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

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

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

#rockbox log for 2007-11-15

00:02:36 Quit barrywardell ()
00:03:39 Quit male ("User disconnected")
00:04:32preglowi like the fact that files are given extension .mp3 no matter what encoder you use :P
00:07:08 Quit robin0800 (" HydraIRC -> <- Would you like to know more?")
00:07:58 Join handmadematters [0] (
00:09:38handmademattersHi all... Anyone in here available for a simple: "how would you start to solve the problem" question
00:09:54 Join Fel [0] (i=4a3b03c8@gateway/web/cgi-irc/
00:10:02Bagderask and you'll see!
00:10:26FelWhat the % rate of the chance that my sansa E dont get detected by rockbox auto install ?
00:10:39linuxstbFel: Which version of rbutil are you using?
00:10:48Felhuh the lastest ?
00:11:02Felthe 1 that we find on website D:
00:11:40Bagderwhat URL?
00:12:04Felno idea lmao
00:12:06Felwhy ?
00:12:15 Quit Fel (Client Quit)
00:12:20 Join Fel [0] (i=4a3b03c8@gateway/web/cgi-irc/
00:12:26Bagderbecause you asked a question
00:12:27Fel1 sec i tell you
00:12:45 Join weezerle [0] (
00:13:04linuxstbGo into the "About" menu in rbutil, and it will tell you its version...
00:13:09Felthe 1 right here
00:13:29linuxstbI just updated the links on that page - which is why I asked you which version you had.
00:13:36handmademattersiRiver H10... Scrollpad... ADC values... I want to get the scrollpad working properly. My problem right now is that i am a little bit annoyed of "writing code" test ist on the DAP "rewrite the code" test it on the Dap... I am very unfamiliar to programming with such unexpactable things like the ADC output... Is there a way to record the values and play around with them outside of the DAP
00:13:41linuxstbFel: Try downloading it again
00:13:46Bagderand there are *4* packages on that page...
00:14:09BagderI assume Fel means the win version then
00:14:56FelSNV Revision 15190 (m.1.0.2)
00:15:01Bagderhandmadematters: no I don't think so, the simulator that can be used for native development of some things don't simulate things at that level
00:15:22linuxstbFel: Try 1.0.3 - that will work for you.
00:15:31handmademattersok so back to my pencil and the papers ;-)
00:16:12Felahh thanks
00:16:37Feli didnt know that there were a new version of this program ..
00:16:50 Quit karashata ("Leaving.")
00:18:36 Quit ompaul (Client Quit)
00:19:39 Quit Fel ("CGI:IRC (EOF)")
00:21:47rasherpreglow: in theory it should, but I don't know how good a job voicefont does of putting them all into one file
00:23:24 Join ender [0] (
00:23:40 Quit handmadematters ()
00:25:50 Quit Febs (Remote closed the connection)
00:26:38preglowrasher: so the current thing is really untested?
00:26:55preglowrasher: i'll just rip out everything but lame and our own speex encoder anyway, probably
00:27:08 Quit Soap (Read error: 110 (Connection timed out))
00:27:38rasherpreglow: I believe I've created a voicefile with it, but I've obviously not been able to test it.
00:27:42preglowjudged by the fact that files are called .mp3 even if you encode with oggenc, i'll just go out on a limb and say no one has used a vorbis voice file yet
00:28:14rasherThat's simply because voicefont expects files to be named .mp3
00:29:02rasherDepending on whether voicefont does anything fancy to the files, or just cat them together, I don't see why it shouldn't work
00:29:19rasherWhy do we need our own encoder?
00:29:35 Quit OlivierBorowski_ ("Konversation terminated!")
00:30:31 Quit Nico_P (Remote closed the connection)
00:31:52preglowrasher: so we don't have to use ogg framing
00:31:57 Join mud [0] (
00:32:06 Nick mud is now known as mud-rb (
00:32:17 Quit mirak (Remote closed the connection)
00:32:20 Quit jgarvey ("Leaving")
00:32:26rasherI guess it'll be put into tools/?
00:32:27 Join TTThomas_ [0] (i=80af320c@gateway/web/cgi-irc/
00:32:44preglowrasher: i have no idea, it depends on svn libspeex, so i don't even know if we're going to go binary only or what
00:33:01 Join male [0] (
00:33:57linuxstb"binary only" ?
00:35:44preglowwell, the simple fact is, right now i use the resampler in speex svn
00:36:12preglowi don't know how to handle such a dependency any other way
00:36:25preglowwe do have libspeex in the rockbox source tree, of course...
00:36:45preglowbut how keen that is on building for system use, i don't know
00:37:17linuxstbIt could just be something in utils/ that a user needs to build and install manually. Similar to things like rbutil, ipodpatcher etc
00:38:46preglowjust sucks badly to have to make voice building any harder than it currently is
00:38:57linuxstbI'm assuming people already need to install a third-party encoder (i.e. lame) to create voice files?
00:39:05rasherThat's easy though
00:39:29rasherCompared to building and installing
00:39:32linuxstbAnd downloading a .exe isn't?
00:39:33 Quit ender (" Eagles may soar, but weasels are seldom sucked into jet engines.")
00:39:48linuxstbI'm not saying people need to download and build - we can provide binaries.
00:39:58 Quit ender` (Read error: 110 (Connection timed out))
00:42:23preglowgood point
00:43:07preglowpoint remains, though, should we require user to download speex svn themselves (don't really need svn, latest beta will do), should we try to use rockbox svn speex, or should we bundle yet another speex tree...
00:43:22linuxstbWhy bundle it at all?
00:43:28rashercan't it simply be built by our build system, like voicefont?
00:43:44DerPapstrbutilqt isn't bundled with the qt library either
00:44:34DerPapstso if someone wants to compile your speex encoder they have to get the library for theirselfes
00:44:43preglowsure, i'm ok with this
00:45:01rasherStill, but voicebuilding ought to be easier than building rbutilqt
00:45:14 Quit Zagor ("Client exiting")
00:45:30pengobundle it
00:45:52pengoand let them delete it if they dont want it
00:46:18pengono one likes hunting around for codecs
00:46:46 Quit jake_ ("Leaving")
00:47:06preglowi could just use regular libspeex, which would be easier to install, but then i'd need to code my own resampler, and that's not happening
00:47:17linuxstbpreglow: Are there many changes to our libspeex? It would seem a waste to have two very similar copies.
00:47:23 Join K-Man666 [0] (
00:47:27linuxstbCan't you just copy and paste their resampler?
00:48:14qwmi love rockbox as much as i love trannies. <3
00:48:45krazykitthank you for sharing.
00:49:37preglowlinuxstb: many changes? no, but the entire build system is gone
00:49:45preglowlinuxstb: i guess i could copy and paste the resampler, yes
00:50:22preglowthat'll break horribly the day distros come bundled with speex_resampler
00:50:35preglowahh, no, speex_resampler is in libspeexdsp
00:50:38preglowand i won't use that
00:51:04preglowok, i'll see about stealing resample.c/h
00:51:04 Quit TTThomas_ ("CGI:IRC (EOF)")
00:52:47linuxstbDoes libspeex need much of a build system, especially for your encoder? i.e. would a simple Makefile do the job?
