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

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

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

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

#rockbox log for 2005-12-14

00:00:01amiconnpreglow: You won't get me to buy an ipod ;)
00:00:22linuxstb_Buy a H10 then...
00:00:30linuxstb_Even less docs..
00:00:36amiconnYes, exactly
00:00:42mirakdoes a function like memcopy is implemented in assembler in gcc according to the target cpu, or does it uses C code ?
00:00:44*amiconn neeeeds datasheets
00:01:31miraktucoz: in Blade Trinity you can see a blind using braille stuffs
00:01:36mirakthat's a reference
00:01:58tucozmirak, ok. I have seen them though. Cool device.
00:02:02linuxstb_^BeN^ Which player are you using?
00:02:24mirakso gcc doesn't support exactly the m5249
00:02:33markunamiconn: Gigabeat perhaps? :)
00:02:52amiconnmarkun: I don't buy overkill hardware
00:02:53markunamiconn: I know.. not enough buttons located on the front.
00:03:00 Quit amiconn (" HydraIRC -> <-")
00:03:02markunAh yes, too much CPU
00:03:25peturCan somebody tell me where the audio mux pin is set for the H1xx?
00:03:41tucozmarkun, the gigabeat is so powerful it almost doesn't qualify as a embedded device
00:04:14markunWe can underclock the CPU of course
00:04:19Midgey34anyone know what is wrong with this line? if(rb->read(fd, score, sizeof(score)) <= 0) break;
00:04:19Midgey34warning: passing arg 2 of pointer to function makes pointer from integer without a cast
00:04:42linuxstb_You need to use &score
00:04:49markunOr maybe have 1 extra CPU boost level :)
00:05:01LinusNMidgey34: what are you doing?
00:05:04 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:05:05tucozheeh, like the older 486's. Wasn't the 486sx a 486dx(?) with the fpu disabled.
00:05:15tucozand underclocked
00:05:39Midgey34LinusN: I'm working on rockblox using the vertical patch as a base
00:05:44LinusNpetur: the audio mux pin is described in the wiki
00:05:50markuntucoz: hey, the gigabeat also doesn't have FPU
00:05:58LinusNMidgey34: there is a highscore api in the plugin library
00:06:10tucozmarkun, grrr. Then it's cool. ;)
00:06:17Midgey34ah, I see, I'll look it up
00:06:24LinusNpetur: ah, you mean where in the code?
00:06:26webguest02markun: Any idea what to look for in this dir that might cause scrolling to get slow?
00:06:34markuntucoz: nog even multi processor :)
00:06:38^BeN^Night All
00:06:59tucozhehe, that is quite cool. multi-processor mp3-players
00:07:12tucozor did you mean not even...?
00:07:21markunwebguest02: can you dump the filenames to a textfile? Then I can recreate the dir and maybe find out what's wrong
00:07:34webguest02I have the dir as 0byte files
00:07:40Midgey34LinusN: is not using the highscore lib frowned upon? I'm using the code from bejeweled with slight modifications
00:07:45markuntucoz: only 1 cpu (unlike the ipods)
00:07:48webguest02where I have already removed some files, and it still does it
00:08:01webguest02I'll zip it up
00:08:04webguest02hang on
00:08:14tucozoh. I am looking forward to see your progress with gigabeat.
00:08:15LinusNpetur: pcm_record.c: pcm_record_mux()
00:08:28markuntucoz: it will have to wait till next year
00:08:49peturyuck. what is it doing there?
00:08:51LinusNMidgey34: not frowned upon, but we would appreciate if the plugins could use the same code
00:09:04linuxstb_markun: That's only 18 days away...
00:09:08LinusNMidgey34: however, no plugin uses it, afaik
00:09:12tucozmarkun, that is soon
00:09:26peturand why is there fmradio.c and fmradio_i2c.c :(
00:09:38LinusNMidgey34: so you are free to redesign it to your liking and make both bejeweled and rockblox use it
00:09:53peturI would have expected stuff like enabling radio (switching mux pin) in fmradio.c
00:10:02LinusNpetur: because different platforms use different fm chips
00:10:12peturthat I can understand
00:10:14LinusNpetur: historical reasons, mostly
00:10:21 Join amiconn [0] (
00:10:34peturmay as well clean up a bit then...
00:10:34LinusNpetur: the mux isn't necessarily a radio function
00:11:10LinusNthe mux controls what audio source is routed to the ADC
00:11:19tucozmarkun, still, do you like the gigabeat as it is though? It looks very neat.
00:11:27peturah - only for recording then
00:11:28LinusNin this case it's radio or line in
00:11:46peturso for radio playback, I don't have to switch anything
00:11:54LinusNfor recording and also radio playback, since we rerout the ADC right to the DAC
00:12:16LinusNyes, you need to set the ADC source to radio
00:12:34markuntucoz: yes, I like it. I made some screenshots of a movie and viewed them on the Gigabeat, looked very nice.
00:12:54 Quit Kohlrabi ("Leaving")
00:13:04peturwell if it's ok for you I'll take some more time and try to clean it up a bit
00:13:14LinusNwhat needs cleaning up?
00:13:27tucozmarkun, the resolution of the screen is very nice. Especially that it has a standard aspect-ratio.
00:13:35peturfmradio.c and fmradio_i2c.c ?
00:13:37tucoz...when tilted that is
00:14:13markuntucoz: The X30 has a nicer screen, but the internals are almost the same, so I think it will not difficult to support it as well.
00:14:39LinusNpetur: fmradio.c is for the archos fm recorder, which talks SPI with the chip
00:14:42tucozmarkun, doesn't yours have 240x320 resolution?
00:14:59LinusNpetur: and fmradio_i2c.c is for the archos ondio and the iriver
00:15:05LinusNwhich talk i2c
00:15:09 Join ehntoo [0] (
00:15:10markunyes, but it's 2.2" and not 2.4
00:15:17tucozI see.
00:15:34peturI saw that, but it's not very clear...
00:15:48tucozhave you recieved the source code yet?
00:15:57markunno, still not
00:16:03LinusNpetur: no it isn't 100% clear
00:16:12LinusNpetur: what do you suggest?
00:16:12webguest02I believe that should work
00:16:29peturit could use something like the lcd-xxx file naming but that would be overkill
00:16:35peturneed to think about it
00:17:47markunpetur: file renaming is not nice with cvs. Wait till we switch over to svn
00:18:30webguest02markun: And please no comments on the music taste!
00:18:38 Quit Moos ("Glory to Rockbox")
00:19:05peturI'll first try to get radio working, then make a common I2C bitbanger.
00:20:03markunwebguest02: yes, it lags here too
00:20:27webguest02Ah, well that's at least something
00:20:53webguest02I'm slowly trying to remove files with suspicious names
00:23:58LinusNpetur: good
00:24:04webguest02down to 94 files now. Still lagging.
00:25:09peturport pins adjusted, probably need to enable radio functionality for H300 somewhere and then read the RB manual how to operate it :D
00:26:32webguest0224 files. Still lagging.
00:27:29Midgey34while I'm here, are there any changes or suggestion people have for the Sokoban patch I wrote the other day?
00:30:43peturuh-oh - adding radio will screw up my settings
00:30:47 Quit ReKleSS (
00:30:59peturenabling the define adds them somewhere in the middle :(
00:31:50peturhehe, seems I need to define some buttons first too
00:33:15linuxstb_petur: You can use the same button definitions as the H100 - just do || CONFIG_KEYPAD == IRIVER_H300_PAD
00:33:37peturalready busy ;)
00:33:59 Nick webguest02 is now known as JackP (
00:41:27peturI get the feeling that code dealing with buttons will need some reworking - many #ifdefs in there
00:41:39peturmany targets...
00:45:04 Part tucoz ("Leaving")
00:46:38NJoinReKleSS [0] (
00:49:02peturdamn... won't boot (shows only RB logo)
00:49:11peturCould it be the settings change?
00:49:25LinusNpetur: maybe
00:49:32peturwill bump my version number
00:49:33LinusNhold rec when you boot next time
00:49:46peturthat should stop it from trying to load
00:49:49 Join nathanh [0] (
00:50:00peturI can't get that trick to work
00:50:00LinusNpetur: that resets the settings, yes
00:50:18peturI tried it yesterday for my brightness thing
00:50:28peturwhen exactly should I start holding REC?
00:52:07JackPBefore the logo comes on, I think
00:52:50peturthat's what I tried: as soon as the bootloader start printing...
00:53:34peturright, any quick hints where the settings version number is defined...
00:54:27nathanhpetur: h300? hold rec while its off, then hold play while still holding rec until iriver has booted
00:54:28peturis it CONFIG_BLOCK_VERSION?
00:55:01peturnathanh: don't want to boor iRiver :D want to reset my settings ...
00:55:48nathanhyeah, ive lost all interest in booting iriver too :-)
00:56:06peturexcept for recording and usbhost!
00:56:18peturAnd I need those damn things
00:57:47peturdamn cygwin is slow :(
01:01:27 Quit ender` (" Oh, I'm also going to hunt you down in real life, break all your fingers and both your wrists, smash your computer, and cut )
01:01:36peturchanged CONFIG_BLOCK_VERSION, still no boot. Let's try without radio...
01:01:48mirakthere is only 96kB of fast ram ?
01:04:36mirakLinusN: how much is it faster to move some data to fast ram, then do the computing on it, move it back to slow ram, instead of just computing in the slow ram ?
01:04:54mirakI mean in your experience for what kind of computing it's worse
01:07:49linuxstbmirak: It will obviously depend on how much that memory is accessed. But the slow ram is very slow. Also, codecs and plugins only have 48kb of iram - the rest is used by other parts of Rockbox.
01:08:22mirakbecause since a frame is bigger alone than the iram
01:08:45mirakonly operations on blocks would have a meaning
01:08:51mirakto be done in fast ram
01:08:58mirak(still on xvid)
01:09:09mirakI have not abandonned yet
01:09:13 Part Sando
01:10:07miraklinuxstb: so for a full image, you need to pass each blocks for each type of frame (Y,U,V)
01:10:16linuxstbThis is the kind of optimisation you can experiment with when you get it working. Don't forget that if you want audio with your video, then the audio codecs will also be fighting for iram - they will struggle without it.
01:10:17mirakfrom slow ram to iram
01:10:37linuxstbIt's not going to be easy...
01:10:59mirakiriver took almost 2 years to achieve that, if not more
01:11:18mirakthe video codec appeared on later firmwares
01:11:22peturgreat... if I build with radio it doesn't boot. If I remove radio, everythings OK :(
01:11:43mirakthey probably also stripped to death the code but you need to know what you are doing to do that
01:11:49LinusNpetur: so there's obviously something wrong with the radio
01:11:55peturstep one: enable radio, do not touch mux
01:11:58LinusNpetur: enable the logf function
01:12:07peturah, where's that?
01:12:20miraklinuxstb: when a codec alone have a binary bigger than iram, how do you proceed ?
01:12:31LinusNwhen you run configure, select "developer" and then logf
01:12:48LinusNyou will then see logf() messages on the remote lcd
01:12:50linuxstbmirak: You specify which functions and which data structures are in iram. The rest are in sdram.
01:13:10peturremote lcd? don't have a remote....
01:13:11linuxstbOnly one or two audio codecs are small enough to fit entirely in iram.
01:14:03miraklinuxstb: so in general is it faster to move the code to iram dynamically ?
01:14:20mirakI guess rockbox doesn't do that
01:14:28nathanhwoo, new cvs doesnt flicker, what was the fix?
01:14:31mirakthat's a bit hard to do probablt
01:14:34***Saving seen data "./dancer.seen"
01:14:47linuxstbWe've never tried moving code to iram. You generally get much better improvements using the iram for data - the Coldfire has a code cache, but no data cache (IIUC)
01:15:04petur_FireFly_ made a ticking patch that solved the flickering :D
01:15:27mirakLinusN: ok thanks, I understand better now
01:15:29linuxstb(I mean we've never tried dynamically moving code to iram)
01:15:44petur(only update the LCD in one place)
01:15:48nathanhoh right
01:16:17miraklinuxstb: so in average it's less a waste to move the data. I don't realise how fast it is to move data from slow ram to iram
01:17:04peturLinusN: logf only goes to the LCD of the remote? Just checking... because then it's useless for me.
01:17:23miraklinuxstb: you will need to write your own compilator to do that
01:17:27LinusNpetur: it is also stored in ram, so you can view it and save it to disk in the debug menu
01:17:29mirakno ?
01:17:45linuxstbYou mean to move code in and out of iram?
01:17:47peturhelps a lot if it doesn't boot :(
01:17:55miraklinuxstb: yes
01:18:08linuxstbNo, I'm sure there are tricks to make that possible.
01:19:15miraklinuxstb: by moving the code of the size of the struct of the function into iram then using
01:19:33mirakthe variable that point to that iram place
01:20:19mirakseems doable like that as far as I have understand what it's possible to do with C
01:21:01amiconnI doubt that moving code will give any speed improvement, I'd expect the opposite
01:21:28mirakamiconn: for code that acces to slow ram or iram data ?
01:21:40amiconn_any_ code
01:22:03mirakwhat lead you to that conclusion ? (just want to learn more)
01:22:06amiconn(1) While moving, the code is treated as data, so there is no caching
01:22:20amiconn(2) After moving, you have to invalidate the icache
01:22:32mirakwhy 2 ?
01:22:52mirakinvalidate what ?
01:23:09amiconn(2) The coldfire has icache, so every code snippet that's run more than once would be cached with high probability
01:23:17amiconnicache == instruction cache
01:23:42amiconnThe last (2) should have been (3)
01:24:16mirakhow does the cpu knows it doing the instruction set more than once ?
01:24:40mirakhow does the cpu knows he is running a particular instruction block more than once ?
01:25:04mirakwell nevermind that's to complicated
01:25:18amiconnIt doesn't know beforehand
01:25:18LinusNmirak: google for "cpu cache"
01:25:50amiconnIt just caches the code it executes, and if it's executed a second time while still in the cache, executes it from there
01:27:07nathanhmirak: you'll need to understand the cache to get video playback
01:27:24peturLinusN: still won't boot without touching the mux pin, so it must be the I2C
01:28:05LinusNi had the same strange hang when i tried to enable recording
01:28:26mirakamiconn: I understand the concept, but was wondering how he was knowing he was about to run the same code again to run it inside the cache and not the ram
01:28:31peturwill continue tomorrow. You need the changes I made?
01:28:41peturguess not
01:28:42 Join Paul_The_Nerd [0] (
01:28:49LinusNpetur: would be nice to see them
01:29:03mirakamiconn: but well that's not necessary to understand how to use it
01:29:08peturwill dump it on the mailing list
01:29:19LinusNpetur: or to me
01:29:27petureven better
01:29:36LinusNmailing list is probably better
01:29:40LinusNnever mnind
01:29:52LinusNlinus at haxx dot se
01:29:52peturbeware - they probably have bad line endings ;)
01:30:07petursorry, W2K user
01:30:13amiconnpetur: Get a decent editor ;)
01:30:29amiconnConTEXT (my favourite on windows)
01:30:37mirakamiconn: in fact I was wondering that in case the code in ram was modified, but I just read that's a nono ro do that now
01:30:44peturusing M$ devstudio atm
01:31:18amiconnmirak: That's what I meant with (2). There is no bus snooping, so the CPU doesn't know the RAM contents has changed
01:31:38mirakamiconn: ah ok
01:32:02nathanhemacs is pretty good, now if only it had a decent text editor
01:32:31miraknathanh: eclipse is not bad with DCT.
01:32:53mirakif you disable indexing
01:35:53 Quit nathanh ("CGI:IRC (EOF)")
01:36:24 Join nathanh [0] (
01:39:13peturLinusN: mail is on its way
01:39:34peturDoesn't the bitbanging suffer from CPU freq changes?
01:40:07peturI mean that delay loop is freq dependent, no?
01:42:27 Part Midgey34
01:43:25 Quit novimon_ (Read error: 104 (Connection reset by peer))
01:44:23 Join mind_less [0] (n=Tim@
01:45:26LinusNpetur: yes
01:46:25peturright now I'm not sure anymore what code gets used: fmradio or fmradio_i2c
01:46:38peturbut that will be for tomorrow
01:47:31 Quit petur ("here today, gone tomorrow")
01:47:55 Join novimon [0] (
01:47:59 Quit Rick (Read error: 104 (Connection reset by peer))
01:49:04 Join Rick [0] (
01:57:11 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
01:59:23nathanhjust a few 100 bytes)?
01:59:35 Quit dpassen1 ()
01:59:44nathanhblah, stupid irc client... i can appreciate that rtc is limited to 44 bytes, but why is hd_bits limited to just a few 100 bytes?
02:05:30 Quit XavierGr (Read error: 104 (Connection reset by peer))
02:07:53linuxstbnathanh: Because it is stored in a sector on the disk. But there is talk of changing that.
02:08:01LinusNthe config sector is 512 bytes
02:08:23nathanhoh right, so its currently not a file, its an actual sector
02:08:28nathanhthe intention is to make it a file?
02:08:35LinusNbut that sector doesn't only contain settings, it also contains the resume info
02:14:39nathanhright, so theres config bits in rtc ram and in a single disk sector
02:15:04nathanhthere are also per-app settings stored in files, each app using its own read/write and parsing routines
02:16:58LinusNnathanh: "app"?
02:17:04nathanh.rock files
02:20:17 Quit Febs ("CGI:IRC")
02:22:55 Quit Rob2222 ()
02:26:28JackPDNS is so annoying. Forums still not working
02:26:30LinusNwee, h300 lcd dma works! :-)
02:27:53amiconnHmm. Is using DMA faster/ as fast as a line-read-optimised CPU transfer would be?
02:28:33amiconnOh, btw, what was the problem?
02:28:50LinusNthe usual stuff, unclear data sheets :-)
02:29:23LinusNthe DONE bit had to be manually reset between transfers
02:32:12LinusNthe sad part is that a line-optimized cpu transfer would be too fast for the lcd
02:32:49LinusNthe dma can do line optimized transfers, but it is too fast at 124mhz
02:33:16PaulJJackP: add the line "" to your HOSTS file
02:33:28LinusNand it's the time between transfers that is the problem, so increasing the waitstates doesn't help
02:37:46JackPPaulJ: Yeah, I know, just annoys me
02:38:45LinusNamiconn: a full lcd update takes 6.7ms in 45mhz
02:39:01Weazel_forum works here
02:39:54amiconnLinusN: I mean line-optimised reading from DRAM, then writing to the LCD according to specs (but in the fast mode with multiples of 4 words)
02:41:54 Quit nathanh ("CGI:IRC (EOF)")
02:43:13 Join DJDD_ [0] (
02:43:59LinusNamiconn: it turns out that we can't use the fast ram mode
02:44:18LinusNbecause the lcd expects CS to be low during the entire transaction
02:44:29LinusN4-word burst
02:45:10amiconnOh, and we can't do that?
02:45:25amiconn(hold CS low I mean)
02:46:09*amiconn notices there are no H3x0 schematics on the website :(
02:46:58 Quit mikearthur (Read error: 104 (Connection reset by peer))
02:47:19LinusNhehe, no, i haven't made any schematics
02:49:07LinusNamiconn: hmmm, i might be wrong about the cs
02:52:25 Join dropandho [0] (
02:52:36dropandhohey all!
02:52:43amiconnBah, we won't benefit from fast mode at 45 MHz
02:53:01dropandhoLinus- im in forum heaven!
02:53:29 Join Sando [0] (
02:53:31amiconnBus cycle time is 88.5 ns, so we need one wait anyway, and then we're at 177ns, for which slow mode is sufficient
02:54:03 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
02:54:24amiconn>wrong<, try again...
