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

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

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

#rockbox log for 2007-03-06

00:00:33 Join BHSPitMonkey [0] (
00:04:41linuxstb_LinusN: Do you agree with removing lcd_update() from the bootloader printf() (unless a button is being held) ? I'm thinking of doing it for the ipods before the next release - apart from the cosmetic improvement, it also increases boot speed noticably.
00:10:19 Quit web-taz ("( :: NoNameScript 4.02 :: )")
00:14:22 Join BoD[] [0] (
00:14:27BoD[]Hi !
00:14:42BoD[]Hey ... the latest build for Ipod Video doesn't work !
00:15:05BoD[]It says "Data abort at 00030760" when playing an mp3
00:15:11BoD[]is it a known bug?
00:16:11 Join B4gd3r [0] (
00:17:16preglowyou need a new bootloader
00:18:24 Join rotator [0] (n=e@rockbox/developer/rotator)
00:18:52BoD[]oh :(
00:18:57 Quit roolku ()
00:19:24BoD[]and the procedure is not trivial, from what I remember
00:19:50preglowit is very trival
00:20:28preglowi assume that will serve you very nicely for this purpose
00:20:52 Quit bospaadje ("great minds run in great circles")
00:22:17 Quit lee-qid ("aufwiederbyebientotsayonara")
00:22:28 Join pearldiver [0] (
00:24:38BoD[]great !!
00:24:44BoD[]thank you very much it works :)
00:25:07BoD[]plus it's prettier now while booting (white text on black) ;)
00:26:35 Join bospaadje [0] (
00:26:37 Join merlin2049er [0] (
00:26:56preglowit'll be prettier yet when linuxstb removes the text altogether
00:27:33mcphailLooking at "sys/timeb.h", am i correct in thinking that I can't use "ftime()"? What is the preferred method for millisecond timings?
00:27:55linuxstb_Speaking of which, anyone have any comments on this patch to do just that?
00:28:00BoD[]hey... what does "recording" do on ipods ?
00:28:32 Part toffe
00:28:35linuxstb_It records via the line-in hidden in your ipod's dock connector.
00:29:06 Join toffe82 [0] (
00:29:12BoD[]ahh right
00:29:52BoD[]i thought it could use the 'speaker out' connector
00:29:55BoD[]that would be funny
00:30:49 Quit bluebrother (Read error: 60 (Operation timed out))
00:31:03preglowlinuxstb_: looks good to me
00:31:08BoD[]anyway thanks very much
00:31:14BoD[]rockbox still rockx
00:31:33zethCan you show pictures on a rockbox-ed nano?
00:31:33dan_alinuxstb_: Have you got any ideas why the old bootloaders might work for me but not for anyone else?
00:31:45zethif so what format should they be in
00:32:00BoD[]byebye all
00:32:04 Quit BoD[] ("rulz")
00:32:14linuxstb_dan_a: Have you tried the official binaries? The release mini/greyscale users are using is from back in January 2006.
00:32:35peturmcphail: use current_tick
00:32:35preglowzeth: jpeg
00:32:39linuxstb_zeth: There's a jpeg viewer.
00:32:49mcphailpetur: thank you
00:32:59zethany idea how I find out what size to put them in?
00:33:07dan_alinuxstb_: I used the official binary from, and the version on it was 20060404, as far as I could see
00:33:14preglowzeth: whatever you want
00:33:22preglowzeth: it does zooming, to some extent
00:33:24 Quit B4gder (Read error: 110 (Connection timed out))
00:33:30linuxstb_dan_a: Ah, OK. Maybe the 4g was later than the others.
00:33:37zethwhat is my native resolution?
00:33:40 Join bluebrother [0] (
00:33:40preglowi've got photos that are digicam size on it
00:33:46zethhow do I find that out?
00:33:50preglowthe native resolution is the ipod resolution, 176x132 or something
00:34:00preglowthat's probably wrong
00:34:17preglowno, it's right, as a matter of fact. yay me
00:34:17linuxstb_Sounds familiar to me.
00:34:20zethsounds right
00:34:39 Quit robin0800 ("Always try to be modest, and be proud about it!")
00:35:07 Quit Febs (Read error: 60 (Operation timed out))
00:35:30preglowso yeah, that gives a fullscreen picture, but anything will go, as long as you have enough memory to deocde it
00:36:04preglowhas anyone started on a multiformat viewer yet? :>
00:36:30 Quit obo ("bye")
00:36:54 Quit bluebrother (Client Quit)
00:37:13 Join Thundercloud [0] (n=thunderc@
00:37:47linuxstb_A gif viewer existed at one point - one of the "plugbox" plugins I think.
00:38:01preglowwho would have imagined that didn't end up in svn
00:38:59pixelmaI thought you could open some more formats with the Rockpaint plugin (but what do I know about colour screen targets ;) )
00:39:07dan_aDo the current bootloader problems affect both generations of iPod Mini?
00:39:39 Quit bun-bun (Read error: 110 (Connection timed out))
00:40:02 Join bonbonthejon [0] (
00:40:28preglowstill no idea why?
00:42:44 Join Nik [0] (
00:43:04Nikanyone know if you can have picture album type things on a video ipod for rockbox?
00:43:17dan_apreglow: No idea why, but we can hack around it by (1) not doing the memcmp(...'portalplayer'...) check AND (2) compiling with GCC4.1.1
00:43:21 Quit ender` (" # Basic IBM dingbats, some of which will never have a purpose clear # to mankind -- /usr/src/linux/drivers/char/cp437.uni")
00:43:35preglowNik: well, you can view .jpg pictures, so...
00:43:49preglowdan_a: well shit
00:43:56preglowi love it when it looks like a compiler bug
00:44:26preglowhad a look at the asm?
00:44:32dan_apreglow: Your idea of love must be different to mine!
00:44:45 Part Nik
00:44:56hcshe's a masochist, apparently
00:45:48dan_aI can write asm, but not read it. There was nothing screamingly, horribly, obviously wrong with it
00:46:18preglowdan_a: so the problem is either retailos being erronously detected in ram, or the entry point of it being somehow different than exptected?
00:47:02linuxstb_No, I think the memcmp is correct - but the act of performing the memcmp upsets retailos...
00:47:15preglowthat sounds really weird
00:47:19preglowis the cache enabled at that point?
00:47:38preglowhave you checked if retailos expects any magic in some of the scratch regs?
00:48:16preglowor in any of the regs, anyway
00:48:17dan_aI can't remember if cache is enabled, but even if the memcmp is performed after the return, or if apple_os.ipod is present it still causes a problem
00:48:54dan_a(in both those cases, memcmp shouldn't be performed)
00:49:09preglowhaven't we had compiler bugs with the bootloader before?
00:49:11linuxstb_preglow: I started to look at that, and it does seem to save a few registers on startup, but I couldn't find where they were being used again.
00:49:36mcphailcurrent_tick is just a global long?
00:49:49preglowi kinda think i remember me having some problems with gcc and the bootloader really early in the ipod port
00:50:07linuxstb_preglow: I would expect it's the same issue. I can't believe gcc has bugs with simple if() statements like that. I remember when I was working on the bootloader, it would randomly stop loading retailos, and then start again.
00:50:46dan_amcphail: Yes
00:51:02linuxstb_mcphail: Yes. But if you're writing a plugin, you access it via *rb->current_tick.
00:51:02preglowand luckily, i can't remember what the issue was
00:51:14preglowi've never tried disassembling retailos, though, so know little of what it expects
00:51:18preglowwhat happens when it fails, anyway?
00:51:22 Join Domonoky [0] (
00:51:40mcphaillinuxstb_: not for a plugin. Just trying to patch the scrollwheel sensitivity on my ipod-3g
00:52:01preglowyou'll need to patch more than that for good performance on 3g
00:52:16mcphailpreglow: :)
00:52:34 Join Peps [0] (
00:52:37mcphailpreglow: i think this will be quite simple...
00:53:03preglowjust wait until you try playing music
00:53:18PepsHi, I just installed rockbox on my ipod video (60gb)
00:53:19mcphailpreglow: yes, i noticed that wa a problem :)
00:53:25linuxstb_preglow: retailos just freezes without displaying anything.
00:54:07preglowlinuxstb_: have you looked at what differs in the call condition when doing the retailos jump?
00:54:17Pepsfirst problem: no theme shows the cover (music art). can anyone help?
00:54:17 Quit Thundercloud_ (Read error: 110 (Connection timed out))
00:54:20 Quit dpro (Read error: 110 (Connection timed out))
00:54:25dan_aIt doesn't freeze for me - the iPod reboots
00:54:53linuxstb_dan_a: Ah, OK. I'm confusing it with rolo :)
00:54:58 Part LinusN
00:55:07 Quit Arathis ("Bye, bye")
00:56:06 Join BigMac [0] (
00:56:33 Join perldiver [0] (
00:56:35BigMacHey is anyone able to access #ubuntu?
00:56:52preglowvery much so
00:57:11BigMacWell it must be this faulty webchat
00:57:27preglowwouldn't know
00:57:30preglowthis isn't one
00:57:47PepsA Q: do I need to keep adding files using gtkpod (I am running ubuntu) for them to play both in rockbox and appel firmware?
00:57:54preglowPeps: no
00:58:09dan_aPeps: Yes
00:58:10preglowPeps: yes
00:58:10linuxstb_preglow: No, I haven't dug that deeply. The bootloader .c file returns to crt0-pp-bl.S, which then jumps both CPUs to the start of DRAM. So the only difference will be what's left in the registers from the .c file, and isn't modified by crt0-pp-bl.S.
00:58:27preglowPeps: misread that, if you want to play in both, then yes you have to
00:58:37Pepshow do I add files in rockbox if I do not care about apple?
00:58:43preglowPeps: just copy them
00:59:20preglowlinuxstb_: yeah, figured that
00:59:37preglowlinuxstb_: and unless the cache is enabled, i can't see much else changing that the register values
00:59:56dan_aOh, swearwords! My new bootloader won't load OF again; it looks like the return doesn't return
00:59:57linuxstb_Maybe simply disabling the lcd_update calls will fix it...
01:00:11preglowit really wouldn't surprise me if the registers are used to pass stuff
01:00:35linuxstb_Nor me. So we've just been lucky so far?
01:00:57preglowtime to update the bootloader
01:01:02linuxstb_Maybe we should simply save the registers on entry, and restore them.
01:01:14preglowlinuxstb_: worth a shot, easy anyway
01:01:34linuxstb_Unless there is also info stored in IRAM, in which case that will need to be backed up...
01:01:41preglowwell, also easy
01:01:56 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
01:02:16preglowBuild bootloader file
01:02:16preglow/home/thomj/proj/rockbox-devel/builds/nano-bl/bootloader/bootloader.bin: Success
01:02:19preglowmake[1]: *** [/home/thomj/proj/rockbox-devel/builds/nano-bl/bootloader-ipodnano.ipod] Error 255
01:02:33preglowi'm used to stuff like that in windows, but...
01:02:54linuxstb_Is tools/scramble compiled?
01:04:23 Quit pearldiver (Read error: 110 (Connection timed out))
01:05:50 Quit merlin2049er (Remote closed the connection)
01:07:46***Saving seen data "./dancer.seen"
01:08:55 Join JdGordon [0] (i=82c20d65@gateway/web/cgi-irc/
01:09:01 Join merlin2049er [0] (
01:13:47JdGordonpetur: you said you couldnt exit the plugin browser?
01:14:12peturwith left
01:14:29peturthe menu button works
01:15:06JdGordonwhat target? (not that is shold make a diff)
01:15:40 Join Llorean [0] (n=Llorean@rockbox/administrator/Llorean)
01:15:45JdGordonah, with follow playlist?
01:16:35peturno idea
01:16:35*JdGordon rereads
01:16:47preglowwell, special
01:16:51preglownow my bootloader hung on usb
01:17:10preglowand there rockbox hung on usb
01:17:30preglowand there it fecking hang before the logo even showed
01:17:30preglowit's a first
01:18:44 Quit muesli__ ("ich will Khe!!!")
01:19:01preglowlinuxstb_: fancy-ass, now the apple logo even lights up a bit before rockbox boots, looks fancy :)
01:19:01 Join Aaron4 [0] (
01:19:29dan_aI'd test it if Linux didn't hate my 4G
01:19:55 Quit petur ("gtg")
01:20:29JdGordondoes anyone know if its possible to add a make tag to exclude plugins?
01:20:31BigMacdan_a:congrats again on the kocop committ, I have never been more happy with a clean sourced build
01:20:32 Quit Aaron4 (Client Quit)
01:21:12LloreanJdGordon: Just edit your SOURCES with an #if 0 encompassing the whole file.
01:21:16dan_aBigMac: Thanks. It didn't go quite as smoothly as I'd hoped, or as badly as I feared
01:21:31JdGordonllorean: yeah, but im after a more permanent solution :p
01:21:39LloreanThat's pretty permanent. :-P
01:21:53JdGordonit would also get commited if i did a svn commit :p
01:22:01dan_aJdGordon: Would "make bin" do it for you?
01:22:46JdGordonI want the bin + codecs... without the plugins
01:22:49 Join Febs [0] (
01:23:39|Rincewind|JdGordon: I think there already exits a make target to exclude the plugins
01:24:15*preglow bed
01:24:21JdGordonisnt make bin the closest?
01:24:48*JdGordon wants to kill whoever thought going to previous screen from the wps was a good idea ;p
01:25:26 Quit entheh ("^~")
01:25:58|Rincewind|JdGordon: I have some issues with the GO_TO stuff in my patch. There are some screens that don't use them (for instance playlist viewer)
01:26:45|Rincewind|it would be nice if every function that displays a screen and responds to buttons uses the GO_TO_xxx defines as return vaulues
01:26:54JdGordonit would....
01:27:24LloreanIsn't the playlist viewer a plugin though?
01:27:50|Rincewind|and there are a few inconsistencies regarding usb handling (on the gui side)
01:28:01|Rincewind|playlist viewer is in the core
01:28:27JdGordonLlorean: its in the core.. but uses the plugin ram which cases the problems
01:28:42|Rincewind|it is called with playlist_viewer() and returns true if usb connection happened, otherwise false
01:29:31*dan_a bed
01:31:00JdGordon|Rincewind|: pretty much all ui screens need to be changed from bool to int... and its going to take a while...
01:31:43|Rincewind|JdGordon: If you start with it, I would like to have playlist viewer first :)
01:32:17 Quit Peps (Read error: 101 (Network is unreachable))
01:32:31|Rincewind|If I can do it, I might try to change this myself and submit a patch
01:33:54|Rincewind|btw, the last time I looked there were stille a few "return SYS_USB_CONNECTED" in gwps.c. Going to usb while in the quickscreen would cause a freeze
01:34:34 Quit Nico_P (Remote closed the connection)
01:35:31JdGordonno it shouldnt
01:35:37JdGordonSYS_USB_CONNECTED == -1
01:35:42JdGordon== GO_TO_ROOT
01:35:46n1sJdGordon: I guess setting ENABLEDPLUGINS=no should disable plugins building but still make codecs but I haven't tested it :-)
01:35:56 Quit CriamosAndy ("( :: NoNameScript 4.03 :: )")
01:36:00n1sin the makefile of course
01:36:09n1sneway, gnight.
01:36:12 Part n1s
01:36:16JdGordonthanks, cya
01:36:27 Quit midkay (Read error: 104 (Connection reset by peer))
01:36:40 Join Peps [0] (
01:38:13 Join Aaron4 [0] (
01:38:23|Rincewind|JdGordon: I just checked, It does crash in the quickscreen from wps
01:39:20JdGordonwell now you know not to plug it in :p
01:40:35 Quit Ribs ("eh eh ehhhh!")
01:40:51 Join dmdfan- [0] (
01:40:51JdGordonpetur: if you read the logs... it does actually exit the plugin browser....
01:41:07JdGordonbut goes back in straight away
01:41:08 Part toffe82
01:41:42mcphailCan anyone with an ipod-3g look at this patch and see if it is ok for submission: (I'm a _very_ amateur C coder!)
01:42:42mcphailactually, that's the wrong one...
01:46:15 Quit jhMikeS (Nick collision from services.)
01:46:22 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
01:46:40JdGordonanyone with commit access wanna do a quickie for me?
01:46:55 Quit dmdfan (Read error: 145 (Connection timed out))
01:56:11JdGordonlinuxstb_: still around?
02:00:36 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
02:01:05 Quit JdGordon ("CGI:IRC")
02:13:46 Join Thundercloud_ [0] (n=thunderc@
02:14:11 Quit Peps (Read error: 101 (Network is unreachable))
02:24:34 Join midkay [0] (n=midkay@rockbox/developer/midkay)
02:25:14 Nick SpAwN is now known as HaSH (n=free@unaffiliated/blazed)
02:27:05 Quit Domonoky (Read error: 104 (Connection reset by peer))
02:27:41 Join fasmaie [0] (
02:30:37 Quit Stalwart (Connection reset by peer)
02:30:46 Join Stalwart [0] (
02:31:19 Quit Thundercloud (Read error: 113 (No route to host))
02:32:28 Join relaxed_ [0] (n=relaxed@unaffiliated/relaxed)
02:34:08 Quit relaxed_ (Client Quit)
02:36:51 Quit relaxed ("Lost terminal")
02:41:51perldiveris anybody using the screen rotate patch for gigabeat?
02:41:53 Join toffe82 [0] (
02:46:43 Quit fasmaie (Remote closed the connection)
02:48:37 Quit inversions (Read error: 110 (Connection timed out))
02:54:01mcphailany comments on this scrollwhell patch for the ipod-3g?
02:58:24 Part pixelma
02:58:26 Part Llorean
03:02:03 Join pacbox [0] (i=4b4b4e36@gateway/web/cgi-irc/
03:02:15 Quit BigMac ("MadPenguin forever!")
03:02:17pacboxanyone here?
03:02:30pacboxI need help with pacbox
03:03:26 Join Peps [0] (
03:04:13pacboxAny advice would be appreciated
03:04:34toffe82what is your problem
03:04:52pacboxi downloaded the Roms that it said i need
03:05:05pacboxpacman.5e etc...
03:05:17pacboxi have them in .rockbox/pacbox
03:05:28pacboxbut when i open it on rockbox it saids roms are missing
03:05:57PepsI am using the latest build for the ipod video, and most of the themes seem to have problems. I see starge numbers and garbeld screens. Is it possible that the new version and the wep files are out of sync?
03:06:11pacboxMy friend had the same problem
03:06:20pacboxhe had to change everything to default
03:06:29toffe82what build of rockbox ?
03:06:39Pepshow do I find out?
03:06:42pacboxFirst build of ipod video
03:07:16Pepsthe version number stays on screen for such a short time
03:07:49***Saving seen data "./dancer.seen"
03:08:25toffe82should be pacman directory , not pacbox
03:08:29Pepsand if the last number is a date this is scary time travel stuff
03:08:59toffe82fron the wiki : Copy these files to the /.rockbox/pacman/ directory on your player.
03:09:24pacboxi am trying that now
03:09:39Pepsany ideas what I can do?
03:09:59Pepspacbox, yes?
03:10:06pacboxMy friend had to set the themes to default and then use fonts +change background colors manually
03:10:22pacboxThemes just wouldn't work for him
03:10:43pacboxI don't know if there is a way to fix that
03:10:46Pepseither the last build is broken or is incompatible
03:11:09Pepsif broken, where do we lodge a bug report?
03:11:50Pepswill give it a try, thanks
03:12:51pacboxAlright pacbox works
03:12:54pacboxthanks toffe
03:12:58pacbox1 more thing
03:13:11pacboxEvery time i boot up Rockbox it says "Database not ready"
03:13:15pacboxWhat does this mean
03:13:24pacboxI can play music and stuff
03:13:43pacboxAnd i went over to the database settings and updated.
03:14:20pacboxAs far as i can tell the message hasn't had any adverse effects, i'm just wondering why it's there
03:14:38toffe82I don't really use it so I cold not answer
03:14:46toffe82try the manual
03:14:59pacboxThanks for all of your help regarding Pacbox
03:15:07 Quit pacbox ("Thanks")
03:19:12 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
03:20:02 Quit Thundercloud_ (Remote closed the connection)
03:23:35 Join webguest59 [0] (i=43bbe7a0@gateway/web/cgi-irc/
03:23:44webguest59does the rockbox view pdf files
03:24:01webguest59does rockbox for ipod video I meant
03:24:22bonbonthejonwebguest59: no
03:25:07 Quit webguest59 (Client Quit)
03:44:06 Quit Peps (Read error: 101 (Network is unreachable))
03:47:30 Quit Aaron4 ()
03:48:17 Join Peps [0] (
03:50:12Pepscan rockbox play video on ipod 5 gen? does not seem like it
04:03:33rotatorPeps: yes, somewhat:
04:12:34Pepsrotator, thanks, does not look very useful on the ipod video at 10fps, and it seem I will need to re-size my movies to 320x240
04:13:08 Quit bonbonthejon (Remote closed the connection)
04:13:26Pepscan I have different fonts for the play back screen and the menues? I do not mind small in the playback but when browsing small fonts are too much for my eyesight
04:13:37rotatornot at the moment
04:16:02 Quit secleinteer (Remote closed the connection)
04:16:20Pepsis there a plugin developer manual somewhere? how hard is it to develop using c++ on ubuntu?
04:18:26rotatorI suggest you take a look at the wiki ;)
04:18:34 Join secleinteer [0] (n=scl@
04:19:37rotatoralso, rockbox is written in c
04:19:56Pepsso no c++?
04:20:47rotatorno, though it should be possible to write plugins in c++ if you really wanted to
04:21:00rotatorthough there's not much point without dynamic memory allocation
04:21:57Pepswhat do you mean? no malloc?