00:53:25preglowmaybe, will of course break sooner or later or on esoteric systems
00:53:41preglowwe'll only want to compile a static lib, that simplifies things very much
00:53:48linuxstbThen the owners of those esoteric systems can fix it..
00:53:57preglowwell, ok, i'll see
00:54:08preglowperhaps i feel motivated too look at it when that is all that stands between me and commiting this stuff
00:54:28preglowmessing with makefiles and build systems is the most boring stuff i can think of
00:54:54linuxstbI don't mind giving it a go...
00:55:51preglowwe'll see, i'll need to comb through the source to look for stuff that'll break
00:55:56preglowthere are a couple of #if 0 here and there
00:57:05linuxstbBTW, do you think the speex encoder has a chance of running on our targets?
00:58:37preglowon arm, kinda limited, on coldfire, perhaps even wb mode
00:59:24 Quit Guile`` (Read error: 104 (Connection reset by peer))
00:59:48 Join Guile`` [0] (n=Guile@
01:01:04preglowcoldfire is just really cut out for speex stuff
01:01:05preglowmostly thanks to emac
01:01:42 Quit scorche|w ("CGI:IRC (Ping timeout)")
01:04:42 Quit Sedgewick ("Money can't buy happiness but it can provide a better class of enemy.")
01:07:37***Saving seen data "./dancer.seen"
01:09:23 Part linuxstb
01:10:58 Quit sarixe ("(EE) Failed to load "quit" module")
01:14:47 Quit mud-rb (Remote closed the connection)
01:15:40 Quit japc (Read error: 104 (Connection reset by peer))
01:22:56*preglow kicks build systems
01:23:29 Join hcs [0] (
01:23:50 Quit n1s ()
01:24:20 Join J3TC- [0] (
01:28:34 Nick fxb is now known as fxb__ (
01:28:45 Part hcs
01:29:52*preglow has a voice file
01:33:28preglowi think i'm just going to gut out the encoder selection stuff, if swcodec, use rbspeexenc, if hwcodec, use lame
01:33:52rasherSure, configure can easily do that
01:33:55preglowone meg english voice at speex −−quality 6
01:36:19preglow800kb at -q 4
01:36:31preglowso voice files are smaller at the same quality, that's for sure
01:36:49preglowthey're actually smaller at better quality
01:39:41 Join ze_ [0] (
01:42:34preglowhow do i do check if a file exists in the tools dir in bash?
01:42:55rasherif [ -f "tools/filename" ]; then
01:42:58preglowif [ -f "$(toolsdir)/lollerksates" ]; then ?
01:43:05rasherYeah, that should do it
01:43:15preglowbhargh, $(toolsdir) is set below voice config
01:43:19preglowi'll just move it, then
01:43:51pregloweh, no, toolsdir should be set by there
01:43:58 Quit ze (Nick collision from services.)
01:43:59 Nick ze_ is now known as ze (
01:44:47rasherI don't think you want to use toolsdir
01:45:02rasherThat's for use in the makefile it seems
01:45:19rasheruse $rootdir/tools instead
01:46:03preglowecho "$(rootdir)/tools"
01:46:13preglowsays: ../../tools/configure: 1613: rootdir: not found
01:46:31rasherThat's because you mean ${rootdir}
01:46:36preglowprobably :>
01:46:40preglowworks without them too
01:46:52rasherYeah, or that
01:47:08markunaway sleep
01:47:10rasher$() is either makefile notation or command execution/substitution
01:47:16markunoops :)
01:48:39preglowi guess i should make configure pass the path to as well, then
01:49:22rasherYeah. Currently it pretty much assumes the executables are in $path, which obviously won't hold true
01:49:29DerPapstmarkun: good night :P
01:49:32rasherPassing a path should work, of course
01:50:03 Quit darkless ("Leaving")
01:50:15preglowwhat's the syntax for checking a variable's contents?
01:50:21preglowi want to check if $swcodec is "yes"
01:50:29preglowgod, how i hate sh language
01:50:37preglowsomeone should rewrite this mess in perl :/
01:50:38rasherif [ "$swcodec" == "yes" ]; then
01:53:45preglowok, i'm gutting away decoder selection, lame is all you have for hwcodec and rbspeexenc is all you have for swcodec, anyone ardently opposed to this?
01:54:50 Join Variable [0] (
01:55:08rasherNot at all
01:55:55rasherI added it at the time because I figured eventually we'd need different encoders, and wrote it to be flexible because I didn't know how or why
01:56:06rasherNow that we have a plan, anything unneeded can be chopped
01:58:36preglowi really, really don't understand how the -e command line option ends up in the $encoder variable
01:58:41preglowplease explain :>
01:58:57 Quit PaulJam (".")
01:59:10preglowforget that
01:59:11preglowi see
01:59:16rasherThat's because of the weird perl thing
01:59:19*preglow is used to Getopt
01:59:42rasherYeah, it's a tad strange, but simpler
01:59:54preglownot than Getopt :>
02:00:38preglowi'll just pass the entire path along with -e=encoder, is that ok?
02:01:07rasherHrm.. I think that'd need some modifications
02:01:19preglowthere's no longer any need for encodewav to know what encoder it's calling anyway
02:01:26preglowwell, i'll try :)
02:01:45rasherWell, we switch() on $encoder, so that won't work with different paths
02:02:02preglowthat's no longer needed
02:02:06preglowand if it were, it just use regexes
02:02:29 Join Soap [0] (n=Soap@rockbox/staff/soap)
02:02:31preglowthat switch is no longer needed, every case statement is the same if i remove "speexenc" and "lame" with $encoder
02:03:06preglowand like i said, if we need it one day, it's just a matter of doing a pattern match on $encoder
02:03:23preglowif ($encoder =~ m/lame$/) {
02:03:24rasherYeah, I forgot how flexible perl's switch() is
02:03:45preglowperl doesn't have a switch normally, this is the first time i've seen it used :)
02:04:01rasherIt's a standard module
02:04:09 Part toffe82
02:04:38 Quit Isolinear (Read error: 110 (Connection timed out))
02:05:25 Quit RaRe ("Quit msgs should be longer.")
02:05:27rasherSuch as: switch ($val) { case (\%hash){ print "entry in hash" } }
02:05:44preglowperl is nifty :)
02:05:47preglowi love it
02:07:35 Quit atsea- (Remote closed the connection)
02:07:58preglowsyntax errors are much loved by bash, i see
02:08:01preglowbloody crap language
02:08:26rasherYou just need to write proper shell!
02:09:11preglowone space too little and it detects an error 200 lines further down!
02:09:45zewell it must've still made some kinda sense up to that point
02:10:20rasherBlame [
02:10:22preglowand this bug i can't even find
02:10:24rasher(I bet)
02:10:29preglowyeah, the fact that [ is a program
02:10:32preglowdon't start me on that
02:10:52rasherOh that's just so clever!