02:54:50amiconnBus cyle time is 44.28 ns ...
02:57:25dropandholinus- im back to bug u!
03:07:52 Join YouCeyE [0] (n=YouCeyE@unaffiliated/youceye)
03:14:38***Saving seen data "./dancer.seen"
03:18:32 Join Paul_The_Nerd [0] (
03:24:40LinusNamiconn: lcd_update() takes roughly 26ms in 45mhz and 10ms in 124mhz
03:26:42lostlogicwow, not much that can be done to optimize that either :(
03:27:51dropandholinus my friend!
03:28:09 Join aliask [0] (
03:28:26dropandhowonderin if you ended up finding the R20 and R22 components on the ticking unit
03:29:34LinusNdropandho: i have measured the ticking board, and the isolation between analog and digital ground is not 100% on the ticking board
03:29:49LinusNi have yet to find out how and why
03:29:52dropandhogot it....i have an odd thought on that one
03:30:05dropandhogot it- thanks for investigating that deeper
03:30:27dropandhoi remember the_winch said that the board was missing R20 and R22....any idea?
03:30:35dropandho(that is theboard u have)
03:30:38LinusNnot really
03:30:59dropandhois he really missing those?
03:32:30LinusNwell, r20 and r22 is missing on my other board as well
03:32:38aliaskIs the bug about rockbox crashing when the "show files" setting in the quickscreen is set to ID3 database known?
03:32:49LinusNaliask: yes
03:33:01aliaskGoodo then.
03:33:07dropandhoso i wonder if we can find a corolation between these missing spots and the tick
03:33:52LinusNdropandho: i doubt it
03:34:24dropandhomy non-technical thought was that if it is the interference b/w the analog and the digital ports....maybe there should be a setting to disable the digital when you dont need you can stop the interference...and stop the ticking
03:35:45LinusNthe digital i/o is used to communicate with the remote
03:36:03LinusNit's not the s/pdif, if that's what you're thinking
03:36:14dropandhoyes- that was what i was thinkin
03:37:07dropandhowhat a total bummer
03:38:11LinusNi'm suspecting that it's a problem with the pcb itself
03:38:29dropandhoyeah...nothing that any software can mend
03:39:14dropandhowhat did the latest tick-fix add?
03:39:35LinusNfewer lcd updates
03:41:02dropandhocoo- the usual
03:41:24dropandhoglad people are still working on this!
03:42:20LinusNgotta get some sleep
03:42:25LinusNcu tomorrow
03:42:41 Part mind_less
03:42:43 Part LinusN
03:43:38 Join amiconn_ [0] (
03:49:08 Quit JackP ("CGI:IRC")
03:49:42dropandhois it me...or did the forum stop responding?
03:51:34Paul_The_NerdIt looks like it's working to me...
03:58:11 Join tvelocity [0] (n=tony@
04:00:30 Quit amiconn (Read error: 110 (Connection timed out))
04:00:31 Nick amiconn_ is now known as amiconn (
04:00:55dropandhoweird- broken over here
04:05:12Paul_The_NerdIt just loves me more than you.
04:52:09 Join scott666 [0] (
04:52:52 Quit scott666 (Client Quit)
04:54:31 Join scott666 [0] (
04:58:48scott666this channel is a lot bigger than i remember it.
04:58:53scott666and just as quiet
04:59:48dropandhothis is the study room sir
04:59:51dropandhoplease keep it down
05:00:16dropandhothe time difference thing is what keeps em quiet i believe
05:00:37scott666well, it was quiet most of the time when i was here regularly too
05:01:13dropandhothey do get chatty at times
05:01:13scott666my arches died a while back
05:01:33scott666so i kinda fell out of the rockbox community
05:01:40dropandhou get a new dap?
05:01:46 Quit tvelocity (Remote closed the connection)
05:01:48scott666looking into that now
05:02:20scott666whats this about rockbox on ipods??
05:02:29 Join tvelocity [0] (n=tony@
05:02:32dropandhoyeah, they are working on that
05:02:36dropandhonot sure where it stands tho
05:02:44dropandhoi know that it seems to be comming along tho
05:03:11scott666hows the iriver port coming along?
05:03:25dropandhono official release for the 120/140
05:03:36dropandhobut it has been solid for many months now
05:03:42dropandhojust cleaning up the loose ends
05:04:00dropandhothe 320/340 just started up maybe a month or 2 ago
05:04:00scott666and the 320? not so much?
05:04:13dropandhoand is rapidly moving along as they can use much of the code base
05:04:20dropandhoi dont know where it stands
05:04:22 Join nathanh [0] (
05:04:27dropandhobut i think it is fairly functional
05:04:32dropandhoi would check out the website
05:04:39dropandhothey have been keeping that up-to-date
05:05:13scott666i check in about once a month
05:05:17scott666i see 1.5 is out
05:05:18scott666thats new
05:05:33dropandho2.5- ya
05:05:52scott666my archos died before 2.4 went official
05:06:10scott666so yeah, havnt really been keeping up
05:06:40dropandhoit's a whole new world!
05:06:47scott666the only thing i use it for now is a portable HD
05:06:57scott666i should load up 2.5
05:07:10dropandhothe audio jacks die?
05:09:07scott666it cna still record too
05:09:14scott666but i havnt had need for that lately
05:12:24dropandhodamn- it looks like my forums link is pointing to ...that's the old one- right?
05:14:40***Saving seen data "./dancer.seen"
05:17:29PaulJthe new ip is
05:20:12scott666question: whats the difference between rockbox.ucl and rombox.ucl? i wonder why i am now back to the old one
05:20:33dropandhoit propgated
05:20:51dropandhodo i need to wait again?
05:21:15scott666answer: rombox doesn't work.
05:31:34PaulJdropandho: you could add the line "" to your hosts file
05:33:14 Quit dropandho ()
05:37:10 Part PaulJ
06:01:33 Quit DreamTactix291 (Read error: 110 (Connection timed out))
06:01:41 Join xbshift [0] (
06:03:39 Quit Strath (Read error: 104 (Connection reset by peer))
06:05:08 Join Strath [0] (
06:08:13 Join calbearfan [0] (
06:10:21calbearfanHi, anyone there to answer a question for a new Rockbox user on an iriver H140?
06:10:31nathanhim around, not sure how much help i can be though
06:11:02scott666im also around, but will almost certainly be useless
06:11:16calbearfanI'm having a problem with several podcasts −− it seems like the gap detection is detecting silence in the middle of a recording and it won't let me play the rest of the podcast. Is this a known issue?
06:11:37calbearfanWhen I play using the normal iriver firmware, it plays fine
06:11:48nathanhdont know about that one
06:11:52nathanhi wasnt aware there was gap detection either
06:12:11 Join Membrillo [0] (
06:12:25 Join lostlogic_ [0] (
06:12:30calbearfanfor example, all the NPR podcasts have this little tune at the beginning of the podcast, then silence. All the podcasts look like they are 4 seconds long using Rockbox, but several minutes long using the iriver firmware.
06:13:00calbearfanI just get the tune, then it moves on to the next song.
06:13:08 Join lostlogi1x [0] (
06:13:08***Alert Mode level 1
06:13:08DBUGEnqueued KICK lostlogicx
06:13:08DBUGEnqueued KICK lostlogic
06:13:08***Alert Mode level 2
06:13:08DBUGEnqueued KICK lostlogic_
06:13:08DBUGEnqueued KICK lostlogi1x
06:13:08***Alert Mode level 3
06:13:35nathanhcant help, you'll need to ask one of the devs
06:13:43nathanhthey'll probably get all excited over a bug like that :-)
06:13:53calbearfanit's very strange, and since it always seems to happen during silence, it seems like it would be some kind of silence detection
06:14:18calbearfansince iriver isn't officially released, I should use the mailing list?
06:14:20 Join webguest52 [0] (
06:14:53webguest52Hi, was reading the logs. As far as I know, there's no silence detection.
06:15:09webguest52My guess is that they're doing something funky with the mp3s
06:15:38calbearfanit's consistent for me. actually, now that I think about it, I think I've only tested with a single podcast in a directory. maybe I'll test a bit more
06:15:44scott666yes, use the mailing list
06:15:54webguest52or wait around here
06:17:09webguest52Do you have an url for these mp3s?
06:17:27calbearfanthat might make sense (something funky with the mp3s). since it is a standard lead-in, they are probably fusing it together with the content, maybe in a nonstandard way.
06:17:42webguest52That's what I'm guessing
06:17:57calbearfanthen maybe it wouldn't be silence, just a screwed up portion of the mp3
06:18:02 Quit NibbIer (Read error: 110 (Connection timed out))
06:18:29calbearfanI also can't seek back into the mp3 from the next podcast
06:19:51calbearfanan NPR feed I have problems with is
06:20:07calbearfan(user-friendly feed name, courtesy of iTunes I imagine)
06:20:31nathanhi wonder if it is two mp3s
06:20:46nathanhthe little mp3 at the start is all rockbox sees, thinks its 4 seconds long
06:22:06calbearfanyep, that's probably it. the one I'm currently testing that has problems is from 11/29 (npr_20051129_atc_04)
06:22:10webguest52calbearfan: Filesize of 2.2mb sound right?
06:22:31webguest52Alright, think I'm getting the latest one
06:23:09***Alert Mode OFF
06:23:47calbearfanI'm going to download that one too and check it out
06:25:35 Join Nibbler [0] (
06:26:12calbearfanyep, same problem with the latest (npr_20051212_me_03.mp3)
06:26:29 Quit lostlogic (Nick collision from services.)
06:26:39 Nick lostlogic_ is now known as lostlogic (
06:26:39DBUGEnqueued KICK lostlogic
06:27:55 Quit lostlogicx (Read error: 110 (Connection timed out))
06:28:06calbearfanhappened to you too? are you doing it on an iriver, or a different platform?
06:29:28calbearfanhey, totally unrelated, what is everyone's favorite theme? any good ones that don't come with the latest builds?
06:30:11webguest52I haven't decided on one yet
06:30:23lostlogicthe wps I made :)
06:30:50calbearfanlostlogic: what's cool about yours?
06:31:06calbearfanother than it's yours? :)
06:31:35lostlogicdisplays complete information (including brief next track info), clean, easy to read, no bitmaps
06:31:43lostlogic <−− last WPS on that thread
06:31:58calbearfancool, i'll give it a try.
06:32:42 Quit webguest52 ("CGI:IRC (EOF)")
06:32:42 Quit nathanh ("CGI:IRC (EOF)")
06:32:57calbearfananybody have a non-iriver that wants to try to see if an mp3 that is failing on an iriver fails for a different platform?
06:33:18calbearfanI imagine it should fail on all though
06:34:07scott666do i have to be able to hear it to know if its failing?
06:34:52 Quit RotAtoR ()
06:35:03calbearfanno. an mp3 that is several minutes long is showing up in rockbox as 4 seconds long. probably because of the way NPR is tacking a standard sound to the beginning of the podcast.
06:35:26calbearfanin the iriver firmware, it plays fine
06:35:47scott666i have an archos FMR with a broken headphone jack
06:38:09calbearfanusing a podcast reader, subscribe to then download the only file in the feed
06:39:08calbearfanlooks like the file is
07:08:53 Join nathanh [0] (
07:14:42***Saving seen data "./dancer.seen"
07:18:38 Quit Membrillo ()
07:19:59 Quit Paul_The_Nerd (Excess Flood)
07:33:10 Join perplexity [0] (n=joust@
07:34:17 Join mind_less [0] (n=Tim@
07:36:22Bgermorning :)
07:36:50scott666just barely
07:39:36dwihnoWee! Morning!
07:40:16dwihnoonly 1h 40 minutes 'til sunrise!
07:41:06scott666this channel got a hell of a lot more idlers since iwas here last
07:41:19thegeek_only 1h 20m until my exam starts
07:41:23thegeek_I'm fucked
07:41:50 Quit nathanh ("CGI:IRC")
07:41:58 Join DreamTactix291 [0] (
07:47:55Bgerthegeek_ wish u luck
07:47:56amiconnlinuxstb: Red builds...
07:49:56thegeek_hehe, thanks Bger
07:50:00thegeek_I need it;)
07:52:41scott666is there a 'whats new in 2.5' page somewhere?
07:53:02Bgerthere was one anecdote about the fact that the bad student learns only 3 topics of 30 and on the exam the theme is one of the 3, and the good student doesn't learn only 3 topics and on the exam he has one of these 3
07:55:10thegeek_in this exam that is certainly true
07:55:14thegeek_discrete math
07:55:26thegeek_so many different and quite separate subjects
07:56:31Bgerscott666 there was something like "changes done since v 2.4" ...
07:57:00BgerBagder ?
07:57:13Bgerscott666 archos user ?
07:57:35 Quit mind_less (Read error: 110 (Connection timed out))
08:06:01scott666i was
08:06:15BgerThis guy Tomasz Malesinski is insane ... to make a emulator alone ...
08:06:18scott666and i was still around when 2.4 came out
08:07:01Bgerscott666 if u have cvs u can get all changes yourself
08:07:50scott666no, i meant more like a summary
08:08:10Bgeru are asking for archos, aren't u
08:08:24scott666or in general
08:08:32scott666my archos is deadish
08:08:36Bgerif so, one of the biggest fixes is the RLD bug...
08:08:42 Join webguest71 [0] (
08:08:43scott666holy crap!
08:08:51scott666thats awesome
08:08:53Bgergeneral... iriver doesn't have 1 release
08:09:07Bgerapplauds goes to amiconn :P
08:13:01 Join einhirn [0] (
08:14:31 Quit webguest71 ("CGI:IRC (EOF)")
08:26:44scott666another quick question: is there a page comparing the specs of the h1x0 and h3x0 somewhere?
08:28:10 Join mind_less [0] (n=Tim@
08:28:40 Join ender` [0] (i=ychat@
08:38:40scott666so the main differences are the button layout and the LCD?
08:39:35Lost-ashand usb2go
08:39:43Lost-ashand the H300 has a built in RTC
08:39:48Lost-ashand can charge from usb iirc
08:39:52 Nick Lost-ash is now known as ashridah (n=ashridah@
08:40:25ashridahyeah. basically the device can be a usb host, so it can read files off some digital cameras.
08:40:35scott666i see
08:40:42ashridah(and pretty much any other usb device, subject to driver support)
08:41:07ashridahrockbox doesn't have support for it yet, but i imagine someone will be insane enough to try eventually :)
08:41:21CtcpIgnored 5 channel CTCP requests in 13 minutes and 31 seconds at the last flood
08:41:21*Bger has insane plans ...
08:41:45scott666is the HD upgradable for either?
08:42:02ashridahnot really.
08:42:16Bgerthe 40GB models are upgradable to 60GB atm, but 80GB are coming
08:42:27Bgerand 20GB models are upgradable to 40GB
08:42:28ashridahnot to the degree that one can do it for the archos's, you need special screw drivers
08:42:32scott666oh right, they use 1.8" disks
08:42:32ashridahbut it can and has been done
08:42:49scott666how special?
08:43:09Bgerscott666 the form factor is more special
08:43:15ashridahjust a set of star screwdrivers.
08:43:21Bgeri mean, this is not the standard laptop 2.5" drive
08:43:26scott666yeah has info on the procedure iirc
08:43:40scott666ok, ill look around there
08:43:54Bgerbut the drives are expensive
08:44:37scott666i should be able to live with 40gb
08:54:41scott666so any idea where i can actually buy one?
08:54:53Bgerh300 ?
08:54:55Bgeror h100 ?
08:55:05scott666h140 im looking for right now
08:55:16scott666only remotes ?
08:58:36scott666im american
08:58:52*Bagder listens to the recorded rockbox review at
08:59:14scott666they have the h340 for $475 + international shipping−−not really ideal
09:02:46Bgerhm, i can't help you then...
09:02:51Bgerask @ MR
09:03:00scott666will do
09:03:07preglowgod, how a week of uninterrupted raining puts me in a foul mood
09:07:24scott666why does a several year old player cost more than a brand new video ipod? :-/
09:07:47Bagderthe rockbox effect! ;-)
09:08:32scott666that effect being that rockbox only runs on players that have been discontinued?
09:09:13Bagderif the ipod dudes are fast enough we might have a short time of actually working on a player that is available
09:09:52preglowsporadic paste
09:09:57preglowi need to reassign that mousebutton
09:10:37preglowbtw, it seems we might be able to get a port up and running in not too long a time
09:10:54preglowit seems we'll be able to run adequately on only one cpu
09:11:03Bagderwow, cool
09:11:09preglowdid a naive test
09:11:22preglowwav2wv runs at 150% realtime on one 75mhz core
09:11:29preglow195% on 120mhz coldfire
09:11:29scott666so is the work being done on the video ipods? the daily build page says nano and 'color'
09:11:45Bagderscott666: no
09:12:07 Part mind_less
09:12:20 Join nathanh [0] (
09:12:45saa[b_r]idereven the H1x0 has a pricey tag now-a-days
09:12:56scott666yeah, tell me about it
09:12:57saa[b_r]idermaybe higher than the H300
09:12:58preglowBagder: on ipod colour and ipod nano
09:13:08preglowscott666: on ipod colour and ipod nano
09:13:19scott666whats an ipod color? couple generations back?
09:13:24scott666or ipod photo?
09:13:33preglowanother name for a photo, or something
09:13:40preglowi'm not completely sure
09:13:50pregloweven the ipodlinux guys haven't figured out the video yet
09:13:56preglowit does some things quite differently
09:13:56saa[b_r]iderit has to be the iPod photo
09:14:00scott666yeah, doesnt shock me
09:14:14preglowthey've got a whole new dsp/cpu they need to figure out how works
09:14:44***Saving seen data "./dancer.seen"
09:15:46 Quit perplexity (Read error: 110 (Connection timed out))
09:16:21 Join perplexity [0] (n=joust@
09:20:08 Join Paul_The_Nerd [0] (
09:24:28 Quit Paul_The_Nerd (Client Quit)
09:31:03 Join use_off [0] (
09:34:44preglowdidn't linus get some docs on some ifp cpu on time?
09:36:07Bagderhm, I remember him getting something but I don't reall the details
09:36:41preglowthis malesinski fellow most certainly seems like somone we want on the team :>
09:36:59Bagderoh yes
09:37:49preglowlinuxstb: btw, i've figured out why disk accesses sometimes work and sometimes not. it turns out it never works after the spin down time has been exceeded
09:46:45 Join Membrillo [0] (
09:55:21merbananthis one is for the scandinavians,
09:56:58preglowclever norwegian, that
09:58:08markunHey, I can even read that :)
09:58:20markunLooks quite a lot like dutch even
10:00:24preglowthere are similarities
10:00:29preglowbut dutch sounds really strange :P
10:01:18 Join Vladoman [0] (
10:01:56markunpreglow: did you receive the crossfeed sample by jlo?
10:02:30 Quit Vlad0man (Read error: 110 (Connection timed out))
10:03:00preglowahh, yes i did
10:03:07preglowi'll do a blind test today
10:03:18preglowjust as soon as i figure out how this program works
10:03:26preglowit's jave, must be blocking my brainwaves
10:03:47markunhe wrote a crossfeed implementation in java?
10:03:49Bagderthat's what java does ;-)
10:04:36preglowthe blind test program
10:04:43preglowabchr or whatever
10:05:03preglowhe did the crossfeeds in reaktor, as is wise
10:06:12markunisn't is also possible with foobar?
10:06:49preglowwell, you can't exactly design effects in foobar
10:06:54preglowyou need to write them in c first
10:06:58pregloweasier with reaktor
10:07:44 Join webguest02 [0] (
10:08:33 Join LinusN [0] (
10:08:41webguest02how is development going for iPod as I may get an iPod for christmas?
