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 2007-07-19

00:00:13Nico_PI'd say around 100K with a reasonable bitmap
00:00:20Nico_Pnot sure
00:00:31stripwaxNico_P : - dumb question but what do the ** represent in your example?
00:00:40bluebrotherhmm. Quite much, but not too much compared with 32MiB buffer size.
00:00:56Nico_Pstripwax: the currently playing part... I took markun's notations
00:02:14stripwaxah sorry
00:03:26pixelmabluebrother: not all targets have 32MB buffer... there are some with 16MB and some with much less...
00:03:34stripwaxright. I'd misunderstood what your example was. Yours is with the metadata at the end
00:03:41Nico_Pstripwax: yes
00:04:03bluebrotherpixelma: right, but I guess you won't want AA on those targets.
00:04:10stripwaxis there a penalty in partitioning the buffer into audio|metadata?
00:04:15bluebrotheror do any of the newer, color targets have smaller buffers?
00:04:23 Quit saratoga ("CGI:IRC (Ping timeout)")
00:04:29pixelmayes, X5 for example
00:04:47stripwaxno, that won't work. right.
00:04:56bluebrotherhmm. Should check the device chart. Thought all color targets have at least 32 MiB
00:04:59*stripwax is finally up to speed, sorry about that
00:05:14pixelmathe X5 has 16MB
00:05:45 Join Neovangl1st [0] (n=Neovangl@
00:05:59pixelmaand I read about someone who likes to store his album art as high res pngs
00:06:01linuxstb_Here's my example with the metadata moving -
00:06:07 Quit Neovanglist (Nick collision from services.)
00:06:12 Nick Neovangl1st is now known as Neovanglist (n=Neovangl@
00:06:25 Quit Arathis ("Bye, bye")
00:06:39pixelmabah... surely that won't work anyway, spoke too soon
00:06:41bluebrotherwell, we could still require users to resize AA to some reasonable size.
00:07:23pixelmabut I could imagine AA on greyscale btw.
00:07:27stripwaxpixelma - hmm, why not? (or just because of png)
00:07:51stripwaxperhaps AA metadata could be resized on load..
00:07:57pixelmayes, I meant because of the png
00:07:58 Quit My_Sic (Read error: 104 (Connection reset by peer))
00:09:20 Join My_Sic [0] (
00:09:23stripwaxlinuxstb - (because I can't remember) - is the buffer just a contiguous array (logically divided until blocks but not physically), or is it an array of pointers->blocks? (i.e. does that imply a memcpy)
00:09:35 Join DrMoos [0] (
00:10:03linuxstb_It's a contiguous array.
00:11:10stripwaxso moving metadata around means invalidating pointers that plugins might have
00:11:44linuxstb_Yes, it means anything using metadata can't assume it will still be there after a yield().
00:13:14markunlinuxstb_: moving makes it a lot easier indeed
00:13:20 Quit bdgraue (Remote closed the connection)
00:13:33*Nico_P is starting to think that too
00:13:48bluebrotherI'm still wondering if it wouldn't be more effecient to just define blocked metadata blocks
00:13:56linuxstb_Yes, but will it complicate the rest of Rockbox a lot?
00:14:01 Quit jgarvey ("Leaving")
00:14:38 Quit davina (Remote closed the connection)
00:14:49bluebrotherif it's capsuled good enough, so the higher audio stuff doesn't know about the buffering?
00:15:15*bluebrother started to draw brackets around old data buffers. Makes it a bit cleaner.
00:15:20markunlinuxstb_: yes, I read the part of "... after a yield()" after I hit enter
00:16:21*bluebrother remembers himself of KISS.
00:17:06linuxstb_The KISS solution is to have a static worst-case buffer for the current track's metadata....
00:17:07 Quit RaRe (Read error: 110 (Connection timed out))
00:17:35Nico_P... like for the codec
00:17:38stripwaxWould using an index to a list of buffers (and reordering the indexes) be easier than copying the metadata around within the buffer?
00:17:48bluebrotheryeah, but there needs to be a KISS solution for something more advanced ;-)
00:18:04 Join RaRe [0] (
00:18:11peturaustriancoder: still here?
00:18:15stripwax(assuming the buffer behaves like a list of buffers and not just one long arbitrary block of memory)
00:18:28bluebrotherstripwax: that sounds like the blocked sections I'm thinking about
00:18:40linuxstb_The buffer does behave like a long arbitrary block of memory...
00:18:54*stripwax must have misunderstood the A1,A2,A3, notation..
00:19:16linuxstb_stripwax: Yes, that's just a simplification of what's going on.
00:19:43linuxstb_In reality you'll buffer blocks of varying size - as big as you can.
00:19:47stripwaxdid it change? vaguely remember the playback code referring to actual buffers
00:19:53stripwaxat some point, in the past
00:20:09 Quit secleinteer (Connection timed out)
00:20:47linuxstb_I think the PCM buffer is split into smaller buffers, but I'm not sure.
00:21:11stripwaxhm, yep - maybe I was thinking of the PCM buffer not the codec buffer. bah.
00:21:17 Quit Siltaar ("Aurevoir -")
00:21:20jhMikeSit's one buffer with an allocation index...not really split
00:23:43 Quit Febs ("ChatZilla [Firefox]")
00:24:23linuxstb_There is one flaw in my plan though - the data that is freed whilst the buffer is being filled can't be used.
00:25:31markunlinuxstb_: I don't see what you mean
00:26:43linuxstb_I mean when there is 1MB of data left in the buffer, you start reading from disk. But when you've finished reading, you may only have 512KB left. IIUC, the current playback code will continue buffering into that newly freed 512KB.
00:27:06linuxstb_(reading and buffering takes many seconds to do)
00:27:32markunah yes, that's impossible if you move the metadata first
00:27:40bluebrothera try by me (linuxstb_'s version on top)
00:29:01stripwaxbluebrother - on line 34, where is AM while A5 is playing?
00:29:48bluebrotherargh. Seems I got myself confused once more :(
00:30:08bluebrotherbut I also wanted to have the metadata at the beginning.
00:30:23bluebrotheror at least look how it will work that way ;-)
00:31:35stripwax:) so the first line would be AM, A1, .. with holes while A1, A2 get played?
00:32:43linuxstb_bluebrother: You also seem to overwrite BM before B has finished playing.
00:34:21jhMikeSis the purpose of metadata movement to keep it valid for the duration of the track if T+M can't fit? (trying to get a clue)
00:35:12linuxstb_1) The metadata needs to be on the buffer for the life of the track; 2) Moving the metadata will stop holes appearing.
00:36:12jhMikeSall of it or can some be discardable after read? are codecs just metadata too?
00:36:25linuxstb_markun's original example demonstrated the hole problem -
00:36:52linuxstb_codecs will be discardable, as they are now. Other info (id3, album art, cuesheets etc) needs to stay available.
00:37:59 Join BHSPitMonkey [0] (
00:38:55bluebrotherok. I just tried to fix my example but it seems I'm too tired. Getting confused again :(
00:39:47 Quit Xerion (Read error: 104 (Connection reset by peer))
00:40:02linuxstb_bluebrother: What were you trying to show?
00:41:10markunlinuxstb_: I might have something.. let me try to 'draw' it out first
00:41:29stripwaxdoes the metadata need to move every time we load in more audio data chunks to keep it at the end? i.e. would there be a huge penalty to go from A1,A2,AM to A2,AM,A3 ?
00:41:41markunlinuxstb_: btw, I updated my original file a bit, is it clearer now?
00:41:48Nico_Plinuxstb_: I really don't know what to do... moving the metadata sounds good but might cause problems for the users of the metadata, but there seems to be no elegant way of not moving it... :(
00:42:04stripwaxbut i guess the problem there is that the codec needs to know that it needs to skip that region
00:42:06linuxstb_stripwax: A2, AM, A3 is bad because A2 and A3 are no longer contiguous.
00:42:21stripwaxyes −− but I'm wondering how 'bad' that really is.
00:43:04linuxstb_striptwax: It doesn't solve the problem anyway, when A2 and A3 are played, you have AM with empty space both before and after it - complicating later buffering.
00:44:01markunlinuxstb_: damn, what I wanted to try would involve a memmove on a part of memory we are currently playing from :(
00:44:34 Quit indro_ (Read error: 104 (Connection reset by peer))
00:44:47linuxstb_Nico_P: I don't know either.... I'm sure you know the "metadata user" parts of Rockbox better than me.
00:45:41stripwaxlinuxstb - hm, I was thinking a lazy move of AM when it matters would resolve that but I've tried out a few examples and can't think of a way to do that reliably :(
00:45:47bluebrotherlinuxstb_: I wanted to show the case where the memory gets fragmented and see how bad it is. It's memory, so random accesses don't cost extra.
00:46:31linuxstb_Well, it makes the buffering code more complex. It also means the guard buffer needs to be used more when giving data to codecs.
00:47:26 Quit HellDragon (Nick collision from services.)
00:48:15 Join entheh [0] (
00:49:56 Join Llorean [0] (
00:50:44markunlinuxstb_: there was an error in my original example, I overwrote AM
00:51:34 Join HellDragon [0] (i=JD@unaffiliated/helldragon)
00:53:54linuxstb_markun: So the problem occurs when starting to buffer B?
00:54:18markunlinuxstb_: this is what my example should have looked like I think:
00:55:17markunin the last line is should say that BM is preventing buffering
00:56:24markundamn, why does it have to be so complex :)
00:56:31 Quit ender` (" Today's BOFH excuse: broadcast packets on wrong frequency")
00:56:32stripwaxmarkun - does that have the same issues as my example (A4,A5 not contiguous, AM has holes either side)
00:56:53stripwaxor does it just stall while there are holes, until there are no longer holes?
00:58:13markunstripwax: it's getting a bit late to think clearly about it :(
00:58:21stripwaxI'm thinking after A4 has played, AM could be moved to where A4 was (so A4 is now at the back of the buffer). But that only works because the buffer doesn't have both BM and B1 buffered.
00:58:38stripwax^AM is now at the back, not A4 like I said..
00:58:42*stripwax is tired too
00:59:22jhMikeSumm...why can't metadata be at the start of buffering and then right-justified with remaining data from a current track? also, handles can be lockable.
01:00:09stripwaxjhMikeS - how much of the buffer would you allocate to metadata? how would you recover it later?
01:00:24Nico_PjhMikeS: I don't see what you mean ?
01:00:40linuxstb_JhMikeS: We need some ASCII art..
01:00:42stripwaxjhMikeS - I'm thinking you mean something like AM,BM,CM,DM | A1,A2,B1,B2,C1,C2, .. -> −−,−−,−−,DM | −−,−−,−−, D1,D2,D3 ...
01:00:57stripwaxwhere | is a virtual division between metadata and audiodata.
01:01:41*jhMikeS will get to the acii art soon. I'd allocate however many bytes are required though.
01:01:49stripwaxin that case, you can't rejoin the metadata partition with the audiodata partition until DM has been consumed and/or you move DM to the beginning (and potentially move audiodata into the reclaimed spot to keep the audiodata as a ring)
01:02:11stripwaxjhMikeS - so that would be like a separate metadata buffer and audiodata buffer?
01:02:41LloreanjhMikeS: Doesn't that kinda depend on a reliable ratio of metadata size to audio size?
01:02:43 Quit ompaul ("night all")
01:02:47jhMikeSA=len 6, AM*A1*A2 A3 => AM*A3*A4 A5 A6 and such
01:02:54LloreanIf you have a single very large file, you've wasted a lot of metadata space
01:03:06LloreanIf you have many, many, many tiny files (SID) you can't store all their tags without rebuffering.
01:03:14stripwaxjhMikeS - ok but what happens when you buffer a bunch of files at once
01:04:05jhMikeSIf they're small, I'm not sure what the problem is.
01:04:36jhMikeSAM A, BM B, CM C, DM D ... etc.
01:06:10*jhMikeS is probably missing a point or two here :)
01:06:12stripwaxjhMikeS - oh so metadata gets memcoped as we move through the audio buffer. this is linuxstb's suggestion.
01:06:17***Saving seen data "./dancer.seen"
01:07:05jhMikeSright...I was following such suggestions. it only happens if a track is incompletely buffered then.
01:07:53stripwaxand redbuffering can't make use of buffer that gets freed up while doing the rebuffering itself
01:08:24stripwaxoh, unless the rebuffering does the memmove in this case?
01:08:34*stripwax thinks that sounds complicated
01:09:07 Join DC1 [0] (
01:09:11 Nick iamben_ is now known as iamben (
01:10:03 Quit obo ("bye")
01:10:05jhMikeSyou mean data played while rebuffering? I had thought about it but not what exactly to do about it except skip the metadata block and fill it up, then move it after being finished.
01:10:31 Part maffe
01:11:51jhMikeSthe goal is to fill the buffer or finish-off the track so it can be moved a reasonable amount or even the exact amount before even starting
01:13:34 Join mpeccorini [0] (
01:13:41stripwaxif the metadata is at the start of the audiodata in the buffer, won't that also mean you need to keep moving the metadata all the time while it's playnig? trying to keep the metadata at the end of the audiodata should mean that gets moved around less
01:13:55 Join mbishop [0] (n=martin@unaffiliated/mbishop)
01:15:19linuxstb_Moving should only be needed before each rebuffering session.
01:15:54stripwaxI guessm, if the buffer is mostly empty you could have the problem that the buffer gets actually empty while you're rebuffering, if you can't take advantage of space freed up while rebuffering.
01:15:56jhMikeSwill it even make any differece? x number of rebufferings x = number of moves <= 2x
01:18:19 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
01:19:20jhMikeSis data going to be copied to a codec's working buffer or decoded in-place? I can't see putting breaks the data not causing problems if the latter.
01:19:42 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:20:07linuxstb_It depends on the codec, but most of them decode in-place. Or at least, perform their internal bitstream reading directly from the audio buffer.
01:21:00linuxstb_i.e. they use the ci->request_buffer() and ci->advance_pointer() API.
01:21:35*ac needs sleep
01:21:40 Quit ac ("Kopete 0.12.4 :")
01:22:28jhMikeSI don't see the problem as any different than assemling an MP3 stream from MPEG packets in mpegplayer really
01:22:45 Part mpeccorini
01:22:54linuxstb_That's not done very efficiently - there's lots of memmoving going on.
01:24:11jhMikeSnot like there use to be. only when the buffer is too empty
01:24:26mbishopI just installed rockbox on my sansa e260, it's very nice :)
01:25:14linuxstb_jhMikeS: Yes, but it's still more than is needed with the current ci->request_buffer API.
01:25:21peturbah, austriancoder left :/
01:25:37*petur fires up gmail
01:26:06jhMikeSif you split a syncword the decoder will choke
01:27:20linuxstb_I'm not saying mpegplayer could do anything different (because of the way the mpeg frames are split), but the same issue doesn't happen with normal audio files.
01:28:20jhMikeSif a metadata block is smack in the middle of an mpeg audio frame it's not any different and the frame must be copied and assembled in another buffer then
01:30:53linuxstb_Yes, that's what the current guard buffer mechanism (to handle the wraparound point) does.