02:12:08 Join karashata [0] (
02:12:26preglowi just can't figure out what's wrong now
02:12:39preglowi'm working around line 490, bash finds the error around line 1599
02:12:40rasherPastebin it and the error
02:13:14preglowpastebin configure?
02:13:24rasherWhy not?
02:13:33preglowshrug, thought perhaps a patch might be neater :)
02:13:37rasherOr just the part you're working on
02:13:38preglowgimme a sec
02:13:44 Quit ze (Read error: 104 (Connection reset by peer))
02:13:55rashersure, that works too, but I lose line numbers unless I start counting
02:14:27 Join ze [0] (
02:15:24preglowi'll spam you the error in private
02:21:22 Join ze_ [0] (
02:21:40 Join webguest36 [0] (i=54975d43@gateway/web/cgi-irc/
02:21:45 Quit webguest36 (Client Quit)
02:21:52 Quit ze (Nick collision from services.)
02:21:59 Nick ze_ is now known as ze (
02:23:24 Quit ivan` (Read error: 104 (Connection reset by peer))
02:27:54 Join ivan` [0] (n=ivan`
02:28:45 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
02:29:08 Join miepchen^schlaf [0] (
02:34:54 Join berninup [0] (
02:36:23 Quit LinuxMafia (Remote closed the connection)
02:37:23 Part berninup
02:38:00 Join billenium [0] (
02:38:29 Join Traveler5 [0] (
02:39:06Traveler5any major games in development that anyone wants to share about?
02:40:42 Join Ebert [0] (
02:40:52Traveler5just curious, I'm kind of working on a rts game
02:44:49 Join Mouser_X [0] (i=cf9bb003@gateway/web/cgi-irc/
02:52:06 Quit karashata ("Leaving.")
02:56:17 Join ___jake [0] (
02:57:50 Quit Traveler5 ("Java user signed off")
02:58:05 Join karashata [0] (
02:58:45 Join webguest88 [0] (i=d06b6e04@gateway/web/cgi-irc/
02:59:07 Quit webguest88 (Client Quit)
02:59:47scorcheDerPapst: around?
03:02:31scorchei saw :)
03:02:44 Nick karashata is now known as kimi-sharamin (
03:03:05DerPapsthehe ;) i guess #ipl
03:03:07 Join Demache [0] (
03:03:23 Nick kimi-sharamin is now known as karashata (
03:04:04 Quit Demache (Client Quit)
03:04:16 Nick karashata is now known as kimi-sharamin (
03:04:45 Nick kimi-sharamin is now known as karashata (
03:07:41***Saving seen data "./dancer.seen"
03:13:04 Quit amiconn (Nick collision from services.)
03:13:13 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
03:15:19 Join hannesd_ [0] (
03:19:12 Join RaRe [0] (
03:20:33 Join DerPapst_______ [0] (i=5b23dcfc@gateway/web/cgi-irc/
03:20:54DerPapst_______scorche: got disconnected
03:21:11 Join webguest23 [0] (i=4a3b03c8@gateway/web/cgi-irc/
03:21:11scorchei didnt say anything anyway
03:21:22DerPapst_______and my irc client refuses to connect to the network :S
03:21:24 Quit webguest23 (Client Quit)
03:21:41DerPapst_______just in case you did ;)
03:22:06 Quit DerPapst (Nick collision from services.)
03:22:26 Nick DerPapst_______ is now known as DerPapst (i=5b23dcfc@gateway/web/cgi-irc/
03:22:59 Join eigma [0] (n=cat@
03:23:13 Quit DerPapst (Nick collision from services.)
03:23:26 Join DerPapst [0] (
03:23:31DerPapstfinally :P
03:24:49DerPapstgood night :)
03:24:53 Quit DerPapst (Client Quit)
03:25:33 Join kkurbju1 [0] (
03:25:47 Nick kkurbju1 is now known as kkurbjun (
03:29:58TTThomasanyone know how to hook a sansa e250r up via linux?
03:30:50kkurbjunhey eigma
03:30:53 Quit hannesd (Read error: 110 (Connection timed out))
03:30:54 Nick hannesd_ is now known as hannesd (
03:30:58 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
03:31:01eigmakkurbjun: at home?
03:31:23kkurbjunand I have time to work on the m:robe tonight :-D
03:31:51kkurbjunyou said that you don't get that crash right?
03:31:56eigmathat's right
03:32:22kkurbjunhuh, I wonder what's causing it on my build.. you're not running with the patched compiler are you?
03:32:31eigmathat's right, unpatched
03:32:31kkurbjunI should try it without the extra options
03:32:42eigmawith -mcpu=arm926ej-s -mfpu=fpa
03:33:28eigmajust built on my machine and transfer it to your m:robe, see if it still happens
03:33:37kkurbjunwill do
03:34:49kkurbjunwow, connection's slow today
03:35:04eigmalemme stop the torrents :)
03:35:22eigmashould be a little better now
03:35:22TTThomaskrazykit: e250r doesn't have an option to change to usb msc mode
03:35:23 Join Guile` [0] (n=Guile@
03:35:30TTThomasat least not that i can find
03:35:31krazykitTTThomas, "rhapsody mode"
03:35:32kkurbjun:), yes much
03:35:35eigmacan I remove the N5002-*.rar files?
03:36:09kkurbjunyep, I have them, I havn't gotten a chance to really look at them though :/
03:36:35TTThomasthat was it. evil r version
03:39:12krazykityou COULD convert it to a vanilla e200 ;-)
03:40:03TTThomasteach me your ways obiwan
03:40:07TTThomasor something
03:42:54krazykityou really want to? it isn't really that useful anymore, now that the r's have an easier install.
03:43:23TTThomasdont know what i want. other than a working rockbox of course
03:43:25*scorche is actually installing to his new R now >_>
03:44:04TTThomasif rockbox works fine on r, might as well leave it that way
03:45:37Ebertare there major compile errors still? on what builds?
03:45:41 Quit Mouser_X ("CGI:IRC (Ping timeout)")
03:47:07TTThomasis there a difference between using the precomiled binary or compiling it yourself?
03:48:06krazykitTTThomas, there is if you add patches from the tracker. if you just compile the vanilla source, though, it's the same
03:48:20scorcheassuming you use the same compilers we do
03:48:21TTThomasmmm.. vanilla
03:48:50TTThomasso i just grab the e200patcher.linux file and run it?
03:49:10krazykitthe manual explains the install process.
03:49:39TTThomasyeah, reading it but not quite understanding
03:50:12TTThomasi'm not sure which one is the precompiled and which is the one i need to compile
03:50:22krazykityou're following this one?
03:50:58krazykitthe answer to your question, then, is "yes"
03:51:02TTThomask, thx
03:53:08kkurbjuneigma: how do you use scp when you have a port on the host also?
03:53:28eigma-p or -P, check the manpage
03:53:47kkurbjunahh, yeah, thanks
03:54:08 Quit Guile`` (Read error: 110 (Connection timed out))
03:55:05TTThomassweet, i broke it already
03:55:18 Quit qweru ("moo")
03:56:48TTThomasblue led is lit, nothing comes on screen, cant get it to power off
03:57:24kkurbjuneigma, hmm, it still freezes with your build too
04:01:14TTThomaswhat does manufacturing mode on the e200r look like? does it say anything or is it just a blank screen?