04:22:43Pepsthat would make wring even simple c++ programs hard
04:22:52Pepsthat would make wring even simple c programs hard
04:23:16Pepsshame, will go and do more reading. thanks
04:23:21rotatornot really, just statically allocate everything to the max size you'll need
04:50:25 Quit Peps (Read error: 101 (Network is unreachable))
05:01:07 Join jhulst [0] (
05:03:44 Join aliask [0] (
05:03:56 Quit miepchen^schlaf (Read error: 54 (Connection reset by peer))
05:04:16 Join miepchen^schlaf [0] (
05:06:02 Join Shaid` [0] (
05:07:53***Saving seen data "./dancer.seen"
05:08:57 Join bonbonthejon [0] (
05:15:09 Quit Shaid (Nick collision from services.)
05:15:10 Nick Shaid` is now known as Shaid (
05:18:56 Quit midgey ()
05:30:13 Join JdGordon [0] (i=dced3920@gateway/web/cgi-irc/
05:31:29 Quit JdGordon (Client Quit)
05:37:06 Quit Rick (Read error: 104 (Connection reset by peer))
05:37:27 Join Rick [0] (
05:48:52 Join Gibbed [0] (
05:48:56 Quit Rick (Nick collision from services.)
05:50:07 Quit MirrorImage ()
05:51:42 Quit merlin2049er (Client Quit)
05:52:36 Join SimonSelki [0] (i=42452c9d@gateway/web/cgi-irc/
05:53:27SimonSelkiHey folks, I had a question about rockbox's playlist function and iTunes
05:54:37SimonSelkiI noticed then when I viewed playlists that had songs that I used iTunes to transfer, the names came up as wierd sequences of letters
05:55:09SimonSelkiI'm sure it's got to do with iTunes splitting up the files and stuff, but is there a way to help this?
05:55:26SimonSelkiI mean, I can't tell which songs are what
05:58:39 Nick Gibbed is now known as Rick (
06:00:35 Quit XavierGr (Nick collision from services.)
06:00:39 Join XavierGr [0] (
06:04:10 Quit SimonSelki ("CGI:IRC (EOF)")
06:05:47 Join shnee [0] (
06:06:40 Join ashridah [0] (
06:10:36 Join Peps [0] (
06:12:08PepsI see this line in the .wps file "%s%m|49|318|%?ia<%ia |%?d2<%d2|(root)>>", but nowhere in the documentation I see what "%m|49|318" is supposed to do and on my screen I actually see those number. anyone knows?
06:12:42 Quit |Rincewind| (Remote closed the connection)
06:13:24aliaskPeps: Which wps is this in
06:13:57Peps Black Glass.wps
06:14:16aliaskI think it's a setting to set the margins, but it's not in official rockbox source
06:14:53Pepsso it means people using this wps are using modified versions?
06:16:01Pepsand I see no mention of an id3 tag for the cover art, yet I saw some wps screen shot showing those. another patch?
06:16:02aliaskBut I havn't really seen much wps code in quite a while, I could be wrong... I'll take a look
06:16:07 Quit Soap (Read error: 104 (Connection reset by peer))
06:16:18ShaidYeah, Album art is an unofficial patch
06:16:19 Join Soap_ [0] (n=Soap@rockbox/staff/soap)
06:16:20aliaskYep, cover art is unofficial
06:16:48Pepsso what do I do except compiling my own rockbox, which I hate to do?
06:17:12Shaiddownload one of the unofficial builds. check the forums
06:17:30Pepswhich formus? I only started today :)
06:18:27Shaiddont really need the session id. :P
06:18:51Shaid is a build that'll do what you need
06:19:57 Quit Rick (Read error: 104 (Connection reset by peer))
06:20:03 Join Rick [0] (
06:21:03Pepsbut those are really old builds. I guess many bugs have been fixed since, no?
06:21:26Shaidwhich means you may want to grab the patches yourself and build it yourself. ;)
06:22:12Shaidthere you go
06:22:14Shaidnewer builds
06:22:20Shaidfrom the guy who wrote the themes/patch
06:28:27PepsShaid, how can I tell if I need the 5g or 5g64MB version?
06:29:13 Join perplexity [0] (n=joust@
06:29:26 Part ashridah
06:29:33Shaiddo you have a 30g or a 60g?
06:29:40Shaid30g is 32meg of ram, 60g has 64meg of ram
06:30:48Pepsand what does oop means?
06:31:25 Join Hoffmann [0] (n=ber@
06:31:32aliaskPeps: cop?
06:31:44aliaskOr oop?
06:31:46Pepsyes (bad eyesight)
06:32:02aliaskBecause ipods have 2 cores in the CPU
06:32:27aliaskSo a COP build will enable the coprocessor
06:32:27PepsI have 60g video, so do I need the cop?
06:32:50aliaskWell, you don't NEED it, but it should make things run a little smoother
06:33:17Pepsstrange that such option have a different build
06:33:55aliaskWell, it hasn't been until recently that it's been at a state where it's been stable enough to be introduced to the main source
06:34:42aliask2 days ago in fact.
06:34:46Pepsanother issue I have is the scroll - what should I set the options to to get as close to the apple one? anything I try the scoll does not accelarate enough
06:35:12aliaskAre you running a COP build?
06:35:19aliaskI hear that helps quite a bit
06:35:26PepsI don't think so
06:35:38aliaskHowever I dont have an iPod so I don't really know about the settings.
06:35:44Pepsso it may be hardware issue and not seting?
06:36:41Pepsok, but I would be surprized as it is such a fundamental thing with any ipod with a scroll wheel
06:36:49lexit's because of the driver
06:36:52lexit's a basic one
06:37:07Pepsthat sounds more like it
06:37:41Pepsand Apple does not provide any specs, so it is hard to write a better driver, right?
06:38:12 Join Gibbed [0] (
06:38:16aliaskQuite true.
06:38:17 Quit Rick (Nick collision from services.)
06:38:53lexi guess reverse engineering is fun
06:38:56PepsIt is amazing to have something like rockbox at all.
06:39:32PepsIn digital cameras no one got further than changing a few basic settings in the canon 300d
06:39:38 Nick Gibbed is now known as Rick (
06:39:44 Quit XavierGr (Nick collision from services.)
06:39:47 Join XavierGr [0] (
06:39:57Pepsand that was several years ago
06:40:32lexit's easier on ipod, the bootloader and the firwmare lies on the hard disk
06:40:41lexon canons the fw is probably on a read-only flash
06:41:57Pepsmost digital cameras have flashable firmware
06:42:06PepsI mean the SLR's
06:42:15lexcan be, but the processor is a problem too
06:42:42lexand it's a lot harder to make a firmware to a camera or a phone than into a music player
06:42:48aliaskNo more so than an iPod cpu i wouldn't think
06:42:52PepsNow, is it safe to dump the julius build on top of the one I have or will I lose my settings and database?
06:43:25aliaskMight lose the settings, but the database will be ok. You can save a config file and reload it though.
06:44:27 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
06:46:15lexi used senab's build and i couldn't play music anymore :(
06:46:32aliaskBut I bet the WPS looked nice!
06:46:35Pepsi don't find the music too
06:48:43PepsI used rolo to load the julius version, now there is no "database" or anything
06:49:44aliaskHow recent was the build?
06:50:15Peps[ ]
06:50:44aliaskLooks like it was built before the root menu came into the code
06:51:01aliaskSo you can access the database mode by changing the file view option to database
06:51:40aliaskEr, not sure... if it's already on it then it should display the database... does it not?
06:52:35aliaskJolly good then.
06:52:51aliaskSorry, watching top gear.
06:53:19aliaskWell, I'm not really sure why it's not working... has the database been built?
06:53:40Pepsyep. now rolo hang waiting for the co- processor
06:54:09aliaskDoes someone who knows more about this want to step in?
06:54:25Pepsafter rolo I get the file menu, and pressing menu goes to settings
06:54:51aliaskYep, that's the old menu system.
06:55:19Pepsso how do I get to the database (it was already built using the new firmware)
06:57:12PepsI guess I will have to live without co-processor and cover art and facny weps until a more recent build comes along
06:57:26aliaskLooks like it sadly
06:57:36Pepsthanks god the reset button works. I had to rebbot about 10 times already
06:57:44aliaskWell, the current official build will have coprocessor support
06:58:11Pepswith the new version, every time it comes up I get the setting menu - can I change that?
06:58:39Pepsthere is no official build forthe ipod video
06:59:08aliaskWell, the current SVN build is what I meant I suppose
06:59:35Pepshow can I tell if I have it or not?
06:59:54aliaskamiconn: Morning!
07:00:03PepsI mean if the build I am running has co-pro in or not?
07:00:11 Quit rotator ()
07:00:28aliaskamiconn: Great work on the M5 by the way.
07:00:50aliaskMy friend has one, I'll see if I can convert him :)
07:01:08aliaskPeps: Good question, anybody else know?
07:07:57***Saving seen data "./dancer.seen"
07:10:57 Quit perplexity (Read error: 110 (Connection timed out))
07:17:27 Quit Rick (Read error: 113 (No route to host))
07:17:43 Join Rick [0] (
07:18:25 Quit actown (Read error: 104 (Connection reset by peer))
07:18:36 Join actown [0] (n=actown@unaffiliated/actown)
07:27:06 Quit billytwowilly (Remote closed the connection)
07:27:25 Quit lini ("lini has no reason")
07:27:32ShaidPeps: Have you installed a new bootloader later?
07:27:46ShaidCause old bootloaders and the new cop code don't quite coexist peacefully together
07:28:59 Quit Rick (Read error: 104 (Connection reset by peer))
07:29:15 Join Rick [0] (
07:29:40 Quit Peps (Read error: 101 (Network is unreachable))
07:32:14 Join lini [0] (i=pugsley@
07:32:25 Quit Rick (Read error: 104 (Connection reset by peer))
07:33:07 Join Rick [0] (
07:35:43 Quit Rick (Read error: 113 (No route to host))
07:36:19 Join Rick [0] (
07:42:46 Quit Rick (Read error: 104 (Connection reset by peer))
07:43:17 Join Rick [0] (
07:43:53 Join Shaid` [0] (
07:50:28jhMikeShmmm...wonder what got the non cf builds ballooning up so much. it's not the pointer use since taking that out for NB_SCREENS == 1 and using arr[0] in their place only saves back 4 bytes. must be from accessing more stuff in a static struct.
07:50:32 Join Shaid`` [0] (
07:50:44 Quit Shaid (Nick collision from services.)
07:50:50 Nick Shaid`` is now known as Shaid (
07:51:47jhMikeSwell, one way to test that
07:57:51amiconnI also wondered that, but it's hard to tell what actually changed because of the heap of formatting changes
07:59:09amiconnIt's also not helpful that the targets which have an lcd remote are also the only coldifre targets, while all single screen targets are either arm or sh
08:02:48axion1I think with AMS support rockbox will really improve greatly. Its already a great program but now it can get even better
08:04:22jhMikeSdoing a quick hard recoding of the others for a single screen didn't help anything...will test another suspicion
08:06:37 Part toffe82
08:09:06 Quit Shaid` (Read error: 110 (Connection timed out))
08:09:09 Join Febs_ [0] (
08:10:39aliaskaxion1: Well, the sansa will probably get sound if the AMS stuff continues. I'm sure they'd like to be able to listen to their mp3 players :)
08:13:53jhMikeSyep, it's the struct accesses. if I move the big buffer out and put the parameters on the stack it goes right down hugely
08:14:58jhMikeScan cut almost 500 bytes on ARM that way
08:15:49jhMikeSany reason the currenly selected character has to persist accross keyboard calls?
08:16:35 Join tucoz [0] (
08:17:04tucozamiconn: how's the lcd on the m5. As slow as the h120 or better?
08:17:32tucozit sure looks like a nice player
08:19:40amiconnIt's the same lcd controller as in the h1x0 but a physically different panel
08:20:10amiconnIt's physically smaller than the h1x0 lcd, and also smaller than the X5 lcd, but the same resolution
08:20:45tucozok. I tried to play chopper on my h120 yesterday, but got nausious because of the blurry display.
08:21:08tucozbesides the lcd, it's an excellent player i must say
08:21:27 Quit perldiver (Read error: 110 (Connection timed out))
08:21:43amiconnChopper is working nicely on the M5 lcd
08:22:06tucozlucky you ;-)
08:22:17axion1I use rockbox on my player in my car often. Just load up my winamp playlist and im good. Database is still taking it's time updating since it's not quite complete yet
08:22:18amiconnCompared to the H1x0, the M5 lacks the radio and s/pdif in/out
08:22:30amiconnTo make up for this, it doesn't need an rtc mod ;)
08:23:23amiconn...and the case looks more elegant and is smaller
08:23:57tucozit sure looks good. I think both the x5 and the m5 looks a lot nicer than the irivers
08:24:10 Join webguest94 [0] (i=c023110a@gateway/web/cgi-irc/
08:24:15amiconnI bet the M5L is the target with the longest battery runtime among all rockbox targets, even a bit longer than the X5L
08:24:17axion1love my X5L
08:24:46axion1natural battery was around 30 hours. Still waiting on improvement :)
08:24:51axion1on rockbox
08:24:58tucozthat is quite impressive
08:25:17 Join decayedcell__ [0] (n=decayed_@
08:25:24amiconnBtw, it's not readily visible from the front scans: the colour design of the M5 and X5 cases are complementary
08:25:33axion1right now on rockbox it's around 16 hours. So I am sure it will be awhile before battery effiency will be improved
08:25:52amiconnThe X5 has a black case with a silver stripe around all 4 sides where the buttons and sockets are
08:26:04amiconnThe M5 is silver with a black stripe
08:26:26tucozaha. so they are a match when you own both
08:27:52tucozwonder why they call the m5 color sound though
08:28:18tucozit might hint to that they actually refer to the sound
08:28:33 Quit Febs (Read error: 110 (Connection timed out))
08:28:46 Join Rob2222 [0] (
08:30:15 Quit decayedcell__ (Remote closed the connection)
08:30:28 Join decayedcell__ [0] (n=decayed_@
08:30:40axion1Even though on my X5L with rockbox I get around 16 hours, I cna't complain at all. I get a great menu and sound. But I'm always looking forward to longer battery code improvments. But sacrafices have to be made sometimes.
08:32:03tucozI would consider my h120 next to unusable without rockbox
08:34:38 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
08:39:12 Join ackbahr [0] (
08:41:09 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
08:41:24 Quit Rick (Read error: 113 (No route to host))
08:41:33 Join Rick [0] (
08:43:06 Join ender` [0] (
08:46:03 Quit Rob222241 (Read error: 110 (Connection timed out))
08:50:47JdGordonhey LinusN, was the power bug in the h300 bootloader fixed? and what was the danger? I think i left it plugged in for an hour or 2 last night with it off before realsing... it was fully charged and still working so just checking...
08:51:57 Join perplexity [0] (n=joust@
08:52:36LinusNthe power bug is fixed, and the potential theoretical danger would be that it went in a (silent) reboot cycle, but it seems to have turned out ok for you
08:53:14JdGordonso it wasnt exploding battery danger?
08:53:28JdGordonor at least pverheating...
08:53:53 Quit ackbahr (Remote closed the connection)
08:54:02JdGordonbah, volume control doesnt work in the sim :'(
08:54:44tucozJdGordon, nice work on the root menu. It feels like it was the right thing to add to rockbox.
08:55:09JdGordon:) not according to some of the comments on the user ml :p
08:55:52Shaidanything that has a major usability change will get complaints.
08:55:55LinusNJdGordon: that's because they are blind, and are confused by changes in general
08:55:55aliaskYeah, I have to echo tucoz, I loaded up a sim recently and went oooooh
08:55:58Shaidused Office 2007 yet? :P
08:56:02 Join thewho [0] (i=c27f0812@gateway/web/cgi-irc/
08:56:08thewhoJdGordon: ping
08:56:14tucozreally? well, don't listen to them.
08:56:23JdGordonI dont :D
08:56:36JdGordoni started a few rude emails but fortunatly decided against sending them :p
08:56:44tucozsome people even prefer the arrow row selector instead of the inverted bar
08:56:57LinusNJdGordon: but the main menu has broken a very important feature for them: that the browser doesn't preselect the played file, and follow playlist is b0rked too
08:56:59tucozi.e. the _tiny_ arrow
08:57:22JdGordonfollow isnt borked... i thought follow worked and off was broked?
08:57:22thewhoJdGordon: observed a few hours ago: when FM is active (I'm in the FM screen) and call up the main menu, it contains "Now playing". Shouldn't it be "Resume playback" (since mp3 playback is not active)?
08:57:23 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
08:57:48LinusNoops, meeting
08:57:49tucozbut isn't that things that are broken, but should work? not features that are removed
08:58:24JdGordonthewho: ok, thanks, ill have a look
08:58:47thewhoJdGordon: ok. This was H120 sim btw (if that matters)
08:59:05 Part thewho ("gtg but will read the log")
08:59:16JdGordonoh? i just checked the h300 sim and it worked and assumed the sim was what was causing it to work
09:08:01***Saving seen data "./dancer.seen"
09:11:05jhMikeSoy...bad for SH
09:18:36dan_alinuxstb: (for the logs) - The latest bootloader works properly on the 4G
09:23:54 Part tucoz
09:25:16 Join pondlife [0] (
09:27:10amiconnjhMikeS: 'volatile' reduces size for arm? That's really odd...
09:27:37jhMikeSindeed :\
09:29:48*jhMikeS is always at war with gcc
09:30:06pondlifeA war you can't win...?
09:30:44jhMikeSheh...what's worse is volatile reduced the size by more than the original increase :P
09:31:14pondlifeThat's better, not worse, surely
09:31:55jhMikeSworse is better with gcc :P
09:32:07pondlifeAh, double -ve!
09:32:24pondlifegcc being one of them
09:35:43jhMikeSoddly enough, * volatile const doesn't seem to help :P
09:38:40 Join ackbahr [0] (
09:39:38 Quit pondlife (Read error: 104 (Connection reset by peer))
09:41:09 Quit ackbahr (Remote closed the connection)
09:43:41 Part decayedcell__
09:44:22 Join decayedcell__ [0] (n=decayed_@
09:45:11 Quit Lynx_ (" HydraIRC -> <- Leading Edge IRC")
09:45:43 Quit Febs_ (Read error: 110 (Connection timed out))
09:49:23 Quit decayedcell__ (Remote closed the connection)
09:49:34 Join decayedcell__ [0] (n=decayed_@
09:49:36LinusNdamn, i tried to reproduce the recording-with-voice-enabled issue on an h120 yesterday, but i failed
09:51:06LinusNreally annoying
09:52:20 Join pondlife [0] (
09:52:29pondlifeJdGordon: ping
09:52:52pondlifeOr whatever
09:53:04*amiconn has a suspicion why * volatile reduces size on arm, but I think we shouldn't use this special trick
09:53:12pondlifeFollow playlist isn't working now.. :(
09:53:37JdGordonoh bloody hell :'(
09:53:48pondlifeYep, well in the H300 sim at least
09:53:54amiconn* volatile makes gcc reload the data everytime, because it thinks the data might change in between
09:54:20pondlifeJdGordon: Go into an album and play track 3, then skip to track 4 and press SELECT...
09:54:25amiconnAnd since arm has very efficient indexed addressing modes, gcc can save registers. But the code will be slower this way
09:54:29JdGordonok.. grrr
09:54:40pondlifeSorry to bear bad news.
09:54:57pondlifeThe interface between browser and root menu is a bit over-complicated
09:55:10JdGordonpondlife: its way too complicated
09:55:14pondlifeCan not the browser handle all of this alone?
09:55:19LinusNthere is room for serioous cleanup there
09:55:22JdGordonnot in the current form
09:55:28pondlifeThe only time you need to pass a path in is for plugin browsing etc, right?
09:55:37 Join bluebrother [0] (i=d2q6WTX5@rockbox/staff/bluebrother)
09:55:44JdGordonthe tree_context is all wrong imho
09:55:44pondlifeI mean in theory
09:56:09pondlifeJust have it passed a NULL on normal entry (meaning, wherever you last were).
09:56:24pondlifeAnd have follow playlist handled between the WPS and browser alone.
09:56:25JdGordonI want to split the tree totaly so all that is needed is a path, nothing else, and the db tree would be handled in tagtree.c instead of tree.c
09:56:40pondlifeHmm, paths are the wrong way IMHO
09:56:49JdGordontree.c is going to be overhauled _soon_
09:57:15pondlifeDo you know how you will handle paths in the Database view?
09:57:20pondlifeIf you have a plan, good
09:57:37JdGordonno, db is a whole different beast
09:58:04pondlifeIf anything they should be sharing more code, surely
09:58:25pondlifeDon't want to end up with 2x browsers if we can help it.
09:58:39pondlifeI've barely thought this through of course!
09:58:42JdGordonif the db could use psuedo-paths it would make things easier
09:59:29jhMikeSamiconn: well, is the code super speed critical there? there's not really much happening.
09:59:32 Join Ribs [0] (n=ribs@
09:59:34pondlifeTricky I suspect.
10:00:07pondlifeJdGordon: Can you not make the browser just keep track of where it is in the current playlist, and only reset if the filter type changes?
10:00:11jhMikeSamiconn: or do you mean using it in general?
10:00:23 Part decayedcell__
10:00:36amiconnjhMikeS: No, but I don't think having esoteric tricks for certain cpus is good. Especially if it's for targets where size doesn't matter that much
10:00:55pondliferoot_menu should not need to know anything about paths at all.
