzilla 0.9.67 [Mozilla rv:1.8a6/20050111]")
21:36:26 Quit fizze`AwAy (Read error: 60 (Operation timed out))
21:47:58 Join XShocK [0] (
21:49:01michiel_Hi XShocK, you wrote the wav playback on the iRiver, didn't you?
21:51:52*HCl thinks he finally got the new lcd mode for rockboy to work
21:51:58HClthe annoying bit is.
21:52:05HClits pretty much the same algorhythm i used earlier.
21:52:22XShocKmichiel_: yes
21:53:02amiconnHCl: The button for the lcd modes on archos should be the on that is not yet used - BUTTON_ON
21:53:06michiel_XShocK: Kudos! So cool to hear rockbox play music
21:53:45XShocKit is not on cvs yet thou..
21:53:55michiel_any idea how much work it is going to be to play mp3's once libmad is > 100%
21:54:16XShocKneed to develop the sound framework.
21:54:37XShocKthat is it.. libmad is actually higher than 100% now
21:54:47michiel_probably quite a bit of work, yes?
21:54:57XShocKat 120mhz it is if i not mistaken 104%
21:55:21michiel_Super! That means it is possible now at least
21:55:36michiel_hope Tremor can be pushed to 100% soon as well :-D
21:56:20 Join Tang [0] (
21:56:33TangHelo :)
21:56:33michiel_wish i could code assembler
21:56:43TangHi Miciel
21:56:56Bagdermichiel_: most of this is plain C, not asm
21:57:23TangHi Badger and HCl
21:57:52michiel_afraid of getting my feet wet. Try to get up to speed again this weekend
21:58:11michiel_code Java, gotten fat and lazy
21:58:12TangChamois told me there were wonderfull progress in mp3 decoding and ome others
21:58:44michiel_progress has been amazing last few days, thanks guys!
21:58:53TangJust installed Rbx on my iHP
21:59:05Tangvery nice :)
21:59:07michiel_my iHP 120 has become my favorite toy again
21:59:16DMJCdidn't someone get flac decoding at 170% of realtime?
21:59:51Stryke`hope musepack is next for optimization
22:00:00michiel_oh well, I've got the simulator working
22:00:10Bagderdon't hope, join in and make it happen!
22:00:17michiel_may work on one or two of the plugins this weekend
22:01:01preglowi've got it running 170% PEAK on ONE file
22:01:15rasheris musepack even available?
22:01:34DMJCwhat's the ogg codec status?
22:01:34rasherAlso, why am I getting two of many plugins?
22:01:52Bagderrasher: you mean in the menu?
22:02:08rasherBagder: yup
22:02:13TangI don't know if it make sense
22:02:15Bagderits a bug
22:02:30Tangbut seems the iHP freeze
22:02:34Bagderpreviously there was a bug that made reach plugin only appear once in the viewers.config file
22:02:42DeadManWill 120Mhz eat more battery?
22:02:51Bagdernow they can appear more than once, for each extension they "own"
22:02:52Tangwith rbx 20050303 when i try to play an audio file
22:02:58Bagdermaking the menu fail
22:03:35DMJCugh, windows
22:03:38DeadManneed to get MP3 decoding lower then really if to make it playback at least as long as iRiver's firmware achieves no? :)
22:03:50DeadManI already went over that aspect though I think
22:04:16preglowthat sentence nearly made my brain explode
22:04:48DeadManI am so eloquent with words am I not? :)
22:05:23Bagdermusepack library source is here:
22:05:30preglowhaha, you are just fine as long as you make sense
22:05:43preglowthat last sentence fried my language center
22:06:02DeadManI thought it made sense. Less Mhz = longer battery
22:06:33DeadManwell iRiver don't run at 120Mhz they run Mp3 decoding a lot lower from what Linus was saying
22:06:54DeadManshrugs. just ignore me most people do
22:07:07Bagderthere are many more factors than just cpu speed
22:07:16DeadManof course :)
22:07:38Bagderlibmusepack is BSD
22:07:40DeadManjust being nitpicky like I always am ;)
22:07:40preglowHCl: woop, does it work well?