10:08:50preglowit's progressing steadily
10:08:56Bgerwebguest02 nano and colour for now
10:08:58preglowwe might have sound pretty sooon
10:09:23preglowsince the arm cores seem to be surprisingly fast
10:09:32Membrilloi got one of the new videos a couple of weeks ago
10:09:41preglowLinusN: didn't you get some docs for some ifp cpu from someone at some time?
10:09:49webguest02There is one thing I am concerned. iPod doesn't use drag and drop to transfer files to the device, you have to use iTunes. So when Rockbox does work on the iPod you still have to use iTunes?
10:10:04LinusNpreglow: yes, from philips
10:10:14preglowwhat core?
10:10:19Membrillono, it will most likely use a different database system that doesnt use the itunes_DB
10:10:41LinusNbut those docs don't tell much about the cpu itself, mainly about the api to their audio framework
10:10:41webguest02I like the black version of the iPod
10:11:00Membrillodoes anyone know how you update the iPod firmware though? is it easily possible?
10:11:00preglowLinusN: riight, but is still of use?
10:11:14preglowMembrillo: hell yes
10:11:20preglowMembrillo: you mean the official firmware?
10:11:27Bagderwebguest02: so far, you can't even use the itunes approach for rockbox on ipod
10:11:34webguest02Also is it possible to repairi a broken headphone socket if the sound only comes out of one side?
10:11:37preglowBagder: should we ever support that? :-)
10:11:42webguest02yea i know bagder, but I was just wondering
10:11:42BagderI doubt it
10:11:47Membrilloi dont really know much about ipods. Is it easy to flash firmware?
10:11:51webguest02iTunes is a horrible program
10:12:00preglowBagder: well, it is possible. there is this libitunesdb
10:12:08preglowMembrillo: you don't flash
10:12:16preglowMembrillo: it's completely disk based
10:12:28preglowMembrillo: which is good, because it makes it almost impossible to brick an ipod
10:12:29 Part scott666
10:12:29Membrilloreally? the ipod FW is disk based... interesting
10:12:39Bagderlike the good old Archos
10:12:43Membrillowhat if you deleted the fw? that would brick it
10:12:45preglowMembrillo: god knows i've tried, and i haven't succeeded yet
10:12:50preglowBagder: apart from being even a bit more clever
10:12:56preglowBagder: you can brick an archos pretty easily, yeah?
10:12:56BagderMembrillo: not at all, the bootloader is in flash
10:12:58webguest02I like to see how iPodbox handles the clickwheel interface
10:13:04Bagderpreglow: no
10:13:10preglowso exactly like ipod, htne
10:13:17preglowthen <-
10:13:30Bagderpreglow: the difference is that the Archos had an internal version in case it found no disk-based
10:13:36preglowahh, right
10:13:46preglowipod's got the disk mode
10:13:54Membrillodoes the ipod wheel respond to tapping in certain locations, or do you have to drag across for a response
10:13:55preglowand the diagnostics
10:14:06preglowMembrillo: drag, but it's not impossible to code tapping support
10:14:22LinusNpreglow: i believe i gave him the docs i had
10:14:52preglowdecoding oggs on a 60mhz arm???
10:14:54Membrillopreglow, so would it be possible to be, for example, holding on left?
10:14:58preglowthat's pretty hardcore
10:15:10Membrilloand get a continuous left response
10:15:23preglowMembrillo: possible, yes, but i don't think it's a very good idea
10:15:26Membrillolike holding a button
10:15:43Membrillowhy? it would be good for games, rather than scrolling around all the time
10:15:45preglowMembrillo: usually it's a bit hard to press a button by accident, but just brushing the wheel is enough for it to register it as a press
10:15:53 Quit webguest02 ("CGI:IRC (EOF)")
10:15:57preglowMembrillo: btw, holding might not work
10:16:06preglowMembrillo: it might just register one press, then nothing more
10:16:10Membrilloah ok
10:16:27Membrilloi guess alternatives are still coming out
10:16:31preglowi wonder if iriver use the epic core at all
10:16:40preglow24 bit dsp, sounds a bit like what i'm using now
10:18:30 Join mikearthur [0] (
10:18:44preglowbet it's easy to get tools for it...
10:19:44 Quit San (Read error: 110 (Connection timed out))
10:20:46 Join linuxstb_ [0] (
10:20:52 Quit edx (Read error: 110 (Connection timed out))
10:29:36 Join Jungti1234 [0] (n=jungti12@
10:31:07linuxstb_preglow: Sorry about breaking your Nano build. I shouldn't make a big commit and then go to bed...
10:32:13preglowdidn't notice they were broken
10:32:40preglowbut feel free to fix ;)
10:32:49preglownot going to compile for a while anyway
10:33:06nathanhhrm, im trying to grok gwps, i got a question
10:35:08 Join tatham [0] (
10:35:21linuxstb_preglow: I'll just disable the two broken plugins (Bejeweled and Sudoku) for the Nano.
10:35:24 Join tathamoddie [0] (
10:35:36linuxstb_They need bitmaps for your size LCD.
10:35:44tathamoddiehowdy folks ... anyone able to answer a H300 question quickly?
10:36:46Bagdertry us
10:36:51nathanhtatha: quickly is guaranteed, accuracy is 50/50
10:37:02tathamoddieok - i just installed the rockbox stuff
10:37:08tathamoddie(i love it already btw)
10:37:12tathamoddiei installed the bootloader
10:37:14tathamoddieworks fine
10:37:25tathamoddiei can see that screen ... and i can revert back to iriver
10:37:50tathamoddiebut in a stupid case of me not reading the doco very well i didnt copy the .rockbox and rockbox.iriver files to the root
10:37:57tathamoddieso it has the bootloader but not the software
10:38:07tathamoddiemeaning it always defaults to iriver firmware
10:38:17tathamoddienow the device wont mount as a hard drive though in win xp
10:38:26tathamoddieyou plug in the USB cable and it just says "Charging"
10:38:32nathanhpress the play button
10:38:35nathanhit will then connect to XP
10:39:04tathamoddiei was pushing and holding
10:39:08tathamoddieturns out you need to tap
10:39:13tathamoddiethanks guys :)
10:39:17tathamoddieyou really do rock
10:39:39nathanhok, so im trying to grok gwps... i can see the parsing code that loads a .wps file into wps_data
10:39:58nathanhi dont grok the relationship between the parsed code and wps_refresh
10:40:18Bagderthe wps code is hairy
10:40:33nathanhis the contents of wps_data a linear representation of the .wps file?
10:40:46Bagderyes I believe it is
10:40:46nathanhso if i iterate over the array of lines, its the same thing as iterating over the .wps file?
10:41:03Bagderexcept that it has skipped comments etc
10:41:15nathanhwhat i want to do is stick a new tag in for color and add a routine to wps_refresh that changes foreground/background
10:41:28nathanhwhich i think will be simpler than adding a new parameter to every existing tag
10:41:47Bagdersounds fair
10:41:52LinusNpreglow: i just studied the SAA7750 data sheets some more, and it doesn't seem like the EPICS7A core is meant to do codec tasks
10:41:55nathanhive dotted lcd_set_foregrounds around the code inside wps_refresh and ive got a funky looking colour display already
10:42:01nathanhbut its not very useful like that
10:42:34LinusNit looks more like postprocessing, like ultrabass, eq and stuff
10:42:44nathanhok, so what i dont understand is where wps_refresh interprets the tags
10:43:03nathanhi see the flags variable, but i dont see how that relates to the tags
10:43:44Bagderhey! we missed the anniversary!
10:43:57Bagder" First mail! ;-)"
10:44:44 Quit use_off ()
10:44:44 Quit tatham ("CGI:IRC (EOF)")
10:49:01preglowLinusN: they do specify it can decompress audio
10:49:06tathamoddie(sidenote: i've just installed rockbox on a mate's h340. he's blind and this is an absolute godsend for him. thanks guys)
10:49:31tathamoddie(is there any way to get it to pronounce artist names / song titles ... we have voice menus working)
10:49:51amiconnLinusN: Last night you said lcd_update() takes 6.7ms at 45 MHz, and later that it takes 26ms. Which value is true, or which value relates to what lcd_update() method/version?
10:49:57Bagdertathamoddie: there's a script available that generates spoken file names
10:50:14Bagderor at least directory names
10:50:18tathamoddiebadger: where am i looking / what am i searching for?
10:51:15tathamoddiethanks again
10:52:11LinusNamiconn: the latter one is correct, i'm afraid
10:52:34LinusNand it's the word-by-word dma version
10:52:34preglowhooray for us! hooray for rockbox!
10:52:54nathanheek, 26ms
10:53:00preglowLinusN: but yeah, the propaganda sheets say they can be usde for decompressing audio and for postprocessing
10:53:13nathanhis that perhaps the cause of the ticking sound that happened with the earlier iriver h300 firmwares and the lcd remotes?
10:53:18preglowLinusN: a 100mips dsp core should indeed be good and solid for decoding audio
10:53:40LinusNwould be nice
10:54:13preglowand nearly impossible
10:54:21preglowunless someone coughs up some datasheets for it
10:54:52linuxstb_preglow: You can safely "cvs update" now if you wish.
10:55:02LinusNthe epics core is located between the arm and the CODEC
10:55:48preglowLinusN: gr8
10:56:02 Quit tvelocity (Remote closed the connection)
10:56:08linuxstb_LinusN: Does it have code in ROM or does it have to be loaded by the main firmware?
10:56:57LinusNit seems like the epic is preprogrammed, but it is "field upgradeable" with a 4kword instruction ram
10:57:01 Quit Membrillo ()
10:57:55 Quit linuxstb_ ("CGI:IRC")
11:00:43LinusNi wish some 3d gfx artist could render some nice gems for bejeweled
11:01:00 Join Polo_o [0] (
11:01:08nathanhwhy dont you nick the ones from gnome bejeweled and downsize them in gimp?
11:01:25preglowLinusN: i wish they'd just steal the ones from the gnome port
11:01:40LinusNare they up for grabs?
11:01:42*nathanh hi5
11:01:46preglowwell, it's gpl...
11:01:56preglowanyway, that's not the problem
11:02:02preglowit was some deal of transparent backgrounds
11:02:51nathanhbagder: they're purdy
11:03:21LinusNgo get'em!
11:03:44BagderI have no sim or device that can show colors ;-)
11:04:03*preglow kicks bagder towards the apple store
11:04:22LinusNi wish an sdl ninja could bless the x11 sim with some colors
11:04:22nathanhhalf tempted to get a nano and run rockbox on it
11:04:28nathanhthe apple fanboys at my work would have kittens
11:04:58preglownathanh: especially in the half working state it is now
11:05:08preglownathanh: it would be like seeing an ipod after a mountain climbing accident
11:05:10nathanhno audio?
11:05:27preglowno audio and glitchy controls
11:06:55preglowthe graphics are svg!
11:06:58 Join tvelocity [0] (n=tony@
11:07:01preglowthen we can scale them as large as wel want
11:07:04LinusNthe h300 file browser scroll still lags, even with the optimized lcd dma
11:07:20nathanhthe lags not so bad, but would be nice if releasing the button flushed the queue
11:07:45nathanhso as to not generate more events
11:08:09preglowLinusN: i have no lagging on nano, why is it so slow on h3x0?
11:08:43LinusN45mhz, lots of bitmap operations in a slow sdram
11:08:56LinusNthat's my theory at least
11:09:05preglowis nano framebuffer in iram, then?
11:09:12LinusNi dunno
11:09:20preglowif it is, it's got to be eating almost all of it
11:09:57preglowyes, 46k
11:10:15LinusNhow big is the nano lcd?
11:10:17preglowwe have 64kb iram for core on nano, though
11:10:29nathanhsheesh, it uses that much iram for a framebuffer
11:10:38nathanhmust be so they can get smooth video
11:10:45Bger77k for h300 iirc
11:10:45nathanhthe iriver video playback is jerky
11:10:46LinusNpreglow: that's a lot smaller than 220x176
11:10:50preglowi gues
11:11:18preglowlinuxstb: does your file browser lag?
11:11:20Bagderthat's 60% of the h3x0 screen area
11:12:10preglowipod video is 320x240...
11:12:24Bagdergigabeat too iirc
11:12:31 Join Kohlrabi [0] (
11:12:31 Quit nathanh ("CGI:IRC (EOF)")
11:12:32preglowi hope they use some special framebuffer ram
11:14:48***Saving seen data "./dancer.seen"
11:15:18LinusNBagder: "bag of shit" :-)
11:15:18amiconnLinusN: My calculations tell me that it should be possible to get H300 lcd_update() timings down to ~10ms with single writes, and to ~8 ms if multiple writes can be made working
11:16:10LinusNso far i've had no success with fast ram mode
11:16:18 Join nathanh [0] (
11:16:28LinusNand the lcd controller can't handle bursts at 124mhz
11:16:44BagderLinusN: I couldn't resist ;-)
11:17:17LinusNso it looks like we'll have toi resort to movem.w for reads and simple move.w for writes
11:17:38amiconn(There is no movem.w, only movem.l
11:18:16Bagdermy dns still resolves the former forum IP
11:18:21amiconnLinusN: Yes.
11:18:27LinusNBagder: indeed
11:18:29preglowBagder: mine too
11:18:38preglowthis is one of the longest dns propagation times i've ever seen
11:18:41LinusNBagder: i cheat with /etc/hosts
11:18:49BagderI'll do that too
11:18:53amiconnBagder: Then something is wrong with the TTL interpretation of your DNS. Works fine here
11:19:13amiconnLinusN: Btw, my timing estimations are for 45 MHz
11:19:21Bagderwell, it is my ISP's DNS
11:20:04 Join nathanh_ [0] (
11:20:04 Quit nathanh (Client Quit)
11:20:16LinusNamiconn: so it would be a movem.l, and a couple of move.w/swap combinations
11:20:42 Nick nathanh_ is now known as nathanh (
11:21:23LinusNthe problem with the fast ram mode seems to be that it updates the GRAM pointer wrong
11:24:09LinusNprobably because the CS is negated for every word
11:25:08linuxstbpreglow: The directory browser doesn't keep up with the buttons - but I don't know if that's a button driver issue or a file browser issue.
11:26:25linuxstbAnd no, we don't use IRAM for the 16-bit framebuffer.
11:26:56amiconnLinusN: Where is CS connected?
11:27:09amiconnWould be helpful if there were schematics...
11:27:18 Quit nathanh ("Quit")
11:27:20LinusNLCD CS -> Coldfire CS1
11:27:28LinusNsame as H1x0
11:27:55LinusNi simply haven't had time to draw schematics
11:29:13LinusNif only we could control the time between accesses with the CS logic...
11:29:32amiconnI see that CS1 is also GPIO58. What if we control that manually?
11:29:49LinusNwe can't, since the data bus is shared
11:31:31 Join nathanh [0] (
11:32:23 Quit mirak (Read error: 110 (Connection timed out))
11:32:24preglowlinuxstb: it doesn't here either, actually, but i suspect that's a button driver issue
11:32:52preglowdoes the file browser/menus etc. empty the button queue before they act?
11:33:03preglowi guess they don't
11:34:04 Join Aramil [0] (n=tony@
11:35:43 Join mirak [0] (
11:37:44 Join KN|stiff [0] (
11:38:04 Quit KN|stiff (Client Quit)
11:38:49*preglow reads up on doxygen
11:39:47amiconnLinusN: (1) Is it essential that the burst write is exactly 4 words, or can it be longer? (2) If you enable the BSTW bit for CS1, does the coldfire still negate CS between words?
11:40:08LinusNamiconn: it must be a multiple of 4
11:40:20amiconn(3) Does the LCD controller use the address bits?
11:40:44LinusN2) I don't know, i'll check
11:41:03amiconnIf the answers are 1-yes, 2-no and 3-no, it should be possible to use line bursts for LCD write
11:41:08LinusN3) yes, A1 is used for register select
11:41:36LinusNanyway, that doesn't matter, since the bursts are too fast
11:41:53amiconnEven with the appropriate number of wait states?
11:42:29LinusNthe wait states isn't the problem
11:42:42LinusNit's the time *between* writes that is the problem
11:43:43LinusNline-burst dma works fine in 45mhz
11:44:02 Quit tvelocity (Read error: 101 (Network is unreachable))
11:45:19amiconnSo then we could use different lcd data writes depending on CPU clock
11:46:55BgerLinusN what about adding a check to lcd_update[_rect] if the backlight is on and only if it's on to update the screen ?
11:50:45nathanhwoot, i have color wps
11:51:25linuxstbnathanh: Do scrolling lines work properly?
11:51:33nathanh? what do you mean ?
11:51:53linuxstbI mean do you just have one fg/bg colour, or do you change them for different parts of the WPS?
11:52:02nathanhdiff colors for diff parts
11:52:12nathanhi have a new tag, insert it anywhere, changes the color of all subsequent tags
11:52:47Bagderand do you have any scrolling parts of the screen?
11:53:00nathanhnothing scrolling
11:53:10 Join XavierGr [0] (
11:53:29Bgermorning, XavierGr ;)
11:53:54Bgerheh, yeah
11:54:06Bgerit's 12:53 in bg and greece
11:55:27linuxstbnathanh: The information about a scrolling line is stored in a "struct scrollinfo" (defined in export/lcd.h). This doesn't include bg/fg colours - which is why I think scrolling will probably not work as you intend.
11:55:30mirak11:56 in France
11:55:42nathanhok, i picked a long filename
11:55:48nathanhit scrolls fine in beautiful green
11:55:58LinusNamiconn: gah, it turns out that the bursts are too fast even in 45mhz
11:56:00mirakBger: are you at the sun time or is their a lag ?
11:56:21linuxstbnathanh: My guess it that scrolling lines use the last set of fg/bg colours you define on the WPS.
11:56:23Bgerno daylight
11:56:34Bgerjust UTC+2
11:57:08 Join KN|stiff [0] (
11:57:10mirakBger: two times a year our governement decide to change the time
11:57:17Bgerhere too
11:57:23LinusNBger: we will add a lcd_enable() function that will handle this, plus the power to the lcd
11:57:25mirakit's supposed to save energy
11:57:28 Quit KN|stiff (Client Quit)
11:57:28Bgerbut it's not in effect
11:57:39mirakit disturbs cows I have heard
11:57:45BgerLinusN anything i can do about this ?
11:58:00mirakcows says meuhh here
11:58:22nathanhclaris says meuhh
11:58:34mirakI had that discussion about animals shoots
11:58:43 Join HypnoticMonkey [0] (
11:58:44LinusNBger: i dunno
11:58:48mirakit's fun
11:58:53mirakbut of topic
12:00:08linuxstbnathanh: What do your new tags look like? Can you give an example?
12:01:02nathanh#C00ff00 is green
12:01:27nathanhim trying to grok this scrolling problem, see if its a showstopper
12:01:36linuxstbSo you use # instead of % for the tag?
12:01:44nathanhblah, sorry, force of habit
12:02:18 Join San [0] (n=sanitari@
12:02:48linuxstbRegarding scrolling, try setting a WPS with two scrolling lines, each using different fg/bg colours.
12:03:14nathanhaye, trying that now
12:03:59*amiconn wonders why a burst write with the appropriate number of wait states can be too fast
12:04:51nathanhboo, its a showstopper, all scrolling lines = color of last line in wps
12:06:09amiconnYes of course
12:06:10linuxstbI think it's relatively easy to fix. You need to add bgcolor and fgcolor to the scrollinfo struct in export/lcd.h (only for targets that have colour LCDs), change the LCD code to store and retrieve the colours from there for scrolling lines.