01:31:06 Quit jhMikeS (Nick collision from services.)
01:31:12 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
01:31:26linuxstb_So ci->request_buffer() is always guaranteed to provide a pointer to up to 32KB of data, sometimes doing it's own copying.
01:31:53linuxstb_So a similar mechansim would be needed for metadata-within-track
01:32:08iambenmbishop: isnt it? installed it last week on my new e280
01:32:11jhMikeSI think the same case can handle spit and wrapped requests
01:32:20jhMikeSsplit even
01:33:20linuxstb_So are you saying you think the metadata should stay where it is first loaded?
01:33:36 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
01:33:50mbishopiamben: puts the default firmware to shame :(
01:33:58jhMikeSnot sure
01:35:09linuxstb_goodnight all.
01:35:11*linuxstb_ sleeps
01:35:15 Quit linuxstb_ ("CGI:IRC")
01:39:03 Join lukaswayne9 [0] (
01:46:58 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
01:48:17 Join homielowe [0] (n=chatzill@
01:50:19 Part homielowe
01:52:33 Quit DrMoos ("Glory to Rockbox")
01:57:41 Join tedrock [0] (
01:58:04 Join Robin0800 [0] (
01:59:20 Quit lukaswayne9 ("Ex-Chat")
02:01:25 Quit miepchen^schlaf ("Verlassend")
02:01:32 Quit Robin0800 (Client Quit)
02:03:38 Quit spiorf (Remote closed the connection)
02:08:56 Join hannesd [0] (
02:09:23 Quit petur ("Zzzzz")
02:09:29 Join safetydan [0] (i=cbca159f@rockbox/developer/safetydan)
02:18:44 Join jhulst [0] (
02:19:25 Quit Nico_P (Remote closed the connection)
02:21:26 Part pixelma
02:45:36 Join rstreeks [0] (
02:48:48TiMiD[FD]it seems that there is a problem building the manual
02:49:07TiMiD[FD]I've got this error :
02:49:07TiMiD[FD]make: /home/kevin/rockbox/manual/platform/ Command not found
02:49:07TiMiD[FD]make: *** [info] Erreur 127
02:58:53XavierGrTiMiD[FD]: that's normal
02:58:57XavierGr(I think)
02:59:03XavierGrthis error appears in the end
02:59:10TiMiD[FD]ah ok
02:59:17TiMiD[FD]so I can ignore it :)
02:59:28XavierGrpretty much yeah
02:59:59TiMiD[FD]but the file it requires exists ... in the tools folder
03:00:40TiMiD[FD]it built the pdf anyway so ...
03:01:02XavierGryes there is something weird with the make file
03:01:10XavierGrbut the pdf works even with that error
03:02:58TiMiD[FD]seems to yes ...
03:03:09TiMiD[FD]still not very elegant
03:03:24TiMiD[FD]but I don't feel like trying to think to fix it :)
03:04:04 Join Febs [0] (
03:06:20***Saving seen data "./dancer.seen"
03:09:18XavierGrpetur: unfortunately the person who told you so was right
03:09:41XavierGrmy iHP-115 doesn't provide any voltage to line-in
03:10:06XavierGrwith a multimeter the H300 showed 3.3V on line-in while iHP-115 0V :(
03:20:28 Quit donutman25 ("ChatZilla [Firefox]")
03:21:15 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
03:23:06iambenis there a guide anywhere for building rockbox on linux? i found a great guide for cygwin, doesnt help me much though
03:24:38TiMiD[FD]iamben: if your distribution doesn't have gcc packages for the cross platform compilers, you can use the script in the tools directory
03:25:14TiMiD[FD]it will build you a developpement environment and install it under /usr/local if I remember
03:25:52iambenheh i've already got an arm dev env for my gp2x, wonder if i can make that work
03:27:27TiMiD[FD]if it doesn't your build will just crash at start
03:27:52 Join SirFunk [0] (
03:28:11iambeni'll just make a new build env, my gp2x is set up so 'gp2x' is the build target
03:29:13TiMiD[FD]all you need is to provide a path to gcc, ld, and so on ...
03:29:41TiMiD[FD]if the gp2x uses the same cpu as your dap, there is no reason it wouldn't work
03:29:59TiMiD[FD]but I admit that compiling gcc is fun !
03:30:01 Join datachild [0] (
03:30:45 Join dandin1 [0] (
03:35:44iambeni dont know a whole lot about gcc outside of x86... i've got a build env set up that i've been using to build for arm9, dont know if that'd work just the same for arm7 (sansa) or not
03:36:37TiMiD[FD]probably not then
03:37:05dandin1I think it would. When you built gcc and binutils you simply specify "arm-elf", right?
03:37:52iambenwouldn't know, i used a build script to set that up too
03:38:06TiMiD[FD]my 68k cross compiled environment didn't worked for iriver
03:38:19TiMiD[FD]it didn't built the coldfire specific libraries
03:38:41TiMiD[FD]so when I booted rockbox compiled with that gcc, it crashed
03:38:51dandin1Well, we could compare the instructions to setup the build environement on the wiki and see if there's any difference.
03:40:13TiMiD[FD]for me I had to build the compiler
03:40:32TiMiD[FD]the previous one I had didn't wored, even if it was for the same arch
03:40:50TiMiD[FD]but it's still worth trying
03:41:11dandin1I think it'd work even if it's arm9 and arm7. It's the same gcc build options for both of them.
03:42:47TiMiD[FD]maybe the gp2x environment removed the arm7 libraries ...
03:46:42 Quit entheh ("^~")
03:46:58 Quit datachild` (Connection timed out)
03:49:10iambenwell its doing the CC blah.c stuff just fine, but now i gotta set up symlinks for ld, ar, ranlib,etc as well
03:55:04 Part rstreeks
03:57:14 Join kubiixaka [0] (
04:06:43 Quit Thundercloud (Remote closed the connection)
04:12:53 Nick Tr1ckY is now known as Tr1ckY^Gone (
04:15:32 Quit XavierGr (Read error: 110 (Connection timed out))
04:15:36 Quit kubiix (Read error: 110 (Connection timed out))
04:16:22 Join EbErT [0] (
04:20:30EbErTim confused about the Gigabeat- can you not sync it via the dock w/ rockbox installed
04:20:43EbErTwhich means it only uses usb 1.1? haven't bought one yet
04:24:39 Quit dandin1 ()
04:30:57jhMikeSEbErT: it connects to USB throught the dock port if that's what you mean
04:31:34EbErTand docking it works fine w/ rockbox installed?
04:31:46EbErTgets usb 2.0?
04:33:08jhMikeSyes, I do believe so.
04:33:52krazykiti'd check for you, but my dock is 12 hours away
04:34:01 Join bonbonthejon [0] (
04:34:28EbErThehe, thanks anyway, just a little confused. I dont care about the dock, just care about USB 2.0 transfer
04:34:45*jhMikeS will try to check but XP tends to while about 1.1 connections and doesn't when connecting the gigabeat
04:34:47 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
04:35:23krazykitoh, it'll do usb2 just fine without the dock, definitely
04:41:24 Join secleinteer [0] (
04:44:31MournBlade1.1 is just well.....painful
04:52:57 Join webguest30 [0] (i=54abcad9@gateway/web/cgi-irc/
04:59:09 Quit DC1 ("$4E75")
04:59:22 Quit hcs (Read error: 110 (Connection timed out))
05:03:32 Join Daishi [0] (
05:04:57iambentrying to build rockbox, its looking for sysfont.h... i read somewhere that the build system is supposed to create this, where am i going wrong?
05:05:39 Quit webguest30 ("CGI:IRC (EOF)")
05:06:22***Saving seen data "./dancer.seen"
05:07:39saratogaiamben: I've gotten that before, usually deleting the build directory and rerunning configure fixes it
05:11:18iambeni extract the rockbox-bleeding.tar.bz2, make the build dir, run the configure, then run make, and it errors on sysfont.h: no such file
05:15:14Presence<3 rockbox so bad. :
05:18:09iambensaratoga: well i dunno why make isnt creating it for me, but i just manually did the command that make was supposed to, and now im past that point... to another error.
05:19:54iambenmake[1]: *** No rule to make target `/home/ben/rockbox-bleeding/newbuild/firmware/drivers/tuner/lv24020lp.o'
05:24:06 Quit kubiixaka (Read error: 104 (Connection reset by peer))
05:26:42safetydaniamben, I think you'll have a much easier time of it if you just use the script
05:29:15iambensafetydan: i did, this is w/ the build env it created
05:29:54krazykitiamben, maybe try using an svn snapshot?
05:30:18krazykitpull it from svn yourself
05:31:20 Quit jhMikeS (Nick collision from services.)
05:31:25safetydaniamben, so you've got the arm compiler in your path, you've created an empty directory for your build, cd'd to that directory, run ../tools/configure and selected your options, and then just typed make?
05:31:26 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
05:31:44iambensafetydan: correct.
05:32:24iambenfirst sign of problems though, is that sysfont.h is not created like its supposed to... i can run that command myself to get around it, but maybe its a sign of a deeper problem?
05:32:52safetydaniamben, I'd suggest grabbing the source from svn if you can. The source tarballs aren't guaranteed to work.
05:32:58iambenits not trying to write anywhere funny outside this build dir is it? im doing this as my normal user
05:33:18safetydanThe build should only make things in your build directory.
05:33:29safetydanyou could always try "make V=1" to see what's being executed
05:33:52iambenyeah i saw that suggested elsewhere, so i tried it... not much useful, lemme see if i can pastebin
05:39:18iambenah hell
05:39:52iambeni was using svn earlier, switched over to the bleeding snapshot, now i go back to svn and it seems to be working
05:41:00 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
05:42:41 Quit saratoga ("CGI:IRC")
05:43:46paughOT question but.. i just read something that implies that you can use external powered mics with the iriver IHP-140 directly. (at least to some extent) can anyone confirm this?
05:46:22iambenthanks for the help, maybe it was just a small problem w/ the snapshot the whole time
05:46:48iambenanyway, i applied a patch, got it to build, and loaded it up on the sansa, all is well
05:53:26 Quit bonbonthejon (Remote closed the connection)
05:54:25 Quit Febs (Read error: 110 (Connection timed out))
05:56:26 Quit tchan ("WeeChat 0.2.6-cvs")
05:56:59 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
05:57:33 Quit rotator ()
06:02:35 Quit MournBlade ("CGI:IRC (EOF)")
06:15:20 Join lazka [0] (
06:18:38 Join aliask [0] (
06:19:11 Quit xNibbler (Read error: 110 (Connection timed out))
06:19:38 Join xNibbler [0] (
06:20:06 Quit lazka (Client Quit)
06:43:48 Join donutman25 [0] (n=chatzill@
06:46:48 Join ptw419 [0] (
06:53:30amiconnJdGordon: around?
07:06:24***Saving seen data "./dancer.seen"
07:09:58ptw419does anyone know how to link a section that has not been called in the main source?
07:13:51amiconnWhat are you trying to do?
07:15:44ptw419link in vectors
07:16:03ptw419arm exception vectors
07:17:07amiconngigabeat S ?
07:17:59ptw419nm i think i got it
07:18:09amiconnYou probably want KEEP() around the vector seciton in the .lds
07:18:14ptw419it was doing it, but i was looking at wrong place in assembly
07:18:25ptw419yeah thats what i was doing, but i was looking at wrong place
07:18:29ptw419thanks though
07:18:51*amiconn wonders why that is #ifndef CPU_ARM in the sources
07:21:04 Quit Daishi (Read error: 110 (Connection timed out))
07:29:04 Part safetydan
07:29:15 Join midkay_ [0] (
07:38:39 Quit atsea- (Read error: 104 (Connection reset by peer))
07:40:05 Join atsea- [0] (i=ariel@gateway/tor/x-5bb18596ef2c4815)
07:41:48TiMiD[FD]amiconn: just my curiosity, but would it be technically possible to make the grayscale lib work on the remotes at the same time as on the main unit ?
07:43:10TiMiD[FD]I suppose the timings are incompatible
07:43:19amiconnIt might be possible to make a grayscale lib for the remote, but you could only use either the main or the remote, not both
07:43:43amiconnThis is because the grayscale lib needs a timer, and there's only one user timer on coldfire
07:44:09TiMiD[FD]if the screens were physically the same, it would be possible then ?
07:44:17amiconnAnd for the iaudio remote, there's too much data to move in greyscale mode
07:44:55amiconnIt might be possible for them when switching the lcd controller to mono mode while the grayscale lib is running, but then it's an all-or-nothing situation
07:44:58TiMiD[FD]I see
07:45:41amiconnThe standard grayscale lib allows to use the grayscale overlay on a part of the screen, with standard output around it
07:45:42 Quit EbErT ()
07:46:06TiMiD[FD]I was actually reading the code of the grayscale lib, that's why I wondered
07:46:27TiMiD[FD]even if I don't see a lot of applications to use grayscale on the remote ...
07:49:04 Quit paugh ("Leaving")
07:49:22 Quit midkay (Read error: 110 (Connection timed out))
07:52:08amiconnEven if the timer could be shared between 2 running grayscale libs (letting it run with a cycle that equals the greatest common denominator of the two LCD's frame frequencies and then only refresh every n resp. m cycles) there would be the problem that the frame transfer for one lcd might delay the other
07:52:39amiconnEspecially since the remote transfer takes quite a while (serial transfer)
07:53:12 Join pondlife [0] (
07:54:10amiconnAnd it would require to disable ticking reduction (temporarily). I'd be curious how that sounds on H1x0s with a loud ticking problem...
07:56:27 Nick midkay_ is now known as midkay (
07:57:29amiconnOr better "zzzzzzzzzzzzzzz" in english
07:59:45TiMiD[FD]if we had color remotes it would be perfect though :)
08:00:18amiconnNo, it wouldn't
08:00:24TiMiD[FD]I wonder if there are daps with color remotes ...
08:00:26amiconnThere's a reason why there are no colour remotes
08:00:35TiMiD[FD]ah ?
08:00:42amiconnThe data transfer would take way too long over a serial link
08:00:45TiMiD[FD]bandwith for transfer ?
08:00:59amiconnAnd parallel links need too many wires
08:01:04TiMiD[FD]even with a faster link ?
08:01:21amiconnWell, they could use a high-speed serial link, but that would require separate chips
08:01:43amiconnAnd the shielding is another problem
08:01:53TiMiD[FD]so we're not gonna see color remotes before a long time I guess ...
08:02:13amiconnSee how some irivers already suffer from the ticking problem. Higher frequencies would make it worse...
08:02:17TiMiD[FD]would be feasible but very expensive right
08:03:15TiMiD[FD]eventually transfering the audio data digitally would avoid shielding :p
08:03:21amiconnEven at the same resolution, a colour remote for the iriver would need 16 times the data as the mono remote
08:03:55amiconnYes, but then you'd need a secondary dac in the remote. Also not necessarily cheap
08:04:26TiMiD[FD]so 17fps
08:04:30TiMiD[FD]according to this page
08:05:13amiconn17fps make the ui appear really sluggish
08:05:38TiMiD[FD]never tried ...