22:07:41Bagderoriginal BSD
22:07:47Bagdernot GPL compatible
22:07:47TanghUM CAN I ASK STHG,
22:07:55HClpreglow: yea
22:07:59Tang(sorry "caps lock")
22:08:02HCli'd like some archos people to try rockboy now
22:08:10HCli'm curious to how much is left of the screen, really
22:08:15HClor whether it works at all
22:08:44preglowTang: just ask
22:08:47TangHow many people actually work on the iRiverport?
22:08:52Tang(thanks preglow)
22:08:52preglowcan't say
22:08:58TangSo many?
22:09:12preglowwell, no, but you don't have to tell anyone that you're working on it
22:09:22TangAh okay
22:09:24preglowpeople just pop in with stuff they've done sometimes
22:09:40Tangokay it's very impressive
22:09:41preglowBagder: mail the guy and ask what the odds of a relicensing is?
22:09:54Bagderperhaps I should
22:10:05Tanghum the guy is Peter Pavlovski
22:10:29Tanghum in fact he 's somewhat famous for
22:10:35preglowyes, yes
22:10:39Tangquite special
22:10:46preglowone can still ask
22:10:46Tangnot very open minded
22:11:00Tangit's foobar coder
22:11:04amiconnHCl: What kind of sound does the real gameboy produce? Is it sample based, or just predefined beeps?
22:11:04preglowyes, i know
22:11:11preglowamiconn: synthesized
22:11:19preglowamiconn: and very basic sample support, afaik
22:11:27Tanganyway you can try
22:11:29HCleither way, gnuboy just gives me a pcm buffer it wants me to play
22:11:30preglowamiconn: square waves and noise, basically
22:12:01Tangmaybe i can also try with a french contact i know which is quite HA friendly
22:12:38preglowi think a single mail will do fine
22:12:41preglowdon't make an issue out of it
22:12:55TangGood luck :)
22:13:41preglowhe will probably get enough of it once people starts learning of the issue anywa
22:14:09Tangin fact i was hoping
22:14:20Tangthat all the noise about rockbox i did on HA
22:14:49Tangwill make Peter Pavlovski contact the team in a good way
22:15:34TangHum if there is no agreement option
22:15:51Tangmaybe you should try contactig Frank Klemm
22:15:53TangMPC author
22:16:06Tang(but not muselib coder...)
22:16:09preglowwhat good will that do?
22:16:15preglowit's libmusepack we need
22:16:31Tangdon't know maybe he participated to the project
22:16:38Tanganyway as far as i know
22:16:47preglowwe'll see, there's no rush
22:16:55rasherhe did
22:16:57preglowgod knows how many hours of work are left on the codecs we already have :/
22:17:05Tanghe had an agrgument with Roberto Amorim
22:17:24rasheraccording to the AUTHO
22:17:26Tangand then he doensn't come at HA anymore
22:17:32rasher.. RS file
22:18:09rasherfrom what I read on some forum, the picking of old-bsd might as well be an accident
22:18:20preglowargh, rockbox doesn't even have a bool, just a conflicting enum
22:18:36rasherif whoever picked it was the same as I saw calling lgpl restricting
22:18:52preglowrasher: mno, i think he has chosen it deliberately
22:19:07preglowbut there's always the possibility
22:19:21preglowwhat are the gripes with the old bsd license again?
22:19:36preglowdidn't that have an advertising clause of some kind?
22:19:40rasherit places additional restrictions on top of gpl
22:19:47rasheryes, indeed
22:19:57*preglow tosks
22:20:02Bagderadvertising clause yes
22:20:22Bagderwhy they made the Modified BSD
22:22:54 Join geoff_o [0] (
22:25:09*HCl prods people
22:25:24HClwho can test rockboy on archos recorder? or archos recorder sim?
22:26:56rasherI guess I could try it on a sim
22:28:37preglowBagder: excellent
22:28:52 Quit DMJC ("Leaving")
22:28:54rasherhow would a hand-crafted license work?
