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 2006-01-17

00:00:48amiconnwindows unzip != winzip
00:01:41preglowright, i misread
00:03:04preglowlinuxstb: i still have noisy glitches with the code you told me to add to your patch
00:03:24linuxstbpreglow: Let me upload a clean patch (if I can find one).
00:03:48preglowdidn't you disable ipod cache in cvs?
00:03:56linuxstbI don't think so.
00:04:06linuxstbOnly for the bootloader I think.
00:04:13preglowthink i dreamt it
00:05:30preglowthings seem awfully stable...
00:05:58preglowamiconn: seems you're good for ipod coding even without bloody owning one
00:05:58 Quit ModernExecutive ()
00:06:10amiconnhehe :)
00:06:13linuxstbWe have him to thank for the lcd code as well :)
00:06:36preglowi've loaded plugins over and over again
00:06:39preglowit works now
00:06:44preglowcache problem solved
00:06:51preglowwhich is a great relief, i can tell you
00:07:00amiconnTeh lcd code also helps the H3x0, which I do have now
00:07:08linuxstbpreglow: Absolutely.
00:07:20linuxstbI'm just glad Linus decided to write a 16-bit driver.
00:07:51amiconnDunno if we will ever switch to 18bit. Even 16bit is slow.
00:07:52preglowbut ok
00:07:57linuxstbpreglow: I think we just need your ATA fix in CVS now....
00:08:10preglowok, what approach should i take? comment away the entire thread or just the sleep magic
00:08:17 Part b0br
00:08:36linuxstbDid the ifp port do anything with that code?
00:08:36preglowi'm for the last
00:08:41preglownot afaik
00:08:47amiconnWhat's the problem with ata?
00:09:05preglowamiconn: sleep causes subsequent accesses to hang on nano
00:09:20preglowamiconn: so far i've fixed it by commenting out the ata thread, but i don't know how clever that is
00:09:30linuxstbThe ifp code doesn't use the standard ata.c anyway, so ignore me.
00:09:51preglowlinuxstb: i'm really, really anxious to see how the codecs perform
00:09:58preglowso far even flac skips, so that's gotta be a driver issue
00:10:17amiconnHmm. Not sleeping may even hurt battery runtime on flkash devices, depending on how the flash works
00:10:25amiconnIf it auto-sleeps instead, it's okay
00:10:41preglowamiconn: well, yeah, it's not like it's a permanent solution
00:10:58amiconnIs there a data sheet for the flash chip?
00:11:16preglowwouldn't matter if there were
00:11:24preglowwe speak to it through an ata interface
00:11:27preglowbut gimme a sec
00:11:30linuxstbpreglow: I think you should just put a #ifndef APPLE_IPOD_NANO around the actual sleep magic, and mark it as a temporary solution until (if) we understand the flash.
00:11:34amiconnYes, but ata != ata
00:11:50amiconne.g. not all devices implement all commands
00:11:52 Join webguest80 [0] (
00:12:10linuxstbBut it seems to do something when told to sleep.
00:12:29preglowamiconn: what i mean is that i doubt the ata interface is in the flash itself
00:12:30 Join `3nergy [0] (
00:20:39BagderTotal Physical Source Lines of Code (SLOC) = 230,432
00:21:02amiconn. . . .
00:21:16Bagderactually, I only counted firmware and apps
00:22:47preglowlinuxstb: find a patch?
00:23:12*Bagder compiles useless numbers for an upcoming talk...
00:23:22Bagderand I'll blab something about Rockbox then
00:23:35linuxstbI'm listening to it now - no noise.
00:23:44linuxstbBut a nasty stutter.
00:24:03 Join TCK- [0] (
00:24:31linuxstbBagder: Speaking of useless numbers, how many $$ did it cost to develop Rockbox?
00:24:45Bagder"Total Estimated Cost to Develop = $ 9,630,799"
00:24:53linuxstbAnd donations so far? :)
00:25:00preglowsmall friggin' wonder, i misunderstood where you wanted me to apply the line
00:25:31 Quit petur ("here today, gone tomorrow")
00:25:51preglowlinuxstb: guess what, now that just sleep is disabled and ata thread is running, rockbox reboots on usb
00:26:02linuxstbNice :)
00:26:07linuxstbDoes it enter USB mode?
00:26:25preglowas in disk mode?
00:26:32preglowjust reboots
00:26:41preglowuntil i enter disk mode myself
00:26:46linuxstbSame as the video then.
00:26:53preglowsmall wonder
00:27:22preglowyeah, audio does work now
00:27:33preglowdoes it block or something?
00:27:41linuxstbYes, I think that's the problem.
00:27:55preglowexplains why everything is just as far away from being realtime
00:28:26 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
00:28:35preglowaac isn't faring too bad compared to on coldfire
00:28:40*amiconn wonders why the codec api struct is that huge
00:28:48 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
00:28:50preglowamiconn: it's unmeeded
00:28:54amiconnMany functions make no sense at all for plugins
00:28:56preglowcontains a lot of stuff that doesn't belong
00:29:04preglowfor codecs, you mean
00:29:15amiconnargh, of course
00:30:54preglowi should really get to bed soon
00:31:38preglowactually wavpack is closes to realtime
00:31:48linuxstbpreglow: If we can get the main CPU responding to FIQs, it would be nice to see if Rockbox could work without the cop.
00:32:00preglownah, flac is equally close
00:32:08preglowi think it will, to some degree
00:32:10preglowwe'll see
00:32:28preglowi'm really bogged down with work these days, but i bet i'll be able to sneak some time in
00:32:31preglowbut now i've gotta go
00:33:11 Quit DreamTactix291 (Read error: 110 (Connection timed out))
00:41:10 Join Linas5 [0] (
00:49:54 Quit TCK- (Read error: 110 (Connection timed out))
00:52:41mirakwith gdb I do that :
00:52:42mirak(gdb) print buffer
00:52:42mirak$7 = (unsigned char *) 0x822d4c0 "n\024"
00:53:07 Quit akaidiot (Read error: 110 (Connection timed out))
00:53:14mirakthen a function is called wich take buffer as parameter
00:53:16mirakmpeg2_buffer(decoder, buffer, buffer + size);
00:53:55miraknow inside the function with signature :
00:53:56mirakvoid mpeg2_buffer (mpeg2dec_t * mpeg2dec, uint8_t * start, uint8_t * end)
00:54:03mirakI do a print start
00:54:17mirakand it says :
00:54:18mirak$12 = (unsigned char *) 0x2 <Address 0x2 out of bounds>
00:54:35mirakhow is it possible ?
00:54:52solexx_um, have I missed integration of mp3 encoding support?
00:55:08solexx_the CVS log suggests there is an mp3 encoder plugin
00:55:19solexx_but I cannot find it in the Plugins Index
00:55:38solexx_(see CVS message from 16 Jan 15:24)
00:56:04*solexx_ doesn't have his iriver around
00:56:23linuxstbsolexx_: It's a "viewer", not a plugin.
00:56:47linuxstbi.e. you long-click on a wav file in the file browser, select "open with" and then select "mp3 encoder"
00:56:59linuxstbThat will then convert the WAV to an mp3.
00:57:00***Saving seen data "./dancer.seen"
00:57:43solexx_linuxstb: thanks for the clarification
00:58:01solexx_I will try it.
00:58:12linuxstbviewers are just a special type of plugin - which is why most people just call it a plugin.
00:58:28 Quit YouCeyE ("Leaving")
00:59:37mirakprint mpeg2dec->buf_start
00:59:37mirakCannot access memory at address 0x222f1e
00:59:53mirakhuh. How can I not acces to that memory ?
01:00:30mirakwhen you run the simulator you don't have a 32meg static ram ?
01:01:42linuxstbNo, the memory will just be parts of your PC's memory.
01:02:13mirakbut the simualtor is in charge of allocating it ?
01:02:41mirakor you must use real mallocs ?
01:03:02linuxstbNo, you don't use real mallocs. get_plugin_buffer and get_audio_buffer will return buffers of approximately the right size.
01:04:03mirakhow is it possible that when I pass variable to a function, the input parmater is not the same when inside of that function ?
01:04:31mirak$7 = (unsigned char *) 0x822d4c0 "n\024" is transformed to a $12 = (unsigned char *) 0x2 <Address 0x2 out of bounds>
01:04:50miraklinuxstb: have you faced that one day ?
01:05:01linuxstbNo, I never use gdb
01:05:23linuxstbWell, I just use it to backtrace segfaults sometimes.
01:05:49linuxstbSo I'm afraid I can't help with gdb debugging.
01:07:06mirakseems that gdbs ault
01:07:12mirakit works outside of it
01:10:12 Join DJDD__ [0] (
01:16:24 Quit kernelsensei (Remote closed the connection)
01:16:28 Join kernelsensei [0] (n=boris@unaffiliated/kernelsensei)
01:16:37markunmirak: can I have your latest mpeg2 source to look at?
01:23:07 Quit actionshrimp ("a bird in the bush is worth two in your house")
01:28:09mirakmarkun: sure
01:28:14mirakmaybe you can help me
01:28:23mirakyou can try the dsl simulator
01:28:28mirakI have put some debug code
01:30:45mirakmarkun: are you sleeping ?
01:31:28mirakso the problem is that for a reason I don't understand some variables a reinitialised to zero
01:31:28markunI was about to go sleeping
01:31:56mirakthe debug output is explicit enough
01:32:14mirakthere is 0,0,0 they should have values
01:32:17mirakI should sleep to
01:33:26markunGood night mirak
01:33:30mirakthen it tries to show what is at this adress and of course it crashes
01:33:46markunvery strange indeed
01:38:04 Quit mikearthur (Remote closed the connection)
01:38:41 Join mikearthur [0] (
01:40:09 Quit ender` (Read error: 110 (Connection timed out))
01:41:23birdfishAlright, I got a 2gig nano, so hopefully I'll be able to start learning C and helping to develop soon.
01:54:07 Join linuxstb_ [0] (
01:55:14 Quit linuxstb (Read error: 110 (Connection timed out))
02:00:21linuxstb_birdfish: You came at a good time. Some important bugfixes for the ipod were committed to CVS today.
02:01:58linuxstb_The first thing you will probably want to do is to set up a Rockbox development environment and compile and install Rockbox on your ipod.
02:04:37Linas5anyone use/own a pair of akg k26p's?
02:13:10 Join DreamTactix291 [0] (
02:22:55 Join NibbIer [0] (
02:23:05 Quit NibbIer (Read error: 104 (Connection reset by peer))
02:26:38 Quit kernelsensei (Connection timed out)
02:35:02 Join akaidiot [0] (
02:35:25 Join ehntoo [0] (
02:35:39 Join XavierGr [0] (
02:57:02***Saving seen data "./dancer.seen"
02:59:08 Quit webguest80 ("CGI:IRC (EOF)")
03:10:27 Quit DreamTactix291 (Read error: 110 (Connection timed out))
03:12:00 Join kernelsensei [0] (n=boris@unaffiliated/kernelsensei)
03:20:40birdfishlinuxstb_: sure thing. But I have to study/take exams this week, so it will be a little while before I can really do much. :/
03:28:52 Join kernel_sensei [0] (
03:32:15 Quit mikearthur (Read error: 104 (Connection reset by peer))
03:32:56 Join mikearthur [0] (
03:35:56 Quit kernelsensei (Connection timed out)
03:35:59 Nick kernel_sensei is now known as kernelsensei (
03:53:05 Quit akaidiot (Connection timed out)
03:53:17BHSPitMonkeylinuxstb_: so is rockbox for ipod profoundly better than it was say, a week ago?
03:53:32linuxstb_Yes. It seems to work reliably now.
03:53:50linuxstb_Give it another go...
03:54:05*BHSPitMonkey is reluctant.
03:54:28linuxstb_Go on, it's easy to backup and then restore your ipodloader2 installation.
03:56:20BHSPitMonkeyis the rockbox bootloader any different? because i already have that image made...
03:57:23linuxstb_No, I think it's the same.
03:57:37BHSPitMonkeyit didn't work too well for me last time ( the bootloader )
03:57:54BHSPitMonkeyi switched a variable so that apple would be default and only rockbox worked
03:57:56linuxstb_Yes, it still has problems loading retailos. This just seems to be a Nano problem though.
03:58:05BHSPitMonkeynot variable, but call
03:58:53linuxstb_See, you are a C programmer. You can add rockbox support to ipodloader2...
03:59:01BHSPitMonkeyi tried
03:59:02BHSPitMonkeyand failed
03:59:05BHSPitMonkeynow it's your turn
03:59:26linuxstb_one day...
04:02:03*BHSPitMonkey dangles a paypal dollar above linuxstb_'s head
04:03:33*linuxstb_ grabs it and goes back to what he was doing
04:04:32*BHSPitMonkey voids the transaction
04:04:39BHSPitMonkeyand what is it you were "doing"?
04:06:44linuxstb_My day job
04:08:03 Quit birdfish ()
04:13:49BHSPitMonkeybeing an irc jockey?
04:15:58 Join hd [0] (
04:19:48 Quit goa (Success)
04:19:49 Nick hd is now known as goa (
04:21:31 Join Jungti1234 [0] (n=jungti12@
04:25:43BHSPitLappygonna try a build now
04:34:10BHSPitLappylinuxstb_: ping
04:37:06BHSPitLappysyetem.c:80:1: unterminated #elif
04:37:38BHSPitLappymake[1]: *** [/home/stephen/rockbox-devel/build/firmware/system.o] Error 1
04:37:57BHSPitLappymake: *** [all] Error 2
04:38:14linuxstb_You need to re-run ../tools/configure
04:38:24linuxstb_Sometimes CVS changes make that necessary.
04:42:26 Join amiconn_ [0] (
04:43:21BHSPitLappyi just did though
04:43:35BHSPitLappyi updated, cleared my build directory (rm -rf), and did configure
04:44:18linuxstb_And you still get that problem?
04:45:12BHSPitLappythat was the original context, yes
04:45:27 Quit bagawk ("leaving")
04:45:38linuxstb_Where were you when you typed "cvs update" ? Did you update the tools directory?
04:47:26linuxstb_Ignore my first answer, that error seems to be a different problem.
04:50:05 Join DreamTactix291 [0] (
04:50:34linuxstb_BHSPitLappy: It compiles file for me, so I'm not sure what your problem is. Try typing "cvs update -dP" in your rockbox-devel directory.