04:01:27scorcheexactly what you described
04:02:14eigmadunno what to tell ya..
04:03:54kkurbjunyeah, no worries, I'll just play around with it more do see what I can find
04:08:06 Quit Thundercloud (Remote closed the connection)
04:08:22TTThomaswhen running sansapatcher am I supposed to be in manufacturing mode still or normal mode?
04:09:24PaulPositionsansapatcher == e200patcher ?
04:09:27 Join iamben [0] (
04:10:17TTThomasum, i believe so, lemme check
04:10:53TTThomasdunno actually. manual calls it sansapatcher
04:11:55TTThomasi messed up, ignore the previous question
04:12:52PaulPositionUhrg... Guess I don't know then.. The wiki page linked above says e200rpatcher.exe but then I don't actually own one so I was just "doublechecking" your info.. :p
04:13:08PaulPosition(which I should not do, just cause more confusion)
04:13:45TTThomasPaulPosition: you got me to read things more carefully, which solved my issue. so thanks!
04:15:39PaulPositionheh... It figures ;)
04:20:32 Join Traveler7 [0] (
04:21:23Traveler7I started looking at the rockbox code, but am still a newb at c. I found drawing functions and know how to use if statements and such.
04:21:44Traveler7my question is what does a simple constructor look like?
04:25:10PaulPositionconstructor? Like, for instancing a class? I don't think there's any, C is NOT an OOP language, unlike C++
04:27:20Traveler7hehe okay, is there like a primary function then, or do I need to look at how each application is initialized and just follow that?
04:27:30 Join sarixe [0] (
04:27:45scorcheTraveler7: do you mean plugins?
04:28:44Traveler7scorche: yeah
04:29:38TTThomasthank you anyone that put up with my stupid questions
04:29:59TTThomase200 install much harder than gigabeat f40 install
04:30:07TTThomaswell, more steps at least
04:30:47krazykitthe e200r is. the e200 install is dead easy.
04:31:26TTThomaswell its working now, not going to mess with what works
04:32:16 Quit Traveler7 ("Java user signed off")
04:32:23 Join Traveler0 [0] (
04:34:30Traveler0I like rockbox, beacause unlike ipodlinux even if one player's offspring are encrypted chances are there still are several to choose from that aren't
04:35:16krazykitrockbox and ipodlinux have different goals, anyway...
04:36:50Traveler0true, I can go with either, I mainly just want games, music, and video on the go but in as small a package as possible
04:36:53 Join Mouser_X [0] (i=cf9bb003@gateway/web/cgi-irc/
04:37:36 Join K-Man666_ [0] (
04:38:12 Quit K-Man666_ (Client Quit)
04:38:21TTThomasdone working on my sansa for the night. now to figure out what the hell ms did to my wife's zune
04:43:19 Join iamben_ [0] (
04:51:36 Quit Variable (Read error: 110 (Connection timed out))
04:53:44 Quit K-Man666 (Read error: 110 (Connection timed out))
04:56:08 Quit iamben (Read error: 110 (Connection timed out))
05:07:01TTThomasmy god. zunes phone support has the worst music ever
05:07:13TTThomasthat should have been community sorry
05:07:43***Saving seen data "./dancer.seen"
05:09:52 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
05:12:43krazykitTTThomas, please stay on topic in here
05:13:28 Join miepchen^schlaf [0] (
05:15:44 Quit bb (Nick collision from services.)
05:15:49 Join bb_ [0] (n=bb@unaffiliated/bb)
05:18:46 Quit donutman25 ("ChatZilla 0.9.79 [Firefox]")
05:21:11 Quit Traveler0 ("Java user signed off")
05:25:28 Join iamben [0] (
05:34:22 Join psycho_maniac [0] (
05:35:55 Quit iamben_ (Read error: 104 (Connection reset by peer))
05:38:32TTThomasmy e250r only shows half the directories under the file menu (music is not one of them) anyone know why or how to make it appear?
05:39:57karashataset the file view to "all"
05:40:31karashatathe folders you can't see are marked as hidden folders, setting the file view to "all" will show them
05:43:30 Join Calcipher [0] (
05:45:09Calcipherhey guys
05:45:21Calcipherany new wonderfull things on the rb front?
05:45:37CalcipherI like poppin in and seeing whats happening
05:45:39psycho_maniaccheck the main page?
05:45:57psycho_maniacor the MajorChanges page on the wiki
05:46:39Calcipheraww, that no fun, I don't just mean implemented stuff, just kinda like what you guys are cookin up atm
05:47:12psycho_maniacdid something happen to the scrolling margins patch?
05:47:32krazykitonly that left margins were comitte
05:47:55psycho_maniacwhat is the FS# for the regular scrolling margins patch? i cant seem to find it in the tracker :S
05:50:03psycho_maniaci figured it would get closed soon
05:50:53Calcipherwow, left margins? what does that mean, what about left margins was changed exactly?
05:54:45karashatathe left margin commited in SVN allows you to specify a left margin for any information you might want to stick beside a piece of album art so it doesn't overlap the art or get overlapped by the art
05:55:08karashatathe limitation being that the art needs to be to the left of the information, and not the other way around
05:55:48karashataI'd assume you could use it for other things you might want to put at the left side of the screen and not have overlapped by writing
05:56:47Calcipheroh I see, didn't know about it since it was related to AA, I don't pay much attention to the AA personally, but its a great add on for folks who like that
05:57:55karashatait's not really related to the AA directly, but the primary reason it's in there is because of the AA commit and people wanting margins to be able to position info beside the art without anything overlapping
05:58:32karashatawould be nice IMO if the margins worked both ways, but the devs were kind enough to give us the left margin instead of nothing
06:00:21psycho_maniacis there a way to show the code of the wps while its being used? (Example i would have this show on the first line: Artist: (%ia) Korn?
06:01:47karashataI would suspect if you put a second % in front of the first, you should get an output of %ia
06:02:02karashatasince %% = % in the WPS code
06:03:35karashataso your WPS code would have to look something like "Artist: (%%ia) %ia" to display your example
06:03:50psycho_maniacthank you
06:04:09karashataadd any scrolling, margins, conditionals, etc. to your liking
06:05:00karashatawhy do you want the WPS tags displayed in the WPS, if I might ask?
06:05:54psycho_maniaci was thinking of making an example wps
06:06:12psycho_maniacyou look at it and it shows what appers and also shows what will appear.
06:06:32karashataahh, interesting idea
06:06:47 Join goffa [0] (n=goffa@
06:13:55 Quit weezerle (Read error: 110 (Connection timed out))
06:14:49 Join weezerle [0] (
06:18:05kkurbjunjhMikeS: are you around
06:18:13kkurbjunI have a question on spinlocks
06:18:44kkurbjunspecifically, I'm wondering if they can be used with interrupts
06:19:58kkurbjunand if they can't is there another way to protect data?
