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

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

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

#rockbox log for 2005-12-06

00:00:20Paul_The_NerdThat's a common complaint
00:00:33Paul_The_NerdI also like the visual style in 1 better. (Possibly just because Shion had glasses)
00:00:50ender`there's two of us
00:00:53 Join muesli_- [0] (
00:01:13Paul_The_NerdIt's that "geek girl in glasses" weakness
00:01:15ender`i only started ep2, but in some shots kos-mos looks just horrible
00:01:31ender`i can't really decide on shion though
00:01:47Paul_The_NerdHave you seen the art for Kos Mos in Ep 3?
00:01:48LinusNPaul_The_Nerd & ender`: while we do appreciate good conversation, we kind of prefer to have rockbox development discussions in this channel
00:01:55Paul_The_NerdSorry LinusN.
00:02:05ender`ok ok :)
00:02:06LinusNno worries
00:02:18cannardso anyways, i was playing xenosaga on rockboy...
00:02:39Paul_The_NerdI don't think the Coldfires are *quite* fast enough
00:02:58ender`hey, if they ported FFX to NES, why not?
00:03:39 Quit yngwi (Read error: 104 (Connection reset by peer))
00:03:47ender`btw, what was gameboy's resolution (and what's H1xx resolution)?
00:03:49Paul_The_NerdY'know, an NES emulator on Rockbox would be kinda neat.
00:04:13Paul_The_NerdGameboy's resolution is vertically 16 pixels more than the H1xx screen, I think. At least, rockbox trims 16 rows iiuc.
00:04:17Paul_The_Nerdrockboy, rather
00:04:29 Join yngwi [0] (
00:04:53Bagdertalk_id(options[gui_select_get_selected(&select)].voice_id, true);
00:04:59Bagderthis is actual code
00:05:15Bagderand gui_select_get_selected() is a macro
00:05:21 Quit matsl ("ERC Version 5.0 (CVS) $Revision: 1.776 $ (IRC client for Emacs)")
00:06:26 Quit yngwi (Client Quit)
00:06:33 Part leftright
00:06:40ender`ok, can anybody tell me what i need to set if i want rockbox to mix the end of the song with 3 seconds of the next song (without fading out/in)?
00:10:58Bagderis that gui code filled with this
00:11:17*Bagder puts on his Code Police badge
00:12:36Paul_The_Nerdender': Mode Mix, 3 seconds for each of the durations, I think?
00:13:06ender`i'll try...
00:13:13Paul_The_NerdI haven't really used it much. I played with it briefly back when the options were all new and shiny, then decided I liked my songs to end crisply.
00:14:01ender`i like them to overlap a bit when shuffling
00:15:16 Join DreamTactix291 [0] (
00:19:01 Quit muesli- (Read error: 110 (Connection timed out))
00:19:16Soul-EIf rockbox has an NES emulator
00:19:17 Join Karny [0] (
00:19:19Soul-EI would play it all day
00:19:36Soul-Eand a SCUMMVM emulator too. I'd just play monkey island and maniac mansion
00:19:55Karnyi hope someone can help me
00:20:00Paul_The_NerdScummVM would apparently be rather hard to port, due to the large amounts of C++ and etc.
00:20:12Soul-EIsn't Rbx programmed in C?
00:20:27KarnyI'm trying to compile rockbox by source... but I ca't seem to find sh-elf-gcc
00:20:28Paul_The_NerdC and C++ aren't the same thing.
00:20:45ender`use a C++ to C translator :)
00:21:00Bagderit using C++ should not be a very big problem
00:21:16Bagderunless it uses very fancy parts of it
00:21:20*Paul_The_Nerd shrugs.
00:21:29Paul_The_NerdThis is just what I've been told.
00:21:42Karnyany ideas?
00:21:54LinusNKarny: windows or linux?
00:22:02KarnyLinusN, linux
00:22:07Karnyubuntu specifically...
00:23:08Karnyawesome!! just what I've been looking for!
00:23:16Karnythanks LinusN!
00:23:31MoosLinusN,markun: think you ReplayGain can be moved in Sounds Settings too (like Crossfeed)?
00:23:32LinusNKarny: where did you look?
00:24:58Mooswith those clipping mode things, it could be cool to have the prevent clipping ReplayGain option in Sound Setting (here all music collection ReplayGain tagged)
00:25:57cannardSoul-E, your best bet would be zork :p
00:26:37LinusNMoos: sounds reasonable
00:26:53KarnyLinusN, I searched the wiki a bit and tired a fair amount of googling
00:26:58*amiconn wants this clipping modes to go away again
00:27:10*amiconn should post to the dev ml
00:27:30KarnyI was also lazy I looked thru a lot of debian and ubuntu repos to see if anyone had rockbox binaries ;)
00:27:33*amiconn has a working patch with dB volume and no artificial volume limit
00:27:35LinusNi'm with you amiconn
00:27:38Moosamiconn: why you didn't reply to len0x in ml, for give your opinion?
00:28:16Moosamiconn: good luck for this
00:28:23LinusNKarny: both those links are in the "documentation" section at
00:29:03amiconnMoos: For one, I was away for 3 days. Then, he changed a thread topic instead of starting a new one. This somewhat distracts attention if you aren't interested in the start of the thread
00:29:18KarnyLinusN, I know... but It's kinda vague as to what the whole process is... sorry for wasting your time
00:29:30Moosamiconn: yes indeed I noticed that too
00:29:48 Join RotAtoR [0] (
00:30:27Paul_The_Nerdamiconn: Is the patch on the patch tracker, or can I download it somewhere?
00:30:46LinusNKarny: don't worry
00:31:16MoosI'm wondering why len0x don't like to speak here ??? :)
00:31:35Moosis the function of this IRC no ?
00:31:44KarnyI might try to write a faq if I get this to work... sure there are other debian/ubuntu/linux users who could use a simple howto
00:32:26amiconnPaul_The_Nerd: Not yet on the tracker, but downloadable:">
00:32:56Moosamiconn: that will be pretty fun
00:33:25Paul_The_NerdThank you amiconn. :)
00:34:27LinusNah, finally usb mode on h300
00:34:40Moosamiconn: at least if so len0x motivated you a bit to work in the dB volume scaling :)
00:34:50MoosLinusN; F├ęlicitations
00:35:21LinusNwill commit later
00:35:25LinusNsleep time
00:35:28LinusNcu all
00:35:31MoosBonne nuit
00:35:36Paul_The_NerdMake Clean doesn't seem to clean everything anymore...
00:35:41 Part LinusN
00:36:03amiconnMoos: Yes, somewhat. For one, the clipping mode options really annoy me. Then, some experiments I did convinced me that we should not limit artificially. We don't do so on the recorders and Ondios for a long time, so why should we do so on the player and iriver?
00:36:21Karny'nother stupid question...
00:36:37amiconn(especially since the maximum possible volume of the iriver is low compared to the archoses)
00:36:46Karnyiriver h320... is that sh-elf or m68k-elf?
00:36:47Moosamiconn: True, I'm totaly with you
00:36:49 Quit DreamTactix291 (Read error: 110 (Connection timed out))
00:37:06amiconndB volume is a thing that both me and Linus wanted for some time
00:37:23amiconnKarny: m68k-elf
00:37:28 Quit muesli_- (Read error: 110 (Connection timed out))
00:38:12amiconnPercent just don't make much sense for a logarithmic scale, and 100% not being 0 dB (on archos) is another reason to switch to dB
00:39:12Karnyamiconn, thanks!
00:39:28Moosall said, go, go for dB switching :)
00:39:36amiconnAll this was on my personal todo list (except the non-limiting thing which resulted from experiments), but with low priority
00:40:02Paul_The_Nerdamiconn: Also, I just tried to apply it against a fresh CVS and I think it didn't fully apply in sound_menu.c
00:40:11amiconnThe drawback of todo lists is that they are always too long and even tend to grow :(
00:40:20Moosamiconn: I assume you've got a big todo list, no?
00:40:27Mooshehe :)
00:40:40amiconnPaul_The_Nerd: Hmm, maybe this is because of the crossfeed move
00:40:48*amiconn goes to update his patch
00:41:11Moosamiconn: how look the unicode thing?
00:41:21Paul_The_NerdI'm not good at reading the .rej file and figuring out how to manually apply
00:41:25*Paul_The_Nerd is not a good Nerd.
00:41:32amiconn(or maybe the gcc4 warnings frenzy? Bagder?)
00:41:49Karnynight all
00:41:58 Quit Karny ("sleep")
00:42:12Bagderif so, then it should be easily to resolve
00:42:17amiconnMoos: Looking good so far, just some glitches
00:42:21 Quit ender` (Read error: 104 (Connection reset by peer))
00:42:32Moosamiconn: good to hear
00:43:49amiconncrossfeed it was.
00:44:28Paul_The_NerdSo, can I redownload then? Purty please?
00:46:09 Quit YouCeyE (Read error: 104 (Connection reset by peer))
00:46:41 Join YouCeyE [0] (
00:46:43amiconnThere. Same URL
00:47:17Paul_The_NerdThank you kindly
00:47:23markunamiconn: a strange thing happened. I fixed the code to reject language files that are too big.. and now it loads greek without problems. Must be doing something wrong.
00:48:47amiconnDoes building the simulator now work on bsd?
00:49:43markunno, there is a new problem. there is a unicode.h in /usr/X11R6/include..
00:51:22 Join tvelocity [0] (
00:53:20*amiconn hates this kind of clashes
00:53:33amiconnI already ran into them numerous times
00:54:06markunShould we rename our unicode.h or try to solve it in a different way?
01:01:03 Quit Rob2222 (Read error: 104 (Connection reset by peer))
01:01:53markunIt would be nice if a language file contained a string with most (or all) of the used chars so they can be preloaded into the cache, just an idea
01:02:21Bagderit would be easy to write a script to do that
01:03:28markunMaybe no so useful for a chinese language file maybe, might become quite a big string :)
01:03:54Bagderwell, it could still be useful if done right
01:04:03Bagderespecially for a language like chinese
01:05:34markunIf we could order the charst with with the most used first we could limit the number of chars to read to the size of the cache
01:05:48Bagdergood idea
01:05:49*markun should learn how to type..
01:06:10 Join edx [0] (
01:07:37markunI'm looking for a function to get the size of a file..
01:07:52*Bagder goes to bed
01:07:58markungood night Bagder
01:08:15amiconnOooh, new vmware version
01:08:24MoosBagder: good night (noticed red builds?)
01:09:34 Join w98 [0] (
01:09:56***Saving seen data "./dancer.seen"
01:12:53w98hey all, quick question - how do i get a copy of the mkboot utility?
01:13:41 Quit w98 (Client Quit)
01:15:10Paul_The_Nerdw98: The source code is available. It's in the tools directory.
01:15:12MoosGood night @ all
01:15:17 Quit Moos ("Glory to Rockbox")
01:19:13 Quit YouCeyE (Remote closed the connection)
01:19:37 Join YouCeyE [0] (
01:20:41markunamiconn: I fixed the language problem. Shall I add a message "Language file too big" when it is? Or just leave it like this because it shouldn't happen?
01:21:28amiconnWhat was the problem?
01:22:22amiconnI don't think it's necessary to add a message, at least not together with unicode
01:22:32markunfirst the language buffer was filled and then it was checked if the file was small enough.. leaving the language buffer corrupted
01:23:33markunI'll commit this fix. When the language file is too big nothing will happen.
01:24:03markunBut we might get some questions from translators if their files stop working
01:24:16markunwithout warning them that the file is too big
01:24:43amiconnThe != itself check should have worked
01:25:11amiconnThe read() call is limited to the buffer size
01:25:55amiconn...but of course, the buffer will be corrupted instead
01:26:15amiconn(and the != check 'wastes' one buffer byte)
01:26:33markunI think I solved it quite ok. Better than reloading the old file.
01:27:21amiconnfilesize() ?
01:27:54markunand renamed filesize to fsize
01:28:33amiconnThe cookie & version check also happen too late in current cvs code
01:28:48markunthey do?
01:28:56 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
01:29:27amiconnIn current cvs, the file gets loaded first, corrupting the currently loaded language.
01:29:53markunso we should read the first 2 bytes to variables first
01:30:00amiconnThis corruption is only resolved if (1) the file is small enough and (2) cookie and version both match
01:31:31 Join Mongey| [0] (n=mongeyc@
01:35:23Mongey|anyone here
01:42:00Mongey|what firmware does a new psp come with
01:42:25markunI don't know. This is #rockbox..
01:42:50markunamiconn: Do the version and cookie have the be in the buffer?
01:42:51Mongey|lol i know but.....meh nvm
01:43:04amiconnmarkun: don't think so
01:43:14markunok, then I'll commit my fix
01:43:29*amiconn is fixing the red recorder builds
01:45:36 Quit Mongey| ()
01:47:12markunGood luck amiconn, I'm going to sleep now
01:48:07amiconnnite markun
01:48:35amiconnThe fix is rather simple, Bagder just overlooked some of the macro calls he wanted to remove
01:49:14markunI still think we might give some feedback if the language file cannot be loaded
01:49:41*linuxstb wonders why Bagder changed some string parameter types from char* to unsigned char*
01:50:52*amiconn thinks this belongs to the gcc4 frenzy
01:51:16 Join w98 [0] (
01:51:28linuxstbI know it's part of the gcc4 fixes, but I'm confused why he chose that solution.
01:51:44linuxstbHe also added casts of the form (unsigned char*)"my string"
01:52:02linuxstbBecause of changing strings to unsigned char*
01:52:32linuxstbs/strings/string parameters/
01:52:32w98hey all, last irc client was acting goofy ... can someone describe how I can get the 'mkboot' utility to patch my H320 firmware?
01:53:08linuxstbw98: It's described here:
01:54:15w98it describes a 'tools' folder that I don't see after unzipping the rockbox software though
01:54:31linuxstbYou need to download the Rockbox source.
01:55:05linuxstbAnd then type "make mkboot" in the tools directory (and similarly for any other tools you need).
01:55:18w98thanks, that's what i was looking for
01:55:36w98are the makefiles set up for a 'make all' instruction?
01:55:51linuxstbOr if you are running Windows or have Wine installed, just use the fwpatcher.exe application.
01:56:12linuxstbw98: Not any more... Tools are now only built when they are needed.
01:56:31linuxstbUp until a few days ago, "make all" would have worked in the tools directory.
01:56:45amiconnUgh... I just got a floating point exception from mpa.codec in the h120 x11 sim (!)
01:57:04linuxstbimpossible :)
01:57:34linuxstbWhat's gcc been doing now?
01:57:40amiconnI just started another track from the browser while music was playing
01:59:35 Quit w98 ()
02:01:30Soul-Eany news on a US firmware bootloader
02:03:44 Quit DangerousDan (Read error: 110 (Connection timed out))
02:09:38markunamiconn: shall I rename unicode.h to utf8.h to resolve conflicts?
02:12:26linuxstbmarkun: Are the routines general unicode, or specifically utf8 ?
02:13:43markunThe all use utf8 in some way, but I would have prefered to keep it as unicode.h
02:14:01markunThe problem is that there is a unicode.h in /usr/X11R6/include
02:14:29*linuxstb is actually reading the patch now...
02:15:01amiconnThis seems to be bsd specific. Both my debian vm and cygwin don't have this unicode.h
02:15:14amiconnStill the problem should be solved
02:15:34linuxstbYes, I searched Debian (dpkg -S) for unicode.h, and no debian packages include a file with that name.
02:15:48amiconncygwin has an utf8.h
02:15:58 Join DreamTactix291 [0] (
02:16:00amiconn(somewhere deep in the perl libs dirs)
02:16:08 Join Shaikoten [0] (n=shaikote@
02:17:33linuxstbOther libraries have things like gunicode.h (glib2) and xmlunicode.h (libxml2). So maybe rbunicode.h ?
02:17:43linuxstb(I don't like it that much)
02:20:14markun"/usr/X11R6/include/unicode.h was installed by package libunicode-0.4_4"
02:23:37linuxstbmarkun: Yes, that's available in Debian as well. My search was wrong...
02:25:20 Join JonSenior [0] (
02:26:00JonSeniorLinusN: What was the solution to the USB problem and how did you track it down?
02:26:32 Join Benacool [0] (
02:26:34JonSeniorI know he's not here, he'll get it when he reads the log file! ;-)
02:28:01 Quit JonSenior (Remote closed the connection)
02:31:08 Join TCK [0] (
02:32:08TCKhey guys, want to hazard a guess at what date Easter Sunday is next year? ;)
02:48:07 Quit webguest49 ("CGI:IRC")
02:49:07 Quit DreamTactix291 (Read error: 110 (Connection timed out))
02:58:20 Join Rob2222 [0] (
03:03:11 Quit DJDD_ ("Trillian (")
03:06:10 Join max [0] (i=max@
03:09:21 Join kkurbjun [0] (
03:10:00***Saving seen data "./dancer.seen"
03:10:26kkurbjunI was wondering if anyone could point me in the direction that I need to go to get color wps' working?