04:50:41BHSPitLappythat's what i did...
04:51:21linuxstb_Well, something went wrong... Have you edited any Rockbox files yourself?
04:51:22BHSPitLappyi came over here, sat down, cd'd to that dir, "cvs update -dP" (it updated), "cd build", "rm -rf *", "../tools/configure",
04:53:42linuxstb_I've just built Rockbox for the Nano without problems...
04:57:06***Saving seen data "./dancer.seen"
04:58:33linuxstb_BHSPitLappy: If you're interested, my build:
05:00:46 Quit amiconn (Read error: 110 (Connection timed out))
05:00:47 Nick amiconn_ is now known as amiconn (
05:01:32 Join Rob2222_ [0] (
05:02:42lostlogicgah, wrong window
05:06:50 Join amiconn_ [0] (
05:15:05BHSPitMonkeylinuxstb_: ok, about to boot it up
05:16:44BHSPitMonkeywow, this is real smooth!
05:16:52BHSPitMonkeymuch better!
05:17:39eEyerockbox works with the ipod?
05:18:10BHSPitMonkeyit has for a while, but NOW it's great
05:18:40eEyeso should i buy an ipod over the x5?
05:18:50eEyewhat features is added to the nano?
05:18:54BHSPitMonkeyyour call/
05:19:07 Quit Rob2222 (Read error: 110 (Connection timed out))
05:21:41 Join Membrillo [0] (n=sam_kill@
05:23:56 Quit Xerion (Read error: 104 (Connection reset by peer))
05:24:40 Join Xerion [0] (
05:25:26 Quit amiconn (Read error: 110 (Connection timed out))
05:25:26 Nick amiconn_ is now known as amiconn (
05:26:21BHSPitMonkeylinuxstb_: themes don't appear to work
05:26:29BHSPitMonkey"Loading..." hang
05:27:45BHSPitMonkeysorry if i'm being naggy, you've got work to do
05:29:27 Quit DreamTactix291 (Read error: 110 (Connection timed out))
05:32:15BHSPitMonkeythere's even a starfield app!
05:32:40linuxstb_That's my favourite plugin...
05:33:35Linas5god damnit
05:33:40Linas5my x5 is broken
05:33:41BHSPitMonkeythe ipl guys are talking about implementing screensavers
05:33:45Linas5fell off my desk
05:33:49BHSPitMonkeysome of these would be good source to steal hehe
05:33:55linuxstb_Linas5: :(
05:34:03Linas5and my sr-60's headphone cable/connector is stronger than the x5's headphone out
05:34:32Linas5the top part of the headphone port thing is pushed up when the connector hit the ground
05:34:45Linas5so it wont make full contact with the headphone thing anymore
05:34:53Linas5think its fixable/rma-able?
05:34:59 Quit perplexity (Read error: 110 (Connection timed out))
05:35:10BHSPitMonkeyoh weird, linuxstb_
05:35:11 Join akaidiot [0] (
05:35:27linuxstb_Linas5: Does it boot and run?
05:35:28BHSPitMonkeyafter running like 15-20 plugins, i can't start any more
05:35:38Linas5it works just fine
05:35:45Linas5just the headphone port thing doesnt make full contact
05:35:51Linas5since half of it is bent out of shape
05:36:01linuxstb_Then it's not too bad then.
05:36:18Linas5think i can use the warrenty service?
05:36:38linuxstb_BHSPitMonkey: Let us know if you can repeat that problem.
05:37:29BHSPitMonkeythemes still freeze no matter what...i'm kind of tired of black-on-blue :/
05:37:55linuxstb_Colour support in Rockbox is still relatively new. You can't change the colours without editing the source.
05:38:07linuxstb_And to answer the next question - firmware/export/lcd.h
05:38:19BHSPitMonkeyno thanks, i don't plan to
05:38:31linuxstb_But it's being worked on - mainly by iriver H300 users
05:38:55BHSPitMonkeywell the colors themselves work in apps...don't see why the setting is that hard
05:39:19BHSPitMonkeyand no sound, right?
05:39:44linuxstb_afraid so. But it's hopefully not too far from getting sound. Then you'll see what Rockbox can do :)
05:40:27BHSPitMonkeywhat, it has something to do with audio then?
05:40:34BHSPitMonkeyi thought it was made by popcap
05:40:45Linas5think cowon would like to repair my x5?
05:41:31BHSPitMonkeyi still don't know the sure-fire way of exiting a plugin
05:41:36BHSPitMonkeyi've been getting lucky
05:42:14BHSPitMonkeysweet, the mystify screensaver!
05:42:21BHSPitMonkeythere should be the beziers one too
05:42:42linuxstb_Yes, the key mappings are not very consistent. That's something quite easy that you could work on if you wanted.
05:43:14BHSPitMonkeydict doesn't work...
05:45:04BHSPitMonkeyhaha, no battery reading right?
05:45:07BHSPitMonkeyalways 0%
05:45:47BHSPitMonkeyooh, where are you in credits?
05:49:00linuxstb_I assume you've found the font browser?
05:50:22BHSPitMonkeyi just finished the plugins
05:50:41linuxstb_General Settings -> Display -> Browse Fonts
05:51:09BHSPitMonkeythats a lot
05:51:11BHSPitMonkeywhat are the best ones
05:51:23 Quit Linas5 (Remote closed the connection)
05:52:17linuxstb_I just stick with the default. I prefer a small font.
05:53:23 Join YouCeyE [0] (
05:53:59Paul_The_NerdI actually usually use 5*8 on H120. But it seems to be just right for how I like my remote wps to look
05:54:26BHSPitMonkeyi like jackash
05:54:52BHSPitMonkeyk, as cool as this is, i'm switching back to loader2 now :(
05:55:25BHSPitMonkeythe usb sense should prompt to reboot...not just do it
05:55:39Rob2222_linuxstb_ isnt bootloader in cvs?
05:56:05linuxstb_hehe. There's an endian bug in the text-file viewer.
05:56:14linuxstb_Every two characters are transposed...
05:56:43linuxstb_Rob2222_: Yes, all the source is in CVS
05:57:16Rob2222_but i get "no such repository"
05:57:25Rob2222_cant i checkout the BL anonymous?
05:57:41linuxstb_BHSPitMonkey: I disagree. If you want to charge, you hold MENU whilst inserting the USB cable. Otherwise Rockbox will attempt to reboot into USB mode.
05:58:01linuxstb_This way, there are no button presses if you want USB mode, and one button press if you want to charge.
05:58:25linuxstb_Rob2222_: If you checkout "rockbox-devel", the bootloader is there.
05:58:31Rob2222_k, thx
06:03:08 Join JoeBorn [0] (
06:03:41 Join jborn_ [0] (
06:03:53 Quit Jungti1234 ()
06:07:37 Part JoeBorn ("Leaving")
06:08:02 Quit jborn_ (Client Quit)
06:09:08 Join JoeBorn [0] (
06:10:00 Join Xerion_ [0] (
06:14:51 Quit Xerion (Read error: 104 (Connection reset by peer))
06:14:52 Nick Xerion_ is now known as Xerion (
06:16:27Membrillodoes anyone know if you can take harddrives out of ipods and put them in H3xxs?
06:16:36Membrillowith a format?
06:16:48Paul_The_NerdI'm pretty sure I've heard of it being done.
06:17:21Membrillothis might be a bit of a good deal then
06:18:21linuxstb_And you could donate the ipod shell to Rockbox for dissection :)
06:18:26Paul_The_NerdIf it is what it claims to be, yes.
06:18:35Membrillono probs
06:19:00Membrilloif it does indeed work fine, i can sell my current 20gb off as a replacement HD as well...
06:20:30Membrilloi cant seem to find any threads on MR
06:21:12Paul_The_NerdI'm pretty sure the 40gb iPods use the exact same drive as the H340, but I can't remember where the source of this information in my head comes from, and it could be merely wishful thinking.
06:21:49Membrillowell, even if they do, i think i will need to find a H340 casing. It will be a struggle getting the 40gb harddrive in my h320
06:22:06 Join Xerion_ [0] (
06:22:48Paul_The_NerdProbably will be.
06:24:00 Quit Xerion (Read error: 104 (Connection reset by peer))
06:24:01 Nick Xerion_ is now known as Xerion (
06:35:25Membrillodoes anyone have a spare H340 casing lying around?
06:35:33Membrilloi doubt it but its worth a shot haha
06:37:07 Join matsl [0] (
06:41:15BHSPitLappyuse a classic game boy case
06:41:34 Quit markun (Remote closed the connection)
06:41:41 Join markun [0] (
06:41:55Membrilloyour not serious are you?
06:42:03BHSPitLappyha, that would be AWESOME
06:42:34Membrilloexcept i wouldnt be able access most of the buttons and wouldnt be able to see about a cm off the screen from all sides
06:44:39 Quit tvelocity ("Leaving")
06:45:28linuxstb_Membrillo: Have you done any tests to compare the sound of your H340 and iPod Video?
06:45:53Membrilloive never even listened to audio on my ipod. Ive only watched video
06:46:00Membrilloi guess i could though
06:47:36Membrilloyou mean, listen to the same file on both players and judge which sounds better?
06:48:35linuxstb_Yes. Preferably an uncompressed file.
06:49:13Membrilloi shall do that
06:49:34Membrillobtw linuxstb: do you know what harddrive the 4gs use? is the same as the H300 players?
06:50:02linuxstb_I'm pretty sure they're the same, but I'm not sure of the model numbers.
06:51:01linuxstb_I have a 60GB ipod (4g color), and I seem to remember someone here saying they had the exact same hard drive working in a h140.
06:51:22 Join Xerion_ [0] (
06:51:45Membrilloah ok
06:52:03Membrillowell, im thinking of buying a broken ipod 40gb 4g monochrome for my h320
06:52:22Paul_The_NerdI'm pretty sure they're the same line of drives
06:52:52 Quit akaidiot (Connection timed out)
06:53:45Membrilloi cant find a model number anywhere
06:54:01Membrilloive googled all over the place
06:54:08Membrillooh well, ill buy it and hope :)
06:56:15 Quit Xerion (Read error: 104 (Connection reset by peer))
06:56:15 Nick Xerion_ is now known as Xerion (
06:56:39 Quit lars|offline (Read error: 104 (Connection reset by peer))
06:56:49 Join _jelle-k [0] (
06:57:07***Saving seen data "./dancer.seen"
06:57:36 Quit jelle-k (Read error: 104 (Connection reset by peer))
06:57:44 Nick _jelle-k is now known as jelle-k (
07:00:54*Paul_The_Nerd is now working on getting dice.rock to work on nano
07:03:30 Quit Membrillo ()
07:04:01 Join ModernExecutive [0] (
07:04:32ModernExecutivewas curious if anyone was in the states and had a bdm wiggler with which they could unbrick my h320
07:06:15linuxstb_ModernExecutive: I think that's extremely unlikely.
07:06:30ModernExecutive... figured it was worth asking. :)
07:07:45 Quit matsl (Remote closed the connection)
07:07:53ModernExecutiveamazingly, shipping to stockholm is only $5.25
07:08:19Paul_The_NerdI just figured something out.
07:08:45Paul_The_NerdThe emergency black and white disk-mode on the iPod nano transfers files REALLY slow, the normal full-color one doesn't.
07:08:51 Join DreamTactix291 [0] (
07:09:42 Quit ModernExecutive ()
07:09:53linuxstb_Paul_The_Nerd: That's useful to know.
07:10:09linuxstb_That should give preglow some motivation to try and fix the bootloader :)
07:10:41Paul_The_NerdMaybe it's using USB 1.1 in the emergency fw?
07:10:57linuxstb_I don't know. It's possible.
07:11:01*Paul_The_Nerd shrugs
07:11:52Paul_The_Nerdlostlogic: Created some alternate keymappings for your dice plugin for iPod, and fixed it for the new plugin loader if you're interested, when you're about.
07:12:08 Join ser_hades [0] (
07:12:39linuxstb_Paul_The_Nerd: Should it work on any ipod now then?
07:12:45Paul_The_NerdI think so.
07:13:22linuxstb_They all share the same IPOD_4G_PAD define - the only real difference between them is LCD size.
07:13:22 Part ser_hades
07:13:38Paul_The_NerdHe was using BUTTON_blah all throughout it, so I changed them to DBUTTON_blah and mapped a few of them to buttons that actually exist on the ipod in an #if
07:13:45Paul_The_NerdYeah, they should work then
07:14:05Paul_The_NerdThe plugin itself is entirely text based, and doesn't use many lines at all. It also auto-centers vertically at least, so it looks decent.
07:14:12linuxstb_You could use DICE_BUTTON_BLAH - that would be more consistent with the other plugins.
07:14:21Paul_The_NerdI'll go do that then
07:15:26Paul_The_NerdSo far it only checks if CONFIG_KEYPAD==IPOD_4G_PAD
07:15:50Paul_The_NerdSo it either uses those, or what it was originally set with.
07:16:36Paul_The_NerdIt only uses up, down, left, right, on, off, and select buttons. are BUTTON_blah not defined for those on any other players?
07:17:25linuxstb_I'm not sure. There are too many targets...
07:18:23Paul_The_NerdI'll just grab some already done ones from another plugin
07:18:34linuxstb_You could check firmware/export/button.h - that defines the buttons for each target.
07:18:46linuxstb_LEFT/RIGHT are the only universal buttons I think.
07:21:57Paul_The_NerdAlright, I just grabbed 'em from bejeweled
07:22:10Paul_The_NerdHow many lines is the smallest screen?
07:22:51jelle-k2 :P
07:23:13Paul_The_NerdThose would the the Players?
07:23:40Paul_The_NerdThe dice plugin seems to require 7 lines, I think
07:26:53 Join B4gder [0] (
07:28:34 Join perplexity [0] (n=joust@
07:40:48Bgermorning :)
07:41:19Paul_The_NerdMornin' Bger
07:43:43Bgerdoes anyone has any idea what's the linux's module for Realtek 8139D (i tried 8139too, but it didn't recognize the card)
07:46:12perplexityis that an ne2k compatible?
07:46:23Bgerdidn't try that yet
07:47:06perplexityJust tickled something back in my grey matter.. try the ne2k-pci module (or whatever it's called these days)
07:47:45Bgerhehe 10x
07:48:02Bgerthis is really strange ...
07:48:11B4gderthere's a r8169.c...