22:29:00Bagderno idea
22:29:08Bagderjust wanted to open up for discussion
22:29:46Bagderwe can in fact add an exception to the GPL
22:29:48coobi swear the advertising clause means you only have to leave the copyright info in the headers?
22:29:54Bagderto allow libmusepack
22:29:57coobor was that some other bsd license
22:30:04Bagdermany projects to that for openssl
22:30:04rasherif they're insisting on not being gpl-compatible, any hand-crafted license would require a rockbox license-change/addition
22:30:10rashercoob: that's the new bsd
22:30:22rasherand indeed the gpl
22:30:46rasherthe old bsd license requires you to reproduce the copyright info in some way for binary distributions
22:30:55rasherlike printing it when you run the code or something
22:31:26coobdoes having it as an unused variable that can be found with a hexeditor count
22:31:33coobor strings bin
22:31:50TangMp3 decoding at 104% realtime and FLAC decoding at 170%. Both at 120MHz.
22:31:56Tang(read on MR)
22:32:12Tangthis mean the past bug at hi freq is away
22:32:28Tang(HD freezing i read)
22:33:26rashercoob: no :)
22:33:55BagderI am a very silly person
22:34:03preglowBagder: are you?
22:34:03rasherHm actually it says to reproduce it in the documentation
22:34:06Bagderthis is the Modified one
22:34:16preglowBagder: then hooray!
22:34:20*Bagder goes to the corner
22:34:27Stryke`success so quickly, ;-)
22:34:28rasherhaha, excellent
22:34:38preglowthese are fast moving times
22:34:51preglowissue resolved a mere five minutes after posting for help
22:35:21TangSo bye all
22:35:24 Join ghode|afk [0] (
22:35:24preglowTang: bye
22:35:28Tangi go good continuation :)
22:35:37Tangbye preglw :)
22:35:41Tangcheers all :)
22:35:43 Quit Tang ("Chatzilla 0.9.66 [Mozilla rv:1.7.5/20041108]")
22:35:48ghode|afkquestion: how do i get a test plugin to compile with rockbox?
22:36:17rasherput it in the SOURCES file
22:36:28rasherand.. that's pretty much it, iirc
22:37:55ghode|afkok thanks.
22:39:20 Nick StrathAFK is now known as Strath (
22:40:07 Quit Quelsaruk ("KVIrc 3.2.0 'Realia'")
22:40:08 Join Tang [0] (
22:40:11Tangback for a notice
22:40:39Tangcheck out this HA validated news
22:40:44Tangif not done yet:
22:40:58TangNow i leave the place bye
22:41:15ghode|afkdoes anyone know why "cvs update -dP" just hangs my screen?
22:42:22HClslow connecting?
22:42:46 Quit Tang (Client Quit)
22:43:17ghode|afkwell i am on 56k but this is hanging for about 10mins now ;/
22:43:31HClthen try again? o.o
22:43:37ghode|afktried ;p
22:43:51HCldunno, got the right repository?
22:44:06ghode|afkno :/
22:45:53ghode|afkok should i be using or ?
22:46:08Bagdereither one
22:46:19Bagderthey are the same IP
22:47:17ghode|afkboth hang, weird thing is it was working earlier today :/
22:48:02preglowcodecs that use malloc...
22:48:03preglowi'll kill them
22:48:52 Quit ghode|afk ()
22:50:23rasheroh boy, a fixed point aac decoder
22:51:10*fuzzie watches the Helix license suck
22:51:40 Quit XShocK (" HydraIRC -> <- The dawn of a new IRC era")
22:51:40coobwhere rasher?
22:51:58rasherlinked from the forum Tang linked
22:54:03preglowthis would rock intensely
22:54:08preglowafaik, no other devices support he-aac
22:54:13 Join pillo [0] (
22:54:32pillohi all
22:55:19pillojust wanted to jump in and say congrats to HCl for his latest RockBoy!
22:55:24HClthanks :p
22:55:34pillocoupled with the patches from pabs it's really awesome
22:55:39rasherjesus, if they'd just put the licensing terms somewhere sensible
22:55:44pilloi WANT that menu ;)
22:56:12HClpillo: if you have an iriver, get the grayscale version from my ftp..