12:06:17amiconnScrolling needs rework
12:06:36nathanhaye, ill do that
12:06:54amiconnlinuxstb: That seems like a half-hearted fix to me
12:07:27linuxstbThe scrollinfo struct should contain all info needed to draw the line.
12:07:30nathanhid rather work it so scrolling lines were rendered as part of format_display
12:08:39amiconnlinuxstb: Yes, but the line based scrolling is too unflexible, plus I'd want some more attributes (like bold/ italic flags)
12:08:51nathanhhrm, i like the cut of your jib
12:09:15linuxstbamiconn: Is there any other solution apart from extending the scrollinfo struct to contain those extra attributes?
12:09:17preglowthen we need italic and bold fonts loaded all the time as well
12:09:30preglowcommence designing the new font system, someone
12:09:37amiconnnathanh: The scrolling has to be done in a separate thread if we don't want it really jumpy. That's the reason why it is part of the driver
12:09:37preglowi want to be able to use several fonts at one time
12:09:44linuxstbWe already have a glyph cache IIUC.
12:09:46nathanhformat_display is too slow to render each frame?
12:09:57amiconnpreglow: Algorithmic emboldening/ italicising
12:10:37amiconnnathanh: It might be slow, but the main point is that the scrlling has to be handled in regular intervals
12:10:59preglowamiconn: that's never pretty
12:10:59nathanhso thats gui_wps_refresh?
12:11:29preglowamiconn: you can't italicise a font algorithmically, just slant it
12:11:32preglowamiconn: italic != slanted
12:11:40amiconnYes I know
12:11:49nathanhstrikes me that the colors will have to be stored in the format_lines
12:12:41XavierGrHello all! Yes it's midday right now here :)
12:12:53LinusNamiconn: the wait states only define the time that CS is low
12:13:17LinusNbut you can't control how long it should take until the next access is started
12:13:32amiconnLinusN: For bursts, CS is low during the whole burst access according to the datasheet
12:13:39LinusNso the CS-low period is nice and long, but the CS-high period is too short
12:14:35LinusNyes, but that probably only works if the target address is auto-incremented
12:14:44LinusNand burst is enabled
12:16:14LinusNin my current dma setup, the sdram data is read in 16-byte bursts, and written with 8 consecutive word writes
12:16:35LinusNand that fails because the time between the writes is too short
12:17:22amiconnDoes the DMA controller allow burst writes without address increment?
12:17:52LinusNthe bursts are not controlled by the dma controller
12:18:16LinusNi just tell it to read entire lines from the source, and the sdram controller does the bursting
12:18:45amiconnWhat if you enable burst write for CS1 and use DMA?
12:18:55LinusNwill try
12:21:05 Join edx [0] (
12:21:17linuxstbamiconn: If adding bgcolor and fgcolor to the scrollinfo struct is "half-hearted", what do you have in mind?
12:21:33LinusNthe data sheet says that bursts are only done if the destination data size is larger than the chip select data size, i.e a 32-bit write to a 16-bit port
12:21:59LinusNand then the cs controller increases the address
12:22:28LinusNwhy couldn't they have put regsel on a higher address pin?
12:24:42LinusNamiconn: so i guess we'll have to try the movem/move/swap hack
12:27:23nathanhlinuxstb: the scroll thread redraws the string each time
12:27:36nathanhwith an optional pixel offset based on the current scroll state
12:28:09nathanhhow about storing the entire string rendered offscreen and blitting the required x/y/w/h of it onscreen
12:28:28nathanhthat way the string can be rendered in its desired color ahead of time
12:28:42linuxstbnathanh: That will require a lot more memory for the scroll buffers.
12:28:49nathanhalso means we could scroll bitmaps
12:28:54Bagderand it doesn't handle transparancy
12:29:11nathanhit could handle transparency with an alpha channel (even a 1bit channel)
12:30:32linuxstbCurrently the scrolling code can handle 26 lines, each containing 255 characters using any size font.
12:30:42linuxstbThat's a large bitmap.
12:31:30nathanhat the very least id like to try blitting the onscreen bitmap to the left/right, then only rendering the bit that needs to be re-rendered
12:32:38Bagderunless there's a bg pic ;-)
12:33:39Bagdermy cell phone has a bg picture present in all screens and it looks nice
12:33:55linuxstbmy cell phone has a bg picture present in all screens and it looks horrible
12:34:09preglowmine has a bg on the main screen
12:34:15preglowlooks horrible, but that's thanks to me
12:34:47nathanhthe current scrolling code flickers because it does lcd_fillrect followed by lcd_putsxyofs
12:35:01nathanhif it rendered offscreen then blitted onscreen, it would not flicker
12:35:13LinusNnathanh: why would that flicker?
12:35:15linuxstbIt probably flickers because the lcd update is slow.
12:35:45LinusNnathanh: all lcd drawing is done offscreen
12:35:49linuxstbnathanh: The LCD is only written do during an lcd_update()
12:36:03nathanhahh, ok
12:36:10nathanhso its flickering because its a cheaparse lcd :-)
12:36:36LinusNno, it's a cheaparse engineer who connected the lcd on the pcb
12:37:17Bgersome poor chinese working for a cup of rice a day
12:37:27LinusNkorean, rather
12:37:43Bgermanifactured by iriver, made in china
12:38:21aliaskBut who ordered the chinese guy to do it that way?
12:38:30aliaskWow this is so deep... and pointless.
12:38:37nathanhim looking at puts_scroll, you can supply arbitrary xpos/ypos, thats the top-left corner
12:38:42Bgerthe last was not said as something funny ...
12:38:55nathanhbut i dont see a value for a clip window... does it render right up to the right edge?
12:40:43nathanhblh, i see that it does
12:41:06nathanhso you can have a scroll that extends from halfway on the screen to the right edge of the screen
12:41:16nathanhbut you cant have a scroll that is completely detached from both edges of the screen
12:45:14 Quit edx (Read error: 110 (Connection timed out))
12:46:26 Quit Aramil ("Leaving")
12:46:36Bagder"realloc: warning: points to free buffer" MPeye uses malloc() ! ;-)
12:46:42 Join Petur [0] (
12:47:28 Join Moos [0] (
12:47:29BgerPetur hi
12:47:34PeturLinusN: is there hope for a bootloader that can do USB?
12:48:23PeturIt came to my mind that I lost a lot of time yesterday evening booting into the iRiver FW when my test crashed
12:48:30Bagdernow why can't I ever remember how to force objdump to dissassemble a binary file...
12:48:39PeturBger: hi, thanks for fixing my patches
12:48:57BgerPetur heh, for nothing
12:49:07Bgerbtw, i suggest you to use cvs, really
12:49:27linuxstbBagder: "-b binary"
12:49:44Bagderright, thanks!
12:49:57linuxstbSomething not obvious from the man page
12:50:02LinusNPetur: yes
12:50:38PeturI like TortoiseCvs, I'm no longer used to command line stuff ;)
12:51:11PeturI think it can put all patches into one file, and after that I'll use a line-ending-fix tool
12:51:35BgerBagder the next to the last string of strings dump is funny
12:52:12Bagderwhich is that?
12:52:25 Quit elinenbe (" HydraIRC -> <- IRC with a difference")
12:52:36BagderI'm a bit surprised there aren't more strings available
12:52:36Peturthis webclient is nice... but it's not good for my growing rockbox addiction
12:53:15Bger"A sibal jola jjajungna jakku ssang soriman nane!!!"
12:53:16BagderI love the paths to source code in the firmware:
12:53:35HypnoticMonkeypetur: agreed
12:53:43BagderBger: it sure sounds korean to me! ;-)
12:54:00Bgermore interesting are "http://", "ftp://" and "file://"
12:55:37Bgerhow this "unsigned char table[] = { ... }; has been included ...
12:56:31preglowmarkun: you get the gigabeat sources yet?
12:57:23markunnot yet
12:58:15markunI posted a message on mygigabeat asking for a defective player, but the chances of that will be pretty slim I guess
12:58:25BagderBger: I'm quite sure it executes code at 0x10000000 -
12:58:55 Part LinusN
13:00:44linuxstbAnyone understand this line: "A sibal jola jjajungna jakku ssang soriman nane!!!"
13:00:56linuxstb(from the mpeye firmware)
13:01:17BagderJungti1234: here?
13:01:51linuxstbThe three exclamation marks are making me curious...
13:01:56linuxstbMust be important
13:02:19Bgerlike "what, the fsck!!!"
13:02:32preglowwe will sure you for looking!!!
13:02:43nathanhor "i like exclamation points!!!"
13:03:16Bgeror "i must stop being lazy!!!"
13:03:56nathanhok, ive haxored the code and i now have scrolling lines in diff colors
13:03:59nathanhi am leetzor
13:04:18nathanhmultilines even work properly
13:04:24nathanhdiff colors for diff times of the multiline
13:05:46Bgernathanh submit patch!!! :)
13:06:04nathanhyou betcha
13:06:16 Join tucoz [0] (
13:06:24 Quit ashridah ("later")
13:08:00linuxstbBger: Sorry, I didn't see you posted the same quote as me five minutes earlier!!!
13:08:16tucozhello, I thought my h120 had died on me today. When I pressed on play nothing happened. Probably out of juice. But, as soon as I put the charger in it started without me pressing anything. Guess it had enough power to register the button, but not enough to start anything else.
13:08:18Bgerhahaha :)
13:09:03preglowcpu running at about 0.5hz
13:09:05Bgerok, tripple exclamation marks are stylish these days ... :)
13:10:19tucozpreglow, 8008 emulation ;)
13:10:37tucozoh, 0.5hz even
13:10:49tucozI thought you said MHz
13:13:52tucozhmm, how did the ifp-porter manage to decode oggs with the original firmware? It is way beyond me to understand his porting approach.
13:14:11Bagderhe's writing an emulator for the cpu
13:14:20tucozIs it supported by the original fw? oggs that is
13:14:29tucozoh, I see :)
13:14:32preglowbut damn
13:14:35preglowogg decoding on 60mhz arm
13:14:39preglowpretty cool
13:14:50***Saving seen data "./dancer.seen"
13:15:00preglowbut btw
13:15:03tucozI thought he somehow managed to add ogg decoding to the original fw. hehe
13:15:07CassandraHmmm. Yoghurt does not improve the looks oif the Nano at all.
13:15:13preglowhis patch says it removes the low bitrate ogg limit
13:15:21preglowdoes it actually work with low bitrate oggs, then?
13:15:23nathanhis there any way to run rockbox normal interface without detaching UMS?
13:15:25tucozI see, anyway. I am impressed
13:15:32preglowCassandra: try coffee
13:15:42Bagdernathanh: no
13:15:46tucozcoffee with lot's of cream and lot's of sugar
13:16:05Cassandrapreglow: Coffee crime! Every drop is sacred.
13:16:11Zagorwe sure have a shedload of ports going now
13:16:20Zagorgoing on
13:16:24tucozI understand the emulator approach. It was the ogg part I didn't get.
13:16:36Bgerso "~" is binary not ?
13:16:36tucozZagor, quite a monster you have created :)
13:17:12BagderBger: it is a one-bit complement operation, yes
13:17:33CassandraWell, it's not so much the number of active ports as the number of useable ones.
13:17:49CassandraMost ports don't seem to make it to that stage.
13:17:50*Bger hides
13:17:56nathanhim really really _really_ impressed with rockbox
13:18:02tucozwas the intention in the beginning to just support one target, or were the multiplatform in your mind from the start?
13:18:07preglowZagor: the more the merrier
13:18:14nathanhgapless playback, 7 seconds ontime, its so much better than the original firmware
13:18:15BgerCassandra i think that even h300 port is pretty usable atm
13:18:23preglownothing like a big multiplatform mess to round off the day
13:18:40Bagdertucoz: we did focus on more than one Archos player from day 1
13:18:45ZagorCassandra: absolutely. I'm just glad people are taking time to try the effort, even if most probably won't complete.
13:18:55CassandraBger, I'd agree with that. If it plays music, it's pretty much useable. Although I don't regard the H300 port as particularly different from the H100.
13:18:57tucozBagder, ok. Not like Linux then.
13:19:11CassandraCertainly no more so that the Recorder port was from the player.
13:19:18nathanhlinux was never multiplatform
13:19:26nathanhthe first version linus even said "this will only ever be for i386"
13:19:38Zagortucoz: that is the reason the two dirs "apps" and "firmware"
13:19:43tucozyes, I've read that discussion. Funny.
13:19:52tucozZagor, I see.
13:19:56nathanhwhats the res of the h300 screen
13:20:10Bgerwhat ? the same LCD controller on h300, nano and photo?
13:20:12Bgernathanh 220x176
13:20:42Bgernathanh see the DeviceChart topic in the wiki
13:20:57Bagderit would be neat to have the ipod color and nano added to that
13:21:10CassandraI didn't realise the Nano LCD controller was the same as the H300.
13:21:35CassandraBagder: I should be able to give that a little more time after Christmas.
13:21:53CassandraAlthough my low level skills are weak and puny.
13:22:12Cassandra(Well, in the new year, really)
13:22:41preglowit'll be playing audio by then, mark my words!
13:22:54*Bger writes down
13:23:15preglowwell, at least if the arm core is as powerful as it looks like right now
13:23:17linuxstbpreglow: Assuming we can fix the ATA problem.
13:23:23preglowlinuxstb: what at problem?
13:23:27CassandraThis is my unconvinced face. It doesn't come over well on IRC, I'm afraid.
13:23:42linuxstbpreglow: You told me that the ATA driver doesn't work properly.
13:23:47linuxstb(on the nano)
13:24:52CassandraAh, is that why I can't boot the original firmware then?
13:25:04linuxstbNo, I think that's a different problem. I'm still investigating.
13:25:13preglow09:37 < preglow> linuxstb: btw, i've figured out why disk accesses sometimes work and sometimes not. it turns out it never works after the spin down time has been exceeded
13:25:38Bagderyes, the flash spinning down can't be good ;-P
13:25:41linuxstbOK, so how do we fix it? Never spin down the disk on the Nano?
13:25:42Cassandralinuxstb, if you want any testing done, mail me or grab me when I'm around.
13:25:52preglowlinuxstb: sounds like a good temp solution to me :-)
13:26:07Bgerone less thread on nano ? :)
13:26:42linuxstbpreglow: OK, I'll look at that and give you a new ata.c to try.
13:26:49CassandraI'd have thought that an instruction to spin down the flash'd just do nothing, myself.
13:27:01preglowCassandra: it might and it might not
13:27:08preglowCassandra: it might have some power preservation function, for example
13:27:29Bagderand obviously it does something
13:27:55CassandraI guess a trawl through the iPodLinux ata code may be in order.
13:32:07 Part tucoz ("Leaving")
13:33:32preglowwell, it's based on ipl code as is
13:34:20 Part Petur
13:35:11linuxstbpreglow: As a quick test, I think you could remove the call to create_thread() in drivers/ata.c, and also remove the queue_post() call in ata_sleep() in the same file.
13:35:24linuxstbThat should prevent the sleep code from ever being called.
13:35:55preglowwill afterwards
13:38:18stevebok. if you were going to buy an mp3 player right now which one would it be? and it has to have >30GB of space...
13:38:41nathanhanybody who is interested, the colorised wps patch is up on sf
13:39:10nathanhsteveb: right now, id buy the iriver h340, cos it runs rockbox!
13:39:26nathanhbut if it didnt run rockbox, i wouldnt recommend the h340
13:39:31stevebheh cool. i used to own an h340.
13:39:34stevebthey are pretty sweet
13:39:43stevebhaving a hard time finding people who sell them in the uk tho
13:39:48nathanhcrappy startup time, no gapless, otherwise they're ok
13:39:52linuxstbsteveb: The best Rockbox target at the moment is the H140 - but they are discontinued.
13:40:21stevebi see that the h340 rockbox stuff is coming on quite fast tho
13:41:17stevebnathanh: start up time doesnt bother me... once its running its running
13:41:55nathanhrockbox on the h340 is great... when i stop the car to duck into the shops, it pauses, when i restart the car, within 2 seconds it has resumed
13:42:15stevebheh cool.
13:42:33linuxstbThe only other choices are the iAudio X5 (port planned, but not really started yet), and the iPod (but not the new video iPod - the version just before that model)
13:42:33steveboh for the record, trying to install flash linux to a partition on the h340 is a Bad Thing (tm)
13:42:42stevebfuck the iPod...
13:42:47stevebi hate those things
13:43:05stevebi do not see why people like that silly wheel thing so much
13:43:30stevebill check out some h340 prices
13:43:42preglowclick wheel = lovely
13:44:08markunsteveb: more control over the scroll speed than with buttons maybe
13:44:26stevebmaybe... it just annoys me
13:46:08HypnoticMonkeyits not very intuitive to scroll downwards by making a semi circle as far as im concerned. i feel that vertical touch pads are superior.
13:46:37preglowi think it's very intuitive
13:46:41preglowplus, it's more efficient
13:46:51preglowanyway, it's just a preference
13:46:58preglowuse whatever floats your boat
13:47:53 Join Febs [0] (
13:48:15 Join edx [0] (
13:48:15stevebno one seems to be selling the ones with the remote any more :-/
13:48:34linuxstbpreglow: I don't know why, but enabling dircache on my ipod caused it to stop booting up.
13:49:13nathanhive had lockup problems on the h340/rockbox with dircache on
13:49:23nathanhrequiring a reset with a pin
13:49:42preglowthe clock now overwrites the hd indicator in the status bar
13:49:45preglowany way to fix that?
13:49:49linuxstbnathanh: When did it lock up?
13:50:09nathanha few seconds after the thing booted
13:50:11linuxstbpreglow: I guess the ipod is the first target with both a clock and virtual LED. The h300 will have the same problem though.
13:50:20nathanhso you can feel the disk chattering away, building the dircache
13:50:22nathanhand a song is playing
13:50:25nathanhthen it just locks solid
13:50:41linuxstbpreglow: I guess it's a simple fix in the status bar code.
13:50:47Zagorthat's because dircache is evil ;-)
13:51:24preglowdircache is god
13:51:31preglowa very gentle and nice god
13:51:38nathanhdircache is awesome, disk doesnt spin up when you move around the tree
13:51:40preglowrockbox without dircache is a pain to use
13:51:47nathanhthats a zillion times better than iriver firmware
13:52:17linuxstbnathanh: So building the cache at the same time as playback causes problems for you sometimes?
13:52:19nathanhi adore the fadeout/fadein when you pause... i owe a beer to whomever wrote that
13:52:28nathanhlinux: yeah, not all the time, just once in a while
13:56:30Slasherinathanh: that is probably caused by too long file names
13:56:47Slasherirecent unicode patch broke rockbox's support with long files
13:57:11 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
13:57:51nathanhahh, k
13:58:11nathanhis anybody working on gui code? id like to collaborate
13:58:33nathanhim looking at these screens -> <- and i want my rockbox to look the same :-)
13:59:08preglowSlasheri: why hasn't that been fixed yet? should be trivial
14:00:18Slasheripreglow: yes, but i don't know.. That doesn't matter whether the dircache is enabled or not.. When you encounter a long file name (for example browsing directories), rockbox instantly hangs needing to reset
14:00:27aliasknathanh: Maybe not look the same, but have the same amount of configureability.
14:00:37preglowyes, i know, why hasn't it been fixed in rockbox?
14:02:14linuxstbSlasheri: What do you mean by too long? You mean longer than MAX_PATH after being converted to utf-8 ?