07:48:23B4gderprobably not good enough for you
07:48:58Bger8139 *should* support it, but maybe the problem is with the "burned" manifacturerid and deviceid
07:49:08Bgeri mean 8139too
07:49:15 Quit Bger ("brb reboot")
07:50:47perplexityyeah.. it does not match any of the pci-ids in ne2k-pci.. wild goose chase there
07:52:00B4gdernothing beats kernel driver hacking in the morning :-)
07:53:23perplexityonly kernel driver hacking a) drunk, b) late at night, and c) under the influence of Pink Floyd.. zen hacking
07:56:39 Join Bger [0] (n=Bager@
07:59:19Bgeryep, seems this is the problem
07:59:31Bger00:0d.0 Ethernet controller: Unknown device 1904:8139 (rev 01)
08:00:01Bger(just updated pci.ids)
08:01:18perplexityhow odd... that should be a very common pci-id..
08:02:09Paul_The_NerdAlright, I think I got LostLogic's dice working properly for Ondio, Recorder, iPod, and H100 and 300 targets.
08:02:33perplexityMind you.. google only gives 3 hits on that id
08:03:28Bgeri'll try to "rape" 8139too to recognize it
08:03:32perplexitywhats the PCI ID of your card? It simply _cannot_ be a 8139D or it may
08:03:32perplexitybe some 8139D variant which is compatible with 8139C and hence its
08:03:32DBUGEnqueued KICK perplexity
08:03:32perplexityRead carefully, 8139D is very much different from 8139C. I dont know how
08:03:32perplexityor why. It's just is that way. If someone knows better then he / she
08:03:34perplexitybetter speak up...
08:03:36perplexityquoted from the web...
08:03:53perplexitylooks like there is no supported driver for the 8139D
08:04:07B4gderpatch patch patch ;-)
08:04:13BgerB4gder :P
08:04:43B4gderI bet you haven't done a kernel patch this week yet
08:05:34B4gderthe hard part would of course be to figure out the hw diff
08:06:09perplexitythe lack of documentation is startling
08:07:08perplexityOh yuk.. seems the standard way to get that card working is using ndiswrapper <retch>
08:08:15 Quit BHSPitMonkey (Read error: 110 (Connection timed out))
08:09:17 Quit BHSPitLappy (Read error: 110 (Connection timed out))
08:09:37Bgerperplexity :
08:09:38Bgerconst static struct {
08:09:39Bger const char *name;
08:09:39DBUGEnqueued KICK Bger
08:09:39Bger u32 version; /* from RTL8139C/RTL8139D docs */
08:09:47Bgerthis is from 8139too.c
08:11:37Bgeri still think it's just a pci-id problem...
08:13:02perplexityin which case it's dead easy, just add the id and fire it up..
08:16:04Bgeris there a way to change the card's pci-id ?
08:16:49B4gderadd that id to the list in the driver
08:17:49Bgeryep, but i prefer to change the id to a more-standard one ... :)
08:18:00Bger(if there is a way)
08:18:06B4gderI doubt you can do that (easily)
08:18:35Bgerah, okay
08:19:30 Join Jungti1234 [0] (n=jungti12@
08:19:53perplexityNo, in most cards the pci-id is in silicon.. there are a rare few that hold it in e2prom, but good luck getting the programming details for those.
08:20:16B4gderand they often do provide different numbers for a reason
08:20:40Bgerok, i'll try messing with the 8139too.c
08:20:49perplexitythat would be a _far_ easier option..
08:21:12Bgerand all this for a $4 crap ...
08:21:34perplexitythis is why I'd love some form of linux hardware compatibility database..
08:22:12Bgerperplexity see, i bought a realtek 8139 because i *know* it's supported
08:22:42perplexityyou mean you *thought* it was supported :)
08:22:54 Join BHSPitLappy [0] (
08:23:05B4gderbuying already supported hw is cheating ;-)
08:23:24CtcpIgnored 2 channel CTCP requests in 6 minutes and 7 seconds at the last flood
08:23:24*B4gder ducks
08:23:29perplexityI do the same almost.. I _always_ buy tulip based pci cards coz I know they _all_ work.. but are a little more expensive than most..
08:24:25perplexityI have 4 usb-wifi dongles at the moment coz I did not know which ones were/were not supported, so I bought all they had available :) Had to tweak one to get it going, the others are still doorstops.. (for now)
08:24:37*B4gder patched the ds1337.c driver the other day...
08:28:46 Join JamesK [0] (
08:29:05 Join DJDD___ [0] (
08:29:07BHSPitLappyfor what now?
08:30:59 Part JamesK
08:31:18 Join JamesK [0] (
08:33:15 Join JamesK_ [0] (
08:38:31 Join einhirn [0] (
08:39:58 Join JamesK__ [0] (
08:40:23B4gderlots of JamesK
08:42:38 Quit DJDD__ (Read error: 110 (Connection timed out))
08:53:55 Quit JamesK (Read error: 110 (Connection timed out))
08:54:30 Join ender` [0] (i=ychat@
08:56:07 Join Kohlrabi [0] (
08:57:08***Saving seen data "./dancer.seen"
09:00:15 Join Membrillo [0] (n=sam_kill@
09:00:23 Quit JamesK_ (Read error: 110 (Connection timed out))
09:02:36Jungti1234Why doesn't it change?
09:02:59B4gderyou mean it should somehow automatically update itself?
09:03:16 Quit JoeBorn ("Leaving")
09:03:21Jungti1234Update that someone does.
09:03:36B4gderif you find mistakes or inaccuracies, please update
09:04:07BHSPitLappyi see "wiki" in that url :P
09:04:08Jungti1234Do I update?
09:04:20 Join LinusN [0] (
09:04:21B4gderJungti1234: its a wiki, it is maintained by all of us
09:04:36Jungti1234I know but
09:04:44BHSPitLappyw i k i!
09:04:48Jungti1234Is worried for fear that I am mistaken.
09:05:08BHSPitLappyif you're wrong, then someone else will see it, get annoyed, and correct it.
09:05:14BHSPitLappyit's wrong NOW, isn't it?
09:05:29BHSPitLappyhow much worse can you make it
09:05:51Jungti1234oops no 'p'
09:06:35Jungti1234ok, I will update it.
09:07:26BHSPitLappyooh, stretched out
09:08:13BHSPitLappyhey, who here would like a rockbox userbar to put in your forum signatures?
09:08:15DBUGEnqueued KICK Jungti1234
09:08:19Membrillo<(^_^<) (>^_^)>
09:08:42BHSPitLappybesides kirby?
09:09:03crwlforum signatures suck, especially ones that have pictures, and it's sad you usually can't turn them off without registering to the forum first
09:09:18Jungti1234I have it
09:09:22crwlbut that's a thin picture, luckily ;)
09:09:28BHSPitLappyi know
09:09:32BHSPitLappythat's why they're so popular.
09:09:44BHSPitLappyand i noticed most people have the straight-out rockbox logo in theirs
09:09:54B4gderthat's one of the most annoying parts of MR
09:10:00B4gderthe huuuuuuuge pics
09:10:18BHSPitLappywe had some people in the iPL forums with a lot of nerve
09:10:19B4gderthen I'm regged and switched that crap off
09:10:33 Quit JamesK__ (Read error: 110 (Connection timed out))
09:11:06BHSPitLappythis one's MY sig though :)
09:11:06Membrilloif theyre restricted to a size, i dont have a problem with them
09:11:08B4gderforums just generally suck
09:11:26Membrillothey should have a limit to the amount of pixels of graphics you can havr
09:11:44BHSPitLappythey should take one very large forum, riddled with hate and flame wars, and then one day just make EVERYONE a mod
09:11:50BHSPitLappyeventually they all die out
09:13:03Paul_The_NerdA last man standing ban-war?
09:14:36Jungti1234I will give up. :(
09:15:09Jungti1234Because I don't English, it's difficult. ;;
09:19:20MembrilloJungti1234: no problems with the downloader?
09:19:43Jungti1234not yet.
09:20:22Membrillogood good
09:20:30Jungti1234It doesn't have icon.
09:21:08Membrillothat doesnt really matter though
09:21:19Membrilloi can send you one with the icon if it REALLY matters
09:22:42 Join ashridah [0] (
09:29:23 Quit RotAtoR ()
09:31:05Jungti1234what is 'Mid-track resume', 'Mid-playlist resume'?
09:31:40 Join DJDD__ [0] (
09:37:20Membrilloi have no idea sorry
09:40:04 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
09:40:08 Join mirfle [0] (
09:41:58mirfleI have a suggestion for the dir cache thing, anyine interested?
09:43:24B4gderJungti1234: it means it can resume in the middle of a track and in the middle of a playlist
09:43:34Jungti1234thank you
09:44:23 Join BHSPitMonkey [0] (
09:44:56 Quit DJDD___ (Read error: 110 (Connection timed out))
09:46:35LinusNwow, the playlist viewer can only handle 200 tracks
09:46:44mirflethe dir cache should be saved as a file on disk after the initial generation, and then loaded from that every time you turn on your player, thus saving alot of disk spinning time. then you add an "update dir cache" option for when changes are done. you can even have rockbox automatically update it whenever it is unmounted from a pc, and after saving a recording.
09:46:53 Join BHSPitLappy2 [0] (
09:47:12LinusNmirfle: i vote against that
09:47:33mirfleLinusN: why?
09:47:39 Quit BHSPitLappy (Nick collision from services.)
09:47:43LinusNthe dir cache must be synchronized at all times
09:47:49 Nick BHSPitLappy2 is now known as BHSPitLappy (
09:48:47LinusNthe user *will* forget to update the cache manually after making changes
09:49:09B4gderI agree
09:49:28mirfleif rockbox synchs it whenever IT does changes, and when it's unmounted, the only unacounted for times are when the original firmware is used to record or delete a file.
09:49:36Slasherimirfle: that will be possible allowed along with a new bootloader including eeprom support
09:50:36mirflewhat does eeprom support change (sorry for the noob question)
09:50:55LinusNSlasheri: how so?
09:50:57Slasheribootloader could mark on the eeprom if we need to rebuild the cache
09:51:06Slasheri(bootloader usb mode or loading original firmware)
09:51:08LinusNhow does the bootloader know that better?
09:51:21Slasheriit knows if the disk might become dirty
09:51:26LinusNso it will rebuld the cache if the original has been started?
09:51:54Slasheriyep, for example
09:52:00LinusNi feel that it is an unneccessary risk to save the cache like that
09:52:16B4gderhow does the code check for updates now?
09:52:27B4gderis there like a hash of the fat or something?
09:52:28LinusNit rescans every time
09:53:06mirflea hash of the fat could be a good idea
09:54:57B4gderI figure the majority of all bootups, the fat would be the exact same as the previous bootup
09:59:23mirakcan't the kernel monitor files wen rockbox is on ?
10:00:15mirakcan the cache be refreshed on exit of UMS mode ?
10:01:39B4gderyes it can
10:02:02B4gderbut when rockbox starts it doesn't know if there are new/removed files
10:04:31LinusNas long as there is an option to run the original firmware, there is a risk that the cache is incoherent
10:05:33B4gderand as long as the bootloader is somewhat independent from rockbox
10:06:02mirflewhy isn't the hash check for the fat reliable enough?
10:10:14 Quit darkless (Client Quit)
10:14:33 Join amiconn_ [0] (
10:17:05 Join stamppot [0] (
10:18:13 Join b0br [0] (
10:18:25Membrillo Jungti1234: yes?
10:18:41Membrillothats old
10:18:48Membrillorecheck the page
10:18:50Jungti1234I can't download.
10:18:58Membrillonew link
10:19:02Membrillocheck the app page
10:19:08Jungti1234I see
10:19:16MembrilloNew rockbox download version available at mistic river
10:21:07linuxstb_I don't know if anyone read my comment in the logs last night, but there is an curious bug on the ipod - each pair of characters is transposed.
10:21:20linuxstb_I mean in the text viewer...
10:21:50MembrilloJungti1234: what?
10:22:34Jungti1234Rockbox bug makes me fun.
10:23:22Membrillowhy what
10:23:32linuxstb_I initially thought this would be an endian issue in the text viewer (reading shorts, converting incorrectly to a pair of chars), but there are no shorts in the viewer. So I'm guessing it's more likely to be an ATA driver bug - but how could it only affect the text viewer?
10:23:33Jungti1234I don't understand
10:23:51Membrilloyou dont understand whay
10:24:00Jungti1234[18:22:54] <01Membrillo> sorry?
10:24:03B4gderlinuxstb_: it sounds unlikely...
10:24:16B4gderis there any unicode involved?
10:24:27linuxstb_B4gder: I know. But that's the only place where I can think that shorts are used and may not be endian-safe.
10:24:47Membrillo<Jungti1234> wahahaha ?????
10:25:00Jungti1234[18:22:43] <04Jungti1234> Rockbox bug makes me fun.
10:25:22linuxstb_I don't know if they are related, but the ipod seems to crash when building the dircache, and also when loading a theme.
10:25:31Membrillorockbox bug?
10:25:35linuxstb_But apart from that, everything's fine.
10:25:49Jungti1234Small problem of rockbox.
10:26:02B4gderhow are they transposed?
10:26:47linuxstb_iLket ih.s
10:26:53linuxstb_Like this.
10:30:03linuxstb_I'm assuming the viewer works on little-endian sims without problems.
10:30:24B4gderI assumse so too
10:32:02 Quit amiconn (Read error: 110 (Connection timed out))
10:32:02 Nick amiconn_ is now known as amiconn (
10:33:17 Join TCK [0] (
10:33:21linuxstb_I see in copy_read_sectors there are two code paths - one for reads to unaligned memory, and one for aligned. Maybe one of those is broken.
10:33:35 Join KN|stiff [0] (i=phhome@
10:34:02B4gderI can't see any obvious flaw in the apps/plugin side at least
10:34:56linuxstb_I can image the unaligned reads being used in the text viewer, dircache and wps loaders, but nowhere else.
10:43:21B4gderhm yes
10:43:33B4gderSWAP_WORDS is defined for SH but not coldfire ?
10:44:04B4gderand not arm
10:44:39B4gderthat can't be right, can it?
10:45:04linuxstb_Reversing the byte order in the unaligned part of copy_read_sectors() fixes the text viewer, and the rest of Rockbox seems to work as well as before.
10:45:17amiconnIt can
10:45:23B4gderhm, the SH code doesn't use the C code
10:45:39B4gderamiconn: how?