08:05:38amiconnYou need at least ~40fps for acceptable performance
08:06:03TiMiD[FD]so they would need roughtly a 3 times higher bandwith
08:07:48TiMiD[FD]would be fun to have a dap with 3 screens
08:08:40scorchewhy?...i rarely look at the one screen..
08:08:45*amiconn wonders what that should be good for
08:09:11TiMiD[FD]because screen api is for 1...n screens :)
08:09:31TiMiD[FD]so rockbox could run on it without changes
08:10:50TiMiD[FD]hey do you have any idea of what could cause a key to behave as another?
08:11:59TiMiD[FD]since maybe 4 month my iriver's remote volume up key behaves like the remote page down key (or directory skip) roughtly one press on 10
08:12:09TiMiD[FD]I wonder if it's hardware related
08:12:28pondlifeTiMiD[FD]: Does it do this in the OF too?
08:13:08TiMiD[FD]the few times I booted into OF I didn't noticed it, but I'm not sure
08:13:17TiMiD[FD]I will try using OF this evening
08:13:55TiMiD[FD]but if it was hardware related I wonder what could be the culpirit ...
08:14:26TiMiD[FD]some resistor that went crazy ?
08:15:06TiMiD[FD](if I recall the iriver's keys are differenciated with different resistor's values)
08:18:59 Quit hannesd (Read error: 110 (Connection timed out))
08:20:23 Join Febs [0] (
08:21:12 Part Sidlet ("Ex-Chat")
08:33:24 Join ender` [0] (
08:33:39*amiconn found that he can make the iaudio remote lcd driver even faster at lower cpu clocks :)
08:37:53TiMiD[FD]amiconn: I was looking at the lcds drivers sources, and I noticed a lot of code is duplicated
08:38:34 Join Bossier [0] (
08:38:57 Nick Bossier is now known as tdtooke (
08:39:42 Join Rob2222 [0] (
08:39:48TiMiD[FD]I wonder if it wouldn't be worth splitting the drawing part in 2 : a low level device specific and a high level that would work the same for every sceens
08:40:33tdtookeI have a quick question for the powers that be. Do you think a patch which adds the option of selecting a specific icon set would be a good idea. I realize we have a .cfg file that sets your icons back to default, I just thought it'd be pretty cool to have that added capability.
08:41:06amiconnTiMiD[FD]: They are already split like this...
08:41:20 Join kubiix [0] (
08:41:23 Join webguest06 [0] (i=18a2ffdf@gateway/web/cgi-irc/
08:46:01 Quit Febs (Read error: 110 (Connection timed out))
08:46:33TiMiD[FD]really ?
08:47:11 Quit webguest06 ("CGI:IRC (EOF)")
08:47:53TiMiD[FD]for example the setmargins function is the same for every driver
08:48:25TiMiD[FD]the firmware/drivers/lcd-* contains a lot of duplicated code
08:48:30amiconnThat will go anyway with viewports
08:49:05amiconnMaybe you're talking about the format specific stuff, not device specific
08:49:59TiMiD[FD]yes format specific stuffs
08:50:39TiMiD[FD]but a lot of things are common
08:50:48TiMiD[FD]like drawline for example
08:51:24 Quit BHSPitMonkey (Read error: 104 (Connection reset by peer))
08:53:15GodEatertdtooke: there's no reason why you can't have a .cfg file that just sets icons too.
08:53:27GodEatertdtooke: so you can have one .cfg file per iconset
08:53:28TiMiD[FD]with viewports it removes indeed a lot of code related to fonts / scroll
08:53:39GodEatertdtooke: then there's no need for an option to select them
08:53:43TiMiD[FD]so it won't be that bad
08:57:10 Quit Rob222241 (Read error: 110 (Connection timed out))
08:57:48bluebrothertdtooke: I would even go further the other direction and remove the "browse wps / rwps" entries
08:57:59bluebrotheras that should be done by the theme anyway.
08:58:11tdtookeAlrighty then, I guess I won't be submitting that one to flyspray :(
08:58:50*amiconn disagrees with bluebrother
08:59:04tdtookeMy preference for iconsets changes every other minute at times so I thought maybe somebody out there would like that idea.
09:00:21tdtookethough using extra .cfgs probably would involve at lot less effort on my part
09:01:48 Join BHSPitMonkey [0] (
09:02:21tdtooketheoretical question, if one were to submit a patch that added a setting would it be required to support all languages before it would be considered for acceptance?
09:06:16amiconnHmm, something made the X5 slower in general since my last lcd framerate measurement... :(
09:06:28***Saving seen data "./dancer.seen"
09:07:54 Join petur [0] (n=petur@rockbox/developer/petur)
09:08:19scorchetdtooke: languages are updated based on english.lang ...i very much doubt there is a single language out there that is fully up-to-date (besides english)
09:08:44TiMiD[FD]tdtooke: I don't think so
09:08:44TiMiD[FD]lot of menu options are not even translated
09:08:44TiMiD[FD]devs only care about the english version, the translators takes care of the rest :)
09:10:10*scorche hears an echo
09:17:29JdGordonamiconn: am now, wassup?
09:18:45amiconnCheck yesterday's log starting at 21:55
09:19:05 Join Febs [0] (
09:20:09peturamiconn: I just had your problem today: changed a seting before powering off and it wasn't saved...
09:21:19amiconnIt does not only happen just before poweroff. There can be as many spinups inbetween, as long as you don't change a setting from elsewhere than the menu
09:21:42amiconnLooks like settings_save() isn't called from the standard settings function
09:21:52 Join desowin [0] (
09:22:08 Quit donutman25 ("ChatZilla [Firefox]")
09:22:28tdtookeOk, thanks guys
09:22:36JdGordonit isnt? it _should_ be
09:22:38 Nick tdtooke is now known as Bossier (
09:22:40JdGordonill check in a few min
09:22:42 Quit Bossier ("( :: NoNameScript 4.02 :: )")
09:22:47*JdGordon starts work on monday!
09:23:37scorcheJdGordon: the link i gave you? =P
09:23:39TiMiD[FD]my name is appearing on the svn log of rockbox
09:23:43scorchei would be excited too!
09:24:21TiMiD[FD]I would really don't want my boss to search on the internet my name and read that page
09:24:39scorchewhy not?
09:25:03TiMiD[FD]then looking at the hour ... noticing what I'm doing during the working hours ....
09:25:22 Join davina [0] (
09:25:42JdGordonamiconn: bah, your right, settings_save isnt called for about 90% of settings, ill fix that now
09:26:10TiMiD[FD]wouldn't it be possible to "fix" that ? :p
09:26:50GodEaterTiMiD[FD]: yes - don't skive off work ;)
09:27:02pondlifeJdGordon: Congratulations on the employment... so you only have 3 days to finish MoB :p
09:27:35JdGordoni hope not
09:27:40TiMiD[FD]I would do that gladly if I had an interresting work, or at least if I had something to do ...
09:27:55GodEaterTiMiD[FD]: unless you have commit access I wouldn't worry about it
09:28:03 Quit aliask ("ChatZilla [Firefox]")
09:28:06TiMiD[FD]I have commit access
09:28:33GodEatersave your commits till you're at home ?
09:28:58scorcheGodEater: he was with us a while back...way before you...he has recently come back into the fold =)
09:29:46GodEaterscorche: I think I just remembered seeing him ask for commit access again, so I figured that ;)
09:31:04pondlifeCan anyone see the problem where skipping tracks clears the pause status on Flyspray? I'm sure it *was* on there, but a search for "Pause" finds nothing much.
09:32:00TiMiD[FD]GodEater: could be a solution :)
09:32:09pondlifeAh, LinusN fixed it.. :(
09:32:22TiMiD[FD]but I'm not at home that often ...
09:36:51GodEaterTiMiD[FD]: cronjob ? :)
09:39:31*pondlife just learnt that Canada has time-zones on the half-hour...
09:41:15 Quit Febs (Read error: 110 (Connection timed out))
09:43:55GodEaterpondlife: think Sri-Lanka is in a half-hour timezone too
09:45:27 Quit Bawitdaba (Read error: 110 (Connection timed out))
09:47:06 Join scorche` [0] (i=Blah@rockbox/administrator/scorche)
09:48:17 Quit scorche (Connection reset by peer)
09:52:01 Nick scorche` is now known as scorche (i=Blah@rockbox/administrator/scorche)
09:54:54JdGordonamiconn: or anyone: feel like testing my patch? im 99% sure it works, but the rest of my tree is fubar so cant test it here
09:59:26 Quit darkless (Remote closed the connection)
10:03:02 Join darkless [0] (
10:04:04GodEaterIs there some magic in onplay.c that decides which context menu entries are appropriate for directories and which are appropriate for files ?
10:04:22JdGordonwell, not magic
10:04:45GodEaterwell, I didn't really think it was Juju, I just can't see how it works
10:05:16GodEaterBagder: T-shirts and mugs have arrived, my porter decided not to steal them for himself :)
10:06:08JdGordononplaymenu_callback is most likely where you want to look
10:06:59GodEaterJdGordon: thanks sir
10:09:03 Join fred_ [0] (n=fred@
10:10:24 Join Lynx_ [0] (
10:11:51GodEaterlooks more like clipboard_callback to me
10:11:58GodEaterunless I've grown even more dense
10:12:05GodEaterwhich is always possible
10:14:00 Quit andrew__1 (Read error: 110 (Connection timed out))
10:14:55 Quit fred_ ("Ex-Chat")
10:15:08 Join fred_88 [0] (n=fred@
10:22:08 Join Nick_Brackley [0] (i=7dff0332@gateway/web/cgi-irc/
10:23:57 Join obo [0] (
10:24:42JdGordonlinuxstb: im just reading yesterday/thismornings log, I still tihnk moving the metadata is a bad idea
10:25:15JdGordonand I dont see why storing it behind the tracks audio would be simpler than storing it before the audio
10:26:46 Nick Tr1ckY^Gone is now known as Tr1ckY (
10:27:24*JdGordon gone... dinner time
10:29:03 Quit fred_88 (Remote closed the connection)
10:33:38 Join fred_88 [0] (n=fred@
10:36:14 Quit scorche (Read error: 104 (Connection reset by peer))
10:37:23 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
10:38:42GodEaterpixelma: good call on that ipod mini post - I didn't think of that
10:39:58 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
10:44:43 Join Wiwie [0] (
10:45:37 Quit Wiwie (Client Quit)
10:47:36 Join Wiwie [0] (
10:48:25 Quit Wiwie (Client Quit)
10:51:56 Quit gtkspert (Read error: 110 (Connection timed out))
10:56:47 Join Rondom [0] (
11:02:03 Join FOAD_ [0] (
11:06:18 Join Entasis [0] (
11:06:30***Saving seen data "./dancer.seen"
11:09:05linuxstbJdGordon: If you don't move the metadata, you need to deal with the current track's metadata floating around in the middle of the empty part of the buffer. If you do move it, then it is always contiguous with the current track's data, and you don't have the complication of dealing with a fragmented buffer.
11:11:32linuxstbGodEater: The checksums are initiialised with the model number from the bootloader. So for example, the mini 1st gen is model number 9, and the 2nd gen is model 11. Hence the difference of 2 in the resulting checksums. So a small difference in the checksum almost always implies the user has downloaded the wrong build.
11:12:09GodEaterlinuxstb: I also noticed he got a bad checksum error right at the top of those messages from the bootloader - is that the same issue ?
11:14:00 Quit FOAD (Read error: 110 (Connection timed out))
11:14:06GodEaterso his issue now is "get it into disk mode" and then "put the right build on" ;)
11:14:37linuxstbAlthough I don't think that's an accurate copy of the error messages - the bad checksum line should be further down.
11:14:57GodEaterI did wonder about that - I've never seen that message appear there
11:15:33linuxstbAh, the text has wrapped back to the top of the screen...
11:16:25linuxstbSo those top four lines actually come after the "Can't load rockbox.ipod" line.
11:17:13GodEaterah ha
11:17:31*GodEater is clearly too used to the enormous screen real estate offered on the video
11:19:24scorcheanyone else need any cloak changes while i am at it?
11:19:30scorchenew devs/etc?
11:19:44*obo raises his hand
11:19:50JdGordonlinuxstb: my idea to handle the fragment problem is to use the current linked list so when we have the problem where the data for one track is in the middle of 2 empty bits, the next track is read into the first bit, then jumps to the bit behind the block, it would all be handled by the same id so it would be transparent to everything but buffering
11:21:21scorcheany others? (speaking for other new devs is alright as well
11:22:22linuxstbJdGordon: Don't forget you will then also need to use a guard buffer to enable the codecs to read the data in contiguous blocks of up to 32KB... It just seems more straightforward to move the metadata to me, but I'm not say your way won't work either.
11:22:28scorcheobo: done
11:23:19JdGordonhmm, yeah, forgot about that problem
11:23:23GodEatershould we put "Support DRM" on the NoDo page ?
11:23:49peturscorche: I think austriancoder hasn't got a cloak yet
11:23:59linuxstbJdGordon: With a fragmented buffer, there will also be more chance of a non-streaming file (one that needs a single contiguous buffer) not fitting, as the buffer will be split into up to 3 parts, rather than up to 2.
11:24:17scorchepetur: he is registered, so he cant be cloaked
11:24:56scorcheerrr...he is now
11:25:04peturyesterday he was reg'd and didn't have a cloak
11:25:16GodEaterany archos experts that know about recording care to take a look at this :
11:25:29linuxstbGodEater: I never like to say never.... IMO the NODO is for things which are not technically possible, so for example the alternative filesystem support should be removed from the NODO now that Rockbox runs from flash on some devices (and some devices like ipods natively support non-FAT filesystems).
11:25:35peturhe's not here now, the nick here is just a ghost I think
11:26:17GodEaterlinuxstb: while I agree DRM is not *technically* impossible, I don't see how we can support it legally ?
11:26:39amiconnGodEater: That's not an archos specific question
11:26:43linuxstbGodEater: Maybe it's legal in some countries...
11:27:07oboscorche: thanks - does saratoga have one?? (trying to remember who else is new to the dev list)
11:27:20amiconnImho the nodo is also for stuff we don't support intentionally
11:27:24scorcheobo: he isnt on, so he cant be cloaked
11:27:31amiconnDRM is one such thing imo
11:28:12scorchepetur: alright...done =)
11:28:20linuxstbI guess I agree - stick DRM on the NODO...
11:29:03*petur answered the recording post
11:29:33amiconnAnd support for other filesystems is also such an intentional nodo - perhaps with the exception that we might want to support hfs on ipods
11:30:23linuxstbWell, if someone implemented ext2 support, I would probably use it...
11:30:26scorcheif anyone else remembers/needs a cloak, let me know (at any time)
11:30:38linuxstbBut it's not something I would want to spend time implementing myself.
11:31:28JdGordonwe need a filesystem plugin api!
11:31:41amiconnAdding support for additional filesystems is pure bloat imho as long as there is no real reason to (like hfs on macpods)
11:31:54amiconnfat(32) is supported by virtually all operating systems
11:32:50scorcheGodEater: "moraley" ?