10:00:57JdGordonpondlife: I have half an idea... it needs more thinking
10:01:02JdGordonback soon
10:01:16pondlifeMe too, put two halves together and you get... idid
10:01:18amiconnI'm not sure how this influences the iFP, but the size differences might be all different there, because it uses only short calls
10:01:40*amiconn dislikes how the db first shows the top-level pseudo folders on entering, and the says 'db not ready' when trying to enter sub folders
10:01:52pondlifeamiconn: That's a bug, surely
10:02:36jhMikeSahh...JdGordon's already using the space back up :P
10:02:58amiconnImho it should show 'db not ready' in the top level, and then ask whether you want to init, and show progress while it's doing that, also asking for the necessary reboot at the end
10:03:31amiconnOf course it should be possible to leave the progress screen and e.g. use plugins, and when re-entering the db view, the progress screen should reappear
10:03:55 Join Lynx_ [0] (
10:04:46JdGordonamiconn: I agree... but if its not done like that coming out of the WPS doesnt work properly if it goes to an un-innited db
10:05:01JdGordonjhMikeS: I only took half your saving...
10:05:20jhMikeSJdGordon: What goes in the other half?
10:05:28 Join linuxstb_ [0] (i=5343d4aa@rockbox/developer/linuxstb)
10:05:33JdGordonflip you for it :p
10:05:36pondlifejhMikeS: I'm sure I can find something ;)
10:06:00jhMikeSWell, Coldfire left a little more room for ya'll :)
10:06:04 Join decayedcell [0] (i=3ba79d47@gateway/web/cgi-irc/
10:06:14*daurnimator slaps JdGordon around a bit with a large trout
10:06:18pondlifejhMikeS: Are you aware of the voice/recording crash issue?
10:06:20linuxstb_dan_a: You're saying that removing the lcd_update() calls makes the bootloader load the apple firmware? Which gcc did you use?
10:07:06jhMikeSpondlife: yes...not sure what to make of it yet or I'd already have put a stop to it. I can't reproduce it very well at all. I mean, voice can lock in just about any context really.
10:07:31JdGordonpondlife: What I would like to do with the tree is replace rockbox_browse() and make it more usable, specifically add a action callback so we can use it as a file save as /open dialog.
10:07:36pondlifejhMikeS: This one makes the recording screen go slower and slower until it stops, but only with larger voice files apparently.
10:07:41jhMikeSbut it it really a crash? iirc it's a lockup
10:07:59*JdGordon shoves the trout down daurnimator's throut!
10:08:00*amiconn spots linuxstb_ around
10:08:03pondlifeJdGordon: That would be useful, but not the primary aim
10:08:09jhMikeSIs this the one from Mar 04?
10:08:10JdGordonof course
10:08:14amiconnlinuxstb_: Did you notice that you broke the h10 and e200 bootloaders?
10:08:17pondlifeJdGordon: Go for KISS firstr
10:08:21daurnimatormy mouth is already full of chicken
10:08:21 Join peppo [0] (
10:08:31peppoany use asking to test the ipod 80gb video code?
10:08:36daurnimatorand you just made me grease up my keyboard by touching i
10:08:39pondlifejhMikeS: Yes, it's a lockup, sorry. I can't repro, but LinusN can. Maybe settings related.
10:08:46linuxstb_amiconn: Argh, no I didn't...
10:08:50JdGordonpondlife: KISS would work out in the long run if that was in from the start, instead of taking it on later
10:09:12pondlifeWell, start with the interfaces...
10:09:47*pondlife barbecues a trout he found lying around
10:09:50jhMikeSI can't imagine what would have a progressive effect on the screen. Once you're recording, voice should be swapped out and having nothing to do with anything.
10:09:52*JdGordon doesnt like the dependancy on tree_context everywhere
10:10:40pondlifejhMikeS: Indeed. I suspect it's a memory corruption somehow. Also LinusN reported that the problem occurs with voice playback disabled, although with the voice file loaded.
10:10:44jhMikeSIf voice made it's way in during record, the codec would get blasted and/or your audio would.
10:11:04 Quit decayedcell ("CGI:IRC (Ping timeout)")
10:11:36linuxstb_amiconn: Fix committed, thanks for the heads-up.
10:12:45 Quit BHSPitLappy (Remote closed the connection)
10:13:09 Quit darkless ("Leaving")
10:13:15 Join darkless [0] (
10:13:15jhMikeSponflife: something I observed may be related though. If I have a voice file set, and then do voice menus...if I reboot, voice doesn't operate and I can't restarting by redoing the settings. I can only get it to work if I set voice menus to "yes" after it's been booted with "no".
10:13:58pondlifejhMikeS: Sorry, run that by me again.
10:14:21pondlifeDo you mean if you change language?
10:14:24jhMikeSok, if I boot with voice menus "yes"...voice doesn't work.
10:14:32pondlifeWhich target?
10:14:45pondlifeHmm, seems ok on my H340
10:14:50jhMikeSIf I boot with "no" then set to "yes" it works until reboot.
10:15:48daurnimatorJdGordon: swap...:P
10:16:02pondlifeIf I boot up with it set to yes, it announces "Files" from the root menu. What's your startup screen?
10:17:04 Join decayedcell [0] (i=3ba79d47@gateway/web/cgi-irc/
10:17:12 Join fejfighter [0] (
10:17:34pondlifejhMikeS: Maybe does point to a race condition at voice init?
10:17:44pondlifei.e works on my target, but not on the sim
10:18:17jhMikeSthe voice implementation is a moving target :\
10:18:34pondlifeHasn't changed much recently has it?
10:18:50jhMikeSnot really, no
10:19:29jhMikeSI put in the separate stop flag for the encoders so they don't collide with stop requests for the decoders, that's it.
10:20:18pondlifeWell this particular bug (i.e. the recording slowdown->stop) appears to be quite old, but only appears with larger voice files (post root-menu).
10:20:30pondlifei.e. old build and new voice file will repro it apparently
10:21:12jhMikeShmmm...interesting. then why the report saying it's a recent problem?
10:21:16pondlifeSmelt like some use of an unincremented audio buffer pointer to me, but I've not looked yet.
10:21:30pondlifeRecent because voice files have got larger recently
10:21:46pondlifeProbably post root-menu, more lang entries?
10:22:24pondlifeWell, gotta do Paid Work now. Sorry to bug.
10:22:34jhMikeSunincremented audio buffer pointer?
10:23:55 Join ackbahr [0] (
10:24:44 Quit Rick (Read error: 113 (No route to host))
10:25:00pondlifejhMikeS: Yes, when a pointer to the start of available buffer is stored, then something else quietly allocates from the start of the buffer
10:25:19pondlifeThis caused similar problems with voice a while back
10:25:35pondlifeResulting not in a lockup, but in a notable slowdown
10:25:43 Join Rick [0] (
10:25:58*pondlife never properly understood how IRAM mapped though
10:25:59jhMikeSJdGordon was talking about using buffer_alloc recently.
10:26:53pondlifeThat's the sort of thing. Basically, there should not be any pointer values persisted anywhere when something else may execute!
10:27:21pondlifeOr perhaps buffer_alloc needs a locking mechanism to panic if something naughty happens?
10:27:42jhMikeSnothing else should execute (or be executing atm) that needs that.
10:27:55pondlifeYou are correct, but maybe something is.
10:28:43jhMikeSI made it really simple, you must call audio_get_buffer(true, &size), _then_ use buffer_alloc once audio has been initialized.
10:28:58jhMikeSso if that rule is broken, things won't go so well
10:29:23 Quit linuxstb_ ("CGI:IRC")
10:29:24pondlifeIndeed. We should have a point where the audio buffer is locked, so further buffer_allocs will fail (or panic)
10:30:04pondlifeI worry that people will regard buffer_alloc like a non-freeable malloc, when there's more to it than that.
10:31:07jhMikeSThe audio buffer state must be BUFFER_STATE_TRASHED for the call to be valid, and it must not be flagged busy (there's no flag yet).
10:32:18jhMikeSYes, buffer lock...the consequences of incorrect procedure should be made obvious and immediate
10:32:59pondlifeI'd implement it, but I'm not sure at what point it should be locked.
10:33:06pondlifeI'm a bit out of date...
10:33:15jhMikeSWhen recording needs it, and when audio initializes it
10:33:35jhMikeSaudio_get_buffer will release an audio lock
10:33:40pondlifeGo for it ;)
10:34:09jhMikeSpanicf when attempting to circumvent? ;)
10:34:35pondlifePretty obvious then :)
10:35:02jhMikeSIt should never occur without bugs anyway.
10:35:37*jhMikeS is afraid of seeing it light up all over the place
10:35:50 Join austriancoder [0] (n=austrian@
10:35:53pondlifeWell, it'll be a useful debugging exercise then
10:39:32jhMikeSI'm pretty buff from exercising so much already.
10:40:45 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:40:54jhMikeSI think it should display the return address into the caller as well...and the thread it was called on. Can do for Coldfire but don't know much about how on other CPUs yet.
10:43:34jhMikeSI have observed a mysterious creeping recording screen on x5 but it happens fast and can't do it reliably.
10:45:36ackbahrGeneral question : do you know if the USB host support is considered for H320? Thanks!
10:47:20linuxstbYes, it's considered, but I don't think anyone is actively working on it at the moment.
10:49:13ackbahrAnyway, there's the lousy original software to use USB host....
10:50:01austriancoderackbahr: i will do it for Googels summer of code
10:50:29bluebrotheraustriancoder: btw, any news from the AMS front?
10:50:53austriancoderbluebrother: in 3 days the meeting will took place
10:51:25bluebrotherah. Thought that date had already passed
10:51:42bluebrothershould check that page more often ;-)
10:52:09 Quit ackbahr (Remote closed the connection)
10:53:48austriancoderbluebrother: so more news will follow around that date
10:54:09 Join JerryLange [0] (
10:59:38linuxstbdan_a: I've built a new set of bootloaders from SVN - can you test on your 4g? I've called these version 1.0 - hopefully we can tell everyone they can safely upgrade to this version.
10:59:44markunaustriancoder: have you talked to petur about implementing a USB stack?
11:00:53linuxstbaustriancoder: There is also interest in USBOTG support in the PortalPlayer targets, so it would be good to work with someone doing the PortalPlayer implementation if such a person exists...
11:01:47austriancoderlinuxstb: i own a portalplayer
11:02:02linuxstbAh, so you're planning on PortalPlayer USBOTG, not H300?
11:02:05austriancodermarkun: no.. is he also interesed in it?
11:02:46austriancoderlinuxstb: i try to write a very portable code.. a friend of my has a H300.. so testing should be no problem
11:05:35markunaustriancoder: he has written one before and was the most likely person to do it for rockbox so far
11:06:46markunaustriancoder: The Gigabeat doesn't have USBOTG, but a normal USB 1.1 controller (can be host or device I think)
11:07:02markunwould be nice if that was supported too
11:07:17austriancoderseems to be very hard to find a Summer of Code project that interests me and that nobody else wants to do..
11:07:40markunpetur doesn't have time to do it I think
11:07:53markunbut you might want to talk to him anyway
11:08:04***Saving seen data "./dancer.seen"
11:09:35austriancoderyep.. i will send him a mail.. maybe he can become my mentor for this project
11:12:10 Join rp- [0] (i=rp@
11:12:57austriancoderhi rp-
11:14:18rp-hi all
11:14:36rp-austriancoder: query
11:14:50austriancoderrp-: one moment.. need to check irssi doku
11:15:01GodEater_markun: USB 1.1??!? Are you sure ?
11:15:09GodEater_that's dog slow
11:15:32rp-austriancoder: alt + number in the statusbar :P
11:15:42amiconnUSB2.0 can also be dog slow
11:16:01austriancoderrp-: cool
11:16:09 Quit GodEater_ ("CGI:IRC")
11:16:24 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
11:16:51markunGodEater_: I'm sure. But the USB2ATA bridge is 2.0.
11:17:20GodEater_ah - that makes more sense
11:17:21markunGodEater_: but the controller being 1.1 is the reason why syncing with WMP is so slow
11:17:34GodEater_I never tried that - so was unaware of it
11:17:50markunnever tried it either
11:19:48markunpeople at mygigabeat were complaining that Toshiba didn't release a firmware update to make MTP mode USB 2.0. I told them it was a hardware limitaion, but I don't think many people believed me.
11:20:17 Quit JerryLange (Read error: 113 (No route to host))
11:21:36*amiconn pings LinusN
11:27:12 Quit decayedcell ("CGI:IRC (EOF)")
11:27:15 Join muesli__ [0] (n=muesli_t@
11:30:29desowinregarding Summer of Code, has rockbox already sent application as mentoring organization ?
11:30:39*jhMikeS wonders if the recordings that exhibited the slowdown were glitch free
11:33:22linuxstbdesowin: I don't believe so.
11:33:28LinusNjhMikeS: when i see the problem on my h300, i don't even need to record anything to see the crawl
11:34:01LinusNi enter the recording screen, and the peak meter is dog slow, and the button response time is horrible
11:34:14LinusNit takes several seconds to exit the recording screen
11:34:15peppoany use asking to test the ipod 80gb-video code?
11:34:36LinusNand if i try to record, it hangs when i try to stop it
11:35:05pondlifeLinusN: I can't repro this with all voice features enabled, I wonder if it's some other setting related too?
11:35:13pondlifeWhat's your start screen set to?
11:35:26pondlifeMine's Main Menu.
11:35:37LinusNpondlife: possibly
11:35:49LinusNi could only reproduce it on my h300, not on my h120
11:36:09pondlifeWith the same config? So timing sensitive too perhaps?
11:36:59 Join kubiixaka [0] (
11:36:59 Quit kubiix (Read error: 131 (Connection reset by peer))
11:37:51 Quit Lynx_ (Read error: 104 (Connection reset by peer))
11:38:38jhMikeSLinusN: I had seen that on the x5 occassionally before way back...but it was like seeing bigfoot.
11:39:08 Join Wiwie [0] (
11:39:26LinusNjhMikeS: yes, it's an elusive bug for sure
11:40:17jhMikeSactually, it happened recently. I could see the peakmeters moving really slow when starting out.
11:40:31 Quit BHSPitMonkey (Remote closed the connection)
11:41:08jhMikeSWhat format?
11:41:13LinusNcould it be that the voice thread goes picking daisies and eats a lot of cpu?
11:41:17markunthe first 2 hint apply to HWCODEC and the last 2 to SWCODEC on GaplessHowTo right?
11:41:20*JdGordon 's rec screen is dog slow also
11:41:37jhMikeSDon't know how. It should just be sitting waiting for the mutex.
11:41:49jhMikeSinside swap_codec
11:42:51LinusNi just reproduced it again
11:43:07LinusNbooted woth voice off, entered recording - all was normal
11:43:08jhMikeSthough it comes to mind that voice could swap itself out inside voice_pcmbuf_insert_callback
11:43:21LinusNthen i enabled voice and tried recording again - slow as hell
11:43:22jhMikeSany format?
11:43:33LinusNjhMikeS: wav
11:44:00jhMikeSok...I seemed to see it with wavpack more often but that might just be that wavpack took so much attention
11:44:46jhMikeSand why h300's giving me trouble all the time? :P
11:45:20JdGordonpondlife: you said i brokle follow playlist again didnt you?
11:45:50pondlifeJdGordon: Yes
11:46:01JdGordonworking fine here....
11:46:07JdGordonafter a whole 1 test :p
11:46:18pondlifeI'm using the sim.. what are you using?
11:46:22LinusNi bet it's a sim vs target thing
11:46:34JdGordonyup, just tried again
11:46:36JdGordonworks correclty
11:46:42pondlifeOK, I'll just rebuild
11:46:49jhMikeSgrrr...wonder where the voice thread is in all that
11:47:18pondlifeOK ... SVN rev 12642 bulding
11:47:22JdGordonpondlife: are you going throught the root menu? or direct wps > tree?
11:47:36pondlifeWPS to tree with SELECT
11:48:00pondlifeThe "normal" way, I think...!!
11:48:25pondlifeThe root menu has no need to be involved in this
11:48:44JdGordonno, but if you were going through then i would understand.. its not supposed to follow in that case
11:48:51pondlifeWhy not?
11:48:54JdGordonbut if your having problems wps > tree then we have a bug
11:49:02JdGordonthat was the decision
11:49:05 Join BHSPitMonkey [0] (
11:49:19pondlifeAh. I'd vote for consistency
11:49:22LinusNwas it?
11:49:34JdGordonwasnt it?
11:49:39LinusNah, now i remember
11:49:40jhMikeSare the voice files being used the newest ones available in the wiki or something else?
11:49:47pondlifeLinusN: What was the logic?
11:49:57pondlifejhMikeS: Something else.
11:50:03pondlifeHang on I'll find the link
11:50:48LinusNi think the logic was that you would expect the root->files way of browsing to retain the position
11:50:49pondlifejhMikeS: Go to
11:50:57jhMikeSI've seen this on H120 as well actually...but more elusive
11:51:10pondlifeTake the 070305 file
11:51:15LinusNi see it every time on my h300 currently
11:51:31pondlifeLinusN: Which of the voice files are you using? David?
11:51:42LinusNcepstral david
11:51:52jhMikeSpondlife: thanks
11:52:04LinusNand i have database and dircache in ram
11:52:10pondlifeNo need for thanks. Until I thank you for fixing it!
11:52:14pondlifeMe too
11:52:29LinusNand the rockboxed theme
11:52:33jhMikeSwell, I've got a big download first :)
11:52:47 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
11:52:53pondlifejhMikeS: Do you have a 7zip unpacker too?
11:52:59jhMikeSoh...well, damn DSL, it's done
11:53:21pondlifeI'm using my own sooper-simple WPS
11:53:33pondlifeNo peak meters, but it does have next track name
11:53:34jhMikeSWinRar does 'em all
11:54:27JdGordonanyone know why the sim baclight is dependant on LCD_DEPTH <= 8?
11:56:17 Quit idnar (Read error: 145 (Connection timed out))
11:57:27jhMikeSthis is a very old bug...all the way back to when I started doing recording stuff
11:58:43 Join decayedcell [0] (i=3ba79d47@gateway/web/cgi-irc/
11:58:46 Join inversions [0] (
11:59:01pondlifejhMikeS: Are you able to repro then?
11:59:24jhMikeSI'm doing a full rebuild. Will try in a minute.
11:59:39pondlifeOK. Keep us posted
11:59:57 Join anathema [0] (
12:01:13jhMikeSthen there's that thing with the boost and voice. that boost shows up even when booting with no voice and then turning voice menus on.
12:01:39JdGordonis dan in the svn logs dan_a ?
12:01:45pondlifeHmm, I've not noticed boost problems recently.
12:02:19jhMikeSthis is old too. It happens via trigger_cpu_boost somehow
12:03:01pondlifeWasn't a boost tracking marker added to the CPU stacks screen?
12:03:31JdGordona boost tracker log was added to logf builds
12:03:36 Quit Wiwie (Read error: 110 (Connection timed out))
12:03:39jhMikeSI find this comment odd: /* We must use queue_wait_w_tmo() because queue_wait() doesn't unboost the CPU */. I think that's backwards.
12:04:04pondlifeSurely both should unboost anyway?
12:04:16pondlifeWhy boost when you're just waiting?
12:04:55*pondlife is guilty of peephole development
12:06:06jhMikeSguess I've got it memory is screwed
12:06:19JdGordonif espeak is 600kb how is it going to be added? or does it have an ultra light lib that we can put in the core?
12:08:34bluebrotherwhy not use it as tsr plugin?
12:09:04 Nick Soap_ is now known as Soap (n=Soap@rockbox/staff/soap)
12:09:11JdGordonrocks can only be 512kb tho...
12:09:19JdGordonand smaller on the archos
12:09:21LinusNthat's not a fixed limit
12:09:22bluebrotherbut that's quite close
12:09:40bluebrotherand afaik the archos aren't powerful enough for a tts anyway
12:09:41LinusNif we need more space to fit the voice codec, then we make some
12:10:05JdGordonso we are going to have 2 speach engines ?
12:10:21jhMikeSbigfoot came to roost
12:10:26JdGordonoh :'(
12:10:34LinusNjhMikeS: :-)
12:10:43jhMikeShad to have the voice active though
12:10:45JdGordonI thought the plan was to replace the current voice with espeak
12:10:46LinusNnail the sucker!
12:10:56bluebrotherpeople on e.g. might ever prefer to use pregenerated clips ...
12:11:05LinusNJdGordon: on swcodec targets, yes
12:11:10pondlifeJdGordon: Have you heard eSpeak?
12:11:13jhMikeS2 times in a row
12:11:26pondlifeThe examples I heard were noway near as good quality
12:11:40JdGordonI havnt heard its output... no
12:11:42pondlifeI would choose pregenerated clips on SWCODEC still
12:12:04jhMikeSooh, every time even after reboot as long as voice menus are enabled
12:12:04pondlifeUnless there's been a big improvement or the clips I heard were particularly bad.
12:12:07bluebrotherwhat is the default tts for voice building? espeak or flite?
12:12:16bluebrotherah, forget me. It was festival
12:12:21DataGhost :)
12:12:35pondlifeFestivsl is even worse sounding than MS Mary
12:12:45pondlifeIMHO, if I could spell it.
12:12:58JdGordonDataGhost: that was in todays paper :)
12:13:06DataGhostand in todays mailbox
12:13:42JdGordonpondlife: did you try again with the newer sim?
12:13:54jhMikeSok, turning voice off and not rebooting still has the problem, but booting with voice off does not.
12:14:07pondlifeNo, I thought you hadn't fixed anything. Here goes...
12:14:18LinusNjhMikeS: exactly
12:14:27JdGordoni havnt but it works here... so maybe you missed a commit or something?
12:14:33pondlifeSounds like buffer_alloc to me!
12:14:47jhMikeSsomethings permanently damaged after a file load
12:16:07pondlifeJdGordon: It works here too now, but I've found another bug...
12:16:21JdGordoncourse you have :'(
12:16:30 Join bun-bun [0] (
12:16:38pondlifeGo into Files and select the 3rd entry
12:16:45pondlifeNow go back into root menu and Database...