14:02:25Slasheripreglow: i think the problem was that increasing MAX_PATH would also increase stack sizes.. But something has to be done on the file system level if it exceeds the MAX_PATH limit
14:02:46Slasherilinuxstb: yes, something like that. I don't know exactly what triggers the crash
14:02:56Slasherionly that long file names causes it
14:03:43nathanhwhy not use dynamically sized strings, rather than an array of [MAX_PATH+1]
14:03:52linuxstbpreglow: Try
14:03:53preglowwe don't do dynamic allocation
14:04:04nathanhwhy not
14:04:09Bgerurgh, this must go in the genreal FAQ
14:04:10Slasherithat would require to use dynamic memory allocation and it is not currently implemented
14:04:14*Bagder prepares for the why-not-malloc talk ;-)
14:04:20preglownathanh: because we would have to reserve the max amount of memory anyway
14:04:27preglownathanh: we might as well just do it statically
14:04:41preglowsomeone desperately needs to put up a faq entry on this
14:04:55*nathanh fears he has struck a nerve
14:05:09Bagdernathanh: we have this conversation with all newcomers
14:05:13linuxstbnathanh: It's a question every new dev asks.
14:05:27Bgernathanh see, we want to use all available memory for buffer
14:05:29nathanhso its a frequently asked question
14:05:33preglowoh yes
14:05:38nathanhif only there was some easy to deal with these frequently asked questions :-)
14:05:49linuxstbIt's already patented
14:06:31nathanhso the buffer cant be dynamically resized, so theres no point in dynamically resizing whats left
14:06:33nathanhis that the argument?
14:07:05preglownathanh: the argument is that if we were to have malloc, we also need to have a malloc buffer, which would mean we can't use that memory for file buffer anyway, so there's no point
14:07:15preglownathanh: in practice, we would use _more_ memory than if we had used static allocation
14:08:00FebsCan anyone tell me what causes the "failed to load archos/_temp_codec.dll" error on the sim? I searched the IRC logs and saw this asked a few times, but never answered.
14:08:01Slasherinathanh: the main problem is the fragmentation of memory, but that could be voided with refreshing the memory pointers
14:08:17preglowSlasheri: i don't see that as the main problem at all, the main problem is that it's pointless
14:08:29 Quit tathamoddie ()
14:08:31Slasheripreglow: hmm, true
14:09:02preglowSlasheri: you a) end up having to worry about memory leaks, b) will use more memory for malloc buffer than you did for static buffers in the first place
14:09:13preglowSlasheri: two negatives and no gains
14:09:23preglowplus, the code gets bigger
14:09:26nathanhnot having dynamic heap is gonna make it hard to implement compositing
14:09:33nathanhwhich is my personal interest for coding
14:09:39preglowno, it's not
14:09:43preglowyou just grab as much memory as you need
14:09:48Slasheriunless we would use that malloc buffer almost for everything allocation, then it would become more effective. But then we would have performance problems and the memory leaks
14:09:52preglowif you think you might need a megabyte, then you grab a megabyte
14:09:57Slasheripreglow: yes
14:10:00nathanhdepends on the wps is the problem
14:10:13nathanha wps with a dozen bitmaps might need 100kb, with a 100 bitmaps might need 1mb
14:10:16Bgernathanh you can ask for the audio buffer
14:10:34Bgerhmm, but yes not for the wps
14:10:36preglowSlasheri: space-wise, using malloc will never be more effective than using static buffers, unless you're willing to run the risk of having malloc return NULL sometimes
14:10:41nathanhim guessing the audio buffer must be contiguous
14:11:17Slasheripreglow: yep, or we could just block the malloc until it succeeded (with the risk of dead-locking whole system)
14:11:26preglowin the end, i can't see any reason
14:11:35preglowif i don't have to use malloc, i'm all the merrier
14:11:35nathanhstatic allocation upfront certainly makes the system easier to debug
14:11:41preglowi hate having to free stuff
14:12:07preglowand why the hell people actually _want_ malloc, is beyond me
14:12:09preglowyou're all crazy!
14:12:11preglowi mean !!!
14:12:31*nathanh regrets asking
14:12:51preglowwe currently waste half a meg on malloc buffer for codecs :///
14:12:57Bgernathanh the problem is not that u ask, but that nearly everyone asks ...
14:12:58preglowi'm sure that can be cut down on, though
14:13:09preglowBger: and that i see it as blatantly obvious...
14:13:28nathanhone benefit of a truly dynamic system - audio buffer too - is that the audio buffer could be a bit larger
14:13:36Bgerbtw, what's the interface between the h300's lcd and cpu ?
14:13:36preglowi wonder if i ever asked about this once upon a time
14:13:36Bagderno way
14:13:38preglowcan't remember it now
14:13:41nathanhbut i guess it dont matter too much
14:13:46preglownathanh: no it wouldn't
14:13:50preglownathanh: it would be smaller
14:13:56nathanhhows that
14:14:04preglowbecause you have to reserve more memory for the malloc buffer
14:14:06nathanhbecause of the overhead?
14:14:19nathanhof storing the free list?
14:14:20linuxstbYou don't want a malloc to fail - so you have to keep the memory unused.
14:14:24Bgerpreglow except the case when the audio buffer is dynamic also
14:14:27nathanho i c
14:14:31preglowbecause of a) you can't ever be sure how much memory is going to be allocated, so you're going to have to overdesign the limit a bit. b) fragmentation will assure that you need even more space for the buffer
14:14:54Bgerbut yes, b)...
14:15:20preglowthe malloc support data itself is also another matter
14:15:32preglowmalloc also makes it difficult to use iram
14:15:33Slasherianother way could be to wait memory to be freed from the audio buffer, and then malloc that.. :)
14:15:34Bgerwhat's the connection between the lcd and cpu in h300 ?
14:15:35preglowso there !!!
14:15:41Bgerparallel or SPI
14:15:49linuxstbSlasheri: Stop!!! :)
14:15:52Slasherihehe :D
14:16:02BgerSlasheri : that was what i wanted to say
14:16:08preglowwould be the best status message ever
14:16:21nathanhdo the .rock files have pre-alloc as well?
14:16:29preglowhaving to wrap every malloc with a lcd_puts that says "Trying to malloc, stand by!!!"
14:16:39preglownathanh: yes
14:16:40linuxstbUnless the status message scrolled, in which case you would have no memory to scroll it...
14:16:41Bgernathanh there are 768kb reserved for plugins
14:16:44Bgeron irivers
14:16:49Bagder"no memory received yet, we continue to wait..."
14:17:06nathanhwhere should i look to see the memory map
14:17:19linuxstbapps/ in your build directory.
14:17:19Bgernathanh u can also get the audiobuffer from a plugin
14:17:39Bgerplugin_get_buffer or something simillar
14:18:00Bgeramiconn ?
14:18:02linuxstbBger: That gets what is left out of the 768kb. There is also get_audio_buffer() or similar.
14:18:45nathanhthat gapless on an ogg album is just _perfect_
14:18:50Bgerlinuxstb u mean what lefts after loading the code of the plugin
14:18:50mirakit's a choice to not have the function headears in the .h ?
14:19:08linuxstbBger: Yes - code and static data
14:19:17Bgeruf, yes
14:21:32nathanhinteresting, a long list of albums scrolls faster when music is playing than when its stopped
14:21:39nathanhim guessing that cpu_burst isnt enabled for list scrolling
14:22:34linuxstbnathanh: Correct. It will be sad if the CPU needs to be boosted just for the GUI code.
14:22:52nathanhit does need to be; button press events when holding the button occur faster than it can scroll
14:23:02nathanhor it needs scroll skipping
14:23:18nathanhcurrently the events queue up and when you let go of the button, it keeps scrolling for a few seconds
14:23:32preglowcredits also scroll faster with music playing
14:24:06linuxstbnathanh: The alternative is to speed up the gui code.
14:24:06 Quit HypnoticMonkey ("CGI:IRC (EOF)")
14:24:35preglowlinuxstb: i wonder how we should handle that with click wheel scrolling
14:24:43preglowlinuxstb: will be very awkward if that is boost dependant
14:25:35nathanhneeds skip; the location in the list is determined by the button presses, but what the list displays is a best effort
14:25:56linuxstbI'm thinking the button driver could send page_down events.
14:26:43linuxstbDid you say that the hardware seems to have its own button queue?
14:27:10nathanhdunno, when i let go of the button it keeps scrolling for a bit
14:27:21nathanhand if i hold the down button, then let go and press left
14:27:28nathanhit scrolls for a bit and then jumps up a level in the tree
14:27:29linuxstbnathanh: Sorry, I was talking to preglow about the iPod :)
14:27:34Bgernathanh regarding scrolling: do you know that u can scroll one "page" at a time if you hold down the play button while pressing up/down
14:27:45nathanhwell that's obvious!!!
14:46:16 Quit nathanh ("Quit")
14:46:56linuxstbAny H300 owners around? Can you charge via USB without entering USB disk mode?
14:48:27mirakdoes rockbox works for big sound files over 30 megs for exemple ?
14:48:57miraklinuxstb: you can't charge and play at the same time
14:49:05mirakor use UMS
14:49:13mirakvia USB
14:49:38preglowmirak: yes
14:49:45linuxstbmirak: You mean that you can either 1) Charge, 2) Use UMS or 3) play music
14:49:54preglowmirak: though it might have bugs, depends on the codecs
14:49:58mirakis there some docs about the audi api ?
14:50:07 Quit Nibbler (Read error: 110 (Connection timed out))
14:50:12linuxstbSadly not.
14:50:16miraklinuxstb: with usb yes
14:50:20 Join Nibbler [0] (
14:50:39mirakyou can't use UMS or listen music and charge at the same time via usb
14:50:48Bgerlinuxstb what about charging ?
14:51:19Bgerh300 can be charged through USB
14:51:22mirakcould you point me where a music reading sequence start in the code ?
14:51:31linuxstbBger: mirak has answered - I was wondering if you could charge via usb at the same time as using the player normally to listen to music.
14:51:48Bgerbut when u charge it this way u can't do anything other with it
14:52:07Bgerbtw, i think it should be possible...
14:52:23mirakit would be cool
14:52:43linuxstbIt's a question for Linus when he's next around.
14:54:03linuxstbI need to explore the same issues on the ipod - but the ipod does UMS in software so it should be more flexible.
14:54:28preglowit does?
14:54:45linuxstbI'm pretty sure it's a routine in flash.
14:55:13markunDoesn't the only have code to go into charging mode when you press a key and insert the USB cable?
14:55:13mirakwow the codec api seems tuff
14:55:20linuxstbAlso, the ipod appears in Linux with a hard disk model name of "ipod" - not the real hard disk model name.
14:55:36linuxstbSo something odd is going on somewhere.
14:55:44mirakit's same for H300
14:55:56preglowlinuxstb: i get the ipod name here
14:55:57mirak /media/H300
14:56:01preglowlinuxstb: that is, what i named it in itunes
14:56:23 Join Jungti1234__ [0] (n=jungti12@
14:56:28Bgerlinuxstb : for the h300 i'm pretty sure this info is written in the config eeprom of the usb->ata bridge
14:56:45Bagderyes, the ondio has the option to not go into UMS mode when you insert the cable
14:56:48linuxstbBger: OK, that's possible.
14:57:11mirakthe codec api redirect data directly to pcm device or is it possible to output to something else ?
14:57:21preglowlinuxstb: that means we can do pretty much cool stuff in usb mode, then...
14:57:23mirakdo we have the choice ?
14:58:11linuxstbpreglow: In theory. I believe the IPL people are working on the usb hardware.
14:58:20BgerJungti1234__ hi. could u translate this "A sibal jola jjajungna jakku ssang soriman nane!!!"
14:58:41Jungti1234__It means bad.
14:58:44Jungti1234__very bad
14:58:50mirakI mean is the codec api reusable for something else than audio ?
14:59:01mirakthan audio stream
14:59:17linuxstbmirak: Are you thinking about video?
14:59:18Bgerhuh ... swear-word ?
14:59:32miraklinuxstb: yes ... I have no global view of what happens
14:59:51miraklinuxstb: some info would be welcolme. Maybe you can tell me where to look first
14:59:55mirakin the code
14:59:57linuxstbYes, the existing codec api will work perfectly well for video with virtually no changes.
15:00:01BgerJungti1234__ ?
15:00:09linuxstbYou will just need a way to disable the display of the WPS.
15:00:16miraklinuxstb: you mean all the buffering part
15:01:05linuxstbThe playback code is responsible for loading the data from disk and buffering it. You basically just access it as a stream via read() commands.
15:01:23Jungti1234__Do you want translation?
15:01:24linuxstbAnd you have a function you can call to send audio data to the DSP.
15:01:50miraklinuxstb: ok. There is some exemple about how to use xvid, the interface is quiet simple
15:01:57linuxstbIf it isn't there already, you will need to add the lcd functions to the codec api so you can write to the screen.
15:02:19BgerJungti1234__ yes, please
15:02:31linuxstbThe codec API is in apps/codecs.[ch] and the plugin API is in apps/plugin.[ch] - you can copy things that are in the plugin API and make them available to codecs.
15:03:12linuxstbBut you probably just need a pointer to the lcd_framebuffer array, and the lcd_update() function.
15:04:44linuxstbmirak: Does the decoder read the input data as a stream or does it do a lot of seeking?
15:05:00Jungti1234__Bger: That's abusive.
15:05:17linuxstbIt comes from the firmware of the MPeye MP3 player.
15:05:27Bgerheh, linuxstb said it before me
15:05:34Bgerhaha so abuses in the firmware
15:05:49Jungti1234__a stream of abuse
15:05:58Bgersomething like "why the f*** are you looking here" ? :)
15:06:06Jungti1234__f u c k
15:06:31Jungti1234__Who used such word?
15:06:50preglowsome witty engineer who just stringed along some curse words
15:07:05Jungti1234__He is bad guy.
15:07:12preglowlike skeletor?
15:07:21linuxstbThat's what using malloc does to you.
15:07:22Zagorooh, nasty. illinstr during playback in archos version 2.5.
15:07:36miraklinuxstb: that's a stream, you put data from file into a buffer, then you give the buffer begining to the decode function. The decode function returns a used_bytes value that help you to know where you are in the buffer, then you call again the decode function with buffer+used_bytes
15:07:53preglowZagor: easily reproducible?
15:08:11miraklinuxstb: output_data is a parameter that returns the decoded frame
15:08:12 Join StrathAFK [0] (
15:08:16Jungti1234__= 아 씨발 졸라 짜증나 자꾸 쌍소리만 나네!!!
15:08:20Zagoryes, too easy. it happens every time I power up (resume = yes)
15:08:22linuxstbmirak: OK, so you can use the get_buffer() function to get a pointer to data in the codec buffer, and the advance_buffer afterwards to increase the codec buffer pointer by used_bytes.
15:08:35Jungti1234__Did you see?
15:09:06linuxstbmirak: The only problem is the buffer wraparound point. Do you know the maximum size of data that the decoder will consume?
15:09:09BgerJungti1234__ only non-meaning letters
15:09:23miraklinuxstb: don't know
15:09:53miraklinuxstb: I guess it can't be bigger than a frame size
15:10:03mirakor two frames
15:10:07linuxstbmirak: You will probably need to decide on a maximum value for that.
15:10:28miraklinuxstb: I am not there yet. but any info are welcome
15:11:06miraklinuxstb: it would be usefull to have the function header in the .h file no ?
15:11:07Jungti1234__Ah motherfucker shit pair voice sounds constantly!!!
15:11:43Jungti1234__Short word: motherfucker
15:11:48linuxstbThe get_buffer() function is guaranteed to return a pointer to a contiguous buffer containing the number of bytes you request ONLY if you request less than GUARD_BUFSIZE bytes (currently 32kb)
15:12:00Jungti1234__Did you understand?
15:12:15BgerJungti1234__ yes, enough
15:12:26preglowwe should start a library of strings found in firmware
15:12:31preglowi bet there's tons of nice stuff
15:12:33linuxstbmirak: I don't understand your question. Which function(s) are you talking about?
15:12:36Jungti1234__It's bad.
15:12:47Jungti1234__Who said such word?
15:13:06BagderJungti1234: it is from the MPeye music player firmware image
15:13:12Jungti1234__I will scare him.
15:13:26miraklinuxstb: since a .c is associated to a .h, the function in the .c can have their header into the .h. sort of to have the interface
15:13:26Jungti1234__hahaha I know MPeye
15:13:34BgerJungti1234__ we don't know ... someone who has been part of the development team of MPeye
15:13:39Jungti1234__The company became dishonor.
15:13:47miraklinuxstb: rather than searching the functions into the .c files
15:13:58Jungti1234__The company was ruined.
15:14:14Bgerwhat ? are you sure ?
15:14:14miraklinuxstb: I have seen that into xvid code, that's easy to undertand and you can take only what you need
15:14:32Bgermirak afaik same is here
15:14:36Jungti1234__MPeye was ruined!!
15:14:41linuxstbmirak: Only the functions exported from the .c file have their prototypes in the .h file.
15:14:43Bgerbut there are many "private" functions
15:14:52miraklinuxstb: ah ok
15:14:53***Saving seen data "./dancer.seen"
15:15:01Bgerand they are not described in the header files
15:15:11miraknope lol
15:15:16Jungti1234__I go to bed.
15:15:18miraknot muich description
15:15:29Bgernite, Jungti1234
15:15:44Jungti1234__You don't say such word.
15:16:03Jungti1234__understand? :(
15:16:05 Quit Strath (Connection timed out)
15:16:22 Part Jungti1234__
15:17:26miraklinuxstb: can I ask you a question ?
15:17:35Bgernite = good night ...
15:17:41linuxstbmirak: Was that the question?
15:18:05Zagorcrash-casuing file (on archos rec20 v2.5):
15:18:13miraklinuxstb: I have some question about how it works.
15:18:32miraklinuxstb: for exemple, where is the starting point from the moment you click on an audio file
15:18:41Zagorno time to debug myself, unfortunately
15:18:42mirakI mean where does it start in the code
15:19:49linuxstbmirak: The playback code is in apps/playback.c - but I don't understand it. I only deal with the actual codecs themselves.
15:20:13linuxstbIf you look at the source to some of the codecs in apps/codecs/ then that may give you a better idea.
15:23:26mirakI don't understand how the code is interfaced
15:23:46 Quit g33 ("they call me snyggve because im so snygg")
15:23:53mirakI don't know generally how dynamic libraries works
15:25:25mirakthat's not easy :-/
15:27:07linuxstbThe codec is loaded from disk into memory by the playback code, and then the playback code calls the codec entry function with a pointer to a "struct codec_api". That struct contains pointers to all the "external" functions that the codec can use.
15:27:10mirakok that's all in th struct
15:30:26miraklinuxstb: ok so on e of the job is to have function that interface the codec side with the api right ?
15:31:32mirakok it's clearer now
15:31:59Bgerwhat ?
15:32:09Bgeri mean, yes ?
15:32:22mirakhem so the codec is a thread in itself ?
15:32:36Jungti1234I mean '"A sibal jola jjajungna jakku ssang soriman nane!!!"
15:32:41mirakok codec_start that's says it all
15:33:03Jungti1234bye :)
15:33:05 Quit Jungti1234 ("Bye~ -")
15:35:07Bgerhow should i interpret this reaction ? :)
15:35:24amiconnBger: yes?
15:35:51Bgeramiconn ... never mind i wanted to ask what's the interface between the lcd and cpu of h300
15:37:04Bgeris it parallel or spi ?
15:38:13 Join webguest26 [0] (
15:38:34webguest26is the bootloader for ipod out?
15:38:41preglownot exactly out
15:38:44preglowbut it works
15:38:52webguest26not out for public ?
15:39:22saa[b_r]iderit's in CVS
15:40:02preglowipod rockbox isn't exactly usable for the general public yet either
15:40:53 Join Rob2222 [0] (
15:40:56webguest26but how about colour support for rockbox h3xx ? anybody working on it ? saabrider?