10:45:47B4gderis the ATA switched on SH?
10:46:27amiconnATA is little endian
10:46:36amiconnSH1 is big endian, so it has to swap bytes
10:46:45amiconnARM is little endian, so no swap
10:46:45B4gderbut coldfire and SH are both big endian
10:46:48B4gderso they should use the same
10:46:52B4gderbut they're not
10:47:00amiconncoldfire is big endian, but the iriver does the swap in hardware
10:48:54linuxstb_amiconn: Do you agree that the part of copy_read_sectors for unaligned reads is currently wrong for the ipod?
10:57:12***Saving seen data "./dancer.seen"
10:57:32LinusNimho, the playlist viewer should always set the cursor to the currently playing track when viewing the current playlist
10:57:49LinusNregardless of the "follow playlist" option
10:57:58CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
10:57:58*B4gder assumed it did...
10:58:24LinusNi think that changed with the multiscreen gui
10:58:56LinusNi'm changing it
10:59:58MembrilloLinusN: what do you think about an option to, while in a playlist, on pressing NAVI it navigates to the location of the current song?
11:00:46Membrillolocation as in location as a file
11:01:00LinusNcool idea
11:01:13amiconnlinuxstb_: I don't see what would be wrong there
11:01:16LinusNnavi plays the selected song
11:01:46BHSPitLappylinuxstb_: i got a desktop build to work! yay!
11:01:48Membrillobut from the WPS
11:01:51BHSPitLappynite nite :(
11:02:34Membrilloinstead of going to last navigated file, could there be an option to go to the currently playing file?
11:02:57B4gderMembrillo: follow-playlist you mean?
11:03:03linuxstb_amiconn: It doesn't do the same as the aligned read. i.e. the aligned read copies a short to memory with bits 0-7 first, and bits 8-15 second. The unaligned read does it the other way around.
11:03:04amiconnMembrillo: Enable 'follow playlist'
11:03:27Membrillowhere is that?
11:03:51Membrillofound it
11:04:09amiconnlinuxstb_: The fun thing is that unaligned read does work right for iriver, i.e. big endian + hardware swap
11:04:57amiconnThat would mean little endian cpu behaves different here. Fun... would need to make some drawings to understand this
11:05:37Membrillois there a button that goes from the WPS to the playlist viewer with the current song highlighted?
11:06:08B4gderthe reason is that for unaligned, it copies the 16bits as-is without SWAP_WORDS
11:06:23B4gderbut for unaligned, it always pick either version and none is as-is
11:06:49B4gdergee i type weird
11:06:57B4gderjust ignore me
11:07:23B4gderthe first line referred to aligned and the second for unaligned
11:09:20B4gderand in little endian, the lowest bits should go into the lower byte
11:10:00B4gderthe problem is the term "swap words" as in the ARM case it isn't swapping
11:11:40Membrillodoes anyone here own shure e2c's?
11:14:57 Quit ashridah ("Leaving")
11:15:35B4gderimho, the proper fix would be to read the unaligned 16 bits first in a defined (little endian) way, and then swap the bytes if SWAP_WORDS is set
11:15:50 Quit perplexity (Read error: 110 (Connection timed out))
11:20:18amiconnB4gder: We don't wanna sacrifice performance
11:21:00 Join BHSPitLappy2 [0] (
11:22:15B4gderperhaps the #ifdef simply could be defined to #if defined(SWAP_WORDS) || defined(ROCKBOX_LITTLE_ENDIAN)
11:23:25B4gderthe SWAP_WORDS define assumes big endian
11:24:59 Join linuxstb__ [0] (
11:26:30 Quit linuxstb_ (Read error: 110 (Connection timed out))
11:27:51Paul_The_NerdHas anyone in here compiled the ifp-790 port?
11:30:27 Part mirfle
11:30:38linuxstb__I've compiled it, but that's all.
11:30:45 Quit TCK ("I shall not cease!")
11:30:55linuxstb__And that was a week ago - so it could be broken.
11:32:17Paul_The_NerdDo I need anything special?
11:32:18 Join perplexity [0] (i=heh4519@
11:32:36Paul_The_NerdOr should a clean cvs and working arm toolchain be enough?
11:32:37preglowi get a makefile error
11:33:08Paul_The_NerdOr there could be other problems... heh.
11:33:44 Join BHSPitMonkey_ [0] (
11:34:22 Quit BHSPitMonkey (Read error: 110 (Connection timed out))
11:34:43 Quit BHSPitLappy (Read error: 110 (Connection timed out))
11:35:31*Paul_The_Nerd hadn't tried it yet.
11:36:11Paul_The_NerdI'm in the middle of compiling gcc 3.4.5 for arm right now, anyway.
11:36:21 Quit linuxstb__ ("Leaving")
11:37:50 Quit KN|stiff ("( :: NoNameScript 3.81 :: )")
11:38:25 Join |Lupin| [0] (
11:38:47|Lupin|Hello, there.
11:39:19|Lupin|My Iriver is alife again, that's great. However itSeems it freezes after a while inThe recording screen. Is anybody here awar of this problem, if reallythere is one ?
11:40:25amiconnpreglow: It seems that file is missing in cvs
11:40:39preglowamiconn: certainly does
11:49:24 Quit Jungti1234 ()
11:50:40 Join NicoFR [0] (
11:57:10 Join webguest67 [0] (
12:14:10 Quit b0br ("CGI:IRC")
12:18:45 Join Kohlrabi [0] (
12:19:38 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
12:21:59Bgerwow :)
12:23:16Membrillodoes anyone know if GAH or GAL harddrives are used in the H340s?
12:23:44Paul_The_NerdWhat he said.
12:30:15Membrillothis IRC channel has to have the highest rate of people saying hello when they join out of them all
12:30:30 Join cYmen [0] (
12:31:09cYmenis it possible to usb-boot from a rockbox?
12:31:53 Quit t0mas (Client Quit)
12:32:38 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
12:32:40Paul_The_NerdWell, you don't really have "a" rockbox. But I believe so far every device that runs Rockbox is a USB mass storage device when connected.
12:34:58Membrilloi think what he means is does it have usb bootloader mode
12:35:09 Join akaidiot [0] (
12:35:10 Quit stamppot ("CGI:IRC (EOF)")
12:35:32 Join stamppot [0] (
12:36:04Paul_The_NerdAh, I was thinking more of "Can I boot a computer off of one" or some such. Like a linux live-thumbdrive-thingy.
12:36:49Paul_The_NerdcYmen: What model of Digital Audio Player are you asking about, and could you ask the question another way?
12:41:19cYmeni want to boot a computer of of my ajr
12:41:35 Part t0mas
12:41:46 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
12:43:10Paul_The_NerdI think it's more dependent on if your computer can boot off of a USB drive than anything about the device. But it's not an area I've really played with at all.
12:43:35cYmenme neither and i agree it should be simple i was just asking out of curiosity
12:43:41cYmenarchos jukebox recorder
12:43:47Membrilloare nice
12:43:57Paul_The_NerdI think it should work though.
12:44:41Membrilloive got no idea about booting OS's
12:44:58Membrilloso you would install say windows or linux to your player?
12:45:13cYmenthat s the idea
12:48:18B4gderif that can boot off a mass-storage usb, it can boot from your recorder
12:51:08Membrillohas anyone tried any of the mario games on H3xx rockboy?
12:52:55 Join tvelocity [0] (
12:56:07NicoFRi tried super mario bros
12:56:11NicoFRit's quite nice
12:56:22NicoFRhaven't tried the latest patch yet though
12:57:14***Saving seen data "./dancer.seen"
12:57:39Membrilloi just remembered that mario was indeed on gameboy (doh!)
12:58:01Membrillothe connection of iriver - rockboy - gameboy - mario just didnt strike me
12:59:26Paul_The_NerdHow high is the volume supposed to go on iPods? (not that I have an audio-enabled build, mind you, just was curious if it's supposed to go up past 0 like it does)
13:03:49 Join petur [0] (
13:05:01*petur is looking for that code police guy ;)
13:05:17BgerB4gder, perplexity : the problem is not only with the pci-id ...
13:05:42BgerACPI: PCI Interrupt 0000:00:0d.0[A] -> GSI 17 (level, low) -> IRQ 19
13:05:42Bger8139too: 0000:00:0d.0: region #0 not a PIO resource, aborting
13:05:42BgerTrying to free nonexistent resource <dfffff00-dfffffff>
13:05:42BgerTrying to free nonexistent resource <0000ec00-0000ecff>
13:06:07Paul_The_Nerdpreglow: You around?
13:06:19Bgerpetur : u mean linus ? :)
13:06:54Bgerall he've done is to remove one var
13:07:36BgerB4gder this was after adding the pciid to the table of 8139too
13:07:53preglowPaul_The_Nerd: yeah, somewhat
13:09:31peturI know.... just want to push him to commit that brightness bugfix. It's a 1 minute job but turned out a 'how to create good patches' course. So I'm here for the exam :D
13:09:42Paul_The_Nerdpreglow: I was just curious if you saw in the logs, or otherwise heard. I've noticed that transfers to the Nano are very significantly slower under the black and white disk-mode using Select/Play when booting, than the Apple_OS disk mode, and Linuxstb thought that might be of interest to you.
13:09:57*Paul_The_Nerd just remembered about that.
13:10:00preglowyeah, saw that
13:10:26preglowjust need to get retailos to boot, then
13:10:31Paul_The_NerdJust makin' sure.
13:10:42Bgerpetur :P
13:10:42preglowwhich will be something of a problem, seeing as how just a couple of ifs make it not boot
13:10:58Bgerwhat bugfix ?
13:11:07Paul_The_NerdJust a couple of ifs?
13:11:31Paul_The_NerdWhat's lost if you want retail os to boot?
13:11:51peturBger: I forgot to disable the IRQs while calling the I2C function
13:12:04amiconnpetur: Btw, the brightness function in cvs (both before & after Linus' change) allows to set brightness 0 and 1
13:12:34preglowPaul_The_Nerd: a couple of checks for which key is pressed is all that makes it hang
13:12:39preglowPaul_The_Nerd: and i have absolutely no idea why that is
13:13:28Paul_The_NerdSo, if you don't check for the key it boots fine?
13:13:32Bgeramiconn ah, then i'm the one to blame :)
13:14:22Paul_The_NerdCould you point me to them?
13:14:36peturamiconn: you mean when calling pcf50606_set_bl_pwm()? I removed that one in my patch, it's now in backlight.c completely (where it belongs)
13:16:23Bgeramiconn : how/when it allows brightness of 0 and 1 ?
13:16:35preglowPaul_The_Nerd: as in what to do to make it work?
13:16:41preglowin bootloader/ipod.c
13:16:43 Join aliask [0] (
13:16:46preglowyou just #if 0 around the button checks
13:16:48preglowin main()
13:16:51peturBger: I think if you call pcf50606_set_bl_pwm() directly
13:17:32peturdon't know why I wrote it that way, changed in in my patch
13:17:41Paul_The_Nerdpreglow: I was just curious where the button checks were. I still don't know my way very well around the source. I figure, it's not like I can make it any worse in the end.
13:17:48preglowPaul_The_Nerd: i kind of figured it's got to be something we do with the hardware that makes retailos not boot, like initialise the button controller wrongly or something, but all i've tried points to that not being so
13:18:22Bgerpetur but we don't call pcf....() directly in any other place (iirc)
13:18:29preglowPaul_The_Nerd: line 405 to 444
13:18:38preglowPaul_The_Nerd: #if 0 or goto around that range, and it'll work
13:18:57Bgerand i don't think that we should do double check
13:19:24peturBger: As I said, I removed that function...
13:19:55peturjust waiting for Sir LinusN to give his blessing :D
13:20:21amiconnbacklight_set_brightness(9 first checks for val < minumum, and *then* masks with 0x0f
13:20:26B4gderpetur: don't forget to slip the money into his hand very descretely
13:20:38amiconnSo if you pass the function a value of 16, 32, 48 ... it will set 0,
13:20:50peturslipping code obviously doesn't help :p
13:20:52amiconnand if you pass 17, 33, 49... it will set 1
13:20:53Bgeramiconn this is linus' change :P
13:21:20amiconnAh, i c
13:21:33amiconnval &= 0x0f; should go before the check
13:21:38LinusNno, that's not my change
13:21:39|Lupin|Hi againGuys. Sorry, I was a way for a while. Did someone reply to what I said about the potential bug relatedtoTheRecording screen ?
13:21:44Bgerhuh ?
13:21:50amiconnLinusN: Yes it is
13:21:51LinusNit was like that before my change
13:22:18LinusNthe check was on val, not ucVal
13:22:24 Join Matze41 [0] (
13:22:35Bgeruh :(
13:22:41amiconnbah :/
13:22:52Bgerso, i'm the one
13:23:04preglow|Lupin|: no
13:23:53Bgersilly me
13:25:04peturwell my patch does it correctly...
13:25:21petur(hint hint :p)
13:25:32Paul_The_NerdPreglow: It's the key check step
13:25:46Paul_The_NerdSomething about the key_pressed function itself.
13:26:52preglowPaul_The_Nerd: i think i concluded that wasn't the problem
13:27:10preglowi think i had that called, then dropped the if checking, and then it suddenly worked
13:27:21preglowbut i'm not certain, so please do check further :P
13:27:23Paul_The_NerdWell, if I set i=BUTTON_MENU everything goes well
13:27:34preglowthat would be the most logical
13:27:46Paul_The_NerdWithout commenting out anything, if I just replace the key check with a pre defined key.
13:31:31peturBger: just checked the histroy, it was you... :p
13:31:52Paul_The_NerdMeanwhile, you're right, if I comment out the ifs, but leave the i=key_pressed() it also still works...
13:32:04preglowthat's probably because that code is marked as unused by gcc
13:32:06preglowand never included
13:32:16B4gdernow we see why they called the cvs annotate option "blame" in SVN ;-)
13:33:49peturgot to go, I'm way past my lunchbreak :(
13:33:54 Part petur
13:34:33|Lupin|C U later all :)
13:34:34 Part |Lupin|
13:36:20BgerB4gder :P
13:38:54*Bger remembers this as his 2nd commit and he was in hurry to correct the mistakes of the first one ...
