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

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

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

#rockbox log for 2005-06-05

00:00:13preglowi only really need mp3/mp2/vorbis
00:00:29preglowi've got about seven mp2 files, so i don't really _NEED_ it, heh
00:00:31amiconnI merely have _one_ album encoed with ogg, and of course I reencoded this to mp3 a while ago
00:00:45preglowreencoding = the devil in disguise
00:00:55preglowi rip all my stuff to vorbis
00:00:58amiconnI can't hear any degradation
00:01:05preglowi'm contemplating switching to musepack
00:02:07preglownah, can't say i usually hear any degradation myself, it just gives me a dirty feeling nonetheless ;)
00:03:36preglowonly transcoding i've evern done with the intent to listen to the resulting files, were the said mp2 files
00:03:41 Quit niobos ("off to bed...")
00:03:44preglowsince the original iriver fw doesn't support mp2
00:03:51amiconnThe .ogg files are ~230 kbps, so quality should be no problem
00:04:29preglowahh, no
00:04:40rasherq10 oggs still don't quite run
00:04:42amiconnre mp2: The funny thing is that the archos _player_ firmware also doesn't support mp2, although the mas is capable of handling it
00:04:46preglowrasher: small wonder
00:04:52preglowrasher: vorbis runs vERY marginally as it is
00:04:54rasherit's pretty close though
00:04:56amiconnThey changed that for recorder
00:05:41rasheriirc, the difference in speed wasn't that large from q1 to q10
00:05:59preglowvorbis is _just_ over 100% realtime
00:06:27preglowif you would have asked me before slasheri's patch, i would have told you to forget having vorbis playback until further optimisations were done
00:06:42rasheryeah, pretty surprising how close q10 is though"> is one
00:07:55preglowvorbis is way too memory intensive
00:08:18preglowi'm really anxious to see what monty comes up with with vorbis2
00:08:18amiconnDoes vorbis use much memory copying?
00:08:32preglowamiconn: i believe the vorbisfile layer does
00:08:38rasherpreglow: try the q10 though - it's quite close
00:08:55 Quit ghode|afk (Read error: 104 (Connection reset by peer))
00:08:57amiconnMabye memcpy and memset optimisation will help...
00:09:08preglowamiconn: yeah, but i'm more hoping to get rid of the copies altogheter
00:13:50 Join coob [0] (
00:14:00coob:D got ogg playing on the ipod!
00:16:28preglowyou are swine indeed, we had to jump through flaming hoops to get it realtime
00:16:38coobthis is the stock tremor
00:16:53preglowbut then again, stock tremor has some arm optimisations
00:17:08coobthis wont run on the older ipods :/
00:17:14preglowis there much difference?
00:18:35preglowtremor uses so much ram it's pretty hard to find out what to stuff in iram
00:18:44coobwell, with the new cpu stuff bern re'd today, ipods using the PP5020's running at 75mhz get ~32 bogomips, PP5002s at 75Mhz get ~ 23 or something (haven't got one, wouldn't know)
00:18:51coobthere's a lowmem branch
00:18:55preglowyeah, i know
00:19:01preglowi haven't tried it yet
00:19:03preglowbut i really should
00:19:06coobbut yeah, this is without using iram at all
00:19:14coobif iram were used, it would speed up quite a bit
00:19:20preglowoh yes
00:19:31preglownothing helps as much as putting things in iram for us
00:19:43preglowordinary ram accesses are shit slow
00:19:59amiconnUse burst mode and icache...
00:20:08preglowof course
00:20:11preglowbut it's not always so easy
00:20:17coobwell, i say we don't use iram... we don't at all for the audio codec, we use it as a pcm buffer
00:20:25preglowi'm going to have pbvas take windows out of iram again
00:20:28preglowand use burst mode for windowing
00:20:40preglowsince that's the only place windows are used, and they occupy tons and tons of memory
00:20:53preglowcoob: that's a waste, really
00:21:01coobit's a necessity
00:21:21coobfor a very convoluted reason i dont have time to get into/fully understand myself
00:21:27coobit's not staying that way heh
00:21:42preglowgood, 'cause it's a waste, like i say
00:21:52preglowthere are temp areas in all codecs that get touched _a lot_
00:22:00preglowwhat do you use for mp3, btw? libmad?
00:22:15coobi now have to write a decent generic player interface /file buffer/playlister :(
00:22:25coobmp3 we use ipp
00:22:35coobbinary only stuff from intel :/
00:22:41preglowfor ARM?
00:22:48coobintel own arm
00:23:03preglowdidn't know that
00:23:08coobhelix's mp3 decoder runs almost as fast
00:23:18cooband we're putting it in as a compile time option
00:23:30coobtoo much work, too little spare time
00:24:01coobhey iis it the iriver h10 that has a pp5020 too?
00:24:07amiconncoob: binary only isn't that nice in an opensource project I'd say
00:24:15coobamiconn: tell me about it.
00:26:05 Join thegeek_ [0] (
00:26:18preglowcoob: yeah, btw, that's h10
00:26:39coobuses the same audio dac too
00:26:42coobas the ipod mini
00:26:45preglowcoob: we just might steal your research if we get enough people wanting to port rockbox to it one day :PP
00:27:04coobwell, there's registers on the wiki and in the source :)
00:27:42coobi was thinking more of porting iPL to the iriver...
00:27:44amiconncoob: I hope to back-port your greyscale-extended cube to rockbox
00:27:59amiconn(soon on iriver, and a bit after that to archos too)
00:28:13coobi use microwindows' handy polygon stuff
00:28:23coobthat was my first ever experiment with C, it's fugly.
00:28:41amiconnI already have a nice little filled-triangle routine
00:28:48amiconnIt's rather simple
00:32:28coobit's in colour on the photo, now we have the cpu running faster i should try putting textures on it
00:32:42 Quit ehntoo (Read error: 104 (Connection reset by peer))
00:33:38 Quit xen` ()
00:35:23preglowbeen a long time since i did texture mapping by hand
00:39:55 Quit zezayer_ (Remote closed the connection)
00:41:56 Join spiralout [0] (
00:43:43 Quit thegeek (Read error: 110 (Connection timed out))
00:44:13 Join midk [0] (
00:47:13 Join asdsd____ [0] (
00:47:33 Part asdsd____
01:01:15thegeek_btw, reencoding files is audible if you use good headphones and a good amp
01:01:22thegeek_it's not very audible
01:01:28thegeek_but it's there
01:01:40thegeek_like you sometimes hear bad stuff in mp3's
01:01:50thegeek_it's just a little more pronounced
01:02:23preglowmy ears aren't exactly gold plated
01:04:09thegeek_neither are mine
01:04:26thegeek_I just bought a sennheiser 595 and built a cmoy
01:04:32thegeek_I can hear the difference now
01:04:43thegeek_(and the cmoy is in no way a good amp)
01:04:49preglowcontemplated getting one of those myself
01:04:55thegeek_a cmoy?
01:05:01thegeek_it's nice
01:05:12thegeek_BOTH my parents heard a _very_ audible difference
01:05:15thegeek_that speaks volumes
01:05:29preglowi've successfully blown my ears out too many times
01:05:30thegeek_my father is actaully hearing impaired
01:05:34preglowhaha, so is mine
01:05:41preglowand i'm not far awaty
01:05:55thegeek_anyway, it really does help a lot
01:06:17thegeek_atleast when you have some != cheap headphones
01:06:33thegeek_the differences is not as pronounced with my koss ur40's
01:06:36preglowthink i paid around 1000nok for mine
01:06:45thegeek_but with the 595's it's a world of difference
01:07:04thegeek_the overall soundimage is more pleasing
01:07:07thegeek_better bass
01:07:15thegeek_not just more
01:07:23thegeek_but more "focused"
01:07:31thegeek_and much better details
01:07:50thegeek_and the cmoy is fun to build too
01:07:55thegeek_I started from scratch
01:08:02thegeek_with no experience with soldering or anything
01:08:03thegeek_and I did it
01:08:11thegeek_with components solely from elfa
01:08:12preglowi've got experience with it, so should be a breeze
01:08:29preglowi've actually got tons of elfa stuff here
01:08:32preglowperhaps some of it's usable
01:08:49thegeek_I made a list
01:08:56thegeek_let me get the url
01:09:45thegeek_that's the best tutorial around
01:10:55thegeek_but grr
01:10:58amiconnBleh. There are numerous source files in crlf format :(
01:11:06thegeek_I'm getting annoyed with my audio files now
01:11:15thegeek_I hear codec artifacts all the time now
01:11:33thegeek_amiconn : can't they all just be run throught a crlf stripper?
01:11:40thegeek_dos2unix or whatever
01:11:46preglowthegeek_: what settings?
01:11:56thegeek_preglow : settings?
01:12:03thegeek_you mean encoding?
01:12:08thegeek_mostly 192kbit
01:12:26preglowvbr mp3 is pretty nice
01:12:31preglowbut very high avg bitrate
01:12:39preglowi have to go very low with vorbis before i can hear it
01:12:41thegeek_I have vbr's too
01:12:42Stryke`aps is pretty perfect
01:12:47preglow128kbps vbr and i usually can't hear a difference
01:12:50thegeek_it's not like one type of encoding stands out as bad
01:12:57thegeek_it's usually more dependent on the music
01:13:07thegeek_I just listened to kings of convenience
01:13:12thegeek_and I could hear some artifacts
01:13:21preglowwell, yeah, but that's a pretty extreme case ;)
01:13:25thegeek_I know
01:13:34thegeek_it was an extreme example to demonstrate my point;)
01:14:01thegeek_I'm worried it might be my soundcard too
01:14:09Stryke`what card?
01:14:10thegeek_I've got a (supposedly) good dac coming in
01:14:15thegeek_soundstorm based
01:14:18thegeek_ac97 shit
01:14:26preglowi wanna buy myself an echo mia midi
01:14:33thegeek_I'm going to use the bitperfect digital/optical out
01:14:38preglowi've got a shitty sb live now
01:14:39thegeek_straight into a dac
01:14:45Stryke`audiophile is a very expensive hobby
01:14:53thegeek_use the kx project drivers
01:14:59thegeek_and use the "back" output of the card
01:14:59preglowthegeek_: i've been using those for three years
01:15:05preglowthegeek_: only reason i've still got the card
01:15:07Stryke`resamble to 48 KHz
01:15:07thegeek_since that output uses a much better opamp
01:15:09thegeek_I do
01:15:23thegeek_Stryke` : hehe, indeed
01:15:26thegeek_but it's fun
01:15:31preglowyeah, i've got them reversed usually
01:15:34thegeek_and I'm leaning towards diy
01:15:39preglowbut not now, since i can't figure out how to do it in linux
01:15:40thegeek_I think I'm going to build the m3 amp
01:18:19preglowthe live is bloody good value with the kx drivers
01:18:36Stryke`Chaintech AV-710 is a good value
01:18:38preglowbut yeah
01:18:44thegeek_depends how you define value
01:18:51thegeek_if you want features kxproject is great
01:18:52preglowi'm prolly gonna buy myself a couple of monitors after i'm done moving
01:18:57thegeek_if you want good sound av-710 is goo
01:18:58preglowand then i'd like a proper audio card as well
01:19:11***Saving seen data "./dancer.seen"
01:20:06amiconnMMC icon on iriver is no more :)
01:20:22HClits about time
01:20:22HCl :X
01:20:28Stryke`was it that annoying?
01:20:29HClgood work :)
01:20:45amiconnStryke`: It doesn't belong there
01:21:04Stryke`when did it appear? HDD access?
01:21:07preglowannoying or no, h1x0 has no mmc
01:21:14amiconnStryke`: yup
01:21:57amiconnI noticed that Linus' HD-defaults-to-off fix does no good on RoLo
01:22:38amiconnThe freshly loaded rockbox will turn off the hd immediately after it was active
01:22:45amiconnThat's definitely bad for the hd
01:22:54preglowperhaps it should be wrapped in a #ifdef BOOTLOADER
01:23:01amiconnHowever, it _is_ a wanted feature in the bootloader
01:23:17amiconnDid you try rolo with a recent build?
01:23:20preglowi seriously doubt he'll have you shot for doing that, so go ahead
01:23:29amiconnThen you'll notice what I mean
01:23:38preglowand if you say it's bad for my drive, i'll take your word for it ;)
01:23:39amiconnBootloader version doesn't matter for that
01:23:44amiconn(I still have v1)
01:24:00preglowme too
01:24:04 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
01:24:16amiconnWell, the drive can stand that when it happens occasionally
01:24:49thegeek_one thing I've always wondered about
01:24:55thegeek_how can the drive park when the power is cut
01:25:01amiconnWhen HD power is cut while the heads aren't parked, the drive has to employ 'emergency park'
01:25:14coobhandbrake parking? :)
01:25:29thegeek_I imagined that handbrake stuff on trains
01:25:31amiconnthegeek_: It uses the spindle motor as a generator
01:25:34thegeek_veritable trainwreck
01:25:44thegeek_how the hell do you know that;)
01:25:53amiconnthegeek_: Reading datasheets
01:26:15amiconnIntel describes pretty clearly how it works, and that it's worse than ordinary parking
01:26:45amiconn..because in that state, it isn't possible to control the current precisely
01:27:13amiconnA drive can stand way more ordinary park/unpark cycles than emergency park cycles
01:35:22 Join asdsd____ [0] (
01:35:29 Part asdsd____
01:40:49 Join tvelocity [0] (
01:43:01*preglow claps wildly
01:43:43amiconnWell that was simple, but I wanted to wait for the compile state of my huge LED commit first
01:44:35MoosCamarohehe amiconn first commit for iriver ;)
01:44:42MoosCamarovery cool
01:44:44amiconnNot exactly
01:44:49preglownot really, no
01:45:05amiconnI already did some fixes for iriver even when I didn't have one
01:45:18MoosCamaroa ok scuse me :)
01:45:32amiconnOf course it's easier to test stuff on da real thing
01:46:58MoosCamaroi know you contribute a lot for the others models
01:48:09amiconnI hope to continue that
01:48:34MoosCamaro:) we hope too ;)
01:49:47cooboh hey! you guys might want to check out libflite for your blind stuff, we got that working on the ipod too, spoken menus = neat
01:49:59coobmemory hog though :(
01:50:09coob4-5meg for 8khz synth
01:50:19amiconnWe have spoken menus on archos
01:50:35coobnot synthed voice though is it?