12:16:49jhMikeSbuffer_alloc? I'm not really sure. This codec recording will self destruct if something uses memory like that.
12:17:10pondlifejhMikeS: Not necessarily recording.
12:17:16jhMikeSGot Warning: 00000002
12:18:40pondlifeJdGordon: Basically moving up and down the Files or Database browser affects the default selection in the other. IMHO it should always start at either the last selection in that particular browser, or the 1st entry when you change.
12:18:55jhMikeSjust an encoder buffer overflow....which is perhaps telling.
12:19:08LinusNpondlife: that's another reason to separate the tree and tag browsers
12:19:31pondlifeYes, but maybe we can just seperate their running data and reuse code
12:19:44LinusNyeah, the tree context should handle that
12:19:48pondlifeI'm all for it, apart from any code size implications.
12:19:57LinusNi'm all for KISS
12:20:02jhMikeSI'll try to let it stop by itself however long that takes
12:20:02pondlifeIf the context can handle it all, we may not need to seperate code
12:20:04JdGordonpondlife: no problems here moving between the 2
12:20:49JdGordonwe already have 2 different code paths... ft_load and tagtree_load ... the only shared code is the display, which imho shouldnt care if its a db or file browser
12:21:23*JdGordon thinks every screens drawing code should be seperate so we can have hi-graphix version and lite versions
12:21:38*markun read some talk about eSpeak!
12:22:11jhMikeSThis could happen if a thread were overprioritized too
12:22:24pondlifeJdGordon: Go into Database and select the top entry, go back into Files, move down one and go back into Database. What is selected?
12:22:58JdGordonhehe so it is... it looks like its only that way
12:22:59markunLinusN: I would like to mentor for soc. TTS interests me a lot.
12:23:05JdGordonmoving in the db doesnt effect the tree
12:23:11JdGordonwhich does make a bit of sense
12:23:12pondlifeJdGordon: It does do
12:23:31JdGordonnot here
12:23:45pondlifeJdGordon: in Database, scroll upwards (and maybe wrap round), it will affect the Files view
12:24:00pondlifeSeems to be some kind of arcane adding machine :)
12:24:08JdGordonanyway, this bug is minor and not worth the effort to fix... dirbrowse() needs rewriting
12:25:07pondlifeJdGordon: I disagree somewhat. A fix would be nice if you can peek in there.
12:25:19JdGordonhehe thought you might say that
12:25:27pondlifeIt used to work (I think).
12:25:39pondlifeYou thought I'd say that too, I bet.
12:25:49pondlifeAnyway back to Work
12:26:13 Quit inversions (Read error: 110 (Connection timed out))
12:26:15JdGordon <- /me just stumbled on this
12:27:19 Join idnar [0] (n=mithrand@unaffiliated/idnar)
12:27:20fejfighterJDGordon: he got his wish now
12:27:33*bluebrother wouldn't have closed that as fixed as there wasn't anything to fix ...
12:27:47bluebrotherbut rather changed :)
12:28:23JdGordonthere isnt really any better option tho... close enough :p
12:29:09 Join akaidiot [0] (
12:29:11bluebrotherno, "out of date" −− superseded by the RB menu :P
12:29:14 Join _Veseliq_ [0] (
12:29:20bluebrothermuch closer IMO
12:29:21 Part akaidiot
12:31:41jhMikeSbuffer_alloc doesn't seem likely here to me...maybe the PCMREC_PARANOID will show something though
12:33:05 Join Thundercloud [0] (n=thunderc@
12:33:41 Join Lynx_ [0] (
12:35:09SoapJdGordon: the other day I tried to make a build for my iPod video w/o TagCache/Database by removing the line "#define HAVE_TAGCACHE" from /firmware/export/config-ipodvideo.h. The build appeared to barf in the menu. Is this a know issue, or is the ability to disable tagcache building no longer supported?
12:35:39JdGordonit should be... probably just missing a few ifdefs in root_menu.c
12:35:50JdGordonscorche was keeping that updated i thought?
12:36:49SoapI can ask him. I'll try again tonight to verify in current SVN before asking, though.
12:37:19JdGordonim pretty sure its still a problem
12:37:35JdGordon5 pages of feature request and nothing interesting :'(
12:38:15 Join lee-qid [0] (
12:38:53pondlifejhMikeS: If it's not really buffer_alloc, maybe more that something is holding onto an invalid buffer pointer?
12:39:06 Quit fejfighter ()
12:39:22pondlifeJdGordon: You looking at new features again?
12:40:05*pondlife thinks it's time for a feature freeze ;-p
12:40:16pondlifeOnly joking
12:40:41safetydanJdGordon, dan in the commit logs is me
12:41:01safetydanand sim backlight depends on indexed colour modes
12:43:17jhMikeSpondlife: something can hold onto anything it's not writing to memory. flushing literally takes forever like a thread is just running way too much.
12:44:37 Join pondlife1 [0] (
12:44:37 Quit pondlife (Read error: 104 (Connection reset by peer))
12:45:18pondlife1jhMikeS: I meant that something took a copy of the ptr to the audio buffer, then a later alloc moved the buffer up.
12:45:35jhMikeSafter recording started?
12:46:17pondlife1Maybe nothing to do with recording at all.. Voice, dircache, database...?
12:46:18jhMikeSonly if it took a huge hunk of memory
12:46:52 Quit GodEater_ ("CGI:IRC (Ping timeout)")
12:46:55JdGordonsafetydan: the 16bit displays cant have backlights in the sim?
12:47:01pondlife1Or happened to be pointing at something important that was only shuffled up when a large voice file was allocated.
12:47:28safetydanJdGordon, basically no. Not with the way it's done at the moment where backlight is a palette swap
12:47:37JdGordon:( ok
12:48:22 Join Nico_P [0] (
12:49:41jhMikeSpondlife1: not quite sure what you mean by that
12:49:50pondlife1Me neither :)
12:50:16 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
12:51:09jhMikeSIt's sooooo like a thread just eating CPU cycles away...interesting the codec was able to overflow the flushing under these conditions
12:51:54pondlife1OK. Say thread A does something that needs to know where the audio buffer is. It is stupid and just takes a copy of the pointer. Thread B then does a buffer_alloc. So Thread A now has a pointer to thread B storage, not the audio buffer....
12:52:19pondlife1That scenario caused a slowdown in the past.
12:52:33pondlife1No idea why, but it was totally reproducible at the time.
12:52:42pondlife1Assumption was that it was IRAM related.
12:52:48jhMikeSyou mean noone ever found out?
12:52:54 Quit anathema (Read error: 113 (No route to host))
12:53:24pondlife1No. The problem was fixed I think. It was the bit of playback.c that sets up the codec buffers IIRC.
12:53:40jhMikeSbig suprise *yawn* :)
12:53:46pondlife1Indeed :(
12:53:57pondlife1I don't have the source in front of me at the moment.
12:54:12pondlife1Hence the vague description.
12:54:18jhMikeSthat parts different now anyway
12:54:31 Quit safetydan ("Ex-Chat")
12:54:43 Quit pondlife1 (Read error: 104 (Connection reset by peer))
12:57:26 Join pondlife [0] (
12:58:43 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
13:00:06jhMikeSIf the recording screen is the startup screen, it doesn't occur.
13:00:34pondlifeHmm amazing how a mere UI option can affect the kernel stability, isn't it!
13:01:37*pondlife thinks the initialisation sequence may need documenting
13:01:37jhMikeSWell, the screen gets started earlier that way...I'm pretty much open to any possiblity 'till I get some results probing stuff to rule things out. Don't know where to start yet.
13:02:53jhMikeSI did reformulate the audio init to a single stage...making it simpler and much smaller...and the last item initialized. But it's just sitting in a patch on my hd right now.
13:03:13pondlifeIt's not just the audio init, it's the whole caboodle.
13:03:31pondlifeDatabase, dircache etc. etc.
13:03:35jhMikeSNothing in init will bother it if it's the last thing though.
13:04:07pondlifeAlways surprised me how many bugs came from having "Resume on startup" enabled.
13:04:46pondlifeMainly because the init sequence wasn't the same for some reason.
13:05:25jhMikeSThe simpler init right before entering the tree would keep the sequence that same always...perhaps I should commit that.
13:05:49pondlifeIt should always be the same, and in no way decided by the start screen.
13:06:54 Join bun-bun_ [0] (
13:07:22jhMikeSThe main init would be long done before entering that stage. Once entering interactive UI code, anything should be able to be shown.
13:08:07***Saving seen data "./dancer.seen"
13:09:52 Quit bun-bun (Read error: 110 (Connection timed out))
13:10:36 Quit linuxstb (Read error: 113 (No route to host))
13:11:13jhMikeScycling in and out of recording doesn't affect the audio buffer size...always 10422040 on the x5
13:11:53JdGordonthe first screen is started from root_menu() which is called in tree_init() (iirc), but could be moved to the very end of main()
13:13:39jhMikeSno screens should ever start until all other init has happened
13:15:42jhMikeSturning voice menus off seem to allow the screen to operate normally reboot...if I start playback before starting recording.
13:17:23jhMikeSinteresting...the voice file space is allocated in either case anyway
13:17:43pondlifeIt will not be unallocated until you reboot
13:17:58pondlifeA standard buffer_alloc thing
13:18:05jhMikeSno buffer pointers are being move...always the same is my point
13:18:30pondlifeYes, but the voice file buffer is no longer being used
13:18:46jhMikeSvoice files don't use buffer alloc...just the thumbnail does
13:18:46LinusNmy guess is that the voice thread goes haywire and eats cpu
13:19:21pondlifejhMikeS: Ah, I thought the main voice file was buffer_alloc-ed nowadays?
13:19:30jhMikeSLinusN: that's in my head too
13:19:53jhMikeSpondlife: no, then nothing can steal voice and use that buffer space
13:20:10pondlifeGood point.
13:20:25*pondlife hasn't looked at that code for months now, and doesn't miss it.
13:20:27jhMikeSRecording does...and besides plugins, it's the only thing that does
13:20:45pondlifeSo voice is still running for some reason...
13:20:53jhMikeSwonder if any plugins will go haywire then
13:21:39pondlifeI thought recording made calls to playback that waited for the voice codec to complete and unload? It shouldn't be doing much more after that.
13:21:57jhMikeSit does...or should be
13:22:18 Join theli_ua [0] (
13:22:23LinusNi suspect that is where it goes wrong somehow
13:22:46pondlifeSo time to put some logfs in the shutdown code...
13:22:59pondlifeAnd pray it doesn't affect the timing too much
13:23:47jhMikeSthough I'm wondering if the voice codec is swapped out somehow at the wrong time and letting audio codecs load...this is where I need some rest and have to "become" the threads :)
13:23:47 Quit decayedcell ("CGI:IRC (EOF)")
13:24:28 Quit theli_ua (Client Quit)
13:26:52pondlifeSounds dangerous :-(
13:27:40pixelmaspeaking of large voice files... it's nice that someone provides updated voice files for now but it's a pity that they won't work on most Archos devices. Even the smallest would be unusable on my OndioFM - I guess it would work on the Player and with luck on a flashed OndioSP...
13:27:41jhMikeSonly if they make a noose
13:28:08pondlifepixelma: Will langv2 help out?
13:29:15LinusNlangv2 will shrink the archos voice files considerably
13:29:20amiconnYes when properly applied to voice
13:29:41*amiconn should really get these per-target voice changes into place
13:30:28pixelmaI'm quite sure - but he could also lower the quality I guess for the moment. The comments say it sounds very good so maybe he could provide one lower quality file (unless it's not understandable it will be better than a few items not voiced)
13:36:04 Join Ne-X-us [0] (
13:37:03pondlifeIs the playlist viewer a plug-in?
13:37:21 Join inversions [0] (
13:37:23pondlifeOK, so PLAY -> WPS should work there then
13:37:41LinusNguess so
13:38:00pondlifeI tried it without thinking, and was surprised when it did nothing.
13:38:43 Join datachild` [0] (
13:45:53jhMikeSthe voice thread appears to be waiting for the mutex like it should or I'd be getting "Haha!" plastered all over the recording screen
13:46:39LinusNjhMikeS: ok, so maybe it something else, like a dma interrupt from hell?
13:46:41jhMikeSI just get the splash when voice first retakes it...just like it should
13:47:32jhMikeScould be...but that would cause a pcm buffer overflow no since each interrupt increments the position. that happened in adding recording to the iPods and it caught the overflow
13:47:38LinusNi see
13:48:32 Quit Nico_P (Remote closed the connection)
13:48:58 Join helppme [0] (i=gotoXY@
13:49:02 Quit datachild (Read error: 110 (Connection timed out))
13:49:23helppmeanyone here who can help me? ive constantly searched the net for answers but it seems as if no ones having the same problem :/
13:49:39pondlifeTry viagra ;)
13:49:47LinusNhelppme: shoot
13:50:06helppmeso here it goes,
13:50:09desowinis it connected with #rockbox ?
13:50:14helppmei installed rockbox perfectly then
13:50:22helppmei tried flipin on some pics and they loaded
13:50:34helppmebut whenever i select mp3s, they dont play, they display some error and then hang
13:50:43desowinwhich player ?
13:50:46LinusNhelppme: which player?
13:50:49pondlifeIPOD with old bootloader?
13:50:54LinusNsounds like it
13:51:41helppmewhat do you mean player? isnt rockbox built in with a player already?
13:51:48*pondlife thinks "Maybe the COP init code could check the bootloader version and warn the user if they've not upgraded...?"
13:51:51desowinwhich device ?
13:52:07desowinso iPod then
13:52:22helppmewhenever i select a song it displays
13:52:32helppmedata abort at 000307c8
13:52:36helppmethen hangs :/
13:53:24LinusNhelppme: which instruction did you follow when you installed the bootloader?
13:53:35desowinhave you been using rockbox before ?
13:53:52helppmethis is my first time using rockbox
13:54:15LinusNhelppme: read the red text on top of that page
13:54:42helppme:/ i missed that
13:54:49desowinyou'll need to install new bootloader, I think
13:54:50LinusNincredible :-)
13:55:02pondlifeNot red enough :)
13:55:14helppmesorry for wasting your time guys, but care to shed some light on where i can find the updated instructions :/
13:55:27pondlifeFollow the link in the red text?
13:55:27GodEater_OT Question: Anyone know if it's possible to get grep to match across line boundaries? I want the a line which matches my pattern, and the following two lines, but no more?
13:55:30LinusNhelppme: there's a link in the red text
13:55:38desowinthere's blue link on red text
13:55:52helppmeokay thanks, :)
13:55:59*LinusN giggles
13:56:01helppmeive read that rockbox doesnt support mp4 right?
13:56:15JdGordonwhat is everyones opinion on ? and assuming it uses too much space, only on the MEM > 8 targets ?
13:56:37jhMikeSmaybe has "Underline Links" turned off?
13:56:44pondlifeJdGordon: I don't like it using 2 lines per option.
13:57:06GodEater_Nor me
13:57:07pixelmait looks ugly...
13:57:17JdGordonhow could you do it in only 1 line tho?
13:57:19pondlifePrefer what we have, where the settings and browser actually look similar.
13:57:19pixelmaand wastes space
13:57:27pondlifeJdGordon: Simple. Don't do it!
13:57:30GodEater_I like that it shows the current setting on the same screen though
13:58:07pondlifeIt's hardly a big benefit. Right-left will show you it at the moment. And makes sense with voice
13:58:32LinusNhelppme: mp4 audio (AAC) works fine
13:58:43pondlifehelppme: As long as it's not DRMed.
13:58:59helppmethen how will i be able to play my mp4 videos? :/ convert them?
13:59:11GodEater_reboot into Apple OS ?
13:59:20LinusNhelppme: use the apple firmware, you can dual boot
13:59:23pondlifehelppme: You'd be best off using the Apple OS for video.
13:59:43pondlifeWow, we all agree on something
13:59:51*GodEater_ notes the date and time
14:00:03pondlife6th March?
14:00:05helppmehow to boot on apple os when im on rockbox? menu+select?
14:00:23GodEater_no, just hold down play to shutdown
14:00:31GodEater_and then restart with hold switch on
14:00:48helppmecool thanks :)
14:00:49*pondlife must try Rockbox on an iPod one day
14:01:07*GodEater_ remembers his lent pledge
14:01:08*JdGordon prefers it on the h300
14:01:15*pondlife too
14:01:27amiconnJdGordon: In-line settings are ugly either way if the screen is small (and small means anything smaller than the H300 / ipod colour lcd resolution)
14:01:37pondlifeI find it hard to imagine a keymap without playing it.
14:01:46JdGordonits nice on the big screens tho
14:01:51pondlifeplaying *with* it
14:02:05pondlifeJdGordon: What's got a big screen then? 5G?
14:02:15JdGordonyeah :p
14:02:15pondlifeI would consider the H300 a small screen
14:02:22JdGordongb is the wrong aspect for it
14:02:24pondlifeBut then I'm somewhat blind
14:02:28JdGordonsorry.. no
14:02:31JdGordongb is perfect
14:02:48GodEater_GB is perfect for the two line approach even
14:03:15GodEater_so - I got a shiny new Cisco IP Phone on my desk yesterday. Anyone fancy hacking the firmware on it ? :)
14:03:17*pondlife suggests, tartly, that bug fixing should take prioirity over ugly new features
14:03:35*pondlife can't spell and apologises.
14:03:58LinusNpondlife: i agree
14:04:01*pixelma likes this attitude :)
14:04:17pondlifeWe are almost at 3.0, right!!?
14:04:23pondlifeSo near....
14:04:36LinusNwe have to tune in on the new menu system and iron out those issues before adding more stuff
14:04:44JdGordononly about 4 bugs ont the tracker for 3.0
14:05:02pondlifeYes, and the USB bootloader one can probably be closed
14:05:21pondlifeLinusN: Do you reckon?
14:05:32LinusNJdGordon: have we discussed the idea of toggling between the current screen and the menu with the Manu button?
14:05:38pondlifeOr wait until you do a release.
14:05:49LinusNMenu even
14:06:01JdGordonIt was decided to do that wasnt it?
14:06:11LinusNi think it would be nice
14:06:21pondlifeSo there's your shiny new feature ;-p
14:06:22LinusNjust wanted to check if we agreed on it
14:06:27JdGordonwtf? my h300 just turned off while charging?
14:06:42LinusNJdGordon: bootloader?
14:06:47jhMikeSit occurrs to me to ask the question: where the codec malloc buffer go in all the swap stuff? either I'm really tired or it's really just ignored alltogether.
14:07:01JdGordonof course by charging i mean the ac is plugged in... even if its not plugged into the wall :p
14:07:02pondlifeIt's not used for MP3 voice IIRC
14:07:10JdGordonfalse alarm :p
14:07:16LinusNJdGordon: lol * 2
14:07:26*pondlife has made that mistake before
14:07:42pondlifejhMikeS: It's probably not an issue until we use speex
14:07:50jhMikeSI never looked to see if mpa codec uses it or not :P
14:08:02 Quit aliask ("Chatzilla 0.9.77 [Firefox]")
14:08:07JdGordonLinusN: did we agree to finally make left not cancel settings?
14:08:09pondlifeWell, it didn't used to ... when I last looked. Maybe it does now
14:08:40JdGordonit did pre august
14:08:46pondlife(sorry that was to jhMikeS)
14:09:20JdGordonLinusN: menu > previous... is that only if your in the actual root menu? or anywhere in the menus?
14:09:37pondlifeAnywhere at all.. WPS, browser...
14:09:39 Quit helppme ()
14:09:49pondlifeMENU once to go to root. MENU again to go back to where you were
14:10:06jhMikeSJdGordon: I agree that I disagreed with the idea, but have grown sort of fond of it...not that you asked me anyway.
14:10:07pondlifeLike the old settings menu did.
14:10:28JdGordonhmmm... this could be tricky
14:10:34 Part pixelma
14:10:36pondlifeSo basically root menu MENU goes back to wherever
14:10:53JdGordonbut in setting? MENU goes where?
14:10:56JdGordonto root or previous?
14:11:08JdGordonsay root and ill fix another treee bug ;p
14:11:09pondlifeRoot I think
14:11:16jhMikeSRoot menu should just go back to a random spot...or at least have an option for it :)
14:11:34pondlifeRecord on H300 should act like a random button
14:11:40*bluebrother still doesn't know of a root menu in Rockbox
14:11:45JdGordonGO_TO_RANDOM ?
14:11:46pondlifeMain menu, sorry
14:12:05pondlifeBut the code is in a file named root_menu.c
14:12:39bluebrotherI know. But users will hang around here, and we can cause great confusion if we permanently use the "wrong" name
14:12:50bluebrotherwhich includes a couple of other people as well.
14:12:51JdGordonRockbox Menu is too long to type in convo
14:13:08ShaidI think the root_menu.c needs to be renamed
14:13:10pondlifeMain Menu - it says so in the Start Screen
14:13:21bluebrotherRB menu is even shorter, and everyone in here should understand that
14:13:43 Join pearldiver [0] (
14:13:54bluebrotherThere _are_ ways to call it in a short and non-confusing way
14:14:08JdGordonShaid: na, code wise it IS the root menu
14:14:25bluebrothercode wise?
14:15:04LinusNi agree that getting back you where you were by pressing Menu again is a little tricky when you are deep down in the settings menu tree
14:15:27bluebrotherif it is I'm not sure if that should considered somewhat broken −− the code should reflect the UI in some way ...
14:16:25 Quit Hoffmann ("Ciao")
14:17:19GodEater_I think clearly we're waiting for a target that comes with 140 buttons, which can each have one setting assigned to them.
14:17:59*Shaid starts the x86-based pc port just for that.