13:47:11LinusNamiconn: i'd like a drawing mode that decreases the brightness of the pixels in the frame buffer, for shadows
13:47:49LinusNor maybe blends the drawed data with the date in the frame buffer
13:52:50 Quit akaidiot (Read error: 110 (Connection timed out))
13:54:37 Join Jungti1234 [0] (n=jungti12@
13:56:46amiconnLinusN: Perhaps lcd_set_transparency() in combination with the existing drawmodes?
13:57:04*amiconn suspects this to be slow as hell
13:57:19amiconnBtw, what do you want to do?
14:03:59LinusN3d look on the interface
14:04:14 Join KN|stiff [0] (i=phhome@
14:04:17 Quit mikearthur (Remote closed the connection)
14:04:25LinusNfor the 3d shadow
14:05:46 Join Skydemon [0] (
14:06:22 Quit mirak (Read error: 110 (Connection timed out))
14:06:55 Join mikearthur [0] (
14:08:58Paul_The_Nerdpreglow: I've run across some strange behaviour, I think...
14:09:39Paul_The_Nerdopto_keypad_read can be called, its return value assigned to a variable, and that variable output during the boot, and with i=BUTTON_MENU it still boots into RetailOS
14:11:57Paul_The_NerdSo, one would assume the problem lies between opto_keypad_read being called, and the return in key_pressed, but to my head that conclusion doesn't make sense.
14:12:00Rob2222_Binutils needs flex installed in order to compile <= Is this still needed?
14:12:15Rob2222_@ coldfire platform
14:12:16preglowPaul_The_Nerd: well, that's just a bunch of ifs
14:12:21Paul_The_NerdI *know*
14:12:23preglowdon't see how they can hurt
14:12:30preglowwhich brings us back to my original conclusion :)
14:12:38preglowgod knows what happens around that point
14:12:43Paul_The_NerdBut I mean, if I can call opto_keypad_read, get a value from it, *display* that value, clearly the problem isn't in it.
14:12:52Paul_The_NerdOr at least, isn't *just* it.
14:13:35B4gderRob2222_: I doubt anyone has bothered to figure that out
14:13:47amiconnLinusN: Do you really think 3D look is what we need now? I'd give that pretty low prio
14:14:03Paul_The_NerdRob2222_: At least on cygwin, I seem to recall seeing "Flex not installed" flash by during the build of binutils
14:14:20Rob2222_yes, i have cygwin here
14:15:08Rob2222_so i need flex, gettext bison and m4, to compile the binutils for iriver (coldfire)?
14:15:18Rob2222_like described
14:15:35B4gderif it says so, then it is probably right
14:15:56Rob2222_hmm, i dont know. i installed cygwin first time yesterday
14:15:58aliaskIn regards to the UI, wouldn't it be better just to have some kind of bitmap or gif file based theme rather than a statically rendered UI?
14:16:07B4gderaliask: no
14:16:14Rob2222_i managed to compile the crosscompiler for arm (ipod) and compile BL and FW
14:16:23aliaskB4gder: Why not?
14:16:28B4gderaliask: it'll never work with all the programmabilty and different fonts and languages
14:16:42B4gderunless you can show us how
14:16:42 Join Febs [0] (
14:16:54LinusNamiconn: of course it isn't what we *need*, i just wanted to give myself a view on how we want it done
14:17:27LinusNin case someone feels like actually doing a pretty interface instead of half-baked color settings
14:17:56B4gderI agree. We need to provide the tools for people to build real stuff with
14:17:59Rob2222_so simply compile binutils 2.16 out of the package wont work for coldfire, right?
14:18:23Membrilloif the tools are there it will also give people the incentive to start work
14:18:39B4gderRob2222_: 2.16 should work
14:19:06B4gderMembrillo: they already started, just doing it the "wrong" way ;-)
14:19:15Membrilloahhh, gotcha
14:19:32Rob2222_B4gder: Without the things described under Coldfire @ ?
14:19:45Membrillois that from the view of a rockbox ol' timer? or someone generally interested in getting good results
14:20:13LinusNMembrillo: the latter
14:20:14B4gderRob2222_: it doesn't mention any other versions or patches for binutils, afaics
14:20:24B4gderMembrillo: we want things done properly
14:20:28amiconnImho, we need the following UI things to be done before adding fancy stuff: (1) complete multiscreen UI (2) wps rendering rewrite (3) scroll thread rewrite, allowing attributes and variable bounding box, not only left margin
14:20:37MembrilloLinusN: Thats all that matters then. Keep up the good work
14:20:52amiconn*Then* wps can have colour support
14:20:53MembrilloB4gder: same to you. I wont disrupt you any more :)
14:21:08LinusNamiconn: time to set up a plan in the wiki then
14:21:34B4gderthe things amiconn mentions don't have to be done in that order or even by the same persons
14:21:52preglowor even at all for lots of gui improvements
14:21:54B4gderthus, doing fancy things wouldn't stand in the way
14:21:59preglowi wouldn't let that block other stuff
14:22:05amiconn(1) is connected with (2) and (2) is somewhat connected with (3)
14:22:34amiconnpreglow: Adding colour now would be an ugly hack
14:22:48B4gderI disagree
14:22:51preglowdepends on what you mean by 'adding color'
14:22:57B4gdernot if you do it properly
14:24:06 Quit Skydemon ("Nettalk6 -")
14:24:23 Quit Matze41 (Read error: 110 (Connection timed out))
14:24:59amiconnB4gder: Doing it properly *requires* attribute support in the scroll thread
14:25:12amiconnOtherwise it won't work for scrolling lines
14:25:13 Join Skydemon [0] (
14:25:53B4gderfor scrolling texts at least
14:26:01B4gdernot for status bar and scrolbars etc
14:27:02LinusNwe should at least express a general idea on how we want it to work
14:27:19LinusNso som eone could implement it
14:27:28B4gderand for some things to even be considerable, we need the GUI primitives
14:31:38 Join leftright [0] (
14:31:43 Join webguest95 [0] (
14:32:35webguest95hi all, i need some help, trying to build the toolchain for compiling rockbox for ipod
14:32:40leftrightLinusN:, your latest commit crashes my player big time :-) when selecting 'view playlist' form context menu crashhhh
14:33:08 Quit Skydemon ()
14:33:32leftrightneeed to do a hard reset, I have re-installed the latest Bleeding edge twice with same effect
14:33:48leftrightoh H140
14:33:49webguest95im running debian sarge, followed the instructions on the wiki for crosscompiler, but get this error when running make in gcc-3.4.5
14:33:53webguest95checking to probe for byte ordering... /home/svante/arm-elf//arm-elf/bin/ld: crt0.o: No such file: No such file or directory
14:34:03webguest95collect2: ld returned 1 exit status
14:34:14webguest95configure: error: unknown endianess - sorry
14:35:30B4gdermany people have done that on Linux/debian
14:36:13webguest95should i build the toolchain as root or as user? does that matter?
14:36:26B4gderit doesn't matter
14:37:30 Join mirak [0] (
14:37:49webguest95ok, ill try some more and get back if i dont get it working, thanks
14:38:05preglowwebguest95: try apt-get install build-essentials
14:38:17webguest95really stoked about seeing rockbox running on my ipod video...
14:38:43preglowunless i'm misinterpreting the error message there...
14:38:56LinusNleftright: i think i know what it is, hang on
14:38:58preglowwhich i probably am
14:41:04 Join Zagor [0] (
14:48:27Paul_The_Nerdpreglow: It's still being very weird at me. If I assign the return from opto_keypad_read() and output it things are fine. If I ever compare it, things are not fine, and it doesn't boot into retailOS.
14:49:28miraka guy helped me to debug libmpeg2. There is something that I didn't exepected .
14:49:31mirakDEBUGF("mpeg2_header_sequence mpeg2dec_t size %d\n",sizeof(mpeg2dec_t));
14:49:53miraksize of mpeg2dec_t is not the same when I call it from two different place
14:50:10mirakthough the included headers are the same
14:50:27preglowPaul_The_Nerd: try comparing the size difference of the code when you just add a simple compare
14:50:34preglowPaul_The_Nerd: to see if gcc does something very drastic when it shouldn't
14:51:17 Quit B4gder ("time to say moo")
14:52:03mirakI am not sure how is it possible to fix that
14:52:11Paul_The_NerdI actually just changed the assignment on i so that i=BUTTON_MENU became i=j (and j was given the value of opto_keypad_read() for my lcd_puts later)
14:52:59Paul_The_Nerdpreglow: It also seems that if I change i=BUTTON_MENU to i=0 it also ceases to work. (I don't have a rockbox.ipod on the hd. So, when it doesn't look for it, and intentionally boots into retailOS it's fine. If it does look for it, fails to find it, and then tries, it doesn't boot into retail os)
14:53:27Paul_The_Nerdthe lcd_puts was later, the j was assigned that value from o_k_r earlier.
14:54:53webguest95im a little bit scared about installing rockbox on my ipod video, how risky is it? what can go wrong?
14:55:33webguest95and i know of course that it cant play audio yet and im aware of that it is almost no meaning in instsalling it at this stage ....)
14:57:15***Saving seen data "./dancer.seen"
14:57:48Paul_The_Nerdwebguest95: On the plus side, it is *very* low risk. I believe the absolute worst you can do is have to resync any music/files over to it.
14:58:26preglowvery hard to break an ipod
14:58:34preglowyou more or less need to drop it
14:58:56preglowPaul_The_Nerd: well, this sounds like it'll be a fun problem
15:00:37Paul_The_Nerdpreglow: It's a willful little device. I'll give it that.
15:00:59preglowdamn little bastard is what it is
15:01:49Paul_The_NerdThat too
15:04:06Paul_The_NerdSonuva... I just don't get it.
15:04:06Paul_The_NerdAt all
15:05:16preglowno, me neither
15:05:54Paul_The_NerdI keep thinking "I'm overlooking something. Logic should allow me to work this out."
15:06:35preglowsure, but it's probably very low-level logic
15:06:43Membrillopreglow, do you know what mechanism apple have put on their ipod harddrives to prevent you using them on other DAPS?
15:06:55preglowMembrillo: you can't use them with other daps?
15:07:00Membrilloapparently not
15:07:07ender`ipod drives are somehow locked
15:07:25ender`apparently you can't replace the drive either
15:07:28Membrilloa guy on MR tried it apparently
15:07:45Membrilloi think you can replace it, but it has to be a legit apple drive
15:08:11ender`well, you can't replace it with an off-the-shelf drive
15:08:22Membrillowhich is dissapointing
15:08:37Membrillofecking apple. protection everywhere against everything
15:08:55preglowand for noe reason...
15:09:10ender`good reason - money, money and more money
15:09:20Membrillobut what other security features could you put on a harddrive beside a file sitting on it?
15:10:04ender`there are some ATA commands to set a password for the drive
15:10:05ender`no idea what actually needs to be done though
15:11:10Membrillothe drive wouldnt be physically different to a regular toshiba drive would it?
15:11:23 Quit aliask ("Chatzilla [Firefox 1.5/2005111116]")
15:11:24ender`no, it's a software (firmware) setting
15:12:16Membrillomaybe when the player is turned off the firmware locks it, and then unlocks it when you turn it on?
15:12:46Membrilloyou could possibly fix it by unplugging the HD while the player is on. although thats a bit risky
15:13:22Paul_The_Nerdpreglow: it gets MORE annoying
15:14:18Paul_The_NerdI can check if i==BUTTON_MENU and output Yes or No depending (using key_pressed to assign i) and then outside the if load retailOS
15:14:21markunMembrillo: I saw a website where they replaced the ipod drive by a 3.5" hdd, I don't think the drive was modified.
15:14:36Paul_The_NerdBut if I try to do it inside it, it fails.
15:14:44Paul_The_NerdLike, if I replace the output of yes with it.
15:15:00Membrillomarkun: thats interesting
15:15:05Membrilloill look around
15:16:11Membrillo "I had a 40GB Toshiba MK4004GAH"
15:16:14markunMembrillo: look at "Preparing the Hard Drive".
15:18:05markunender`: where did you read about the locked ipod drives?
15:18:51preglowLinusN: did you forget to remove debug.h from that commit? :)
15:18:53Membrillohmmm so stored in one of those partitions must be a harddrive locking system
15:19:16LinusNpreglow: yes, but that doesn't matter
15:19:35MembrilloStep666 on mistic river said: There was a thread around here a while back, in which a guy had bought an Apple-branded Toshiba drive (basically, one out of an iPod) and it wouldn't work in his DAP. Apple do something to their drives that, unfortunately, makes them incompatable with other players.
15:19:38preglowno, i can't say i see that hapening
15:19:47ender`markun: i've seen it on some forums, and one or two webshops that sell 1,8in hdds had "This drive might not work in ipod"
15:20:54markunender`: maybe they can't get it to work because of the special partitions
15:21:10Membrillothat sucks. I was going to buy a dead ipod and use the HD in my iriver
15:21:20ender`that might be possible, too - i don't have an ipod
15:21:35Membrilloi have an ipod video, but im not taking that apart
15:21:41ender`i've seen references to the ata lock feature somewhere, but those might have been just guesses
15:21:58Membrilloanyway, im off. thanks for the info guys
15:22:07Membrillothat super ipod article was intersting
15:22:13 Quit Membrillo ()
15:26:21 Quit dwihno (Remote closed the connection)
15:26:53 Join dwihno [0] (n=dw@
15:29:51Paul_The_NerdIt has something to do with the loading of retailos being conditional.
15:30:08 Part LinusN
15:30:13Paul_The_NerdPreglow: As far as I can tell, you can do whatever you want as long as there's nothing conditional about whether it loads..
15:30:56mirakmarkun: hey
15:31:01mirakmarkun: you tried the thing ?
15:31:03markunhi mirak
15:31:08markunno, didn't try it yet
15:32:06preglowwell, now that is kind of weird
15:32:24preglowi'd very, very much love to have a hardware debugger right now
15:33:15 Quit Febs (Read error: 110 (Connection timed out))
15:33:51mirakmarkun: I got some help from a guy on #c, he told me to check if the type was the same by cheking the size of the type and he was right it's not the same. that's bizarre
15:34:31Paul_The_NerdWell, at least it doesn't have anything to do with how input is handled.
15:38:25Paul_The_NerdThis is the point where I very definitely don't know enough. At all.
15:38:43markunmirak: even stranger that it worked when you wrote the first test
15:40:16 Quit KN|stiff ("( :: NoNameScript 3.81 :: )")
15:41:31preglowi'm not exactly used to code behaving like it's insane either
15:41:51Bgerhow someone could reinit the mouse under X ?