01:50:47amiconnIt's a pre-made clip collection
01:51:01coobah, but it has to spell out song names?
01:51:24amiconnSpell out, tell the file number + type, or play a clip file
01:51:54amiconnThere's a script to generate clip files for directories, files, or both
01:52:08coobsynth is increbly cool i must say :)
01:52:18coobbut sspelt right.
01:53:03amiconnWell, pre-made clips are not as high-tech as real synthesis, but I think they sound better it's the only solution on archos, unless someone manages to write complex code for the MAS' DSP
01:54:08coobwell i would personally hate to make clips of all 3528 songs, 778 artists and 261 albums in my collection :)
01:54:39amiconnIt's as simple as running a script on the PC and waiting a couple of minutes
01:55:16amiconnApart from that, I only use clips for folders, not for files
01:56:27coobwai so the script uses a synth prog?
01:57:01amiconnThe windows solution is a vbscript using any sapi5 compatible voice
01:57:10coobahhh gotcha
01:57:16coobi thought you meant manually recording clips :)
01:57:25amiconnYou could do that as well
01:57:39amiconnThe clips are simply mp3s with a different extension
02:04:24preglow4-5 meg for a synth?
02:04:24rasheramiconn: the file browser is way faster now.. I'm thinking the only change related to this was the led fix?
02:04:34coobyep :/
02:07:33amiconnrasher: Maybe the removal of some code changed the caching behaviour of some other code
02:07:56rasherIt's really very noticable
02:10:10 Join ashridah [0] (
02:21:34MoosCamarogood night all
02:21:46 Quit preglow ("bed")
02:21:53 Part MoosCamaro
02:21:57 Quit spiralout ("ChatZilla [Firefox 1.0.4/20050511]")
02:23:18 Join Strath [0] (
02:41:06 Join Stryke` [0] (
02:41:48 Quit StrathAFK (Read error: 110 (Connection timed out))
02:42:49 Join textchimp [0] (
02:49:05textchimpis slasheri's audio codec in cvs yet?
02:49:32Stryke`nope, only as a patch
02:50:41Stryke`i've found people are generally pretty good at providing a compile for you if you ask, they were for me
02:51:54 Quit midk ("later...")
02:55:35textchimpi can apply the patch myself....but the sooner it's in cvs, the sooner other people will start working on it too
02:55:56Stryke`sure, that should happen within the week, i'd imagine
02:55:56coobwhatd oes it do?
02:56:41Stryke`MP3, Vorbis playback, basically
03:02:20 Join Musicmad_ [0] (
03:05:39HCladhd people are horrid.
03:06:53coobnothing a little ritalin can't fix :)
03:07:29Stryke`over diagnosed disorder
03:10:52 Quit einhirn (Read error: 60 (Operation timed out))
03:19:15***Saving seen data "./dancer.seen"
03:28:00Stryke`does anyone have a center-scrolling patched, Slasheri patched build post-amiconn's MMC icon fix?
03:28:54 Join amiconn_ [0] (
03:43:59textchimpdoes anyone know what kind of file names rockbox gives to recorded files?
03:45:41 Quit amiconn (Read error: 110 (Connection timed out))
03:45:41 Nick amiconn_ is now known as amiconn (
03:58:33 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
04:05:21 Join QT_ [0] (as@area51.users.madwifi)
04:09:30 Quit RED_M_CHIU ("( :: NoNameScript 3.81 :: )")
04:12:20 Join Musicmad [0] (
04:18:16 Quit QT (Read error: 110 (Connection timed out))
04:28:26 Join kramerica [0] (
04:29:41 Quit Musicmad_ (Read error: 110 (Connection timed out))
04:42:57 Quit Musicmad (Read error: 110 (Connection timed out))
04:42:57 Quit kramerica (Read error: 104 (Connection reset by peer))
04:42:58 Quit textchimp (Read error: 54 (Connection reset by peer))
05:09:02 Quit tvelocity ("Leaving")
05:15:55 Join DMJC [0] (
05:19:18***Saving seen data "./dancer.seen"
05:31:03 Join textchimp [0] (
06:33:08 Quit Dave99 ("CGI:IRC (EOF)")
06:33:08 Quit DomZ ("CGI:IRC (EOF)")
06:42:29 Quit textchimp (Read error: 60 (Operation timed out))
07:19:19***Saving seen data "./dancer.seen"
07:40:41 Join webguest26 [0] (
07:41:39 Quit webguest26 (Client Quit)
07:51:30 Join Musicmad [0] (
07:59:15 Quit Strath (Read error: 110 (Connection timed out))
08:38:53 Join kurzhaarrocker [0] (
08:39:25kurzhaarrockermoin amiconn
08:39:43kurzhaarrockerIn the cvs I read "When starting a recording, wait a bit longer before grabbing a header"
08:39:56kurzhaarrockerDoes that include incresed latency when recording triggered?
08:40:29kurzhaarrocker-include +imply
08:41:27amiconnnope, it doesn't
08:41:36kurzhaarrockeryay! :D
08:41:58kurzhaarrockerThanks for the improvements.
08:42:57amiconnHowever, I had to 'brake down' the peakmeter a bit
08:43:08kurzhaarrockerI noticed that
08:44:20kurzhaarrockerI'm surprised that it makes the trigger led blink more regularily
08:44:57amiconnThis is because of another change (the delay generation
08:45:24amiconnThere was an interference of 2 frequencies
08:46:32kurzhaarrockerOne being the blink freq and the other the screen update freq?
08:47:13amiconnThe 'blink mask' should make the LED change state every 16 ticks, but the loop round trip was always ~10 ticks (due to peak_meter_peek_get_button)...
08:47:30amiconnSo the state change could only ever happen after 10 or 20 ticks
08:48:10kurzhaarrockerok, I understand
08:51:18amiconnRecording should now be as glitch-free as possible
08:51:29amiconnI still have some ideas for improvement
08:52:05kurzhaarrockerAre you talking about the dictataphone trigger mode? :)
08:52:14amiconnThat's your task ;)
08:53:15amiconnI'm thinking of the following: (1) Always prepend an id3v2 header, even at (automatic or manual) file splits
08:54:07amiconn(2) Add a xing header if at all possible
08:54:23kurzhaarrocker(That should apply to the split editor, too - in my dreams at least)
08:54:55amiconnCurrently, this is not done with > 1 file (i.e. split recording), or when prerecording
08:55:32amiconn(3) Improve the frame boudary handling
08:56:06kurzhaarrockerThose headers would be generated empty and filled with life after the recording has finished I assume?
08:57:10amiconnId3v2 is always an empty header, it's just reserving the space so that later tagging doesn't need to extend the file
08:57:54amiconnIt's just necessary to prepend the header within the buffer after the old file has been fully saved
08:58:52amiconnXing header generation is already done, but only if you're recording one file from start to stop without prerecording, for 2 reasons
08:59:26amiconn(1) It needs the frame count, but the frame count at the start is nonzero if prerecording
09:00:30amiconn(2) It uses the audio buffer to construct the header, unconditionally overwriting it, so no next recording can run meanwhile
09:01:36amiconn(1) is easy, just store the frame count at start. For prerecording, this means adding a frame count to each buffer pointer
09:03:08amiconn(2) will need a separate buffer
09:04:12kurzhaarrockerWhy prepend the id3v2 tag in the buffer instead of writing it directly to the new file?
09:05:17 Quit Musicmad (Read error: 54 (Connection reset by peer))
09:05:26amiconnDelayed file open
09:19:24***Saving seen data "./dancer.seen"
09:40:46ashridahi've got a set of mp3s that refuse to play with the last patch slasheri had put out
09:41:31ashridahit displays the WPS screen, but then immediately stops and ditches me back out
09:45:50 Nick ashridah is now known as Lost-ash (
09:53:44 Quit kurzhaarrocker ()
09:56:47 Join t0mas [0] (
10:01:01bill2or3there's a bunch of H340's on ebay
10:01:23bill2or3closing soon, and they're < $300. tempting
10:07:22 Join Allex [0] (
10:08:09 Quit Allex (Client Quit)
10:16:47SlasheriLost-ash: could you send on of those mp3s?
10:24:43 Join Harpy [0] (
10:24:52Lost-ashSlasheri: i can toss one up someplace. it'll take me a little while tho
10:24:53Lost-ashi'm on dialup.
10:25:19SlasheriAh, ok :)
10:28:29 Join LinusN [0] (
10:28:30Lost-ashhopefully i should have one for you in 10-15 minutes
10:28:58Slasherihi LinusN :)
10:29:03LinusNamiconn: skipping the id3v2 header in the subsequent files can be a good thing
10:29:16LinusNmakes it easier to join the files
10:29:48amiconnBut why would you split then in the first place?
10:30:07LinusNand the frame count is only valid until the frame counter reaches the limit
10:30:13amiconnI know
10:30:23LinusNamiconn: you have a point there
10:30:39amiconnThis still doesn't forbid to calculate a xing header as long as the counter doesn't reach the limit
10:30:53LinusNSlasheri: hi btw
10:31:03amiconnWe just have to store the start count and calculate the difference
10:31:15amiconn...when prerecording, or after a split
10:31:42SlasheriLinusN: the patch should be ready for commit is there is not any big problems
10:32:03LinusNSlasheri: good, i'll see if i can find the time to do it later today
10:32:14Slasheriok, thanks :)
10:32:15LinusNSlasheri: does it compile with the simulator?
10:32:24Slasherii haven't tried that yet
10:32:27LinusNdo that
10:32:43LinusNi'm afraid it won't compile
10:32:56Slasherihmm, i will fix it then :)
10:36:14amiconnLinusN: Btw, the xing header creation bug was a really nasty one. Fortunately Daniel implemented logf(), it helped a lot...
10:36:59LinusNi gotta go, cu around
10:37:05 Part LinusN
10:49:55 Part amiconn
10:56:21 Join niobos [0] (
10:56:29nioboshi all
11:16:28Lost-ashSlasheri: dude, sorry it took me so long, i got distracted by dinner
11:16:36Slasherinp :)
11:16:38Slasheriah, thanks :)
11:17:30Slasherii just corrected wps bindings and now pressing select key will get you back to file browser. I will upload updated patch today
11:19:25***Saving seen data "./dancer.seen"
11:26:35crwllooks like your patch doesn't work against current CVS
11:28:00Slashericrwl: Hmm, what is the problem?
11:28:11SlasheriI just compiled it against current cvs and it seems to work :)
11:28:21crwlpatching file firmware/sound.c
11:28:24crwlHunk #1 FAILED at 458.
11:28:28SlasheriAah, that..
11:28:45SlasheriI really don't know what is wrong with that patch but you have to manually patch that file :/
11:37:32SlasheriLost-ash: Hmm, my current patch will play the song you send
11:42:34 Join Mirfle [0] (
11:44:12 Nick t0mas is now known as t0mas| (
11:44:33 Join t0mas [0] (
11:45:45 Part t0mas
11:55:02Lost-ashSlasheri: that's cool. i'm using an older one from the 3rd or so
11:56:44Slasheriyou should wait that i will release a new patch today, it will include some codec switching bugfixes also
11:58:30 Join t0mas_ [0] (
11:58:48 Part t0mas_
12:00:36bipak_to which clock speed does rockbox set the coldfire cpu atm.?
12:01:06t0mas|48 mhz
12:01:10t0mas|and when needed 120
12:01:35bipak_ok :)
12:02:41Mirflewhat's the lowest it can theoretically be set to?
12:03:35Mirfleso you can have a "soft" poweroff, and set a timer or something like that, right?
12:03:35 Join bipak [0] (
12:04:07t0mas|yes, that was a thought of me and preglow...
12:04:12t0mas|but we never tested it...
12:04:36t0mas|the iriver can run for a very long time... when the disk is stopped... and the cpu is set to 11 mhz
12:06:03 Nick t0mas| is now known as t0mas (
12:06:08Lost-asheven longer if it's on A/C power.
12:08:31*Lost-ash hands himself the cookie of obviousness
12:08:34*Lost-ash munches
12:08:46SlasheriHmm, has anybody measured what is the minimum power consumption at 11 MHz when everything unnecessary are powered off?