15:41:02saa[b_r]iderbtw, i told a friend of mine about the iPod port, and I'm trying to convince him to help you guys out with the coding :) only he has a 3G iPod
15:41:34 Quit aliask ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
15:41:41preglowsaa[b_r]ider: hooray! we need someone to port to 3g :P
15:42:03saa[b_r]idercolor is implemented in some of the plug-ins in the H300...
15:42:26webguest26yeah i saw that in misticriver :D
15:43:01saa[b_r]iderpreglow: to be honest, I'm not an iPod advocate, but Rockbox is cool, and I'd like to see it spread :)
15:43:30 Quit Nibbler ("life is like a rental car, you fuck it up, and give it back.")
15:43:46 Join Nibbler [0] (
15:44:17webguest26yeah but i mean if we someday can have our own bitmaps on the background of rockbox?
15:44:35saa[b_r]iderany one know how colors are changed in plug-ins?
15:45:06 Quit Rob2222 (Read error: 104 (Connection reset by peer))
15:45:25saa[b_r]iderwebguest: a lot of us think it would be great to have color in the H300... but many other things are more important. just wait
15:46:02webguest26like what ?;)
15:46:22Bgerlike radio, recording ...
15:46:33Bgerfaster lcd updates ...
15:46:38Bgerbootloader USB mode ...
15:46:47saa[b_r]iderif you noticed, there's also the flickering screen, and the low battery life...
15:47:00saa[b_r]iderI'd think these things are more important ;)
15:47:24Bgerespecially the last :)
15:47:40webguest26okey i get it;)
15:47:45saa[b_r]ideras someone said, it should at least be on par with the H1x0 port, then they can start thinking of other stuff
15:48:29 Join Rob2222 [0] (
15:49:46saa[b_r]ideri'm not a programmer, so bare with me... how are .rock files edited?
15:49:58saa[b_r]iderhex editor?
15:50:23Bgersaa[b_r]ider ?
15:50:30Bgeryou usually edit the source
15:50:35Bgerand after that recompile
15:50:36 Join PaulJ [0] (
15:50:54saa[b_r]idermakes more sense :)
15:51:36saa[b_r]iderwhat would the source for pong be?
15:52:17saa[b_r]iderI swear, I must be blind :| thanks Bger :)
15:52:23Bgerhehe for nothing
15:54:43linuxstbDoes anyone actually play pong? I don't think I could persuade a friend it's a good way to pass the time...
15:55:06saa[b_r]iderprobably snake is better...
15:55:35saa[b_r]iderI found it difficult to get the bats to move quickly enough to where the ball is heading
15:55:49saa[b_r]iderespecially that you do it for both players!
15:56:12linuxstbBut you get a nice "comet" effect on the h140 lcd.
15:56:50saa[b_r]iderand you get nice colors on the H340 LCD :D
15:57:48saa[b_r]iderlinuxstb: j/k ;) "comet" you mean like a trail?
16:03:13linuxstbsaa[b_r]ider: Yes.
16:07:44 Quit webguest26 ("CGI:IRC")
16:08:07preglowi wonder what the cop does now, when an interrupt occurs
16:08:46preglowperhaps i need to enable interrupts for that too
16:10:03linuxstbZagor: Your nice illinstr.mp3 file causes the iriver to just freeze when I try and play it. Now I need a paper clip...
16:10:59preglowwhat does foobar say about it?
16:11:08linuxstbIt must be an id3v2 parsing bug.
16:11:15Zagorlinuxstb: probably, yes
16:11:36preglowi just curled it to stdio
16:11:49linuxstbUse wget :)
16:12:13linuxstb(Bagder's away, it's OK)
16:12:34preglowwho the flaming fuck even connects the pc speaker anymore?
16:12:40preglowgod, that was annoying
16:15:20linuxstbHave you had a chance to try disabling the ATA sleep command yet?
16:18:32preglowwill soon
16:18:44stevebis rockbox on the h340 functional? if i wanted to try it what should i expect?
16:19:29linuxstbYou can read the forum dedicated to rockbox on the h340 here:
16:19:38preglowif you're lucky
16:19:39linuxstbOr the ones at
16:19:47preglowoy, it loads now
16:19:49steveboh thanks
16:19:55*linuxstb forgets that he has changed his /etc/hosts
16:20:08stevebi love that
16:20:29stevebi couldnt understand why my pc was going to localhost for a website that should be on my server
16:20:33stevebuntil i remembered...
16:24:40 Quit San (Read error: 110 (Connection timed out))
16:24:53markunlinuxstb: pong on ipod would be cool. The wheel makes it more like the original game.
16:27:32linuxstbI would prefer breakout
16:27:55linuxstbHow would two players use an ipod?
16:27:57markuna car racing simulator! :)
16:29:33markunIf you could make a voip ap somehow you could use the wheel as an old fashion dial :)
16:30:26preglowlistening to some crossfeed samples now
16:30:31preglowlooking forward to see what's what
16:34:45markuncan't you tell by the volume difference?
16:35:00preglowguess i can, though it wasn't as drastic as i remember
16:35:07preglow2.wav sounds the worst of them
16:35:51preglowi think i like 3.wav best
16:38:41 Join amar [0] (
16:39:17preglowwhat was the setup again? one was the crossfeed we have now, one was the new one, and the last was...?
16:39:50markundon't know. Maybe 1 is the original?
16:40:49preglowi assume lsd1.wav is the original, yes
16:40:55preglowbut can't remember what the remaining one is
16:41:50preglowanywho, gotta dinner
16:47:20 Join lamed [0] (
16:50:51stevebare the rockbox forums down?
16:51:57stevebi cant connect :-/
16:52:17Lynx_works for me...
16:52:32stevebwhat ip is it resolving to?
16:52:38Zagorthe address has changed
16:52:44Zagorip address
16:52:53stevebto what?
16:54:01Lynx_when the h340 is i usb mode, will it use usb power?
16:56:17 Join Kohlrabi [0] (
16:57:55lamedxaviergr, can you give me t0mas's email in a private message?
17:01:45 Join Mark__ [0] (
17:07:49 Quit Febs (Read error: 110 (Connection timed out))
17:14:57***Saving seen data "./dancer.seen"
17:22:19miraklinuxstb preglow a scratch turn table on the ipod
17:22:26miraka virtual scratch turn table
17:22:46mirakszouick chou szouick
17:23:17mirakhum not preglow, markun
17:33:06 Join muesli_- [0] (
17:33:51lamedguys, can recorders do f1 + on button?
17:34:32 Join _FireFly_ [0] (
17:37:40Weazel_what you mean?
17:37:41lamedwasn't i clear with my question? - can I bind BUTTON_F1 | BUTTON_ON for the recorder?
17:40:55linuxstblamed: apps/viewer.c does it, so I guess the answer is yes.
17:41:25lamedlazy me.. tt
17:47:03preglowmirak: hah
17:47:08preglowmirak: that's a bloody _GREAT_ idea
17:48:03lamedmay i ask a db question?
17:48:38preglowjust ask
17:50:01 Quit YouCeyE ("Leaving")
17:50:26lamedwith the current db volume system, does increasing the volume by one, increase the output strength by 1 decibel?
17:54:42preglowrelative to the previous level
17:54:51lamedso what did it actually did before?
17:55:00preglowsomething more arbitrary
17:55:13preglowi'm not completely certain
17:55:19lamedcould it 'devide' a db?
17:55:29lamedwould the hardware alow that? -cool
17:55:31perplexitySo in theory.. at 0db with flat eq it should not clip.. _ever_ ?
17:55:40preglowlamed: the hardware allows finer adjustments than one db
17:55:40lamedi should clip
17:55:48preglowperplexity: correct
17:56:00preglowperplexity: the only clipping will clipping that is already in the audio
17:56:12perplexityOk then.. the natural extension of that says that at -10dB with 10dB Bass gain it should not clip ?
17:56:12preglowinsert 'be' after 'will'
17:56:25perplexityOk.. it does..
17:56:35preglowbut be aware that it may clip _before_ entering the hardware
17:56:40lamedpreglow: so, by what you said now, file audio level could not exceed 0 db right?
17:56:51perplexityand these are clean mp3's/ogg's that play nicely if I crank the vol another 3db down
17:56:57preglowlamed: no, it can't, but it will very often try
17:57:04preglowlamed: resulting in clipping in the decoder, which is a different matter
17:57:08preglowlamed: which can be fixed by replaygaimn
17:57:16 Quit einhirn (Read error: 110 (Connection timed out))
17:57:32preglowperplexity: yes, but you'd need a software gain control for that
17:57:42preglowcurrently, the only way to fix it is by using replaygain
17:57:51lamedcould a different decorder do it right - without clipping?
17:58:09preglowyou would have to 1) replaygain, or 2) gain the audio down _before_ encoding
17:58:51perplexityUmm.. ok.. hang on.. what I meant was the track in question is nicely balanced and plays with no distortion if I run it at -13dB with 10dB bass boost.. but with -10dB and 10dB bass boost it clips.. this is a software clip? Where is the eq applied on a h3x0 ?
17:58:57lamedah.. so you mean there is a limitation on the file... which will cause the waveform to be corrupted
17:59:40preglowperplexity: no, it wont clip at -10db and 10db bass boost
17:59:43preglowperplexity: it will _just_ not clip
17:59:50lamedand to a different matter: I'm in a hurry, i've made a pitch screen fix. who can accept it... fast please i'm late for a bus
17:59:51preglowperplexity: but that's only the eq we're talking about
18:00:00preglowlamed: just put in somewhere
18:00:01preglowpatch tracker
18:00:03preglowor dcc it to me
18:00:07perplexityOk.. well it is distorting in some way..
18:00:11lamedhow do i dcc?
18:00:23preglowlamed: can't you put it in the patch tracker?
18:00:51lamedpreglow: it's down. what's your
18:00:57preglowmail it to
18:01:28perplexityI guess it's possible my cassette converter is low enough impedance that it's causing the output stage on the headphone amp to clip..
18:01:36_FireFly_sf isn't really down but slow
18:03:07lamedit's sent cya alll I"m sooo late....
18:03:43preglownot everyday i receive a file named pitch.patch :P
18:03:54_FireFly_nice name
18:04:13_FireFly_it sounds like plitsch-platsch if you speak it ;)
18:04:19preglowhahah, yes
18:06:48 Join San [0] (n=sanitari@
18:08:30 Quit lamed ("CGI:IRC (Ping timeout)")
18:08:58 Join Febs [0] (
18:10:50 Join muesli- [0] (
18:11:14 Join Acksaw [0] (
18:11:15mirakhem I am a bit stuck on how to integrate the source code to rockbox
18:11:25mirakabout the make etcetera
18:11:53perplexityLook at the way the other codecs do it mirak.. it's not terribly difficult
18:12:30mirakI have some lacunes with make
18:12:47mirakI will try
18:16:21linuxstbpreglow: Are you going to commit your button driver to CVS soon, or can I have a patch? I've optimistically started work on the audio driver...
18:17:18preglowi'll work on it now
18:17:28preglowthe blind guy who reviewed us is back
18:17:31preglownow with a dead unit
18:21:52lostlogi1xperplexity: any luck with that replacement of the imdct in vorbis?
18:22:01 Nick lostlogi1x is now known as lostlogicx (
18:22:01DBUGEnqueued KICK lostlogicx
18:22:58perplexityHave got it working.. but I'm actually busy learning about the math behind mdct and fft at the moment so I can get serious about optimising.. so I'm away from the code and nose down in the books
18:23:09mirakis flac usable actually ?
18:24:10preglowmirak: usable how?
18:24:12linuxstbOf course.
18:24:22preglowthe format itself is decent, and rockbox support is pretty good
18:24:40lostlogicxperplexity: neat −− way over my head.
18:24:42mirakpreglow: usable in term of power
18:24:47preglowmirak: in term of battery, you mean?
18:24:49preglowmirak: if so: yes
18:24:51linuxstbYou can get about 12-13 hours playback with FLAC
18:24:56mirakpreglow: in term of cpu
18:24:58mirakbut seems yes
18:25:00preglowmirak: i actually believe flac will keep my player going longer than ogg
18:25:11linuxstbThat's a h140 with a standard battery and "flac -8" encoded files.
18:25:24 Quit muesli_- (Read error: 110 (Connection timed out))
18:25:30perplexityit's way over my head at the moment also lostlogicx, i've had to go back to all my math and relearn a lot of it.. don't use it you lose it, and i've not used it in 15 years
18:25:55preglowthe clever part is having interests such as these while you learn it
18:25:59mirakperplexity: I have lost it in six months
18:26:03preglowit even made me apply myself in school
18:26:09preglowwhich did not happen often
18:26:11perplexityindeed preglow :)
18:26:39perplexityI have changed my minor at university (I'm studying part time) to include some heavier math courses.. I need to sharpen the old brain again
18:27:03mirakyou are retired ?
18:27:25perplexitynot even close.. only 31
18:27:47mirakand you did fft in high school ???
18:28:02miraknot even, in schoo
18:28:04perplexityyep.. year 12 math and physics..
18:28:22mirakhuh ! where do you live ?
18:28:23preglowperplexity: it really is worthwhile if you get to apply it
18:28:33preglowmath isn't that hard, you just need to practice
18:28:35perplexityI'm living in Dubai now, but I'm Australian
18:28:52perplexityindeed preglow, and you need an interest to make you want to learn.. rockbox has done that nicely
18:29:36mirakperplexity: so how old where you exactly when you did fourier ?
18:29:56perplexityerr.. about 17
18:30:13perplexitybut it was for my final year physics project, so it's not something the whole school studied
18:31:08mirakthat's pretty much advanced indeed for 17
18:31:17perplexityI was bored :)
18:32:17perplexitymirak back at that time I only understood a little of the math, it was more application of the algorithms.. now I'm learning the math in depth
18:32:42preglowi learnt about ffts and stuff around then
18:32:45preglowbut it was on my spare time
18:32:59preglowtook me a while before i truly understood it, though
18:33:09perplexityWe had an entire 6 months free time to do our final project, so I had heaps of time to learn
18:33:11 Quit muesli- (Read error: 104 (Connection reset by peer))
18:34:10perplexityI don't claim to understand anything at the moment, but the picture is slowly coming into focus.. more time and contemplation over a scotch required yet
18:35:38 Part Polo_o
18:36:19perplexityOk, I'm crook as a dog, so gonna have an early night.. later all :)
18:36:23 Quit perplexity ("*pop*")
18:37:46 Quit Acksaw ()
18:38:34mirakhem what's the difference between shared and static library ?
18:38:46preglowshared libraries are located in separate files
18:39:02preglowso they don't occupy space in each program that uses them
18:39:40mirakin the case of rockbox it's a static library ?
18:40:10preglowno point in implementing shared libs on a platform with just one application :-)
18:40:15preglowshared libraries are bit more inefficient
18:41:00mirakhem I see no configure files in the codecs folders, what's the approach used to "configure"
18:41:33preglowthere's just one configure, and that's for all of rockbox
18:41:50mirakI have the codec source folder of xvid
18:41:55linuxstbmirak: You need to remove all that autoconf bloat
18:42:25mirakit's all in a buildir, it should be easy
18:42:38mirakhowever recreating something for it to compile ...
18:43:00linuxstbYou should probably run configure once to get a config.h file, and then manually change that config.h file for Rockbox. You should also rename it to something like xvid_config.h
18:43:22linuxstb(to avoid conflicts with Rockbox's own config.h)
18:43:47amiconnw00t! Nextup now offers ScanSoft RealSpeak voices :-)
18:44:40amiconnThese are really good quality, and are available for a lot more languages that AT&T
18:46:17 Join DrumRBoy|Away [0] (
18:49:55markunmirak: I would prefer ffmpeg over xvid because ffmpeg has asm optimisations for ARM
18:50:27mirakffmpeg is a bit big
18:52:02markunyou don't need all of it, only the parts relevant to mpeg-4
18:52:28mirakmarkun: yes but stripping it is an extra job I am not even sure to be able to do
18:53:05mirakI think that ffmpeg uses just an interface to xvidcore
18:53:57linuxstbmarkun: I'm sure I've asked this before, but does Toshiba's firmware play any video on the gigabeat?
18:54:16markunno, it doesn't
18:54:33 Quit Kohlrabi ("Leaving")
18:54:45 Join Kohlrabi [0] (
18:55:38mirakmarkun: where is the code of the codec ?
18:55:51mirakI have ffmpeg I started to look at it
18:56:03mirakthen I focused on xvidcore
18:56:29markunmirak: I'm looking at this file now:
18:56:51linuxstbDoes xvidcore just decode video frames? Or does it deal with the avi (or whatever) container and the demultiplexing?
18:58:13 Quit San (Read error: 110 (Connection timed out))
19:02:32 Quit Febs ("CGI:IRC (EOF)")
19:02:33mirakonly the xvid stream I think
19:02:52preglowbah, this takes forever
19:02:57preglowi should probably just make you a patch
19:03:21 Join Mmmm [0] (
19:03:22miraka patch for xvid ? preglow ?
19:03:27preglowtalking to linuxstb
19:03:43mirakit was a bit arch
19:04:03MmmmFirefly: you know that patch I tested yesterday?
19:04:05 Quit drumrboy (Read error: 110 (Connection timed out))
19:04:32mirakmarkun: don't know what to say, I only see an encoder
19:04:37mirakmarkun: in that file
19:04:49markunmirak: yes, most code if for encoding
19:05:22mirakmarkun: well anyway the hardest part is probably to write a xvid.c file for rockbox
19:05:38mirakI guess the codecs can be used similarly
19:06:59 Part XavierGr
19:08:44_FireFly_Mmmm: yepp
19:09:45MmmmWell... I've been using it all day and it has actually fixed my skipping ogg problem!!! :o
19:09:52mirakmarkun: xvidcore is easier to understand, I don't know about you
19:10:00_FireFly_Mmmm: really ??
19:10:42_FireFly_have you synced with the latest cvs when testing this patch ??
19:11:54MmmmYes!! Its funny how i said that all you had to do was fix that, and all along you already had! Havent synced yet.
19:12:17_FireFly_a nice sideeffect :)
19:12:56MmmmIt certainly is.. hee hee... I cant believe it!
19:13:20_FireFly_i only want to fix the image-flicker and now it seams that it also fix a skipping problem with oggs ;)
19:13:45MmmmI suppose it must have taken a bit of load off the processor eh? You're a genius...
19:14:01markunmirak: good luck, I have to go now
19:14:03_FireFly_Mmmm: yeah it seams so
19:14:35_FireFly_if i had cvs-write rights i think i would commit it :)
19:14:58***Saving seen data "./dancer.seen"
19:15:14MmmmIt should be done! Noone else seemed to have the same problem as me though..
19:16:01_FireFly_now it seams this patch has 3 advantages 1. it solves the image-flicker-problem(at least for the h1xx remote-lcd) 2. reduces the code-size about 84 Bytes 3. it solves a skipping-problem with ogg-files
19:16:53MmmmYou would think that any self respecting dev would commit something like that in a shot eh? ;)
19:17:44 Join Acksaw [0] (
19:17:56Acksawis muesli here?
19:18:22_FireFly_currently not or he has changes is nick in the chan
19:19:47MmmmAlthough, firefly, I must admit that i hadnt tested before i applied that patch, It could have been the previous flicker commit!
19:20:15MmmmSo I will test that now....
19:20:37_FireFly_Mmmm: this patch is the simplified version of the two patcher
19:21:50_FireFly_the first one had solved the flicker-problem but introduce a louder remote-ticking-noise on "infected" h1xx-devices
19:22:17MmmmYes.. What I mean is that the unsimplified version may have already fixed the skipping problem.