14:18:27 Join thewho [0] (i=c27f0812@gateway/web/cgi-irc/
14:18:32GodEater_only 113 buttons here
14:18:49jhMikeSmad_layer_III, receive, and mad_decoder_run call malloc, the first calls calloc too :\
14:19:28thewhorg. setting values in the same screen: could it be done so that the setting name and its value are shown every 2 secs for the setting under cursor?
14:19:37ShaidI go sleep
14:20:01bluebrotherthewho: there was a patch that used two lines: the first showed the setting and the second showed the value right aligned
14:20:09*bluebrother liked that patch pretty well
14:21:14thewhobluebrother: that could be a problem for smaller screens.
14:21:46 Quit Shaid ("I will see you, in the end. And I will laugh at your pain...")
14:22:03thewhoBuf for large screens (newer ipods, gigabeat) it would be also ok if each setting would take up two lines. Couldn't we manage this with #ifs?
14:22:12bluebrotheron small screens you usually won't want to display the values in the same screen
14:22:16pondlifebluebrother: We were discussing it earlier... problem for small screens or large fonts.
14:22:32bluebrotherpondlife: earlier today? Ok, going to check the logs
14:22:49pondlifeAbout 13:00
14:22:56bluebrotherthewho: that patch added a setting for one or two-line display.
14:23:44 Quit cadu (Read error: 104 (Connection reset by peer))
14:23:57amiconnThere is already a screen that uses in-line settings, and imho it is the ugliest screen in rockbox
14:24:06thewhobluebrother: both are displayed at the same time? Then it would also be a problem on small screens
14:24:24thewhoWhy not show them
14:24:45jhMikeSwell, it's ugly but it's the implementation
14:24:47thewhoLike in the filebrowser: the file name under cursor is crolled
14:25:06bluebrotherscrolling is evil :P
14:25:18bluebrotherthe h100 has a quite slow lcd. Scrolling lines are a PITA
14:26:08thewhobluebrother: not being able to see the whole file name is an even greter PITA
14:26:41bluebrotherI know, but if there is enough room that shouldn't get restricted by merging two values on the same line
14:27:48thewhobluebrother: you can never know since a large (and wide) font can be used
14:27:48 Part jhulst ("Kopete 0.12.4 :")
14:27:51linuxstb_Sorry if someone has mentioned it, but what about large LCDs with small remote LCDs?
14:28:20JdGordonscreen dependant...
14:28:24thewholinuxstb_: also good point. Showing them would also solve that
14:28:48thewhoIs there a standard abbrev. for (like I18N)?
14:28:55bluebrotherhow exact do you mean that?
14:29:33JdGordondoes this look more KISS than the current root_menu() switch?
14:29:55JdGordonthat replaces everything untill the comment above global_status. ...
14:30:47thewhobluebrother: for the setting under the cursor, first the setting name is displayed for 2 secs, then its value for 2 secs, then the name again etc. Or for 1 sec. Or waht you like
14:31:23bluebrotherhmm. Not sure if I like that, but I would try it if there is a patch
14:31:42bluebrotherjust looked at the logs, I really liked the way of the screenshot JdGordon posted
14:32:13bluebrotherit's fine for my players display, and that patch allowed to turn it off.
14:33:58bluebrotherah, found the patch. FS #5833 if someone else is looking.
14:34:23JdGordonthat wont patch properly anymore so wouldnt bother
14:34:25 Quit inversions (Read error: 110 (Connection timed out))
14:34:40JdGordoniirc he brought in some horrible macros also to do ti
14:34:52bluebrotherI know. But it has some discussion attached ;-)
14:35:12bluebrotherJdGordon: I can't believe someone will be able beating you at macros :P
14:36:09bluebrother was the forums discussion about it
14:36:58*JdGordon hopes for a green delta table...
14:37:32bluebrotherbtw, the trigger setup screen already has some values-in-menu style
14:39:00preglowhrm, didn't v2rec use to be green?
14:39:32thewhobluebrother: when exactly were the screenshots sent?
14:40:15bluebrotherthewho: see the forum topic ... last august
14:40:54bluebrotheralso have a look at the first screenshot in that topic ... it's a one-line style
14:41:14JdGordonpreglow: not for a loong time
14:42:18thewhobluebrother: yes, looks fine. If only the font is not too wide...
14:43:19bluebrotherif it's possible switching between one and two line displays that should be a good solution for all combinations
14:43:23JdGordonsweet... that saved a whoppiong 32 bytes :D
14:43:35bluebrothermaybe have a setting one line / two lines / separate lists
14:43:54bluebrotherso one can still have the old style
14:43:56JdGordonand on that note im going to bed... cyaz
14:43:57thewhobluebrother: two lines per setting would be a waste IMHO
14:44:06 Quit JdGordon ("Konversation terminated!")
14:44:21 Quit XavierGr ()
14:44:30bluebrotherI like the two lines per setting as it gives me a much faster overview and isn't that crowded with larger fonts
14:45:13bluebrotherand one line for setting name and value is sometimes quite short on smaller displays. Even for the h100
14:45:30bluebrothertried that myself the times that patch applied cleanly
14:46:19thewhobluebrother: I use nimbus-19 on my H120. That will probably explain my concerns :-)
14:47:59*bluebrother doesn't even have nimbus-19 on the device
14:48:17austriancoderfor this some kind of "abstraction" of the rendering would be cool.. so everybody can easily change the look of the UI - maybe with an extended wps or skins
14:48:19 Nick datachild` is now known as datachild (
14:48:34bluebrotheryeah, it would be a waste for that large fonts. But I don't see a problem if you can turn it off.
14:48:40markunaustriancoder: I think so too
14:49:28bluebrotherputting the value in the lists for that lcd with that large fonts will be painful too.
14:50:05austriancodermarkun: i think we are the only two - so we need to start hacking on our own
14:50:40thewhobluebrother: that's why I thought about showing them one at a time
14:50:45preglowanyone know what datasheet we used for writing the wm8975 driver?
14:51:02bluebrotherfor the font I use (nedore-9) the two line approach looks best
14:51:30*bluebrother still thinks having optional different views would be nice
14:51:40linuxstb_Any ipod users willing to test the latest bootloader release? Windows version of ipodpatcher including the new bootloaders is here - - linux (32-bit) version is here - and the bootloaders themselves are here -
14:52:02thewhoSo there could be 4 options: none, in the same line, two separate lines, and alternate
14:52:02linuxstb_Especially useful are mini (1st or 2nd gen) and 4g greyscale owners to test if the original firmware can be started.
14:52:15linuxstb_preglow: Is that the Nano's codec?
14:52:26preglowlinuxstb_: aye
14:52:26linuxstb_preglow: Then I think it's the 8750.
14:52:43bluebrotherwith "none" meaning separate list of options as we currently have
14:53:00 Quit GodEater_ ("CGI:IRC (Ping timeout)")
14:53:18thewhobluebrother: yep. The current state is "none"
14:53:34preglowlinuxstb_: just thought i'd check out if there's a good reason bass and treble controls work like they do
14:53:45preglowif not, i'll replace them with the software ones, since they're pretty unusable
14:54:00bluebrotherthat setting will definitely be related to the used font / theme, so it should be a setting
14:54:45thewhobluebrother: you mean "a _theme_ setting"?
14:55:19bluebrotheryes, as themes usually set the font (i.e. require a specific font)
14:55:41thewhobluebrother: agreed
14:55:54linuxstb_Shouldn't it be up to the user to decide if it looks good or bad with the current font/language/LCD size etc etc
14:56:32bluebrotheryes. That's what I meant: depending on the settings the user chose the display style might be different
14:56:50thewholinuxstb_: yes. The user sets it as he/she likes. And then the setting is saved along with the theme if "sava settings" is performed
14:57:08thewho*"save theme"
14:57:33bluebrotherthe way themes work now it wouldn't be strictly required to save that setting −− the user can add it if he wants to
14:57:51bluebrotherbut if we should save that per default along the theme file is a different thing.
14:58:09bluebrotherat least I'd like to have the style options.
14:59:15thewhobluebrother: IMHO we should
14:59:57 Join Febs_ [0] (
14:59:59 Nick Febs_ is now known as Febs (
15:00:06 Quit thewho ("CGI:IRC")
15:02:13 Join SirFunk [0] (
15:05:33pregloweh, wtf
15:05:38preglowauto resume on alarm no longer works
15:05:46preglowno wonder i didn't wake up when i wanted to today
15:06:57 Part Ne-X-us
15:08:08***Saving seen data "./dancer.seen"
15:10:04markunpreglow: do you keep your headphones on all night?
15:10:42preglowmarkun: they're on the floor
15:10:46Kasperlejust connect it to your stereo :)
15:11:02 Quit perplexity (Read error: 110 (Connection timed out))
15:12:08markunKasperle: I discovered that headphones + cardboard cylinders also work somewhat as speakers last week :)
15:12:19peppoany use asking to test the ipod 80gb-video code? is it in svn?
15:12:21*preglow summons jdgordon
15:12:22peppoare there any snapshots?
15:13:12Kasperlemarkun: hehe, nice :D
15:13:41Kasperlemarkun: getting the wakeup timer _and_ the piezo to work would be nice
15:19:58preglowLinusN: still agree that playback resume should be the action taken on alarm wakeup? seems jdgordon removed it when he commited rootmenu
15:20:36linuxstb_peppo: No, there is no code to test. The only person with the code is the person writing it, and I'm sure that as soon as it's working on his ipod, it will be committed to SVN. Before that point, there is nothing to test.
15:20:44LinusNit should probably go to the selected screen
15:20:51LinusNthe start screen
15:21:10LinusNthen you can select the recording screen and have it record, can't you?
15:21:41preglowi doubt the recording screen will automatically record
15:21:57preglowbut i don't want my player to always resume playback
15:22:01peppolinuxstb_, okay. thanks for the information
15:22:07preglowi want it to do that only when waking up, and that doesn't sound too far-fetched
15:22:12LinusNi was under the impression that the "recording screen at startup" setting was for that purpose
15:22:19preglowwell, i'll try
15:22:30preglowi doubt it
15:22:45 Part peppo ("part")
15:23:22LinusNpreglow: in any case, scheduled recording should be implemented in a better way anyway
15:23:36preglowstart screen doesn't even stick here
15:23:45preglowit always starts in the root
15:23:50preglowperhaps i should update my build
15:24:50preglowbut yeah, _some_ action should always be implicit on alarm
15:24:57preglowit just waking up to the root menu is pointless
15:25:16pondlifepreglow: It can loudly say "Files" to you :~
15:25:19 Join theli_ua [0] (
15:25:32preglowor start formatting the hd
15:25:34preglowthat'd make me wake
15:26:03 Quit theli_ua (Client Quit)
15:27:18preglowB4gd3r: yo, the left menu on the webpage seems to be an older version while viewing posts in the ml archive
15:31:35 Quit austriancoder ("Lost terminal")
15:31:49preglowLinusN: setting recording as starting screen does absolutely nothing. the settings is even forgotten
15:33:47LinusNsounds like a bug
15:35:07preglowi can only find minimal code using the setting at all
15:35:29preglowthough root_menu() does return it
15:36:16preglowand the return value of that is ignored.......
15:36:28preglowi'm not exactly impressed
15:36:34 Part LinusN
15:37:12preglowa, misnamed variable
15:38:07 Quit jaebird (Read error: 110 (Connection timed out))
15:38:10preglowbut there's no handler for GO_TO_RECSCREEN anyway
15:39:03amiconnbluebrother: The trigger screen is this most ugly screen I was talking about...
15:39:58amiconnI had another idea for _showing_ the value of the setting under cursor (there would still be a separate screen for setting it)
15:41:19amiconnWell, in fact it's not really my idea because it's just what my mobile phone does: it shows the current value of the setting under cursor in a little framed box (similar to our splash boxes) in the line below the cursor (or above the cursor if the cursor is on the last line)
15:41:40pondlifeHmm, preview pane?
15:42:16amiconnFor changing the setting you have to enter it, like in rockbox
15:42:40 Join Criamos [0] (
15:44:34 Join jaebird [0] (
15:46:14 Quit bluebrother (Read error: 104 (Connection reset by peer))
15:47:44amiconnlinuxstb_: Just swapping the condition fixes loading retailos from file?? That sounds more than fishy...
15:49:04 Join bluebrother [0] (
15:49:08preglowthere seems to be something extremely finnicky about retailos loading
15:49:14preglowwe've had trouble with it before
15:49:24preglowalways very mysterious shit
15:49:48preglowi think perhaps retailso expects some (probably scratch) registers to contain info
15:52:23amiconnWhat about just saving the whole register file on bootloader entry, and restoring it before calling retailos?
15:53:21preglowlinuxstb said he'd try that
15:53:36 Quit rp- ("leaving")
15:53:36preglowmight want to do the same with iram
15:56:09preglowthe bootloader button detection is bloody driving me crazy
15:57:03 Join XavierGr [0] (
15:58:34 Join wookieface [0] (
15:58:44bluebrotheramiconn: seems my network connection is causing trouble atm. Just looked in the log and found this idea for showing the value inline. Sounds nice.
15:59:32bluebrothersimilar to the two lined approach I like but without inline changing of the value
16:00:03wookiefacehi, i get this error message when trying to install the bootloader on an iPod video: Short read in disk_read
16:00:33linuxstb_amiconn: No, the fix I committed was because the bootloader was checking for a value > 0 for success, whereas 0 was the code for success.
16:01:14linuxstb_But dan_a said this morning that my commit to disable lcd_update() caused retailos to be happy again on the minis and 4g greyscale. But I've not had confirmation that it's still working after my other changes.
16:01:53 Quit Febs (Read error: 110 (Connection timed out))
16:01:59linuxstb_You're right that it's fishy, but it's hard for me to test as it works fine on both my ipods.
16:02:29preglowthe nano clicks something awful on boot compared to retailos
16:04:19linuxstb_Which part of a Nano can click?
16:04:40pregloweh, wm8975 :>
16:04:47linuxstb_Ah :)
16:04:58preglowincreasing the wait time for vref makes it somewhat better
16:05:03linuxstb_I was thinking of hard-disk clicking...
16:05:06 Join Domonoky [0] (
16:05:09preglowbut spending one second in waiting for sound to stabilise is a bit muc
16:06:23linuxstb_I always turn on my ipod before putting the headphones in, so never notice things like that.
16:07:32 Join perplexity [0] (
16:12:40preglowdon't often notice it myself, but is very annoying when i do
16:12:45preglowthe nano emits several clicks on start
16:13:00preglowdespite that the driver follows the datasheet recommendations fairly well
16:13:26linuxstb_Assuming the datasheet is right for the 8975...
16:13:53preglowthe treble bass controls definitely aren't
16:14:00preglowthey sound like a gain, basically
16:15:48pregloweh, now wait a minute
16:16:18preglowno, it's right
16:17:35preglowthe controls would be pretty shitty even if they worked
16:17:36preglowdon't map well onto the rockbox bass/treble range at all
16:18:53 Join perl|work [0] (
16:20:43wookiefacehi, i get this error message when trying to install the bootloader on an iPod video in linux: Short read in disk_read
16:22:01linuxstb_What command are you typing to get that message?
16:26:03preglowwhat the hell does "TBD" in datasheets mean?
16:26:06preglowto be disclosed?
16:26:14linuxstb_to be decided?
16:26:30amiconnto be determined afaik
16:26:32preglowdoesn't really matter much, i guess, the actual figure is nowhere to be seen anyway
16:26:43amiconnI.e. not yet known at the time of writing
16:26:54preglowand then they go ahead and publish the datasheet
16:28:11 Quit webguest94 ("CGI:IRC (EOF)")
16:31:09Domonokyhi, is the M5 Bootloader installation the same as X5, only other bootloader ?
16:34:45Domonokyand the M5 bootloader should be copied to so rbutil can install it.. :-)
16:35:11preglowlinuxstb_: so, ipod4g uses the same codec as in the nano? does bass/treble sound like they actually work there?
16:35:40linuxstb_I think so, yes.
16:36:03amiconnDomonoky: Yes, the procedure is exactly the same, only the filename is different
16:36:04preglowcare to test? i'm going to just use sw tone controls on nano
16:36:44linuxstb_amiconn: Do you know if the M5 port has any users yet?
16:36:53Domonokyamiconn: nice, so for M5 support in rbutil i only need to change the ini file :-)
16:37:03linuxstb_preglow: Sure.
16:37:23 Quit wookieface (Remote closed the connection)
16:37:47 Join Febs_ [0] (n=chatzill@
16:37:49 Nick Febs_ is now known as Febs (n=chatzill@
16:40:01 Join Arathis [0] (
16:48:15 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
16:48:41linuxstb_preglow: I've just compared my Color with my H140, and the bass setting definitely behaves differently. On the ipod, the overall volume increases, but there does seem to be more bass. On the h140, the overall volume feels the same when the bass is increased.
16:49:01preglowthere does seem to be more bass indeed, but there also is more of everything else, so...
16:49:19 Quit scorche (Read error: 104 (Connection reset by peer))
16:49:22preglowi'm thinking of having a quick jump to windows to do an actual measurement later on
16:49:44 Join scorche [0] (i=ScorchE@rockbox/administrator/scorche)
16:49:55preglowbut ok, if i end up using sw tone controls on nano, i'll also do so on color and 4g, since the problem seems to be the same
16:50:26*Domonoky just commited M5 support in rbutil..
16:50:28markunpreglow: same on the Gigabeat
16:50:47Domonokynow only the bootloader needs to be copied to the download server
16:50:51linuxstb_Is the Gigabeat's codec unique to the Gigabeat (amongst Rockbox targets)?
16:51:04preglowthat also has 8975? i thought that had wm87 something
16:51:30markunit's register compatible with the 8975
16:51:57markunit's the 8751L
16:52:08amiconnThese wolfson codecs are mysterious...
16:52:15preglowso it actually seems like they just have really bad filters
16:54:00amiconnAt least the 8711/21/31 are known to have no tone controls
16:54:21 Join SimonSelki [0] (i=aa4c14fd@gateway/web/cgi-irc/
16:54:46SimonSelkiHey guys, I had a quick question about file viewing in rockbox
16:54:52markunI don't really hear a diffrence for all the negative dB values in the bass control
16:55:20markunsame for trebble
16:55:28preglowi can hear it, but only in overall volume
16:55:45markunI don't even hear a change in volume
16:55:53GodEater_linuxstb_: odd news from the apple forum
16:56:11SimonSelkiI noticed when I was viewing playlists that had songs transfered over using iTunes, their names were a scrambled bunch of letters.
16:56:21linuxstb_GodEater_: Which news? That the bootloader is mysteriously working again?
16:56:31GodEater_yes that news - especialy if you didn't fix it
16:56:41linuxstb_SimonSelki: Yes, that's what itunes does to your filenames.
16:57:32SimonSelkiI guessed, but is there anything I can do about it? I'm looking at a playlist right now and can't tell which songs are what
16:58:11GodEater_SimonSelki: only delete the iTunes music from the iPod and copy it over again properly
16:58:29GodEater_but then you won't be able to use AppleOS anymore
16:58:40linuxstb_At the moment, the only thing you can do is to stop using itunes and copy the files normally. But I think someone was working on looking up the track information from the Database when viewing the files - there's a patch in the patch tracker.
16:59:26SimonSelkidamn. If I do that, I won't be able to listen to them in the apple firmware, which I use a lot when my battery is low
16:59:37SimonSelkiI have a nano, so it's low a lot
17:02:50 Quit ender` (" One cup of ramen, / Two cups of boiling water, / Three minutes -- dinner.")
17:03:00markunpreglow: does the software tone control use the EQ or are the filters separate?
17:03:43 Quit GodEater_ ("CGI:IRC (Ping timeout)")
17:04:31 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
17:05:44 Quit lee-qid ("aufwiederbyebientotsayonara")
17:07:30Domonokywhat screen dimensions does the m5 have ?
17:08:12***Saving seen data "./dancer.seen"
17:08:40amiconn160x128 4gray
17:09:09amiconnSame resolution & depth as the h1x0, and same resolution as the X5 (but different depth)
17:11:06GodEater_that's handy - so there's a bunch of pre-done WPS's for it already
17:11:39linuxstb_Has anyone reported that when resuming a long mp3 file (I had the problem with a 68 minute mp2 yesterday) the elapsed time starts at 0, even though playback has resumed at the correct position mid-track?
17:12:37 Join toffe82 [0] (
17:15:19GodEater_I don't have any that long I don't think...
17:16:56GodEater_I have some 45 minute ones, are they likely to exhibit the problem ?
17:17:15linuxstb_I've no idea. I'm assuming it's not a problem with "normal" length files though.
17:17:37GodEater_does it happen when resuming anywhere ?
17:17:48GodEater_or do I need to try it when I'm already 20 odd minutes in ?
17:18:27linuxstb_I noticed it yesterday when I was about 20-30 minutes in. I stopped playback and shutdown my player. A few hours later, I turned it back on and resumed.
17:18:43GodEater_I'll give that a try - but it will take a while of course
17:19:00linuxstb_I'll test again now to see if simply seeking 30 minutes in, and then stopping will cause the problem.
17:20:52GodEater_resuming at 2:44 didn't trigger the error
17:20:54linuxstb_Yes, same problem. The progress bar seems to flash up to the correct location, but then immediately go back to zero once playback starts.
17:21:06linuxstb_(I seeked about 32 minutes into a 68 minute MP2).
17:21:48GodEater_I just seeked 24 minutes in - it's still fine here
17:21:55linuxstb_It's possibly just an mp2 issue...
17:22:19GodEater_just went to 32 minutes here
17:22:28GodEater_no - still fine
17:22:39*Domonoky has 2 xpm files (for rbutil) which somehow dont like diffing, or svn diff doesnt like the files.. does anybody have a idea ?