12:09:05SlasheriThat would be a great feature to have even a clock
12:09:07 Join einhirn [0] (
12:09:22*t0mas will try to
12:09:28t0masbut how can I get the cpu at 11 mhz?
12:09:42SlasheriLook at the cpu_boost code
12:09:54Slasherior set_cpu_frequency or something like that
12:09:55t0masok... I can try to add it to the debug menu
12:10:09t0maspower down the disk... clear the screen... and set the cpu to 11 mhz
12:10:19SlasheriBut you should also power off disk, display and backlight to get accurate measurements
12:10:27t0masnot display...
12:10:32t0masthen I won't know when it stops...
12:10:42SlasheriAh :)
12:10:49SlasheriYou will test how long it will run?
12:10:58t0maswell... I will ask rasher to test ;)
12:11:02t0mashe did all other battery tests too
12:11:13*t0mas looks up the code to edit
12:11:13 Quit bipak_ (Read error: 60 (Operation timed out))
12:12:17SlasheriI think that i will buy an ionity battery from Maybe i will do some real measurements also when i change the battery
12:13:18*t0mas slaps kdevelop...
12:13:22t0mascrashing :X
12:13:30Slasheritry use kate :)
12:13:34Slasheriit's great and stable
12:13:56*Lost-ash huggles eclipse
12:14:04Lost-ashthe CDT plugin's a bit dodgy tho
12:14:09Lost-ashnice for javarse
12:14:17t0mashm? CDT works great here...
12:14:40Lost-asht0mas: not on 3.1rc1 it isn't.
12:14:49Lost-ashthe debugger acts weird
12:14:53t0masI still have 3.0
12:16:22Lost-ashyeah. 3.1 gets released at the end of the month. some plugins aren't completely stable yet
12:16:29Lost-ashalthough 34.1rc1 by itself works nicely
12:16:52Lost-ash3.1rc1 even
12:21:37t0masI need linus :)
12:21:43t0masvoid set_cpu_frequency(long frequency);
12:21:54t0masto what should I set it when I need 11 mhz?
12:21:59t0masas it isn't really 11 hmz
12:23:01Slasherit0mas: look from the debug menu. If you press down the joystic, the frequency will be changed to ~11 MHz
12:23:15t0masah ok
12:23:28t0masI only knew the boost option
12:23:48t0maswhen I switch it to 11 mhz in debug... and then run a plugin to disable backlight and disk...
12:24:00t0masthen I just need a simple image on screen to see if it's still working
12:24:03t0masand count the time
12:25:05t0masit isn't going to 11 mhz...
12:25:25t0masit's 119952000 = 120 mhz
12:25:36t0masand 47980800
12:25:39t0mas= 48 Mhz
12:26:01SlasheriMaybe it would be even better if you just completely power of the display and power it on for a short time in periods
12:26:43SlasheriNo, it's 11 MHz when you press down the joystick
12:26:46t0masyes.. but then I really need to pay my sister to look at it :P
12:26:57Slasheri11 289 600 Hz
12:27:07t0maserm? when I press down... it sets boostcount to 0, and stays at 48
12:27:33Slasheriwhen i do that, it will change to the frequency above
12:27:46t0masyou have a debug build?
12:27:49t0masor a very old one?
12:28:07Slasherihmm, i have a fresh build with logf debug-option enabled
12:28:09crwli think pressing rec button changed to 11 MHz here
12:28:10Slasherinothing more
12:29:43t0mashm... I'll try the rec button
12:29:44 Join Sucka [0] (
12:30:19t0masSlasheri: do you have thick fingers? ;)
12:30:37Slasherit0mas: Hmm, no? :D
12:30:43t0mashitting rec and joystick at once? ;)
12:30:56Slasherii don't do that ;)
12:30:59t0mashaha ok
12:31:05t0masI have it at 11 hmz now :)
12:31:09Slasherigreat :)
12:32:06t0maswhy don't we run it at 11 mhz constantly?
12:32:11t0masit's still working...
12:32:41Slasheriit's not possible to feed pcm fifo at that frequency
12:32:53t0masah.. we need 48 for that?
12:32:54Slasheriyou may try and the audio will get horrible distorted
12:32:59SlasheriI think so
12:33:08t0masand we can't set it to 22 or something?
12:33:17t0masso it can just fill pcm buffers... and nothing else?
12:33:19Slasherimaybe we can :)
12:33:23Slasheriyou should try :)
12:33:30t0massetting it to 22?
12:33:34Slasherifor example
12:33:35t0masI guess Linus knows...
12:34:07t0masa plugin disabling the backlight... and stopping the disk
12:34:23Slasheribut when audio is not playing, i think the frequency can be kept at 11
12:34:42t0masbut for testing... I can just set the frequency to 11 mhz in debug
12:34:49Slasherihowever, there is no point in having that frequency while audio is playing, because we wont get any power savings
12:34:57t0maswhy not?
12:35:04t0masswitching to 120 some times for decoding?
12:35:16SlasheriWe have constantly change the frequency to 120 MHz for a short periouds because even 44 is not enough for codecs
12:35:42t0masbut decoding on 120 mhz... untill the memory is full... and then drop back to 22...
12:35:47t0maswouldn't that work?
12:35:57Slasheriit does that already with 44 MHz
12:36:20t0masok, then optimizing codecs to keep the bursts short... and checking 22 might save power?
12:36:24Slasheriand while we are at 44, full decoding work is still being done until pcm buffers reaches watermark and we change to higher frequency
12:36:35t0masah ok
12:36:36SlasheriSo 22 should not give any power savings
12:37:24Slasheriif we can get a really good optimized codecs that can run below 44 MHz, then it will save some amounts of power
12:38:31 Quit Mirfle (Read error: 104 (Connection reset by peer))
12:38:40t0mascan I control the led on iriver?
12:38:43t0masor is that hardware?
12:38:58 Nick Lost-ash is now known as ashridah (
12:39:03Slasherii think you can control the hdd led by software
12:39:15Slasherithat might be a good idea, you could disable the display
12:39:15 Join Shagnar [0] (
12:39:47t0maslike this:
12:39:48t0mas- Set frequency to 11 (debug menu)
12:39:48t0mas- Stop disk (rb->ata_sleep())
12:39:48DBUGEnqueued KICK t0mas
12:39:48t0mas- Disable backlight (rb->backlight_off())
12:39:48t0mas- Clear screen (rb->lcd_clear_display())
12:39:48***Alert Mode level 1
12:39:50t0mas- Sleep 300 seconds... flash hdd led
12:40:06t0masor can I disable the screen more than just clearing it?
12:40:24Slasheriyou should check if there is a way to completely power it off
12:40:38t0masany other place then Linus to check that?
12:40:57Slasherii don't know..
12:41:07t0masto what is the LCD controller connected?
12:41:36Slasherii don't think it's on the i2c
12:41:49Slasheriprobably directly connected to cpu pio lines
12:41:52t0masok, then there's no way for me to disable it...
12:42:00t0masas I don't know how to control those...
12:42:08Slasherii think you have to write some control code to the display to do that
12:42:30t0masin lcd.h
12:42:30t0masextern void lcd_write_command( int byte );
12:43:10Slasherilook at lcd-h100.c
12:43:22SlasheriThere are a lot off power control defines
12:43:28 Join Mirfle [0] (
12:43:45 Join webguest58 [0] (
12:44:53Slasherimaybe you could try: lcd_write_command(LCD_CNTL_OSC_ON_OFF | 0); lcd_write_command(LCD_CNTL_OFF_MODE | 0); lcd_write_command(LCD_CNTL_ON_OFF | 0);
12:44:56Slasherithat _might_ do it
12:45:31t0maswhat's OSC?
12:45:52t0masand why the | 0 ?
12:46:21Slasheriit will set it to "off" (not necessary)
12:46:29Slasheriand | 1 would set to on
12:46:38Slasheribut be careful if you try change those
12:46:41t0masLCD_CNTL_OFF_MODE | 0 -> LCD_CNTL_OFF_MODE | 1
12:46:55t0masor not?
12:47:11Slasheriyou disable the display and don't care about re-enabling it
12:47:25CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
12:47:25*t0mas takes a look at the shutdown code...
12:48:40Slasherimaybe a better solution would be to not enable the display at all at boot
12:49:05t0masyes... but then I will have to edit the bootloader
12:49:19Slasheriah, that's correct and you should no try that
12:49:31t0masno... the only one who can try that is linus
12:49:49***Alert Mode OFF
12:50:31 Join MoosCamaro [0] (
12:50:38MoosCamaroHi all
12:50:45t0masand disable the backlight
12:50:48t0masmight work?
12:51:02t0masnot as good as disabling it fully... but we can better wait for Linus to test that...
12:51:14Slasheriit wont disable oscillator or display power
12:51:33t0mashm... how sure are you about disabling the oscillator?
12:51:33Slasheribut maybe you should ask linus how to do that
12:51:41Slasheriplease ask linus :)
12:51:41t0masyes, I'll do...
12:52:06Mirflenoob question: what's an oscillator?
12:52:45SlasheriMirfle: it generates constant frequency to drive some logic. In this case it should control display update
12:53:04t0masSlasheri: where do you think that HDD led is?
12:53:14t0masI guess no-one wrote code for it untill now?
12:53:15Mirfleahh... thanks :)
12:53:28Slasheriin led.c
12:53:40t0masthat isn't archos specific??
12:54:01Slasheriah, it might be
12:54:09t0masI'll test
12:54:44SlasheriI think there is no way to control it..
12:54:46t0mas#ifdef HAVE_LED
12:54:48t0masthat's not iriver
12:54:56SlasheriLooks like the led might be done on hardware level
12:55:04t0mashm... both?
12:55:13t0mas(as I don't care wich led I control...)
12:55:26Slasheriat least the charge led is hardware only
12:55:47t0mashm... can I go to usb mode w/o spinning up the disk?
12:56:04t0masthen I can go in and out of usb mode every 10 minutes... and check my syslog for how long it was running...
12:56:11Slasheriyou can quite easily spin up the disk but _that_ will really eat power :D
12:56:26t0masno, I wanted to do it w/o spinning up the disk
12:56:48Slasherii don't know if that's possible..
12:57:54t0masI'll first try it with some simple text on screen... and no backligt, disk and cpu
12:58:09Slasheriyes, that is a good start
12:58:11t0masand wait for linus to explain how to disable the lcd... and make some other signal
13:05:46t0mas /* stop the disk */
13:05:46t0mas rb->ata_sleep();
13:05:53t0maswhy doesn't it power off the disk?
13:08:12 Quit webguest58 ("CGI:IRC")
13:10:28t0masit waits for the timeout...
13:12:04 Join LinusN [0] (
13:12:10 Quit edx (Read error: 110 (Connection timed out))
13:12:13t0mashi LinusN
13:12:21t0mashave you read the logs?
13:12:29t0masor should I ask some questions again? :)
13:12:48SlasheriLinusN: would you like to see the new patch? Now simulator build will compile fine :)
13:12:56LinusNSlasheri: yes please
13:12:59Slasheriok :)
13:13:04LinusNt0mas: i have read the logs
13:13:20t0masok, and you know how to do it? :)
13:13:24LinusNwhy do you want to disable the display?
13:13:34t0masto save power
13:13:58LinusNi don
13:14:06 Join leftright [0] (
13:14:10LinusN't think it'll save that much power
13:14:23LinusNbut we'd have to test of course
13:14:36t0masand the led? do you know if we can control it?
13:14:46LinusNand no, i don't know more than you do about how to turn it off
13:15:09LinusNthe green led can't be controlled
13:15:32t0masand the red one?
13:15:36SlasheriLinusN: now it's updated,
13:15:41LinusNand i don't see why we would disable the red one
13:15:48LinusNSlasheri: great
13:15:50t0masno, I want to enable it once in a while
13:16:00t0masso I can see the unit is still running... while the screen is off
13:16:12LinusNthe red led is controlled by the hard drive
13:16:51LinusNwe have discussed the cpu frequency
13:17:13LinusNwe should probably have three levels
13:17:29LinusNone idle (11mhz), one "normal" (48mhz) and one boost (120 or even 140)
13:17:50LinusNthe normal one can probably be less that 48
13:18:35t0masyes... if it can decode and feed pcm at the same time... it can be lower
13:18:43t0masto just feed pcm data
13:18:44LinusNt0mas: how do you plan on measuring?
13:18:54t0maspay my sister to look at it for several hours...
13:18:56t0masor ask rasher
13:19:03LinusNor use a multimeter?
13:19:12t0masthen I'll have to take it apart first
13:19:20LinusNof course
13:19:28***Saving seen data "./dancer.seen"
13:20:01LinusNi have plans of adding an ampere meter inside my 120
13:20:04t0mashm... I'm not good at that kind of things... normally it won't close after I opened it :)
13:20:22LinusNand read it with the A/D converter
13:20:43t0mashm... and then? you have a perfect battery indicator?
13:20:49t0masor is there some other use?
13:22:03t0mas(oh and the usb mode idea? is it possible to get in usb mode w/o spinning up the disk? so linux sees the device... writes an error to syslog and I know it was still running?)