11:32:59GodEaterscorche: shoot me - my spell checker is on the blink
11:33:11*scorche shoots GodEater
11:33:20*GodEater bleeds quietly on to the floor
11:33:32JdGordoni just cleaned that you bugger!
11:34:10*GodEater coughs and points at scorche
11:34:27*scorche sits in his chair and cleans his gun
11:36:17 Join Nico_P [0] (
11:39:05GodEaterdo we offer dual boot on the H10 ?
11:40:06pixelmaI think so
11:45:51linuxstbGodEater: Looking at bootloader/main-pp.c, it looks like we do - hold LEFT as you boot (same as Sansa) and it will load /System/OF.mi4 or /System/OF.bin
11:46:17GodEatergood enough
11:53:20 Quit JdGordon ("Konversation terminated!")
11:56:25 Join maffe [0] (
12:00:21 Part maffe
12:00:25Nico_Pscorche: is it too late for a cloack ?
12:00:34 Join maffe [0] (
12:00:49scorcheit never is...although it may take a bit longer for you...
12:02:34pondlifeHmm, I missed JdGordon? I don't think he finished his sentence on
12:03:12Nico_Pscorche: thanks :)
12:03:33*Nico_P learned about the extistence of cloaks two days ago
12:04:05Nico_Pwhat should I do to use it ?
12:04:08scorche(underscore isn't possbile in cloak)
12:04:18scorcheas long as you are identified, it will apear
12:04:25Nico_Pit has :) thanks
12:05:02 Join JdGordon [0] (
12:05:11linuxstbpondlife: I think I can guess what he meant - that config files are not a replacement for when you want the ability to choose between many values in many settings - you would need hundreds of config files.
12:05:15linuxstbAh, he's back ;)
12:05:29pondlifeJdGordon: I was just saying - did you finish your comment:
12:05:57JdGordonyes an no
12:07:09pondlifeI'd personally rather have browse .CFG files instead of any quick menu.
12:07:19pondlifeany configurable quick menu, I mean
12:12:05JdGordonyeah, i sort of lost my train of though and couldnt be bothered finishign the sentance.. so thought that comment was good enough
12:12:26 Join Lear [0] (i=chatzill@rockbox/developer/lear)
12:13:12Nico_PJdGordon: should've deleted the comma :)
12:13:16GodEaterif you clear the backdrop in a plugin, does it get restored to the original when the plugin exits automatically, or do you have to remember what it was and set it explicitly ?
12:13:37Nico_PGodEater: it will get restored
12:13:42GodEaterthank you
12:16:23 Quit jhMikeS (Nick collision from services.)
12:16:29 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
12:16:54 Join jgarvey [0] (
12:25:02 Join PaulJam [0] (
12:27:23 Quit Rondom (Read error: 113 (No route to host))
12:53:11 Quit linuxstb (Read error: 110 (Connection timed out))
13:06:32***Saving seen data "./dancer.seen"
13:07:16 Quit JdGordon (Read error: 104 (Connection reset by peer))
13:08:42 Join MournBlade [0] (i=4401597c@gateway/web/cgi-irc/
13:11:34 Quit austriancoder (Remote closed the connection)
13:13:43 Quit Nick_Brackley ("CGI:IRC")
13:16:57 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:22:31 Join MoFF [0] (
13:23:44MoFFhello, is anybody about?
13:24:07 Join Thundercloud [0] (
13:24:40MoFFi was going to ask if divx/xvid is supported in rockbox for the ipod video 80gb?
13:25:02scorcheit isnt
13:25:14MoFFthank you
13:25:31MoFFis it impossible to implement too?
13:25:45 Join Febs [0] (
13:26:37GodEaternot impossible - but v. difficult
13:26:59GodEaterwe're having trouble optimising mpeg2 - going for mpeg4 is going to make life even more challenging
13:27:28*scorche looks over in the direction of the broadcom building a few min away
13:27:41GodEaterscorche: you've been through their trash already I take it ? ;)
13:27:58linuxstbMoFF: The Apple firmware makes use of a completely undocumented video processing chip to perform the video decoding. Rockbox doesn't touch that chip and just uses the main CPU, which obviously is far less powerful.
13:28:01scorcheno, but i *could*
13:29:18MoFFthank you all for answering my question so well :D
13:29:26Davo_Dinkumyay for dumpster diving!
13:30:06MoFFso you say you're trying to implement mpeg2 eh... that's interesting
13:30:37scorchewell, technically, it is already implemented
13:30:50scorchewhich is why we said optimizing =)
13:32:01MoFFwow :o
13:32:23*amiconn wonders what cpu type and rating his mobile phone has
13:32:32 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
13:32:55amiconnIt can play mpeg 4 video, with realtime scaling
13:33:18MoFFforgive me if this comes out sounding really silly, but, would that mean i could watch a dvd on the ipod with rockbox?
13:33:27 Quit linuxstb (Nick collision from services.)
13:33:27*MoFF blushes :s
13:33:30 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
13:33:39scorcheMoFF: if formatted right, yes
13:33:43linuxstbpetur: Looking at your properties.c plugin, I always thought that dircache should be mostly transparent to the apps/ code - so does the normal readdir() not use dircache?
13:34:13linuxstbMoFF: You need to resize the video to your LCD's resolution, plus also convert the audio to 44.1KHz MP3 (or MP2).
13:34:42MoFFthat is really impressive
13:35:04peturhmmm can't remember, there must have been a reason I did it like that. Can it be the normal calls don't use dircache?
13:36:01linuxstbLooking at the code, it seems that way - but that seems really odd to me...
13:36:26 Quit jhulst (Remote closed the connection)
13:36:34*scorche uses on his phone
13:43:16MoFFbear with me, yet another question which might pain you; if i install rockbox, and for some reason i need to send my ipod in for repairs (it is an apple product after all), will i be able to delete rockbox so as to not void my warranty?
13:43:32 Join andrew_ [0] (
13:43:33scorcheassuming the device is usable, yes
13:44:12*MoFF is just scared because of previous troubles with ipods
13:44:30 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
13:44:32Davo_Dinkumheh yeah
13:44:50JI talked to somebody in my local apple shop, he said if the device won't start the drive then warranty isn't voided regardless of what's on the drive... but don't quote this!
13:45:31 Quit Febs (Remote closed the connection)
13:45:32KasperleMoFF: if you can still reset your ipod to disk mode, itunes can fully restore it to how apple would like it to be
13:45:40Davo_Dinkumyeah rockbox goes onto the hdd on ipods
13:45:46Davo_Dinkumit doesnt touch anything else
13:46:46MoFFi just want rockbox for MPEG-2 playback :D
13:47:09scorchewhich ipod so you have?
13:47:24MoFFiPod Video 80GB
13:47:49PaulJamMoFF: the apple firmware is much better for video playback on g5 ipods
13:47:55scorcheif you are looking for decent video playback, you will find yourself being much happier int he apple firmware tbh
13:48:24MoFFah ok :)
13:49:07Kasperlejust dual boot
13:53:54 Quit linuxstb (Nick collision from services.)
13:53:56 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
14:00:31 Join Febs [0] (
14:07:15 Quit jgarvey ("Leaving")
14:09:05 Quit Nico_P (Remote closed the connection)
14:14:58 Part MoFF
14:23:44 Quit andrew_ (Nick collision from services.)
14:23:47 Join andrew_ [0] (
14:23:50 Quit andrew_ (Nick collision from services.)
14:28:52 Join andrew__1 [0] (
14:35:58 Join Nick_Brackley [0] (i=7dff0332@gateway/web/cgi-irc/
14:38:18 Quit idnar (Nick collision from services.)
14:38:20 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
14:38:50 Quit maffe (Remote closed the connection)
14:44:49 Join toke [0] (
14:44:53 Quit toke (Client Quit)
14:49:43 Quit SirFunk (Connection timed out)
14:49:44*GodEater has added a note about DRM to NoDo and would be grateful for feedback.
14:55:32 Part pixelma
14:56:01PaulJamisn't "crippled" a little bit harsh?
14:56:31Davo_Dinkumwho said that?
14:56:49GodEaterI believe that's the accepted way to describe them in OSS circles
14:57:08Davo_Dinkumdescribe what?
14:57:21GodEaterhahaha - no
14:57:28GodEaterfiles encrypted with DRM technology
14:57:58Davo_Dinkumno, crippled isn't too strong a word
14:58:04TiMiD[FD]is there a way to activate logging on the device (DEBUGF) without using the remote screen ?
14:58:05Davo_Dinkumit suits DRM well
14:59:14GodEaterPaulJam: if you can think of a better term by all means lets hear it
14:59:42PaulJamGodEater: DRM encrypted?
14:59:44bluebrotherDRM = Digital Restrictions Management.
15:00:11bluebrotherdoesn't the terms tells itself? ;-)
15:00:23linuxstbAnyone (Slasheri?) know why some of the apps/ code is calling readdir() and some is calling readdir_cached() ? IIUC, that a mistake and the intention is that all code should be using readdir_cached() ?
15:00:30MournBladeI kinda like virus infected
15:00:39pondlifeRestrictions in F/OSS, Rights in RIAA..
15:00:44GodEaterMournBlade: that's inaccurate though
15:01:02GodEatercrippled is accurate IMO
15:01:18GodEaterit's of no use unless you possess the key to use it
15:01:54bluebrotherpondlife: RIAA is lying −− it restricts the user in all cases :o
15:02:16linuxstbTiMiD[FD]: Only on some devices - those with a serial port. On most devices you can only use logf.
15:02:35GodEaterbluebrother: it's your right to be frustrated ;)
15:02:55PaulJamI mean i don't like DRM either, but in my opinion the wiki should be more or less objektive and i think "DRM crippled" is a quite subjective term.
15:03:03linuxstbIsn't it just referring to the rights of the copyright holder, rather than the rights of the consumer?
15:03:11bluebrotherwell, luckily I never bought such cr*p. But those "copy control"led CDs can be annoying enough.
15:03:14GodEaterlinuxstb: that's not the way they try to sell it
15:03:37GodEaterDRM somehow empowers the consumer in RIAA's world
15:03:45pondlifebluebrother: If I receive a CC CD, I send it right back.
15:04:01pondlifeMany of them won't play in my car stereo.
15:04:12pondlifeI ordered Red Book, dammit!
15:04:15linuxstbGodEater: Surely no-one believes that? ;)
15:04:29GodEaterlinuxstb: no-one with half a brain cell
15:04:44TiMiD[FD]linuxstb: too bad :( the stack overflow I'm trying to track doesn't occurs when rockbox is buit in debug mode ...
15:05:35TiMiD[FD]linuxstb: for the readdir it seems that almost all the plugins are using it ...
15:06:02linuxstbTiMiD[FD]: Using the original readdir() you mean, rather than the dircache version?
15:06:35***Saving seen data "./dancer.seen"
15:06:53TiMiD[FD]seems so
15:07:10TiMiD[FD]the plugins call a rb->PREFIX(readdir)
15:07:27TiMiD[FD]which is initialized to a pointer to the non cached version
15:07:37GodEaterlinuxstb and I feel it should be using the cached version
15:07:40linuxstbYes, that's another complication - the protection in the sim.
15:08:10TiMiD[FD]should use readdir_cached instead anyway
15:09:58TiMiD[FD]readdir_cached is redefined to readdir on targets without dircache
15:10:00TiMiD[FD]so ...
15:10:50TiMiD[FD]actually in the plugin api both cached an non cached versions are present ...
15:11:10TiMiD[FD]doesn't makes sense
15:11:24peturpondlife: I observed that the last CD's I bought no longer contain the official CD logo. Philips once complained the protected cd's are not following the standard and should not carry the logo. The last disks did not contain CC however...
15:11:30TiMiD[FD]wouldn't be hard to fix
15:12:02 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
15:12:51 Quit linuxstb (Nick collision from services.)
15:12:53 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
15:12:57TiMiD[FD]some of them even make the test toknow whether to use dircache or not ...
15:13:09PaulJampetur: maybe those were "enhanced" CDs containing some bonus material.
15:13:24GodEaterTiMiD[FD]: those are the ones which prompted the question :)
15:13:32peturTiMiD[FD]: what happens on targets without HAVE_DIRCACHE?
15:13:45linuxstb#define readdir_cached readdir
15:14:05 Join Arathis [0] (
15:14:42TiMiD[FD]so unless people want to make the disk spin uselessly this should be corrected :p
15:15:07linuxstbUnless there's something else going on we're not thinking of...
15:15:08TiMiD[FD]grep tells me 28 changes to make
15:15:21*linuxstb would prefer to ask Slasheri about it.
15:15:23TiMiD[FD]#ifdef HAVE_DIRCACHE while(0 != (entry = rb->readdir_cached(dir)))
15:15:23TiMiD[FD]#else while(0 != (entry = rb->readdir(dir)))
15:15:39TiMiD[FD]when I see that I wonder if there is something to understand ...
15:15:53TiMiD[FD]also opendir
15:16:03TiMiD[FD]vs opendir_cached
15:16:16GodEaterproperties.c ?
15:16:17peturaren't the structs different?
15:16:20TiMiD[FD]well all the api
15:16:31TiMiD[FD]it's wrapped
15:16:50linuxstbIMO, it would have been nicer to keep the POSIX API in apps/ and hide the fact that dircache versions are being used, rather than hide the POSIX API and use the dircache version...
15:17:00pondlifeIt's annoying that the app needs to specify the dircached API... why can't that be hidden lower down?
15:17:03*GodEater agres
15:17:14TiMiD[FD]could be fixed also
15:17:35TiMiD[FD]there are not so muchplaces which make use of dir reading
15:18:46 Join webguest20 [0] (i=ca36dcc4@gateway/web/cgi-irc/
15:19:18 Quit MournBlade ("CGI:IRC (EOF)")
15:21:15 Quit webguest20 (Client Quit)
15:21:34Soapdesowin, can we remove your unsupported build since the patches have been commited?
15:26:59TiMiD[FD]linuxstb: if there is no objection, I could maybe fix that tmorrow during my boredom at work ...
15:27:16TiMiD[FD]use the posix api in the application
15:27:42TiMiD[FD]and rename the readdir to readdir_uncached for example ?
15:28:40desowinSoap: yes, please do it
15:32:41 Join idnar [0] (i=mithrand@unaffiliated/idnar)
15:33:16Soapdesowin, Loving QTScrobbler, one serious request, one half-hearted request.
15:33:44SoapSerious = Could you put the version number in "Help -> About"?
15:34:34SoapHalf-hearted = How unethical would it be to have a (hidden?) option to force change a Skipped track to Listened?
15:34:42linuxstbTiMiD[FD]: I can't think of a reason why it would be bad, but I think we should double-check with Slasheri that there isn't a reason he didn't do that originally. Maybe it was simply so his code was less intrusive on the existing FAT code.
15:35:10peturmaybe amiconn knows more...
15:35:42TiMiD[FD]well I'm going to bed anyway ...