06:20:47JdGordoniusnt jhMikeS trying to fase ou spinlocks?
06:20:59kkurbjunI didn't know that
06:21:06kkurbjunwhat's the replacement?
06:21:45JdGordonr15593 Portal player i2c driver: More struct spinlock phaseout.
06:22:02kkurbjunor do you know another way to protect data in an interrupt/function that works with that interrupt data.. hmm, I'll look at that revision
06:22:42JdGordonspinlock in a interrupt probably isnt a good idea....
06:23:29eigmakkurbjun: alternatively, temporarily disable the UART interrupts while you're updating the ringbuffer from the main thread
06:24:34kkurbjunyeah, that could work also, I'm interested in the non-locking fifo structures
06:24:44kkurbjunthanks for the link
06:24:57eigmaI think the lack of CS background will be felt even more acutely while reading on that subject.. :)
06:25:36eigmalockless data structures are a fairly advanced topic afaik
06:25:46kkurbjun:), I have a bit of a CS education, just some fundamentals of algorithms, data structures..
06:26:26 Quit Calcipher ("I-n-v-i-s-i-o-n 2.0 Build 3515 with A Pack Fix By")
06:28:34eigmaJdGordon: I've got the DSP adding numbers :)
06:28:48psycho_maniackinda a neat idea:
06:29:17JdGordonpsycho_maniac: as a demo of the avilable tags?
06:29:38psycho_maniacyeah. a kind of example wps
06:29:59*JdGordon thinks that one should be ported to all targets and go in
06:30:27psycho_maniacthat dumb was a dirty write up. didnt. just wanted one as an example.
06:30:41JdGordoneigma: using free tools?
06:30:43kkurbjunyeah, I think I'll just disable the uart interrupt while I'm reading
06:30:57*karashata notes that the "album" tag isn't displayed anywhere
06:31:58eigmaJdGordon: almost free :)
06:32:06eigmafree as in beer, not free as in speech ;)
06:32:10psycho_maniaci never really even see the composer tag being used.
06:32:20*karashata also notes, neither is the "title" tag...
06:32:32JdGordoneigma: good enough for me :)
06:32:37karashataartist, album, and title are the three big ones
06:32:46karashatacomposer tags half the time aren't even filled in
06:32:58*eigma dives again to try to get some data out to the DAC
06:33:14karashatagenre... kinda depends on who's tagging the files, some people care about it, some don't
06:36:33JdGordonanyone know if the NODO is in the wiki? or just the sources?
06:37:28JdGordonah, it is there.. just not linked on the docs index
06:39:36krazykitmaybe it should be linked there?
06:40:23eigmakkurbjun: do you know how to put the ARM to sleep, to wait for interrupts?
06:41:15eigmachrist, there's too many {system,kernel}-{arm,system,}.{c.h} files!
06:41:37kkurbjunI havn't looked into it too much, but there is a way to put the core to sleep
06:41:52kkurbjunI don't know what the requirements are for interrupts
06:42:10JdGordonbah, im blind.. it is there :p
06:43:15psycho_maniaci think its in the new ports page. Thats not the correct name though.
06:44:26JdGordonits under "contribution information" in the index.. which is a bit meh....
06:46:24psycho_maniaci remember reading it somwhere.
06:46:42 Join webguest47 [0] (i=429d9d92@gateway/web/cgi-irc/
06:49:15kkurbjun"To enter SLEEP mode the CPU must enter the Wait-for-interrupt state using register 7
06:49:15kkurbjunin CP15 of the CPU system control coprocessor. This ensures that the CPU is in a
06:49:15kkurbjunlow-power state."
06:49:24eigmaI *just* found that too
06:49:33kkurbjunthat's from the 926 technical reference
06:49:35eigmaI was grepping through linux sources
06:49:44psycho_maniacgetting creative now. see the last line in the wps
06:50:09webguest47I was wondering if anybody had any knowledge of using a microSDHC card with a Sansa e260 and rockbox?
06:50:41krazykitwebguest47, it works as long as you load music on the microSDHC card with a card reader.
06:50:55krazykitand the microSDHC card will NOT work with the OF, only with rockbox.
06:51:10webguest47right - that is what I have read - only with rockbox
06:51:35webguest47so I need to preload the card with my files and then pop it in the player and boot up rocbox?
06:52:13webguest47when I init the database, I guess it will find the files on the SD card?
06:52:44krazykiti believe it will, but i don't use the database.
06:53:39webguest47ok - sweet... That helps me a lot. You just saved me like 3-4 hours more of digging.. Thanks!!
06:54:50psycho_maniacwas the left margins patch comitted before or after 2007-11-13?
06:55:03maleYay for %m!
06:55:07kkurbjuneigma, are you implementing the thread_sleep function?
06:55:18kkurbjunor core_sleep
06:56:46eigmajust wanted something to replace for(;;)nop; because that just sucked :P
06:57:26eigmalooks like I might have to settle for that though
06:57:46 Quit hannesd (Connection timed out)
06:59:00 Quit Ebert ()
06:59:44 Quit webguest47 ("CGI:IRC (EOF)")
07:02:18psycho_maniacI think i may have to make a couple example themes because you really cant show EVERYTHING on just one screen. i have a great idea for the new left margin tag
07:02:52psycho_maniaci could show everything if i really wanted to but the text would be very small
07:06:29 Quit sarixe ("(EE) Failed to load "quit" module")
07:06:59 Join WebGeist [0] (i=4a87b329@gateway/web/cgi-irc/
07:07:46***Saving seen data "./dancer.seen"
07:08:23psycho_maniac%m|x| means where the text will start to display correct?
07:09:43psycho_maniacat the point you tell it (IE: %m|160|)
07:10:38karashatathat's where the margin is, and it will (if told to) scroll between that point and the right side of the screen
07:11:31 Join criznach [0] (
07:11:38 Quit karashata ("I will *SO* make you regret that... Later... *is a lazy dragon*")
07:18:14psycho_maniaci am having problems with the margin tag. this is what i have: #%s%m|160|text starting at half the screen
07:21:25psycho_maniacbesides the # what would prevent that from working? it just loads the default wps and when i put # in front of it. the wps displays correctly.
07:22:02PaulPosition%m is a non scrolling margin, %s tells it to scroll... ?
07:23:25 Join goffa_ [0] (n=goffa@
07:25:04 Quit Mouser_X ("CGI:IRC (EOF)")
07:25:35PaulPositionI think you only need the %s|160 bit for a margin+scrolling text.. and %m for margin but non scrolling.
07:25:47 Join colin_ [0] (
07:26:06psycho_maniacok well i just tried %m and it still didnt work.
07:26:44PaulPositionFS #8135 - add an optional "left margin" parameter to the %s WPS tag - e.g. %s|100|. This patch uses parts of the scroll-margins patch (FS #2954), but is much smaller, only offers a left-margin, and only affects the WPS code.
07:27:18PaulPositionAllow setting a margin on a non-scrolling line by using %m|margin| instead of %s|margin|. This allows to easily place dynamic info next to album art.