13:22:48LinusNt0mas: with the ampere meter i don't have to read the battery, it tells me how much current it is drawing
13:23:17LinusNyes, you can get into usb mode with the disk powered off
13:24:43LinusNgotta go again, cu later
13:25:02 Part LinusN
13:26:35 Join webguest78 [0] (
13:27:12t0mas<LinusN> yes, you can get into usb mode with the disk powered off <−− now just find out how :)
13:38:40 Quit webguest78 ("CGI:IRC")
13:45:05*niobos is off to lunch
13:48:17 Part leftright
13:51:33 Join zezayer [0] (~chatzilla@
13:56:01*HCl yawns
13:56:04HClsleep ish good.
13:56:39 Part Mirfle
14:03:31HClis Slasheri's patch committed yet?
14:03:34*ashridah peeks at the updated patch
14:03:51ashridahHCl: nah, updated version on his website tho
14:05:11Slasheriashridah: there is not yet very big improvements
14:05:53*ashridah cracks whip :)
14:05:56Slasherii try to get seeking of mp3s working and then i will update it again
14:06:23HCl :P
14:06:37*HCl dreamt he got further with the runtimedatabase
14:08:52HClbut i havent
14:09:15HClmostly cause i don't want to make a patch upon a patch thats still being developed
14:11:34 Join LinusN [0] (
14:12:00LinusNSlasheri: don't add more functionality before i commit
14:12:13SlasheriLinusN: ok :)
14:12:29HClwhen will you commit? :)
14:12:30LinusNregarding seeking, you should have a look at mpeg.c
14:12:43Slasheriah, i will do that
14:12:49LinusNi'll hopefully commit later this evening
14:12:55LinusNgotta go
14:12:57 Part LinusN
14:17:16 Join Lost-ash [0] (
14:18:03 Quit zezayer (Remote closed the connection)
14:18:22 Quit ashridah (Nick collision from services.)
14:18:25 Nick Lost-ash is now known as ashridah (
14:28:00 Join dapureplaya [0] (
14:33:16 Join nobby-BRB [0] (
14:33:20nobby-BRBI LOVE YOU ALL :)
14:33:27 Nick nobby-BRB is now known as nobby (
14:33:38dapureplayalol nice. break the silence :p
14:33:42 Nick nobby is now known as nobbeh (
14:33:51 Nick Sucka is now known as Sucka`away (
14:33:51*ashridah bops
14:33:52t0masok... tnx... but I have a gf...
14:33:53ashridahno silence here
14:34:16nobbehrockbox on iriver now plays music with an unofficial build apparantly
14:34:33dapureplayai'm using it right now
14:34:33t0masand Linus will commit it this evening... if he has some time
14:34:35nobbehim charging mine now so i can reflash it with the new bootloader
14:34:50dapureplayawhy don't u just keep it plugged in while yoiu flash it?
14:35:00nobbehyeah, i could do that... :)
14:35:06*nobbeh does
14:35:15nobbehTY for the common sence :)
14:35:25*nobbeh does happy dance
14:35:47dapureplayaomg....gapless is so awesome on my H140 :D
14:36:39dapureplayayou gotta try this out nobbeh. :-P
14:36:59nobbehi will as soon as i untangle this goddamn cable :S
14:37:01ashridahnobbeh: you won't need to flash it to get sound working if you've already got a rockbox bootloader installed.
14:37:20nobbehany point in a reflash then?
14:37:24ashridahunless you're using a partcularly old version that is
14:37:32dapureplayajust copy the files over
14:37:38nobbehi'm using the last CVS one i think
14:37:47ashridahnobbeh: well, recently someone fixed it to allow one to use iriver's 1.65 firmware without breaking ogg
14:37:53ashridahand a few other minor fixes
14:38:16nobbehi have 1.63 with the old loader afaik
14:38:19ashridahbut if you're using what's currently on the wiki, that's safe enough. no point in using a potentially broken CVS build.
14:38:25Slasherinew bootloader has also usb mode. you should reflash if you have older one
14:38:44nobbehi have usb mode in rockbox
14:38:51Slasherinobody should ever build a bootloader from cvs
14:40:03nobbehITS PLAYING!
14:40:14dapureplayalol right
14:40:27dapureplayacalm down
14:40:36nobbehdont wanna :P
14:40:47dapureplayawell i'm sure u dun wanna drop dead either :p
14:41:31nobbehim gonna go get a shower and get dressed. Its 20 to 2 :P
14:41:47 Nick nobbeh is now known as n[o]bbeh (
14:43:57dapureplayaslasheri: I know u only just roughly worked on the track skipping just recently, but i'm kinda getting this thing where i just skip once but it skips like 3 or 4 tracks at a time. Is it suppose to do that?
14:44:44Slasherino, it shouldn't do that. Are you using the latest patch?
14:45:00 Join edx [0] (
14:45:12dapureplayaBy patch you mean?
14:46:23SlasheriI recommend you wait little more while it gets to cvs and then you could use the daily builds from cvs
14:48:36HCln[o]bbeh: you need the new bootloader cause the old one has mirroring of 16mb of ram
14:48:41dapureplayayeah fair enough. I mean it only did it when i first started using it.
14:48:59ashridahyes, without having to merge in a patch by hand (which, btw, looks like it should have cleanly applied, not sure why it didn't)
14:49:35Slasheridapureplaya: ah, ok. That bug was on first versions but i think it's fixed now
14:49:51dapureplayaso i have the first version?
14:50:08Slasheriit might be possible you don't have the most recent version
14:50:12HClam i the only one who hasn't tried Slasheri's patch? :P
14:50:33dapureplayawell i'm using it right now HCl
14:50:49HClmost of the time, i use my iriver as portable harddisk with my music collection
14:50:56HCli only actually play music on it when i'm in my car or otherwise not home
14:50:57HCl :)
14:51:00crwlthe current mp3 implementation seems to be a little too gapless :)
14:51:12HCltoo gapless?
14:51:15dapureplayawhat do u mean by that?
14:51:25crwlwell, there are no gaps between songs even if there should be
14:51:27ashridahyeah, it seems to cut off the end of songs in some cases i find.
14:51:35dapureplayahow much of it?
14:51:50crwlmaybe half a second or a little less, i'm not sure
14:51:54dapureplayawell i tried out gapless for both it's fine. unless i'm deaf.
14:52:02Slashericrwl: Hmm, that is caused only if last song has no gap at end of the song
14:52:37HClis there an option yet to insert gaps?
14:52:46crwlSlasheri, ok
14:52:46Slasherinot yet :D
14:53:43*HCl bites his game.
14:53:55HClcrash. i log back on. dead.
14:53:57crwlthe newest album of SOAD which i, hmm, acquired from the 'net, is maybe about five seconds shorter with current implementation than it should be :)
14:54:17dapureplayadamn i better get back to my assignment. all this talk is getting me excited. bbl guys.
14:54:24 Join preglow [0] (
14:55:22preglowcrwl: crosscheck that with foobar if you can
14:55:46crwlwell, i can't
14:56:17preglowwhy not?
14:56:49crwldon't have a windows machine handy
14:57:10dapureplayaSlasheri: just one more quick Q before i go, has an OGG 'now playing' screen been implemented yet?
14:57:50Slasherino it has not
14:57:54Slasheriwill be implemented later :)
14:58:10crwli wonder if WPS will be upgraded to support custom tags or will just the "standard" vorbis tags be mapped to the existing id3 tags and others be thrown off
14:58:19 Join DomZ [0] (
14:58:44 Nick n[o]bbeh is now known as nobbeh (
14:58:58Slashericrwl: I think we have to change the wps code to support better vorbis and other formats
14:59:08crwlSlasheri, it'd be sweet
14:59:10nobbehwhy does the current database have "album number" but no "album year"?
14:59:22dapureplayaanwyay i'll be back when i finish up this dread of an electronics assignment. due tomorrow. man i hate deadlines. cya
14:59:33crwlthough, just the usual artist/composer/title/trackno/genre etc. tags are what 99% of people use, i guess
14:59:37crwlincluding me
15:00:18Slasherinobbeh: The wps can be customized and i think you can add album year into it if you want to
15:00:34nobbehyeah, but the tag database makes no mention of
15:00:45nobbehwhat if i want to search by year?
15:00:45SlasheriAh, hmm..
15:00:54Slasherii don't know about that
15:01:00 Nick dapureplaya is now known as dapureplayaBUSY (
15:02:55nobbehon the default irirver WPS
15:03:04nobbehwhat are teh 2 lines separated by a dotted line?
15:03:30SlasheriI think that's a peak meter. Not supported yet
15:03:37 Join tucoz [0] (
15:03:47tucozHCl: are you there?
15:03:51nobbehlooks horrible
15:04:33nobbehalso, stuff like volume change lags badly
15:05:15tucozHCl:I noticed something, When browse for albums using the db (i.e. Albums->), albums starting with a number, are sorted before the <> tags
15:05:36tucozlike <no album tag>
15:05:55tucozHCl: and I also got two <no album tag> in there
15:06:14dapureplayaBUSYuh oh. i'm on the shutting down screen and it's still playing music. :S
15:06:29dapureplayaBUSYit won't shut down! lol
15:08:16dapureplayaBUSYoh there it goes. heheh. =P
15:13:12HCltucoz: yea, i saw that.
15:13:37HClits hard to change sorting.
15:13:44tucozi guess
15:13:46HCltwo no album tags can occur
15:13:56tucozHmm, why is that?
15:14:04HClan album is associated to an directory
15:14:10tucozok, I see
15:14:14HClif two albums are in two different directories
15:14:17HCland they don't have an album tag..
15:14:46*HCl twitches at his game. having lost EVERYTHING cause of a fucking crash of the game client.
15:14:56niobosHCl: are albums related to dir's? or is it the songDB-program that makes the relation?
15:15:05HClsongdb program.
15:15:06niobosin the DB-structure I don't see dir's anywhere...
15:15:10tucozWould it be an idea to have the directory as a tag instead?
15:15:21HCltucoz: there's an option for that.
15:15:25HCltucoz: −−dirisalbumname
15:15:56tucozHCl: but is that for all albums or only those where no tags are found?
15:15:56*niobos is back off to his electromagnetism... :-(
15:16:06HClonly for those where no tags are found
15:16:13preglowniobos: i feel for you
15:16:24tucozOk, then it's ok.
15:16:50HClthere's also dirisalbum, which states that all files in a directory belong to the same album
15:16:58HClregardless of their tags
15:17:24niobospreglow: thx
15:17:37tucozok, cool. I just ran it wo dir, and saw the options
15:19:29***Saving seen data "./dancer.seen"
15:19:44tucozHas anyone of you noticed that sometimes (very seldom) when booting rockbox, a loud sound emerges through the headphones?
15:19:45 Quit ashridah ("Leaving")
15:23:30tucozGuess it's some initialization code missing.
15:23:50tucozeither in the bootloader, or when rockbox starts loading
15:24:10tucozbecause it's very early in the boot-up process.
15:31:46HClbleach is cool
15:31:48HCl :/
15:34:05t0masBC F4 13 00
15:34:07HClsup t0mas
15:34:11t0maswhat's that if it's big endian?
15:34:17t0mas(it's a long)
15:35:13*t0mas starts thinking :P
15:35:21t0masB = 11
15:35:38t0masso 1011
15:35:39HClits mostly that i don't know the order of big endian
15:35:43HClaside from that its not too hard
15:35:54t0mas10111100 = BC
15:36:24t0mas4 + 8 + 16 + 32 = 60
15:36:34t0mas+ 128 = 182
15:36:39*HCl frowns lightly
15:36:44HClyou have calculators for that t0mas :P
15:36:52HCl15:31 <+cu2q_> 0xBC = 188
15:36:56t0masyes, but I'm to lazy to get one...
15:37:02t0masthat's better
15:37:08t0maswhat channel? :)
15:37:11HClum. on ircnet
15:37:20HClbut you can query it on ircnet too
15:38:02Slashericonverting hex to binary or vice versa is similar as long as you keep in mind that one hex-character is 4 bits in binary
15:38:10t0masyes, I know...
15:38:55t0masbig endian = bytes in screen order right?
15:39:17t0masand little endian is "reversed"
15:39:19SlasheriMSB bytes comes first in big endian
15:39:42t0masok so this is a negative number?
15:39:58t0masas it starts with 10111100
15:40:57SlasheriHmm, should be if you are using signed values. Keep also in mind that bit order is always the same regardless of endianess
15:41:03HClshould i just give you the number
15:41:11HClits like what, a 5 line c program
15:41:15t0mas-1124855040 ?
15:42:48t0mashm... I don't need the number...
15:43:01t0masI'm debugging some weird thing in my dictionary plugin
15:43:17t0masyes HCl but what if you read it in reverse? :)
15:43:40t0masso 00 13 F4 BC
15:43:51t0masok, that's what I want :)
15:44:18t0masbut in file it's written as BC F4 13 00
15:44:32t0mason rockbox (big endian) that would be -1124855040 right?
15:44:44t0masso I have to reverse it... before writing it to a file...
15:44:56HClthere are C macros for that in the rockbox code
15:44:56t0masso in rockbox it gets read right...
15:44:59HClin database.h, i think.
15:45:18t0masyes, but I'm creating the file...
15:45:25t0masso I can better reverse it at the pc...
15:45:39HClyes, i'm just saying that if you need c code for it, its in database.h O.o.
15:45:44t0masok, tnx
15:47:39niobosQuestion: on x86, pointers are 32bit. On that ColdFire also?