03:10:37kkurbjunI have a color bmp loader made as a plugin
03:11:04kkurbjunand it works fine on the H300
03:11:26 Join XavierGr [0] (
03:14:20golf7nice job man, i would start working on plugins, but all i know is basic php
03:14:29golf7if i could learn a real language, i could start
03:14:59 Quit tvelocity ("Leaving")
03:15:19kkurbjunI've been looking at the gwps-common, but Im not sure what needs to be changed beyond that.. I'm suspecting alot
03:15:50kkurbjunthanks, I don't know php, so you've got one up there : )
03:28:39golf7its pretty basic
03:29:18golf7not much markup required just what you want to do, so you dont have to do alot of like setting what type of coding your doing
03:31:00 Join DJDD_ [0] (
03:40:06 Join amiconn_ [0] (
03:42:25 Nick Lost-ash is now known as ashridah (n=ashridah@
03:44:27 Quit Kingstone (Read error: 110 (Connection timed out))
03:54:57 Quit max ()
03:57:40 Quit amiconn (Read error: 110 (Connection timed out))
03:57:40 Nick amiconn_ is now known as amiconn (
04:03:03 Join Jungti1234 [0] (n=jungti12@
04:04:59 Quit Jungti1234 (Client Quit)
04:12:12 Join Paul_The_Nerd [0] (
04:17:32kkurbjunwell, I got a color backround to show in the wps
04:18:08kkurbjunthe text fashes alot now though
04:20:01 Nick ashridah is now known as Lost-ash (n=ashridah@
04:20:41kkurbjunalso, again if someone can reply to this later: is there a difference between lcd_mono_bitmap and lcd_bitmap on the other devices. I've started changing the image buffers to the data type fb_data, which I don't think will affect other devices, but I'm not sure
04:22:40 Quit Sucka ("a bird in the bush is worth two in your house")
04:28:47 Join DreamTactix291 [0] (
04:34:06 Join Kyl3 [0] (
04:38:42 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
04:43:16golf7nice work man, i hope you get the flashing worked out
04:48:24 Quit Kyl3 ()
04:48:49Lost-ashthe flashing really needs to be solved with a proper z-buffer, or something, i suspect, and/or region updates.
04:51:47kkurbjunthat was what I was thinking (more a z buffer), but I havn't found the code that does the updates yet
04:52:05kkurbjunright now I'm just checking to see if it'll work ok without a backgroung and color icons
04:53:54Lost-ashbut there's flashing on the remote that doesn't seem to be coming from z-issues
04:54:08Lost-ashthat said, i haven't updated in a few days
04:57:14 Nick Lost-ash is now known as ashridah (n=ashridah@
04:59:15kkurbjunwhat theme are you using?
04:59:28kkurbjunand player
05:00:52ashridahH140, and i've been using boxes.
05:00:58ashridahhaven't looked at some of the newer ones
05:01:12 Join Rob2222_ [0] (
05:01:19kkurbjunahh, there's a background image on the remote for boxes as well..
05:01:31kkurbjundoes it flash if you move the background image?
05:01:44kkurbjun(out of the wps directory)
05:05:36ashridahi haven't tried. haven't got my player handy. that's all the way upstairs :)
05:06:22kkurbjun: )
05:10:02***Saving seen data "./dancer.seen"
05:12:16 Quit Rob2222 (Read error: 104 (Connection reset by peer))
05:31:48lostlogicgah, how am I getting more ticks allocated to functions than the time totals
05:31:53lostlogicI must be fecking it up.
05:32:09ashridahone of the things i used to love about gprof.
05:32:24ashridahthe sample rate was 10ms/sample
05:32:36ashridahso functions running quicker than that used to get bogus numbers
05:32:42ashridahso my programs would run 2-3 seconds
05:32:51ashridahand gprof would say "this program ran for 30 seconds"
05:32:55ashridahand i'd go "wtf?!"
05:33:00lostlogichehe, do you think that 100us timer pulses is reasonable on the coldfire?
05:33:15ashridah.1ms? hmm.
05:33:34lostlogicactually... now that I think of it, I base that on CPU_FREQ so it's somewhat longer at 90mhz than 100us
05:33:38ashridahkeep in mind that the cpu speed changes tho
05:33:48ashridahyou might need to record boosts too?
05:33:54ashridahso you can change the calculation rates
05:33:56lostlogicdoesn't change during profiling, always boosted cuz profiling is SOOO SLOOOOWWW
05:34:21lostlogicI have a todo on my forum post about this to setup a timer that specifically doesn't recalculate timer interval on boost changes for profiling, but it's a low priority.
05:34:31ashridahheh. i remember bringing my issues up with my lecturer about gprof. he didn't know what to do. useless tool.
05:34:50ashridahso i basically submitted a report with analysis, and a footnot that said "these numbers are completely inaccurate due to XYZ"
05:34:59ashridahstill got perfect marks. silly
05:35:03lostlogicI submitted some like that in school
05:35:12lostlogicteachers like it when people are smart enough to know that stuff.
05:35:31ashridahthe thing that annoyed me was that he never really addressed it
05:35:48ashridahall i got was some other moron going "use gprof, it rocks!"
05:35:59ashridaheven tho i WAS using it, and it doesn't. ;)
05:36:10ashridahwish the uni had sprung for a decent commercial compiler
05:36:35lostlogicmeh, they should have just done a project to improve gcc... :)
05:38:11ashridahat second year level C programming?
05:38:27lostlogicnot for _you_ to do it, but as a project at the UNI
05:38:47ashridahah, they dropped compiler-based subjects at my uni, totally pissed off
05:39:04ashridahclosest we get is computing theory, which has grammars
05:43:52lostlogicyay, fixed another profiling bug −− I'm a retard.
05:47:28lostlogic(I was never stopping timers, ugh)
06:04:32 Quit Benacool ()
06:09:12 Join Kingstone [0] (n=sg1@
06:12:38lostlogic*ponders* profiler is telling me that ov_time_total is taking up 99% of profiling time...
06:12:43lostlogicI _think_ something may be buggy here.
06:16:19 Nick ashridah is now known as Lost-ash (n=ashridah@
06:18:22lostlogicrecursiveness breaks my sheeeeiiiitttt
06:20:48Soul-Ecan someone tell me my ip
06:21:00Soul-Ei forget this channel is logged
06:26:40XavierGryou can see your ip at what's my ip cpm
06:29:52 Join ModernExecutive [0] (
06:31:22ModernExecutivesoul-e: you there?
06:38:03 Nick Lost-ash is now known as ashridah (n=ashridah@
06:47:19ModernExecutivedid you get my msgs?
06:48:01*ashridah blinks
06:48:03ashridahModernExecutive: yes?
06:50:09 Quit markun (Remote closed the connection)
06:50:16 Join markun [0] (
06:52:29Soul-Eno news that i know of
06:52:43ModernExecutivei saw that you inquired like an hour ago
06:53:05ModernExecutive& got no answer
06:54:30 Nick ashridah is now known as Lost-ash (n=ashridah@
06:54:58 Join hshah [0] (
07:00:00 Quit Soul-E ("CGI:IRC (Ping timeout)")
07:10:05***Saving seen data "./dancer.seen"
07:12:15 Quit RotAtoR ()
07:25:02 Quit ModernExecutive (Read error: 110 (Connection timed out))
07:35:38 Quit edx (Read error: 110 (Connection timed out))
07:40:23 Join MrShlee [0] (
07:40:30 Part Midgey34
08:04:44amiconnkkurbjun: (1) There is a difference between lcd_mono_bitmap and lcd_bitmap on all units except the pure b&w ones
08:05:21amiconn(2) For proper colour support in wps, the bmp loader is the simplest part
08:11:21 Join ender` [0] (i=ychat@
08:12:41MrShleeI've noticed on the H3XX. that the LCD isn't actually turned off.. the backlight is off but you can see see the faded updated lcd.
08:16:18amiconnA lot more needs to be changed if we want to do it right, like allowing colour attributes for text
08:17:17amiconnThe scroll thread is one such place that needs work, and imho we should switch to textboxes for scrolling at the same time
08:18:28amiconn(3) I have some ideas for colour support, e.g. possible global background image support in the LCD driver, with almost no slowdown
08:20:22amiconnMy idea for the BMP loader was to make it work similar to bmp2rb, but without all the intermediate buffers
08:21:08amiconnIt should allow all BMP colour depths for input, and output either a mono bitmap or a native bitmap (greyscale or colour)
08:21:36amiconnFor b&w targets both output formats are identical
08:25:13 Join Bger [0] (n=Bager@
08:51:31 Quit hshah (Read error: 110 (Connection timed out))
09:02:02preglowi like the idea of stopping audio when skipping
09:02:05preglowSlasheri: implement!
09:02:14BagderI agree
09:02:28preglowi often wonder whether i actually managed to press the button or not
09:02:34preglowdamn slippery joystick
09:02:35BagderRockbox does so on Archos since years back, definitely a good thing
09:02:47Slasheripreglow: yes, i will do that (at least when crossfading is not activated and instant audio skipping is not possible to do)
09:03:07preglowwhen crossfading is on i agree it should not stop audio
09:03:16Slasheribut i still haven't figured out how to detect if instant skipping can be done or not
09:03:42Slasheriprobably easiest is to check that if we need to buffer from disk, then stop audio
09:03:45preglowwe should also implement some way to fade out audio, so we don't get clicks
09:04:39preglowand also figure out/agree on how to fix the dsp latency issue
09:04:44*Bagder reads the monthly chunk of OpenNeo commits
09:05:13 Join einhirn [0] (
09:06:12Bagderamiconn: thanks for fixing my red build tonight
09:06:28preglowmarkun: you know, it IS a shelving filter
09:07:14preglowat least kind of
09:08:22amiconnBagder: np. The code police was a bit sloppy it seems ;)
09:08:34preglowmarkun: at the coefs you use, it's got just three db attenuation at lower freqs
09:08:48*Bagder returns his badge and goes to the corner
09:10:09***Saving seen data "./dancer.seen"
09:10:21XavierGrwhat lcd_set_drawmode does?
09:11:40 Join TCK- [0] (
09:12:39XavierGrI am looking into this because I use the lcd_putsxy in a loop to render statistics. But sometimes letters are garbled, or previous info is left on the screen.
09:13:35XavierGrnice page thanks
09:14:28 Join edx [0] (
09:15:50 Quit cannard ("2.0 Build 3515")
09:25:10 Quit TCK (Read error: 110 (Connection timed out))
09:33:53 Join DJDD__ [0] (
09:38:09Bagderamiconn: I'm supporting your volume patch approach
09:38:57preglowhell, me too, since it was all i wanted in the first place :-)
09:39:11preglowis the volume displayed in decibel as well, now?
09:39:12preglowor perct?
09:43:14preglowSlasheri: btw, did you find out anything about the queueing issue?
09:43:15 Join Vlad0man [0] (
09:43:51Slasheripreglow: ah, forgot that.. i will check that today :)
09:44:04preglowplease do, it's extremely annoying behaviour :/
09:44:14preglowand slicer battery life to shreds
09:46:23amiconnBagder: Look at the dev ml...
09:49:01Bagderand now you can look ;-)
09:49:46preglowamiconn: but yeah, is volume displayed as decibel as well?
09:50:21preglowwell, it's not a problem for me
09:50:34preglowsome users might wonder, but hell, they'll probably use fancy wpses anyway
09:50:56preglowit's got my support
09:51:33amiconnAh, I need to adjust the wps tag(s) as well...
09:51:50preglowthink i'll give it a spin now
09:51:57preglowif i can find an usb port on this frigging thing
09:52:24 Quit DJDD_ (Read error: 110 (Connection timed out))
09:52:36XavierGrpreglow is there any reason to advance playback speed by steps or 0.1?
09:52:43XavierGrit is way too small.
09:53:26preglowXavierGr: i just think we need an acceleration scheme there, so it advances faster if you keep the button pressed
09:53:35preglowwe need a small playback speed step for serious use
09:54:14XavierGryes but it is very difficult to make significant changes in small intervals with this.
09:54:18 Join pabs_ [0] (
09:54:32XavierGrand very awkward to have to hold Play button.
09:54:55XavierGrat least make the steps larger or make it on a menu. (Playback or Sound)
09:55:03 Quit pabs (Read error: 104 (Connection reset by peer))
09:57:28XavierGrand there that I was saying that I will never use playback speed; It is very handy now that I play midi files ineed.
09:57:38preglowthe steps wont be made larger, but like i said, we could implement some kind of acceleration scheme
09:58:11XavierGrthat would be good too.
09:58:32preglowmight look at it later, but no time now
09:58:54XavierGrjust mentioning.
10:00:21 Quit Vladoman (Read error: 110 (Connection timed out))
10:00:39 Nick pabs_ is now known as pabs (
10:01:54preglowwindows REFUSES to let me unmount the h120
10:02:22Bagderits not easy to let go of rockbox
10:02:25Bagderit sticks
10:02:50XavierGrI never unmount my ihp from windows
10:02:56XavierGrI just pull the plug.
10:03:00preglowseems there was an hour long write cache operation
10:03:13preglowXavierGr: i did that, rockbox now hangs on booting
10:03:19XavierGrAh, I have set caching to off.
10:03:35XavierGrwindows cache is EVIL
10:04:21preglowthe entire .rockbox dir is heavily mangled
10:05:15XavierGrSeriously though with windows cache to off, I had never problems with "unplug and play".
10:05:30preglowi thought that was the default
10:05:30preglowbut no
10:05:33preglowapparently not
10:05:46XavierGrno cache is always on/
10:05:50preglownot on xp
10:06:07preglowi can't delete my wps folder now
10:06:37preglowand scandisk doesn't work
10:06:38XavierGrcan you start the original fw?
10:06:58Bagderthere's on original fw?
10:06:59XavierGryou could always make a quick format from the iriver fw, it comes handy sometimes.
10:07:06preglowok, so how the hell do i turn caching off?
10:07:31XavierGrwait a little to remember it.
10:07:40XavierGrgo to my computer -> properties
10:07:44preglowi'm _not_ formatting it, it'll be weeks before i've got the means to fill it again
10:08:38XavierGrhardware and the go to your hard drives
10:08:49XavierGr(you must have the ihp plugged in)
10:09:19XavierGrchoose ihp's HD and select properties
10:09:38XavierGrthen go to policies tab and check the "optimize for quick removal"
10:10:23preglowpolicies tab isn't there
10:11:12XavierGrdid you clicked properties on the ihp's HD?
10:11:30XavierGrok another route then
10:11:38XavierGrgo to My computer
10:11:50XavierGrright click your ihp and select properties
10:12:24XavierGrgo to hardware tab and select "again" ihp's HD.
10:12:36preglowno policy tab there either
10:12:43XavierGrThen click properties and the the policies tab. It should be there.
10:12:49preglowit's not
10:12:51XavierGrI have it in front of my eyes
10:12:55preglowit's not here :P
10:13:00XavierGrdo you have many services disabled?
10:13:00markunpreglow: Is it easy to make a low-pass shelving filter?
10:13:01preglowSpecial Windows
10:13:11preglowmarkun: it's not hard, at least
10:13:39XavierGrpreglow: can you see a policies tab for your other HDs?
10:13:44preglowXavierGr: no
10:14:10XavierGrWhat windows do you have?
10:14:17preglowbut bah
10:14:24preglowi'll go find another computer and try something there
10:14:39preglowperhaps i can at least fix the drive errors
10:14:43preglowdon't know, not my computer
10:15:13XavierGrlucky you...
10:16:30linuxstbBagder: Did you read my questions in the log about your char/unsigned char changes? I was wondering why you decided to make (some) strings unsigned char.
10:17:11BagderI didn't do any closer analyses, I've just been scanning through the sources attempting to stomp out the warnings
10:17:37Bagderin the cases where there are/were mixed uses, I mostly did the smallest change
10:17:57Bagderin general, we should remove lots of unsigned char * uses
10:18:07Bagderwhich I think my changes make clear
10:19:11BagderI'm tempted to switch off the warning inhibit soon
10:19:15linuxstbBut, for example, in gui/splash.c you changed the type of a string parameter (fmt) from char to unsigned char.
10:19:33Bagderbecause that is almost always used with a str() input
10:19:34linuxstbWhich means you have to do (unsigned char*)"my string"
10:19:42Bagderwhich is made to return an usigned char *
10:20:19linuxstbShouldn't str() be changed then? All the functions in string.h (and gcc's interpretation of literal strings) make strings plain char*.
10:20:25Bagderyes imho it should
10:20:37Bagderbut that's a big change
10:21:19Bagderwe should use "char *" for strings all over, unless especially motivated
10:21:48Bagderthe thing is, I feel a slight need to rush in my changes
10:21:53Bagderto disable the warning inhibit
10:22:04Bagdersince I see the same errors are going in while I'm working on this
10:22:43linuxstbBut the unicode patch could change things - it defines strings (with good reason) to be unsigned char.
10:23:13Bagderyes, when the unicode patch is applied we should have it sorted or sort it out
10:23:18XavierGrok tell me your opinion on a performance matter.
10:23:57preglowgot it fixed
10:24:09preglowgod damn piece of shit computer here obviously doesn't work
10:24:40XavierGrCurrently the midi2wav plugin is set to 30,000 bytes output. It is too low IMHO and I will get it fixed with a big pointer buffer. But can this small buffer affect the performance of the plugin?
10:25:05BagderXavierGr: it can, but I doubt you'll notice any significant speed change
10:25:42XavierGrso the disk doesn't hog down the cpu?
10:26:08XavierGrI mean that maybe the cpu waits the okay from disk to continue processing.
10:26:54linuxstbBagder: With unicode in mind, I think you were right to change strings to unsigned char. We just need to carry that through everywhere - which will always mean casting to (char*) when using the string.h functions.
10:26:58BagderXavierGr: you'll find out when you increase the buffer size
10:27:23preglowamiconn: nice patch, i like it
10:27:49XavierGrwell then I will. But I am not an optimization expert so this is the only I thing I can do for this poor soul.
10:27:54linuxstbXavierGr: Simply do a test with disk writing disabled, and see what difference there is. In my early codec tests, I never found the writing to make much of a difference.