15:36:14TiMiD[FD]amiconn know every parts of the rb code ;)
15:37:00TiMiD[FD]by the way increasing the main stack size of 0x50 fixes the stack overflow in the mpegplayer on the h1x0
15:37:03 Join aliask [0] (
15:37:05TiMiD[FD]should I commit that ?
15:38:46 Join The-Compiler [0] (
15:39:03PaulJamSoap: can't you achieve your second request by editing the .scrobble.log file?
15:39:09 Quit Nick_Brackley ("CGI:IRC (Ping timeout)")
15:40:03 Quit idnar_ (Connection timed out)
15:40:04 Join donutman25 [0] (n=chatzill@
15:41:14oboSoap: or you can actually edit the items in the QTableWidget - but I'll add a version into the about window
15:41:43oboif desowin doesn't beat me to it ;)
15:42:03TiMiD[FD]obo: are you rewriting the rbutils with qt ?
15:43:13desowinTiMiD[FD]: I came with that idea, but I didn't started it yet
15:43:37TiMiD[FD]that would be nice :)
15:43:39oboTiMiD[FD]: no, but I've thought about it :)
15:43:54TiMiD[FD]I hate gtk
15:44:02desowinwx != gtk
15:44:12TiMiD[FD]wx is based on gtk
15:44:19TiMiD[FD]under linux
15:44:28GodEaterI've not seen a wxQt module certainly
15:44:34TiMiD[FD]and qt is more sexy
15:44:44TiMiD[FD]from the programmer point of view
15:44:44 Join indro_ [0] (
15:44:46indro_hi all
15:45:00 Join XavierGr [0] (
15:45:08XavierGrpetur: ping
15:45:11desowinTiMiD[FD]: from c++ programmer point of view of course
15:45:14indro_anyone using amarok?
15:45:25TiMiD[FD]of course
15:45:26peturXavierGr: I saw your reply, thanks for testing
15:45:31XavierGrah okay nice
15:45:31TiMiD[FD]java is better anyway :)
15:45:36 Join godzirra [0] (
15:45:39godzirraheya guys.
15:45:50XavierGrpetur: you've considered to buy an H115?
15:45:53*desowin stops as he would soon start flamewar
15:46:11TiMiD[FD]just kidding ...
15:46:14godzirraI've used rockbox a lot in the past and am getting ready to try it out on my sansa e280. If I decide I -don't- like it (unlikely, but still), can the rockbox boot loader be uninstaled?
15:46:59desowinindro_: well, it's rockbox channel, anyway, just state your question
15:47:20peturXavierGr: no, a fellow taper bought one as replacement for his broken h120 and wondered if he got a broken one. Our target table doesn't show this (must still edit the wiki)
15:47:45indro_I dont know which modul I should take with ipod/rockbox.
15:47:57indro_any idea?
15:48:32desowinthere's no one which I'm aware of, just copy files directly
15:48:39Kasperleindro_: you mean, to have amarok use it as a "media device"?
15:48:49XavierGrdoes anyone know if H115 is flashable with rockbox?
15:49:09desowinsomeone would need to make amarok plugin to support rockbox database, it would be great
15:49:10godzirraI don't, but you can always exchange it if it gets bricked. }:)
15:49:12XavierGrpetur: the result I got was with a very old build, though I think it doesn't matter
15:49:59peturXavierGr: it is a hardware thing, the plug power is always there (or not, so it seems)
15:50:21indro_Kasperle: right
15:50:33Kasperleindro_: use the "generic audio player"
15:50:58indro_Kasperle: okay, thx!
15:51:01Kasperleindro_: the iPod plugin is only useful if you used iTunes to transfer your music to the ipod
15:51:26indro_never used iTunes. ;)
15:51:30Kasperlegood :D
15:51:36Kasperlegeneric audio player it is, then :)
15:54:16Soapthanks for the tip obo
15:54:21amiconnTiMiD[FD]: The stack overflow needs to be fixed the proper way, i.e. by decreasing stack usage
15:54:26SoapI hadn't figured that one out.
15:54:36TiMiD[FD]a jvm plugin would be nice, I wonder if it's feasible
15:55:15Kasperlethe java micro edition vm is hard enough to compile on full-blown platforms ;)
15:55:24Kasperlecvm or what's it called?
15:56:12TiMiD[FD]j2me ?
15:56:43Kasperlethough it kind of worked on some pentium 166 linux box with ~ 64MB RAM or so
15:57:09TiMiD[FD]it works on my cellphone too
15:57:28TiMiD[FD]which for sure doesn't have 64mb of ram
15:57:43TiMiD[FD]with java going gpl ...
15:58:15Kasperlewell, i wonder when sun will finally do that
15:58:37TiMiD[FD]amiconn: I was trying to figure out what was eating the stack, but unfortunately I gave up :/
15:58:40Kasperlethis is what i once used on a small embedded x86 board :
15:59:13TiMiD[FD]j2meis suposed to go gpl too
15:59:49Kasperlei think there's more than one j2me?
15:59:52Kasperledifferent "profiles"?
16:00:08TiMiD[FD]of course this wouldn't a work for people without a strong motivation
16:00:33*petur would like to ban java from this channel
16:03:16*XavierGr hates java due to his stupid so called "java-professor"
16:03:45*bluebrother hates java because of this stupid we-don't-need-pointers thing
16:04:20XavierGrthe final exam was writing GUI java programs by hand (in a text book) and of course I failed it :(
16:04:25TiMiD[FD]but java is a very nice ptograming language
16:04:39*petur sighs
16:04:54*bluebrother did some java recently and disagrees.
16:05:05bluebrotherhey, how about some java coffee? ;-)
16:05:20XavierGrwhy should I learn by heart all those awt and swing objects anyway? but as I said my professor is brain dead
16:06:13TiMiD[FD]XavierGr: you would prefer to code your gui in asm pixel by pixel maybe ?
16:06:40XavierGrno I would prefer to code my gui program on a pc with an IDE (even in java)
16:06:55XavierGrbut the teacher said to no to final exam in a PC
16:07:14TiMiD[FD]you're lucky then ^^
16:07:24TiMiD[FD]an exam on pc would take more time
16:07:34XavierGryeah but way more easy
16:07:41TiMiD[FD]you don't do the debugging when programing on paper ...
16:08:14*petur had to write pascal and asm programs on paper for exams, with points lost for both compile errors and bugs
16:08:16Kasperlebut you need to get it right without a debugger, too
16:08:32Kasperlemy uni tought me modula-3 in year one
16:08:37Kasperletalking about useless crap
16:09:04bluebrothersome stupid script deleted the source files I created the last two days :(
16:09:52 Quit aliask ("ChatZilla [Firefox]")
16:10:54TiMiD[FD]bluebrother: have you tried to recover the file ?
16:11:12bluebrotherwell, I'm on a journaling file system ...
16:11:23bluebrotherand the file is still present. It's just empty :(
16:11:50TiMiD[FD]ah so it was emptied ^^
16:13:37 Quit indro (Remote closed the connection)
16:13:49TiMiD[FD]could have worked if you were able to remount the partition read only after the deletion ...
16:14:11 Quit relaxed ("")
16:15:09 Join saratoga [0] (i=980398fe@gateway/web/cgi-irc/
16:17:20TiMiD[FD]amiconn: what do you think about the apps that doesn't use the dircache api ?
16:17:43 Join SirFunk [0] (
16:18:05 Quit donutman25 (Read error: 104 (Connection reset by peer))
16:19:49 Join indro [0] (i=indro@
16:20:09TiMiD[FD]some apps as well as lots of plugins doesn't use the dircache api
16:20:12*amiconn wants to stay away as far as possible from java (re the java discussion)
16:20:39amiconnAfaiu apps don't need to use dircache specifically. It is used automatically when enabled
16:20:42TiMiD[FD](they use readdir instead of readdir_cached for ex)
16:21:17TiMiD[FD]it's used when they call the api explicitely
16:21:33TiMiD[FD]if they call the old functions they don't use it
16:21:43TiMiD[FD]linuxstb noticed that
16:22:14TiMiD[FD]sounds more like a bug than something wanted in my opinion
16:23:15amiconnI have no idea. I though it was fully wrapped
16:23:27*amiconn wouldn't hesitate to drop dircache
16:24:04TiMiD[FD]it was also proposed that apps should use the posix api, eventually wrapped by dircache
16:24:19amiconnThey do...
16:24:22TiMiD[FD]instead of calling directly the dircache functions (a cosmetic change)
16:24:29TiMiD[FD]ah no
16:25:30TiMiD[FD]they call dircache functions
16:25:39TiMiD[FD]for mostof them
16:25:59*amiconn suggests to ask Slasheri
16:26:36TiMiD[FD]well I'm going to bed anyway
16:27:04amiconnIt might be that the 2 functions are due to the dircache api hiding the on-disk cache file
16:27:14TiMiD[FD]that would be something I could do tomorrow if I had approvement
16:27:18amiconn(only present with rockbox in flash on H1x0)
16:28:58amiconnWhen rockbox is flashed, rockbox has full control over what happens on disk. So it writes the dircache contents to a file on shutdown, and reloads dircache from that file on next boot (if disk is clean, i.e. no intermediate usb access)
16:28:58TiMiD[FD]it would be more logical to present the posix api to the apps
16:29:20amiconnSo readdir_cached() hides that file, but there must be a way to find it
16:29:21TiMiD[FD]yes I know that :)
16:29:42TiMiD[FD]but apps doesn't need to find it ...
16:29:48amiconnI would appreciate simplification here
16:30:16amiconnThis cache-on-disk still sounds fishy to me, and hence should go away _imho_
16:30:17TiMiD[FD]doesn't needs to be hidden too
16:30:48TiMiD[FD]if we don't store it on dsk where can we put it ?
16:30:58amiconnNowhere. Just scan on boot
16:31:08TiMiD[FD]takes a lot of time ...
16:31:11 Quit ansivirus ("Leaving")
16:31:37amiconnA lot?
16:31:39TiMiD[FD]for flashed units
16:32:02 Quit Entasis (Read error: 104 (Connection reset by peer))
16:32:06amiconnMaximum observed was ~1 minute. That was with a cache size near the imposed maximum
16:32:07TiMiD[FD]on disks filled with files
16:32:14amiconnBut it's a background scan, so what?
16:32:17 Part fred_88
16:32:26TiMiD[FD]1minute of disk activity at each startup is a lot
16:32:32TiMiD[FD]it eats battery
16:32:43amiconnWith a sane amount of files, it's more like 10 seconds
16:32:58 Quit indro (Remote closed the connection)
16:33:12 Join indro [0] (i=indro@
16:33:23TiMiD[FD]anyway this is hidden in the dircache subsystem so ...
16:33:31TiMiD[FD]the question was
16:33:44amiconnAnyway, dircache still shows numerous odd effects, and it's about the first thing to disable for me if it doesn't behave...
16:34:02TiMiD[FD]to know wether apps have to access the dirs without using dircache in some cases
16:34:20amiconnAsk Slasheri....
16:34:31TiMiD[FD]but he isn't here
16:35:03amiconnI don't know much about dircache interactions and implications, and I'm not interested enough in dircache to dig into it
16:35:18TiMiD[FD]it seems that people who are not aware of dircache program using the posix api, which makes sens
16:35:28TiMiD[FD]I understand that
16:36:11TiMiD[FD]the app layer would still be cleaner accessing dircache through the posix api
16:36:36TiMiD[FD]so if Slasheri could comment on that
16:36:46TiMiD[FD]anyway time to sleep
16:38:09SlasheriTiMiD[FD]: yes, the dircache should be wrapped directly over opendir and so on
16:38:37Slasheriat beginning it wasn't stable enough to do that, but now it should be certainly possible
16:38:54TiMiD[FD]I"ll look into that later then
16:39:01Slasherisounds good :)
16:39:20linuxstbSlasheri: So all the apps/ code currently using readdir() should be using readdir_cached() ?
16:39:21TiMiD[FD]what about apps accessing dirs without dircache ?
16:39:23TiMiD[FD]mistakes ?
16:39:42*petur wonders if that would further avoid spinups in the recording screen
16:40:23Slasherilinuxstb: most likely, but firmware/ code could be different
16:40:42linuxstb"most likely" ? ;)
16:40:52TiMiD[FD]so we would have the uncached api renamed to readdir_uncached in fw
16:40:54Slasheriand of course dircache itself can't use it's own readdir function :)
16:41:08linuxstbI'm purely talking about the apps/ code.
16:41:11Slasherilinuxstb: i think so, but haven't checked to make sure of that :)
16:41:15 Quit Tr1ckY ("( :: NoNameScript 4.02 :: )")
16:41:18linuxstb(including plugins)
16:41:36Slasherilinuxstb: yeah, then i see no reason why it couldn't be wrapped directly to readdir
16:41:55linuxstbCurrently the plugin API exposes both readdir() and readdir_cached() (and related functions). It seems to be it should just expose readdir_cached() ?
16:42:06linuxstb(or the equivalent if they get renamed).
16:42:15Slasherilinuxstb: indeed, or just the readdir
16:42:33TiMiD[FD]doesn't sounds too heavy to change
16:42:35Slasheriand there could be some syscall if it _really_ would necessary to access the non-cached version directly
16:42:40*linuxstb thinks about bidding on a 1st gen ipod, currently 26 UKP on ebay with a few hours left
16:42:42Slasheribut i doubt that
16:42:59 Join JdGordon [0] (
16:43:02linuxstbamiconn: Are you close to buying any 1st/2nd gen ipods/
16:45:51amiconnI'm trying to
16:46:06amiconnOne 1st gen auction ends in <3 hours
16:46:07linuxstbAny idea what they are selling for?
16:46:17amiconnNot yet
16:46:48amiconn3rd gen went for around 100EUR (both 20GB and 40GB)
16:48:14linuxstbThe UK one I'm looking at is a 10GB first gen. I'm not sure how much I would be willing to pay though...
16:49:35JdGordonis the info in DeviceChart for the e200 correct? it says it has alarm wakeup and is 100mhz?
16:52:57amiconnPP5024 100MHz is correct as PP5024 is just a PP5022 with tacked on AS3514
16:53:14amiconnBut the of runs it at 80MHz
16:53:23*linuxstb spots a 20GB first gen with a buy it now price of 60UKP
16:54:26amiconnThe 1st due 1st gen is at EUR40 atm (~2h 30 min to go)
16:55:00godzirraI've used rockbox a lot in the past and am getting ready to try it out on my sansa e280. If I decide I -don't- like it (unlikely, but still), can the rockbox boot loader be uninstalled? (Sorry if someone already answered this)
16:56:39godzirraOddly, there are almost no wps's for the E200 :/
16:59:49Kasperlehrm. talking about wps's
17:00:12godzirraAhh cool. thanks linuxstb
17:00:13Kasperlei need to look into why some themes i downloaded only set menu icons, fonts, colors, background image, but _not_ the wps
17:01:43 Join low_light [0] (i=c730190b@gateway/web/cgi-irc/
17:03:11LloreanKasperle: Usually this means they do try to set the wps, but there's bad context in the WPS so it fails.
17:05:56amiconnlinuxstb: Do you know whether the 1st gens are single platter or dual platter. Or did they even use both?