07:31:01psycho_maniacalright. the wps does not work on my player either so i have something messed up
07:31:33 Quit goffa (Read error: 110 (Connection timed out))
07:33:02psycho_maniacany ideas?
07:35:32PaulPositionHmm, been a while I haven't played with wps.. conditionals maybe? If you have the dev. environement (cygwin or vmware) I think there's a wps-debugger somewhere in there.
07:36:10psycho_maniacyeah i couldnt get that to work
07:36:18PaulPositionOr try very very basic (read bare) wps with only a few lines. If it checks, add some more, etc.
07:36:38PaulPosition(that's how I used to work)
07:37:44psycho_maniaci just have this on a line and it still has errors somehow %m|20|
07:38:51psycho_maniacand then i added this and it still dont work: %m|20|%cl:%cM:%cS - %cm/%cd/%cY
07:41:12psycho_maniaci figured it out and PaulPosition said it but i didnt pay attention
07:44:32 Quit Mxxd ("leaving")
07:45:07psycho_maniacnow is there a way to have words and then have the margin start? (ie: x scrolls, y does not) yyyy - xxxxx
07:53:25 Quit WebGeist ("CGI:IRC (EOF)")
07:58:01 Quit PaulPosition ()
08:01:53amiconnkkurbjun: Did you check the register value I asked for menwhile on your mini g1 ?
08:02:59psycho_maniacfinially figured it out. how about this idea?
08:07:52 Join Lambuntu [0] (
08:11:17jhMikeSkkurnjun: what do you wan't to do? if you want to sync with interrupts, set_irq_level is the way
08:14:17 Quit eigma ()
08:18:47 Quit Lambuntu (Remote closed the connection)
08:21:49 Join Rob222241 [0] (
08:24:01 Join atsea- [0] (i=atsea-@gateway/tor/x-d11fc7c104eb39ab)
08:26:31 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
08:32:32 Join foolsh [0] (
08:39:13 Quit Rob2222 (Read error: 110 (Connection timed out))
08:43:54 Join FOAD_ [0] (
08:44:19amiconnThe PP 'mono' lcd bridge does have configurable speed :D
08:45:03amiconnThis means we can get significant lcd update speedups on ipod 1st...4th gen, mini 1st gen, and Sansa c200
08:51:57 Quit colin_ ("")
08:58:03 Join ender` [0] (
08:59:48 Quit FOAD (Read error: 110 (Connection timed out))
08:59:49 Nick FOAD_ is now known as FOAD (
09:02:18 Join Lynx [0] (
09:04:41GodEateramiconn: are you trying to reach a FPS of ∞ ?
09:05:46amiconnOf what?
09:05:53GodEaterinfinity :)
09:05:55*amiconn sees only gibberish chars
09:06:15amiconnDid you check the fps table, how slow the c200 lcd update is?
09:06:27GodEaterno to be fair I didn't
09:06:44GodEaterI just felt I ought to remind you of the laws of physics, in case you were thinking of breaking them. The penalties are harsh :)
09:07:38 Join Zagor [0] (
09:07:47***Saving seen data "./dancer.seen"
09:08:09GodEaternow that I've checked the LcdFrameRates page though - I see what you mean
09:11:47 Join petur [0] (n=petur@rockbox/developer/petur)
09:13:37 Quit atsea- (Remote closed the connection)
09:17:13GodEatergod damn windows you stupid stupid OS!
09:18:32 Quit Lynx_ (Read error: 110 (Connection timed out))
09:18:32 Nick Lynx is now known as Lynx_ (
09:19:02amiconnSeems we can roughly double that speed. Maybe more - need to experiment further
09:22:20 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
09:24:32 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
09:26:07*GodEater wanted to extended a partition in windows using the diskpart command. He only wanted to extend it by 10GB. He issues the command "extend ?" hoping for a help message on how to do this. Diskpart complained that the argument was invalid. So he issued the "extend" command, hoping it would shed more light on how to use the command. It extended the volume all the way to the end of the disk. There is no way to undo this. Bugger.
09:26:51GodEaterand why god why, is this command not available in the disk admin tool? Where there's a nice gui.
09:28:04 Join Wildboy99 [0] (n=chatzill@
09:29:09 Part Wildboy99
09:33:12 Quit crwll ("remontti")
09:36:59ZagorGodEater: isn't windows/dos command parameters usually prefixed by / ? /me recalls that is the whole reason they went with \ for path delimiters.
09:37:36Zagornot that it makes the default action of "extend" any less stupid
09:37:49 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
09:39:07pondlifeMaybe a silly question, but do MAS targets use pcmbuf.c, or is that SWCODEC only?
09:39:42pondlifeI'm fairly sure they don't...
09:40:29 Quit daurnimator (Remote closed the connection)
09:40:51LinusNthey don't
09:40:52GodEaterZagor: unfortunately, Microsoft have now decided to give us discrete "shells" for certain functions. So "Diskpart.exe" is what you run, and then that presents you with a "DISKPART>" prompt. So the "extend" command is only valid there, and sadly "extend /?" does exactly the same as "extend ?" did.
09:40:57 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
09:41:37GodEaterthe only place I found words on how to specify the arguments to extend was on the microsoft website, after the fact :(
09:41:45Zagoroh lovely
09:42:28Jreminds me of managing my l2 switch via command line!
09:42:35GodEaterI will no doubt get a kicking from our Windows server team later today for eating up all that valuable free disk space :(
09:43:28*petur always uses 3rd party tools for partitionmanagement on windows (PQMagic)
09:44:02GodEaterpetur: not allowed to install it on servers here :(
09:44:25GodEaterbesides which, the diskpart command (if you use it right) is much cleaner, since it doesn't require a reboot
09:44:38GodEaterthe volume just extends on the fly
09:45:18*GodEater grabs the now wildly flailing topic steering wheel and attempts to gain control over the direction of discussion once more
09:45:33peturhow about this:
09:45:33 Join homielowe [0] (
09:45:49peturplayback/buffering screwed up this morning
09:46:18pondlifeJdGordon: Please could you try the updated keyclick patch - I'd like to commit it soon...
09:47:01peturI was at about 10-15 seconds near the end of a song when I resumed this morning and decided to listen to the song again, so I skipped back. It played about 10-15 seconds of the beginning and then skipped automatically to the next song.
09:47:03pondlifeSince owning a Gigabeat (thanks to bascule), the keyclick has shot up in priority for me :)
09:48:02 Quit Lynx_ (" HydraIRC -> <- The dawn of a new IRC era")
09:49:03pondlifepetur: Resume from off, or stop?
09:49:11peturresme from off
09:49:23pondlifeStart screen straight to WPS?
09:49:35peturboot, then press play
09:49:46 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
09:49:58pondlifeIs this with lostlogic's latest update?
09:50:06peturprobably not
09:50:07pondlifeI think the first one might have broken stuff...