15:48:47niobosk, i'll use sizeof(void*) then
15:48:54preglowbut never assume that
15:48:56preglowalways use sizeof
15:49:54niobosbecause I have a datastructure that contains OR pointers to struct... OR file-offsets...
15:50:18*niobos just realizes the word union...
15:52:56preglowyes, union
15:56:26 Join webguest27 [0] (
16:03:50t0masHCl? this fdprintf(fd, "0x%x", var) you use in dynarec... can I use that for outputting a long too?
16:06:14 Part tucoz ("Leaving")
16:10:46 Join Chamois [0] (
16:11:56 Quit webguest27 ("CGI:IRC (Ping timeout)")
16:16:59 Join tvelocity [0] (
16:27:29 Part DMJC ("Leaving")
16:27:29 Quit preglow (Read error: 104 (Connection reset by peer))
16:28:33 Join preglow [0] (
16:28:51 Join XavierGr [0] (
16:29:18niobosif I want to keep an array of pointers in C, I could do so with void* array[];
16:29:46niobosbut wouldn't that be a waste of space? since array[] itself is already an array of pointers (to pointers)?
16:30:06t0maswhere are you doing it?
16:30:24niobosbut it's a general question
16:30:26XavierGrhi all!
16:30:31niobosHi XavierGr
16:30:40 Part nobbeh
16:30:43t0masniobos: yes, but in a plugin you have a pluginbuffer... so you can "malloc" in there
16:30:47XavierGrDoes anyone knows if we can change the WPS font?
16:31:11t0masXavierGr: you can change the global font...
16:31:32niobosI'll figure it out, thx
16:32:04XavierGrthen I shouldnt change WPS code
16:32:10XavierGrgeai tvelocity!
16:32:22t0masniobos: but afaik an array is just 1 pointer
16:32:28t0masnot an array of pointers...
16:32:34 Join asdsd____ [0] (
16:32:39 Part asdsd____
16:32:46XavierGrI will make some test to see if we can utilize all of iriver's screen in WPS
16:32:49t0masand array[5] is just the first pointer + 5 * sizeof(object)
16:32:58t0masXavierGr: check out the board
16:33:00niobost0mas: yeah, I'm drawing the pointers on a sheet op paper to get it...
16:33:09t0masah ok
16:33:17niobosit always confuses me when I do things like void* array[] and stuff
16:33:59XavierGrok I will look to it
16:34:56t0mashm... I need some core dev...
16:35:08XavierGrnice! t0mas!
16:35:18tvelocitywps uses all of the screen on my iriver.... I just used a bigger font:P
16:35:19t0maswant to implement that hack for real... using the bmp code I wrote some time ago
16:35:28t0masbut I don't know what memory I can use for it...
16:35:36XavierGrdid you change the code of wps.c?
16:36:04XavierGrthen I shouldnt touch it. bravo...
16:36:12t0masyou don't have to
16:36:18t0masit should be able to read a .bmp file
16:36:22t0masand display that as a background
16:36:57 Join Mouss [0] (
16:36:58 Join zezayer [0] (~chatzilla@
16:37:07XavierGrI will better look some of snake.c code to adapt for iriver
16:37:17 Quit thegeek_ (Read error: 54 (Connection reset by peer))
16:37:19t0masisn't that already done?
16:37:24t0masor you mean snake2?
16:37:53XavierGrsnake 1 no, there is an invisible wall and then you die!
16:37:55t0mas <−− there's a patch for snake
16:38:13t0masHCl wrote it
16:38:17XavierGrnice again
16:38:20t0masbut didn't commit it...
16:38:29*t0mas prods HCl: why isn't it in CVS?
16:38:39XavierGryeah they must commit it to know what to change
16:38:52t0masit's listed here:
16:40:36XavierGrthere are not many things to touch. That's good. Only a couple of games.
16:40:39t0masSlasheri: you're using the normal wps code in wpc.c now right?
16:40:54t0masXavierGr: can you write perl?
16:41:30XavierGrunfortunately no, I only now a little c and visual basic :(
16:41:37Slasherit0mas: yes, i have made no changes to the wps code
16:41:39tvelocityi know some perl & python if there's anything i can do
16:41:56t0masyes... for the dictionary stuff... we have a conversion script for wordnet sources
16:42:20t0masbut there is a widespread format, the dict format... so if someone could write a dict2rdf perl script...
16:42:57tvelocityi isn't there allready a C tool for that?
16:43:10t0masno, that rdf2binary
16:43:44t0masthe C tool converts the output of the perl scripts to the format used in the plugin... for fast searching
16:44:19t0mastvelocity: and have some dict files...
16:44:26t0masand that's what we need
16:45:03t0masso short: word[tab]description, alls words in lowercase and the words should be sorted
16:45:27t0mas(and to make it easier: tools/ is an example)
16:46:02 Quit MoosCamaro (Read error: 60 (Operation timed out))
16:46:29tvelocitydoesn't sound very difficult
16:46:42t0masnot if you know perl :)
16:47:15t0masI have to learn it somewhere in my vacation... as it's really irritating to have some perl scripts... and not being able to edit them
16:47:38tvelocityi do know *some* perl... but i'm not as fluent in perl as I am in python...
16:48:29t0masI only know some types of basic... php and C
16:48:52t0masand I don't know perl and python because I use php/bash scripting for all my shell trouble :)
16:50:24tvelocitywhat for is the dict being used?
16:50:45t0maslooking up words...
16:51:08tvelocityerr, yeah, but why do you need looking up words in rockbox?
16:51:40tvelocityfor spell checking, or for something like T9 maybe?
16:51:41t0massome people have dict readers on their PDA... I have an unusable one on my phone... so that's why I created one for rockbox too..
16:52:06tvelocityoh... i see
16:53:37XavierGrdo we know in which corner of the screen is the (0,0) pixel cordination?
16:53:46t0masleft top
16:55:46tvelocityso t0mas we need a perl script that converts a standard dict like the ones on, to a word[tab]description format... that's all?
16:56:20t0masyes that's all
16:56:41t0maswe've split it like this to make it easy for people to write these scripts...
16:56:57t0masoh... and don't forget to convert it to lowercase... and sort it...
16:57:01t0masbut that's easy too...
16:57:45tvelocityi'll take a look then
16:57:51t0mask tnx
16:58:09tvelocitywhy perl anyway?:P
16:58:17t0masbecause amiconn likes perl :P
16:58:29tvelocityhm... ok
16:59:03HCl :P
16:59:27HClt0mas: i dunno, should i commit it?
16:59:46XavierGryeah commit
16:59:46t0masif there's no good reason not to do it...
16:59:47HCli had gotten the impression it wasn't good enough because people hadn't committed it
16:59:53HCli made it before i got cvs access
16:59:59t0masah ok
17:00:51t0masHCl: does it still work on archos with the path?
17:01:03HClcan you try on a sim?
17:01:22t0masno... only have iriver sim
17:01:32t0masbut maybe someone here with an archos can?
17:02:05XavierGrhmm somewhere on the wiki we must say that sim screen is true sim only when the size of the sim window is not modified
17:02:09XavierGrI can test it
17:03:18 Quit preglow ("Lost terminal")
17:03:33 Join preglow [0] (
17:03:40t0masah preglow found back his terminal ;)
17:03:48t0masXavierGr: what do you mean?
17:04:22preglowwhat happened?
17:04:30t0masyou lost your terminal
17:04:36preglowi noticed :P
17:04:38XavierGrwell I was mading some test in the sim about the calculator
17:04:39t0mastime to clean your room ;)
17:04:42t0mas* preglow has quit ("Lost terminal")
17:05:11XavierGrand whenver I pressed the maximize button of the windows the screen seemed the same but it was not
17:05:30t0masoh ok...
17:05:40XavierGrSome of the pixel sizes are changed so it seems that you have some pixels empty
17:05:41t0masyou can resize it... but not with the max butoon
17:07:09XavierGrI had a serious confusion with that
17:07:37XavierGrI wanted to test it pixel by pixel so I maximized and the screen was slightly different
17:07:54XavierGrthen I thought that it has to be the changes I made but it wasnt this
17:09:25XavierGrwith the maximize of the screen there are some slight changes in pixel orientation
17:10:51XavierGrso HCl do you want me to test your patch for the archos
17:11:12XavierGrbut I think that all that have a cygwin enviroment can do that in the sim
17:11:56 Quit dapureplayaBUSY ()
17:12:13HClyea, please
17:12:27HCli mostly just increased the allowed screensize iirc..
17:12:33XavierGrI have to download the patch right?
17:12:33HClrather than scaling it
17:12:53XavierGryes but did you made an if statement acorrding to the target
17:13:11HCli don't remember, he.h
17:13:45XavierGrdo you have the snake.c it would more easy for me to apply changes (except if you made changes in other files too)
17:13:56HCleh, no, it should be in cvs..
17:14:10HCli'm a bit busy at the moment, can't really help, sorry
17:14:14coobis there a badger.c? and a mushroom.c?
17:17:07 Join bagawk [0] (~Lee@bagawk.user)
17:17:11XavierGrwell I meant if he had the modified snake.c file not the original. Obviously if the patch hasnt made CVS then this c file is the original. I will get and apply the patch
17:19:31***Saving seen data "./dancer.seen"
17:21:13 Quit zezayer (Remote closed the connection)
17:21:35 Join jamesshuang [0] (
17:21:49jamesshuangHey everyone, just letting you know about a bug from mysticforums
17:22:41jamesshuangseems that with the latest slasheri patch, trying to view a playlist gives an IllInstr at 32F40196
17:23:13HClmarkun reported that before..
17:23:35crwlit's related to plugins not yet working while playing, yes?
17:24:10Slasherijamesshuang: thanks, that is a known problem and will be fixed later
17:24:16jamesshuangahh.... I see.. i knew the plugins crashed while playing, but I didn't know the playlists were generated by plugins
17:24:58Slasherii think the view playlist does nothing to do with plugins (i don't know yet why it will crash)
17:25:29HCli don't quite understand how playlists work at all..
17:27:16 Join zezayer [0] (~chatzilla@
17:28:55XavierGrHCl: Your patch works perfectly fine for all the archos that shares a 112*64 screen! Commit as soon as possible :)
17:29:10preglowlinus said he'd commit it today, yes?
17:29:44XavierGryes preglow at last the moment you was waiting!
17:45:45 Join muesli- [0] (
17:48:05muesli-is there another possibility to get access to this without using
17:49:07bipakfor read only, there is a log in the rockbox site :)
17:49:24jamesshuangthere's always the webclient
17:52:46 Join zezayer_ [0] (~chatzilla@
17:54:48 Join thegeek [0] (
17:57:39 Quit zezayer (Read error: 60 (Operation timed out))
17:57:39 Nick zezayer_ is now known as zezayer (~chatzilla@
18:03:24 Quit Shagnar (Read error: 110 (Connection timed out))
18:05:24 Join webguest29 [0] (
18:10:40bipakhmm, where can i buy motorola coldfire C's? :)
18:10:48bipakcant find anything on google
18:11:48bipakfound :)
18:12:20preglowwhy'd you want that???
18:12:34bipakjust wanted to know the price :>
18:12:45preglowdunno, 10$ ?
18:12:58bipak20 Euro in germany
18:13:12preglowmore sane, probably
18:13:46bipaki thought they would be more expensive
18:14:28preglowembedded processors aren't very expensive
18:16:42 Quit webguest29 ("CGI:IRC")
18:17:13bipakyes, i do some hobby stuff with an Atmel AVR C, thats costs about 3 Euro :)
18:27:23 Join Strath [0] (
18:28:39preglowi've uses a lot of avrs myself
18:28:43preglowused, even
18:29:12niobosfor the songDB thing in C: what should I do?
18:29:30niobosamiconn said it should run on low memory (=the device itself)
18:29:51niobosso during the DB construct, (at least) some items should be in file instead of RAM
18:30:08niobosbut should I do a black/white: either all in RAM or all on disk?
18:30:22niobosor play a bit: these entries are in RAM, these on disk?
18:31:09preglowthat's up to you
18:31:10niobos(I prefere the black/white, the entire array is either on disk or in ram)
18:31:20preglowentire array on disk will be very, very slow
18:31:30niobosI know...
18:31:37preglowa hybrid would be best
18:31:44preglowyou can't have it all in ram
18:32:00preglowdatabases can probably easily be above 2 megs
18:32:11niobosBut "some" on disk means an extra overhead to remember what is in ram and what on disk
18:32:41preglowwell, i have no idea of how this thing is built up, so can't help you there
18:33:05niobosI don't think it'll be that slow...
18:33:14niobosonly sorting will take longer
18:33:32niobosbut anyway it will as soon as some are on disk...
18:33:48niobosI might benchmark it, see what it does on my x86 machine and extrapolate that
18:34:08 Nick Mouss is now known as MoosCamaro (
18:34:27*niobos gets back to programming
18:37:21XavierGrwhat are you programming niobos?