10:28:30preglowdoubt it'll make a difference
10:28:34BagderXavierGr: real performance gains are made by improving the stuff that take the most time and is used most frequently
10:28:53Bagderoddly enough
10:29:20XavierGrnow that you mention it: I was fiddling with filenames (wanted to output to a new file instead of a fixed one) and it had errors so there was no file output but the results were the same. That's dissapointing. :(
10:29:45XavierGrBagder: You mean loops and stuff?
10:29:55Bagderand math
10:30:34XavierGrYes but it probably has to come to ASM base optimization, so no go for me. But I will try for the fun of it nevertheless.
10:31:13 Join KN|stiff [0] (i=phhome@
10:31:16Bagderdidn't the midi code still suffer from remaing floats?
10:32:12XavierGrI don't know to tell you the truth. I just raised it from the dead when I found the old xxx2wav files in the attic.
10:32:40 Quit MrShlee ("Changed by popular Request - "I am the heart of your darkness, you stinking oompa-loompa morons!"")
10:32:44XavierGrNow I am just fiddling around to have user input and stuff.
10:32:48preglowwhat? midi plugin uses xxx2wav?
10:33:14XavierGrand it is currently dead.
10:33:22XavierGrNot completely though
10:33:31preglowwell, the coder vanished, so...
10:33:46XavierGreven without the xxx2wav it will work but it will not output anything
10:33:49XavierGrnot a file nor sound
10:34:00XavierGrI emailed the coder.
10:34:18XavierGrHe is interested about it, but he has no free time or a target to test.
10:34:47XavierGrHe said that he will get back to it when he finds the time or buy an actuall Rockbox compatible DAP.
10:35:44 Join LinusN [0] (
10:36:40LinusNi'm getting pretty tired of all these ifdefs for different hardwares
10:37:10XavierGrHi Linus, what's your advice then?
10:37:17Bagderwe should strive to decrease ifdefs
10:37:27LinusNwell, we don't have much of a choice really
10:37:46LinusNmy major source of grief is the ifdef policy
10:37:55Bagderthe aim should be to push ifdefs to lower-level code, and attempt to keep higher-level ifdef-free
10:38:06LinusN"check for features, not for platforms"
10:38:28BagderI disagree with that conclusion
10:38:38LinusNthat policy works for the higher layers, but not the lower
10:38:40BagderI would even argue that policy has helped us a lot
10:39:10LinusNfor instance, we will have plenty of platforms with the uda1380
10:39:24LinusNbut it is connected in different ways on all of them
10:39:43Bagderperhaps we should make target/ dirs
10:39:48Bagderfor target-specific code
10:39:53Bagderdrivers etc
10:40:04Bagderand cpu/ for cpu
10:40:07LinusNor allow #ifdef IRIVER_H100_SERIES in the drivers
10:40:39Bagderyes, but the problem with that is that people will read those and re-use them in other code
10:40:48Bagderso they'll spread
10:41:09Bagderperhaps we should name them something that sounds driver-ish
10:41:17LinusNso we should let it be a mess in the drivers to prevent a mess in onther places?
10:41:33Bagderdid I say that?
10:41:48LinusN(in response to your "the problem is..."
10:42:16Bagderwell, you can't pretend the problems don't exist, you need to think of them
10:42:22Bagderand then consider what the proper action is
10:42:37BagderI'm not saying what to do, I'm just thinking out loud
10:42:38LinusNi have a few suggestions
10:42:41preglowLinusN: doing some x5 work?
10:42:50LinusNpreglow: no
10:43:32LinusNbut none of them reduces the #ifdefs in any significant way
10:44:31LinusNone approach would be a target/ dir, with all the gory hardware details
10:45:08LinusNsmall functions for all the necessary stuff, like enabling ata power, usb power etc
10:45:19LinusNmight even be macros
10:45:20Bagderthat's how u-boot does it and I like that
10:45:35Bagder(it supports ~200 boards and ~30 cpus...)
10:46:10*preglow supports anything leading to fewer ifdefs
10:46:37Bagderthe linux kernel does it like that too
10:46:47preglowyes, and let's split crt0.S while we're at it
10:46:48LinusNi'll ponder about the target/ approach for a while
10:47:19BagderLinusN: and a separate cpu/ dir for the cpu-specific things that are shared between targets using the same cpu
10:47:33 Join ingenius [0] (
10:47:42Bagderasm-optimized things etc
10:47:56LinusNgood idea
10:49:04*linuxstb waves goodbye to Rockbox's cvs history...
10:49:34linuxstbBut we shouldn't let CVS stop us from doing the right thing.
10:49:38LinusNa good opportunity to go subversion :-)
10:51:58Bagderwell, it takes some admin tasks on server-side to go subversion nicely
10:53:33 Nick Vlad0man is now known as Vladoman (
10:55:14Bagdercoffee time
10:57:09preglowdamn straight
11:03:09markunand with subversion please change the names of apps/player and apps/recorder
11:08:35*amiconn has no problems with ifdefs in drivers as long as they aren't spread all over, but grouped nicely
11:08:50amiconnbacklight.c being an example...
11:09:28preglownot always possible
11:10:13***Saving seen data "./dancer.seen"
11:12:07amiconnWe also employ the empty macro method in a number of places.
11:12:21amiconnThere are more places where this method can be used
11:13:00 Join Polo_o [0] (
11:13:08Bagderthe empty-macro method is a fair approach to make the code easier to read
11:13:36BagderI'm not usually a fan of it, but when it can remove ifdefs I'm in favour of it
11:14:03LinusNcpu_boost() is a nice example
11:14:32LinusNlcd_update() should have been a nice example
11:15:23Bagderyeah, lcd_update() should be fixed like that
11:15:43amiconnAfaik, lcd_update() already uses this method
11:15:47 Quit Bger (Read error: 104 (Connection reset by peer))
11:16:21amiconn...but there's the problem that it doesn't work in plugins
11:19:28 Join muesli_- [0] (
11:20:24*Lost-ash hands LinusN a copy of cvs2svn
11:20:28 Nick Lost-ash is now known as ashridah (n=ashridah@
11:22:49preglowashridah: what time is it over there now?
11:23:49 Join Jungti1234 [0] (n=jungti12@
11:23:55ashridahnot all that late, really.
11:25:10ashridahit claims to be 5:30am on the shell i'm using tho ;)
11:25:19preglowwell bah
11:25:28preglowi'm waiting for some aussie to mail me some dev tools
11:25:34preglowbut that obviously wont happen before tomorrow, then
11:25:41ashridahmail as in snail-mail?
11:25:45preglowno, email
11:25:53preglowsnailmail would have taken forever
11:26:35ashridahi'm curious as to what kind of dev tools you don't already have
11:26:44preglowtools for motorola 56000
11:27:24LinusNpreglow: which target?
11:27:26amiconnLinusN: Do you have an idea what may cause this:
11:27:37preglowLinusN: some proprietary thing, lake huron
11:27:52preglowit's got a bunch of 56002s in it
11:28:06LinusNamiconn: bad soldering is my guess
11:28:23amiconnI observed similar behaviour on my recorder v1 (spurious BUTTON_OFF events) lately
11:28:33amiconnIt only happens when batteries are low
11:28:37LinusNwe decreased the button poll rate because of this phenomenon
11:29:03amiconnThis didn't happen some weeks ago...
11:29:24LinusNmaybe we have increased it again? :-)
11:30:22amiconnThe strange thing is that while I would suspect changes in the button driver, I can't find any which might cause this behaviour
11:31:37 Join criis [0] (
11:31:39amiconnThe spurious OFF events are also annoying when playing bejeweled...
11:31:44LinusNit may be caused by a voltage drop for some other reason
11:35:34amiconnIs it possible that my backlight handling cleanup causes this?
11:36:04amiconnThere used to be a (dirty) call to backlight_tick() from button_read()
11:36:34amiconnNow backlight_tick() is properly registered as a separate tick task
11:37:27amiconnThat may cause it to be called before the button tick
11:37:46LinusNbut how can that cause a spurious OFF?
11:38:24amiconnSwitching the backlight might cause voltage changes
11:38:41LinusNamiconn: ay, yes
11:38:41amiconnHmm, but then the backlight tick just sends events to the backlight thread...
11:39:34amiconnWhat puzzles me most is that it causes OFF events, but no others
11:39:45amiconnOFF is connected to a port pin...
11:40:39preglowbtw, shouldn't current_tick be declared volatile?
11:40:58amiconnIt's simple to check whether it's connected to the backlight thing, I should do so
11:41:14amiconnJust set backlight to always on (or always off)
11:41:18preglowand why isn't it unsigned?
11:42:20amiconnIt's signed to make it easier to compute time differences
11:42:26LinusNpreglow: why volatile?
11:42:39ashridahisn't volatile basically ignored by the compiler these days?
11:42:46ashridahor am i thinking of 'register'
11:42:46Bagderashridah: indeed not
11:42:51Bagderregister is, yes
11:42:53preglowi thought the optimiser would really need volatile as a hint
11:42:55ashridahaah, okay
11:42:57LinusNashridah: you mean "register"
11:42:57preglowwhat if current_tick changes in a function
11:43:15LinusNpreglow: *any* global variable can change in a function
11:43:18*Bagder searched for days for a bug involving volatile the other day...
11:43:22LinusNthe compiler knows that
11:43:25preglowLinusN: alrightey
11:43:38preglowso what's the use of 'volatile', then?
11:43:45Bagderhardware bound stuff
11:43:45LinusNfor example
11:43:59Bagder*ptr = 0x200000;
11:44:02Bagder*ptr= 0
11:44:03Bagder*ptr= 0
11:44:04LinusNreading a FIFO
11:44:09ashridahio ports, basically stuff it shouldn't expect to be cachable iirc
11:44:38 Join Febs [0] (
11:44:45preglowso it's basically a write thing?
11:45:01preglowsince all global variables will be read each time anyway
11:45:03BagderI'd say it is mostly used for writing
11:45:04LinusNlet's say you poll a port pin
11:45:24Bagderno they won't
11:45:33XavierGrchar *buf_pt = (char *)codec_malloc(BUFFER_SIZE); <- is this right? BUFFER_SIZE is 1MB.
11:45:37Bagderthe compiler will keep them in registers within the same function
11:45:42Bagderunless volatile
11:45:44preglowXavierGr: you can't malloc that much memory
11:45:48preglowXavierGr: please don't malloc anything
11:46:01 Quit KN|stiff ("( :: NoNameScript 3.81 :: )")
11:46:15preglowuse of malloc in rockbox is forbidden unless you really need it
11:46:22preglowand even then we'll just correct you and tell you to stop being silly ;)
11:46:32XavierGrpreglow it is for the midi2wav plugin.
11:46:41preglowXavierGr: i know, but nothing in rockbox should use mallocs
11:46:42LinusNpolling current_tick without calling a function in between will fail
11:47:00preglowXavierGr: the only places in rockbox we use mallocs, is in codecs, because we haven't bothered to remove them yet
11:47:00XavierGrwhy you included then in the first place?
11:47:06LinusNfortunately, we usually call yield() or sleep() between polls
11:47:10preglowXavierGr: because of the codecs, we didn't want to rewrite it all
11:47:30XavierGrbut I am just testing anyway. And it is like a codec so...
11:47:46 Join Bger [0] (n=Bager@
11:47:49preglowXavierGr: but no, you don't have much malloc memory anyway
11:48:15XavierGrBut even if I can't allocate that much shouldn't it return a null?
11:48:23linuxstbXavierGr: Are you working on a plugin or a codec?
11:48:26 Join justsomeperson [0] (
11:48:35preglowBagder: you basially contradict linus here, he said the compiler knows global vars can change at any time
11:48:42Bagderhe's wrong
11:48:45XavierGrplugin midi2wav
11:48:53preglowwhich makes me right! hooray!
11:48:57Bagderyou can easily prove that
11:49:00linuxstbXavierGr: In which case, you can use the audio buffer.
11:49:06LinusNlook what i just wrote
11:49:16LinusNpolling current_tick without calling a function in between will fail
11:49:18Bagderoutside functions, they can change
11:49:19LinusNfortunately, we usually call yield() or sleep() between polls
11:49:57preglowLinusN: yeah, i thought about that, i was just trying to clarify whether you'd thought of it or not
11:50:01linuxstbXavierGr: There is a function called something like get_audio_buffer() - the jpeg viewer uses it for example.
11:50:04XavierGrlinuxstb: *mp3buf ?
11:50:22preglowXavierGr: you're going to use the mp3 buffer for storage space once the midi codec is turned into a real codec anyway
11:50:47preglowi'm just used to seeing all variables which are manipulated by isrs declared volatile
11:51:02preglowand then: why isn't it unsigned? :P
11:51:14XavierGrxxx2wave.c has an *mp3buf that relates to the audio buffer. (or so the comments say so.)
11:51:34ashridahpreglow: to allow differences to be calculated easily apparently
11:51:36LinusNpreglow: i believe it makes it easier to handle time comparisons
11:51:47preglowvalid point
11:51:56amiconnpreglow: [11:44:11] <amiconn> It's signed to make it easier to compute time differences
11:51:57preglowavoid casting hell
11:52:27ashridahso, does rockbox hang if the variable rolls over? :)
11:52:51BagderI doubt many users have experienced a tick counter roll-over ;-)
11:53:11preglowbtw, we still have a 'prevent clipping' option we can use to include bass and treble clipping
11:53:35ashridahBagder: what, no-one's kept their player running for more than 400 days? (assuming it's a signed int, not a long or something)
11:53:57preglowshould we use that for that purpose as well, or should we just ignore the bass and treble limiting people are going to be nagging about?
11:54:08Bagderashridah: people are such reboot freaks ;-)
11:54:54ashridah<MR crowd> ooh! new cvs commit, new bleeding edge! it seems faster. is it faster?
11:55:03ashridah<rockbox> uh, that was a spanish language update.
11:56:58amiconnashridah: It's signed long on all current targets
11:57:19ashridahah. so more like the year 3000 :)
11:57:41Bagdersome 248 days I believe
11:57:47Bagderour longs are 32 bit
11:57:53linuxstbashridah: It's in units of 10ms (i.e. 100 ticks/sec)
11:57:58*ashridah should have been more specific
11:58:13LinusNpreglow: maybe some people would line a "prevent clipping" option
11:58:19ashridahof course, counting ticks, not seconds.
11:59:08amiconn*If* we want to implement clipping prevention, it should be done at the app level, properly indicating the changed values
11:59:26amiconnNo behind-the-scenes mangling...
12:00:03LinusNamiconn: does the app have enough knowledge about the dac to adjust properly?
12:00:41 Nick ashridah is now known as Lost-ash (n=ashridah@
12:07:06linuxstbI'm just thinking aloud, but what about a sound settings screen that makes everything (volume, bass, treble, loudness etc) visible together, with some indication where a certain combination could cause clipping. If a "prevent clipping" mode was added, then this would visibly auto-adjust the other settings when a change to one setting introduced clipping.
12:08:25LinusNlinuxstb: i like that idea
12:09:11XavierGrNow just great! I used the ready buffer from xxx2wav and the wave file is all garbage sounds.
12:09:59XavierGrI will try the get_audio_buffer.
12:09:59LinusNlinuxstb: the problem still remains, the application doesn't have the necessary knowledge about the dac to do those adjustments
12:10:18justsomepersonhaving sound adjustments at app level is a good idea
12:11:22justsomepersonin fact that what I'd do if I had time originally instead of current "prevent clipping" option
12:11:51justsomepersonwhat knowledge we don't have about DAC ?
12:12:40LinusNthe application has only the set_xxx() API
12:13:04LinusNit doesn't know the limits and the relations between the bass/treble and the volume
12:13:15LinusNand not anything about the prescaler
12:13:18markunWill the volume display be changed to dB? I like it
12:13:25justsomepersonset everything as %
12:13:34LinusNmarkun: yes
12:14:45LinusNof course, we could extend the sound driver api with get() methods
12:15:02LinusNthen the app can easily find out what has changed and display it
12:15:09markunThe Gigabeat's firmware is getting worse and worse compared to rockbox :)
12:15:21Bagderdon't they all?
12:15:34 Join KN|stiff [0] (i=phhome@
12:15:35XavierGrwave2wv seems to use 2MB buffer (or 1)?
12:15:47 Part Sando
12:16:09linuxstbCould we write a function called something like prevent_clipping(&vol,&bass,&treble,ADJUST_BASS|ADJUST_TREBLE) which will adjust the specified settings to a combination that won't clip.
12:16:20markunBagder: they only advantage some have is unicode support.. but not for long :)
12:16:31 Join wxls [0] (
12:16:41 Join Sando [0] (
12:16:45Bagderwell, some would say WMA support or video playback of course
12:17:02 Nick wxls is now known as kamsa (
12:17:04justsomepersonlinuxstb: good idea
12:17:25XavierGror DRM. But I hate DRM
12:17:30kamsadoes the h300 series have usb support now? (referrring to the latest build update)
12:17:41justsomepersonbut this is essentially what my patch does internally...
12:19:01markunLet's develope our own DRM and force other companies to use it to be rockbox compliant :)
12:19:27 Part kamsa
12:21:01 Join Zak1392 [0] (
12:22:00 Join leftright [0] (
12:23:06Zak1392haha, USB mode is cool. thanks linus ;)
12:24:14LinusNZak1392: you're welcome
12:24:17leftrightreplay gain already has a "prevent clipping" menu, this conflicts with the "prevent clipping" menu under volume, very confusing
12:24:24Jungti1234XavierGr, Why do you hate DRM?