17:22:46GodEater_I have no .mp2 files to try it with
17:23:16linuxstb_Are you shutting down, or just stopping and resuming?
17:23:27GodEater_shutting down - as per your instructions
17:23:41linuxstb_:) No-one ever does what I say...
17:23:56GodEater_you're thinking of the forums :)
17:24:16linuxstb_I try not to.
17:24:25GodEater_that way sleepless nights lie
17:24:34GodEater_or possibly pace
17:25:07 Join webguest14 [0] (i=43bbe7a0@gateway/web/cgi-irc/
17:25:45webguest14is putting rockbox to sleep the same as shutting down, and if not will it help save battery power?
17:26:06GodEater_rockbox has no "sleep", it's either on or off
17:26:06linuxstb_Rockbox can't be put to sleep, only shut down.
17:29:11preglowmarkun: separate
17:29:19preglowmarkun: different design routines and all
17:30:47 Quit webguest14 ("CGI:IRC (EOF)")
17:31:47linuxstb_Hmm, even seeking 2 minutes into my long MP2 breaks resume...
17:32:12GodEater_is it an audio book or something ?
17:32:26linuxstb_A concert recording from digital radio.
17:32:41 Join ackbahr [0] (
17:33:03 Nick jborn_ is now known as JoeBorn (
17:33:08linuxstb_The other characteristics are that it's 48KHz and 160kbps CBR.
17:33:35linuxstb_(and yes, 160kbps MP2 does sound as bad as you would imagine)
17:33:47*linuxstb_ curses UK broadcasters
17:34:05linuxstb_Why can't we have nice 320kbps streams like Germany...
17:34:23GodEater_because then George would stop talking to Tony at dinner
17:34:57 Join tvelocity [0] (
17:35:03linuxstb_I need to buy and install an analogue FM aerial...
17:35:32GodEater_do you know - I can't remember the last time I listened to the radio
17:35:37 Join mattzz [0] (
17:35:38 Join Stalwart^ [0] (
17:35:38 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
17:35:57 Quit Thundercloud (Read error: 113 (No route to host))
17:37:10linuxstb_October 2004 was when I stopped listening regularly...
17:37:29 Join Juice^ [0] (n=Juice@
17:37:32GodEater_you seem to remember it well
17:39:05linuxstb_That was when John Peel died.
17:40:18GodEater_can't say I listened to him much
17:40:19preglowrip :/
17:40:49*GodEater_ contemplates a trip to Brick Lane one evening this week
17:41:05 Quit ackbahr (Remote closed the connection)
17:45:43 Quit SimonSelki ("CGI:IRC")
17:46:35 Quit tvelocity ("Αποχώρησε")
17:46:37 Quit Stalwart (Read error: 110 (Connection timed out))
17:46:47 Join Vyrus [0] (
17:48:56 Join x1jmp [0] (
17:48:59markunlinuxstb_: would be nice if the bootloader would detect when it's too old for a rockbox version: "Please update your bootloader"
17:49:27markunor too new for the installed rockbox
17:50:09linuxstb_Yes - dan_a thought he had done that, but it seems that it only worked on his ipod...
17:50:30linuxstb_Well, not exactly that, but falling back to running all threads on the main CPU if the COP isn't active.
17:50:45 Quit Vyrus001 (Read error: 131 (Connection reset by peer))
17:51:20linuxstb_But non-compatible changes like that are very rare - I can't recall it happening on any other targets.
17:51:50markunI think it would be a nice feature in general, it happend a few times on the Gigabeat
17:52:53markunusually the cache just doesn't work anymore and rockbox becomes slow
17:53:26linuxstb_We already (on all swcodec targets except gigabeat!) read a header from the rockbox.XXX file, so I guess that could specify a minimum/maximum bootloader version.
17:54:19linuxstb_But introducing that would in itself break all current bootloaders...
17:54:36 Join Stalwart_ [0] (
17:54:39linuxstb_Maybe something to consider for a 3.0 release.
17:54:53preglowi love the fact that saving common DIV64 expressions in variables saves a considerable amount of bytes
17:54:59*preglow dropkicks gcc
17:55:22markunand introduce a header for rockbox.gigabeat at the same time (as we forgot to do that)
17:56:15 Quit Ribs ("eh eh ehhhh!")
17:59:54 Quit Stalwart^ (Read error: 110 (Connection timed out))
18:00:00GodEater_I think it would be acceptable to do that sooner rather than later on the gigabeat - it's probably the smallest user pool currently
18:01:12Domonokyif someone likes to try: domonoky/"> <- rbutil with Theme installer (win binary)
18:01:29markunGodEater_: I wonder howmany there are..
18:01:52preglowmarkun: so, should i enable sw tone controls for gigabeat too?
18:02:06GodEater_we could probably get stats from the number of downloads of the current bootloader
18:02:14markunpreglow: unless there is a way to fix the hw tone controls..
18:02:30preglowwell, i've had a look, and what the driver does is identical to what the datasheet specifies
18:03:37 Join Thundercloud [0] (n=thunderc@
18:04:00linuxstb_Domonoky: Is there a forum thread about rbutil? I think you'll get more testers amongt users than developers :)
18:05:04Domonokyjeah i will put it in the forum
18:10:40amiconnlinuxstb_: DAB streams are never 320kbps. Afaik the specified maximum is 192kbps
18:10:55amiconn...and that's what most dab stations are using here
18:11:23amiconn192kbps or even 160kbps mp2 streams may sound quite good, it depends a lot on the encoder
18:13:13 Join linuxstb [0] (i=5343d4aa@rockbox/developer/linuxstb)
18:13:43linuxstbamiconn:: Your satellite digital radio (DVB) is generally 320kbps.
18:13:57amiconnAh ok
18:14:15*amiconn never dealt with dvb for audio
18:14:43linuxstbUK radio stations generally broadcast at 128kbps via DAB, and 160kbps-192kbps via satellite (DVB), with some good and bad exceptions.
18:15:11amiconnI don't like the idea of radio functions tacked on to a medium that's meant for tv
18:15:26amiconn...and dab is made for mobile reception, unlike dvb
18:16:11linuxstbTrue, but DVB is good for in-home radio - you can get a far larger choice of stations (from all over Europe) via satellite.
18:16:22 Join Rondom [0] (
18:17:06linuxstbAnd (in theory at least), satellite has more capacity, allowing higher bitrates.
18:17:12amiconnThat's only true with satellite reception
18:17:23amiconn...and web radio works worldwide
18:18:24linuxstbIt does, but generally with low bitrates, at least at the moment.
18:20:41 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:20:46amiconnThe first 29 stations on are 320kbps mp3
18:23:31 Join Ribs [0] (n=ribs@
18:24:39linuxstbOK, but the stations I listen to aren't - DVB satellite is the highest quality digital source.
18:26:55toffe82Domonoky: I am trying now to download the themes
18:29:05Juice^anyone know if the wakeup alarm will ever work on the sansa? (when the sound is working..)
18:30:03toffe82Domonoky: seems to be working, I try one theme, don't have time to try mor ;)
18:31:31linuxstbJuice^: Does the Sansa have a clock?
18:31:40Juice^linuxstb: yep
18:32:21 Join mattzz_ [0] (
18:32:29Juice^linuxstb: real time clock yes, and its working atm
18:32:51linuxstbIt works in Rockbox?
18:35:21linuxstbAh, the RTC seems to be part of the AS3514 - so I guess we'll know when/if AMS give us datasheets.
18:35:44Juice^oh ok.. cool
18:35:54linuxstbAccording to this page, there's an alarm -
18:36:00linuxstbSo there's hope.
18:36:39Juice^that link didnt work for me
18:37:06linuxstbHmm, try googling for as3514
18:38:15Juice^are you able to open this link?
18:38:25Juice^maby the ams servers are down
18:38:43 Join barrywardell [0] (
18:39:24 Join theli_ua [0] (
18:39:29amiconnJuice^: Link working fine here
18:39:48Juice^amiconn: ok - probably a routing error then, i'll try a proxy :)
18:39:59Nibbierworking here, too
18:40:06Nibbieri'd mail it to you if you wish
18:40:44theli_uaam i wrong or there are now bootloader builds with cop support for mini2g?
18:41:36linuxstbNot released yet though - they need testing on the minis.
18:41:47linuxstbAre you in Windows, Linux or Mac OS X?
18:41:53Juice^oh thanks Nibbier, but i got it working with a proxy here
18:41:54amiconnI've built a mini2g bootloader from svn yesterday
18:41:59amiconnworks fine.
18:42:04linuxstbWith the Apple firmware?
18:42:09amiconn(starting of is untested)
18:42:18linuxstbThat's the main thing that needs testing...
18:42:20theli_uaamiconn, can you put it for download somewhere?
18:42:21amiconnI have no retailos in my firmware partition
18:42:27amiconnIt's banished
18:42:50linuxstbtheli_ua: The bootloaders I want to release are here -
18:43:38amiconnlinuxstb: Btw, when I first tried cop-enabled rockbox on my mini, it just hung. For some reason I thought my bootloader was new enough, but it wasn't
18:44:19 Quit Rick (Read error: 104 (Connection reset by peer))
18:44:30 Join Rick [0] (
18:44:31*amiconn wonders what use retailos is on a rockboxed ipod
18:44:47theli_ualinuxstb, strange .. now my mini is stuck at apple logo .... before updating bl it was stuck at rockbox logo :)
18:44:49amiconnThe only ipod where it makes some sense to me is the video
18:45:17amiconn(because of ... video)
18:45:39amiconntheli_ua: g1 or g2?
18:46:06linuxstbHow did you install it? With "ipodpatcher -a" ?
18:46:20theli_uawith dd :D
18:46:30theli_uaipod_fw + dd ...
18:46:39linuxstbThat's why then...
18:46:53theli_uai see .. rules have changed ;(
18:46:58*amiconn installs with ipodpatcher -wf :P
18:47:02theli_uai need to use ipodpatcher?
18:47:21linuxstbBest to restore a clean firmware partition first.
18:47:44linuxstbAre you on linux?
18:48:05theli_uaheh ... rockbox changed much since i closely followed it :( .... i wonder if my commit account is still valid :)
18:48:31theli_ualinuxstb, yeah .. on linux
18:48:59 Join ender` [0] (
18:49:01linuxstbThen just run this (no options) - (after restoring your firmware partition)
18:49:57 Quit mattzz (Read error: 110 (Connection timed out))
18:50:03 Join ackbahr [0] (
18:50:04amiconnlinuxstb: Are these bootloaders any different from svn?
18:50:20*amiconn wonders whether he should test booting the of
18:50:25linuxstbBut because of the odd issues, I want those specific binaries tested.
18:50:44amiconnI'm lazy and I don't really see the point, but features should work properly
18:51:36amiconnBtw, don't know whether you read the logs; cop support doesn't seem to help anything wrt performance
18:51:41linuxstbI agree with you, but the majority of ipod users seem to want to dual-boot.
18:51:53amiconnIt even feels slightly slower than without
18:52:12amiconnBoost ratio during mp3 playback doesn't change
18:52:32linuxstbI haven't really looked at that side of it.
18:52:47theli_uai wonder ... how do ipodpatcher "scans"
18:53:04amiconnThe strange thing is that scrolling through lists feels sluggish now, which it never did in single-core operation on the mini
18:53:10linuxstbIt simply goes through /dev/sda to /dev/sdz
18:53:19theli_uathat's why it doesn't sees mine
18:53:27theli_ua/dev/ipod* :D
18:54:58linuxstbThen run "ipodpatcher /dev/ipod −−install"
18:55:19linuxstb(the bootloaders are embedded inside ipodpatcher)
18:55:25theli_uai've tried specifying /dev/ipod
18:55:27theli_ua Drive is not an iPod, aborting
18:55:46theli_uaoh .. i've broke my first partition .. nice
18:58:08 Join CriamosAndy [0] (
18:58:53 Quit Arathis ("Bye, bye")
19:00:11theli_ualinuxstb, "[ERR] Drive is not an iPod, aborting" why? how it "detects" of this is an ipod?
19:00:49 Nick Stalwart_ is now known as Stalwart (
19:01:09 Quit _Veseliq_ (Remote closed the connection)
19:01:33linuxstbIt's a combination of magic in the MBR and magic in the firmware partition - it errs on the side of caution.
19:01:56barrywardellamiconn: there were reports of that in the past with the cop patch, but I thought they had been resolved
19:02:21barrywardellone version I tested quite a while ago felt like that
19:02:35 Quit pondlife ("disconnected has pondlife")
19:03:01barrywardellI think people were suggesting "lockstepping" as an explanation
19:03:13 Quit JoeBorn (Read error: 104 (Connection reset by peer))
19:04:51theli_uawhat "type" should boot partition of ipod have?
19:05:46linuxstbCan you just restore a standard MBR?
19:05:57linuxstbThe firmware partition should be type 0.
19:06:10theli_uai only have a whole dump of ipod
19:06:19 Quit ackbahr (Remote closed the connection)
19:06:21theli_ualinuxstb, thanks .. that's what i've set :(
19:06:27 Quit GodEater_ ("CGI:IRC (Ping timeout)")
19:07:53theli_ualinuxstb, i'll try to dd only needed parts from whole disk dump
19:08:13***Saving seen data "./dancer.seen"
19:09:12 Quit mattzz_ ("Leaving")
19:11:21theli_ualinuxstb, i've recovered boot partition, retail OS works . ipodpatcher still can't recognize device as an iPod
19:11:48 Join ackbahr [0] (
19:14:35linuxstbCan you restore a standard MBR?
19:14:45theli_uait's standard
19:14:48 Quit Rick (Connection reset by peer)
19:14:49theli_uai think
19:14:55 Join OgMaciel [0] (n=omaciel@ubuntu/member/gnukemist)
19:15:00 Join Rick [0] (
19:15:02linuxstbI mean one from
19:15:09theli_uaok ... i'll try to dd one sector from disk dump (didn't want to unpack it :( )
19:15:21theli_uaoh . tah one
19:15:59 Quit perplexity ("* out *")
19:16:19theli_uai've restored that one .... [ERR] Drive is not an iPod, aborting
19:16:45 Quit Criamos (Read error: 110 (Connection timed out))
19:17:56linuxstbSo you restored the original MBR and original firmware partition?
19:18:45linuxstbThen all I can suggest is compiling ipodpatcher yourself from SVN (in the rbutil/ipodpatcher) directory, and seeing where it failes.
19:19:43linuxstb(adding printfs...)
19:19:52theli_uai wonder how can i test whether i have an account (i had in cvs) in svn?
19:20:13theli_uawithout commiting , of course
19:20:22linuxstbIf you didn't email Bagder with an SVN password, you don't have an account.
19:20:59theli_uaso account's didn't mograte from cvs to svn? :-/
19:21:22theli_uanice :)
19:21:22linuxstbNot automatically, no.
19:22:16*GodEater just received an enormous parcel from amazon. Sadly it was about 16 times the size of its contents
19:26:21 Quit pilled ("changing servers")
19:26:52 Quit Ribs ("eh eh ehhhh!")
19:27:13 Join Ribs [0] (n=ribs@
19:27:27 Join pilled [0] (
19:29:11 Join clay [0] (
19:30:15clayit's terrible, rockbox dosn't work on archos 104
19:31:27 Nick clay is now known as pH_7 (
19:31:44 Nick pH_7 is now known as clay (
19:31:57 Quit Lear ("Chatzilla 0.9.77 [Firefox]")
19:32:51 Quit ackbahr (Remote closed the connection)
19:34:46GodEaterthe english language is dying the death of a thousand cuts in our forums
19:36:11claywhat ?
19:37:17GodEaterlinuxstb: congratulations on not fixing the bootloader bug :)
19:38:06 Join Nico_P [0] (
19:38:33GodEaterit looks like the new ipodpatcher with prompts and built in bootloaders is popular too
19:40:20linuxstbThen I wonder what people will make of rbutil when it's released...
19:40:20 Quit Rondom (Read error: 110 (Connection timed out))
19:40:51GodEateryeah - it's kind of shame that's been polished up at the same time
19:40:57GodEaterDomonoky has stolen your thunder :)
19:41:24GodEatermind you, it means Domonoky gets to answer all the install support questions now :)
19:41:27 Join web-taz [0] (
19:41:39 Join Rondom [0] (
19:41:47theli_ualinuxstb, well.. ipodpatcher obviously fails at
19:41:48theli_uaif ((memcmp(&sectorbuf[71],"iPod",4) != 0) && (memcmp(&sectorbuf[0x40],"This is your Apple iPod. You probably do not want to boot from it!",66) != 0) )
19:41:49 Quit ender` (" In politics, absurdity is not a handicap. -- Napoleon Bonaparte")
19:42:04GodEatertalking of which - Domonoky, I'd have started a whole new thread for it if I were you
19:42:27theli_ualinuxstb, but removing the check i've installed bl successfully
19:42:46claywhen rockbox on archos 104 ? :D
19:42:50GodEatersounds like your bootpartition was a bit broken
19:42:55linuxstbtheli_ua: Then it didn't like the MBR...
19:43:02Domonokyi will make a new Thread for rbutil when its got a bit of testing..
19:43:05GodEaterclay: when someone decides to start working on it
19:43:23linuxstbDoes the MBR you installed have either of those strings at those offsets?
19:43:30theli_ualinuxstb, i understand . but i've restored both mbr and boot part.. and i don't have things it looks for
19:44:00GodEaterit's not an HFS+ iPod is it ?
19:44:47theli_uaGodEater, i think it was a long time ago .... i do not actually remember
19:44:59 Join obo [0] (
19:45:25clayso who want to work on it ?
19:45:40linuxstbtheli_ua: Which MBR did you download?
19:45:53GodEaterclay: usually people who own one
19:46:09clayyou GodEater, nasty man
19:46:20preglowmarkun: the crossfeed filter calculation i handed you last time was completely wrong...
19:46:59linuxstbtheli_ua: Seems that MBR isn't an original...
19:46:59GodEaternasty? I'm just telling you when and if it'll happen ?
19:50:48 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:51:27x1jmpWhy do the menu functions not reset the lcd margin? Sounds like a bug to me...
19:51:47 Join inversions [0] (
19:52:22x1jmpI'm now having problems with the remote, lcd_remote_setmargins is currently missing in the plugin API
19:54:41Learx1jmp: wasn't that fixed the other day? At least something related to that was, I think...
19:55:45clayGodEater: are you a god ?
19:56:03zeclay: you are what you eat :p
19:56:12linuxstbx1jmp: Functions are generally only added to the plugin API when needed - so feel free to add anything you need.
19:56:41 Join lukaswayne9 [0] (n=lukas@
19:57:24 Join spiorf [0] (
19:57:44x1jmplinuxstb: IMHO it's better to fix the menu instead of extending the plugin API, and no, it's not fixed, at least in my simulator build I compiled from SVN
19:58:28spiorfhi, i would like to install the bootloader on an ipod nano 1st gen with 1.3 firmware
19:58:35spiorfit is possible or i have to downgrade?
19:59:40linuxstbx1jmp: I agree, but you just said the plugin API was your problem...
20:00:05linuxstbspiorf: Rockbox and the Rockbox install tools work with any version of Apple's firmware.
20:00:49Learx1jmp: is it a problem for one or both margins?
20:01:43x1jmplinuxstb: the plugin API is used as a workaround in the plugins now
20:02:11x1jmpLear: I just discovered it on the top margin
20:02:42LearHm... Menus doesn't seem to change y-margin, and x-margin should be properly restored...
20:03:41dan_alinuxstb: Your bootloader works for me
20:04:11 Quit Ribs ("eh eh ehhhh!")
20:04:12linuxstbdan_a: Good news, but I almost wish it didn't, so we could try and debug it...
20:04:39LearSo I'd guess you're suffering from the statusbar margin, which isn't changed by the menus as such.
20:05:17 Quit lukaswayne9 ("Ex-Chat")
20:06:04 Join vinca [0] (
20:06:05 Quit theli_ua ("quit")
20:06:08x1jmpLear: you're right, looks as it's the statusbar
20:06:15dan_alinuxstb: The old ones are still in SVN...
20:06:24 Quit jhMikeS (Nick collision from services.)
20:06:30 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
20:06:40linuxstbAnd when I said "we", I meant someone with a 4g or mini... ;)
20:07:07*preglow hands the tool of the year award to linux for ipodpatchet
20:07:17preglowalso linuxstb
20:07:53linuxstbThanks, I'm happy with it. Imagine making ipod users upgrade their bootloaders without it...
20:07:53preglowi remember i usually couldn't be bothered with upgrading bootloaders thanks to how cumbersome it used to be...
20:08:01preglowyeah, exactly
20:08:19bonbonthejonlinuxstb:what do you need with someone with a 4g?
20:08:42linuxstbTo work out the reason why the last version of the bootloader didn't load the Apple firmware.
20:09:08bonbonthejonoh, I dont use the apple firmware anyway
20:09:57x1jmpLear: I just disabled the statusbar for the menu and margin is gone, but IMHO it's still a bug in the menu function
20:10:05preglowlinuxstb: debugging it really would be clever anyway
20:10:21preglowlinuxstb: having to worry about maybe breaking retailos support everytime we alter the bootloader isn't very optimal
20:11:17LearStatusbar is generally handled by the screen in question. In some it is shown, in some not...
20:11:40 Join ender` [0] (
20:12:11linuxstbpreglow: By debugging, I mean working out why Retailos wasn't working, so we don't have to worry in the future.
20:12:19preglowyeah, what i mean too
20:13:18linuxstbSorry, I misread your sentence as "wouldn't be clever"...
20:13:35*linuxstb goes to reading classes
20:13:42 Part vinca
20:14:07dan_aDo we know what the possibilities are? A scratch register not being what OF expects, hardware being set up incorrectly, an exception...
20:14:39linuxstbI would rule out hardware setup, as that should be identical in all versions.