22:56:38pilloHCl: yes, I was referring to that one ;)
22:56:44HClok :p
22:57:16Bagderrasher: and
22:57:27pilloHCl: your University is going to notice some additional traffic these days...
22:58:16HClnah, ;p
22:58:34HClthey won't notice unless i hit more than 15gb traffic per week
22:58:39HCland even then, they don't really care
22:58:52HClone of my flatmates accidentally uploaded 800gb in 5 days,
22:58:58HClbut because it was an accident, they didn't care xD
22:59:01pilloomg :)
22:59:41rasherBagder: but does it say which of those apply to the he-aac?
23:00:08Bagderthe forum post said "dual-licensed under RCSL/RPSL"
23:00:17pillocan I ask some of the big guys to test my viewer patch (#1152363) and maybe add it to mainline rockbox?
23:00:29pilloi've only tried it on iriver...
23:00:32rasherah, I missed that
23:01:25Bagderrasher: and I believe RPSL is labeled open source, while RCSL is not listed by at least
23:02:31rasherthe license claims that GPL is compatible
23:03:13rasherbut that's the wrong license claiming that doesn't list them
23:03:54BagderI checked there for it too and failed
23:04:08rasherbut that's not an extensive list obviously
23:04:17***Saving seen data "./dancer.seen"
23:05:25[FlaT]Heidel <- RPSL
23:06:02*rasher checks debian-legal
23:06:44rasherI'm pretty sure there are licenses there that are not gpl compatible
23:10:37[FlaT]Heidelnot fully compatible
23:10:40 Join linuxstb [0] (
23:10:41[FlaT]Heideli think
23:13:54rasherthis does not make sense..
23:14:34 Quit linuxstb ("Leaving")
23:14:46rasheroh.. nevermind.
23:16:22*rasher waits
23:16:37preglowi was talking to linuxstb :/
23:17:09 Join linuxstb [0] (
23:17:16preglowlinuxstb: ehlo
23:17:19rasherNo, that sure doesn't look gpl compatible
23:17:41preglowhow is xxx2wav.c linked againt the codecs? i can't find it anywhere
23:18:18preglowin makesfiles, that is
23:18:54amiconnxxx2wav is part of the plugin library
23:19:10 Quit linuxstb (Client Quit)
23:20:05pilloamiconn, bagder: can I ask some of you big guys to test my viewer patch (#1152363) on archos, and maybe add it to mainline rockbox?
23:20:13preglowi can't get this shit working, i've included codec.h, and ld STILL complains about bloody malloc missing
23:20:28 Join toolmanwv [0] (
23:20:32preglowunless i'm mistaken, it should have been complaining about codec_malloc
23:21:04amiconnpreglow: Does any of these include stdlib.h? If yes - this must be removed
23:21:16 Join linuxstb [0] (
23:23:23 Join Seed [0] (
23:23:37 Quit frank_ ("Leaving")
23:24:00Rickor afternoon
23:25:14Strathyup rick, one of those depending on how bad my lag is ;)
23:26:22Strath*** Ping delay from Strath: 1 minute 7.858 seconds
23:27:09linuxstbpreglow: what are you working on? Adding a new codec library?
23:28:06HClhey amiconn
23:28:27HClamiconn: can you checkout current rockboy and do some tests whether its still working on archos?
23:29:27amiconnDid you add the modes for archos as well?
23:30:07 Quit jyp (Read error: 110 (Connection timed out))
23:30:18 Part pillo ("Kopete 0.9.2 :")
23:30:19amiconnYou could try yourself in the simulator too, but let's see whether I find some time
23:30:21HClthough there is no button for it yet
23:30:27HCldon't have access to a simulator
23:30:38amiconnStill working... though at home now
23:30:41HCli'm on a hacked wireless connection that doesn't have the bandwidth to vnc to my server
23:31:42amiconnrasher: u there?
23:31:54preglowlinuxstb: a new mp3 decoder
23:32:03preglowamiconn: i've removed all the standard libs
23:32:08preglowamiconn: headers, i mean
23:32:26linuxstbHave you written an "mp3towav" test plugin for it?