18:38:37 Join Stryke` [0] (
18:39:35preglowniobos: you can't extrapolate from that
18:39:51preglowniobos: your os probably has a very nifty disk caching mechanism
18:41:47 Join muesli__ [0] (
18:45:41 Quit bagawk ("Leaving")
18:46:20bipakthe link is down
18:48:26 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
19:00:11 Quit muesli- (Read error: 110 (Connection timed out))
19:00:25jamesshuanghey guys, I just saw this on mysticforums... seems that again with the slasheri patch, rockboy doesn't seem to work anymore
19:01:16jamesshuanggoes through the init process, seems to work, but then crashes on a blank screen
19:01:32tvelocityi think they allready know that
19:02:51 Join muesli- [0] (
19:03:17jamesshuanghmm... does slasheri's patch really change the plugin structure THAT much? I'm not touching any audio, fresh boot, then immediately attempting to load a rom crashes
19:09:58 Join Mumble [0] (
19:10:07tvelocitywell i'm not a dev and don't know, but shit happens all the time... i'm sure they'll look into it when the time is right. we can't have everything at once right? ;)
19:11:18Slasherijamesshuang: Hmm, it will crash even when you are not playing audio?
19:11:34Slasherijamesshuang: Could you send a rom to me and i will test it
19:15:50zezayerMario rom here:
19:16:19zezayeror pokemon
19:16:24tvelocityi really hope nintendo is not reading the logs :P
19:17:29zezayerif they ask, it wasnt me ;)
19:18:15Slasherithanks, i will test those
19:18:50zezayerthey both worked on the unpatched versions of rbx
19:18:52 Quit muesli__ (Read error: 110 (Connection timed out))
19:19:05Mumbleis there a way to stop the sound playing (when playing gb/gbc roms) so the CPU can process the graphics faster? or it doesnt work like that?
19:19:21Mumble(games like pokemon seem to run quite slowly) or thats the way it is?
19:19:33***Saving seen data "./dancer.seen"
19:20:09preglowhcl says he'll optimise it sooner or later
19:20:14preglowthere's a _LOT_ to be done for it
19:22:01XavierGrsnake 2 adjustment for the iriver is very strange
19:22:33XavierGrthe archos screen ratio is 1.75 while iriver's 1.25
19:23:08XavierGrThat will be a problem if we want to utilize the whole screen and keep current levels
19:23:36Mumblealso are gameboy screens bigger than the H1X0's? as some games seem slightly squashed or distorted.?
19:23:46XavierGrthe board of the game is 28*16
19:24:14XavierGrwith a multiplier of 4 this goes to 112*64
19:24:35zezayeryer GB screens are higher, turning hold on/pff changes how the screen is adjusted
19:24:58zezayeroff , typo
19:24:59preglowgameboy screen is sligthy taller
19:25:04Mumblekk, sorry
19:25:45XavierGrare you talking to me zezayer because I am not talking about GB emu
19:25:55SlasheriHmm, rockboy crashes due to some pcm output structure change
19:25:58Slasheriinvestigating that
19:26:08zezayerno mumble's comment above
19:26:53 Quit DomZ ("CGI:IRC (EOF)")
19:26:53XavierGranyway as I was saying 28*5=140 and 112*6=168
19:27:06bipakdo i need to do any special to get rockbox to play a mp3 file? :> im using the cvs rockbox.iriver file
19:27:22preglowneed a patch
19:27:26XavierGreither we will not utilize all the screen or we will need float values!
19:27:40bipakah, i thought it's in cvs allready :)
19:27:43tvelocityor you can wait for the patches to be commited to cvs
19:27:46preglowit's due today
19:28:07tvelocityor someone could send you a binary
19:28:23bipakyeah, would be nice :)
19:28:32zezayerbipak read this:
19:28:52zezayerthere are some patched fw's there to download
19:29:23zezayer[ post 788 ]
19:30:09t0masXavierGr: what?
19:30:19t0mas"either we will not utilize all the screen or we will need float values!" <−− ??
19:30:35XavierGrfor the adaptation of snake2 into the iriver screen
19:30:59XavierGrsnake 2 uses a 2d board of 28*16
19:31:27XavierGrthus when drawing to archos the coder multiplie the values by four to get the whole screen
19:31:35t0masthe comment in the todo page is that you need to change the maps isn't it?
19:32:12XavierGrwell I was said not to do that
19:32:17t0masah ok
19:32:23t0masthen leave out a little part of the screen?
19:32:34t0masor put a score indicator on top?
19:32:40tvelocityyou could fill the screen with some statistic etc
19:33:00XavierGrI will need float numbers for multipliers if I want to have a significant size gain
19:33:07bill2or3looks like I won an H340 on ebay
19:33:28XavierGr28 * 6 = 168
19:33:35bipakzezayer: works :D
19:33:44t0mas28 * 5 = 140 ?
19:33:55t0masI see the problem
19:34:07tvelocitywhat's the resolution of the iriver LCD?
19:34:11XavierGrand 28 * 5 = 140 (too little gain) can I type 28 * 5.7
19:34:39zezayergdgd bipak - enjoy
19:34:47 Quit muesli- (Read error: 110 (Connection timed out))
19:34:47t0masXavierGr: that will be 160
19:34:48XavierGrbut I dont think I can type that can I?
19:34:59t0masI guess you can... but it will be slow...
19:35:06XavierGrwhy slow?
19:35:19t0masgcc implements some floating point stuff then
19:35:22t0maswich is slow..
19:35:41XavierGrI will test it anyway
19:35:51t0maswhy not use fixed point math?
19:35:55t0masif you really care...
19:35:56tvelocityyou could use *6, but every 3 blocks use *5 instead?
19:35:59bipakwhy does it fade in/out when pressing the pause key? isn't that usefull i think
19:37:25tvelocitybipak, that is configurable
19:37:26XavierGrHaha the sim runs fine but another problem occured.
19:37:39XavierGrWhere can I upload a pic to see?
19:37:42t0masXavierGr: width = (long) (28 << 8) * (5.7 * 256) >> 8;
19:37:47t0masgives a fixed point width
19:38:06t0masshifted 8 positions to the left...
19:38:27t0masbut it's a little complicated for something like this...
19:38:44XavierGrwell porting is almost done but I will show you what is the problem...
19:38:56t0massorry: width = (long) ((28 << 8) * (5.7 * 256)) >> 8;
19:39:05t0masnot sure if the first one works :)
19:41:39 Join muesli- [0] (
19:42:34XavierGrhere it is. Weird eh?
19:43:33t0masit's not supposed to be dashed is it?
19:44:36CoCoLUSjust tried playing a few oggs...
19:44:57XavierGrt0mas:no it's not. Well I guess that the 5.7 is to blame
19:45:12CoCoLUSand the first few seconds of a track always skip
19:45:14t0masyou get 1/2 pixels...
19:45:19t0masand you can't draw them...
19:45:49t0mashow much height do you have left when you use 5 ?
19:45:50XavierGrwill your suggestion solve this?
19:46:06t0masdon't think so... the lcd still is setup in pixels...
19:46:15tvelocitywhat about my suggestion?
19:46:15XavierGr28 pixels
19:46:19t0masso in my way you can just paint your dashes a little faster :)
19:46:26tvelocityuse *6, but every 3 blocks use *5 instead?
19:46:30t0mas160x128 right?
19:46:42t0mastvelocity: that would make the code more complicated...
19:46:47XavierGrwell I could do tvelocity suggestion but I will have to rewrite all the code
19:46:54tvelocityindeed, but it would solve the problem...:P
19:46:56t0masXavierGr: try with 5 and center it?
19:47:10t0masand check if you can put some nice statistics here and there...
19:47:18t0masto make it look like it's ment to be this way ;)
19:47:29XavierGryeah like velocity m/s haha
19:47:48t0masno.. but score? and running time or something would be nice?
19:48:35tvelocitya graphic of a snake would be w00t :P
19:48:55t0maslol, that might be a good idea...
19:49:47t0mas28 * 5 = 140, and 16 * 5 = 80
19:49:54t0masthe screen is 160x128
19:50:24t0mas16 * 7 = 112
19:50:48t0masso we have left 8 pixels on top... 8 on bottom and 10 left and 10 right?
19:51:00XavierGrbut then I will spoil the aspect ratio
19:51:09tvelocityasymetrical blocks on a snake game would not be really intuitive
19:51:19t0masok, then 140x80
19:51:23zezayertry it??
19:51:28t0masmake a bar of 20 px on top... with the score
19:51:37t0mascan someone paint a snake for me? :)
19:51:42t0masjust bmp?
19:52:15t0mas8 pixels left, 8 pixels right... 8 pixels bottom? and then 40 pixels score / image on top?
19:52:23 Part Mumble
19:52:39XavierGrso you think *5 centered will be better?
19:52:48t0masand then paint an image around it
19:53:06XavierGror I could do *5 with various information on the side
19:53:22XavierGr(to avoid centering me lazy :P)
19:53:35t0masno... center it...
19:53:41t0masI like the image idea...
19:54:49zezayerwhat size snake image do u want?
19:56:01t0maswell... it has to be around the play field
19:56:33t0masyou have 10 pixels on the left... 10 on the right (10 + 10 + (28 * 5) = 160)
19:56:59 Join courtc [0] (
19:57:00XavierGryeah but t0mas now that I test it *5 is really small for the iriver
19:57:31 Join Musicmad [0] (
19:57:33t0masthe old version was playable... to with some graphics... it shouldn't be a problem
19:57:45t0maszezayer: and say 10 pixels on the bottom too?
19:57:53 Join C-Keen [0] (
19:58:21t0masand then keep a nice rectacle on top open for score info?
19:58:29t0mas(no problem to put that inside some image)
19:59:23XavierGrwhat do you think?
19:59:26C-Keenafter openning up my archos 6000 player for the first time my harddisk seems to only click when booting up. Is this an indicator for bad soldering or a dead disk?
19:59:41t0masXavierGr: with some image and score info... I think it's good...
19:59:51t0masbut maybe try the 6 scale for vertical?
19:59:57t0mastest if it makes it unplayable?
20:00:43t0maszezayer: do you think you can make some nice backgroundimage for that?
20:01:03zezayerwill try, no promises
20:01:20t0masno problem :)
20:01:42zezayerso for left and right 112*10 ?
20:02:10t0masthe image width should be 160
20:02:12t0masheight 128
20:02:22t0masand you can use 10 px left and right....
20:02:36t0masand bottom 10 is ok too
20:03:11t0masthen you have 128 - 80 - 10 = 38 pixels left for the top bar... for some image... and maybe score info inside the image?
20:03:32XavierGrwell we have 16*5=80 128-80=48 pixels vertical
20:03:43CoCoLUShas someone done some battery tests with the latest firmware?
20:03:50t0masXavierGr: yes, 10 on bottom
20:03:59t0masso 38 left for nice top image and score info
20:04:03XavierGrand 28*5=140 160-140=20pixels width
20:04:04t0masCoCoLUS: no
20:04:13t0masXavierGr: again, 10 - 10
20:04:24t0mashm.. better idea...
20:04:38zezayerso 10 * 128 for the sides and 10*160 for the bottom
20:04:40t0maszezayer: the play field is 140x80
20:04:54t0masand the screen is 160x128
20:05:06zezayerah k
20:05:08t0masso just leave out a field somewhere... of 140x80
20:05:21t0masand maybe some place in the top for the text
20:05:32XavierGrin the center I suppose
20:06:21XavierGrthought looking at the previous picture I think that If we can make the gaps dissapear it would be better
20:06:33XavierGrand add the icon on the bootom of the screen
20:06:52zezayershali wait then
20:08:29t0masXavierGr: you would still have a weird walking snake...
20:08:50CoCoLUShey it's a snake, they do walk in curves ;)
20:08:50t0masbecause with the idea of tvelocity it will walk some parts faster than other...
20:09:34XavierGryeah you are right :(
20:09:48XavierGrthen I will try to center it
20:14:53 Quit muesli- (Read error: 110 (Connection timed out))
20:15:13XavierGrcenter it both vertical and horizontal?
20:15:21niobosHCl: are you on?
20:15:47HClohno! its me
20:15:48HClyes o.o
20:15:51HClonly just.
20:15:53HCli just got home
20:16:27niobosthe DB-sorting: can I use stricmp(), or could there be non-printable chars in the names?
20:16:57HCldoesn't stricmp do non printable chars as well?
20:17:05niobosdefinitely not \0
20:17:09HClyea, okay.
20:17:18niobosso jsut use stricmp then
20:17:18HClwell. you don't have to sort on \0
20:17:35HClthere's a pretty neat tag library in c++ out there
20:17:37XavierGrt0mas:are you here?
20:17:44HClmight want to look it up, it should be a pretty nice backend
20:18:05t0mas3 channels screaming... and I'm trying to learn some management stuff...
20:18:08t0masas I have a test tomorrow
20:18:19XavierGrI am sorry then
20:18:24niobosHCl: do you know the name?
20:18:30t0masXavierGr: no problem, and yes
20:18:37HClhold on, i'll look it up
20:18:47t0masbut wait for zezayer
20:18:51XavierGrDo you think that centering will change the orientation in archos targets?
20:18:58t0maslet him paint some background... and fit the board in there?
20:19:04niobosoh, btw: can anyone tell me how to make irssi BEEP when someone calls me?
20:19:18t0masI think we'll have to do those things in #Ifdef's
20:19:19XavierGrwe should implement a big ifdef code for the iriver
20:19:21t0masfor iriver only
20:19:28t0masok, we agree ;)
20:19:38 Join Tangleding [0] (
20:20:15TangledingHi HCl
20:20:20HClhello :)
20:20:36XavierGrI will write it first for the iRiver and then i will adjust the ifdef's
20:20:41XavierGrhi Taang
20:20:43Tangledinghum anyone knows if now Slashey patch is comitted in CVS,
20:20:48crwlniobos, /set beep_msg_level hilight and then set your terminal to beep when it receives a beep signal
20:20:48TangledingHi XavierGR
20:21:08XavierGrNot yet
20:21:08t0masTangleding: subscribe to the cvs mailinglist...