20:14:49dan_aGCC being mad
20:15:14linuxstbMy best guess is the OF expecting values in registers and/or iram.
20:15:21dan_aI don't understand why it should work for the colour targets but not for greyscales
20:15:31linuxstbWhich could also explain Rolo not working...
20:16:00linuxstbMaybe the greyscale targets are using less IRAM, so a vital part isn't overwritten...
20:16:23dan_aBut then that would mean that the colour ipods should have the problem
20:16:33 Join vinca [0] (
20:17:46linuxstbI started looking at the dissassembly of retailos, and on both the Color and Video, R0, R2 and R3 are being saved on startup (in IP, R8 and R9 respectively).
20:17:57 Quit Febs ("Chatzilla 0.9.77 [Firefox]")
20:17:59vincahi boys, could you tell me what to do when ipodpacher is telling me that my ipod is not an ipod? its the newest 80gb ipod and i'm running mac osx..
20:18:53linuxstbOn the COP on both targets, R0 has a startup value of 0x40000050. Other values differ between Color and Video.
20:19:23linuxstbvinca: Rockbox doesn't work on the 80GB ipod (but ipodpatcher should...)
20:19:27dan_alinuxstb: It's probably worth saving those, then. Although lots of compilers will save those anyway
20:19:44spiorflinuxstb, IpodInstallationFromLinux misses the part about ipod nano >= 1.2 users. i have to do the extra things (like IpodInstallation page says) or not?
20:19:57preglowlinuxstb: then the contents are probably important
20:20:05preglowlinuxstb: it would never save the scratch regs if not
20:20:11dan_avinca: Are you pointing it to the device or the partition (/dev/sda or /dev/sdaX)?
20:20:44preglowthat is, "never" is a bit too strong...
20:20:47linuxstbspirof: Read the notes at the top of that page.
20:20:59pixelmavinca: the 80 GB isn't supported yet
20:21:39amiconnlinuxstb: Didn't check the ipod .map files, but the m5 bootloader uses *more* iram than the x5 bootloader
20:22:08 Quit clay (Read error: 110 (Connection timed out))
20:22:21vincathanks to all, i wasted a bit of my time and i don't want to waste yours.. i hope i'll be back when time comes..
20:22:31 Part vinca
20:22:43linuxstbamiconn: Yes, thinking about it, I retract that guess...
20:23:30linuxstbBTW, the register saving is done for both CPUs...
20:25:35linuxstbamiconn: I retract that retraction - I meant to say the greyscales are using more IRAM. Don't their framebuffers go in IRAM?
20:26:09linuxstbdan_a: If you re-enable the lcd_update() in the bootloader, is it broken again?
20:26:40linuxstbBut the writing is happening still...
20:27:27dan_alinuxstb: I'll test
20:27:39linuxstbMaybe the stack...
20:29:35preglowit should ignore the stack quite thorougly
20:30:07linuxstbI mean maybe the bootloader's stack was overwriting something important in iram.
20:30:31preglowah, it's iram, of course
20:30:32dan_aThat just worked fine
20:30:50preglowr0 sounds like it's used to point to some data from the bootloader
20:30:54preglowi wonder what it stuffs there
20:31:07preglowjtag, oh jtag
20:32:30jhMikeSoy, found the bug that makes recording crawl :)
20:32:37linuxstbThis is what I found so far -
20:33:15jhMikeS..and I feel stupid for not immediately knowing the's so obvious
20:33:28preglowif i had a dime for everytime i felt like that...
20:33:52preglowlinuxstb: cpu and cop get different values? fancy
20:33:54x1jmpSTATUSBAR_HEIGHT is defined twice :)
20:33:55jhMikeSit's in pcmbuf_under_watermark where it boosts the codec thread priority
20:34:03x1jmpin statusbar.h and icons.h
20:35:00jhMikeSit never unboosts because the pcm buffer is never full enough to untrigger the codec thread boost...hence the fact the encoders were running very fast even though everything else was inital excess thread threory was right though.
20:35:42 Quit Rondom ("Ex-Chat")
20:35:56preglowlinuxstb: 0x60007000 is the address to sleep the cpu, weird
20:37:21jhMikeSso, now what to do about it ...
20:37:48*dan_a pops out for 30 minutes
20:42:52jhMikeSdid the typical voice clip size shrink (better trimming)?
20:43:08jhMikeSor length rather
20:44:43 Quit atsea- (Remote closed the connection)
20:45:17 Quit Rick (Read error: 113 (No route to host))
20:45:42jhMikeSguess I get the thread_get_current function I always wanted afterall :)
20:46:06 Join Rick [0] (
20:47:21 Join atsea- [0] (i=ariel@gateway/tor/x-39023d6da28e6b14)
20:48:40 Join BigMac [0] (
20:52:04BigMacAnyone have a working build/patch of the scrolling margin with the current svn
20:53:58 Quit barrywardell ()
20:54:02jhMikeSpreglow: libmad in our usage never actually tries to use malloc and calloc right? Checking it seems to suggest that it doesn't - never saw it do so myself.
20:54:19 Quit Rick (Connection reset by peer)
20:54:22 Join rotator [0] (n=e@rockbox/developer/rotator)
20:54:27 Join Rick [0] (
20:57:07preglowjhMikeS: correct
20:57:14preglowjhMikeS: the one spot that uses it is fixed in mpa.c
20:58:59jhMikeSgood, cause the malloc buffer is not part of the swapping
20:59:43 Join Febs [0] (n=chatzill@
21:08:14***Saving seen data "./dancer.seen"
21:15:23jhMikeSIt's hard to see the spc codec boosting the codec thread now :)
21:15:40preglowhow's it perform on ipods?
21:16:03jhMikeSCan't play everything with echo
21:16:18 Quit netmasta10bt (Read error: 110 (Connection timed out))
21:16:30jhMikeSAfter some more assembly updates it'll run on coldfire almost where it was before adding echo
21:16:53LearAnyone buildt festival? Seems to wait for input, but I don't know what...
21:17:08Lear(Under Cygwin...)
21:17:32 Quit muesli__ (Read error: 104 (Connection reset by peer))
21:18:16 Join Domonoky_ [0] (
21:18:40 Join Gibbed [0] (
21:18:44 Quit Rick (Nick collision from services.)
21:18:58 Nick Gibbed is now known as Rick (
21:19:48linuxstbIs anyone around who has access to
21:20:21Domonoky_it works for me
21:20:30preglowhere too
21:20:33linuxstbI mean write access...
21:21:10jhMikeShmmm...voice seems to get reset to "no" after rolo :\
21:23:09x1jmpLear: Can you tell me where the statusbar sets the margins? I can't find anything related to the margins and statusbar
21:24:00preglowwhat's the point in the value parameter to the set_sound callbacks? it seems the global_settings struct memebers are updated anyway
21:24:05preglowis it just a binary size measure?
21:24:07Learx1jmp: no specific call, but typically something like this: screens[i].setmargins() with proper values.
21:24:09amiconnjhMikeS: RoLo doesn't do a clean shutdown. You need to make sure settings are saved
21:24:52amiconnI usually do this by browsing a dir
21:24:53 Join |Rincewind| [0] (
21:25:06amiconn(of course I have dircache disabled)
21:25:13jhMikeSamiconn: but it was shutdown clean several times before with voice "yes"
21:25:15LearHuh? Made a voice makefile, it builds a rockbox binary...
21:25:22jhMikeSI have no dircache either
21:25:31linuxstbLear: You have to type "make voice"
21:25:32amiconnAh, that's indeed strange
21:25:41 Join decayedcell__ [0] (n=decayed_@
21:26:11LearAh, so it is a normal makefile, but with a voice target? Unlike the manual makefile?
21:26:49 Quit Domonoky (Read error: 60 (Operation timed out))
21:29:30 Quit miepchen^schlaf (Read error: 60 (Operation timed out))
21:29:45*amiconn doesn't lie how these extra targets mix up 'configure' and make targets
21:29:46 Join miepchen^schlaf [0] (
21:29:52x1jmpLear: that's what I tried, but I have no success
21:30:28Learx1jmp: how so? But see e.g. bookmark.c for an example.
21:30:38FebsHmm, why is Rockbox scanning my disk on every single startup?
21:30:40 Quit decayedcell__ (Remote closed the connection)
21:31:44x1jmpLear: yes, but none of it seems to relate to statusbar and menu
21:32:30Learx1jmp: calls to setmargin with STATUSBAR_HEIGHT as possible argument, like line 499.
21:32:49amiconnx1jmp: What's your actual problem?
21:33:50 Join lee-qid [0] (
21:33:54 Join Llorean [0] (
21:34:10x1jmpamiconn: when calling the menu with statusbar enabled in a plugin, the y-margin isn't reset
21:34:46LearGah, lame configure script to bail on paths with spaces in them... :p
21:34:56 Join decayedcell__ [0] (n=decayed_@
21:35:18 Join robin0800 [0] (
21:36:02 Join muesli__ [0] (n=muesli_t@
21:36:08amiconnx1jmp: Afaik it's the responsibility of the caller to reset the margins atm
21:36:38amiconnI think it's the list code where the margins are set
21:36:39LearOk, so what do I type when the festival make says "Scripts: (sh) (prl) festival_server"?
21:37:07Learlist.c sets (and re-sets) the x-margin, doesn't change the y-margin.
21:37:54x1jmpamiconn: When calling a menu, I think it's bad to have to reset the y-margin afterwards
21:38:51x1jmpLear: I think I've just located it in gui_textarea_clear, now trying a fix
21:40:10robin0800with the album art patch and the new bmp resize patch both updated is it possiable to put them together and then commit them?
21:40:57 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
21:41:00Learx1jmp: Ah, yes, that part does set the y margin. Should be reset by the list code, I think.
21:41:14LearAnd look who just choose this moment to join us... :)
21:41:21preglowJdGordon: you've changed alarm wakeup behaviour with the root menu commit5
21:41:26 Part decayedcell__
21:41:33dan_arobin0800: The general view at the moment is that the album art patch is not the right way to do things - we want more general code which can be used throughout Rockbox
21:42:47robin0800dan_a: could you elaborate a little?
21:43:06preglowJdGordon: also, the start screen setting doesn't seem to stick here
21:43:15 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
21:43:44linuxstbrobin0800: It's always been said that album art won't be committed until it's storing the images in the audio buffer - which is much more efficient than the existing implementation.
21:44:58 Join entheh [0] (
21:45:21dan_arobin0800: We also want to be able to split the screen up into "viewports" (or windows) - so the status bar could be one viewport, the WPS could be a viewport with a viewport inside it to hold the album art, etc.
21:46:32robin0800dan_a: is viewport being worked on at present?
21:47:11preglowit's another one of those concepts that we'd really like, but never seems to materialise
21:47:19preglowjust like metadata on buffer, heh
21:48:05robin0800well then why not commit the patches?
21:49:08preglowwell, because if we do that, there's an even smaller chance we'll ever get it done the proper way
21:49:14 Join hannesd_ [0] (
21:49:45Lloreanrobin0800: Album Art is hardly critical functionality. It can wait until someone implements the prerequisites.
21:50:33dan_arobin0800: People are starting to think about that, but taking it out after putting it in will be more work than doing it correctly before committing it
21:51:13Lloreanrobin0800: Besides, metadata-on-buffer is still essential for album art working the 'right' way. The current patch is considered 'wrong' in that aspect.
21:51:28JdGordonpreglow: which commit? last nights? or the origional one?
21:51:33robin0800perhaps not but it is a defult function in the OF
21:51:44preglowJdGordon: the original one
21:51:53Lloreanrobin0800: Rockbox isn't the original firmware.
21:52:07preglowJdGordon: just check the check_alarm_started() call in the original tree.c versus the remains of it you have included in root_menu.c
21:52:09Lloreanrobin0800: "The original firmware did it" is not a reason to do things the wrong way.
21:52:20FebsPlus, that's not true for all Rockbox platforms.
21:53:43LloreanJdGordon: and may interest you.
21:53:50preglowJdGordon: previously, wake on alarm meant rockbox would always resume playing audio. we might want to extend that to do the default action in the start screen, but until that's in place, i'd like it if i can still use my ipod as a wakeup alarm without setting auto-resume on :)
21:54:21robin0800I know but some people may like the wps to look like the original just look at some of the themes
21:54:25JdGordonget alarm working on h300 and ill make alaram work perfectly :p
21:54:37preglowi believe linus is on that
21:54:59amiconnalarm should also be possible on X5 and M5
21:55:02 Quit SirFunk (Remote closed the connection)
21:55:09Lloreanrobin0800: Then those people should get working on implementing metadata on buffer, so that it can be done the right way.
21:55:14amiconnSame chip, after all
21:55:41Lloreanrobin0800: Again, Album Art is not a critical feature. There is no reason to do it the wrong way as a 'holdover' until it can be implemented the right way.
21:56:56JdGordonbut cuesheet was?
21:57:34robin0800ok I can wait i'll continue to patch it I just hope you don't loose people who can't patch
21:58:00Lloreanrobin0800: If the single most important feature to people is Album Art, I'm surprised they're using Rockbox in the first place.
21:58:09preglowrobin0800: if they can't stand using rockbox because of lacking album art, then we wouldn't have kept them for long anyway
21:58:10JdGordonanyone now like the idea of showing <No Files> in an empty folder in the browser? I know its requested but not sure why it was never implemented
21:58:21preglowJdGordon: sounds like a good idea to me
21:58:37dan_arobin0800: People make unsupported builds for people who can't patch. Album Art is something that we want, and it will come when it's done right
21:58:47dan_aJdGordon: Thumbs up from me.
21:59:43LloreanWhy not something like ".."? That way browsing can be done without a 'left' button. I know it seems silly right now, but it would allow for the possibility that some targets may have very, very few buttons.
21:59:58 Join Rondom [0] (
22:00:09JdGordonput .. in every folder?
22:00:17LloreanOr at least few usable ones (say, controlling a device with VoiceUI using a very limited remote)
22:00:34LloreanThat's my personal suggestion, I don't know how popular it'd be. ;)
22:01:23linuxstbAre people familiar with ".." away from the command-line?
22:01:36JdGordoni wouldnt think so
22:01:45pixelmaI think I'd confuse a "no files" entry as "oh there is a file" at first glance
22:01:45preglowme neither
22:01:58Llorean<Back> then?
22:02:09preglowpixelma: sounds like a very quick glance
22:02:21JdGordonok, well Im going to have breaky so im not late to uni, you guys aregue and ill read the logs for the decision
22:02:23preglowLlorean: much better
22:02:24hcsmaybe the angled arrow like on a return key?
22:02:29linuxstbMaybe <No Files> shouldn't be highlighted.
22:02:30LloreanIt just seems to me that a navigational entry would both offer additional functionality, a little bit of future proofing, and resolve the empty folder issue all at once.
22:02:36preglowJdGordon: fixed alarm?
22:03:02JdGordonuni in 45min... i have a 2 hour break at 11.. ill look then
22:03:06preglowlinuxstb: well, it's not a file...
22:03:08pixelmawell I see a "<MMC1>" entry quite often...
22:03:24robin0800yes the return key and to always take you back
22:03:30amiconnEmpty folder issue?
22:04:28amiconnAn empty folder must be empty, imho
22:04:44amiconnAn extra entry to make it non-empty sounds like a silly thing to me
22:05:12linuxstbAt least with the current directory name as a title, an empty dir isn't a blank screen any more.
22:05:19linuxstb(if that's enabled)
22:06:11amiconnAn empty dir never was a blank screen for me
22:06:35amiconn...because I always have the status bar enabled
22:06:54amiconnBut even without that, empty is empty and should be shown as empty
22:06:56robin0800how do you end up with empty directories?
22:07:29linuxstbEasy if you've switched file view to "supported" and have other files on your disk.
22:07:38hcsrobin0800: as yogi berra might say: a directory is empty until you put something in it
22:07:42robin0800I don't have any
22:07:54LearInteresting. If I type "make install", and there is no install target in the makefile, but there is an file in the directory, make will copy it to "install" and mark it as executable.
22:08:20amiconnThe . and .. dir entries are a silly workaround to save early ms-dos from remembering the current dir.
22:08:43amiconnOn fat they're actual entries...
22:09:03preglow. and .. is older than dos
22:09:06robin0800yes use the retun symbol
22:09:40amiconnLear: That's part of make's sophisticated guessing what you want to do
22:09:46preglowi think they're actual entries in more fses than fat
22:09:50linuxstbBut anyway, I think I agree with amiconn that empty directories should stay empty.
22:09:58robin0800and always up one level when used
22:10:10preglowlinuxstb: would "no files" make it any less empty?
22:10:14preglowit's just a hint
22:10:28preglowwe could even center it on the screen or something to help differentiate it
22:10:38amiconnpreglow: Afaik the only file system which actually has . and .. entries is (v)fat
22:10:39linuxstbThat could be nice.
22:10:47pixelmaisn't that quite obvious that if you enter a directory and then the screen is blank that you've entered an empty directory?
22:10:57preglowpixelma: not according to all the people who've asked us
22:11:08linuxstbPeople have asked?
22:11:28amiconnDirectory traversing using . and .. may be older than that, though
22:11:33preglowamiconn: it is
22:11:37dan_aIn database view, files without tags show up as "<untagged>", not ""
22:11:58amiconndan_a: That's very different
22:12:02Lloreandan_a: But those are still real files.
22:12:03*pixelma sighs and gives up UI discussions :/
22:12:18LloreanThey're there, just not described, as opposed to being something that isn't there
22:12:43 Quit BHSPitMonkey (Remote closed the connection)
22:12:48 Join petur [0] (n=petur@rockbox/developer/petur)
22:14:05amiconnAn empty folder just shows like that in windows explorer, nautilus, amiga workbench, ... : empty
22:14:20 Quit BigMac ("Leaving")
22:14:29amiconnWhy should that be different on an mp3 player?
22:14:36preglowthose have further hints, though
22:14:41preglowexplorer has a previous dir entry
22:14:49Lloreanamiconn: In those you have a frame around the file browser still, it isn't a completely blank screen.
22:14:58LloreanIn Rockbox you have the possibility of a 100% blank screen.
22:15:24LloreanWhile that is the user's responsibility, it still seems to confuse them.
22:15:26preglowand the fact that people have been confused by that before, and several times, shows that you can't compare
22:15:51preglowpeople never get confused by an empty dir in windows, but they do in rockbox, hence further hints showing it's empty is obviously needed
22:16:06linuxstbBut people didn't even realise they were in a file browser in the past...
22:16:13linuxstbNow they should.
22:16:25dan_aamiconn: What problems would displaying something to show that they are in an empty directory bring?
22:16:25pixelmaI think the "show path" is enough - then they'll see that they are in a dir...
22:16:50linuxstbIs "show path" enabled by default?
22:16:57preglowat least i've never seen it
22:17:16linuxstbI think as long as the defaults make it clear, then that's enough. If the user disables things, it's their fault.
22:17:32amiconndan_a: It would clutter the screen, and can be mistaken for an actual entry
22:17:39preglowamiconn: clutter an empty screen???
22:17:40amiconnYOu have to look more close
22:17:51dan_alinuxstb: When I choose "files", it takes a while for the disk to spin up - I often press to go into it twice and end up in the "Calendars" folder of the iPod. The first couple of times it happened I got confused
22:18:06preglowamiconn: would "no files" listed in the middle of the screen have any chance of you mistaking it for a file?
22:18:21linuxstbHow about simply enabling "show path" by default?
22:18:42pixelmaI could agree on that
22:18:43LloreanI'm still worried that themers may disable 'show path'
22:18:50amiconnlinuxstb: Could be done, except for the player where it should be off by default
22:18:56LloreanBut I think 'show path' is the best option without adding text somewhere
22:19:04 Join JoeBorn [0] (
22:19:14amiconnPlayer users should know what they're doing ;)
22:19:22linuxstbamiconn: Sure. I would expect player users are clueful enough..
22:19:33preglowi guess show path by default is nice
22:19:45preglowbut i still think "no files" would be a good thing to have
22:19:51preglowno chance of any mistakes
22:20:02amiconnBtw, the player of just shows a single file at at time in its browser because it uses the other line for (pretty useless) info...
22:20:02preglowcan't be more verbose, and it should annoy absolutely no one
22:21:55linuxstbpreglow: A "no files" message wouldn't upset me, but I'm still not convinced it's needed now we have a separate menu item for the file browser, and the current directory shown.
22:22:32amiconnIirc the sub-browsers do show a 'no files' message and exit afterwards
22:22:36 Join webguest64 [0] (i=5ba87612@gateway/web/cgi-irc/
22:22:43 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
22:23:14linuxstbLinusN: Can I trouble you to update the ipod bootloaders and ipodpatcher on
22:23:21webguest64Hello, I am french, sorry for my bad english
22:23:45webguest64De you speak french?
22:23:59preglowlinuxstb: needed? no, but if we even already have the string in the lang files...
22:24:09amiconnLinusN: Domonoky_ also wants the m5 bootloader on to make it usable for rbutil
22:24:13 Quit Lear ("Chatzilla 0.9.77 [Firefox]")
22:24:44markunpreglow: do you have a new crossfeed patch?
22:24:45linuxstbpreglow: You're obviously more strongly in favour than I am against, so you win.
22:24:50LinusNamiconn: ok, will fix
22:25:06preglowlinuxstb: good, too bad i have no intention of implementing it :P
22:25:23preglowmarkun: yes, i do!
22:25:23preglowmarkun: keen to test?
22:25:33preglowmarkun: this one should sound exactly the same too
22:25:34LloreanLinusN: 250k is larger than the fonts he's testing it with, so it really can't *not* work fine.
22:26:08JdGordon250K is too much
22:26:12LloreanLinusN: The fonts he was experiencing problems with were about 208k, so at 250 it just loads the whole thing into memory up front.