12:25:15linuxstbJungti1234: I think it's obvious - DRM removes your freedom to do what you want with the music you buy.
12:25:44Zak1392hey linus, do you know why rockbox sucks up so much battery life?
12:25:54Jungti1234I think so.
12:26:03LinusNZak1392: no
12:26:18LinusNZak1392: some people claim it's the lcd, but i haven't measured
12:26:30Zak1392any ideas at all?
12:27:03LinusNit could be the lcd, but it may very well be a port pin collision somewhere, draining power and heating up the cpu
12:27:04Bagderthose people blaiming the LCD hardly knows that anyway, they just guess
12:27:24Bagdersince it is an obvious difference to the original fw
12:27:44Bagderbtw, can the original h300 firmware shut off the lcd on demand?
12:27:52amiconnLinusN: I think it should be possible to do those adjustments at the app level. Just make sure current_volume + MAX(bass_boost, treble_boost) <= max_volume
12:28:03Zak1392but sometimes when the battery goes down to about 60% in rockbox it shows full battery in iriver firmware
12:28:05amiconnThe thing is, I start to believe that it doesn't make sense
12:28:34BagderZak1392: well, you shouldn'y judge battery life based on ther percentage shown
12:28:40Bagderonly pure run-time counts
12:28:54LinusNthe battery level hasn't been calibrated on the h300
12:29:13LinusNamiconn: doesn't make sense?
12:29:20amiconnWe will have (and already have) more factors than just treble and bass which influence the point of possible clipping
12:29:30LinusNoh yes
12:29:43Zak1392but sometimes it shows the battery down in iriver firmware too
12:29:55Bagderagain, only pure run-time counts
12:30:13amiconnLoudness on MAS3587/MAS3539, probably software EQ on iriver, replaygain on archos...
12:30:16 Quit muesli_- (Read error: 110 (Connection timed out))
12:30:26LinusNamiconn: true, true...
12:30:42amiconnFurthermore, quiet tracks don't clip as early as loud tracks (of course)
12:31:57Zak1392how come it says 2h 36min remaining at 100% battery life but when it goes down to 98% it says 8h 47min?
12:32:04amiconnYou see, I changed my opionion. We shouldn't limit the volume artificially. If it clips, the user will notice and reduce the volume
12:32:18LinusNstill, i think it could be useful with a mode where the user can crank the volume safely
12:32:23amiconnThis is in fact the case with most other equipment
12:32:31BagderZak1392: what in everything we write doesn't reach you?
12:33:00BagderZak1392: you can help us fix that, or you can continue asking those questions
12:33:25linuxstbThe behaviour some users want (iriver firmware users) is to automatically scale back the bass when the volume is increased, and also to automatically return the bass to the high setting when volume is decreased.
12:33:31XavierGrI agree with Linus on that
12:33:57BagderI guess the problem is to define "safely"
12:33:57LinusNi wouldn't use that mode, but i believe it has a use
12:34:07Zak1392bagder: it's just that 2h is kinda scary lol
12:34:13LinusNclipping can destroy a tweeter
12:34:21LinusN(so i've heard)
12:34:26BagderZak1392: why? If the estimate is wrong?
12:34:58Zak1392it just freaked me out when i first saw it
12:35:05leftrightwhyy not treat the volume as most hifi's do, with no behind the scenes trickery, turn your hifi up too much distortion, simple
12:35:25amiconnleftright: That's exactly how my patch handles it
12:35:32Bagderthat's what I think too
12:35:36Bagderlet's do that first
12:35:53Bagderand only if it reaaaaally gets annoying, add a weird way to avoid it
12:36:36justsomepersonI can already see flaming in the forum about that :)
12:36:41amiconnI doubt that it will get annoying. We do this on archos recorders for years...
12:37:07Bagderjustsomeperson: users will complain no matter what we change, that's not a reason to not change
12:37:25 Part criis
12:37:25linuxstbI agree - let's take this one step at a time. amiconn's patch is a much better basis for whatever we may want to change in the future.
12:37:26justsomepersonfirst users ask for a feature for several months, then they get it for couple of days and then it gest taken away :)
12:37:32LinusNyou just switch between complaining crowds
12:37:36Bagderjustsomeperson: I disagree
12:37:46Bagderjustsomeperson: they mostly wanted the previous behaviour removed
12:37:49leftrightI honestly dont expect the bass/treble to change if I fiddle with the volume,
12:37:50Bagderand this still has that removed
12:38:19justsomepersonnot a single user complained about new options I did... but whatever...
12:38:19Bagderand again: users will whine no matter
12:38:46Bagderjustsomeperson: you haven't been around here enough then!
12:39:31justsomepersonI haven't, but the option now is very customizable, however I do agree that internally it looks ugly...
12:39:38leftrightprevent clipping appears twice in different menu's why, do they do the same thing
12:41:02Bagderjustsomeperson: yes, but the existance of the option itself is also a problem
12:41:08leftrightI prefer the tradtional "audiophile method of volume control, which I can control
12:41:16justsomepersonbut all user sees is the ability to do whatever he wants - app level options do make more sense though, I agree
12:41:53justsomepersonleftright: which is why OFF option should be always there
12:42:18justsomepersonbadger: why option is a problem ?
12:42:45Bagderwe have too many
12:42:52justsomepersonso what ?
12:43:05Bagderso users are confused, can't find them, can't understand them
12:43:20justsomepersonif you worry about user-friendliness - half of the option should be removed...
12:43:23Bagderwe fight hard to limit the amount
12:43:55justsomepersonits a personal choice what do you want to change
12:43:55Bagderwell, we certainly don't improve usability by adding more
12:44:08BagderI mean, if avoidable
12:44:17leftrightjustsomeperson: a lot of the confusion appeard because no-one knew/understood what your menu options did, and the wiki wasn't updated
12:44:17 Join webguest87 [0] (
12:44:20justsomepersonIMHO its too late for that :)
12:44:28 Quit webguest87 (Client Quit)
12:44:33Bagdernothing is too late
12:44:35justsomepersonwhy not have "expert options"
12:44:50justsomepersonin one place insteat of all over options
12:45:07Bagderwhat would these expert options do?
12:45:14LinusNi don't like that idea
12:45:17preglowanyway, a general prevend clipping option would be nice
12:45:26justsomepersonit will hide all tricky options from newbie if you don't want to see them
12:45:28preglowwonder if it should be default on or not :>
12:45:34Bagderso what is a "tricky" option?
12:45:46LinusNmostly because there will be an eternal discussion about what options are "expert" or not
12:45:54amiconnpreglow: The problem is how to detect potential clipping...
12:46:00Bagderand all users will fiddle with the expert options anyway
12:46:29LinusNi think options are good, but only if really necessary
12:46:33justsomepersonno one would agrue (I mean endusers) where to put option as long as they can access them - believe me
12:46:36 Join newnick [0] (
12:46:43BagderI don't
12:46:54Bagderusers _always_ argue about where the options and menu items are
12:46:55preglowamiconn: assume worst case
12:47:03LinusN"i want xxx to be reachable with one click" - heard that before?
12:47:25Bagderjustsomeperson: there are currently several threads on exactly that topic going on
12:47:48justsomepersonyes, but that's basic functionality - clipping is not one of them...
12:47:55LinusN"why not use the xxx button to do yyy instead?"
12:48:17LinusNall users have different opinions on what basic functionality is
12:48:26justsomepersonlike I want radio to be accesible in one click is a valid request for instance...
12:48:37BagderI don't think you've been around long enough to make such general statements about rockboc users
12:48:37justsomepersonwhile the same don't applyfor clipping
12:49:02XavierGrthat's why TiMiD wanted to make a full configurable button map and menu positions! :p But that is way too difficult and tricky
12:49:04Bagderthese discussions always comearound
12:49:09justsomepersonI know users in general - they do listen to reasoning...
12:49:19Bagderjustsomeperson: yes, but want the results their own way
12:49:37Bagderjust like everyone else
12:49:51justsomepersonas long as they can do it in any way - they will make piece in the end
12:49:58 Join `3nergy [0] (
12:50:10amiconnpreglow: Well, if we assume worst case: 24 dB of bass boost together with +12 dB of some eq band means reducing volume by 36 dB
12:50:18amiconnThat's too quiet even for my taste
12:50:19preglowamiconn: don't reduce volume, just clip bass
12:50:43preglowamiconn: i think volume should always be the last thing to be capped
12:50:50LinusNme too
12:50:51preglowmore bass doesn't matter to me if the volume is lowered as an effect
12:50:55justsomepersonme too
12:51:18LinusNlet's cite fletcher-munson or whatever they were called
12:51:28amiconnReduceing the bass is possible... (when indicated properly) ... but how would you do this for the EQ
12:51:59preglowamiconn: exactly like that? if the volume is 0dB and someone tries to amplify, then hell no if you've got "prevent clipping" on
12:52:19preglowthe eq will come with a preamp so you can fix the levels
12:52:28amiconnAt 0 dB volume there is no problem, but at slightly lower values there is
12:53:04amiconnEQ band and hardware treble/bass might be additive or not, depending on which bands they act on
12:53:26preglowahh, indeed, again, we'll pretty much have to assume worst case
12:53:33preglowassume they overlap and prevent it
12:53:44amiconnE.g. 12 dB of bass boost + 12 dB of lowest EQ band mean a maximum boost of 24 dB
12:53:58preglowi don't think "prevent clipping" is an option for people who would dislike such behaviour anyway
12:54:16amiconn...but 12 dB of bass boost and +12 dB of highest EQ band mean a maximum of 12 dB
12:54:24preglowyou use "prevent clipping" if you want to avoid clipping at all costs
12:54:32linuxstbDoes it make sense to enable both the software and hardware EQs together?
12:54:40preglowwell, perhaps
12:54:44amiconnIt might
12:54:47preglowi don't know what the hardware eq filters looks like
12:54:52preglowthey might be one pole shelving filters
12:54:59preglowin which case they'll sound slightly different than mine
12:55:07amiconnTreble boost range is rather low on the UDA
12:56:18linuxstbFrom a user's point of view, it could be confusing to have bass settings in two different places.
12:57:06XavierGrWe can always say to them RTFM!!!
12:57:10preglowso do you suggest we leave the uda ones out?
12:57:17preglowthey're still useful, since they take no cpu
12:57:17LinusNi would
12:57:23preglowi kind of would
12:57:30preglowbut like i said, they're free, no extra battery required
12:57:33preglowand that's a good point on a dap
12:57:38linuxstbI would prefer them to both be adjustable on the same screen.
12:57:54Zak1392see ya guys
12:57:55LinusNhow does it work in the original firmware=
12:58:00preglowLinusN: that's all they have
12:58:02preglowLinusN: the uda ones
12:58:16Bagderand then the SRS stuff etc
12:58:18linuxstbSo somehow combine the software and hardware EQ settings - but I've no idea how.
12:58:23amiconnpreglow, LinusN: Perhaps make them selectable?
12:58:27preglowbut yeah, if we put the uda ones in the eq screen, then the proper ones wont really make sense for most users
12:58:31 Quit Zak1392 ()
12:58:40amiconnIf EQ is enabled, only use EQ, else use UDA treble/bass
12:58:48LinusNdon't they have those silly Jazz, Rock presets? Are they just uda settings?
12:58:56Bagderyes, I vote for the selectable approach
12:58:58preglowyou can choose
12:59:00Bagderuse either one
12:59:05preglowbut yeah, selectable approach sounds good
12:59:07LinusNamiconn: makes sens
12:59:36preglowbut yeah
12:59:43preglowsoftware eq will introduce a volume control
12:59:48preglowwill this fit better some place else?
12:59:54preglowi don't think so
13:00:03amiconnpreglow: Why?
13:00:15amiconnMy idea is as follows:
13:00:29preglowbecause eqs usually have them
13:00:35preglowpeople expect them
13:00:35 Part leftright
13:00:57preglowand they might be nice if you want to do some manual prescaler action
13:01:03amiconnThe EQ needs to scale down the whole signal by the same value as the highest boost in one band, to prevent clipping
13:01:24preglowyes, the good old prescale/postscale thing
13:01:32amiconnThis can be auto-corrected by boosting the main volume, like it's done for the UDA/MAS3507 presclaer now
13:02:14amiconnThe EQ just needs a way to tell it that it should reduce prescaling, to avoid global volume cap
13:03:22preglowbtw: now that we will have an extra software gainer, it becomes possible with amplitudes higher than 0dB
13:03:42preglowi think it should be possible to have higher gains than 0dB if "prevent clipping" is not enabled
13:03:54preglowhmm, or that might not be the correct way
13:04:07preglowit'd be too unobvious that 0dB isn't the "real" maximum
13:04:46amiconnIt's no different from the treble/bass boost as I see it
13:05:06preglowno, it is no different
13:05:18amiconnThe only difference is that you can apply positive gain values to all EQ bands
13:05:51 Quit justsomeperson ("CGI:IRC")
13:05:52preglowbut about this software gainer thing, am i the only one that wants to be able to digitally amplify the signal?
13:05:56amiconn...but that doesn't change the meaning of the main volume control, imho.
13:05:57preglowi've got some albums that are really quiet
13:06:13amiconnReplaygain does in fact the same
13:06:32preglowdoes what? gain the signal?
13:06:33 Join Zak1392 [0] (
13:06:50preglowsure it does, but i want to be able to control intuitively, which does not mean entering the menu jungle and adjusting the preamp
13:07:22preglowwhich again does only work for files that are tagged with replaygain, which almost none of my files are
13:09:29amiconnSame here, as I don't use replaygain
13:09:40markunreplaygain is nice
13:09:40preglowi'd plain and simple just like to be able to use positive db values, just like on some of the archoses, with the software gain taking action once you go over 0dB
13:09:45 Quit Zak1392 (Client Quit)
13:10:15preglowbut yeah, we can just plain and simple use the replaygain gainer for this
13:10:17***Saving seen data "./dancer.seen"
13:10:49amiconnThe archoses all have the ability to go above 0 dB, that's just how the datasheets define 0 dB for the respective DACs
13:10:52preglownice, but i don't like it on daps,
13:16:35 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
13:16:54markunpreglow: why not?
13:17:24preglowmarkun: in general, it makes music too quiet
13:19:01markunFor me it works great. Maybe my earphones are a bit more efficient.
13:19:27 Join webguest85 [0] (
13:20:22preglowmy earphones just broke, so i need new ones ://////
13:20:27preglowany recommendations?
13:21:14markunI'm looking myself. I like my PortaPros, but only for quiet environments like the library.
13:21:52 Join leftright [0] (
13:21:59preglowi never heard an earplug i liked
13:22:02amiconnI think it depends a lot on what you want and how much money you are willing to spend
13:22:03preglowi'd like to, though
13:22:11markunI would like to try the Sharp MD33, but I can't find a place to buy them.
13:22:12*amiconn uses Sennheiser MX450
13:22:15preglowi'm willing to spend a bit of money, i listen to music a lot
13:22:16leftrightsony noise cancelling earbuds are good
13:22:22 Quit TCK- (Read error: 113 (No route to host))
13:22:36leftrightnot the headphones
13:23:06XavierGrI use shure e2c, very good phones IMHO.
13:23:09preglowthese in-ear type earplugs have better sound than the usual ones, then?
13:23:14preglowbecause usual ones sound like shite
13:23:20leftrightI have UE10's and Sony noise cancelling buds, the sony sound is quite good
13:23:52XavierGrIndeed. I can hear details that I couldn't imagine with those earphones
13:24:30leftrightI also have the normal sony earbuds, they are complete trash
13:24:32*amiconn can't imagine using in-canal phones
13:24:37Bagdersounds you can't imagine? you mean like voices in your head? B-]
13:25:10preglowbut he can imagine those, you see
13:25:14leftrightthe sony's aren't very intrusive, the noise cancelling one's that is
13:25:20XavierGrBagder: Yeah something like that :D
13:25:28preglowhe's speaking about the sounds he doesn't hear in his head, few though they may be
13:25:48Bagderhead-external sounds
13:25:48amiconnNo. First, I think these things can't be comfortable. Then, the sealing would irritate me like hell
13:26:02XavierGrI oreder the sony mdr ex71 for ease of use.
13:26:15preglowi find most earplugs extremely uncomfortable when using them for long
13:26:18XavierGrDo they have that annoying neck strap?
13:26:20leftrightstrangely not so with the sony's amiconn
13:26:35XavierGrYou know the one that one channel cord is longer than the other...
13:26:42amiconnI'm already irritated by closed classic heaphones. I always prefer open ones
13:26:43leftrightI also have ER6's, which do hurt
13:27:06leftrightand sound way to bright, dont like their sound either
13:28:27leftrightUE make some lower priced models as well, but I cant vouch their tonal quality,
13:28:42 Quit t0mas ("brb")
13:29:20 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
13:30:37 Join nitromx [0] (
13:31:07 Quit nitromx (Client Quit)
13:31:20linuxstbpreglow: Can I have your opinion on this patch:
13:31:33linuxstbHe's added a _lot_ of assembler to the decoder - possibly too much IMO.
13:31:37preglowcertainly, gimme a sec
13:31:48preglowhow does it perform as it is?
13:32:16linuxstbIt's extremely fast already. Only a tiny boost with the CPU at 24MHz IIRC
13:32:53XavierGrNow THAT is totally strange.
13:33:03preglowhe also does a ton of c based opts
13:33:17XavierGrPlugin-Api didn't got boost those days right?
13:34:01XavierGrMy old compiled jpeg.rock (with same plugin verion) half-worked.