15:42:15mirakmarkun: well it works on pc but not in the sim or rockbox
15:42:26mirakmarkun: it never worked in rockbox
15:44:05preglowPaul_The_Nerd: the last time something behaved completely batshit crazy on ipod, the coprocessor was the culprit
15:44:14preglowbut this time it should be soundly asleep...
15:46:34Bgerreally, is there a way to reinit a PS/2 mouse under X ?
15:46:44Paul_The_Nerd i=key_pressed();
15:46:44Paul_The_Nerd entry = tblp->addr + tblp->entryOffset;
15:46:44Paul_The_Nerd if (imageno || ((int)tblp->addr & 0xffffff) != 0) {
15:46:44DBUGEnqueued KICK Paul_The_Nerd
15:46:44Paul_The_Nerd memmove16(tblp->addr, tblp->addr + tblp->devOffset - padding,
15:46:44Paul_The_Nerd tblp->len);
15:46:44***Alert Mode level 1
15:46:44Paul_The_Nerd }
15:46:46Paul_The_Nerd if(i==BUTTON_MENU)
15:46:48Paul_The_Nerd return entry;
15:46:54Paul_The_NerdAfter that, the rest of the code to load linux/rockbox is normal
15:47:01Paul_The_NerdI moved them before it, out of curiosity.
15:47:05 Part leftright
15:47:09Paul_The_NerdIf you take if(i==BUTTON_MENU) out, everything is fine.
15:47:14 Quit Bger ("brb")
15:47:52Paul_The_NerdOtherwise, it'll load rockbox but not retailOS.
15:48:06Paul_The_NerdAnd I've yet to find a way to get retailOS to load on a missing rockbox.ipod
15:48:25mirakPaul_The_Nerd: use a pastebin
15:48:26preglowi don't know whether working around it like that is very clever
15:48:43preglowi expect we'll skip into this problem again sooner or later
15:48:50preglowperhaps we should try with gcc4
15:48:51Paul_The_NerdI wasn't going for a permanent solution, so much as just seeing if I could force something to happen.
15:48:57mirak Paul_The_Nerd
15:49:23mirakthe code looks nicer
15:53:30Paul_The_Nerd <−−- If you comment out the second line, it doesn't work. Otherwise, it loads the retailOS.
15:53:33 Join Bger [0] (n=Bager@
15:54:18Paul_The_NerdAh well.
15:54:53Paul_The_NerdAt least some other possibilities as to what *are* causing it have increased unlikeliness.
15:55:57Paul_The_NerdBut it irritates me that I can't cheat past it.
15:56:06preglowthat might still remove the code to key_pressed
15:56:45***Alert Mode OFF
15:56:53preglowi'll probably need to disassemble the code to see if key_pressed is called
15:57:15Paul_The_NerdWell, I tried putting an if statement in between the two lines, if(i==BUTTON_MENU) lcd_puts(0,line++,"Pressed"); and with the button pressed or not, retailOS still loaded.
15:57:23preglow'cause if it is hardware state that isn't working out for us, then it gets a whole lot easier
15:58:02preglowso you do get a proper button reading, and still manage to load retail?
15:58:18Paul_The_NerdThat was about what I said.
15:58:27preglowi'll switch to gcc4 afterwards and see if that works better
15:58:34preglowif i have time
16:02:51Paul_The_NerdTo try that would I need to do anything special, or just get gcc-4 source and build the appropriate crosscompiler, and replace my current ver with it?
16:04:13preglowthat <-
16:04:28preglowdon't need to replace it, just make sure it's before the regular one in your pathj
16:05:08Paul_The_NerdWell, I was planning on just moving the folder for 3.4.5 out of the path, then installing 4.0.2.
16:05:16preglowsounds neat
16:05:47Paul_The_Nerd'cuz apparently I'm not sleeping, so I may as well give it a shot.
16:06:12preglowapparently you're not
16:06:31Paul_The_NerdWell, I need an all nighter to fix my schedule before class resumed, so this works out nicely, in a way.
16:09:02Rob2222_Im trying to compile a H320 version, first time for myself, but i get: dsp.c:476: error: can't find a register in class `ADDR_REGS' while reloading `asm'
16:09:07Rob2222_anyone know why?
16:09:54preglowyou're doing a debug build
16:10:20pregloweither that or ifdef that part of dsp.c out
16:10:40Rob2222_no, i was building a debug version, youre right
16:11:01preglowif you don't need it, you might as well do a normal build
16:11:03Rob2222_thought the player will set up a logfile with interesting debug infos.
16:11:27preglowwell, unless you're a coder, you don't need it
16:11:40preglowand if you are a coder, you should have no problems with ifdeffing away the offending asm part in dsp.c
16:11:51Rob2222_i want to code... thats cause i set up the cross compiler... ^^
16:12:00XavierGrHaha Paul_The_Nerd kust like me. My schedule is so farked up that once in a while I need an all-nighter to fix it. :)
16:12:12Rob2222_at least i want to start get into it
16:12:23Paul_The_NerdXavierGr: Exactly.
16:13:18Rob2222_preglow: ah ok, theres a wring ifdef in dsp.c
16:14:01Rob2222_but i didnt got, whats the differend between debug und log option. will have a look at it
16:14:29preglowa log build is one kind of a debug build
16:15:35Rob2222_debug with an extra log file i think?
16:19:44 Quit aegray ("leaving")
16:30:36 Join _FireFly_ [0] (
16:34:17 Join akaidiot [0] (
16:35:15 Nick solexx_ is now known as solexx (
16:35:55 Join ripnetuk [0] (
16:38:03Rob2222_why the hell my "normal" self build RB from CVS is freezing on my H300?!
16:39:00Rob2222_from cvs without changes
16:40:03Paul_The_NerdWhat's it doing?
16:42:24Rob2222_starting, and freezing after ca 15 seconds.
16:43:27Rob2222_ill trying a rebuild
16:43:35Paul_The_NerdOdd. I know my builds started giving me IllInstr freezes (it wouldn't load rockbox at all, as soon as it left the bootloader it happened.) I upgraded to GCC 3.4.5 and it stopped. (Well, I ended up doing a new cygwin install)
16:47:20 Quit ripnetuk (Remote closed the connection)
16:47:57Rob2222_I'm thinking about, if it would work, to boot different rockbox versions regarding which button combination is pressed at startup. Maybe UP-PLAY for /.rockbox1/rocbox.iriver
16:48:40Rob2222_but would the plugions then still access /.rockbox/...
16:49:12Paul_The_NerdWell, why would you want different versions?
16:49:25CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:49:25*amiconn can't imagine what this should be good for
16:50:43Rob2222_atm i like the H300-opimized builds for using
16:51:08Rob2222_but i would like to start my own programming on the cvs version
16:55:54_FireFly_Rob2222_: you can load other version of rb within rb
16:56:11Paul_The_NerdYeah, if you have other .iriver files, you can click on them, no?
16:56:19Paul_The_Nerdor .whatever
16:56:25Paul_The_NerdDepending on your platform
16:56:47Rob2222__FireFly_: yes, but how should i take the directory structure for that?
16:56:54Rob2222_.rockbox for the main
16:56:57_FireFly_only if the new firmware file isn't incompatible to the codec and plugins in .rockbox
16:57:02Rob2222_and .rockbox1 for the 2nd?
16:57:13_FireFly_no only .rockbox
16:57:16***Saving seen data "./dancer.seen"
16:57:45Rob2222_so even if the .iriver file is in maybe .rockbox1, all accesses going to .rockbox, right?
16:58:04_FireFly_because the path is hardcoded
16:58:16amiconnThen leave your standard-use rockbox.iriver as-is, and copy your test builds as <something>.iriver
16:58:19Rob2222_where is the configuration stored? language etc?
16:58:31 Quit Zagor ("Client exiting")
16:58:32amiconnYou can the RoLo <something>.iriver for testing
16:59:14Rob2222_amiconn: will the CVS builded version run without problems in a H300 optimized .rockbox/?
16:59:43Rob2222_only the plugins/viewer maybe incompatibility
17:02:05Rob2222_ok, ill simply try.
17:03:13 Join BHSPitLappy [0] (
17:03:31 Quit Kohlrabi (Nick collision from services.)
17:03:37 Join Kohlriba [0] (
17:04:23Rob2222_are the preferences are stored on the disk?
17:07:22 Quit markun ("leaving")
17:07:55 Quit wubbla (Remote closed the connection)
17:08:26 Join wubbla [0] (
17:11:36_FireFly_Rob2222_: the configuration is stored in a block on the hd in track 0
17:12:35Paul_The_NerdPreglow: Victory! Kinda...
17:12:58Paul_The_NerdWhen built with GCC 4.0.2, if rockbox.ipod is not present, it goes properly into the Apple FW
17:13:30Paul_The_NerdBut, I can no longer time my button presses properly, it seems, as if it's present, I can't seem to get it to go into anything BUT rockbox.
17:14:06 Quit webguest95 ("CGI:IRC (EOF)")
17:15:10Paul_The_NerdAlright, my fingers have improved. It works preglow.
17:18:05 Quit BHSPitLappy2 (Read error: 110 (Connection timed out))
17:18:43 Join BHSPitLappy2 [0] (
17:22:58 Quit YouCeyE ("Leaving")
17:24:20Rob2222_hmm, whats a hunk?
17:25:08lostlogica section of a patch
17:25:12 Quit BHSPitLappy (Connection timed out)
17:25:35Rob2222_ah hi lostlogic
17:25:36preglowPaul_The_Nerd: ok, so we've got a bloody compiler bug on our hands
17:25:41preglowain't that just dandy
17:25:54preglowi'll take that as a bona fide sign for me to upgrade
17:25:54Paul_The_NerdAt least it can be made to work.
17:25:59preglowand see what other differences there are
17:26:50Paul_The_NerdWell, the .bin was very slightly smaller
17:27:28preglowthis gets better
17:28:02 Nick Rob2222_ is now known as Rob2222 (
17:44:18 Nick Rob2222 is now known as Rob2222_ (
17:45:00 Join Moos [0] (
17:45:48 Quit DJDD__ ("Trillian (")
17:47:44 Nick Rob2222_ is now known as Rob2222 (
17:53:13Paul_The_NerdPreglow: Is there something I need to do besides patching it with that audio.diff to get music playback in its current not-so-happy state?
17:59:57Paul_The_NerdHrm, somehow I didn't rebuild properly.
18:00:56Paul_The_NerdNevermind then. Got it.
18:02:00 Quit Jungti1234 ()
18:05:50 Join DangerousDan [0] (
18:10:12 Join darkless [0] (
18:10:54 Join daire [0] (
18:15:41 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
18:15:48Paul_The_NerdYes daire?
18:16:16 Join DreamTactix291 [0] (
18:17:44dairesorry new to irc, nobody talkin?
18:18:12Paul_The_NerdThere's often periods of quiet in here.
18:18:28 Quit daire ("Disconnecting")
18:24:29 Quit tvelocity ("Leaving")
18:25:07 Join Mmmm [0] (
18:25:28MmmmAmiconn...are you there?
18:27:33Mmmmamiconn: you were interested in the recording screen gain method right?
18:31:15 Join Febs [0] (
18:37:32 Join ghode|afk [0] (
18:37:35 Quit Paul_The_Nerd ("Leaving.")
18:38:54 Quit Febs ("CGI:IRC (EOF)")
18:39:30 Join ender [0] (i=ychat@
18:40:22 Join Febs [0] (
18:42:15 Quit webguest67 ("CGI:IRC (EOF)")
18:43:54 Join linuxstb [0] (
18:47:08 Quit ender` (Read error: 110 (Connection timed out))
18:53:19linuxstbpreglow: Did I understand the logs correctly that upgrading to gcc 4 fixes the Nano bootloader problems?
18:57:20***Saving seen data "./dancer.seen"
19:03:17 Join webguest33 [0] (
19:08:42preglowlinuxstb: yes
19:08:51preglowthat is in accordance with my understanding
19:17:58 Nick ender is now known as ender` (i=ychat@
19:21:39 Quit akaidiot (Read error: 110 (Connection timed out))
19:24:00 Join Bger_ [0] (n=Bager@
19:36:38Papricahome/guest/rockbox-status/simu/apps/settings.o(.text+0x1b38): In function `settings_apply':
19:36:38Papricahome/guest/rockbox-status/apps/settings.c:990: undefined reference to `_set_bat_x_pos'
19:36:38Papricahome/guest/rockbox-status/simu/apps/settings_menu.o(.text+0x348): In function `baterry_x_pos':
19:36:38Papricahome/guest/rockbox-status/apps/settings_menu.c:264: undefined reference to `_set_bat_x_pos'
19:36:42Papricawhat it means?
19:39:49 Quit linuxstb ("CGI:IRC (EOF)")
19:39:50 Join birdfish [0] (
19:40:01 Quit birdfish (Client Quit)
19:40:57 Quit Mmmm ()
19:43:23 Join birdfish [0] (
19:44:21 Join SteL [0] (
19:51:54 Join webguest76 [0] (
19:58:46 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:06:14 Join Gibbed [0] (
20:06:18 Quit Rick (Nick collision from services.)
20:06:30 Nick Gibbed is now known as Rick (
20:09:17 Quit webguest76 ("CGI:IRC")
20:09:47BagderPaprica: it means you use functions that it can't find
20:10:00Bagderit == the linker
20:14:00Papricaso where i need to defined it?
20:14:25Bagderyou need to use functions that exist, or add the functions that don't exist
20:15:12Papricai add one
20:15:18Papricaand this s the message
20:15:48Papricai add it to statusbar.h
20:15:56 Quit birdfish (Read error: 104 (Connection reset by peer))
20:18:21 Join birdfish [0] (
20:19:09 Quit Febs ("CGI:IRC (EOF)")
20:21:16preglowBagder: so yeah, we might be needing a new gcc on the server :>
20:21:26preglowfor arm, that is
20:22:04Bagderlatest gcc 4 then?
20:22:22preglowbut no stress
20:22:26preglowi'll tell you when i'm sure
20:22:52preglowbut it compiling code that works sure is a good indication it'll happen
20:23:50 Join Mmmm [0] (
20:33:51 Join lyon8 [0] (
20:34:47lyon8thank you very much for doing this great work!!!