19:22:20_FireFly_the second one had solved the introduces louder ticking-noice from the first patch
19:22:43_FireFly_Mmmm: could be
19:23:02 Join rh [0] (
19:23:08preglowlinuxstb: you doing a proper audio driver?
19:23:26Mmmm3 patches yes? 1st increased ticking, 2nd is in cvs 3'rd i tested. i'm talking about 2'nd :)
19:24:07Acksawanyone here good with pc hardware.
19:24:13_FireFly_1st gots commited then the louder ticking was discovered then i made the second one which gots also commited
19:24:43MmmmThat's what i thought :D
19:24:51 Quit Cassandra (" HydraIRC -> <- Go on, try it!")
19:24:58Mmmmhee hee.. do you know what i mean?
19:27:02rhanybody here who knows the status of the fm-radion on h3xx?
19:30:33 Join linuxstb_ [0] (
19:30:46linuxstb_preglow: There's no point doing anything else.
19:33:25preglowwell, you were talking about a dummy driver
19:34:34linuxstb_I'll probably get to that stage first.
19:36:32 Quit rh ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
19:38:16preglowahh, sixth day of non-stop rain
19:38:37preglowat least i can stay inside without any second thoughts
19:38:39_FireFly_what ?? really 60 days
19:38:52MmmmFirefly: ok, tested and it seems that the version already in cvs (patch2) has fixed the problem...and that's not all. I was having difficulty seeking in oggs too and that is also fixed!
19:38:54preglowsixth <-
19:39:05_FireFly_preglow: yeah
19:40:31_FireFly_preglow: i have sixth mixedup with sixty
19:40:39preglowso i see
19:40:53_FireFly_sometime i*m reading to fast
19:42:30_FireFly_Mmmm: also with the simplified version ??
19:43:32MmmmYep, simplified version from a users point of view is identical.
19:43:50_FireFly_good to hear :)
19:46:46 Join Strath [0] (
19:46:47MmmmI've gotta go now...tons of thanks firefly... Now, how can I add to the playlist on the fly with the remote... hmmmm.. :D
19:47:27_FireFly_Mmmm: long press play on the remote is the same as long press select(joystick)
19:47:48 Quit StrathAFK (Read error: 104 (Connection reset by peer))
19:47:54MmmmWOW... heh heh...silly me! cheers ;)
19:48:48 Quit Mmmm ("DINNER TIME!!")
19:49:33 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
19:52:06 Join RotAtoR [0] (
19:52:06 Quit zeero (Read error: 110 (Connection timed out))
19:55:51preglowRotAtoR: yo, plase work on prettier gems in bejeweled :-)
19:56:09preglowRotAtoR: gweled has svg format graphics, so should be a breeze to make custom size jewel bitmaps
19:56:56Moosreally good idea
19:59:46wubblaanyone using mp3gain?
20:00:03wubblawould you guys recommend to use this tool?
20:00:13Moossorry, foobar2k here
20:02:01 Join petur [0] (
20:05:44PaulJwubbla: do you want to change the volume of the mp3s directly or do you only want to add rplaygaintags?
20:07:41wubblaPaulJ: i'd prefer replaygain tags (as it doesn't change the files and it's reversible AFAIK)...
20:08:17RotAtoRpreglow: I'm looking at that right now :)
20:08:19wubblaand as rockbox does support replaygain, there shouldn't be any point in changing the mp3s directly, right?
20:08:45RotAtoRand looking at getting bejeweled working well on that *&#%$ ipod click wheel
20:09:06PaulJthen i would suggest to use foobar2000, because mp3gain adds the replaygaininfo to apev2 tags and rockbox wouldnt recognize them
20:09:08RotAtoRjust finished exams, so now I have a little time :D
20:09:48wubblaPaulJ: you're sure about that?
20:10:47MoosPaulJ: with last version of mp3gain, it add tag to APEv2, and doesn't suported yet for mp3's in Rockbox
20:10:55Moosoops wubbla :)
20:11:16PaulJi'm using version 1.2.4
20:11:37PaulJthe helpfile says "MP3Gain stores "Analysis" and "Undo" information in special tags inside the mp3 file itself. These tags are in the APEv2 format."
20:12:11Moosyeah and don't apev2 tag for mp3 yet
20:12:18wubblaPaulJ: ok, that's an argument :-)
20:12:19Moosin Rockbox
20:13:09wubblaPaulJ: then i'll look into foobar2000's replaygain support...
20:15:02wubblaPaulJ: can you point me to a howto?
20:15:27 Quit DJDD_ (Read error: 104 (Connection reset by peer))
20:16:10RotAtoRhmm, does anyone know of a good tool to convert an svg graphic into a raster graphic in windows?
20:19:11wubbladpassen1: wow! thanks!
20:22:31 Quit Acksaw (Read error: 110 (Connection timed out))
20:29:46 Join AnInternetUser [0] (
20:30:27AnInternetUserhello can anyone explain exactly what ata error 11 is?
20:30:55 Quit Rob2222 (Read error: 104 (Connection reset by peer))
20:35:45preglowlinuxstb_: if i'm stalling you, i can make you a patch
20:36:00preglowmight not finish it today either
20:36:23preglowi've broken it now, but i kept the old tree around so i can make you a patch
20:37:25 Join xmixahlx [0] (n=xmixahlx@
20:37:59preglowi never made a patch from cvs before, is it just a matter of doing a cvs diff ?
20:38:11AnInternetUsercan anyone explain the difference between ata error 11 and 32?
20:38:21_FireFly_preglow: cvs diff -u > file :)
20:38:23_FireFly_-u for unified diff
20:38:31preglowi've got -u as default
20:40:24 Join San [0] (n=sanitari@
20:41:45RotAtoRhmm, it looks like rgb565 bitmaps are byte swapped between iriver and ipod?
20:41:47linuxstb_preglow: Thanks.
20:41:56linuxstb_RotAtoR: Yes :)
20:41:59preglowjust keep the patch around so you can undo it
20:42:07RotAtoRso is there a good way to determine which to use?
20:42:17RotAtoRin code, that is
20:42:25RotAtoRahh, ok, thanks
20:42:34_FireFly_does it define RGB or GBR ??
20:42:35linuxstb_RGB565 and RGB565_BYTESWAPPED (I think - check lcd.h)
20:42:43_FireFly_or so :)
20:42:47RotAtoRthanks :)
20:42:57linuxstb_It's not GBR
20:43:06preglowBGR, yes?
20:43:18_FireFly_ups yepp
20:43:21amiconnNot BGR either
20:43:30_FireFly_whats then ??
20:43:38linuxstb_RGB565 byte swapped
20:43:44amiconnIt's more like G3B5R5G3
20:44:02preglowthe purdiest format in the world
20:44:06RotAtoRhaha, ;)
20:44:12preglowmakes jesus cry
20:44:21 Quit AnInternetUser ()
20:44:23_FireFly_amiconn: ah right it isn't RGB888
20:44:25 Join hshah [0] (
20:44:27linuxstb_All to save us some byte-swapping when updating the LCD.
20:44:40preglowi say it's worth it
20:44:55preglowesp since we don't have a byte swap instruction
20:45:04hshahis it just me or have the forums been down for a few days?
20:45:13linuxstb_I agree - it makes RotAtoR's life harder though.
20:45:15preglowhshah: they're up now, depending on your is
20:45:15_FireFly_hshah: it was shorly up
20:45:40peturAnybody around that knows what read_hw_mask() does?
20:45:40hshahok - thanks
20:46:02peturexcept mask = *(short *)0x020000fc;
20:46:05hshahnow... can anyone commit my wps for me - TiMiD was going to do it... but hes like disappeared off the face of the Earth...
20:46:20amiconnpetur: read_hw_mask() is archos only
20:47:30RotAtoRall these various bitmap types and screen sizes are making the bejeweled code almost 50% different bitmap definitions
20:47:45amiconnIt reads a certain byte in ROM that tells us about subtle hardware differences
20:47:52preglowcan't you just use a macro per value to make the compiler bitswap for you?
20:47:55peturgot to read that code that inits the radio 3 times - it's not at all clear what it tries to do to the hardware :(
20:48:35peturoh... it skips it completely ;)
20:48:38linuxstb_preglow: I was thinking that - it would mean adding a new output format to bmp2rb (or merging the two existing 16-bit formats)
20:48:48RotAtoRpreglow: hmm, i'm not sure how i'd go about doing that
20:48:54preglowlike '#define _(x) dothebitswap(x) \n short mycolourtable[] = { _(0xvalue1), _(0xvalue2), ... }
20:49:35preglowwhat would also work is just building the bitmap headers at compile time
20:49:43preglowfrom bitmaps
20:49:55 Join Lear [0] (
20:49:57Bagderthat's actually a rather good idea
20:49:57CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
20:49:57*linuxstb_ prods Bagder
20:50:09RotAtoRpreglow: too fancy for me :)
20:50:24Bagdersince the lcd variations are gonna add up even more in the future
20:50:37preglowyou'll still have various bitmaps, though
20:50:40preglowsince bitmap size will change
20:50:52preglowbut yeah, colour wise we'll be fine
20:50:59linuxstb_We will hopefully get another 2-bit greyscale format as well soon with the greyscale ipods
20:51:26preglowbut stuff like differing formats for the same dimension and rough display type will work
20:52:20linuxstb_But does anyone still have the original bitmaps that have been used everywhere? It would mean generating the bmp files for the existing bitmaps
20:52:27preglowand if implemented properly in the build system, it'll be prettier than what we do now anyway
20:52:55preglowlinuxstb_: if not, it's not worse than just making new bmps from the headers
20:54:22wubblareplaygain doesn't seem to work for me.... :'-(
20:54:33*preglow observes that button_tick() seems to take care of REPEAT
20:54:45preglowwubbla: have you enabled it?
20:55:12wubblapreglow: i enabled it in the settings menu
20:55:19preglowgood, then what format do you use?
20:55:25wubblapreglow: mp3
20:55:31preglowdo we support replaygain for mp3?
20:55:40wubblapreglow: i think so...?!
20:55:50dpassen1preglow: yes, if written in an ID3v2 tag
20:55:55preglowwell, is it?
20:56:02wubblai hope so
20:56:10 Join PcHeRo [0] (
20:56:16dpassen1does replaygain info appear in Show ID3 Info
20:56:43PcHeRoanybody know something about doom on h120??
20:56:53preglowPcHeRo: i know it doesn't exist
20:56:53wubbladpassen1: track/album gain is there
20:57:02amiconnpreglow: Yes it does (for ages ;) )
20:57:02preglowthen it should work
20:57:25preglowthat's archos!
20:57:28markunWe could just add the APEv2 tag support, very easy, but somehow I think some people in here will not like that.
20:57:47dpassen1markun: with unicode supported, it seems that APEv2 tag support should be added
20:57:53_FireFly_markun: why ?
20:57:54preglowmarkun: well, it screws up the tag priority
20:58:07_FireFly_ah right
20:58:36_FireFly_an comprimise would be to only add apev2 support for replay-gain
20:58:40preglowi like apev2, but i think we should not encourage use of it with mp3
20:58:46_FireFly_for mp3-files
20:58:47wubblaisn't the pre-amp option used to increase the perceived "volume"?
20:58:54markunIf people add apev2 tags they should have the highest priority I think
20:58:55preglowwubbla: well, yes...
20:59:11wubblapreglow: this setting doesn't work for me at all
20:59:17preglowwubbla: then replaygain is not enabled
20:59:22dpassen1i'm not even sure why there is an option for ID3 tag priority to be honest
20:59:28wubblapreglow: 0dB and 12dB gives no difference...
20:59:30preglowwubbla: either because you haven't enabled the option, or because rockbox does not use it
20:59:31PcHeRois it difficult to write the archos code of doom to the h120 code
20:59:39preglowPcHeRo: it's a hoax...
20:59:53preglowit even says so on the page
20:59:58preglowThis is an April Fools joke.
21:00:09preglowon a red background
21:00:18wubblamaybe it's the "prevent clipping" setting that makes problems?
21:00:22wubblano idea
21:00:45wubblamaybe a reboot helps :-D
21:00:47_FireFly_wubbla: if you don't know in the latest build prevent clipping is gone
21:01:00dpassen1there is a different prevent clipping in the replaygain options menu
21:01:02PcHeRoohh lol
21:01:11markun_FireFly_: not for replaygain
21:01:19_FireFly_markun: yepp i know
21:01:20 Quit linuxstb_ ("CGI:IRC")
21:01:32_FireFly_but the other is gone :)
21:01:46wubblareplaygain type is set to album gain
21:01:56preglowwubbla: so track and album gain is displayed in id3 info?
21:02:00preglowwubbla: what values do you see?
21:02:03wubblapreglow: yes
21:02:24 Quit San (Read error: 110 (Connection timed out))
21:02:41wubblapreglow: track: -0.58 / album: -0.44
21:04:08wubblaall i want is to increase volume, as i simply cannot hear anything at all at -40dB volume setting...
21:04:21dpassen1so raise the volume, -40 is very very low
21:04:30preglowif you want to _increase_ volume, replaygain is not the way to go
21:04:35preglowreplaygain will most often decrease it
21:04:42preglowunless you manually use the preamp
21:04:53wubblapreglow: i intended to use the pre-amp option
21:05:06wubblaso that all my music plays at the same volume level...
21:05:48*preglow summons Lear
21:05:53preglowLear: there you are!
21:06:05preglowso, what's up with this mp3 replaygain thing?
21:06:14preglowLear: and how are the tremor opts going?
21:06:17Learwhat about it? I use it all the time...
21:06:27preglowwubbla says he can't get it to work
21:06:34pregloweven though track and album gain are parsed
21:06:45wubblaLear: basically the pre-amp options doesn't seem to work for me...
21:06:48_FireFly_wubbla: why do you have the volume to a such low level (-40 db) ??
21:06:50PaulJon my h320 replaygain works with mp3, but i have never used the preamp option
21:06:56preglowif preamp doesn't work, that means replgaygain doesn't work
21:07:08Learpreglow: For the small thing I tested, I hadn't checked the which cases were executed most often, so my code wasn't that fast...
21:07:21wubblai might add that i'm using a h320...
21:07:33LearBut I think I can remove 1-2 percentage points of boost... :)
21:07:55LearI don't use pre-amp myself, so let me check that...
21:08:00amiconn[20:57:17] <Bagder> <== At least the archos can show the intro video of Doom 3, as .rvf :-P
21:08:21Bagderand a neat video it is
21:08:21amiconn(That was not yet possible when this joke was made)
21:08:30dpassen1to test replaygain settings, do you need to restart playback? reboot? or should it take place when the buffer is refilled?
21:08:36Bagderit would have made a more realistic joke
21:08:41LearNope, nothing of the sort.
21:08:56wubblai've already rebooted...
21:08:58Leardpassen1: you only need to exit the menu..
21:09:07dpassen1so it should take place immediately
21:09:41LearThere's a small delay due to buffering, so after exiting one level in the menu, it can take a few seconds.
21:10:09 Join solexx__ [0] (
21:10:41*RotAtoR puts on evil music and begins working with 100's of lines of code of bitmap definitions >:D
21:10:46wubblaLear: so the pre-amp does work for you?
21:11:07 Join Rob2222 [0] (
21:11:13preglowlinuxstb: does the patch apply? i've got to go real soon
21:11:43 Join Mmmm [0] (
21:11:56_FireFly_btw what about my simplified version of the image-flicker-patch
21:11:57PaulJpreamp seems to work on my h320
21:11:59Learwubbla: yes, pre-amp works fine.
21:12:17linuxstbpreglow: Yes, it applied perfectly. I can't compile at the moment because of my audio changes though.
21:12:54wubblahum. :'-(
21:13:26_FireFly_wubbla: maybe you could upload one of your mp3's so that other can test it
21:13:29*amiconn has an idea how to solve the delay problem for EQ and other DSP effects
21:13:48_FireFly_to check if it is maybe an problem with the files
21:13:50 Quit Rob2222 (Read error: 104 (Connection reset by peer))
21:13:51PaulJwubbla: what value did you set the preamp
21:14:23wubblaPaulJ: even 12dB didn't change anythin...
21:14:34*Lear wonders what amiconn is smoking this time ;)
21:14:43PaulJok, this should be audible
21:14:48wubbla_FireFly_: will do that...
21:15:00***Saving seen data "./dancer.seen"
21:16:06preglowamiconn: spill it!
21:17:10amiconnWell, the large delays will only occur with crossfade enabled, right? Without crossfade, the PCM buffer is only half a second of audio
21:17:21preglowit's far too long without crossfade as well
21:17:34preglowlonger than half a second
21:17:36amiconnSo we could do the following with sufficiently fast codecs:
21:18:02Learwubbla: pre-amp is only applied for files with replaygain and if replaygain is enabled.
21:18:30wubblaLear: yes, i'm aware of that
21:18:31amiconn- Keep the buffer level roughly at half a second during the track. Only fill it completely a short time before track change
21:18:41linuxstbpreglow: Your patch runs fine, thanks.
21:19:23preglowamiconn: i don't like it, the cpu spike protection vanishes
21:19:23 Quit PcHeRo ("CGI:IRC (EOF)")
21:19:58preglowi don't like the inconsistency either
21:20:08preglowpeople will wonder why the hell changes are fast some times and not other times
21:20:41preglowamiconn: and what if someone skips a track=?
21:20:52preglowamiconn: the entire buffer will have to be filled before rockbox can comply
21:21:13amiconnHmm, does crossfade also work when skipping tracks?
21:21:23amiconnI didn't think it does
21:21:29preglowi'm not sure
21:21:32preglowi never use crossfade
21:21:32*amiconn never used crossfade
21:21:32 Quit solexx_ (Read error: 110 (Connection timed out))
21:22:45preglowi think the voice ui shouldn't interact with audio buffering like it does now
21:22:56preglowit should probably just mix it's changes right at the beginning of the buffer
21:23:27amiconnvoice UI needs really low latency
21:23:27preglowif you do a long string of commands now, all of them spoken in a row
21:23:33preglowpretty unusable
21:23:40preglowbut ok
21:23:42preglowi really need to run
21:25:01amiconnOn archos, switching to a new menu item preempts the old clip and immediately plays the new one. This doesn't work on iriver yet
21:25:14 Join zeero [0] (
21:25:54mirakif I want a static lib out of a codec how do I proceed ?
21:25:55wubblamaybe i should also paste my rockbox config-file?
21:26:23wubblaLear: please check my config...
21:26:46_FireFly_wubbla: if its an config problem then try to reset the settings and enable only replay-gain
21:27:12mirakI don't understand the compile chain
21:27:58mirakI have run a configure with a generic target to see how look like the configurated sources. Only a file called gets created
21:28:32miraknow from that file I get defines that I should set in a .h file right ?
21:28:51linuxstbAre you talking about the xvidcore build system, or Rockbox?
21:29:00Learwubbla: looks good. Volume is low, but not that low (depending on headphones...).
21:29:06mirakxvid part
21:29:24miraklinuxstb: in each codec there is a make file. I adapted that make file to xvid
21:29:35 Join Rob2222 [0] (
21:29:37wubblaLear: this is a mp3 with added replaygain tags:
21:29:48 Join Benacool [0] (
21:29:49miraknow I don't get how exactly the xvid sources are build
21:30:13miraklinuxstb: the static lib is created from the rockbox interface file should be xvid.c
21:30:15linuxstbThe first thing I would do is to take the xvidcore library, and one of the example programs, and try to compile it as a standalone program outside of both Rockbox and the xvidcore build system.