22:26:20JdGordon150K should be max going by the standard fonts
22:26:27*JdGordon gone
22:26:36 Quit JdGordon ("Konversation terminated!")
22:26:52markunpreglow: hm, I could compile it now and then test later
22:27:02 Quit spiorf (Read error: 110 (Connection timed out))
22:27:30markunbut if you say it's the same on ARM targets I can just wait for the commit
22:27:43webguest64Why rockbox does not compatible the NW-HD5 Sony?
22:27:53preglowmarkun: really just want to make sure, the coefs aren't exactly the same, but the results look the same when i graph them
22:28:11Lloreanwebguest64: Same reason it's not compatible with so many other players. Someone who owns one has to port it.
22:28:19 Join Ribs [0] (n=ribs@
22:29:50LinusNlinuxstb: where are the originals?
22:29:56 Join Id2ndR [0] (
22:30:07webguest64A lot of utlisateurs of the NW-HD5 sony likes this mp3 player but sonicstage is bad. And the full ums is many good
22:30:26Id2ndRHi everybody, I've just seen that Rockbox now works on iaudio M5L
22:30:36Id2ndRThanks for this !! :)
22:30:41linuxstbLinusN: The bootloaders themselves are here - - can you put the zip file itself in /bootloader/ipod/ and also unzip the contents in the same place.
22:30:59Id2ndRMaybe you can help me : I'm looking for the bootloader I have to use
22:31:13Lloreanwebguest64: Ports do not happen by request. People who own the device have to do it for themselves.
22:31:34Id2ndRI don't know f I can use one of these available at
22:32:06LinusNId2ndR: yes you can use the m5 loader i just put there :-)
22:32:42linuxstbLinusN The Windows version of ipodpatcher is - can you unzip it and put the .exe in /bootloader/ipod/ipodpatcher/win32/
22:33:00linuxstbThe Linux 32-bit x86 version is at a similar URL, replace win32 with linux32..
22:33:11Id2ndRLinusN, thanks, I've got an M5L, the M5 is the good one too ?
22:34:05pixelmayes - it only has a different battery
22:34:39linuxstbLinusN: And finally, the Mac OS X version is here -
22:34:48pixelmaId2ndR: the M5 that was used to port it was an M5L ;)
22:35:07Id2ndROk ! thanks pixelma :)
22:35:13 Quit webguest64 ("CGI:IRC (Ping timeout)")
22:35:25bluebrotherLinusN: I'd like to suggest bascule to get close powers on FS −− he added a couple of good comments lately and I think he could help keeping the tracker clean.
22:35:39amiconnId2ndR: In fact the M5 version was developed on an M5L
22:36:45linuxstbLlorean: I don't know if you read in the IRC logs, but the ipod bootloaders are now silent - they don't display anything unless there's an error (or you hold the RIGHT button).
22:37:15Lloreanlinuxstb: I'd missed the bit about the right button. Thanks.
22:37:27 Join webguest64 [0] (i=5ba87612@gateway/web/cgi-irc/
22:37:32Lloreanlinuxstb: And apparently they mysteriously work again for loading Apple OS
22:37:47webguest64thank you for your Llorean answer. I will like to do it by me, but I do not know large thing there. The life is unjust! lol
22:38:17pixelmabluebrother: you don't know by chance how I'd have to go on with the svg to use it for the M5 manual?
22:38:25LinusNbluebrother: ok, i'll se if i can fix that
22:39:50LinusNbluebrother: bascule is now added to the developers list in flyspray
22:40:01bluebrotherpixelma: I don't know exactly what tucoz did. I know he added the text using inkscape
22:40:16preglowmarkun: if you want to test
22:40:16linuxstbamiconn: Could you build a 64-bit ipodpatcher from the latest SVN with the new bootloaders?
22:40:16bluebrotherLinusN: thanks.
22:41:07Id2ndRamiconn, pixelma LinusN , it works for me. (Not surprising wasn't it ? ;)) Thanks all
22:41:40markunpreglow: thanks
22:42:04pixelmabluebrother: I have that part done (just copying it over from the X5.svg - though I'm not sure if it's exactly the same font). He told me that he made it an eps first and ran epstopdf (but don't know which settings etc.)
22:42:21 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
22:42:31 Quit amiconn (Nick collision from services.)
22:42:33 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:42:37mcphailDoes the simulator emulate the limitations and glitches of the target cpu?
22:42:42preglowmarkun: like i said, the filter coefs don't end up exactly the same, but the frequency response of the filters look identical
22:42:42bluebrotherI remember he had some steps, but I don't know what he did exactly.
22:42:52 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
22:42:54preglowmarkun: let me know if it doesn't sound good, i'll commit it if it does
22:42:57pixelmathat's a pity
22:42:57bluebrotherI already thought about adding the conversion as additional target to the makefile
22:43:05linuxstbmcphail: No, it's Rockbox code compiled to run on your PC's CPU, using SDL for I/O.
22:43:12 Join Thundercloud_ [0] (n=thunderc@
22:43:12 Join miepchen^schlaf [0] (
22:43:23jhMikeSpreglow: what's are you messing with? the crossfeed? You said it doesn't hit on the asm routines right?
22:43:24bluebrotherthat would create a new dependency at least to inkscape, but if it's only intended to be run manually by some dev I think that would be ok
22:43:25mcphaillinuxstb: ok. that's a pity
22:43:57preglowjhMikeS: yeah, i just got obsessed with cutting out the old crossfeed filter design routine. it's pretty similar to the sw tone control routine, so i'm replacing it with that
22:44:33linuxstbbluebrother: I think just documenting it in a README, or a small "svg2pdf" script would be enough.
22:45:14preglowjhMikeS: i'm just not much in a mood for asm right now :/
22:45:52pixelmabluebrother: the problem is that ATM I can't even post a link to it because my servers of my ISP seem to be down, so my webspace isn't available...
22:46:12 Quit linuxstb ("CGI:IRC")
22:46:12amiconnSchmei doch mal rber...
22:46:19amiconnoops :)
22:46:23 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
22:46:35preglowanother bootloader related bug
22:46:43 Join miepchen^schlaf [0] (
22:46:43 Quit Rick (Read error: 113 (No route to host))
22:46:50 Join Rick [0] (
22:46:55jhMikeShehe...the filters themselves in the asm will need updating? not sure how you can change the filters and not need to update the asm unless you mean "design" as in coefficient selection. :\
22:46:57 Join webguest33 [0] (i=5ba87612@gateway/web/cgi-irc/
22:47:19 Quit webguest33 (Client Quit)
22:49:12preglowthat's the design i mean, yes
22:50:03jhMikeSassembly is an easier task than dealing with gcc tripping over itself all the time :)
22:50:25 Join blueworm [0] (
22:50:45 Part webguest64
22:52:26 Quit ompaul ("night")
22:54:06 Quit web-taz ("( :: NoNameScript 4.02 :: )")
22:54:55bluebrotherpixelma: I think I got it again ... he opened the file with inkscape, selected everything and used "save as / eps"
22:55:26bluebrotherrunning inkscape -A doesn't catch the text. I guess this is because the text is outside of the page
22:55:38jhMikeSpreglow: I'm getting the feeling you're just waiting to hand all the arm asm off to me when my Gigabeat arrives. ;D
22:56:13preglowjhMikeS: not really, no, it's what i'm planning to do when i'm finished with the stuff my head just doesn't refuse to let go
22:56:50markunjhMikeS: you bought one? Is it from toffe's collection?
22:56:53jhMikeSme 2
22:57:01jhMikeSmarkun: yes
22:57:29 Quit Thundercloud (Read error: 110 (Connection timed out))
22:57:33markunHe told me he sold 3 already in here :)
22:57:48preglowhas he got any more? :>
22:58:28jhMikeShe probably does...I promised to help to some work for the S model too but I want backlogged stuff out of the way
22:58:43bluebrotherswitching the page to landscape seems to help. I can not get the text exported as well, but the image gets larger than necessary
22:59:30markunpreglow: he keeps buying broken ones to recombine them
22:59:56amiconnbluebrother:"> and
23:00:16pixelmabluebrother: I already did the step and adapted to the right area - the question is how to make the right pdf
23:00:37preglowlooks really nice
23:01:05*amiconn needed to tell his apache about image/svg+xml
23:01:19toffe82I am here :)
23:01:47markunpreglow: so, what kind of Gigabeat are you looking for?
23:01:51toffe82somebody wants something ;)
23:01:57preglowmarkun: not really looking for one, just wondering
23:02:23markunpreglow: hm, I just told toffe82 he had a new customer :)
23:02:45toffe82I don't have too much left in good shape, but I could probably do F10 or F20 complete with dock
23:03:22toffe82I still have a F40 but customize, half black half silver, looks nice :)
23:04:11Lloreanamiconn, pixelma: Does the M5 actually say 'Color Sound' like the X5 does?
23:04:21LloreanThat's just strange.
23:04:33jhMikeSColor Sound...hrm...why not a larger lcd?
23:04:38LloreanWhy not a color LCD?
23:04:51jhMikeSooh...that's ever creepier
23:04:52toffe82jhMikeS: yours arrive tomorrow ;)
23:05:12jhMikeStoffe82: I'm waiting impatiently...:)
23:05:23preglowjhMikeS: started learning arm asm yet? :>
23:05:24 Part perl|work
23:06:12jhMikeSpreglow: no...that won't work without a device to reality ARM datasheets contain the IS ref?
23:06:33preglowjhMikeS: nah, you want the arm arm doc
23:06:52preglowcan't remember where that is anymore
23:07:02jhMikeSpreglow: where's that? should probably get a jump
23:07:02 Join thewho [0] (i=95e13c38@gateway/web/cgi-irc/
23:07:07preglowi think altera had it, but i no longer have a link'er mind then...suppose I'll Google around
23:07:13bluebrotherpixelma: just using epstopdf seems to do the trick here.
23:07:28preglowjhMikeS: i've got it so i could web it
23:07:47pixelmaLlorean: I think it has a metaphoric meaning - like "vivid" sound or something
23:07:50thewhoIs the #ifndef in radio.c:428 needed? It causes the main menu item to be "Now playing" when the main menu is called from the radio screen IMHO
23:08:00jhMikeSok...on your site?
23:08:07*jhMikeS saved a link to that
23:08:08pixelmabluebrother: I don't have that tool here...
23:08:19***Saving seen data "./dancer.seen"
23:08:38bluebrothershould I do the conversion and check it in? If there is something missing tucoz could still change it.
23:08:52pixelmawould be very kind :)
23:08:53bluebrotherthe result looks fine here
23:08:57preglowjhMikeS: yeah, on its way
23:09:14preglowjhMikeS: will url you in a sec
23:09:47jhMikeSI'm looking in the directory now :))
23:10:00preglowit won't end up there
23:10:01thewhoOn the sim, audio playback is not stopped when FM radio is started.
23:10:22jhMikeSahh..the first one you didn't put there and I was all prepared :(
23:10:41LinusNlinuxstb_: i have uploaded all the bootloaders and patches
23:11:39thewhoBut I don't know the sim well enough to create a patch
23:12:10toffe82markun : do you have the link for the arm doc I found the other day ?
23:12:47bluebrotherhmm. What's the correct mimetype for the svg? image/svg+xml? Or just leave it as text?
23:13:20markuntoffe82: let me look. It was on
23:13:58linuxstb_LinusN: Thanks. But did you forget the Linux version? Same URL as the Windows version, but s/win32/linux32/
23:14:21toffe82yes, it would be good to post it on the wiki, I didn't check the copyright but could be a good thing if it is possible
23:15:12 Join aleska [0] (n=aleska@
23:15:49 Part thewho ("Someone with more knowledge about the sim: please look at radio.c:428 and take away the #ifndef if it's not needed")
23:15:55markuntoffe82: I believe you can request if from ARM on a CD. I found some thread about the legalisty and it didn't seem a problem to post it somewhere.
23:16:15 Join Peps [0] (
23:17:04markuntoffe82: I also found it here. With ZIP it's only 3.5MB!
23:17:15 Join webguest32 [0] (i=53556757@gateway/web/cgi-irc/
23:17:19toffe82markun: I was just asking because it seems that more people are going to work on arm
23:17:26LinusNlinuxstb_: done
23:17:34Pepswith The latest ipod build, when I initialzed the database it got the tracks but my playlists were not converted. is there an easy way to do that?
23:17:57markuntoffe82: yes, all the newer DAPs seem to be ARM based
23:18:15 Join BHSPitMonkey [0] (
23:18:16toffe82put some link on the wiki
23:18:47LloreanPeps: To play m3u playlists you just need to access them via filetree, or the playlist catalog feature. They don't belong in the database.
23:19:50Pepsyes, but how do I get them out as correct m3u, i.e. one that maches whatever rockbox convention is?
23:20:38LloreanPeps: As long as the m3u points to the files on the device, and is an m3u as per the standard format, it'll work. I'm not really sure what you're asking now.
23:21:03 Join stripwax [0] (
23:21:34PepsI created playlists using gtkpod. I want to convert them to rockbox, but I have no idea where there are located on the ipod
23:21:41preglowyou're coming to tell us you've decrypted the nano 2g firmware, right :>
23:21:56stripwaxhaha :-) sorry to disappoint
23:22:00LloreanPeps: GTKpod probably uses the Apple Database. The playlists are probably embedded in the xml database it uses.
23:22:26Pepsso that is the q, any utility which will extract them from there?
23:22:27LloreanPeps: You'll need to find a program or tool that can convert playlists from the retail apple firmware's format to the standard M3U format.
23:22:27stripwaxjust lurking actually.
23:22:50LloreanPeps: There was a patch in the tracker for a plugin that did it, you can probably find it with a search if it's still there
23:23:44PepsLlorean, do u remember the name of the plugin?
23:23:49 Quit Febs ("Chatzilla 0.9.77 [Firefox]")
23:24:17LloreanPeps: Firstly, I'm a "you" not a "u", and no, I do not. As I said, you'll have to search for it.
23:24:43linuxstb_LinusN: Thanks again.
23:24:47PepsLlorean, thanks, sorry about my typing abilities
23:24:57 Join funky [0] (n=repulse@unaffiliated/funky)
23:25:35bluebrotherpixelma: committed.
23:25:52pixelmathank you! :)
23:26:10webguest32is it possible that compiling the current svn source fails on a nano build ? (i am getting "undefined references.." errors in settings.list.o)
23:26:22bluebrotherhmm, there seems to be something suboptimal with the transparency in the png for the html version.
23:26:36bluebrotherbut we can ask tucoz about that. At least the manual now builds.
23:26:45bluebrotheryou're welcome
23:29:27Lloreanwebguest32: What arm-elf-gcc version do you have?
23:29:44linuxstb_webguest32: The build servers don't report any problems - (the Nano column is green - meaning no errors or warnings)
23:30:36webguest32Llorean: gcc version 4.0.2
23:31:03Lloreanwebguest32: Please upgrade to 4.0.3 as it's the current recommended version, and then see if your errors remain.
23:31:19 Part LinusN
23:31:40webguest32Llorean: i see. ok thanks.
23:36:41 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
23:36:44jhMikeSalmost finished reading it ... quick read :P
23:36:50 Join miepchen^schlaf [0] (
23:37:21stripwaxLlorean, looks like the latest available from the cygwin package is 4.0.2-1 ?
23:37:28peturis it me or does rockbox no longer detect a new file after usb disconnect (since root menu)?
23:37:37preglowpetur: it's the same here
23:37:50bluebrotherPeps: iirc that patch was a tool for the host PC, not a plugin. And it got posted in the wiki unless my memory serves me wrong
23:37:51jhMikeSpetur: me too
23:38:40amiconnpetur: This detection only ever worked when browsing the root
23:38:51Pepsbluebrother, if you can remeber a name that will make googling it possible that would be most appreciated
23:38:59amiconnSo you need to go there and it *might* notice the new bootfile
23:39:01peturamiconn: that's what I suspect now...
23:39:10amiconnIt always worked that way
23:40:36Lloreanstripwax: Look again, 4.0.3 should be available.
23:42:02stripwaxLlorean - hm, looks like I was still using rather than . I checked in the wiki and noticed that the user url has changed..
23:42:17stripwaxLlorean - but yes, you're right, it is there (and updating now ;-)
23:42:30jhMikeSyes, the rooty thing should be changed imo
23:43:12markuntoffe82: with 7zip the document is only 2.9MB
23:43:32 Quit obo ("bye")
23:43:39jhMikeShave it check the firmware checksum too ... suppose I'll repeat that alot 'till everyone agrees with me :)
23:43:42 Quit Rondom ("Ex-Chat")
23:43:46bluebrotherPeps: seen the "search" link on the top of the wiki pages? Tried searching for "itunes"?
23:44:03 Quit petur ("gtg")
23:44:07bluebrotherfound it easily that way, it's the third match: ConvertiTunesDBtoTagCache
23:44:09webguest32Llorean: the vmware image was updated last month ... do you reckon i should download that .. or a dist-upgrade of my old vmare image should do the trick ?
23:44:10linuxstb_Peps: The "itunes playlist" tools was indeed a plugin -
23:44:24Lloreanwebguest32: You're best off just downloading the new image.
23:44:39bluebrotheroh, you're looking for playlists?
23:44:55bluebrothermissed that part.
23:45:01jhMikeSpreglow: the Q flag is of no use in DSP?
23:45:02webguest32Llorean: ok cheers.
23:45:09preglowjhMikeS: arm?
23:45:14preglowjhMikeS: if so, we haven't got it anyway
23:45:16toffe82markun :can you put it on the wiki ? is there a page with dev doc on the wiki ?
23:45:27preglowjhMikeS: if we had an arm5 core: hell yes it would be of use
23:45:31preglowjhMikeS: but we only have arm4t
23:45:34jhMikeSit's always like that
23:45:39preglowindeed :/
23:45:53preglowarm5 is very nice for dsp
23:46:06bluebrotherPeps: FS #4771
23:46:15Pepslinuxstb_, thanks, but does this mean that I need to compile it myself (praying that it is not out of date?)
23:46:37 Quit Juice^ ("Leaving")
23:46:45bluebrotherPeps: yes.
23:46:45linuxstb_Peps: Yes, it does. And it will almost certainly be out of date - it's a year old.
23:47:03bluebrotherthe last comment even tells it's out of date and doesn't apply cleanly anymore
23:47:43*bluebrother still waits for his ipod to arrive
23:47:47Pepswell, i jumped into this only yesterday, I am not sure I want to try and compile rockbox at this stage :(
23:48:07bluebrothercan't itunes export its playlists to m3u?
23:48:10linuxstb_bluebrother: Which ipod are you getting?
23:48:16 Quit kubiixaka ("Miranda IM! Smaller, Faster, Easier.")
23:48:34PepsI will check, I am using gtkpod on ubuntu, does not have itunes
23:48:35linuxstb_bluebrother: If it can, they won't be any use in Rockbox, as they would most likely use the PC filenames, not the obfuscated ipod filenames.
23:48:38Pepsor windows :)
23:49:02 Join kubiix [0] (
23:49:20bluebrothermini, but I still don't know if it's 1G or 2G.
23:49:41bluebrotherhmm, that's a valid point.
23:50:07Pepsanother Q. I love rockbox but scrollwhhel performance is hidious on my 60gb ipod video. is there anything I can do?
23:50:24linuxstb_No, just wait until it's improved in Rockbox...
23:50:41*linuxstb_ looks towards Slasheri
23:51:14preglowdamn, i'd almost forgotten that slasheri was working on improving that
23:52:11Pepslinuxstb_, so it is a known issue?
23:53:02linuxstb_Most people wouldn't call it hideous, but it's accepted it can be improved.
23:53:20Pepsgtkpod was able to export the playlist. I see something like (/media/ipod/iPod_Control/Music/F36/WRSY.mp3). I assume rockbox did not move the files, so that should work?
23:53:21linuxstb_(or maybe they would...)
23:53:58linuxstb_Peps: You may have to remove the "/media/ipod" part at the start of the names, but other than that, yes they should work.
23:54:00PepsI have long lists and accelaration does not work well. takes ages to get to the z's
23:54:06 Quit OgMaciel ("mv OgMaciel $HOME")
23:54:10bluebrotherlibitunesdb seems to be a good start for writing a converter for the itunesdb playlists ...
23:54:12linuxstb_Peps: Then scroll backwards from the top.
23:54:35Pepslinuxstb_, you are a true hacker :)
23:55:00Pepslinuxstb_, still, even getting to the d's is a struggle
23:55:18bluebrotherthe ipod doesn't have paged scrolling?
23:55:36linuxstb_I don't think so - not enough buttons.
23:55:39Pepswhats that? i may be missing something here
23:55:56bluebrothertoo bad −− that is a quite handy function
23:56:14linuxstb_It shouldn't be needed if the scrollwheel accelerated properly - which it doesn't at the moment.
23:56:15mcphailbluebrother: the apple firmware is good enough that it isn't needed
23:57:02Pepslinuxstb_, any chance it will happen in the near future, or is this low priority
23:57:07Lloreanmcphail: What does the Apple Firmware have to do with it at all?
23:57:08bluebrothermcphail: might be, but e.g. on my h120 it's really nice. Especially because of the slow lcd.
23:57:29bluebrotherLlorean: I think he referred to scrolling long lists
23:57:46mcphailLlorean: the default firmware scrolling is fast and smooth. Good for scrolling long lists.
23:57:49jhMikeShaving the dest operand on the left look backward to me now
23:58:20mcphailLlorean: hence no need for "paged scrolling"
23:58:42Lloreanmcphail: Yeah, but the fact that the original firmware doesn't need it has nothing to do with whether or not Rockbox has it, or why.
23:58:44mcphailPeps: i hacked my 3g scrolwheel last night. Works better but still not exactly smooth

Previous day | Next day