13:34:14preglowlinuxstb: have you/he tested this? i can't even find the macsr init
13:34:20XavierGrIt could display the pics and load picures while playing music.
13:34:44XavierGrBut it couldn't scroll pictures. I recompiled it and all is fine. But why this happened?
13:35:38linuxstbpreglow: He's tested it, but I haven't. He also said he can't find any SHN files using LPC compression - so that part is untested. I'm not sure if the other parts even use the EMAC.
13:36:15 Part leftright
13:37:37preglowi'd also love to know what performance improvement we'll see
13:37:43preglowcouple of notes you can relay/fix, whatever: 1. macsr needs to be initalised 2. i don't like the mcd5249_ prefix, use emac_ or something instead
13:38:39preglowman, this patch will really remove us from the ffmpeg code base
13:40:29preglowbtw, the suggestion about different peak meter colours sounds decent. i'd love to see more use of grayscales/colours
13:44:01linuxstbpreglow: Yes, the patch does make it very different from the original decoder. I think I'll ask the author to reduce the number of changes to just those that make a significant speed improvement. I'll mention your notes as well.
13:45:25preglowsounds nice
13:45:59 Join tvelocity [0] (
13:46:04preglowand ueah
13:46:09preglowtell him to test the lpc code somehow
13:46:15preglowi don't want untested code commited
13:46:52Jungti1234You help me.
13:47:20 Join ep0ch| [0] (n=ep0ch|
13:47:22Jungti1234Error happened during compile.
13:47:30pregloware you still using patches?
13:47:46ep0ch|hello :)
13:47:54markunpreglow: I told him to cvs update with the unicode patch still applied..
13:48:15Jungti1234hi ep0ch|
13:48:15preglowmarkun: pretty please, someone commit the unicode patch soon :///
13:48:49linuxstbpreglow: Yes, he can generate test files with lpc by using the "-l lpc_order" option to shorten when encoding the files. (the default value is 0 - i.e. no lpc)
13:48:59preglowthen there's no reason for not testing
13:49:07preglowi seriously doubt the code works, esp. without the macsr init
13:49:13markunThere are a few small problems, but maybe I can fix them after the commit.
13:49:45ep0ch|i dont suppose someone fancies adding "current bitrate" to the audio thread debug screen? i.e. to see what the current bitrate is for vbr files.
13:52:41chopped_porkyes *please* commit unicode at last, it's been too many days in state of maybe-it'll-get-commited-today-or-tommorow :)
13:53:26Bagderit was never "maybe today" until about yesterday ;-)
13:53:44chopped_porkmy bad
13:53:48amiconnLinusN: It seems that you fixed the dual ISD power for H1x0 as well, but the comment is incomplete now
13:54:07LinusNwhich one?
13:54:53markunLinusN: should I commit unicode and fix the remaining problems later?
13:55:17Bagderare the remaining ones any serious?
13:55:58LinusNmarkun: how does it affect voice and plugins?
13:56:14chopped_porkoh and btw, i've been using someones build with unicode to give it a shot (it worked great), and that build included linus' patch for left-from-browser-goes-to-wps and surprisingly found it extremely useful and intuitional - is that going to make into official rockbox source someday?
13:56:43markunthe viewer is not ready yet. Only reads utf-8 (so 7-bit ascii works). There is a problem with the starmap font (segfaults). The credidts on the player have a small glitch
13:57:12markunLinusN: I didn't try any voice related things.
13:57:16linuxstbep0ch|: How would you define "current bitrate" ?
13:58:24LinusNchopped_pork: the left-to-wps feature doesn't get my vote
13:59:03amiconnVoice does work as far as I tested it
13:59:10chopped_porkLinusN: oh. i thought since it's your patch you must've found it useful, but the rest of the team was against that or something :)
13:59:21amiconnmarkun: Did you fix the text alignment in the button bar?
13:59:37LinusNi only made it out of pure kindness
13:59:49chopped_pork:D three cheers for that, thank you:)
13:59:54markunamiconn: yes, I did
13:59:59amiconnmarkun: Also, we do need a smaller unicode font
14:00:19amiconnAre there any such fonts?
14:00:26Bagderwell, we don't need to have all quirks fixed before the commit
14:00:33amiconnYes, sure
14:00:48 Quit Seed (Nick collision from services.)
14:00:53 Join Seedy [0] (
14:00:54markunamiconn: well, I have some smaller unicode fonts (5x7 is the smallest) but it doesn't have japanese of course :(
14:01:00amiconnThe font cache might profit from having a real memmove(). Motivation for working on it...
14:01:27amiconnI already have fully optimised memmove() for SH1
14:01:38preglowamiconn: and memcpy for coldfire, that's half the work
14:01:39preglowgo go go!
14:02:11amiconnYes, memmove() is basically dual memcpy(), one working forward and the other working backward
14:02:18amiconn...and a selector between them
14:02:36markunI'll just commit unicode then, ok? ami, linus?
14:02:36 Join webguest28 [0] (
14:02:40amiconnWith asm, memcpy() is just a secondary entry point into memmove()
14:03:03webguest28please commit this unicode!!!!
14:03:06 Quit webguest28 (Client Quit)
14:03:20amiconnmarkun: Don't forget to convert the .lang. files. I could do the cleanup & version bump later
14:03:22preglowguerilla requesting
14:03:35markunamiconn: just do it in 1 big commit?
14:04:08*Bagder buckles up
14:04:24*linuxstb goes for a long lunch
14:04:37amiconnOh, iirc there's some work for a script magician as well. Bagder?
14:04:55Bagderto do what?
14:05:08amiconnThe codepage files need to be packaged in the distribution .zip
14:05:27*Bger goes with "unicode" screaming people
14:05:34BagderI can work on that when the files are in
14:06:30amiconnBger: The unicode patch does not yet touch the builtin font
14:06:48 Quit webguest85 ("CGI:IRC (Ping timeout)")
14:06:56amiconnmarkun: Imho that's an area that will need to be worked on next
14:07:09*LinusN hides and awaits the unicode impact
14:07:18amiconnSupport an on-disk default font, preferably using a builtin ascii part
14:07:38Bgeramiconn: guessed that it's this way, but ... :)
14:07:41 Quit KN|stiff (Read error: 104 (Connection reset by peer))
14:07:48 Join djgalak [0] (
14:08:01Bgerplease, built-in cyr and greek :)
14:08:01markunLinusN: I'm preparing a patch to apply to my commit dir, could take a while because I don't want to commit the wrong things of course.
14:08:42markunI'll add the right people to the credit file later, ok?
14:08:43BgerLinusN: 10x for the USB mode!! :)
14:08:52amiconnBger: Built-in non-ascii makes no sense. The .lng files need to be read from disk anyway
14:08:55XavierGrAh ye me Greel :)
14:08:55XavierGrnice typing
14:09:03LinusNmarkun: we expect a few "oops" commits, it's a tradition
14:09:09Bgerhehehe :)
14:09:14Bgerespecially for Linus:)
14:09:19preglowi sure as hell would with something as large as this
14:09:24*preglow hoists the unicode flag
14:09:41amiconn(but unicode is another reason to work on cyrillic + greek for the player)
14:09:53amiconnOh, and katakana for new-lcd players
14:09:57XavierGrnow who will catch first the MR post?
14:10:56 Join Mongey| [0] (
14:11:02 Nick Seedy is now known as Seed (
14:11:48Bgeramiconn: why need for .lng from disk ?
14:12:08Bgeru choose the built-in lang compiletime ... or i'm wrong
14:12:08 Quit djgalak (Client Quit)
14:12:11XavierGrRockbox Song Current Timer: 15:00 Rockbox Song Total Timer: 4:04 :((
14:14:06 Join webguest28 [0] (
14:14:51linuxstbLinusN: Thinking about the RTC, do you have any preferences for implementing it on the H300? You suggested either disabling interrupts during the read or reading it from an interrupt.
14:16:03markunShall I throw in some unicode fonts or do it in the next commit?
14:17:26LinusNlinuxstb: i think it doesn't matter, so do it with disabled interrupts
14:17:57LinusNhowever, it might be a good idea to read one byte at a time and enable the irq in between
14:18:08LinusNmarkun: add some fonts too
14:18:14 Quit ender` (Read error: 110 (Connection timed out))
14:19:31Mongey|(h300)just to let you know (if you dont already) when charging it charges with the i-river firmware unless you've already booted to RB
14:20:27LinusNMongey|: what do you mean?
14:20:37Bagderthat's how it works
14:20:46Bagdersince you didn't switch it on, it can't start rockbox
14:21:22BagderLinusN: I assume he means the charging screen it starts when you charge
14:21:24LinusNMongey|: you mean it starts the original firmware when you insert the charger?
14:21:50Mongey|unless youve already booted to RockBox
14:22:05LinusNthat's because the h300 bootloader only start rockbox if it was started with the ON button
14:22:25LinusNthat will change
14:23:10Mongey|LinusN; any idea when rockbox will be optimised for the h3x0
14:23:50Bagdersee topic
14:24:10preglowwe should have logbot spam the topic to the channel every ten seconds
14:25:10*Mongey| thinks how to alter his question so it not an esimate release date
14:25:30Bagderdon't bother
14:25:39Bagderthe answer is the same
14:25:52Bagderwe don't know
14:26:01Bagderif you help us work on it, it'll be sooner
14:26:03LinusNMongey|: the thing is that we can't predict anything, since it depends on how much spare time we have
14:26:04BgerMongey|: rockbox is a free-time work
14:26:16Mongey|i know
14:26:22*Mongey| = 14
14:26:28LinusNspare time is a scarce resource
14:26:38Bagder14 is a good age, no work and no serious obligations
14:26:42Bagderlots of spare time
14:26:46Bagderlots of rockbox
14:26:46LinusNespecially when youre a 36-year old family father
14:27:04Mongey|i cant code
14:27:25BagderI couldn't either once, then I learned
14:27:29Mongey|what do i need to learn?
14:27:33Jungti1234I love everybody. :)
14:27:43Bagderincidently, I started programming when I was 14! ;-)
14:27:50LinusNme too
14:27:55 Quit t0mas ("brb, reboot")
14:28:03Mongey|i can program a cnc lathe
14:28:05markunok, unicode is being committed..
14:28:07preglowit's happening!
14:28:15*LinusN ducks
14:28:16Bgerreally happening :))
14:28:33markundon't let there be red builds..
14:28:33*LinusN puts his party hat on
14:28:43LinusNmay the source be with us!
14:28:46*Bagder does the rare unicode dance
14:28:52Bgerplease, God ... no red builds :P
14:29:11markunJungti1234: I will do the latest korean translation and japanese tranalation next.
14:29:42Mongey|will gba roms work
14:30:02markunMongey|: no
14:30:25Bagdernothing is forever
14:30:42Bagderbecause all things change
14:30:52Jungti1234I want it.
14:30:53*LinusN detects a philosophical discussion
14:31:21*Bagder works from home, unshaved in his underwear and utters philosophical phrases
14:31:55preglowlike a true philosopher
14:31:56markunBagder: can you see if one of the builds has failed before it shows up on the website?
14:32:02Bgergo and dress quickly, you, bad boy! :P
14:32:19Bagdermarkun: I can, but not easily
14:32:32markunok, then I will just have to be patient.
14:33:33Bagderor force everone present to cvsup and rebuild
14:34:58LinusNmarkun: any fonts?
14:35:07Bger5x7 at least :)
14:35:27markunLinusN: I updated some of the fonts to their unicode versions and added unifont.bdf
14:36:01markunI will have to sort out the hebrew fonts later. Some of them came from unicode fonts.
14:36:34*Bagder tries a recorderv1 build
14:36:40Bgersilly (rtfm) q: what's the diff between P and M
14:37:03Bgerthat is, while doing cvs update
14:37:06Bagderls -l ajbrec.ajz
14:37:06Bagder-rw-r−−r−− 1 daniel daniel 127986 Dec 6 14:36 ajbrec.ajz
14:37:42BagderBger: M merges with your local changes
14:38:08XavierGrwell I posted on MR I hope that you will not revert the commit. ;D
14:38:17BgerP just updates ?
14:38:21BagderBger: yes
14:38:26Bger10x :)
14:38:31XavierGrC is bad
14:38:41XavierGrI got one today.
14:38:54LinusNmarkun: why codepages.h?
14:38:54Jungti1234markun: maybe 'Dotum' is unicode font.
14:39:33markunLinusN: hm, I don't know. phaedrus' work. I'll take a look
14:39:54linuxstbmarkun: The iPod and Archos Player both build fine.
14:39:55LinusNseems silly to have a huge array declaration in a .h file
14:40:19LinusNbetter make it a .c file and let the linker do the job
14:40:34 Join NicoFR [0] (
14:41:02LinusNor put the whole shebang in codepages.c
14:44:16Bagdererror: firmware image is 205754 bytes while max size is 204800!
14:45:15XavierGrI thought that you managed to split the image, no?
14:45:35Bagderthat's the uncompressed size
14:45:53Bagderit compresses it fine down to 127000
14:47:27markunBagder: it still shows the error?
14:47:55Bagderyes, I was just thinking we should change the output
14:48:04Bagdersince it now isn't really an error
14:48:08Bagdermore an information
14:51:25markunLinusN: I also have a big array in arabjoin.h should I change that into a .c file as well?
14:51:35LinusNplease do
14:51:47 Quit tvelocity ("Leaving")
14:52:19markunalso in the codecs there are big tables in .h files..
14:52:38markunWhat exatcly is the problem? Will it compile slower?
14:52:55LinusNyes, but we try to not modify the codecs, to be able to patch them
14:53:20LinusNmarkun: no, but it's generally a bad think to declare data or code in a .h file
14:53:45LinusNsince it will have side effects if you include it in two files
14:55:56NicoFRLinusN: "H300 bootloader USB mode" : does this mean there is a new bootloader to come ?
14:56:10LinusNyes, but it isn't ready
14:56:26NicoFRjust wanted to be sure
14:56:37NicoFRanyway thanks a lot for USB suuport :)
14:57:12XavierGrhehe even at USB Rockbox is better....
14:57:46LinusNit was a pain to develop before i got it working
14:58:11LinusNthe iriver firmware takes ages to scan the damn hard drive all the time
14:59:12NicoFRit's much better the rockbox way...
14:59:20LinusNi think so too
15:00:48NicoFRand do you already know what will be your next "step" ?
15:01:06Jungti1234Bye all
15:01:10LinusNthat will probably be the cpu frequency
15:01:12Jungti1234I go to bed
15:01:22LinusNnite Jungti1234
15:01:38NicoFRgood night
15:01:46Jungti1234Have a nice Day.
15:01:47 Quit Jungti1234 ("bye")
15:02:18preglowrockbox even does the scanning better
15:02:25NicoFRLinusN: have you found anything about what is making it unstable at more than 90 MHz ?
15:02:26preglowslasheri's disk scan is tons and tons faster that iriver's
15:02:33LinusNNicoFR: no
15:02:49LinusNpreglow: indeed
15:02:50 Join DJDD_ [0] (
15:02:59 Nick Lynx_awy is now known as Lynx_ (
15:03:28NicoFRwell thanks for answering my questions :)
15:07:36 Join tucoz [0] (
15:08:06tucozcongratulations markun. Massive commit :)
15:08:49markunyes, we've been waiting for a long time :)
15:09:11tucozhehe, I am impressed that you have kept maintaining the patch for so long
15:10:22***Saving seen data "./dancer.seen"
15:11:22lostlogicLinusN: any idea how to make USB or power connection not automagically boot iRiver firmware on the H340 (now that USB mode is going to work)?
15:11:43LinusNlostlogic: that will be in v4
15:12:02lostlogicwow, rock on.
15:12:39preglowred build
15:12:58preglowa nice one at that :-)
15:13:14Bagderregion IRAM is full
15:13:29preglowdoes bootloader even _USE_ iram?
15:13:34preglowi thought it only used it internally
15:13:41lostlogicpreglow: I've fixed several ugly bugs in my profiling stuff, but now I'm stuck on recursion breaking timing.
15:13:43LinusNit runs entirely in iram
15:13:54preglowyes, exactly
15:14:22preglowlostlogic: oh?
15:14:57*linuxstb wonders why the ipod bootloader still is OK
15:15:26tucozNow that the space issue is solved, what do you think of linuxstb's solution to have user configurable menus (.menu-files)? I have been using that for a while.
15:15:28lostlogicpreglow: yeah, in order to time individual functions right, I remember each function's parent as I descend... recursion causes breakage.
15:16:27markundamn, the bootloaders..
15:16:52preglowtucoz: i'd like something like it, but i don't think that particular solution is very elegant
15:16:55LinusNunicode.o takes a huge amount of ram
15:17:41tucozpreglow, I think it is quite elegant, but it does force the user to have a certain directory structure which is perhaps not that nice.
15:17:47preglowtucoz: exactly
15:17:59preglowplus, it breaks if you change language
15:18:06preglowyou need to rename everything
15:18:24tucozah, that was the problem. Right. Ok, I then I am with you on that one.
15:18:30Bagdermarkun: unicode.c needs some code policing
15:18:35Bagderlike 80 columns
15:18:50*Bagder runs off
15:19:38 Quit DJDD__ (Read error: 110 (Connection timed out))
15:21:36tucozNow only mod-playback and I am satisfied :)
15:22:26 Join ender` [0] (i=ychat@
15:27:24ep0ch|how did the user configurable menus work again? you have one file in a directory for each menu item? or one plain text file in a directory file that contains a list of all the menu items?
15:28:08LinusNwhat configurable menus?