21:30:24 Join nathanh [0] (
21:30:28miraklinuxstb: I did that
21:30:39mirakbut it failed
21:30:57mirakI will focus on that
21:31:01linuxstb:) Do you mean you tried to do that but failed?
21:31:44mirakin fact I did something like gcc -I path_to_xvidlib exemple.c
21:31:52mirakand got loads of errors
21:31:55mirakthen I went to bed
21:32:04mirakit was yesterday I think
21:32:35dpassen1wubbla: mp3 is fine
21:35:18wubbladpassen1: pre-amp does work for you?
21:35:45dpassen1ill check, but im on a 120
21:35:53miraklinuxstb: there is a problem with the width variable
21:35:56MmmmDoes anyone know the DNS for the forums? I think i'm going to edit my hosts file, its just too painful otherwise!
21:36:01mirakit's not declared
21:36:17PaulJwubbla: did you go back to the wps after changing the preamp?
21:36:27wubblaPaulJ: yes
21:36:47 Join saratoga [0] (
21:37:38linuxstbmirak: What width variable are you talking about?
21:37:59miraklinuxstb: you have the code ?
21:38:23miraklinuxstb: line 577 in xvid_decraw.c
21:38:36dpassen1replaygain and the preamp definitely work here
21:38:52linuxstbmirak: No, I don't have the code, sorry.
21:38:56miraklinuxstb: it's not declared it must be something else
21:39:31_FireFly_mirak: do you have some errors which say that he can't find some *.h files ??
21:42:12mirakno I fixed the width error, I think they didn't used the good variable
21:42:24mirakthat's strange
21:42:25MmmmAhh tis ok, found the DNS:
21:42:54miraknow I got undefined reference to some methods
21:43:04wubblagrrr.... :-/
21:43:23_FireFly_mirak: do you have given the compile which libs you should link ??
21:43:24mirakmy build like is just gcc -I ../src/ xvid_decraw.c with ../src/ pointing to the codec source
21:43:33mirak_FireFly_: I am clueless for that
21:43:54wubblais there any h300 user able to confirm that the replaygain's pre-amp option does work?
21:43:58mirak_FireFly_: I said him what to include, I am not sure it must be done like that
21:44:02dpassen1wubbla: test with a file with a more pronounced replaygain
21:44:19wubbladpassen1: what do you mean?
21:44:27PaulJwubbla: on my h320 the preamp works with your file too
21:44:45dpassen1nevermind, i thought you were testing with a -0.xx replaygained file
21:44:46wubblaPaulJ: you're using latest CVS?
21:44:53mirak_FireFly_: my goal is to use to not use external libs
21:45:00mirakto not use
21:45:08linuxstbmirak: Have you written your own Makefile?
21:45:12_FireFly_mirak: i think you should learn how to use libs (static or dynamic)
21:45:19miraklinuxstb: not yet, I just try to build the exemple
21:45:44_FireFly_mirak: i think the example-code from the xvid-core needs the lib
21:46:18MmmmWubbla: do you have prevent clipping on? If so have you tried turning it off?
21:46:43_FireFly_i think he should also try a newer build
21:46:43PaulJwubbla: i'm using a version from yesterday. i have applied the brightness patch, but this shoult not affect replaygain
21:47:05wubblaPaulJ: can you upload your rockbox.iriver as well as the .rockbox folder (in a zip/rar/whatelse archive preferably) to
21:47:08_FireFly_because the overall prevent-clipping option is gone in the newer builds
21:47:22mirak_FireFly_: so I need xvid-core-dev or something like that with the headers ?
21:47:35_FireFly_mirak: i think do you have the sources
21:47:43mirakI have the sources yes
21:47:55_FireFly_then compile it which will generate the lib
21:48:14mirak-do ?
21:48:25mirakoh sorry :)
21:48:25_FireFly_and then with this lib and the headers you can compile the example-programm
21:48:35mirakI tough it was gcc option ;)
21:48:44_FireFly_gcc is used :)
21:49:02_FireFly_but i think the sources you have will generate a lib which can be used in other programms
21:49:18 Join Cassandra [0] (
21:49:47mirak_FireFly_: but why can't I just build a static binary with the lib just as includes ?
21:50:47nathanhlib as includes?
21:51:04_FireFly_mirak: you can but you must either generate a static-lib or include the sources to you build
21:51:21mirak_FireFly_: I want to do that
21:51:26_FireFly_mirak: i think you should realy learn how to use libs
21:51:32MmmmWubbla: if you have prevent clippping on (in the replaygain menu) and your track is clipping, then I think the preamp won't do anything.
21:52:03wubblaPaulJ: thank you so much!
21:52:14nathanhmirak: read this -> <- its an awesome book, free copy online, that will teach you all you need to do so as to do what you want
21:52:26_FireFly_mirak: normaly there is a configure-script(when under linux/*nix) in the sources
21:52:28mirakthank you nathanh
21:52:29PaulJwubbla: the colours are also changed
21:52:33nathanhi bought the book because i liked it so much
21:56:24mirakwhat I don't get is why from the moment it includes the main header file of the codec, why doesn't it find all the other methods
21:56:48mirakxvid.c file reference a bunch of .h files
21:57:02_FireFly_mirak: because you need also the compiled code of the functions
21:57:08mirakhow to just make the exemple programme be aware of that ?
21:57:29_FireFly_mirak: is there a configure-script in the sources
21:57:40mirak_FireFly_: yes
21:57:45_FireFly_then run it
21:57:47wubblaPaulJ: with your build pre-amp works!
21:57:56wubblaPaulJ: thanks!
21:58:03_FireFly_wubbla: then try a newer build :)
21:58:25_FireFly_i think the problem was that he had used a old one
21:58:26mirak_FireFly_: I must also put xvid.c on the gcc line somethig like that ?
21:58:30wubbla_FireFly_: i compiled the latest cvs version on my own
21:58:31_FireFly_a old build
21:58:40Maximegreat RotAtoR
21:58:52nathanhrot: nice
21:59:01_FireFly_mirak: the configure-script will create a makefile
21:59:12_FireFly_after the script has ended simply run make
21:59:30_FireFly_then i guess you will get a lib
22:00:23nathanhxvid lib uses malloc, its not going to work without major mods
22:00:37nathanhits also a huge memory hog
22:00:51_FireFly_nathanh: as many codes for non embedded systems
22:01:25_FireFly_does dynamic memory allocation
22:02:28_FireFly_wubbla: if you had run the latest build then why did you had an prevent-clipping option outside the replaygain-setting menu ??
22:02:44_FireFly_or i miss something
22:02:59wubbla_FireFly_: i used the clipping option in the replaygain menu...
22:03:48_FireFly_wubbla: you had said something different
22:03:58mirak_FireFly_: I have built it
22:04:02mirakthe lib
22:04:23miraknathanh: yes we are aware of that
22:04:48miraknathanh: I am at least. the succes of this operation is very uncertain
22:05:14mirakbut I am learning new stuffs, that's fine
22:05:25wubblaReplaygain scan status (9/12749)
22:05:27_FireFly_mirak: yeah many new stuff :)
22:05:31wubblathat might take a while :-D
22:05:36CassandraTheoretically there's no reason why you couldn't include a malloc implementation within a plugin.
22:05:57mirak_FireFly_: ok so as far as I understand, now ld needs to know that libxvid exist to be able to feed the exemple with it ?
22:06:31mirakCassandra: I have read the code, and we could put the buffer static I think we will see
22:06:35_FireFly_the gcc option for this is -l(lower L) and the name of the lib without the lib-prefix and the .so/.a suffix
22:07:00_FireFly_and with -L you give the search patch
22:07:03_FireFly_for libs
22:07:22mirakok so with -L I am not obliged to install the lib
22:07:56_FireFly_gcc -I ../src/ -L ../src -lxvid xvid_decraw.c
22:08:21_FireFly_maybe the libxvid is in a subdir of ../src
22:08:31_FireFly_then you need to change it accordingly
22:08:31DrumRBoy|Awaythis is so cool... i didnt realize USB was supported in RB now... awesome
22:08:45_FireFly_mirak: yepp
22:08:47mirak_FireFly_: yep it's in the builddir
22:08:58_FireFly_mirak: is the lib an *.so or *.a file ??
22:10:23_FireFly_hmm then you have to install the lib your create an run-script which adds the path to the lib-loder-search path or compile the lib as a static one (*.a)
22:10:32_FireFly_the last would i prefer
22:11:26amiconnThere is no shared library support in rockbox, so static lib is the only option
22:11:27_FireFly_is in the output from configure −−help any line which has static lib(or similar)
22:11:30mirakbut with that, I will able to run the exemple even without the xvidlib installed ?
22:11:53_FireFly_mirak: yepp because static libs are linked into the programm
22:12:12_FireFly_-> the size grows is bigger then when using dynamic libs
22:12:58_FireFly_mirak: it seams that you have to tell configure to make static-libs instead of dynamic
22:13:16_FireFly_there is normaly an option/parameter for this
22:14:25_FireFly_amiconn: have you yet found some time for the remote-setting patch ??
22:15:25peturanybody familiar with the soft-I2C written for iRiver?
22:15:37mirak_FireFly_: can't find that option
22:16:04_FireFly_can you post somewhere the output of configure −−help ??
22:16:13 Quit nathanh ("Quit")
22:17:25peturhmmmm... If somebody told me the radio on the H1xx works, I wouldn't believe him...
22:17:36_FireFly_petur: why 횩?
22:18:14peturthe soft I2C getack() looks wrong... and that's where it hangs on H3xx
22:18:27peturBut it's the same code as the H1xx
22:18:38peturSo I would expect it to be ok
22:18:49_FireFly_petur: really ?
22:18:54 Quit ender` (" cd /pub && get beer")
22:19:02_FireFly_maybe there is something different
22:19:09 Quit Rob2222 (Read error: 104 (Connection reset by peer))
22:20:06peturgetack does the same wait loop as ack, meaning it's generating SCL
22:20:21peturthe other soft I2C implementations don't do that at all...
22:20:54_FireFly_petur: maybe the hardware differs in that case
22:21:24_FireFly_so that the code for the h1xx doesn't use the right output/input-pins
22:23:35petur_FireFly_: you mean the radio is connected to other pins? I already changed that, unless Linus measured it wrong (but he double-checked)
22:24:03peturI'm gonna change it and see what it brings :D
22:24:18*petur hopes it brings radio
22:24:59_FireFly_mirak: try configure −−enable-static
22:25:17 Join jlo [0] (
22:25:27jlohello all
22:25:44_FireFly_petur: if LinusN say that the pins are the same then it that should right
22:26:12_FireFly_jlo: hi
22:26:37petur_FireFly_: he said the pins mentioned in the wiki are correct - they differ from the H1xx
22:26:37 Join ender` [0] (i=ychat@
22:27:28_FireFly_petur: or so but maybe the pins must be interchanged
22:27:44mirak_FireFly_: ok I have the .a file
22:27:59_FireFly_also the .so file ??
22:28:12mirakyes both
22:28:26_FireFly_if so then delete the .so file or move it
22:28:33amiconnpetur: The comments tells you why the loop is repeatedly setting SCL high and switching to input again
22:28:40_FireFly_because the linker will use the .so file if he find it
22:28:56mirakok then ?
22:28:59 Quit hshah (Read error: 110 (Connection timed out))
22:29:25_FireFly_mirak: then this should work : gcc -I ../src/ -L ../src -lxvid xvid_decraw.c
22:29:36*petur learns to read comments
22:29:38_FireFly_amiconn: what about my question ??
22:30:29_FireFly_mirak: only the path for the -L parameter must you set it to the right parth in which the .a file is
22:31:12peturwell, I'm stuck there: it never sees SCL as high :(
22:32:11mirakgcc -I ../src/ -L ../build/generic/\=build/ -lxvidcore xvid_decraw.c
22:32:38mirakin the make file he uses -lxvidcore
22:32:42peturthis is bad
22:33:05mirak /tmp/ccQfGPHD.o: dans la fonction 짬 dec_init 쨩:
22:33:06mirakxvid_decraw.c:(.text+0x13a9): r챕f챕rence ind챕finie vers 짬 xvid_global 쨩
22:33:06_FireFly_mirak: what is the name of the .a file ??
22:33:26peturno it isn't (wrong pin description) - nothing happened...
22:33:43_FireFly_mirak: then it should be -lxvidcore
22:34:01mirakI use that it fails with the above error
22:34:16_FireFly_mirak: i didn't understadn frensh ;)
22:34:32mirakr챕f챕rence ind챕finie vers −−-> undefined reference to
22:35:08_FireFly_then something else is missing
22:36:08_FireFly_mirak: do you find an function-definition or a var-definition with the name xvid_global
22:36:16_FireFly_in the sources do you have
22:36:38_FireFly_amiconn: ??
22:36:48amiconnNo, sorry :(
22:37:18_FireFly_also no comments in my gentoo-dev-ml-thread
22:37:36_FireFly_nobody has any comments about this it seams :)
22:37:40_FireFly_or no time
22:39:10mirakextern int xvid_global(void *handle, int opt, void *param1, void *param2);
22:39:15mirakit's in the code
22:39:42miraklib is not loaded I get the same error when not using -L
22:39:54 Join nathanh [0] (
22:40:02jlois somebody working on spdif recording (H1xx) ? : recording.c seems not far from
22:40:29_FireFly_mirak: does he say that he can't find the lib ??
22:40:43mirakI past you the output
22:43:02mirakLANG=en grep -n xvid_global libxvidcore.a
22:43:02mirakBinary file libxvidcore.a matches
22:43:09mirakit's in it ^^
22:43:41_FireFly_mirak: yeah but the gcc doesn't try to link it or doesn't say that i can't find the lib
22:44:10_FireFly_try this -llibxvidcore.a
22:44:51 Join San [0] (
22:44:56peturright... this I2C thing calls for measurement on the signals. I'll mail Linus
22:45:06mirak/usr/bin/ld: cannot find -llibxvidcore.a
22:45:20_FireFly_have you set the -L parameter ??
22:45:46mirakLANG=en gcc -I ../src/ -L ../build/generic/\=build/ -llibxvidcore.a xvid_decraw.c
22:45:47 Join DangerousDan [0] (
22:45:56mirakthat's the command
22:46:23_FireFly_maybe he has problems that the last subdirdir starts with an '='
22:46:48_FireFly_try to copy it to the same location as the source-file and set -L ./
22:52:18 Quit Slasheri (
22:56:34 Quit Mmmm ()
22:57:58 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
22:58:22_FireFly_mirak: do you have xvid installed in your system ??
22:59:02_FireFly_i think this defines are also needed: -DARCH_IS_IA32 -DARCH_IS_32BIT -DARCH_IS_LITTLE_ENDIAN -fPIC
22:59:29_FireFly_these are from the Makefile-output in the example-dir
22:59:54_FireFly_then it should at least compile then -lxvidcore
23:00:08_FireFly_with -lxvidcore
23:00:17mirakI use this defines where ?
23:00:21NJoinSlasheri [0] (
23:00:41_FireFly_similar to the -l -L parameters
23:00:54_FireFly_gcc -DARCH_IS_IA32 -DARCH_IS_32BIT -DARCH_IS_LITTLE_ENDIAN -fPIC ......
23:01:55mirakstill not
23:02:04mirak LANG=en gcc -I ../src/ -L ./ -lxvidcore -DARCH_IS_IA32 -DARCH_IS_32BIT -DARCH_IS_LITTLE_ENDIAN -fPIC xvid_decraw.c
23:02:07mirakI about to give up
23:02:32_FireFly_-lc -lm are missing
23:02:57 Join Midgey34 [0] (
23:03:16mirakLANG=en gcc -I ../src/ -L ./ -lxvidcore -lc -lm -DARCH_IS_IA32 -DARCH_IS_32BIT -DARCH_IS_LITTLE_ENDIAN -fPIC xvid_decraw.c
23:03:24mirakbouhhhh :'(
23:03:30_FireFly_hmm strange
23:03:33Bagderthe -ls are only needed when linking, not mere compiling
23:04:08_FireFly_Bagder: yepp but he compiles and links with one cmd-line
23:04:19Bagderah, right
23:04:22Bagderno -c
23:04:42*Bagder shuts up
23:04:48_FireFly_mirak: have you tried to run make in the example dir ??
23:04:53_FireFly_does this works ??
23:05:02mirakit didn't worked
23:05:15mirakgcc -o xvid_encraw xvid_encraw.o -lc -lm -lxvidcore
23:05:15mirak/usr/bin/ld: ne peut trouver -lxvidcore
23:05:47mirakbut since the lib is not installed
23:05:51mirakthe dev lib at least
23:06:13_FireFly_edit the Makefile and add to the LDFLAGS-var the -L./ parameter
23:06:16 Join Febs [0] (
23:07:06mirak_FireFly_: you are a god
23:07:09mirakit worked
23:07:26_FireFly_btw have you tried this ?? : LANG=en gcc -I ../src/ -L ./ -lc -lm -lxvidcore -DARCH_IS_IA32 -DARCH_IS_32BIT -DARCH_IS_LITTLE_ENDIAN -fPIC xvid_decraw.c
23:07:39_FireFly_the order of the -l parameters are importend
23:09:23mirak_FireFly_: ope
23:09:27miraknope it doesn't work
23:10:31_FireFly_strange because it is the same line
23:11:01_FireFly_i have to go cu
23:11:15_FireFly_and a good night :)
23:11:16mirakbye thank you
23:11:27 Quit _FireFly_ ("Leaving")
23:13:49mirakgcc -o xvid_decraw xvid_decraw.o -lc -lm -lxvidcore -L./
23:13:53mirakthis line alone works
23:15:02***Saving seen data "./dancer.seen"
23:21:13 Join webguest84 [0] (
23:21:36 Quit Mark__ ("Leaving")
23:24:08 Join Rob2222 [0] (
23:24:58miraknope it doesn't a .o was left
23:30:26jlogood night
23:30:28 Quit jlo ()
23:34:27 Part webguest84
23:37:13 Join tvelocity [0] (n=tony@
23:40:33peturno use waiting for Linus here, I'm off to bed. Goodnight!
23:40:45 Quit petur ("here today, gone tomorrow")
23:40:55 Quit nathanh ("Quit")
23:44:47 Join linuxstb_ [0] (
23:46:39linuxstb_mirak: You may find this a good place to start:
23:47:00linuxstb_I've stripped xvidcore to the bare minimum and written my own simple Makefile
23:47:50linuxstb_It's going to take a lot of work to get that working in Rockbox though.
23:49:08mirakyou stole my job !
23:49:45linuxstb_No, I don't want that job
23:50:26dwihnoan xvid playback test thing?
23:51:03linuxstb_Just the xvid library with a (PC based) test app
23:52:02dwihnoah, okay...
23:52:08dwihnoImagine xvid playback on target :)
23:52:23linuxstb_Maybe on the gigabeat...
23:52:32Bagderx5 plays xvid
23:52:37Bagder13 fps
23:52:51linuxstb_Not with xvidcore I expect
23:53:07dwihnoBagder: stock firmware, I presume?
23:53:15Bagderdwihno: indeed
23:53:47dwihnoIs any of the CPUs fast enough for decoding of lo-res vids?
23:53:54mirakwhen I see the cube running on the H300 I suspect the screen to have to much latency over 10 fps
23:54:07 Join webguest46 [0] (
23:54:09mirakbut if that's the same screen than ipod video
23:54:21mirakmaybe it's ok
23:54:44linuxstb_The h300 isn't the same as the video ipod.
23:55:10webguest46 looks really good!
23:55:26dwihnoAh, stupid me didn't consider screen latency
23:56:24mirakwebguest46: that's on what device ?
23:57:10webguest46RotAtoR posted that earlier on
23:57:33mirakit's on current build ?

Previous day | Next day