17:06:17XavierGramiconn: do you have an idea of what it takes to make the Ondio boot up automatically when someone plugs in the USB cable?
17:06:37***Saving seen data "./dancer.seen"
17:06:41Kasperledoes rockbox like spaces in wps file (and dir-)names? also, what about DOS-style CRLF in wps files? these are the first two things I noticed
17:07:26amiconndos line endings should be handled properly. Spaces in the path shouldn't matter
17:07:40 Join nls [0] (
17:08:26nlsscorche: I read in the logs that you were handing out cloaks, I want one too :)
17:09:20 Join lazka [0] (
17:11:14XavierGrif that's true I would want one too if it is no hassle for you :P
17:12:44low_lightamiconn: on the pp5020, 0x70000000 should be "PP50"?
17:13:35amiconnThis is wrong in the ipl wiki
17:15:20amiconnThe PP5020E(space) string is actually a single "little endian string", i.e. it starts at 0x70000004 and extends to 0x70000000
17:17:24low_lightyes, that's what I was getting
17:19:19linuxstbamiconn: I have no idea (about 1st/2nd gen disks)
17:20:22linuxstbamiconn: ipodpatcher will need updating to work with the 1st/2nd gens though - the firmware partition format is different. But the old ipod_fw install method should work in the meantime.
17:21:17amiconnI hope you could help with adapting ipodpatcher?
17:21:43linuxstbSure, even if I don't buy one, I can do it based on a disk image.
17:22:14amiconnWe'll need firewire detection for the 1st gen/2nd gen
17:22:17amiconnShould be easy
17:22:34 Join WaddaDoer [0] (
17:22:36linuxstbIIUC, the 3rd gen is also missing any kind of disk mode detection as well.
17:22:46amiconnThe only thing I am uncertain about is whether we want a separate firewire screen or not
17:23:41linuxstbFor devices with only a firewire connection, we could give them a different logo. But yes, I'm not sure about devices with both types of connection.
17:24:00amiconnE.g. the minis
17:24:27amiconnI know how to detect firewire and firewire power on mini 1st gen, mini 2nd gen, and the video (power only)
17:26:07 Quit funky ("Changing server")
17:26:41 Part WaddaDoer
17:27:06 Join WaddaDoer [0] (
17:27:27 Quit low_light ("CGI:IRC")
17:27:41 Part WaddaDoer
17:32:13 Join Soul-Slayer [0] (n=Administ@
17:37:04 Quit PaulJam (".")
17:40:08 Quit idnar (Nick collision from services.)
17:40:10 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
17:40:56 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
17:42:44 Join My_Sic [0] (
17:45:52 Quit JdGordon (Remote closed the connection)
17:52:42 Join spiorf [0] (
17:53:45 Quit Nico_P (Remote closed the connection)
17:59:23 Join Bawitdaba [0] (
18:00:36 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
18:04:49 Quit petur ("work->home")
18:06:20 Join Jens [0] (i=Jens@pdpc/supporter/active/Jens)
18:06:31*The-Compiler hugs all the Rockbox-devs here - great work!
18:23:20 Part xNibbler ("bei uns hat noch niemals nicht koana koan durscht ned ghabt")
18:24:00 Join jhulst [0] (n=jhulst@
18:28:21 Quit obo ("bye")
18:35:08 Join Domonoky [0] (
18:37:19 Quit indro_ ("Lost terminal")
18:41:15 Quit Febs (Read error: 110 (Connection timed out))
18:43:03 Join iriver320user [0] (i=560555d4@gateway/web/cgi-irc/
18:43:09iriver320userhello everyone
18:44:37 Quit iriver320user (Client Quit)
18:44:53 Quit jhulst (Read error: 104 (Connection reset by peer))
18:49:18 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
18:52:55lazkai just saw this on another channel: what about an IRC bot that posts new created bugs/patches and SVN commits?
18:53:39 Join obo [0] (n=obo@rockbox/developer/obo)
18:54:34Lloreanlazka: Why not just use the RSS feeds instead of having it spam the channel?
18:55:31lazkathat's a good point..
18:56:28*dionoea sees a reason: runing IRC is easier than runing IRC + an RSS reader :)
18:58:53*amiconn sees no reason for either one
18:59:06Soul-SlayerWhats wrong with just looking at the front page?
18:59:18amiconnexactly what I a mthinking...
19:00:28godzirraSo another question... Sansa E200 series is supported right? I assume I won't lose my music after installing rockbox since its al ready drag and drop?
19:01:10 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
19:01:33godzirraAwesomeness. Installing now then.
19:01:38jhMikeSamiconn: the sansa can be run at 100MHz? :)
19:02:35 Join Ikuma [0] (
19:03:09amiconnjhMikeS: Well, the CPU is specced for 100MHz. It might be that it gets too hot in the Sansa though
19:03:25dionoeaSoul-Slayer: I think that it depends on the projects. Some projects don't use web interfaces a lot, so having other means of communication like commit emails or IRC messages help in keeping people up-to-date with what's happening. (plus you usually lurk on chans ... but don't reload the same webpage every 5 minute). But then that's only my pov.
19:03:28 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
19:04:00godzirrait says its installing.
19:04:04godzirraHow long does it take to install
19:04:18 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
19:04:31godzirraand does it display a message saying it was installed successfully?
19:04:33Soul-Slayergodzirra: Are you using rbutil or just unzipping?
19:04:37 Join Febs [0] (n=chatzill@
19:04:49godzirraI was using the sansapatcher found in the installation section
19:04:58godzirrahuh... it just told me my device was disconnected.
19:05:04godzirrathen connected again
19:05:10Soul-SlayerOh of course, forgot that
19:05:12godzirrabut sansapatcher said nothing after I hit "i" to install.
19:05:22godzirraIs that normal?
19:05:30Kasperlei think this channel is already pretty high-traffic. adding such a bot wouldn't really help
19:06:04godzirraahh it just exited.
19:06:10godzirraand is now displaying a blue screen of death. :(
19:06:26Soul-SlayerEject your sansa and try booting it
19:06:34Soul-SlayerIf it boots into rockbox, it worked, otherwise it didn't, simple :P
19:06:40***Saving seen data "./dancer.seen"
19:06:43godzirraI'm trying to now. It did boot the rockbox bootloader.
19:06:44Kasperleif people want it so much, you could have it sit in just a chan like #rockbox-commits or so
19:07:04 Join Xerion [0] (
19:07:06Kasperleas a compromise
19:07:22godzirraWhew. that ugly bluescreen after it finished instaling the bootloader was scary.
19:07:29godzirramight want to mention that on the manual :)
19:07:43Soul-SlayerI think it's probably a one off
19:07:49Soul-SlayerOtherwise someone would have sorted it out
19:08:20Lloreangodzirra: What bluescreen?
19:08:46LloreanLike, what exactly did it say?
19:09:29godzirraLlorean: nothing.
19:09:35godzirrait looked like a "your thing has crashed" screen.
19:09:36godzirrait was blue
19:09:42godzirraand was kind of ...staticky.
19:09:46godzirrawas very odd.
19:09:51godzirraSoul-Slayer: could be.
19:10:38LloreanThat screen isn't usual.
19:10:42godzirraalthough I can't seem to extract the .rockbox directory to my player...
19:10:53godzirrawhat usb mode should it be in?
19:10:53godzirrafor that?
19:11:10LloreanIt needs to be in MSC for everything Rockbox related.
19:12:02Soul-Slayerdionoea: All very true, but this chat is more for discussing undocumented features and how devs could get feature such and such working, having all the commits posted up here would be kinda needless spam in my opinion. Another channel devoted to it, perhaps, if someone was interested in taking the time to get a bot to parse it all properly. But the rockbox main page refreshes incredibly fast, if I want to know I can find out
19:13:11godzirraAn rss feed would be better for that imo.
19:13:14 Quit linuxstb (Read error: 113 (No route to host))
19:13:19Soul-SlayerAlready exists doesn't it?
19:13:22godzirraThat way anyone could subscribe and not have to be connected on irc.
19:13:31dionoeayeah, i agree for the rockbox case. People use the web interface / trackers a lot so it's ok.
19:13:33godzirraOh. then a bot is totally unnecessary... again imho :)
19:13:35LloreanYeah, there's an RSS feed for SVN changes, and an RSS for tasks added to the tracker.
19:14:04dionoeaI was just mentioning why it could be useful to other projects.
19:14:16Soul-SlayerOr I suppose a bot that PMed you commits, if you requested it...
19:14:22dionoea(where most people complain when the bot unexpectedly quits :) )
19:16:15tumuis HowToWriteCodecs wikipage uptodate? i was planning on adding .mod support using dumb lib
19:18:33godzirraAfter installing rockbox and rebooting into it...
19:18:38godzirraI got an error when plugging in usb.
19:18:47godzirra"One of the usb devices attached to this computer has malfunctioned"
19:19:04godzirraand no usb device popped up
19:19:22godzirra"and windows did not recognize it. For assistance in solving this problem, click this message.
19:19:22tumudoesn't support usb yet
19:19:33tumuyou have to use the old fw to upload files
19:19:36godzirraso I have to reboot into the .. ok.
19:20:28 Join dandin1 [0] (
19:21:04 Quit Bawitdaba (Read error: 110 (Connection timed out))
19:22:05godzirraDoes it support the sd slot?
19:23:25godzirraHow can I check the status of usb? I assume its being worked on?
19:24:03tumui don't know if anyone is working on it currently
19:24:44tumuatleast i haven't noticed anything being done on it for past few months
19:25:24dionoeait's being worked on as a Google Sumer of Code project
19:25:33Lloreangodzirra: It's the same case as the iPod USB.
19:30:14godzirraLlorean: Yeah, just irritating to have to reboot into sandisk firmware everytime. ;)
19:30:39godzirrawith the ability to run in drag and drop usb mode, I may just stick with sandisks firmware until then.
19:30:51 Join GodEater__ [0] (
19:31:31LloreanRebooting into Sandisk's firmware isn't exactly hard.
19:31:40Llorean1. Turn off device. 2. Insert USB cable.
19:32:38godzirraoh, it boots up automatically if the usb cable is inserted?
19:32:48tumuis the codec compiling only supported in flat folders?
19:33:07tumuso do i need to take the original sources and put all necessary files in one folder?
19:33:49Lloreangodzirra: Yes, and if the bootloader detects a usb cable it chooses to boot the retail firmware. Which will exit and reboot to Rockbox when you eject and unplug after transferring files
19:36:03tumubefore i venture into hacking the codec main makefile into submission :)
19:36:08godzirraLlorean: ok, cool. it didnt reboot when I plugged it in, it just moved into a weird usb page that didnt actually work, so i didnt try that.
19:36:15godzirrabut it was already on as well.
19:36:31LloreanYes, once it's on you just get the USB logo and nothing happens.
19:37:10godzirraYeah, thats what I was getting.
19:38:06 Join Bawitdaba [0] (
19:38:56tumuno one who knows about the make procedure on codecs around?
19:39:04 Quit lazka ("I'm off now")
19:39:38bluebrothertumu: why do you want to build the codecs separately?
19:39:49tumui want to add new codec
19:40:01bluebrotherthey need to match the build, so I don't see any reason why someone wants to build it on its own.
19:40:12bluebrotherwell, there is a wiki page about writing codecs.
19:40:20tumuthe instructions on wiki are very terse
19:40:35tumuand what i can see from the actual rockbox sources, it needs more than the wiki says
19:40:45saratogajust decoded a 20kbps wma file in fixed point
19:40:49 Join bdgraue [0] (
19:40:58bluebrotherthat's possible. I haven#t worked on codecs.
19:41:01tumusure, i can waste time learning on how rockbox does it, but i thought if someone could tell me
19:41:21tumuor update the wiki
19:41:27bluebrotherchecked the KnowledgeMap wiki page?
19:41:37 Join robin0800 [0] (
19:41:40tumujust the HowToWriteCodecs page..
19:41:42tumuthere is another?
19:42:14bluebrotherwell, this page holds (or should hold ;-) a list of names and their knowledge of Rockbox.
19:42:45saratogatumu: you want to talk to linuxstb
19:42:54bluebrotheraccording to that, preglow and Lear are your men.
19:43:02saratogahe did almost all of the wma API stuff for me
19:43:09saratogasince i couldn't figure it out
19:43:32saratogawhat codec anyway
19:44:33bluebrotherah, dumb. We had that in svn for a while.
19:44:42LearYep, pretty sure about that.
19:44:48bluebrotherbut as it wasn't working it got removed at some point.
19:44:58tumuwell, i plan to get it working :)
19:45:10saratogathe old svn code might be useful
19:45:24LearAs for how to write a codec, best way is to look on a working one.
19:45:28*bluebrother was about to suggest the same
19:46:04tumuLear, i can do that, just that it would speed things up knowing what i need to do
19:46:21tumui'm not novice in coding :)
19:46:55bluebrotherdumb was removed on 11th Dec.
19:47:34 Quit linuxstb_ (Read error: 110 (Connection timed out))
19:48:06LearHm, the wiki shows all places where information about the codec needs to be added.
19:48:46tumuLear, do i flat the original sources to a single folder?
19:49:18tumualso do i need to copy and edit makefile from existing codec to make it compile?
19:49:44bluebrothertumu: check the old svn version −− it compiled those days
19:49:54bluebrotherit just wasn't working as codec
19:51:09tumulooks like that old svn only had the dumb sources added
19:51:12tumunothing else done
19:51:20tumumissing makefiles/sources etc
19:53:02bluebrotherwell, the build process was building it. Removing dumb made compiling quite faster ;-)
19:53:17tumuit was?
19:53:37bluebrotherwell, I don't know the details. I just remember that it was built.
19:54:39tumuhmm i see
19:54:44tumuthe original makefile was hacked
19:54:55 Quit robin0800 (Read error: 104 (Connection reset by peer))
19:55:41tumunot sure if that is the current practise nowadays
19:55:45tumui can do that tho :D
20:00:17tumuno comments?
20:01:31bluebrotherwell, either nobody objects or those who would do aren't around ;-)
20:03:11tumuyeah.. tho no use submitting patches if they don't conform to current practise
20:03:50bluebrotherwell, you could start with the old Makefile and ask later. If there are objections you'll always can adjust the Makefile
20:04:26bluebrotherand as it worked that way I don't think there will be too big objections if you do it the same way.
20:06:43tumumaybe, tho the original makefile contains lot of crap not needed
20:11:33 Quit saratoga ("CGI:IRC")
20:13:52 Quit ptw419 ()
20:16:47 Quit GodEater (Nick collision from services.)
20:16:55 Nick GodEater__ is now known as GodEater (
20:17:19 Join GodEater__ [0] (
20:19:17 Join saratoga [0] (i=980398fe@gateway/web/cgi-irc/
20:20:29 Join mirak [0] (
20:22:41 Join jgarvey [0] (
20:27:54 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
20:32:27 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
20:36:53 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
20:48:27 Join robin0800 [0] (
20:50:22 Join Siltaar [0] (
20:50:44 Join robin_0800 [0] (
20:54:43 Part Ikuma
20:55:18 Quit amiconn (Read error: 110 (Connection timed out))
20:55:18 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
20:56:44 Join [1]robin0800 [0] (
20:56:51 Join godzirra_ [0] (
20:57:01 Join tucoz [0] (n=martin@rockbox/staff/tucoz)
21:01:39tucozbluebrother, are you here?