15:29:08 Join hshah [0] (
15:29:15tucozep0ch|, the menu-files work as shortcuts to for instance "General settings". I.e. you place a "General" in your root, and when you select that file you are sent to "general settins"
15:29:24linuxstbLinusN: This patch (by me) -
15:29:32linuxstb(it's a few weeks old but will probably still apply)
15:29:56tucozep0ch|, or "Browse" or whatever.
15:29:58LinusNwhy gzipped?
15:30:06hshahnow when is that TiMiD going to commit my WPS?
15:30:30ep0ch|hmm so its not possible to rearrange the menu? i.e. they will always be alphabetical?
15:31:07LinusNmy browser doesn't unpack .gz files automatically, so i can't review the patch online
15:31:24linuxstbLinusN: I'll fix it for you...
15:31:37tucozep0ch|, unfortunately not :\
15:31:45linuxstbLinusN: Fixed.
15:32:20tucozBut, as preglow said. It breaks if language is changed.
15:33:03ep0ch|and as he said it would mess with my direcotory structure
15:33:19linuxstbI've never thought of this patch as the _only_ way to customise menus. Just one way. Personally, I like having some things integrated into my directory structure.
15:33:42LinusNand it strcmp's all the strings in the language array
15:34:07linuxstbLinusN: So? It doesn't need to be efficient.
15:34:21linuxstbAs long as it reacts immediately to the user's button press - which is does.
15:34:23LinusNjust feels...brutal :-)
15:34:35tucozep0ch|, no it will not mess with you directory structure. It is in a way more usefull if you organize your root like "Browse Music", "Browse Data", "General Settings" etc.
15:34:56linuxstbLinusN: Yes, but brutal is sometimes good :)
15:35:36ep0ch|i think one text file in '.rockbox/' called something like '' would be good idea. And when you hit previous in the file system root this file creates the root menu, i'm finding it difficult to explain
15:35:48tucozYou could have a "Settings"-folder, where you keep the files
15:36:17ep0ch|tucoz: i like my root directory holding all my artists
15:36:46LinusNi must say that it's quite a nice and simple approach
15:37:07XavierGrspeaking of it I must update Greek.lang and correct some mistakes....
15:37:21tucozep0ch|, then this solution is not good for you. But, if you look at the root of the player as not really the root, and instead keep you music root under "Browse Music", then you get the point.
15:37:32linuxstbI think I worked it out that the code impact was about 160 bytes, plus about 20 bytes for each supported menu item. (on Coldfire)
15:37:34ep0ch|i may try that
15:37:48ep0ch|but ordering of the menu is a problem
15:37:54tucozep0ch|, true
15:38:04preglowcould just include order in the filename
15:38:10ep0ch|1. 2. 3. etc
15:38:33preglowlanguage is a bigger issue
15:38:35ep0ch|ugly imho
15:39:25linuxstbpreglow: I agree. But how many users would use both this feature and frequently change languages?
15:40:20preglowwho would do that at all
15:40:24preglowit's not a big issue
15:40:42ep0ch|ok i'll give this patch a whirl
15:43:02 Quit NicoFR ()
15:47:32markunIs it possible to do multiple commits with different messsages in 1 go?
15:49:27 Join MrShlee [0] (
15:50:39linuxstbmarkun: I don't think so. But you could enter all your cvs commit commands on one line, separated by ";" - which will mean _almost_ simultaneous commits.
15:51:13 Quit webguest28 ("CGI:IRC")
15:51:21 Join webguest28 [0] (
15:52:12 Join Paul_The_Nerd [0] (
15:54:00MrShleeany idea when the next bleeding edge is going to finish compiling?
15:55:07MrShleeohh well. I'll wait and read somemore source
15:58:45MrShleeDelete folder doesnt work on H3xx. Doesnt detect the the accept button input
15:58:59preglowdo you press play or the middle button?
15:59:03preglowthe middle button is what's correct
15:59:16MrShleeplay acording to the LCD
15:59:21preglowyes, but do what i told you
15:59:24preglowand you'll see it works
15:59:27preglowthe string is wrong
15:59:32preglowand will be until we have langv2 going
15:59:53MrShleeok that worked.. Cheers
16:00:04preglowspent a couple of days wondering about that myself
16:02:01ep0ch|regarding the user configurable menu, how can i get the radio on the root?
16:02:11ep0ch|i don't think .fmr is associated with anything
16:02:29linuxstbep0ch|: Create a file called "FM" (unless I forgot it in my patch)
16:02:54preglowmarkun: what's up with the master file warnings in japanese.lang?
16:02:58linuxstbep0ch|: Yes, that patch doesn't include FM radio. Just a minute...
16:03:03ep0ch|:) cool
16:03:07preglowand shouldn't the filename be kanjii, btw? :)
16:03:24ep0ch|but shouldn't .fmr files be associated with the radio somehow?
16:04:10XavierGrabout fmr
16:04:17markunpreglow: I'll take a look at it. The filename could be in kanji yes :)
16:04:25XavierGrIn my previous patch fmr files could trigger the radio and load the fmr preset.
16:04:44ep0ch|ahhh thats what i was used to i think
16:04:56XavierGrIn my new patch (as I was told that this is not consistant) I removed this behaviour
16:05:41 Quit MrShlee ("Changed by popular Request - "I am the heart of your darkness, you stinking oompa-loompa morons!"")
16:05:41XavierGrI also changed the way that fmrs are handled. Now save will occur only on exit instead in every change of the list.
16:06:16XavierGrStill, I think that triggering the radio screen with an fmr file is quite cool and neat. IMHO
16:06:47 Join Moos [0] (
16:07:05MoosHello everyone !
16:07:45MoosNice day for Rockbox, Unicode it's amost here, hooray !
16:07:58markunhi Moos!
16:08:11Moos(thanks to markun & phaedrus)
16:08:19MoosSlaute markun :)
16:08:27Moosoops Salute
16:08:40Moosbegining to be dislexique??? :)
16:09:32markunHm, I'm messing things up with my latest commit :(
16:09:42pregloware ye now
16:10:11linuxstbep0ch|: (with FM Radio):
16:10:13markunwill the be a build for every commit?
16:10:59linuxstbmarkun: A build is started every 60 seconds, if needed. Each build includes all changes since the last build (IIUC)
16:11:33ep0ch|linuxstb: cheers, the configurable menu is pretty neat actually :)
16:12:14webguest28is it possible to write capital letters with the morse-code input?
16:12:16Paul_The_NerdIs there anything special I need to do for the unicode support to function properly? Like, the unicode.txt that had been linked to in here yesterday, it does not display correctly on rbx.
16:12:21tucozhmm, I get an error on when trying to build the h120 sim. on arabjoin
16:12:39markuntucoz: yes, I broke it
16:13:17preglowwebguest28: you could ask slasheri to add a toggle for caps
16:13:27preglowmakes we want to learn morse code, this
16:13:45 Part tucoz ("Leaving")
16:20:01markunlinuxstb: is the previous build aborted?
16:20:13linuxstbmarkun: No.
16:20:32markunPaul_The_Nerd: I just added unifont.bdf. It should display all the strings in unicode.txt
16:20:34linuxstbA new build only starts when the previous one finishes.
16:20:35 Join actionshrimp [0] (
16:20:42markunbut the viewer needs some work
16:22:51ep0ch|linuxstb: i also think that "recording" should be included in your menu patch (no rush though)
16:22:53 Join Kohlrabi [0] (
16:23:35 Join edx__ [0] (
16:25:25Mongey|woah saw2 is amazing
16:27:36 Join BBub [0] (
16:29:15preglowunless you're speaking about the aphex twin album, i have no idea what that is
16:30:32preglowhah, tarantino
16:30:49ep0ch|i haven't seen saw or saw2, do you recommend i should watch saw before i try saw 2?
16:31:12Mongey|saw 1 first
16:31:19ep0ch|ah ok
16:31:39Mongey|saw2 ties in at the with saw1
16:32:16Mongey| ^end
16:35:47 Quit Mongey| ()
16:40:10linuxstbep0ch|: Which recording option(s) would you use? There are three possiblities - "Recording", "Recording screen" and "Recording Settings"
16:40:29markunLinusN: the H100 bootloader still doesn't compile
16:40:32 Quit edx (Read error: 110 (Connection timed out))
16:41:18ep0ch|hmm, thats a tough one
16:42:03ep0ch|i would say Recording
16:42:12 Join webguest84 [0] (
16:43:09 Quit thegeek (Read error: 104 (Connection reset by peer))
16:43:09 Part webguest84
16:43:32ep0ch|oh someone should capatilize that 's' in "Recording screen'"
16:44:41BBubyou only capitalize names
16:44:52ep0ch|because it isnt consistent
16:47:49 Join webguest80 [0] (
16:48:49webguest80hmm, I cant manually scan for radio stations (H140), is there a trick to this ?
16:49:07 Join thegeek [0] (
16:49:36webguest80I enter FM mode, then laft press joiystrick to scan and nothing happens
16:49:56BBubare you in scanning mode?
16:50:12BBubif not, hold play >1 second
16:50:33webguest80I'm not refering to auto scan, but Manual scan
16:50:43preglowdoes it say 'mode: scan'
16:51:09webguest80it says mode preset
16:51:18preglowthen press and hold play
16:51:22preglowthen try again
16:51:25BBublike i said :P
16:51:43webguest80the morse code key thingy comes up
16:52:27BBubnot on my iriver
16:52:50preglowthen you're doing something wrwong
16:52:54preglowenter radio screen
16:52:58preglowpress play for a second
16:53:02preglownot a/b
16:53:07preglowthen it should switch to mode: scan
16:53:41BBuband play is in this case not the joystick-button, but the one on the left side of the player
16:53:49webguest80I have done exactly that and it says, "no presets"
16:53:58preglowyou're not pressing play
16:54:05preglowyou're pressing select
16:54:08preglowyou need to press play
16:54:20webguest80and it has know lockedup, yippee
16:54:23ep0ch|the one above stop
16:54:30preglowno it has not, press stop
16:54:44ep0ch|it has a |>|| on it
16:55:15preglowSlasheri: morse mode is damn cool :PP
16:55:32ep0ch|how do i get into morse code input?
16:55:46preglowpress a/b when in virtual keyboard mode
16:55:57ep0ch|oh wicked :)
16:56:14webguest80sorry but I cant get scan mode, it stays in preset mode, *sigh*
16:56:23preglowwebguest80: ok
16:56:27preglowwebguest80: top right button
16:56:30preglowwith the play symbol
16:56:40preglowyou keep that pressed until it says 'mode: scan'
16:56:52preglowshould just need a half second or something
16:57:05ep0ch|good job the morse code screen has a lookup table on screen :)
16:57:28preglowep0ch|: yeah, that's a real life saverf
16:57:30Moosep0ch: very fine for who want to learn it :)
16:57:38linuxstbYes, I was about to ask for that, but then saw it was already there...
16:57:46webguest80ok, now I have it, I was too quick with the play button, thanks gents
16:57:46ep0ch|yes gives me something to do on the train :)
16:57:56Mooshehe :)
16:58:06Mooslot of plugins too ;)
16:58:23ep0ch|would be even better if another button could be used other than joystick, pressing the joystick is tiring ;)
16:58:57Moosthe irivers joystick isn't very good pitty
16:59:20webguest80the trick is to press and hols the Play buttin for >=.5sec
16:59:43BBub[16:50:18] <BBub> if not, hold play >1 second <- ;)
17:00:10ep0ch|wow that came from the future dude
17:00:19BBubfor you maybe :)
17:00:57webguest80so are the shares of microsoft going up, seeing as you have an insight into the future
17:01:21ep0ch|down with 'em i say
17:02:09 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:02:14 Join einhirn [0] (
17:02:33webguest80I didn't think that the play button had two functions, pause and mode select, stopped looking after I saw it did pause
17:04:02ep0ch|hmm a sortof problem with the user menus is you must have the display filetype to "supported".
17:04:50webguest80and the side play button should perhaps be referred to the "pause"button, technically h140 has two play buttons, just my .02c
17:04:53ep0ch|i changed it by mistake and wondered where all the menus went
17:06:52ep0ch|i think "play" should be known as "play" and the joystick button known as "select", but i don't have 0.2c.
17:07:38webguest80yes but jotstick has 5 different select positions
17:07:57webguest80where press = play
17:08:16ep0ch|eh? it has one select button and 4 directions :p
17:09:25ep0ch|time for tea
17:10:15 Part webguest80
17:10:23***Saving seen data "./dancer.seen"
17:16:24 Nick edx__ is now known as edx (
17:20:13 Join bobwise [0] (
17:22:47 Join webguest87 [0] (
17:23:17webguest87hi there guys! just want to say how amazing rockbox is and how much i appreciate your volunatry work :D
17:24:21chopped_pork:DDDDD unicode
17:25:06webguest87Oh, and i was just wondering, I have just aquired an iriver which was completly water damaged (everything corroded) but amazing the HD worked, and bdecause i had a spare mainboard i was able to get this working (still not sure on teh abttery) ANYWAY! what is the LCD screen for the iriver h100 series? model number would be great !!!!!!!!!!!!!!!!
17:25:23preglowi don't think it's that easy
17:25:23webguest87i REALLY need this :(
17:25:33ep0ch|forgive me for being so ignorant, but what is so good about unicode?
17:25:49preglowep0ch|: 1. you can forget about code pages, 2. it supports all languages
17:26:19Paul_The_NerdOkay, I was bored, so now there's a .zip of the FMPresets page with each city in its own FMR, organized under a friendly directory structure.
17:27:13webguest87hello? sorry to be annoying but i reall have searched the web for hours trying to find replacement LCDs for my H120, i just thougt that with you guys writing drivers for the hting you'd at least know what LCD it was?
17:27:14markunep0ch|: you can use the same fonts for different language files
17:28:11ep0ch|okayyy so you need unicode compatible fonts?
17:28:31linuxstbwebguest87: Have you looked at the Rockbox Wiki?
17:28:37ep0ch|tell you what i'll wikipedia it :)
17:29:47 Quit bobwise ("CGI:IRC (Ping timeout)")
17:30:37 Join bobwise [0] (
17:31:13webguest87thankyou very much, didnt see that lol!
17:31:30preglowi don't think it'll be easy to get
17:31:49linuxstbI'm curious, does anyone _not_ use the dircache?
17:31:56preglowamiconn :>
17:32:03linuxstbObviously :)
17:32:13preglowbut no, i can't live without it
17:32:19*amiconn doesn't like the dircache
17:32:43linuxstbI'm just thinking it should be enabled by default, now that it's proved itself. (assuming it has)
17:32:44amiconnI'd say 90% of the time the dirscan is done for nothing
17:33:12amiconn...since I moften just connect the player to my car stereo, power on & resume
17:33:35markunI like dircache
17:33:41preglowi believe i actually save battery using it
17:33:45preglowi browse pretty much
17:34:00Bgeri like it too
17:34:07amiconnApart from that, only 1 of my 5 rockboxes has dircache
17:34:14markunI have all my music on my iriver and browse a lot when I try to decide what I want to listen to
17:34:18amiconn...and it's still one of the lesser used
17:34:29preglowwasn't someone going to test dircache on an archos?
17:34:31*amiconn still uses recorder v1 most of the time
17:34:49amiconnYes, me, but that's definitely for testing only
17:35:00preglowyeah, but did it work?
17:35:07amiconnI did not yet try it
17:35:31amiconnIt might be of some use for those who applied the 8 MB mod
17:35:42preglowyou can't fit more than 8 megs?
17:35:52preglowi still think the default two megs sounds absurdly small
17:36:13 Quit webguest87 ("CGI:IRC")
17:36:16amiconnThe CPU supports a maximum of 16MB, but afaik there are no suitable chips
17:36:57amiconnHey, the archos player was designed around (in?) the year 2000
17:37:06preglowi know
17:37:08amiconn..and the recorder shortly after that
17:37:11preglowi still think it sounds small for its time
17:37:28 Join webguest08 [0] (
17:37:52preglowbut it's of no consequence
17:37:55webguest08hi is linus around?
17:37:56markunAre there any DAPs with more than 32MB now?
17:37:57preglowdoubt i'll ever get one anyway
17:38:04linuxstbStrangely, the 60GB version of the new iPod has 64MB of RAM, but the 30GB version only has 32MB...
17:38:08preglowmarkun: yeah, i believe a g5 ipod has 64meg
17:38:32webguest08i finally did what I said i would (get an igp100) just need somewhere to send it
17:38:56 Quit bobwise ("CGI:IRC (EOF)")
17:40:40ep0ch|is that the round one?
17:41:01webguest08a while ago i posted a big file of photos
17:41:07webguest08taken of my own igp taken apart
17:41:17webguest08ive finally gotten an extra one
17:41:32 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
17:41:36ep0ch|which country are they available in?
17:41:43ep0ch|or were
17:41:44webguest08just need to collect it from postal depot and then i can send it to linus or someone who can look inside
17:41:53 Join bobwise [0] (
17:42:22webguest08well i think they were available from everywhere but phased out eventually in march 94 or so?
17:42:32amiconnmarkun: Would it be possible to have a secondary .lang file for japanese? japanese-katakana.lang ...
17:43:30Bgerwebguest08?? march 1994 ?
17:43:56ep0ch|ahead of its time then :)
17:44:27webguest08lol probably
17:44:59webguest08i thought was same as h1x0's except a smaller version but apparently it was the granddaddy coming even before those
17:46:07amiconnLinusN: Something is still wrong with the h100 bootloader. h120 is okay
17:46:50webguest08 at the bottom is the pics i uploaded a while ago of igp
17:49:06preglowfebs has done a lengthy port on the mailing list
17:49:10preglowgotta say i pretty much agree with him
17:50:26webguest08last time i spoke on irc about this, i asked if igp was possible considering that rockbox still supports those little archos flash players and ami said should be easy because technology is basically same as h1x0. has anything changed this?