20:21:09preglowTangleding: no, it might be commited later today
20:21:10TangledingWant to know if i can edit the fornoobs sections
20:21:12t0masor check the homepage
20:21:17Tangledinghum okay
20:21:20niobosthx crwl
20:21:24Tangledingso i guess i should wait
20:21:26t0masthen you'll be the first to know :)
20:21:55TangledingI ccan add a usefull link for frenchies
20:21:55XavierGrTangleding: But the snake plugin is commited! Yay! ;P
20:22:08t0masand the dictionary fix too ;)
20:22:17XavierGrnext is snake 2!
20:22:22MoosCamaroHi Tang
20:22:26TangledingHi Moos
20:22:38MoosCamaroca va?
20:22:44TangledingCa va oui
20:22:51Tangledingoups sorry in private
20:23:47nioboscould anyone say my name to test that beep-stuff?
20:24:01tvelocityniobos, niobos niobos niobos niobos
20:24:09niobosok, tvelocity you win
20:24:10HCl :P
20:24:50tvelocityrockbox shows average bitrate! pwnz! :D
20:24:58 Nick Sucka`away is now known as Sucka (
20:25:46HClit does?
20:26:06niobosHCl: for the database: should all names end in \0? or do we save that 1 byte?
20:26:08TangledingRockbox does?
20:26:16Tangleding(sorry really)
20:26:24HClall names have a constant length..
20:26:45HClthey *should* be 0 terminated, but i think with the current implementation its not required...
20:26:47niobosso don't all end in \0
20:26:52Tangleding(was a jokes HCl)
20:26:53HCleither that or i forgot that in the java version
20:26:54HCllet me check
20:27:17TangledingI've some troubles with java on my desktop
20:27:24HClthats a bug in my java implementation
20:27:30HClall strings should be 0 terminated
20:27:33HCllet me fix this...
20:27:53niobosok, so if the longest item is "test", there should be 5 bytes reserved for all
20:28:00t0maszezayer? have some raw idea for us?
20:28:06HCl5 bytes. plus 32 bit alignment
20:28:08HCl= 8 bytes
20:28:14t0masso we know where to put the playfield?
20:28:27HClall text is 32bit aligned
20:28:28nioboshmm... that alignment.... I forgot
20:28:42HCli highly suggest you look at the sources of both the perl and java implementation
20:28:47HClno need to reinvent the wheel
20:28:47HCl :)
20:28:51niobosI know
20:29:02niobosbut I DO need to reinvent A LOT, since C != OO...
20:29:06HCl :/
20:29:21*HCl goes to fix his java implementation
20:29:24niobosbut it's ok, I'm making progress
20:29:36*niobos is away: phone
20:30:44XavierGrt0mas:If I center it both horizontal and vertical it will be 10 - 10 horizontal size in eache side and 24 - 24 vertical size in each side
20:31:02XavierGrso zeyzayer did you got it right
20:31:39t0masXavierGr: bot 10 - 10 and 38 - 10 looks nicer I think...
20:31:47t0masas we van fill the 38 in top with score info?
20:32:57XavierGrso I will change y how much by 10?
20:33:15t0mas38 - 24 = 14
20:33:28t0masso from the center position.. move it 14 px down
20:33:49XavierGrwait a sec...
20:34:16XavierGrwe have 16*5=80 128-80=48
20:34:24XavierGr48 free pixels vertical
20:34:37XavierGrhow many pixels you want down
20:34:52XavierGrand 10 for the score?
20:34:58t0masoh wait
20:35:16XavierGrwith 24 it is centered vertically
20:35:50t0masI'll paint it...
20:35:54t0masas this gets confusing
20:36:25tvelocityshouldn't the alignment stuff be decided from whoever designs the bitmap?
20:36:52tvelocityso he has more freedom on what he can do
20:36:54XavierGrt0mas: I will run a test now from the sim wait
20:37:16t0mastvelocity: yes, it should
20:37:22t0masbe we don't like waiting ;)
20:37:30bill2or3is the iRiver H3xx remote worth getting?
20:37:45tvelocitylol :D
20:37:51zezayerguys ive got people round 4 dinner, so let me know the dimensions later.
20:38:12 Nick zezayer is now known as zezayer_away (~chatzilla@
20:38:43t0masXavierGr: margins: left, right, bottom: 10
20:38:44t0mastop: 38
20:38:50Tangledingthe H3xx remote isn't supported well by H1xx
20:39:04t0masso move the field right by 10px and down by 38px
20:39:13Tangledingwill need a patch for Rbx as for iRiver fw
20:40:03XavierGrhere it is horizontally centered and vertical 10 pixels down
20:40:40bill2or3tangled, I mean for an H340.. just wondering if I should get a remote at all.
20:40:44XavierGrso 38 down and closed right?
20:40:50Tangledingoh okay sorry
20:40:53Tangledingi misunderstood
20:41:15XavierGrt0mas got it
20:41:26bill2or3do you have a remote?
20:42:09HCldoesn't the 3xx series come with a remote?
20:42:48Tangledingnon LCD one HCl
20:43:11t0masXavierGr: hm?
20:43:31t0masXavierGr: that picture is ok... just move it a little down
20:43:38t0masthe horizontal centering is ok
20:43:49 Join Mr_Wik [0] (~Mr_Wik@
20:43:58Mr_Wikhi all !
20:44:26Tangledinghi Wik
20:44:41Mr_Wikhi Tang !
20:47:15zezayer_awayonly back 4 a min, does it have to be black and white or can it be 4shades?
20:47:56t0masblack and white...
20:48:03t0masgrayscale sill isn't working...
20:48:13Tangledingwhat B&W?
20:48:22t0maszezayer_away: have 10 seconds?
20:49:30t0massave that one...
20:49:34t0masand use it as a template?
20:49:52 Join Aison [0] (
20:49:54t0masthe game part should be left as is... and you can paint everything else..
20:55:05 Join muesli- [0] (
20:59:41XavierGryes but there is also some gap in the sideways not only the top
20:59:43 Quit muesli- (Read error: 54 (Connection reset by peer))
21:00:20XavierGr10 pixels gap left and and another 10 pixels right look here:
21:00:34Slasherii fixed the rockboy issue, it was kinda tricky..
21:00:41XavierGror do you think it's too small to draw anything?
21:00:48Slasherii update the patch soon
21:01:03t0maswith color...
21:01:23t0masthe blue part can be painted
21:01:29XavierGrok now it is nice
21:01:33t0mas(the other one wasn't that clear... I know :))
21:01:42XavierGrgot it zezayer_away?
21:01:56t0masyes, he got it... but he's eating :)
21:04:00t0masXavierGr: can you send me the code? :)
21:04:41HClgod, i love what the belgian minister of external affairs said about balkenende
21:05:08XavierGrt0mas okay tomas wait a bit to fix the center code!
21:05:17niobosHCl: I know... I'm belgian
21:05:29HCl :D
21:05:33HClhe's so right XD
21:05:35tvelocityhm, just had a crash iwth RB...
21:05:46niobosI agree
21:05:56t0masHCl: he looks like hp? that?
21:05:59tvelocitywhat did he say?
21:06:02t0masthen I must agree with you :)
21:06:05niobost0mas: yep
21:06:10t0mastvelocity: that he looked like harry potter
21:06:27tvelocityand who is he? any pics?:P
21:06:37tvelocitythat balkenende guy
21:06:47HCl :P
21:06:49HClhold on.
21:07:06HCl :P
21:07:26HClwhy on earth did the dutch people vote for him anyways
21:07:34 Quit Tangleding ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
21:07:34t0masjup... don't know
21:07:35HClhe's one of those extreme christian people.
21:07:51t0maswell... not that extreme... we have one even worse...
21:08:20HCl :p
21:13:48XavierGrt0mas: What sort of Info do you want to implement
21:14:01t0masand time since last killed?
21:14:18XavierGrtime since last killed?
21:14:47XavierGrwhy do we need such a thing
21:14:57XavierGror do you mean overall play game
21:14:58t0masto fill space? ;)
21:15:14t0masis there a score counter?
21:15:23XavierGryes there is
21:15:24t0masthen that would be nice to put somewhere...
21:15:33t0mascheck where it's updated?
21:15:49t0masand put it on screen somewhere? :)
21:15:50XavierGreh it would be good to write how many apples he has eat
21:16:20XavierGrTo the bottom of the screen eh?
21:16:58t0maswhy the bottom?
21:17:05t0maswe have 38 pixels in the top...
21:17:08XavierGreh yes I am sorry
21:17:14XavierGrI got confused
21:17:15t0masand zezayer_away will leave some space for it :)
21:19:35***Saving seen data "./dancer.seen"
21:21:46XavierGrwell I have to go for dinner now. I want to eat something. See you later.
21:21:55 Nick XavierGr is now known as XavierGr|Dinner (
21:24:48 Join [solid] [0] (~solid@
21:25:06t0mashave a nice dinner
21:26:10*HCl grins.
21:26:24*HCl played hakkuhbar - gabbertje really loud cause one of his flatmates shaved his head ;p
21:29:34 Join muesli- [0] (
21:33:44preglownever heard that
21:34:15t0masdon't do it then preglow :P
21:34:33t0masit's the skinhead jogging suit jumping music here...
21:36:05preglowthe 'gabber' part sounds so promising :P
21:36:51tvelocityi didn't know that gabba still exists:P
21:37:27preglowoh, it does
21:39:58niobosHCl: for the sorting: you sort the Objects
21:40:41niobosBut I might as well sort the entries in the binary-file-format, since the name-field is always the first one
21:40:46niobosor shouldn't I?
21:43:08 Nick XavierGr|Dinner is now known as XavierGr (
21:43:12XavierGrI am back
21:43:15HCli sort the objects, and the sorting is defined by the objects compareTo methods
21:43:26HClwhich i've defined as case insensitive name compare, for most of them
21:43:36niobosk, but since I don't know if the object is goint to be in memory or in file
21:43:48niobosit is easyer to first make them equal-length
21:43:55niobosthen convert them to the file-format
21:44:02niobosand then sort them
21:44:10HClsorting on disk?
21:44:18niobosnot realy
21:44:53niobossorting = switching pointers around
21:45:01niobosso it doesn't matter if they're on disk or ram
21:45:25HCltemporary file?
21:45:47preglowsorting exclusively on disk will be so very, very slow
21:45:54niobosI know
21:46:11niobosbut sorting in RAM does require an amount of ram that the device might not have
21:46:20preglowis there no way to do it in pieces?
21:46:22niobosso I prefer slow sorting to no sorting at all
21:46:31preglowi think amiconn mentioned using a merge sort for something like that
21:46:32 Quit muesli- (Read error: 113 (No route to host))
21:46:43HClyea, you can do it in pieces on disk.
21:46:51niobosmerge-sort would do...
21:46:53preglowthen that's what you'll do
21:47:03preglowyou can't access the disk for every sort compare
21:47:09preglowthat'll be prohibitively slow
21:47:10XavierGrhmm high score in snake2 is lost whenever you exit the plug-in
21:47:12niobosbut my question is: should I (merge)sort on the objects?
21:47:19niobosor can I do it on the final file-format?
21:47:28HCli don't think it should matter
21:47:33HClas long as the result is sorted.
21:47:34XavierGrwe should make it create a file to store high score
21:47:52nioboswell, it doesn't matter, as long as the first field is the one you want to sort
21:47:59niobos(which is the case in V3)
21:48:02t0masI'm working on a gui tool... rockbox for noobs
21:48:06niobosbut I don't know if you plan to change it
21:48:30t0masis it ok to copy your firmware patcher code?
21:48:34HCli'd prefer if you wouldn't rely on it.
21:48:41HClit would maintain code flexibility
21:49:04HClundocumented requirements like that is what makes code unmaintainable
21:49:17niobostrue, but doing merge-sort will be more difficult then: since I'd have to parse the file into fields, sort and put it back into file
21:50:16niobosOk, I'll implement it in the "general" way
21:51:03preglowt0mas: of course it's ok
22:07:08XavierGrdoes anyone knows what exactly lcd_getstringsize does or where I can find information about it?
22:07:28HClit finds the size of a string, in pixels.
22:07:32HClwith the current font
22:08:07XavierGrI thought so but what are the arguments that needs?
22:09:41t0masXavierGr: see plugin.h
22:10:07t0masint lcd_getstringsize (const unsigned char *str, int *w, int *h);
22:10:32t0masso you have to pass it pointers to a string and two ints... and it will fill in the ints with the size of your string
22:11:05XavierGrwhy 2 ints?
22:11:21XavierGrthe size needs 1
22:11:38t0maswidth and height
22:11:44t0masint width, height
22:11:51t0maschar *string = "something";
22:12:11t0masrb->lcd_getstringsize (string, &width, &height);
22:12:15XavierGrso I need to have declared 2 pointer ints before i call it right?
22:12:20t0masthat would set width and height to the size of string
22:12:58t0mas(so declare normal ints... not pointers)
22:13:04 Join webguest98 [0] (
22:14:06XavierGrthat would be better if there was a function like width=getwstringsize("Hello");
22:17:11[solid]XavierGr: you could write a macro, couln't you?