09:50:37*petur admits it might even be a very old revision (before the weekend)
09:50:39pondlifeYou'd have been unlucky to update in that 5 min window though :)
09:50:58pondlifer15626 is the bad 'un
09:51:29peturI should first se if I can reproduce and then load the latest and try again
09:52:10psycho_maniac7307 refers to all players that use sw-codecs?
09:52:35 Join Casainho [0] (
09:52:48Nico_Ppetur: I think that's one of my bugs, which I then fixxed
09:52:53pondlifepsycho_maniac: Yep
09:53:23CasainhoHello to all!! :-)
09:53:45Casainhocan anyone take me out one doughty?
09:53:46pondlifepsycho_maniac: But it's most useful on players which have little tactile feedback, like the Gigabeats.
09:54:16*petur discovers he runs r15503 and hides
09:54:57Casainhoto have USB OTG in RockBox, we need to have one USB interface on processor with OTG capabilities??
09:55:23BagderCasainho: yes, or an external oth chip
09:56:03CasainhoIt's no possible to implement OTG by software?
09:56:13markunare we really implementing USB OTG or just USB host?
09:56:20psycho_maniacI know this has been asked severl times. What is stopping 7738 from being comitted?
09:56:20peturNico_P, pondlife: very reproducable on r15503, will now upgrade and try again
09:56:27Bagdermarkun: Casainho is on his own venture...
09:56:46pondlifepetur: Keep us informed :)
09:56:48BagderCasainho: you need software support too sure, but the hw needs to support it
09:57:22Casainhookok - I am confused, OTG is diferent form USB host?
09:57:32Casainhowhat is desirable?
09:57:40Bagderyou tell me, its your player
09:57:49Casainhoto be able to copy files from one device to other?
09:58:13Casainhoor to read/copy files from player to pen flash drive?
09:58:18LinusNis "Wheelspeed" a word?
09:58:29pondlifeLinusN: Not really, no
09:58:29psycho_maniacI dont think so
09:58:30Casainhoor USB Hardisk drive
09:58:47pondlifeHow about accessing a file system directly on a USB disk?
09:58:50LinusNshouldn't it be "wheel speed" at least?
09:58:55 Quit TMM ("Ex-Chat")
09:59:01pondlifeLinusN: Yes
09:59:26markunCasainho: with USB OTG a device can be either host or device.
09:59:52petur`Nico_P, pondlife: current SVN still has the bug!
09:59:55LinusNdo we have any issues left with 7738?
09:59:57Casainhoand without OTG, can be host?
10:00:00Nico_Poh damn
10:00:07markunCasainho: depends on the hardware
10:00:09pondlifepetur: Go Flyspray Go!
10:00:24*petur goes
10:00:31BagderCasainho: try ask google or wikipedia for details
10:00:49pondlifepetur: Which device? H180?
10:01:12psycho_maniacLinusN: did you check flyspray? one person complained about a current error but it has been re'synced to current svn.
10:01:15pondlifeI hope to have one of those soon.. Have ordered my drive....
10:01:28LinusNpsycho_maniac: yes i checked flyspray
10:01:53CasainhoBagder: okok
10:02:01*Nico_P attempts to reproduce
10:02:02LinusNi'm talking about problems with the actual acceleration
10:02:02pondlifeLinusN: Does #7738 still have settings?
10:02:17LinusNpondlife: looks like it
10:02:39psycho_maniacoh. i dont think so just its a big laggy but that is the same with normal scrolling because i believe that has to deal with cpu boost.
10:02:41pondlifeIIRC the plan was to decide if we could have fixed values (per target, perhaps).
10:02:46Nico_Ppetur: oh it was actually the same track?
10:03:15Casainhomarkun: can you please see the quicly datasheet, and tell me if this processor USB would support MAster and Salve USB?
10:03:18Nico_Pthen it might be the dreaded FS #8092
10:03:25Nico_Pnot sure though
10:03:54GodEaterpondlife: I think the current default values it uses are fine on all clickwheel ipods
10:04:02GodEaterall it needs is the setting to be removed now
10:04:14peturFS 8160
10:05:27pondlifeNico_P: Could you perhaps put more explanation on FS #8092... I don't quite understand the report. Is it a buffering or playback issue? Can't it just do a full rebuffer, or am I misunderstanding the problem?
10:05:43Nico_Ppetur: I'm almost sure it's a manifestation of FS #8092
10:05:51Nico_Ppondlife: buffering
10:05:53*pondlife becomes a teddy bear for a while
10:06:06pondlifeSo, why not just do a full rebuffer?
10:06:25pondlifeI suppose my basic question is
10:06:35pondlifeWhat is the purpose of the routine rebuffer_handle().
10:06:48desowinlinuxstb: you're getting famous - ;-)
10:06:53Nico_Ppondlife: it's to go back to a point in the track which we don't have
10:07:10pondlifeOK, so I'd expect a full rebuffer, right?
10:07:19Nico_Ppondlife: the problem with it is that we most probably don't have enough space left to write all of the required data
10:07:26Nico_Ppondlife: yeah, that would make sense
10:07:47pondlifeSo what goes wrong? A race?
10:07:54Nico_Pthing is we do have quite a lot of useful data... scrapping it sounds like a waste
10:08:15Nico_Ppondlife: no, it's actually a sign that some things work well :)
10:08:35pondlifeAny other option is likely to complicate stuff horribly, no?
10:08:52 Join Casainho_ [0] (
10:08:52pondlifeThis is not a common condition in "normal playback"..
10:09:12Nico_Ppondlife: I was thinking of using the callback stuff we discussed with jhMikeS to allow the buffering thread to close handles and notify the audio thread about it
10:09:15pondlifeOnly a rewind (or back skip) off the start of buffered data will trigger it, right?
10:09:26Nico_Pthat way the buffering thread could make room to rebuffer the track entierly
10:09:33Nico_Ppondlife: yes
10:10:01pondlifeSounds complicated. Why do we need to close handles? I'd just start afresh..
10:10:22Nico_Pyeah maye... could be a fix
10:11:00Nico_Panyway I need to go to class
10:11:10pondlifeFor the occasional waste of disk spinup time when someone rewinds back in the first buffered track, it's worth KISSing.
10:11:26Nico_Pyeah probanly
10:13:03*Nico_P is off
10:13:06 Quit Nico_P (Remote closed the connection)
10:14:43 Quit criznach ("ChatZilla 0.9.79 [Firefox]")
10:15:13JdGordonpondlife: i dont remeber having a rpoblem with that patch.. the soud is better now?
10:15:20psycho_maniacIs the piezo patch ever going to get comitted? just wondering what would be better, 7307 or 5111. Are they two different things?
10:15:42pondlifeTwo different things, although #7307 could make use of #5111
10:16:30pondlife#5111 would improve the sound :)
10:16:49pondlifeAh, I'm wrong
10:17:05pondlife#5111 does the same thing for iPods
10:17:28pondlifeBut I don't have an iPod
10:17:31JdGordonpondlife: your patch has whitespace changes?
10:17:50pondlifeYes, ‎lots of them.. I should remove and commit those seperately
10:18:00pondlifeLots of trailing whitespace removal.