17:51:10preglowwebguest08: not much, no
17:51:18preglowjust need to find someone to do it
17:51:24preglowi think linus has his hands full
17:51:51markunamiconn: for the player? I don't know enough about japanese to decide if it's useful
17:52:59webguest08k well i dont mind. i think he mentioned a problem in that no firmware was available but that going through some port or something (same as how he messed around with h1x0 originally) he would be able to get it
17:53:45webguest08im just happy to send it. even if nothing is done, it will be worth a look hardware-wise probably
17:54:00webguest08and in any case what am i gonna do with 2 igps :)
17:54:52preglowyou write all japanese with katakana, i think
17:55:27linuxstbwebguest08: Do you know the basic details about the hardware, such as which CPU it uses? Also, is the firmware upgradeable in the same way as the H1xx/H3xx?
17:55:44webguest08well coldfire same as h1xx
17:56:06webguest08firmware upgrade option is in there but no extra firmware upgrades were released to my knowledge
17:56:51webguest08or iriver america may of been doing one but igp was phased out before it got finished or they withheld it knowing it was being phased out
17:57:25webguest08screen is identical to h1x0 and so is the gui
18:05:54 Nick Lynx_ is now known as Lynx_awy (
18:08:27 Quit bobwise ("CGI:IRC (EOF)")
18:15:35 Join webguest54 [0] (
18:16:05ep0ch|preglow: you know the crossfeed filter constants, what format are the hex values in? i want to try the suggested values but i can't figure out how to convert them.
18:18:19preglowep0ch|: they're in signed 0.31 fixed point
18:18:24ep0ch|would 1.0 = 0x80000000L?
18:18:33preglow-1 = 0x80000000
18:18:44preglow1.0 = 0x7fffffff
18:19:16 Join wubbla_ [0] (
18:19:42BBub <- ;(
18:19:44ep0ch|ahhh i get it :) ty
18:19:54BBubanyone got an idea where to get a replacement part?
18:20:55preglowep0ch|: that is, 0x7fffffff isn't 1.0, but 0.999998whatever
18:21:23BgerBBub: i suggest u to sell it on ebay and buy another one from there
18:21:23ep0ch|yes i just completely forgot about the signed bit :)
18:21:48BBubBger: u think anyone would pay enought for it? ;)
18:22:00Bgerwho knows ...
18:22:29BBubmaybe i could ask iriver, its only 1 month over waranty
18:22:54Bgerthey usually don't repair after waranty (iirc) :(
18:24:18Bgeranyway, u won't be charged for trying
18:24:44markunpreglow: isn't 0x80000000L = -0.0 ?
18:25:36markunI thought 0xffffffffL = -0.99999..
18:25:36 Quit webguest08 ("CGI:IRC (EOF)")
18:29:16 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
18:31:42 Quit wubbla (Read error: 110 (Connection timed out))
18:32:14preglowmarkun: no
18:32:23preglowmarkun: well, the last is true, but not the first
18:32:56preglowmarkun: if what you said was true, then 0x8000 would not be -32768, which it is
18:33:12 Quit webguest54 ("CGI:IRC (EOF)")
18:33:57ep0ch|to change the delay i need to modify the size of the delay array and index?
18:34:08 Quit newnick ("CGI:IRC")
18:34:12preglowep0ch|: correct
18:34:42preglowmarkun: nothing of what you say is true, btw :)
18:34:52preglowmarkun: 0xffffffff = -0.00000001
18:35:01ep0ch|no wonder there is attenuation ;)
18:35:47preglowmarkun: you've got to remember the numbers after the sign bit are in twos complement format
18:36:00ep0ch|ok i'll change the size of the array to 15 from 13
18:37:07ep0ch|asm question for you preglow, #13? that is a constant?
18:39:12 Join webguest25 [0] (
18:40:14preglowahh, you'd better hack the c version
18:40:17preglowand yes, it is
18:40:22ep0ch|hacked both
18:40:32preglowbut the c version is the reference
18:40:47ep0ch|dont worry i don't plan on commiting as i can't :p
18:45:52ep0ch|ok there is no attenuation now, but it sounds like the output is 6 bits instead of 16 bits :(
18:46:22ep0ch|ok slight increase in attenuation i think
18:47:40ep0ch|my values are probably wrong
18:50:31 Quit chopped_pork ("leaving")
18:55:34 Join hshah_ [0] (
18:55:35 Quit hshah (Read error: 104 (Connection reset by peer))
18:59:52 Quit t0mas ("brb")
19:04:27kkurbjunamiconn: so would it make sense to try and eliminate the lcd_mono_bitmap calls, using only lcd_bitmap and in the process have the bmp loader just handle storing everything into a player native format?
19:04:37webguest25WoW! UnicodE! way the go guys... (I'm updating and building for the last half an houre, is it just me or cvs update is pretty darn long?)
19:07:12 Nick webguest25 is now known as lamed (
19:10:26***Saving seen data "./dancer.seen"
19:10:35 Join webguest54 [0] (
19:14:05Bgerlamed: many changes, big fonts...
19:14:06 Part webguest54
19:14:44 Join webguest75 [0] (
19:15:21 Join hshah__ [0] (
19:15:22 Quit hshah_ (Read error: 104 (Connection reset by peer))
19:18:47 Quit webguest75 (Client Quit)
19:20:56 Join webguest43 [0] (
19:21:06 Join dpassen1 [0] (
19:22:10webguest43Hi everyone and thank you for your great effort on supporting the iRiver player!
19:22:12 Join Lear [0] (
19:22:48webguest43just wondering if anyone know if there is anyplace to buy H-120 or H-140 players ?
19:24:00lamedbger: yeah, but i got ah... wait
19:24:28lamedcvs [update aborted]: error writing to server: Connection reset by peer
19:25:15lamedwhat the hell happend? (it was later using a script to update all the target builds so i ended up waiting for nothing)
19:26:46lamedok. now i had it updating.
19:28:36 Join webguest17 [0] (
19:29:38 Quit webguest17 (Client Quit)
19:30:09 Join _FireFly_ [0] (
19:33:46 Join webguest71 [0] (
19:34:17webguest71webguest43, there's aH140 for sale on ebay right now
19:37:07 Part webguest71
19:39:23 Part Polo_o
19:44:11markunincredible that people start complaining about the lack of korean so short after unicode was added.
19:44:42markunYou would think they would be thankful.. but no.
19:46:24_FireFly_markun: that is a common problem :)
19:46:37markunpisses me off
19:47:01markunbut well, a lot of people told me they love me, so that's nice :)
19:47:27_FireFly_i just look if this breaks my other patches ;)
19:49:06_FireFly_markun: currently only two files have conflicts :)
19:49:22markunwhat kind of patches do you maintain?
19:50:10_FireFly_markun: additional remote-settings, a new wps-file-loader and combined-bitmap-support
19:50:25 Join DangerousDan [0] (
19:50:59 Join tvelocity [0] (
19:51:12 Quit BBub (Read error: 110 (Connection timed out))
19:51:34Bgermarkun: i love you :D
19:51:38_FireFly_markun did you noticed that the text of the bookmark on stop( if set to ask) is printed twice on the main screen of the h1xx ??
19:51:56markunno, I didn't
19:52:44_FireFly_at least the description which button to press
19:52:49_FireFly_is pronted twice
19:52:57 Join Maxime` [0] (
19:54:03Learmarkun: code police hint: remove "int w, h;" in firmware/mpeg.c. :)
19:54:18LearBtw, why preload just those fields? Looks a bit odd, IMHO...
19:57:23markunI don't know exactly why only those fields. Everything that will be on the screen should be preloaded.
19:58:37 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
20:01:13LearSo... WPS files don't need to be in UTF-8?
20:01:45markunThey do
20:02:17 Join webguest70 [0] (
20:02:17markunalthough I didn't test WPS' with UTF-8 in them
20:02:45 Join dukezed [0] (
20:02:47 Quit webguest70 (Client Quit)
20:02:49LearBut the ones I looked at only contained ASCII, so the editor didn't mark the files as UTF-8. :)
20:03:52 Quit dukezed (Client Quit)
20:03:55markunIt doesn't matter. I don't even know what happens if the file starts with a UTF-8 mark..
20:05:59LearConvbdf dumps here too...
20:06:03 Join muesli_- [0] (
20:07:35markunhi muesli
20:07:53muesli_-hi markun
20:08:00markunconvbdf has a problem with the starmap file, or also others?
20:09:54 Quit Maxime (Read error: 110 (Connection timed out))
20:17:26_FireFly_markun a wps-file with an utf-8 start mark loads with only a minor problem
20:17:40markunit displays a little box?
20:17:44_FireFly_it seams that my new wps-loader needs to ignore this flag
20:17:49_FireFly_no the display is right
20:17:58_FireFly_as far as i can see in the sim
20:18:26_FireFly_only the first line of my wps didn't gets parsed correctly
20:18:30preglowi'm out of here, later all
20:18:32 Join ^Guest37784 [0] (
20:18:36_FireFly_preglow: cu
20:19:15 Quit ^Guest37784 (Client Quit)
20:19:36Bagdermarkun: very good job on the unicode stuff
20:19:53markunthanks, also phaedrus961 did a lot
20:19:55BagderI made it front page news
20:20:11ep0ch|any idea if it would be possible to shutdown rockbox when in usb mode?
20:21:49markunI'm adding a menu to the viewer and removing most of the shortkeys, does anyone object?
20:21:53_FireFly_markun: a 0xef was added to the start of the file when saving it as utf-8 :)
20:22:06_FireFly_me not
20:22:26ep0ch|preglow: still here?
20:23:35ep0ch|preglow: for the log. i can't seem to change your asm crossfeed to use a 15 sample delay. the C one seems fine.
20:24:20ep0ch|(well i can change it, but it sounds rough)
20:25:40 Join Gibbed [0] (
20:25:41 Quit Rick (Nick collision from services.)
20:25:49 Nick Gibbed is now known as Rick (
20:31:54 Quit tvelocity ("Leaving")
20:34:28 Join tim66 [0] (n=tim@
20:37:01 Join DrMoos [0] (
20:37:01 Quit Moos (Read error: 104 (Connection reset by peer))
20:37:47 Quit lamed ("CGI:IRC")
20:38:18 Nick DrMoos is now known as Moos (
20:39:30 Join linuxstb_ [0] (
20:40:29 Join JonSenior [0] (
20:41:01_FireFly_ok the new wps-file-loader handles now correctly the utf-8 mark (BOM) :)
20:43:36muesli_-l8er...schalke will win...
20:44:11_FireFly_it doesn't bother me ;)
20:46:04_FireFly_but only the BOM is currently ignored
20:48:41 Quit webguest43 ("CGI:IRC (EOF)")
20:53:50 Quit linuxstb (Read error: 113 (No route to host))
21:02:30 Quit muesli_- (Read error: 110 (Connection timed out))
21:03:18 Join solexx [0] (
21:07:44^BeN^mmm what is preferable in the h300 iriver firmware, usb hub or desktop connection mod?
21:08:31 Nick linuxstb_ is now known as linuxstb (
21:10:28***Saving seen data "./dancer.seen"
21:20:16markunwho has experience with menus in plugins?
21:20:36 Quit solexx___ (Success)
21:22:30 Join |Bger| [0] (n=Bager@
21:23:42 Join Lear_ [0] (
21:25:00linuxstbmarkun: I added the menus to the plugin interface for Sudoku
21:26:56 Quit `3nergy (SendQ exceeded)
21:32:28 Quit Lear (Read error: 110 (Connection timed out))
21:32:59 Quit Shaikoten (Read error: 110 (Connection timed out))
21:33:37Lear_Hm... Is the .rockbox/codepages folder supposed to be empty?
21:33:56 Join Drumr|away [0] (
21:34:27 Nick Lear_ is now known as Lear (
21:34:57 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
21:34:59markunLear_: no, it shouldn't
21:36:04_FireFly_but it's empty
21:36:08Drumr|awayim getting a 'filesystem error' on my rockbox sim
21:36:19Drumr|awaywhat can i do to fix this?
21:36:40LearMaybe a victim of the convbdf crash? (Just guessing...)
21:37:47_FireFly_Lear: no i think the problem is that er is no *.cp files in the sources
21:37:51LearHm.. No tools/codepages either... Maybe a full checkout is needed...
21:38:00LearSomething like that, yes...
21:38:27markunThere is also no codepages in the bleeding edge build..
21:38:59LearBleh, last I checked, there were no bloody codecs in the bleeding edge... :P
21:39:17_FireFly_markun where in the cvs-tree should the codepages be
21:39:49markunIt should be generated when you do 'make zip' from /tools/codepages.c
21:40:05LearI suspect a lack of ../tools/configure here... :)
21:40:12_FireFly_i had reconfigured
21:40:26_FireFly_but a empty codepage-folder
21:40:36markunI have to go now. Will take a look at it tonight.
21:41:03LearBut it looks like tools/codepages.c isn't buildt (I'm guessing it should be...)
21:42:13_FireFly_yepp it isn*t build and trys to call it
21:42:22 Quit webguest28 ("CGI:IRC")
21:42:29 Join webguest28 [0] (
21:42:41LearSo tools/configure needs updating then...
21:43:49_FireFly_i think so because the makfile in the build-dir as no codepage-rule
21:44:36 Nick hshah__ is now known as hshah (
21:45:04_FireFly_yepp it needs an update
21:45:19_FireFly_i think codepage should be made for all targets
21:48:20LearYep, that worked better.
21:48:34LearNow I have five files there at least.
21:52:35SlasheriHmm, tagcache loading seems to work now. Initial build takes ~7 minutes and loading from disk into ram ~10s and 700 KiB extra ram
21:52:42Slasherinight :) _>
21:53:03Slasheriand all loading is of course transparent
21:54:06linuxstbSlasheri: Have you tried building the same database with the Java or Perl tool on your PC?
21:54:38Slasherilinuxstb: yes, it takes about same time or longer (usb 2)
21:55:04Slasheri(i tried the tool, not exactly same database but almost)
21:55:07Slashericu ->
21:55:30LearDoes that mean yet another thread? :)
22:01:36 Quit Kohlrabi (Nick collision from services.)
22:01:42 Join Kohlriba [0] (
22:03:06 Join muesli_- [0] (
22:04:40 Quit hshah ("Leaving")
22:12:07 Join Jungti1234 [0] (n=jungti12@
22:15:48 Quit ingenius ("Leaving")
22:25:59 Nick Lost-ash is now known as ashridah (n=ashridah@
22:29:05 Join San [0] (
22:29:18Sanhow do I compile a plugin for rockbox?
22:29:31Sando I put it in the rockbox folder and compile the whole thing?
22:30:20ashridahit's a little more complicated than that. You'll need to put it in an appropriate directory, and at least modify SOURCES
22:30:26ashridahpossibly some other its as well
22:30:59*San will have a look
22:31:56ashridahbut yeah, most plugins go in the apps/plugins/ directory
22:32:27Sani know that
22:32:39ashridahwell, you said 'rockbox folder'
22:32:43ashridahthat's a little vague
22:32:53 Join leftright [0] (
22:32:57Sanwell, i was typing fast
22:33:05Sani am trying to make pacman for RB
22:33:24amiconnThe unicode rush didn't cause any cvs conflict with my volume patch :)
22:33:53leftrightSan ; the first sticky in the forum is "The simplified guide to compiling rocknox"
22:34:12Sanok, i'll have a look
22:34:16Santhanks guys
22:35:43leftrightare you going to commit your volume patch amiconn ?
22:36:00amiconnI'm a bit undecided...
22:36:14amiconnIt seems to have support from most other devs
22:36:22Moosamiconn: please do :)
22:36:35amiconnI would also like to commit it before my language file cleanup
22:36:56_FireFly_and the missing remote-settings what about these ??
22:37:07amiconnI should at least write a reply in my own thread
22:37:17amiconn...but I can't decide which mail to reply to
22:37:20leftrightwell lenox's options does have a off feature, which I understand is similar to your mod, why nit refine his patch
22:38:23Moosamiconn: your first one?
22:38:33amiconnThe main points are that (1) it has too many options. Scaling back volume doesn't make much sense, and scaling back the non-current adds even more confusion
22:38:51amiconnJust off and scaling back bass/treble might make sense
22:38:54leftright?me puts hands in air and gives up
22:39:13amiconn(2) It does it the wrong way, behind the scenes
22:39:41amiconnThe good thing with switching to dB for volume is that a prevent clipping option is possible without much knowledge about the dac
22:40:02*leftright nods and agrees with everything amiconn says
22:40:11Mooshere too :)
22:40:27amiconnSo 'prevent clipping' can be implemented at the app level, uniform for all platforms and with proper indication
22:41:09_FireFly_iirc if the volume is >0db then clipping is possible
22:41:44amiconnClipping is possible if current_volume + MAX(bass_boost, treble_boost) > max_volume
22:41:59amiconnSince all values will be in dB, this is simple...