21:01:40 Join Phill [0] (
21:02:04 Join [2]robin0800 [0] (
21:02:04***Alert Mode level 1
21:02:04DBUGEnqueued KICK robin0800
21:02:04DBUGEnqueued KICK robin_0800
21:02:04***Alert Mode level 2
21:02:04DBUGEnqueued KICK [1]robin0800
21:02:04DBUGEnqueued KICK [2]robin0800
21:02:04***Alert Mode level 3
21:04:34 Join [3]robin0800 [0] (
21:04:34***Alert Mode level 4
21:04:34***Alert Mode level 5
21:04:34DBUGEnqueued KICK [3]robin0800
21:04:34***Alert Mode level 6
21:04:34***Alert Mode level 7
21:04:34***Alert Mode level 8
21:04:34***Alert Mode level 9
21:05:40 Quit [3]robin0800 (Read error: 104 (Connection reset by peer))
21:06:44***Saving seen data "./dancer.seen"
21:07:28 Quit robin0800 (Read error: 110 (Connection timed out))
21:08:01 Quit godzirra (Read error: 110 (Connection timed out))
21:10:10tumuhm, any reason why make is not entering new codec folder even when makefile, sources and edit of root codec makefile is done?
21:11:49 Quit Phill ()
21:11:58tumuonly thing missing is the stub codec.c file in codec root
21:12:20tumubut i doubt that would prevent it compiling the actual codec sources
21:13:17 Quit robin_0800 (Read error: 110 (Connection timed out))
21:14:35***Alert Mode OFF
21:18:47tucozi think you have to change the configure script as well, but i might remember wrong
21:18:59 Quit [1]robin0800 (Read error: 110 (Connection timed out))
21:19:15tucozcheck the initial commit of the wma codec, what files were changed then
21:25:05tumucan't see anything relevant in configure
21:25:32tucoztumu, ask linuxstb when he comes around. he has written several codecs
21:26:21 Join My_Sic [0] (
21:26:41 Quit [2]robin0800 (Read error: 110 (Connection timed out))
21:26:55tucoztumu, did you check what was committed in the wma-codec initial commit? otherwise, check out the july 3d on this page:
21:27:27tucoztumu, yeah, you have to add the codec to the codecs in configure
21:28:05tumuwhere is this line?
21:28:37tucozjust check the diffs in the commit i mentioned ^^
21:29:30tumumy copy of tools/configure has no such
21:30:17amiconntucoz: This was removed on Monday
21:30:27tucozaha, sorry
21:33:23 Join low_light [0] (i=c730180b@gateway/web/cgi-irc/
21:33:29amiconnThis should tell you what's neede
21:33:47tucozamiconn, just took a peek on the frontpage of :)
21:34:24amiconn3 days and still on the frontpage, tsss....
21:35:16tucoztumu, are you porting a new codec?
21:35:53CtcpIgnored 5 channel CTCP requests in 2 minutes and 30 seconds at the last flood
21:35:53*amiconn received his H10 today :)
21:35:59tumui must be doing something wrong as i have basically done the same as with wma codec
21:36:02amiconnNow for some more PP5020 hacking...
21:36:22tucoztumu, cool. may i ask what codec that is?
21:36:51tumu.mod support
21:36:54tucozoh. we had that in rockbox earlier. that is, it compiled but didn't run
21:37:07low_lighttumu: did you add a SOURCES file to your codec directory?
21:37:13tumui did add
21:37:15Arathisanother dev with an h10 :3
21:37:25Arathisamiconn: 5&gb or 20gb?
21:37:57Arathisglad they are not soo different ..
21:38:21amiconnI hope to be able to find out what's needed to make the radio work
21:38:33tumumake is not entering the folder and i'm not getting any errors either
21:39:13amiconnBut first I need to install rockbox...
21:39:20amiconn...on my target #9 :P
21:40:47 Join PaulJam [0] (
21:41:13Arathisamiconn: is the freeze-error fixed allready? last time I checked it wasn't :(
21:41:37Domonokyamiconn: thats only 9 out of 23 targets . :-)
21:41:59amiconnI will (hopefully) receive target #10 out of 24 soon though :)
21:42:59mbishopDid anyone ever screw around with rockbox on a neuros?
21:44:41GodEatermbishop: there's a thread on the forums about it
21:45:08mbishopAh, well I have a (slightly broken) neuros II if someone wanted to screw around with one
21:45:22pixelmaalso a bit of info here (don't know how outdated this one is though)
21:45:36*jhMikeS is still waiting for word on an H10 deal :\
21:46:21 Quit Nico_P (Remote closed the connection)
21:46:50jhMikeSamiconn: I ran into something interesting where calling sleep(0) from the COP caused a prefecth abort at C0EDBABE on e200. This could be telling of something. No idea what atm.
21:46:53nlstumu: have you seen this ?
21:49:41nlsjhMikeS: minor nitpick, didn't you break the plugin api back compatiblity for single core swcodec targets without bumping bersion number ?
21:49:43jhMikeSamiconn: MpegPlayer has cross-core mailboxes now (not the PP built-in ones yet). Try testing that.
21:50:20jhMikeSnls: it only should break the compatibility with test_codec which isn't standard. nothing else should notice.
21:50:42nlsjhMikeS: ah, ok :-)
21:52:37amiconnBumping the api version wouldn't have hurt
21:53:43tumunls, haven't tho i'd like it to support more formats and be accurate in its output
21:54:07tumunls, so any old player sources aren't exactly what i'm looking for
21:54:51amiconntumu: Ideally we want both (after the playback buffering rework)
21:55:26 Join Tr1ckY [0] (
21:59:28 Quit low_light ("CGI:IRC (Ping timeout)")
22:00:47 Join low_light [0] (i=c730190b@gateway/web/cgi-irc/
22:02:29tucoztumu, i think dumb in rockbox would be excellent :)
22:02:54tumuwell, the make is no cooperating with me, so have to wait for someone more details on it
22:03:12 Quit Febs (Read error: 110 (Connection timed out))
22:03:42low_lighttumu: I thought dumb was full of mallocs and floats
22:04:06tumucould be
22:04:21tumui have time to fix it
22:04:22 Quit low_light (Client Quit)
22:04:39 Join low_light [0] (i=c730190b@gateway/web/cgi-irc/
22:07:27amiconnHmm, H10 USB is quite slow
22:09:12 Quit pondlife ("night all")
22:12:00amiconnWhen did the time setting screen become so buggy? :(
22:12:26nlsamiconn: could be my fault, what is the problem?
22:12:53amiconnThe first letter of each part in the date line is cut off
22:13:18bluebrothertucoz: now I'm in.
22:13:29amiconnSo I have to set "007 uly 9" if I want to set 2007 July 19 (and guess I got the missing parts right)
22:14:10tucozbluebrother, hi. i read christis remarks on the manualstructuretalk page, but do not agree in full to what she has to say
22:14:23nlsamiconn, I'll look into it, might have to do it tomorrow tho
22:14:25tucozi agree that the manual should be simple (hence the quick start chapter)
22:14:32amiconnI think it depends on the font. Nimbus 10 here
22:15:01tucozbut other than that, i think the manual should be a reference on what rockbox has to offer
22:15:04bluebrotherwell, I also don't agree 100%.
22:15:04 Quit jhMikeS (Connection reset by peer)
22:15:08amiconnAnother JdGordon bug :(
22:15:35bluebrotherIMO Rockbox is some highly technical thing, so we should be able assuming some basic knowledge by the user
22:15:49bluebrotherincluding that he read about what it exactly is ;-)
22:16:13tucozi especially don't find rockbox to be a highly complicated piece of software. it might have been in the past, but with the introduction of the main menu i find it quite intuitive and simple
22:16:44tucoz(with a wealth of complex options and configuration possibilities)
22:16:46bluebrotherwell, "highly complicated" also depends heavily on the technical background of the reader / user
22:17:06nlsamiconn: ah, nasty assumption that monthnames are 3 letters, but I am quite sure that we assumed that before too, maybe it just dropped the last char then... (german uses longer names than english...)
22:17:17 Join Wiwie [0] (
22:17:22amiconnIt's not only the month...
22:17:26tucozit does, but the main menu makes it quite intuitive, don't you think?
22:17:32bluebrotherbut it could be simpler. Like "browse wps" might be confusing, as users should always load complete themes
22:17:39amiconnAnd the _first_ char is missing, not the last
22:18:00tucozyes, it could. but to play a track is quite easy
22:18:10nlsamiconn: yes I noticed, will fix tomorrow, now sleep, gnight
22:18:19bluebrotherthe main menu adjust it more to the way other firmwares are doing it, i.e. it comes more to the way the user expects.
22:18:19 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
22:18:21tucozit might be the customization that is hard.
22:18:30tucozchange font, wps etc
22:18:57bluebrotheryes. But maybe we could think about shipping some standard configuration files, depending on the usage?
22:19:09bluebrotherrbutil could even install a basic config.cfg
22:19:19tucozi think we could/should
22:19:37bluebrotherlike give an installation option "install Ipod style configuration / Rockbox style"
22:20:10tucozor call it optimized configuration ;)
22:20:30bluebrotherthat default configurations could be put somewhere on the download server, so we not even need to add something to the builds.
22:20:37bluebrotherhmm, I'm starting to like that idea.
22:21:12tucozi like it as well. we can configure it to use a pretty theme, turn on some sensible defaults and so on
22:21:40bluebrotheryes. And I bet quite some Ipod users won't even bother looking at the settings ;-)
22:22:13tucozand i don't think such an option would be too hard to convince the core devs of as well, as it doesn't interfere with the core rockbox
22:22:22tucozi mean svn
22:22:57bluebrotheras we have cfg files for the runtime settings now −− I think rbutil could safely do such things.
22:22:57tucozbut still provides newbies with something that sounds like a BoS, but doesn't look like it
22:23:10bluebrotherwithout real objections ;-)
22:23:59tucozi wonder what a first time gigabeater thinks of the first boot into rockbox. the font must be almost invisible
22:24:22bluebrother"has it booted? Had something happened?" ;-)
22:24:24 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
22:24:30 Quit Arathis ("Bye, bye")
22:24:46tucozis there a wiki where we can discuss this?
22:25:11bluebrothermaybe the RockboxUtility wiki page?
22:25:19tucozand agree on certain aspects such as default theme etc
22:25:44*Domonoky detects talk about rbutil :-)
22:25:56*bluebrother waves at Domonoky
22:26:44 Join Wiwie [0] (
22:27:15bluebrotherI also had this idea of having rbutil only shipping a base functionality and plugins to add additional features
22:27:29bluebrotherbut I guess it isn't worth the work ...
22:27:39tucozto strip out plugins?
22:27:50bluebrotherno, plugins for rbutil.
22:28:16Domonokyqt has a plugin system, but wx... i dont know.. :-)
22:28:18bluebrotherthat way people could implement an unsupported builds plugin
22:28:24tucozi think rbutil should be as simple as possible
22:28:27bluebrotherDomonoky: I'm already playing around with Qt ;-)
22:28:27tucozbluebrother, hehe
22:29:00Domonokyi also would like an rbutil in Qt.. but someone has to do the work :-)
22:29:04bluebrotherbut I haven't started with the network stuff. Wanted to do this today, but my tree crashed :(
22:29:37bluebrotherwell, atm it's mainly a programming exercise for me. Once I have the basic installation working (if I ever get that far) I'll release it somewhere
22:30:14bluebrotherand then we'll see if there is enough interest / support in switching toolkits.
22:30:15tucozthat would be fun, and a nice opportunity to learn some qt
22:30:27bluebrotherthat's what I'm doing it for
22:30:49bluebrotherlearning a programming language without an interesting project isn't half as productive
22:30:59tucozi spend to little time in linux these days, have to change that habit
22:31:57ender`whee, my sister's ipod mini died again
22:32:00bluebrotherDomonoky: btw, I added a "Quick Start" page to rbutil. Is this worth to look into?
22:32:55Domonokybluebrother: it would be nice if you could try to get the basic functions of rbutil in Qt.. like downloading function, unzipping.. settings ..usw and a GUI, then we could slowly port the rbutil functionality
22:33:02 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
22:33:16bluebrotherthat was my intention.
22:33:34bluebrotherbut I'm quite new to Qt / cpp so it takes some time.
22:33:53Domonokybluebrother: with Quick Start, you mean some help text to start ?
22:34:18bluebrotherno, some buttons for a first start: "Full install", "basic install", "update"
22:34:40bluebrotherso you don't have to install bootloader, build and fonts separately (but still can by using the installation tab)
22:34:58Domonokyah, jeah thats nice.. it would also be worth to rethink to ui, if you are rebuilding it :-)
22:35:50bluebrotherwell, I'm currently using designer to create the gui, so changing shouldn't be much work.
22:36:05bluebrother(and I need to admit that I start to like the concept of designer)
22:36:11tucozwhat version of qt are you developing in?
22:36:18bluebrotherqt 4.3
22:36:27Domonokythe momentary UI evolved with the features rbutil got, maybe there is a better way to arrange the UI..
22:37:18tucozthe current UI works, but it feels like it could follow ui standards a bit more
22:37:36bluebrotherin what way? More wizard-like?
22:37:55Domonokyit was wizard like, and it wasnt good :-)
22:38:00*amiconn wonders whether rockbox on H10 could reboot into OF like it does on ipod
22:38:13bluebrotherI know ;-)
22:38:17 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
22:38:28bluebrotherthe wizard was too limited.
22:38:29*amiconn will put on some music and try to reproduce freezing effects
22:39:35bluebrotherhmm, what's the default port of svn?
22:40:00ender`depends on server
22:40:11Domonokygoogle tells me default is 3690 ;:-)
22:40:31ender`80 if it's through http/webdav, 443 if it's https, no idea what the svnserve is
22:41:42 Join My_Sic [0] (
22:41:46jhMikeSamiconn: any chance you are going to want to try "the patch" any time in the near future? lemme know and I'll resync.
22:42:07 Quit andrew__1 (Read error: 110 (Connection timed out))
22:42:10 Quit The-Compiler (Read error: 104 (Connection reset by peer))
22:43:04amiconn"the patch"?
22:43:28jhMikeSthat core lock one
22:44:30amiconnI tried it on mini 1st gen, where it didn't work
22:44:37amiconnDid you change something meanwhile?
22:44:44bluebrotherDomonoky: you're working on windows, right?
22:44:48*ender` yawns
22:45:18 Quit Tr1ckY ("( :: NoNameScript 4.02 :: )")
22:45:24Domonokybut that shouldnt be a problem with Qt..
22:45:44 Quit SirFunk (Connection timed out)
22:45:57bluebrotherno −− I was just thinking if I can get some scm running on that uni box. But I'm only a user on that machine.
22:45:58jhMikeSamiconn: no, since there's no reason I've found it shouldn't. it might need a resync and that's all.