23:32:27 Join jyp [0] (
23:32:33rasheramiconn: yup
23:32:40amiconnpreglow: hmrpf. Then it shouldn't complain when you include codec.h instead
23:32:41preglowlinuxstb: i'm working on it
23:33:00preglowamiconn: but what about stuff like memcpy?
23:33:08preglowamiconn: it uses that as well, and that isn't declared in codec.h
23:34:12amiconnYou'd need an own central header for the codec, or add some more #defines to codec.h
23:34:41amiconnAH, wrong
23:34:52pabsHCl: you really should have hte menu on select, not on mode (for the iriver)
23:34:56rasher < it's in a slashdot comment... but it still looks interesting
23:35:11pabsHCl: it's the easiest button to accidentally hit on the h100, so it should point at a non-game button if possible
23:35:24HCli'll consider it
23:35:27rasherpabs: menu on mode is consistant with rockbox
23:35:31pabsHCl: granted, select is the best button on the gameboy to have there, if any
23:35:48amiconnrasher: I hacked together a funny thing - your logo.rock is now running on the archos player too (you know, the one with the charcell lcd)
23:36:00rasheroh dear
23:36:08pabsHCl: i'll jsut finish the keybinding stuff so this isn't an issue
23:36:19amiconnIt's a somewhat tinly logo though - 16x7 pixels
23:36:24pabsrasher: true, but we don't quit on stop, and that'd be consistent with rockbox as well
23:36:40rasherpabs: if I had my way, it would :)
23:37:11pabsrasher: given the button requirements of rockboy, taht's not really feasable
23:37:37coobhow many buttons do you have
23:37:43HClnot enough.
23:37:49HClwe have 4 buttons
23:37:59HCland we need A,B, select and start
23:38:20coobpossible to detect two at a time?
23:38:32HClonly A + any other
23:38:47HCltechnically, hold could be used but its not very practical
23:38:53HClits assigned to mode switching at the moment
23:38:54pabsHCl: if i get a break from wrk i'll do the keybindings stuff this evening, otherwise it'll have to wait until tomorrow
23:39:01HClyea, take your time
23:39:06HCli'm not really in a hurry either
23:39:15HClmy jaw still hurts like hell from those wisdom teeth being pulled
23:39:26HClso i just do things slowly and when i feel like it
23:39:32rasherDon't mention those
23:39:41HClsorry o.o
23:39:45rasherMine are threatening to annoy me
23:39:54rasherThey might hear you :)
23:39:57HCl :p
23:40:42 Quit jyp ("poof!")
23:41:28Rickis srand() handled automatically by rockbox or should I initialize it with a seed?
23:42:13coobwhy does that always sound vaguely sexual :/
23:42:32HClsomebody whack the pervertedness out of coob
23:45:37amiconnpreglow: The mem*() functions are part of xxx2wav, so the references should be resolved when linking. You need to include the standard header that declares them though.
23:45:55amiconnThat's why I didn't remove all standard headers from the other codecs btw, only stdlib.h is evil
23:48:14rasherrasher: I think you have to seed it
23:48:47rasherrick: that was for you :)
23:49:40 Quit midk ("Leaving")
23:52:35rasheramiconn: Does it run in the sim? patch? I have to see this..
23:53:21amiconnrasher: I'll integrate it into the main logo.rock (made a temporary copy to try my ideas), then commit it
23:53:54amiconnNeed to do variable button definitions for that.
23:54:09amiconn(Player has no up & down nor off)
23:56:01amiconnI also want to change the delta increase / decrease slightly
23:56:14rasherplease do so
23:56:27rasherit seems... strange
23:56:40amiconnCurrently it just adds or subtracts one regardless of direction, so the effect of ech direction pair is switched when the logo bounces
23:57:52amiconnMy change will make the Up button always increase the absolute delta, and the Down button decrease it (down to 0)
23:58:03amiconnLeft/ Right equivalent
23:58:24rashersounds reasonable
23:58:55amiconnAnd yes... the player logo.rock works in the simulator, although a bit slow

Previous day | Next day