20:35:14 Part lyon8
20:35:49 Join webguest63 [0] (
20:35:51 Join matsl [0] (
20:36:09 Join Matze41 [0] (
20:41:59 Join webguest72 [0] (
20:42:07 Join einhirn [0] (
20:42:34webguest72join #banc
20:42:50 Part webguest72
20:44:43preglowyes, yes that's definitely going to happen
20:49:21preglowjoining #banc
20:52:04webguest63the mp3 encoder seams to be realtime ready.. anyone tried to adapt it to recording?
20:52:47Bagderwe've been waiting for you to do it! ;-)
20:53:39preglowit is ready and good to go
20:53:44preglowwe just need a recording codec api
20:53:49preglowand i believe someone started working on that
20:54:31preglowand i've got a couple more ideas for optimising as well, so it'll be faster
20:55:42 Join Gibbed [0] (
20:55:44 Quit Rick (Nick collision from services.)
20:55:56 Nick Gibbed is now known as Rick (
20:57:24***Saving seen data "./dancer.seen"
20:58:56webguest63preglow: cool
21:00:38Mmmmamiconn: you fancy trying out a patch? It sorts out the recording gain adjustments:
21:02:48MmmmI reckon that's how it should be... what do you think?
21:02:55 Join akaidiot [0] (
21:04:27 Quit kernelsensei (Read error: 104 (Connection reset by peer))
21:04:32 Join kernel_sensei [0] (n=boris@
21:04:43 Nick kernel_sensei is now known as kernelsensei (n=boris@
21:10:58 Quit kernelsensei (Remote closed the connection)
21:11:00 Join kernelsensei [0] (n=boris@unaffiliated/kernelsensei)
21:11:57*amiconn wants bigger fonts
21:12:14amiconn..way bigger fonts, for in-car use of H3x0
21:12:57*lostlogic pokes and prods devs to give feedback about or merge several of his older patches (profiling, ab repeat, sokoban level cache (or the later combined version), numeric file sort, RTC in WPS) and new better battery smoothing ... please?
21:13:23 Quit Matze41 ("Miranda IM! Smaller, Faster, Easier.")
21:13:31preglowahh, yeah, there's the profiling patch
21:13:35 Join tucoz [0] (
21:14:35*tucoz just read what lostlogic wrote and thinks that rockbox might just need another h3x0 developer with cvs access
21:14:55lostlogic(it only matters to me because the more work I do the harder it is to extract the particular stuff I'm working on from my development tree :-\)
21:14:59preglowyes, we do need that
21:15:05XavierGrSomeone that will look to patches definitely.
21:15:18preglowbut lots of this isn't even h3x0 specific
21:15:51lostlogicI volunteer... if I have the requisite whatevers you are looking for...
21:16:49tucozWell, not only this. But the ever increasing volume of patches and plugins for the colour targets needs a person with time and skill to review those.
21:17:35tucozof course, some of the plugins need to be useful on the other targets as well.
21:17:53*Moos is looking at lostlogic side :)
21:17:53Bagdermost of those patches are drive-by patches that aren't very useful
21:18:06 Join NightCat [0] (
21:18:12NightCatHi all!
21:18:21tucozOh, I see. Well, no crap should be commited :)
21:18:21lostlogicBagder: 'drive-by'?
21:18:35BagderI'm not referring to your patches, lostlogic
21:18:42tucozas in hacks, right?
21:19:00lostlogicBagder: phew! :)
21:19:18Bagderas in quick changes that work for one target, possibly break others and that are done in a way we don't approve
21:20:27NightCatI want to know basic constructions (like round, trunc, div, mod, case, repeat, while, for) shoud I look only for books or I can use C++ books?
21:21:06Bagder ?
21:21:11XavierGrYes but that's the point. I have 3 patches pending and I didn't got one simple response as to if these patches can be commited. Some patches are indeed drive-by but others are not. Someone must be able to look each one and inform the other devs if the patch is worthwhile considering.
21:21:38NightCatBagder: Yes. RB uses C. No?
21:21:40Bagderyes, we need more people to review and comment patches
21:21:53BagderNightCat: yes
21:21:55tucozXavierGr, like your battery plugin?
21:21:59Bagderis related to C?
21:22:14MoosBagder: lostlogic is volunteer :)
21:22:18lostlogicyeah, I've tried to comment on some to help people along, but of coruse I'm not a dev so my comments are just general patch etiquette stuff
21:22:18XavierGrAlso current Devs have too little time for patches and I cant every once in a while ask them too look something else.
21:23:07NightCatBagder: oh sorry is C :)
21:23:29NightCatBagder: My keyboardswitcher works fine :)
21:23:51tucozAnyway, the way things have progressed in the past (regarding devs with cvs access), will probably work fine in the future as well.
21:23:57NightCatBagder: And what you mean by "Yes"? I can use C++?
21:24:26XavierGr1) Battery benchmark, (Bger said he will look into this) 2) JPEG filescroller (I have done this 3 times with the lates beeing the best) 3) Multiple FM presets handling. All of them was commented as a nice addition (when I programmed them) but some remain in the tracker for more than 6 months
21:24:27BagderNightCat: Rockbox is C and we prefer C
21:25:01NightCatBut I think the basic constr. is the same/
21:25:22MoosXavierGr: I'm sympathize :(
21:25:24lostlogicNightCat: the basic syntax is the same, but there are definitely differences between the two.
21:25:32tucozXavierGr: regarding 2). Sounds like a good thing that you did it over again ;)
21:26:21Moostucoz: isn't you worked on latex manuel?
21:26:27NightCatlostlogic: I just can choose the book to leanr C...
21:26:27XavierGrYes the first implementation had many bugs. The last one takes all the files from the dir buffer, so no more big arrrays and such (or memory problems)
21:26:33tucozMoos: Yes, but I have very little time now.
21:27:08Moostucoz: not a complain, just wondering if it was you ;)
21:27:18tucozSo, the proposal is accessible from the webpage, but I do not have time to work on it for some weeks now.
21:27:45lostlogicNightCat: not sure I follow −− if you learn C++, you will have a big headstart on learning C, but a lot of the stuff taught in C++ beginning classes will also not directly apply basically any library functions are different between the two, the syntax is the same, there are also differences in argument passing (pointer vs. reference)
21:27:55Moostucoz: like other patchs mentioned, this usuful thing is pending
21:27:57tucozMoos: I got that. Just wanted to tell you why I haven't worked on it for a while.
21:28:08 Part webguest63
21:28:18tucozMoos: you know latex right?
21:28:28MoosNo :)
21:28:33 Quit webguest33 ("CGI:IRC")
21:28:45NightCatI find many books about C++ but still can't find C book (or FAQ or something it doesn't matter...)...
21:29:19Moostucoz:but when this thing will be commit, maybe that will interest people to "finish" it
21:29:23NightCatI just want to find some info about basic C...
21:29:23 Join lImbus [0] (
21:29:32lImbushi all
21:29:43tucozMoos: sure. It really isn't any harder than editing the wiki.
21:30:13NightCatCause I don't know C but I have programming experience in Delphi (Pascal), PHP...
21:30:44tucozNightCat: Maybe you can find something interesting at this page:
21:32:36tucozor maybe this:
21:33:06NightCatThere are a lot of C++ and C# :)
21:33:39NightCatOh the second is something like what I want...
21:33:52Bger_NightCat why don't just search for a "brief c tutorial" or something similar, if you already know delphi ...
21:35:21tucozbye bye
21:35:24 Part tucoz
21:35:34NightCatI am searching it right now but in russia :)
21:35:46Bger_in russian ?
21:35:47NightCattucoz bye
21:35:50 Join webguest22 [0] (
21:35:57NightCatBger_: Yes
21:36:08NightCatBger_: I'm an russian one :)
21:36:20webguest22hi all, what's the default behavior of an iPod with rockbox on it when you start it?
21:36:32webguest22does it boot into apples original fw?
21:38:40 Quit einhirn (Read error: 104 (Connection reset by peer))
21:38:57preglowlostlogic: the profiling patch in the tracker good to go?
21:39:43lostlogicpreglow: the latest version on the tracker is what 'should' be committed −− with parsing scripts in perl and in the build system and all that.
21:39:56amiconnFixed my silliness :)
21:40:06preglowi'll have a look at it right now
21:40:24preglowno java left?
21:40:32 Quit Kohlriba (Read error: 104 (Connection reset by peer))
21:40:33lostlogicpreglow: right
21:41:29webguest22does it work like the iriver, that you hold a button to boot into rockbox? or reverse?
21:43:16NightCatOh YES I founded russian C tutorial :) If someone iterested:
21:43:48 Join Midgey34 [0] (
21:44:12lostlogicpreglow: as I think you were discussing with someone, it still ICEs if enabled for 2 plugins :-\
21:45:05lostlogic*codecs (faad and alac)
21:45:10Bger_NightCat i think the first part (1) will be enough for you
21:45:43Midgey34hmm, anyone care to explain cvsignore?
21:46:06lostlogicMidgey34: files CVS ignores to not clutter up it's output with built files
21:46:31lostlogicfiles that should never end up in the CVS repository
21:46:36preglowlostlogic: yeah, not much we can do about that
21:46:38Bger_NightCat why not start reading first something like this
21:46:39preglowgcc is flakey
21:46:52Midgey34so if I put apps/plugins in .cvsignore will it not try to update the plugins dir?
21:46:56preglowwe can always hope the gcc people were true when they said coldfire is going to see some attention now
21:47:30lostlogicpreglow: I hope!
21:47:54lostlogicMidgey34: not sure if it works on directories like that, or on files that are already in the repo, sorry.
21:48:20 Join Kohlrabi [0] (
21:48:45pregloweven so, it'll be yonks before we see any of those improvements
21:48:57NightCatBger_: Cause it's an english tutor...
21:49:49NightCatBger_: I know some english but it's much better to read on my native language :)
21:50:31NightCatBut I don't make my choise yet :)
21:54:04XavierGrIs there such a think as a logic function minimizer (or somethink that will output the truth table of a logic equation)?
21:54:16 Quit webguest22 ("CGI:IRC (EOF)")
21:55:07Bger_XavierGr ?
21:55:58lostlogicXavierGr: yes, such things exist... but you could just amke a kernot(sp???) map for it and do it by hand :-D
21:56:07 Quit solexx (Read error: 104 (Connection reset by peer))
21:56:51XavierGrlostlogic obviously, on some occassion is very difficult to follow kernot's map.
21:57:02 Join solexx [0] (
21:57:22XavierGralso it would be way too faster to do it on PC.
21:57:53XavierGrlostlogic: Do you know names of such programs?
21:58:19lostlogic google is your friend
21:58:50XavierGrwell I was searching but I wasn't sure for what to search.
21:59:42lostlogicXavierGr: believe it or not, your phrase "logic function minimizer" seems to do the trick
22:00:11XavierGrI used it here but not on google, lol
22:00:44lostlogicI hate to support evil (or not so evil) empires, but google is the best damn search engine :-\
22:01:12XavierGrI agree
22:01:34XavierGrThough I haven't use anything else for my searches.
22:04:21preglowlostlogic: don't know if we should modify all these codecs for profiling when only a few devs will ever need it. it's not that much work to modify a codec for profiling when you want to, no?
22:05:11 Join elinenbe [0] (
22:05:31lostlogicpreglow: that's up to you −− the changes don't cause any harm when profiling isn't activated, actually make the return path of the codec better by my standards.
22:06:51preglowthe return path change i was thinking of keeping anyway
22:07:06preglowbut ok, how does one profile plugins?
22:07:14preglowis that included?
22:07:17amiconnAnything I should know?
22:07:35amiconn(I'm currently looking into the loader adjustment for codecs)
22:07:41lostlogicyeah, it's not activated on any plugins, but I made the header and C file needed to use it to profile part or all of a plugin
22:07:56lostlogicyou'd just have to turn on the profiling system and call the profile start and profile stop functions int he plugin api
22:08:08preglowamiconn: doesn't look like it
22:08:46preglowalso, calling a variable errno isn't very clever :-)
22:09:19preglowwe're probably never going to have any conflicts, but
22:10:06lostlogicdid I add that, or just use it when it was already used? You're quite right.
22:10:18preglowadded it inside of plugins
22:10:32preglowehh, codecs
22:10:45preglowlike i said, nothing major
22:10:53lostlogicyeah, that should be named retval or something of the sort
22:10:59preglowjust that i don't like calling things errno, being a "reserved" name and all
22:11:36NightCatCould anybody of you add to the plugins API function to get sound levels of playing song?
22:11:45preglowalso, the header for profile_plugin.c is wrong
22:11:53preglowas in the file header, it credits amiconn
22:11:59preglowand says it contains lcd routines
22:12:06amiconnhehe :)
22:12:53lostlogicpreglow: are you fixing some of this and going to commit, or do you want me to make the changes and send a new patch?
22:13:11preglowprefer it if you fixed it, it is your stuff after all, and i don't want to mess anything up
22:13:24preglowjust reading the patch and mentioning odds and ends i'd like fixed before it's commited
22:13:34lostlogicyep yep, will make fixes.
22:14:13 Join eagle730 [0] (
22:14:49 Join webguest61 [0] (
22:14:50 Quit eagle730 (Client Quit)
22:14:58preglowwill second hand id strings mess anything up, or will cvs just replace it with a new one for a new file?
22:15:07lostlogicjust replaces.
22:15:36 Quit webguest61 (Client Quit)
22:15:46preglowmokay, we might have a license problem here
22:16:02 Join webguest01 [0] (
22:16:29lostlogicwith the fact that I used the GCC gprof code for a basis?
22:16:33 Quit webguest01 (Client Quit)
22:17:23 Join jonnybe [0] (
22:20:23preglowBagder: profile patch, line 1231, if you've got time
22:20:36preglowBagder: compatible? looks like bsd with advertise clause to me
22:20:49lostlogiccrikey −− how's it in GCC then?
22:21:00preglowi have no idea
22:21:23preglowi'm also not an expert on licenses
22:21:27preglowwhich is why i ask the one who is :)
22:21:40 Part jonnybe
22:21:59 Quit akaidiot (Connection timed out)
22:22:03preglowif it's in gcc, it does sound a mite unlikely, yes
22:27:20preglowshouldn't the configure option only be called Profiling?
22:27:29preglowit's not like it's codec specific
22:29:39*preglow finds out his own private tree isn't the best tree to apply patches to...
22:39:01preglowmy gcc4 experiment most surely went good
22:39:03*preglow kicks freebsd
22:39:48 Join uski [0] (
22:39:52uskihi !