22:42:07Moosyeah indeed
22:42:07_FireFly_ah ok
22:42:22 Quit Rick (Read error: 104 (Connection reset by peer))
22:42:43amiconnmax_volume is 0 dB for iriver, +12 dB for archos recorders/ondios, and +18 dB for archos player
22:43:16amiconn0 dB is just equal to the 0 dB definition of the DAC manufacturer, and should resemble standard line level
22:43:20 Join NicoFR [0] (
22:43:23_FireFly_then was my recall a bit true but only for iriver :)
22:44:00NicoFRhi everyone
22:44:27 Quit ep0ch| ("gone")
22:44:29NicoFRun franšais :)
22:44:37leftrightmethinks the masses will be very confused by the db display, "why does the volime increase if it goes towards zero"? type questions
22:44:54MoosNicoFR: yes I'm (parisian one) ;)
22:45:01NicoFRmoi aussi
22:45:25NicoFRI was wondering wether rockbox "knows" when the battery is full on the H300...
22:45:34_FireFly_afaik there will be a conversion fn to convert db-values to percent-values
22:45:54leftrightas then the masses will be appeased for aure
22:45:54dpassen1as a non-developer but normal user, i would love a dB scale
22:45:59NicoFRsometimes when I switch back to iriver, it tells me the battery is full so I was wondering if rockbox stops charging
22:46:06 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
22:46:17NicoFRbecause it nevers seems to
22:46:17BagderNicoFR: rockbox doesn't control charging on the iriver models
22:46:19ashridahNicoFR: iirc, the iriver platforms haven't implemented the charging-stop notification. the hardware itself knows when to stop tho
22:46:20amiconnwhoa! 'make zip' result almost doubled in size (recorder)
22:46:51_FireFly_there a many wpses afaik
22:46:54 Join DangerousDan [0] (
22:47:01leftrightheh, great zipper that is
22:47:03NicoFRok so there should be no risk in leaving it to charge then ?
22:47:07amiconn_FireFly_: None for recorder
22:47:30BagderNicoFR: none added by Rockbox at least
22:47:32SanAnyone interested in porting DOOM for the H300? I have the source if you want it
22:47:39 Quit DangerousDan (Client Quit)
22:47:40_FireFly_amiconn: oops
22:47:54NicoFRBagder: thanks
22:48:04amiconnBig unicode fonts plus the codepages
22:48:19 Join DangerousDan [0] (
22:48:43 Quit NicoFR ()
22:49:10tim66a dB based volume control would be great to parallel the gain settings for record. at least for me
22:49:16amiconnThe whole self-extractor is <1200 bytes
22:49:36ashridahamiconn: so how much smaller is the firmware with the compression now?
22:49:37Moosamiconn: everyone seems to be with you for your dB scalling :)
22:49:47Bagderamiconn: I really like that decompressor work of yours, very nice!
22:49:49ashridah(on the apropriate platform, of course)
22:49:50tim66I often use the iriver as a pass-thru device, so being able to adjust the gain and 'know' how much to adjust the volume by would be great
22:49:51amiconn(including the almost-empty 512 byte vector table)
22:50:04Bagderashridah: 204K => 127KB
22:50:18 Nick ashridah is now known as Lost-ash (n=ashridah@
22:50:23amiconnashridah: Compression ratio is very close to gzip
22:50:52amiconn...but this .ucl decompressor is small and fast, just ~260 bytes of code
22:51:23amiconnThis is a bit of cheating, because uclpack contains several algorithms, but our decompressor only supports one
22:51:43amiconn(the one that we found to work best with SH binaries)
22:51:50_FireFly_amiconn: why to support then more then needed :)
22:52:35amiconnBagder: Did you try running a compressed .ajz on target?
22:52:56amiconnCould be interesting to test on an unflashed target
22:53:03Bagderhaven't even started my archos in months
22:53:34amiconnThe archos descrambler is rather slow, maybe loading a compressed .ajz plus extraction is faster than loading an uncompressed .ajz ...
22:53:52 Join Rick [0] (
22:55:06*linuxstb almost deafens himself by turning off clipping prevention with volume and bass both at 100%
22:57:50amiconnBagder: Regarding 'make' in tools, there are some quirks. (1) Would it be possible to silence the output when there's nothing to do? Or is there a reason to display the output?
22:58:07amiconn(2) codepages.c isn't built automatically
22:58:46_FireFly_amiconn: yepp configure wasn't updated to compile codepages
22:58:52Bagdertools to build must be added to configure
22:58:56Bagderthe toolset variable
22:59:11Bagderand the (1) is simply because that is the default output
22:59:23_FireFly_and it seams markun had forgotten to add it to this var
22:59:37Bagderyes, the tools change is rather recent
22:59:42BagderI can fix it
22:59:57Bagderdoes the player build need codepages too?
23:00:23amiconnHmm, perhaps not
23:00:24_FireFly_afaik all targets need codepages
23:00:28dpassen1amiconn: are you hosting the dB scale patched build anywhere?
23:00:37amiconnThe buitlin codepages might be sufficient
23:00:48Bagderok, so I don't do it for the player for now
23:00:54amiconn_FireFly_: He wants a patched build
23:01:09Leargood point, it can't really show more then latin1 anyway, and that's a hack. :)
23:01:18_FireFly_oops again
23:01:28amiconnLear: It can and will...
23:01:43amiconnI plan adding cyrillic and greek, and katakana for the new lcd
23:01:52LearNew LCD??
23:02:01amiconnThere are two lcd types
23:02:16amiconnThe newer type has katakana built-in
23:02:16LearAh, you mean the 8 "soft-char" ones...
23:02:46LearBut even greek, at eight (or four) chars at a time... Is it really useful?
23:03:45BagderLear: remember that amiconn is slightly insane ;-)
23:04:07amiconnLear: The thing is that quite some greek and cyrillic characters look the same (or almost the same) as latin characters
23:04:47amiconnI 'just' need to extend the dynamic remapping which is already there to support almost full latin1
23:04:56LearHm... Maybe they do... Haven't really looked closely at that.
23:05:06Bagderfixed configure
23:06:34LearIf you feel like some more code policing, there's a small thing in mpeg.c... C98 variable declaration. :)
23:07:21amiconnNot only in mpeg.c ...
23:07:39Bagderyou mean mixed with code?
23:07:40linuxstbLear: I've looked at that bit of code you mentioned, and I think the w and h vars can just be removed completely, and NULL passed to the lcd_getstringsize() functions.
23:07:51LearExactly, done like that in playback.c.
23:08:42amiconndpassen1: Building it now... H120/140 ?
23:09:08Bagderbookmark.c:212: warning: ISO C90 forbids mixed declarations and code
23:09:12amiconnThey're the same
23:09:13Bagderlanguage.c:57: warning: ISO C90 forbids mixed declarations and code
23:09:20amiconnBagder: sound.c ... I know
23:09:41amiconnIt was like that before
23:09:59amiconndpassen1: Only special thing would have been h100
23:10:03 Quit San (Read error: 110 (Connection timed out))
23:10:23dpassen1yeah, understood
23:10:31***Saving seen data "./dancer.seen"
23:10:43Jungti1234help me
23:11:12Bagderthe question is, should we remain C89 compatible or not?
23:11:43CassandraSo this new unicode support - does convbdf read unicode values from the bdf fonts?
23:12:02CassandraBagder, can't think of a good reason why we should.
23:12:05amiconnBagder: It looks more clean most of the time, but with a lot of ifdefs it might be different
23:12:16CassandraDoesn't really gain us anything in terms of portability, does it?
23:12:30 Quit muesli_- (Read error: 104 (Connection reset by peer))
23:12:32Bagderamiconn: ifdefs should not be required for it at all
23:12:42BagderCassandra: only for possibly new targets/compilers
23:13:13amiconnI mean, if there are a lot of ifdefs, it might help readability to allow C99-style declarations
23:13:17Jungti1234Why does error happen?
23:13:26BagderI'm with you now, misunderstood
23:13:27t0masfeature proposal: sleep mode
23:13:29amiconnThat's the reason for them in sound.c afaics
23:13:29CassandraBagder, well techically I think making it compile without warnings under gcc does more for practical portability.
23:13:41t0masa mode to make it shutdown after the current song
23:13:44 Join linuxstb_ [0] (
23:13:55t0masor make it beep after the song, wait 20 secs for a button... and then shutdown
23:13:59Bagderamiconn: that's very often easily fixed with just a pair of { and } in the right place
23:14:00t0masideal for podcast listening etc
23:14:04Cassandrat0mas, it's already there. It's called "idle timer".
23:14:18t0masyeah, but the beep idea might be usefull...
23:14:42t0masand with "skip to next dir" on it never gets idle
23:14:49amiconnt0mas: Just enable the voice UI for file names, will have almost the same effect
23:15:04 Quit linuxstb (Nick collision from services.)
23:15:08 Nick linuxstb_ is now known as linuxstb (
23:15:11CassandraYou see, this is why we should never have bowed to pressure and included "skip to next dir".
23:15:12amiconnIt will announce the current file after returning to the browser
23:15:38CassandraI don't think it's a good idea, t0mas. Seems like extra complexity for little to no gain.
23:15:54t0masyeah, it is possible to disable the skip to next thing...
23:15:57t0masand then use the idle timer
23:16:01amiconnWake-up timer will be more useful imho
23:16:04CassandraBut that's just a personal opinion.
23:16:12*t0mas is still thinking about this...
23:16:16Cassandraamiconn, we have that.
23:16:18 Quit Sando (Read error: 104 (Connection reset by peer))
23:16:18*leftright votes for wakeup timer
23:16:21CassandraAt least on Archos.
23:16:21t0masmaybe a beep every 10 minutes, and a button to press to keep it on
23:16:23Jungti1234Can no one help me?
23:16:33t0masit would make listening to audio books in bed so much easier
23:16:34amiconnFeasibility depends on how low we can get the power consumption
23:16:41t0mas(sleeptimer never gets my sleep time right :))
23:17:12amiconnWake-up timer is on my todo list
23:17:18t0masamiconn: on iriver we can get quite low iirc... someone checked how long it could run w/o disk at 11 mhz...
23:17:31t0masif we even disable the screen it would be even longer
23:17:39amiconnCassandra: Yes, we have it on archos fmr/v2, and v1 with mod
23:17:49 Join `3nergy [0] (
23:17:50 Quit DangerousDan (Read error: 110 (Connection timed out))
23:17:55Bagdert0mas: not to mention if you leave the charger connected... :-)
23:18:09t0maseven better :)
23:18:27CassandraJungti1234, not sure - I assume you're using Bluechip's Windows dev kit? That's kind of a black box for us, I'm afraid.
23:18:28t0masthen it will last until the next snowstorm
23:18:32t0mas(in Holland at least :P)
23:18:43amiconnt0mas: I plan to do more. Put lcd in power-save, run timer code in IRAM only and put SDRAM to sleep, and even run _lower_ than 11 MHz if that helps saving power
23:18:59amiconnOh, and the tick interval will be reduced to 1 Hz while sleeping
23:19:00t0mascan we get < 11?
23:19:14t0masI tought disabling PLL was as low as we could get
23:19:17Cassandraamiconn, Ah, so you're thinking of implementing software RTC for H100?
23:19:22amiconnIf my calculations were right, we can get down to 6.25 MHz
23:19:33CassandraIf so, it gets my vote, for whatever my vote's worth.
23:19:44CassandraI'd considered doing it myself.
23:19:52t0masCassandra: so have I
23:19:59amiconn...but that needs the PLL, so it might be running at 11 MHz saves more power, because it allows to disable the PLL
23:20:18Jungti1234Cassandra: What meaning is it?
23:20:25t0masamiconn: I can lookup some things I've already tried... so you can focus on the cpu speed and IRAM exec thing
23:20:33amiconnCassandra: No RTC, just a wake-up timer
23:21:29amiconnI think a delay of 1 day or more is possible without running out of power
23:21:33CassandraThat's not so good, amiconn. It's kind of hard to use in conjunction with the sleep timer.
23:21:34amiconn(standard battery)
23:22:18 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
23:22:39CassandraJungti1234, looks like some sort of Windows / UNIX incompatibilty to me, but I couldn't say for sure.
23:23:25 Join hardeep [0] (i=hardeeps@norge.freeshell.ORG)
23:23:32CassandraHmmm. I was thinking you'd probably get weeks of standby time - is it really that power intensive?
23:23:44Jungti1234yes that's right.
23:24:40t0masJungti1234: try ./configure
23:24:58Jungti1234../ ?
23:25:22t0mas(calling it directly)
23:26:07amiconnCassandra: I really don't know yet. Need to do some experiments, measuring power consumption
23:26:18Jungti1234../tools/configure: 417: Syntax error: redirection unexpected
23:26:46t0masare you using a "vanilla" devkit?
23:26:46BagderJungti1234: check your cvs update output
23:26:50Bagderthat is a cvs conflict
23:26:53t0masor do you use cygwin?
23:26:57Bagderload the file, edit it
23:26:57t0masoooooooh wait
23:27:14Bagdert0mas: configure is a shell script
23:27:27t0masJungti1234: easy way: remove tools/configure
23:27:36t0masand 'cvs update' in that directory
23:27:45t0masthen it will reconstruct it
23:27:46Jungti1234ok wait
23:28:03t0mas(Bagder: I don't know perl, so I expected it to be some perl thing :))
23:29:44*t0mas is right back
23:29:56 Quit t0mas ("just have to restart xdm")
23:30:06linuxstbOuch again. "decreasing" the volume from -84dB to 0dB isn't good. Did someone have a fix for the wrapping numerical settings?
23:30:31BagderI think _FireFly_ had one
23:31:00CassandraBut the current volume control doesn't wrap.
23:31:01 Join amar [0] (
23:31:02 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
23:31:12linuxstbCassandra: It does in Sound Settings -> Volume
23:31:18t0masbck :)
23:31:26CassandraAh. :(
23:34:17t0masBagder? About 'The volume issues'... isn't it possible to calculate where the volume stops getting higher
23:34:22t0masand to make the setting stop there?
23:34:41_FireFly_linuxstb: it can be found here :
23:34:55t0masso people notice that with bass boost on +20 the volume setting doesn't go any higher than (for example) 70
23:35:00amarjust want to check; when programming for rockbox, you can't use floting point/real numbers
23:35:03Bagdert0mas: I leave the audio magic to amiconn and preglow, they master those areas far better than I do
23:35:23t0masit's more of an interface thing :)
23:35:36Bagderamar: correct
23:35:39t0masmakeing the setting stop at 70% is that's what we're limiting it too in a lower level...
23:35:45linuxstb_FireFly_: Do you need someone to commit it, or do you have CVS access? (I've lost track now)
23:35:46t0masthen it's clear what's happening..
23:35:54Bagderamar: you _can_ in fact use floats since they are emulated, but there's a huge penalty
23:36:30amarand another piont, if you do Int1/int2, it rounds to the nearist integer?
23:36:38_FireFly_LinusN: i have no cvs-acces
23:36:45_FireFly_linuxstb: i mean
23:37:00Bagderamar: yes
23:37:12Bagderah no
23:37:14linuxstbI always thought it did "floor" ?
23:37:16Bagderit doesn't round at all
23:37:21Bagderme no thinking
23:37:36Bagderamar: its plain integers, 19/10 = 1
23:38:02_FireFly_linuxstb: the patch won't be compile because some macros gots deleted i will make a new-one
23:38:03amarso it rounds down to the nearest integer
23:38:09Bagdernot nearest
23:38:20Bagderit cuts off any "decimals"
23:38:48 Quit Moos ("Glory to Rockbox")
23:39:00_FireFly_amar: eg when the result of a calculation is 5.9234 then only the 5 will be saved in an integer var
23:39:18amarBagder:yea, it heads down until it finds an integer
23:39:57t0masthat got me once calculating prices...
23:40:18t0massometimes you have to calculate in the right order
23:40:35t0mas(first multiply, then devide to keep some "decimals")
23:40:53linuxstbAnd then you get overflow problems...
23:41:14t0masnot with reasonable prices ;)
23:41:23linuxstbAh yes, the Lira is no more.
23:41:39t0masbut yeah... that's why I think everybody is used to devide first...
23:42:28amiconnconvbdf.exe.stackdump files are cluttering my build dirs :/
23:43:07 Join DangerousDan [0] (
23:43:42Jungti1234ah thanks guys
23:44:11_FireFly_linuxstb: ok updated
23:44:35_FireFly_linuxstb: you can find the patch on the same place as a had posted some minutes ago :)
23:44:55Lost-ashhmm. i just got a bleeding edge build and selected the 'boxes' wps (where'd themes go btw) and the unit hung while showing the wps's content
23:45:32Jungti1234Makefile:14: *** missing separator. Stop
23:45:52Jungti1234make: ***[tools] Error 2
23:45:54Lost-ashrepeatable too
23:46:14_FireFly_Lost-ash: in my build it doesn't hang at least in the sim
23:46:58t0mastime to go to bed :)
23:47:01t0massee you all tomorrow
23:47:11Jungti1234good night
23:47:53Lost-ash_FireFly_: it's doing it no matter which WPS i select, on my H1xx
23:48:25_FireFly_Lost-ash: try to reset the settings
23:48:36Jungti1234I go to the school.
23:48:48 Quit Jungti1234 ("bye")
23:48:53 Nick Lost-ash is now known as ashridah (n=ashridah@
23:50:30ashridahdamnit, i'll bbs. resetting the settings made no difference, since the bleeding edge had already done that
23:50:35 Nick ashridah is now known as Lost-ash (n=ashridah@
23:51:39_FireFly_i will test the latest build on my target after the build is done
23:57:00 Join ep0ch| [0] (n=ep0ch|
23:57:19dpassen1amiconn: the build works perfectly, i hope it gets committed
23:57:21*amiconn is updating all his rockboxes to unicode
23:58:05ep0ch|i can't seem to apply the new volume patch, anything special i need to know about?

Previous day | Next day