22:46:14bluebrotherand my local machine isn't accessible from the internet (except via ipv6)
22:46:29 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
22:47:08jhMikeSsince, I've managed to find a way to blow up PP5024, I guess I have something to work on in the meantime in the scheduler
22:47:10 Join relaxed_ [0] (n=relaxed@unaffiliated/relaxed)
22:48:13 Part nls
22:49:59 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
22:50:42amiconnjhMikeS: Blow up PP5024??
22:51:04Domonokybluebrother: hm, maybe some can provide us with a svn for rbtutilQT, or maybe a branch in the rockbox svn .. :-)
22:51:22jhMikeSI called sleep(0) from the COP thread in mpegplayer ... prefetch abort at C0EDBABE :)
22:51:34amiconnAH, that one
22:51:34BagderDomonoky: why a branch, why not just a different subdir?
22:51:45*amiconn wonders where rockbox uses that magic
22:51:52bluebrotherI was also thinking about git, but I don't think that is an option on windows.
22:51:57jhMikeSit's build into the pp hardware
22:52:00DomonokyBagder: that would also work.. perhaps its better.. :-)
22:52:08Bagderno need to complicated it, imho
22:52:09jhMikeSthat one and CACAD0D0
22:52:11 Join relaxed__ [0] (
22:52:44*amiconn wants b490f517
22:53:08*Domonoky could use git in the the vm with linux.. but thats another scm for me to learn :-)
22:55:18saratogaanyone know when preglow will be back?
22:55:54 Part tucoz ("god natt")
22:56:06*bluebrother starts windows
22:56:45jhMikeSCACAD0D0 reads out from meaningless registers it seems and the other trying to execute from somewhere
22:57:12*Domonoky goes sleeping
22:57:15amiconnHmm, meaningless registers is usefull
22:57:31amiconnMaybe I should try to make a register map
22:57:41 Quit Domonoky (Read error: 104 (Connection reset by peer))
22:58:08 Quit RaRe (Read error: 104 (Connection reset by peer))
22:58:18jhMikeSI've used that to map small areas. C0EDBABE seems related to "coprocessor" somehow (of course)
22:58:41LloreanI've gotten C0EDBABE in the past
22:58:58LloreanI'm trying to remember when though
22:59:21*jhMikeS remembers but not re: pp chips :)
22:59:52LloreanI think it was iPod Nano, actually
23:00:06LloreanYeah, it was
23:00:11 Quit linuxstb (Remote closed the connection)
23:00:12LloreanDEADBEEE and C0EDBABE on the Nano
23:00:26jhMikeSDEADBEEE ?
23:00:31 Join linuxstb [0] (i=5343d4aa@rockbox/developer/linuxstb)
23:01:01LloreanIt may have been "DEADBEEF" and I read wrong, though
23:01:02jhMikeSnever seen that
23:01:05 Join Davide-NYC [0] (
23:01:05 Quit desowin ("use linux")
23:01:14*linuxstb got distracted this evening and missed a 1st gen ipod sell on ebay for 26UKP :(
23:01:23LloreanIt was a prefetch abort.
23:01:53LloreanIt was a bug relating to voice.
23:01:56jhMikeSmaybe that's shown from the CPU? (guess)
23:02:14 Part low_light
23:03:02LloreanIf I recall the address was deemed to be coincidental (or at least, unrelated to the actual problem)
23:03:04LloreanSo yeah.
23:03:47 Quit relaxed__ ("")
23:03:49 Quit relaxed (Read error: 113 (No route to host))
23:04:32 Quit relaxed_ (No route to host)
23:04:43 Join stripwax [0] (
23:04:54Davide-NYCHas anyone thought about implementing UNILINK functionality into rockbox?
23:05:06pixelmaDavide-NYC: re. forum post - the owner of the FM recorder doesn't have so much choice in "encodings" - only "quality" (but channels, different input too) - just fyi
23:05:09Davide-NYCI searched the logs and found a couple of quick references
23:05:27 Join RaRe [0] (
23:05:47pixelmaatm - there is a wavrecord plugin too, though
23:05:50*GodEater directs linuxstb at his messages in the forum
23:05:50Davide-NYCpixelma: Hopefully "old yeller" will figure it out
23:06:36linuxstbDavide-NYC: What's UNILINK ?
23:06:44 Quit thegeek (Read error: 110 (Connection timed out))
23:06:46***Saving seen data "./dancer.seen"
23:06:53Davide-NYCunlink is a sony CD changer communication protocol.
23:06:53stripwaxsony interconnect protocol?
23:07:06Davide-NYCgnunilink exists (open source)
23:07:15pixelmaI really didn't want to post in that thread anymore...
23:07:31Davide-NYCand there is this (anciant) page of a rockbox player controlling a sony Head Unit
23:07:43stripwaxmmm, not the other way around?
23:07:57Davide-NYCstripwax: yes, the other way around, sorry
23:08:15stripwaxvery cool
23:08:21*stripwax would find two uses for this!
23:08:55Davide-NYCI just bought a POS car that has such a stereo unit
23:09:33stripwaxI have an old spare
23:09:50amiconnIt's impossible to leave the audio debug screen on H10
23:10:38 Quit mirak ("Ex-Chat")
23:10:45linuxstbGodEater: What messages?
23:10:55 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
23:11:03Davide-NYCstripwax: in theory could this be done without having to cobble together a GNUnilink dongle?
23:11:06stripwaxDavide-NYC - so, how would unilink "in rockbox" work? you'd have to have a physical interface between the rockbox-enabled DAP and the sony headunit
23:11:29Davide-NYCwell I was thinking the remotes maybe?
23:11:41GodEaterlinuxstb: on the forum ?
23:11:46stripwaxok, so you still need to build that interface
23:11:47GodEaterI sent you one earlier
23:11:50stripwax^you would
23:12:01Davide-NYCOK, not the end of the world
23:12:03stripwaxipod remote support is not even in rockbox yet
23:12:33GodEaterat least I thought I did
23:12:36tumulinuxstb, is there anything particular to do to have make compile new codec sources?
23:12:43GodEaterif you haven't got it - I've no idea where it's gone
23:13:07 Join petur [0] (n=petur@rockbox/developer/petur)
23:13:27Davide-NYCstripwax: I know, neither is the gigabeat port stuff. But there is always a beginning.
23:13:34linuxstbGodEater: Ah, private message ;)
23:13:47GodEatersorry yes
23:14:20linuxstbtumu: Did you create a .c file for your codec in apps/codecs/ ? I read the logs, and maybe your lib isn't being built because no target is depending on it.
23:14:41tumulinuxstb, i have not done that, only added the sources
23:14:48tumuis it dependent on that?
23:15:07tumuwon't try to compile the codec sources without root .c?
23:15:41linuxstbtumu: But my approach to porting codecs has always been to get it working as a standalone C program first, and working in the Rockbox way - i.e. no (or very limited) malloc, no floats, and processing data from a memory buffer rather than reading from dsk directly.
23:15:59Davide-NYCI'm going to drop a feature request on the tracker for this Unilink stuff.
23:16:22linuxstbtumu: Only after I've done that (which is far easier outside of Rockbox), would I try to integrate it.
23:16:28tumulinuxstb, well, if i only could get it to error out because of original sources
23:16:49tumuit is not recognizing it currently at all
23:17:04tumunot much sense in doing work on the sources if compiles don't do anything
23:17:11*amiconn notices that the H10 lcd is (somewhat) readable without backlight :)
23:17:18tumuthat is, not even giving errors :)
23:17:19stripwaxDavide-NYC - mm, isn't the feature a hardware one?
23:17:39stripwaxDAPs don't usually have a freeform-serial output/input
23:17:39Davide-NYCmaybe. I must be missing something
23:18:00Soul-Slayeramiconn: Only in broad daylight at the right angle :p
23:18:25amiconnFluorescent light only atm...
23:18:34Davide-NYCso some hardware has to be designed to convert the unilink data to the dap's remote codes for the corresponding functions?
23:18:45 Quit dandin1 ()
23:18:55*Davide-NYC confused (as usual)
23:19:16Soul-SlayerMust be somewhat intense then, I can only just make out the seek bar on mine in my artificial likght
23:19:39 Quit GodEater ("User pushed the X - because it's Xtra, baby")
23:19:43tumuscorche, read it, it is not current
23:20:09scorchehrm.... hcs seemes to be using it just fine a bit ago...
23:20:51tumumy problem is that the new sources are not getting compiled
23:21:03tumueven after adding supposedly all necessary changes
23:21:18stripwaxDavide-NYC - yep, and that's what gnunilink seems to be - the hardware
23:21:27tumuwhich are more than what the wiki page tells
23:21:33Lloreantumu: There's a GBS patch in the tracker, you could look at what it does.
23:21:47Lloreantumu: Also, will you be updating the wiki page then with the things you've discovered need to be added?
23:22:17tumuif i can get this working first :)
23:22:55linuxstbtumu: What have you added to apps/codecs/Makefile?
23:23:16*jhMikeS wonders why encoders aren't included in that wiki page. it's similar.
23:23:17tumumade a copy of existing codec entries
23:23:27tumuand edited accordingly
23:23:29saratogalinuxstb: I decoded a 20kbps WMA file with only a tiny bit of distortion
23:23:35saratogagetting close
23:23:44tumuonly thing i'm missing is the root stub for codec
23:23:53linuxstbsarataga: Bet they sound great :) But congratulations.
23:24:08 Join dandin1 [0] (
23:24:37linuxstbtumu: Yes, but did you add a line such as "$(OBJDIR)/mpa.elf : $(OBJDIR)/mpa.o $(BUILDDIR)/libmad.a $(OBJDIR)/codec_crt0.o" - i.e. one that specifies your library needs to be built.
23:24:52*jhMikeS just remembered adding encoders is way more complicated than decoders :P
23:25:12*linuxstb just greps the entire source for "flac" and does the same thing for any new codec
23:26:28tumulinuxstb, yes, i added such, made a copy of the spc codec ones
23:28:00linuxstbAnd added your codec .c file to apps/codecs/SOURCES ?
23:28:13linuxstbIf not, then it won't appear in the list of targets to build.
23:28:16tumui have not done that
23:28:20linuxstbAnd hence your lib won't be built.
23:28:22tumubecause i still don't have the stub .c
23:28:32tumuso it is dependent on the stub .c
23:28:51linuxstbYes - look at the "all:" line in the Makefile, and work back from there.
23:29:16amiconn'make' is lazy, and only builds what's necessary for the explicit final output(s)
23:29:19linuxstbi.e. all: specifies $(ROCKS), which is simply the .c files listed in SOURCES, renamed from .c to .codec
23:29:34tumulinuxstb, can you update the wiki page about making codecs then? :)
23:29:51tumucould have saved some time knowing this
23:30:19scorchetumu: that /is/ in that page...
23:30:51 Join Wiwie [0] (
23:30:57tumuit doesn't say anything about adding folders of source
23:31:14tumuonly about one source file
23:31:14amiconnHmm. The H10 freeze isn't a plain freeze, it seems
23:31:50 Join SirFunk [0] (
23:32:36Soul-SlayerA flavoured freeze you mean :o? Please say it's chocolate :O
23:33:00tumulinuxstb, ya, working nicely now
23:33:09tumujust as expected
23:33:23stripwaxamiconn - sometimes i find ipod 5g debug menu can't be quit too
23:33:32tumulinuxstb, thanks for your input
23:34:07 Join thegeek [0] (
23:34:46Davide-NYCFor the core devs: Is there the potential for having serial IO (with or without hardware modification) on all rockboxable players?
23:35:04Davide-NYCIf not, which ones would be left out?
23:35:31bluebrotherDavide-NYC: h100 and h300 don't have a serial port exposed
23:35:44bluebrotherIpods have one, and afaik archos too.
23:35:44stripwaxbluebrother - but with a mod they could have..
23:35:51bluebrotherdon't know about other players.
23:36:02Soul-SlayerDo ANY players have one exposed o.0?
23:36:07bluebrotherindeed, you could mod it. But where to put the connector?
23:36:17bluebrotheryes, Ipods. Though the dock connector
23:36:24Soul-SlayerAh right
23:36:38Soul-SlayerWell I'm sure my gigabeat and H10 don't have one
23:36:39bluebrotherand afaik the archos uses a serial connection for their remote. Don't know which ones exactly though
23:36:45Soul-SlayerNot that I have a cradle for my H10 though
23:37:41Davide-NYCbluebrother: on the bottom plate?
23:38:01 Quit davina (Remote closed the connection)
23:38:02Davide-NYCmaybe between the reset and power plug?
23:38:21bluebrotherI'm not sure if this would work nicely. But it might be possible.
23:39:59Davide-NYCreplace the mini usb with a mini usn a/b(five pins
23:39:59linuxstbCouldn't you use a USB serial adapter on the H300?
23:40:28*Davide-NYC doesn;t have an H300 anymore so couldn't say
23:41:19 Quit saratoga ("CGI:IRC (Ping timeout)")
23:42:04stripwaxlinuxstb - would rockbox support that?
23:42:18Davide-NYCthere are five connectors on the mini usb plug currently. AFAIK only four are connected and used. does that extra pin help us or are we a pin short?
23:42:24XavierGr strange I tried to build the rtc build and it stopped with an error on libwma
23:42:43linuxstbstripwax: When Rockbox has USB support, yes...
23:42:45amiconnSoul-Slayer: A plain freeze would mean the H10 just stops responding and playing, but display content stays the same
23:43:11pixelmaXavierGr: when did you last reconfigure - had that too today with a sim
23:43:19amiconnAnd the latter isn't the case.
23:43:26XavierGrI just reconfigured
23:43:33XavierGrnow I am trying with make very clean
23:43:52amiconnReconfigure + make clean should work
23:43:52XavierGryup that did the trick
23:44:26XavierGrtime to see if the RTC mod works on my player
23:44:47*amiconn has a suspicion when these freezes happen
23:44:49XavierGrI am pretty scared about it
23:45:32 Quit Xerion (" ")
23:45:45 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
23:47:03 Quit Jens ()
23:47:06stripwaxXavierGr - if your iriver works in OF you have nothing to be scared of :)
23:47:31 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
23:47:42 Join rotator [0] (n=e@rockbox/developer/rotator)
23:47:44Davide-NYCno serial on the iaudio X5?
23:48:23 Quit midkay ("Leaving")
23:49:53pixelmabluebrother: not all Archos have a remote - the Ondios don't, not sure about the Player
23:50:43amiconnOnly the Player and the Recorder v1 can use the archos remote
23:51:42 Quit jhMikeS ("Meow!")
23:51:47amiconnThe FM recorder and recorder v2 can't (no 4th contact in the headphone socket), and the Ondios couldn't even be modded to use it
23:52:16amiconnThe 2nd serial of the SH1 in the Ondio is occupied by the MMC interface
23:53:22 Join entheh [0] (
23:55:15 Quit linuxstb ("CGI:IRC")
23:55:30 Quit pixelma (" brb")
23:57:19 Quit jgarvey ("Leaving")

Previous day | Next day