22:40:04 Quit NicoFR ()
22:41:20*amiconn wonders why some codecs define both rb and ci
22:41:32amiconnThey point to the very same struct ...
22:41:46preglowbecause of DEBUGF, mainly
22:41:56preglowwhich for some numbskull reason is hardwired to use rb
22:42:09preglowi've changed those i can to use ci only, i think
22:42:38amiconnFor codecs it can be re-hardwired to ci
22:42:53preglowor it can take the struct as a parameter, like the good lord would want us to
22:43:10amiconnCodecs use the definition in codecs.h, plugins use plugin.h
22:43:25amiconnWhy codecs.h defines them using rb-> is beyond me
22:43:27preglowi'm not very much a fan of just assuming a variable name
22:43:31 Quit thegeek (Read error: 110 (Connection timed out))
22:43:38 Join Skydemon [0] (
22:44:35preglowpatch for id3v1 problem just came in
22:45:07 Quit Moos ("Glory to Rockbox")
22:45:20lostlogicpreglow: updated patch with retval and that header fixed.
22:45:54preglowgood, i'll check out a fresh tree and try it
22:46:19 Join Moos [0] (
22:46:33BHSPitMonkey_hey preggerz
22:47:35*amiconn hopes he didn't mess up something
22:47:47amiconnBlind conversion of codec loader...
22:48:01lostlogicooh! excitement!
22:48:05lostlogic*gets ready to build*
22:48:10 Join RotAtoR [0] (
22:48:32preglowlostlogic: probably means you're going to need to sumbit a new patch again, heh
22:48:47amiconnNot yet committed, patience...
22:48:50preglowahh, right
22:49:57lostlogicpreglow: if necessary, that's fine, I have profile already separated out into it's own tree because it's big and ugly.
22:52:30 Join Adam_ [0] (n=Adam@
22:54:23*amiconn wonders why libfaad is rebuilt when only rockbox core stuff was changed
22:54:48*Adam_ wonders what amiconn is talking about
22:54:54preglowamiconn: deps broken
22:55:05preglowfaad's been riddled with dep problems since commital
22:55:11 Join thegeek [0] (
22:55:16Adam_well I have consolidated the H10 information into one section on this page:
22:55:26Adam_but now I need someone to dig in... any takers?
22:57:25***Saving seen data "./dancer.seen"
23:00:02 Quit uwe_ (Read error: 110 (Connection timed out))
23:00:37Adam_guess not...
23:00:45 Join uwe_ [0] (
23:01:01 Join Paul_The_Nerd [0] (
23:01:04 Quit Bger_ ("[BX] That's Mister BitchX to you")
23:01:12 Quit matsl (Remote closed the connection)
23:01:23preglowlostlogic: when is profile.out written?
23:03:25 Join petur [0] (
23:03:46 Quit NightCat ()
23:04:17lostlogicwhen the profstop is called
23:05:16preglowyou've gotta give me a quick lowdown on how to run the reader here
23:05:24preglowseems i have no clue
23:05:49lostlogicdid you run it w/o arguments to get help?
23:06:08preglowgot help, yes
23:06:36preglowtry passing it maps and objects and all, just spots a ton of no symbol found
23:07:14lostlogicso it's /path/to/profile.out /path/to/ /path/to/libcodec.a[...] 0
23:07:26preglowahh, right
23:07:33preglowprofile.out is text?
23:07:34preglownice <-
23:08:03lostlogicyeah −− didn't make sense to me to print it out binary and have to parse that (the way gprof does).
23:08:21lostlogicI guess for applications with 1000s of functions being profiled it makes sense, but we don't have that situation on rockbox
23:08:27 Quit Adam_ ()
23:09:12preglowso profile.out libmad.a 0 ?
23:09:23preglowdoesn't resolve names properly if so
23:09:27preglowi just get addresses for symbols
23:09:27 Join webguest33 [0] (
23:10:28 Quit webguest33 (Client Quit)
23:13:03lostlogichmm, yes, that should work :-\ doesn't contain a whole lot of function name info
23:13:24lostlogicdidn't rebuild in between or anything?
23:13:32lostlogicthat's why it needs the .a file
23:13:38lostlogicthe map provides the section offsets
23:13:45lostlogicthe .a provides the symbol offsets from the sections
23:13:54preglowCalls: 00001900 Ticks: 00002367 Symbol: 0x32fd1e34
23:13:55preglowlines like that
23:14:26lostlogiccan you send me a zip build and the .out file to poke at?
23:15:04preglowsure, what about map and .a files?
23:15:24lostlogicyeah, skip the zip and send those instead :)
23:17:32 Quit Mmmm ()
23:17:47lostlogic403 on the .out
23:19:15lostlogicpreglow: must be an issue with paths or something...
23:19:19lostlogicCalls: 00000015 Ticks: 00000475 Symbol: mad_layer_II(layer12.o)
23:19:20lostlogicCalls: 00000405 Ticks: 00001583 Symbol: III_reorder(layer3.o)
23:20:04preglowthis is plain wrong
23:20:08preglowi was playing a layer3 file
23:20:38preglowsmall chances of mad_layer_II being called
23:21:16lostlogicit was only called 15 times?
23:21:28lostlogicCalls: 00239762 Ticks: 00009405 Symbol: III_overlap(layer3.o)
23:21:28lostlogicCalls: 01658402 Ticks: 00029651 Symbol: mad_bit_read(bit.o)
23:21:31lostlogicare the two most called
23:21:45preglowit should be called 0 times
23:21:54preglowmost called, probably
23:21:57preglowwhich have the most ticks?
23:22:04 Quit _FireFly_ ("Leaving")
23:22:55lostlogicCalls: 00001915 Ticks: 00036189 Symbol: synth_full(synth.o)
23:22:55lostlogicCalls: 00007600 Ticks: 00067052 Symbol: III_huffdecode(layer3.o)
23:23:21 Quit petur ("here today, gone tomorrow")
23:23:29preglowdoesn't look too unreasonable
23:23:52preglowbut yeah, why doesn't it work here?
23:23:56preglowdo you depend on some outside programs?
23:24:25preglowobjdump is used, i see
23:24:55lostlogicthat's the only one, and it should be standardized :-\
23:25:13preglowit's a version from 2002 here
23:25:24preglowmaybe that's what's breaking stuff
23:25:34amiconnDo you profile sim codecs?
23:25:51preglowprofiling target codecs now
23:25:52amiconnOtherwise I would expect the correct xyz-elf-objdump to be used
23:26:00preglownah, it's all elf
23:26:15lostlogicGNU objdump 2.16.1
23:26:15lostlogicCopyright 2005 Free Software Foundation, Inc.
23:26:18preglowbut yeah, okish point
23:26:32lostlogicand it's just the one from /usr/bin not m68k-elf bin
23:26:40preglowi'll try to change that
23:26:43preglowand see what happens then
23:26:48amiconnI'm no elf expert, but when I tried using standard objdump on cross-compiled targets, I got strange results
23:27:20preglowlostlogic: yup, that was your problem there
23:27:26preglowworks fine and dandy now
23:27:41lostlogicjust by upgrading, or by using m68k-elf-objdump?
23:27:53preglowby using the target objdump
23:28:02preglowhow to autodetect target is a nice question
23:28:15 Nick BHSPitLappy2 is now known as BHSPitLappy (
23:28:40amiconnpreglow: What's the input for the tool? profile.out and ???
23:29:11 Nick BHSPitMonkey_ is now known as BHSPitMonkey (
23:29:14lostlogicpreglow: it's an endianness issue
23:29:15preglowamiconn: profile.out, the codec map file and the codec library .a file
23:29:21lostlogicI think
23:29:38lostlogicpreglow: is your host system big endian?
23:29:44preglowlittle endian
23:29:50lostlogicmine is and so is m68k
23:29:58preglowoh? what're you on?
23:30:16lostlogic... athlon...
23:30:25preglowathlon is little endian
23:30:28lostlogicor maybe... it's just the tool being able to detect it?
23:30:40preglowathlon sure as hell is little endian
23:30:47lostlogicwhen I use host-objdump, it shows: bit.o: file format elf32-big
23:30:49preglowlots of changes since binutils 2.12
23:30:53amiconnpreglow: You're on amd64, right?
23:31:05preglowamiconn: nope, compiling remote
23:31:09preglowamiconn: pentium3, i believe
23:31:18Bagderat least 'file' can tell what system the elf is for
23:31:27preglowBagder: did you look at my license question?
23:31:39 Join ashridah [0] (
23:31:48Bagderthat gcc one?
23:32:11preglowprofile patch gcc one
23:32:32preglowlostlogic: still, i find it extremely weird that layer_II functions are being called
23:32:56Bagderpreglow: gcc is all GPL and LGPL so it should be fine
23:33:03preglowBagder: no, this was bsd
23:33:09preglowahh, right, yeah
23:33:25preglowbut there is a gcc license, requiring a copyright message distributed with binary releases
23:33:36preglows/gcc/bsd/ :&
23:33:40preglowi'm tired
23:33:50Bagderwhere is it (URL) ?
23:34:11preglowtop patch
23:34:12preglowline 12xx
23:34:14preglowyou'll see it
23:34:17preglowbig header chunk
23:34:58preglowlostlogic: small wonder, there's a layer2 file right after my test file
23:34:59BHSPitLappypreglow: anything exciting happening
23:35:07lostlogicpreglow: hah!
23:35:14preglowlostlogic: i should start making a playlist containing just one file for these tests...
23:35:20lostlogicyeah, that's what I had to do.
23:35:20preglowBHSPitLappy: no, looking at some patches right now
23:36:12 Quit t0mas ("Doei")
23:36:56preglowBagder: any opinions?
23:37:11Bagderpreglow: I think it is fine
23:37:26preglowokies, but what about the binary thing? why can we disregard that?
23:37:35 Part SteL
23:37:38lostlogicbecause gnu did it first? :-D
23:37:40mirakI asked that lattely, so I still ask in just in case. I call that line at two place in the code DEBUGF("mpeg2_header_sequence mpeg2dec_t size %d\n",sizeof(mpeg2dec_t)); but the sizeof value is different
23:38:04Bagderno, it only say we reproduce the copyright "in documentation and/or other materials provided with the distribution"
23:38:08preglowmirak: then you're doing something funky
23:38:10mirakhow is it possible that a type have his size changing ?
23:38:13BagderI think that is fine
23:38:22lostlogicbecause we include source?
23:38:32preglowwe don't include source with the distribution
23:38:36mirakpreglow: hum like what ?
23:38:45preglowmirak: absolutely no idea, but types don't change size, that's for certain
23:38:47 Join einhirn [0] (
23:39:09mirakthat's really incredible
23:40:14 Join TCK- [0] (
23:40:17 Quit BHSPitMonkey (Client Quit)
23:40:19Bagderthe license is basically the Modified BSD
23:40:25Bagderdeemed GPL compatible by FSF
23:40:30preglowthen hooray
23:40:34preglowall's fine
23:40:45 Join BHSPitMonkey [0] (
23:41:22preglowbut i can see a glass of whisky and a bed in my immediate future, so unless there's something pressing
23:41:32amiconnBagder: That reminds me... should we include the imporved mersenne twister (random.c), or not?
23:41:36Bagdergo in peace my son
23:41:42preglowgood, gnight
23:41:46lostlogicnight preglow
23:42:02Bagderamiconn: any significant improvement for us?
23:42:07Bagder(I don't remember)
23:42:19amiconnBetter seeding
23:42:39amiconnWe already discussed this some time ago, with no final decision iirc
23:43:01amiconnJust that sentence reminded me: "in documentation and/or other materials provided with the distribution"
23:43:14mirakin reality the headers of functions are useless ?
23:43:45mirakare they only usefull to create libraries ?
23:44:08Bagdermirak: what are you talking about?
23:44:53amiconnmirak: The size of a non-intrinsic type can change depending on which headers are included, because the type definition may change
23:45:06mirakI hadded a test fuction but with no header in the .h and I can call it from another .c file
23:45:21amiconnYou can try to deliberately redefine the type to something arbitrary, then gcc should complain about the clashing definitions
23:45:40amiconn..and hopefully tell you where the original definition came from
23:46:18Bagderor just compile with -E
23:46:18*amiconn has the new codec loader running in the sim and on h140
23:47:30mirakso for exemple if you use the type BLA in a struct X wich is of size 8, if you use the struct X in another .c file but that you redefined BLA to size 4, then the X struct will have a modified size ?
23:48:13mirakBagder: where do I put this -E ?
23:48:17 Quit Skydemon ("bye bye")
23:48:32Bagdergcc -E file.c > dump
23:48:44Bagderand preprocesses the file only
23:48:51mirakif the name are the same
23:49:04mirakhow can I know something have changed ?
23:49:12amiconn'Show ID3 info' really shows a lot, but the codec type is missing...
23:49:41Bagdermirak you get the preprocessed file, use search with a text editor and check
23:50:14amiconnImho the 'deliberate clash' method works faster
23:50:27Bagdermost probably, yes
23:50:41amiconngcc will (hopefully) tell you what line in which file contains the definition
23:51:06mirakamiconn: I am not sure about the clash method. How do you proceed ?
23:51:30 Quit ehntoo ("Leaving")
23:52:54 Join Skydemon [0] (
23:53:03mirakam wondering if I should include plugin.h also in the headers
23:53:08Moosamiconn: will you commit the mersenne twister?
23:53:16mirakshould I ?
23:55:19Moosamiconn: it could be cool if there aren't license conflict, is there?
23:55:57Bagdercool for what?
23:56:34Moosfor Rockbox
23:56:45Bagderwe already have a mersenne twister
23:56:52Bagderthis is a modified one
23:56:58Bagderor rather improved
23:56:59Moosan improuved one I mean
23:57:11Bagderso the question is, what difference will it make to the ordinary user
23:57:43amiconnWell, some people reported less-than-optimal shuffling
23:57:58MoosBagder: can improvment, not be good for the user?
23:58:01Bagderis that really such a huge difference with the new one?
23:58:04amiconnI'm using the improved mersenne twister in my local builds
23:58:10lostlogicmy dice plugin would love it :-D
23:58:11mirakamiconn: should I include rockbox api in the headears also or is it useless ?
23:58:19Moosamiconn: please commit it
23:58:41BagderMoos: well, some improvements aren't visible

Previous day | Next day