22:22:20 Quit webguest98 ("CGI:IRC")
22:22:54XavierGrsorry newbe here
22:23:57niobosXavierGr: join the club... I'm in too
22:24:43*[solid] stfu's
22:29:26 Quit preglow ("leaving")
22:29:59t0masI was looking for preglow :(
22:30:46t0masin his firmware patches
22:30:58t0mashe uses a compiled in resource...
22:31:05t0masbut I need a crossplatform way to do that...
22:31:11t0masand I fear it's not possible
22:31:23t0masas the linux/windows/mac binary format differ to much...
22:31:54niobosIs there a prefered way to return errors? just integers? enum? #define?
22:32:27t0masjust ints afaik
22:32:31t0masand some info to debugf
22:32:42t0masDEBUGF("Error - Some text... \n");
22:32:45t0masreturn -1;
22:33:07niobosbut just -1, not ERR_NO_MEM or ...
22:33:26t0masin a plugin?
22:33:34t0masthen you can invent your own error codes...
22:33:53t0masand there is a standard for returning error's from other system calls
22:34:11niobosno in the songDB
22:34:11t0maslike: lseek returns -3, then you return -30 + your own error code
22:37:09niobosI don't know where they plan on putting the song-db in; it might become a plugin
22:37:24niobosright now I'm just writing it to run on the PC
22:37:36XavierGrdo you know a quick way to center a given string. I use the getstringsize to get the size in pixels and the type:rb->lcd_putsxy(LCD_WIDTH/2-stringwidth,y)
22:41:48[solid]um... wouldn't (LCD_WIDTH-stringwidth)/2 work?
22:43:37 Join amiconn [0] (
22:44:00XavierGr[solid]: maybe I will try
22:46:21XavierGramiconn: is putsxy has a concatenate character in order to type ("string" & variablestring) ?
22:47:00XavierGrlets say that i want to print a string in the screen
22:47:09XavierGrI would use the putsxy function
22:47:58XavierGrbut in my case I want to print a string that i will give inside the function (with quotes) and concatenate to that a string that comes from a variable
22:48:45amiconnUse snprintf() to compose your output string in a buffer, then use lcd_putsxy()
22:49:17t0masand watch out for bufferoverflows
22:49:22t0masamiconn? undefined reference to `iriver_decode(char*, char*, unsigned, striptype)
22:49:26t0maswhen linking to iriver.o
22:49:32t0mashow's that possible?
22:50:21amiconnwut are you talking about??
22:50:28*amiconn is puzzled
22:50:45t0masI'm trying to make a tool...
22:51:08t0masfor windows people... to patch firmware (using preglows code) and to make a songDB and to convert dictionarys
22:51:22t0masand this rbx-tool.cpp includes iriver.h
22:51:24t0masthen I do
22:51:31t0masgcc -c -o iriver.o iriver.c
22:51:33t0masand then:
22:51:35t0masgcc rbx-tool.cpp iriver.o -o rbx-tool
22:51:41t0masgives me that linker error :|
22:51:49t0mas"undefined reference to `iriver_decode(char*, char*, unsigned, striptype)"
22:51:57amiconnI have no idea
22:52:21t0masshit :)
22:52:39amiconnHmm, the second command you gave looks a bit odd to me
22:53:14amiconnIt looks like you try to tell it to compile rbx-tool.cpp, and then link with iriver.o, but all in one line
22:53:26amiconnI'm not sure whether this is supposed to work
22:53:41t0masfor other thins it does...
22:55:21t0masld rbx-tool.o iriver.o -o rbx-tool
22:55:23t0mas"undefined reference to `iriver_decode(char*, char*, unsigned int, striptype)"
22:58:04XavierGrwhat snprintf does exactly i am confused!
22:58:34t0masint snprintf (char *buf, size_t size, const char *fmt, ...);
22:58:41[solid]XavierGr: the man page has great explanation of all printf functions
22:58:47t0masyou define a buffer...
22:58:52t0maschar buffer[32];
22:58:59t0masand a score:
22:59:11t0masint score = 10;
22:59:37t0massnprintf(buffer, sizeof(buffer), "Score: %d", score);
22:59:48t0masand then buffer becomes Score: 10
22:59:51XavierGrahh great!
23:00:00t0masnot sure if score should be &score
23:00:11XavierGr[solid]:where is the man page?
23:01:02XavierGrI got only the API and many functions are missing from there
23:01:19[solid]XavierGr: oh, so i guess you're on windows... kinda sucks, eh? :)
23:01:59t0masXavierGr: that docs should be updated... I started something about it on the mailinglist...
23:02:25XavierGr[solid]: how did you guessed that?
23:03:23t0mas<XavierGr> [solid]:where is the man page? <−− from that I guess
23:03:38t0masI can just type "man printf" and have docs about it :)
23:03:41[solid]years of experience :D
23:03:41t0mason windows.. you can't
23:03:56t0maseven man snprintf :D
23:04:01amiconnt0mas: you can... if you have cygwin installed
23:04:01t0masdidn't know linux knew that one too
23:04:12t0masamiconn: ok, but that's not really windows :)
23:04:13[solid]the printf manual is one of the best ever
23:04:24[solid]it's the one i learned most from ^^
23:04:41t0mas[23:00:25] <t0mas> not sure if score should be &score
23:04:41t0mas[23:00:27] <t0mas> amiconn?
23:04:42amiconnt0mas: yeah, but it is _on_ windows
23:04:53amiconnNo &here
23:04:56Slasheripatch updated, new version may work or not. Fixed rockboy issue, and seeking mp3 files works better
23:05:03XavierGrwell I have installed the Rockbox Devkit but it seems that is not there
23:05:05Slasherigood night :)
23:05:13XavierGrGood Night
23:05:26XavierGrand sweet dreams Slasheri. ;P
23:05:38t0masah... great idea.. have a management test tomorrow...
23:05:41amiconnXavierGr: The devkit is a really cut down cygwin install. Many things are missing
23:05:44t0masso I'm going to bed too :)
23:05:51t0masgood night
23:05:59XavierGrGood night to you too t0mas.
23:06:08 Quit t0mas ("good luck XavierGr")
23:06:34XavierGr*sweet* as Eric says!
23:06:37[solid]Slasheri: thanks for all your work:D didn't have the chance to say that yet
23:14:32MusicmadXavierGr: have you got a make zip you could send me of the newest patch?
23:19:37***Saving seen data "./dancer.seen"
23:20:43XavierGrno but I could make one
23:22:28Musicmadplease :)
23:24:41XavierGreh I forgot the url to slasheris site
23:25:20 Join ashridah [0] (
23:26:34Ismothere you go
23:28:27[solid]question... how do you devs feel about bugreports at the current state of progress? annoyed by them, don't really care or is it a good idea to post'em here?
23:30:18 Quit Chamois ("Leaving")
23:30:43 Join thomj [0] (
23:31:02thomj[solid]: depends on what kind of bug we're talking about
23:31:05 Nick thomj is now known as preglow (
23:38:33[solid]for example 'went here and there, pressed that, did that and w00t! illinstr 0x666'
23:38:44preglowmore like what part of rockbox it is about
23:38:51[solid]or 'did this and got a lockup'
23:39:04preglowif it's about something that's under heavy development, like audio playback, you can pretty much just mention it here
23:39:09preglowand we'll let you know if we're aware of it
23:39:15[solid]okay, cool
23:40:46[solid]i've been checking out the daily builds, well, daily for some time so i thought i might be just a little useful
23:40:49XavierGrI am making it right now
23:41:38[solid]but on the other hand i guess you are pretty flooded by bug reports, including good ones and ones that are more feature requests or implications of not reading some faq or such
23:41:48XavierGrpatched successfully now compiling
23:41:52preglowjust mention it, now that you're on irc, it doesn't take much time
23:42:41niobosI need a 16-bit integer; how should I do that in C?
23:42:50niobosasume short = 16bit (BAD, I guess)
23:42:57niobosuse char[2]?
23:42:59[solid]i'll try to to the best i can (not having m68k cross-compiling set up = relying on peoples builds)
23:43:02niobosuse #define
23:43:02amiconnshort is always 16 bit
23:43:14niobosamiconn: ALWAYS? on all platforms?
23:43:42preglowmore or less
23:43:45amiconnYes afaik (unlike long, which might be 64 bit on 64 bit platforms)
23:44:01preglowi really think someone should put c99 type in rockbox some place
23:44:09preglowunless they're internal to gcc already
23:44:23preglowisn't it self-explanatory?
23:44:29niobosisn't it "cleaner" to typedef unsigned short u16 in 1 central file?
23:44:32preglowif you KNOW you need 16 bits, why not make it verbatim?
23:44:53MusicmadXavierGr: sounds great!
23:44:54amiconnerm, because there is 'short'?
23:45:14amiconnI don't get why so many people want just another word for the same thing
23:45:19preglowyes, like the way you used 'int' before, you mean?
23:45:29niobosand if I need 32bit int? what should I use then?
23:45:44preglowthat didn't exactly save jyp any time
23:45:47niobos23:43:46 < amiconn> Yes afaik (unlike long, which might be 64 bit on 64 bit platforms)
23:46:23preglowyou don't know for certain short will always be 16 bits
23:46:29amiconnyes, but I don't think we'll see a 64 bit based player any time soon
23:46:30preglowit is very, very probable, yes
23:46:35niobospreglow: that's why I asked it
23:46:47 Quit ashridah (Read error: 145 (Connection timed out))
23:46:52preglowbut it doesn't hurt to be explicit when you can be
23:46:53niobosno, but you might see a 64bit computer
23:47:02niobosand sinde the songDB might run on that computer...
23:47:19preglowshorts are 16 bits on 64 bit computers as well, afaik, but still, yeah
23:47:31amiconnpreglow: The int usage was lazy coding (because rockbox platforms were all 32 bit before gmini)
23:47:36amiconnStill, lazy coding
23:48:01XavierGramiconn:how can I reset my configuration file in rockbox?
23:48:11amiconnpreglow: If long is 64 bit on 64 bit archs, how do you specify a 32 bit int?
23:48:12preglowyeah, but you get my point, there's absolutely nothing to be lost by being explicit, and there's a rare gain here and there
23:48:18niobosI still think it would be cleaner to have a config.h or something with 'typedef ... u32'
23:48:27preglowamiconn: with int ?
23:48:44amiconnNo, int is _always_ native int format,
23:48:46preglowniobos: my point is that there already are types like that in c99, they're called int32_t, etc
23:48:57amiconni.e. 16 bit on 16 bit archs, 32 bit on 32 bit archs etc
23:49:10preglowamiconn: no, int is 32 bit for x86-64
23:49:16niobosso I might just use int32_t then... (didn't know they existed)
23:49:19preglowamiconn: yes
23:49:23preglowamiconn: i am 100% certain
23:49:58preglowamiconn: it is something of an exception to the rule, i believe, goes to tell how little you can depend on these things
23:50:22niobosI'll use int32_t then
23:50:30preglowniobos: i don't think you can
23:50:47amiconnThe problem with typedef is that every coder has its own opinion how these types should be called
23:51:02XavierGrdoes anyone know how can I reset my configurations in Rockbox?
23:51:10amiconnSo imho it's better to completely stay away from typedef
23:51:24preglowamiconn: well, i don't see how that's a problem when they're internal to the compiler
23:51:28preglowamiconn: like the c99 types are
23:51:44niobosamiconn: ok, fine; but I want an 32bit-integer on RockBox, i386, x32, x64, ... hoe do I do that?
23:52:46XavierGrnevermind I found it
23:53:12preglowniobos: int, point is, that'll break on powerpc, for example
23:53:27amiconnint won't work
23:53:35preglowahh, then you need long
23:53:40preglowand that'll break on x86_64
23:53:53niobosint32_t looked good.......
23:53:55preglowOR, int32_t, which works everywhere
23:54:19preglowniobos: yes, but we'll need stdint.h for that, and rockbox hasn't got it
23:54:25niobospreglow: is int32_t signed or unsigned? or can I use unsigned int32_t (too lazy to check myself)
23:54:34preglowniobos: signed, uint32_t is unsigned
23:54:44niobosk thx
23:54:45amiconnniobos: Why do you need_exactly_ 32 bits?
23:55:02amiconnAH ok
23:55:03niobosbecause the offsets are EXAXTLY 32bit
23:55:21niobos(or 16 bits. depending)
23:55:41amiconnFor the data structures, you can't simply use int or long or whatever
23:55:53amiconn...since you also have to deal with endianess
23:56:16niobosI use macro's for that: BE32 will convert (or not) between BE and native-format
23:56:31nioboswell, stole those from one or another CVS-file
23:56:37 Join Manyoufear [0] (
23:56:53niobosI could use char[4]...
23:56:57 Nick Manyoufear is now known as lunchbox (
23:57:06amiconnI think using these macros will work fine
23:57:42niobosthos macro's only work on integer-32's since they jump around with << and >> and &
23:58:08amiconnpreglow: I don't know whether you tried compiling the simulator, but I'd guess some things won't work correctly on x86-64...
23:58:52preglowamiconn: now that you mention it...
23:58:59preglowamiconn: i remember giving up on making it work in this computer

Previous day | Next day