10:18:10pondlifeAnd some less-readable X&Y
10:18:44JdGordonok, testing the patch on my sansa...
10:18:56*GodEater grumles again about no-one commiting his tab->spaces fix up patch again
10:19:18pondlifeJdGordon: Try with and without repeat enable on a wheel device...
10:21:21LinusNCasainho: i'm not a usb guru, but it looks like it should be able to handle otg
10:21:29pondlifeJdGordon: AAAAAAAAHHHHHH!!! ???
10:21:32JdGordonbloody thing just zaped my ear with a beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
10:21:40JdGordonhad to turn of keyclick to stop it
10:22:14pondlifeWeird, I reduced the click length...
10:22:18JdGordonits not so easy to reproduce
10:22:25JdGordonno, its a beep that doesnt stop
10:22:37JdGordononly happens sometimes after pressing a button
10:22:43JdGordon(as in the release event)
10:22:52JdGordonok, just did it on the wheel
10:23:01pondlifeRelease events are always ignored.
10:23:54JdGordonits the high pitch one
10:24:11pondlifeHigh pitched? It's a fixed beep.
10:24:17markunCasainho_: I think it's just "usb device"
10:24:27pondlifepcmbuf_beep(5000, 2, 2500*volume (0-3))
10:24:32JdGordonno, im getting 2 distinct beeps.. a low click and a high pitch squeel
10:24:56JdGordoncould be a e200 thing though
10:25:04pondlifeWith repeat off, or on?
10:25:05JdGordonpressing anything after it happens stops it
10:25:33pondlifeSome kind of timer failure in the beep code?
10:25:56pondlifeMaybe try increasing the length back up to 3
10:26:09pondlifeaction.c line 131
10:26:16 Quit Casainho (Read error: 110 (Connection timed out))
10:26:22pondlifeSearch for pcmbuf_beep
10:26:28JdGordonwhich param?
10:26:29pondlifeChange the 2 to a 3 or 5
10:26:40pondlifeThat's the length
10:26:48Casainho_markun: Thanks.
10:26:50Casainho_I find this info on AVR freaks forum:
10:26:51Casainho_"If you can find a external usb harddisve that supports USB OTG then you could connect it to the UC3 series. Else you could wait for the AP720x series that includes a usb host controller and also runs with 1,8V ram insted of 3,3!!! But are probably not lanched before sometimes in 2008"
10:26:53Casainho_So I think in a few months this device family will have OTG :-)
10:27:11 Quit tedrock (Client Quit)
10:27:20BagderCasainho_: hah, there will of course always be newer hw
10:27:29Bagderyou'll be a fool if you sit waiting for such stuff though
10:27:37Bagderas they could just as well just never appear
10:27:40 Join tedrock [0] (
10:27:41Bagderor not work
10:27:51JdGordonpondlife: changing to 5 didnt change anything
10:27:56JdGordonpossibly made it worse even
10:28:06 Join AceNik [0] (n=AceNik@
10:28:11pondlifeJust for sanity, could you comment that line out?
10:28:20Casainho_well, from Atmel I don't find a processor that fullfills needs...
10:28:20pondlifei.e. ; // pcmbuf_beep
10:28:36BagderCasainho_: then pick something else
10:28:37pondlife(Sorry, I meant ; /* ... */ of course :)
10:28:56JdGordonof course :)
10:29:05JdGordon/ would have given a compile error anyway
10:29:17JdGordonactually.. no it wouldnt
10:29:20JdGordonit wouldnt work though
10:29:40pondlifeYou need the ;
10:30:07JdGordonthats definatly the problem
10:30:20 Part AceNik
10:30:52pondlifeCould you perhaps try some kind of logf to see if it's constantly getting called or whether the beep code is busted?
10:32:44Casainho_Bagder: Okok, I don't know others... - I will have to find... - I know that this AVR32 have free/open Tools to develop and have a lot of community support
10:33:02BagderAVR32 is just the cpu core
10:33:04 Join CaptainSquid [0] (
10:33:26BagderCasainho_: but it's your choice
10:34:05Casainho_in this case the core and the peripherals - USB and DAC
10:34:52 Join spiorf [0] (
10:35:12Bagderthere are many other such SoCs
10:35:20Bagderwith tools, docs and community behind
10:35:27Bagdercoldfire imx31
10:35:29Casainho_okok, I must find them :-)
10:35:34Bagdermarwell PXA series
10:35:41Bagderyada yada
10:35:49Casainho_can you please leave links to them in TWiki page?
10:35:59Bagderthere are about a billion ARM-based
10:36:02Bagderno I can't
10:36:06LinusNCasainho_: remember that we have no AVR32 port of rockbox, so we would have to do all codec optimizations all over again
10:36:08Bagderyou do the work
10:36:30JdGordonpondlife: its not being called repeatedly
10:36:38BagderLinusN: very good point there
10:37:05BagderCasainho_: in other words: go ARM...
10:37:29JdGordonI put a splash in which counts every tme its called to see whats happening (with a timeout of 0) and its working as expected
10:37:37Bagderhm imx31 is not coldfire but freescale, oops ;-)
10:37:40pondlifeJdGordon: What if you turn off keyclick and turn on the track skip beep?
10:37:53Casainho_okok - should go ARM and find one with DAC and USB OTG, free/open dev tools, hand assembled...
10:37:56JdGordontrack skip beep?
10:38:15pondlifeYes, under Playback > Beep Volume
10:38:34pondlifeAnother place where pcmbuf_beep is used.
10:38:48JdGordonno brain damaging beep
10:38:51pondlifeThen play a track and try skipping
10:39:07JdGordonyep, no probs
10:39:11pondlifeAh, does the keyclick work ok during playback?
10:39:29LinusNimx31 is tempting, but i'm afraid it's expensive and afaics only available in BGA packaging
10:39:49Bagderthere we go again... DIY won't work
10:40:04JdGordonpondlife: yes, works fine with music playing
10:40:10 Quit linuxstb ("Leaving")
10:40:13pondlifeBut not when you stop?
10:40:19LinusNCasainho_: it's not *that* important to have all features integrated in the mcu
10:40:27pondlifeOr is it an initial thing only?
10:40:37LinusNCasainho_: but it sure is nice
10:40:41JdGordonno, comes back when music is stopped
10:40:51BagderCasainho_: the tools are the gnu tool chain, they're already here
10:40:51pondlifeThe playback skip beep is 100 long, much more than 2... Try setting the keyclick length to 100?
10:41:10JdGordonthis hasnt shown up on any other target?
10:41:11BagderCasainho_: and we have rockbox already being optimized for ARM so that's already here
10:41:19pondlifeNot on H340 or Gigabeat
10:41:39*markun still things that Blackfin is a cool platform :)
10:41:41pondlifeThe sim doesn't keyclick in lists, because there's no interrupts.
10:41:43JdGordonmaybe its the way data is sent to the dac? maybe it keeps playing the same sample because we arnt tlling it to stop?
10:41:56Casainho_Bagder: so, looks like ARM is a smart choice ;-) - but with DAC and OTG will simple board and soldering, and possible problems in hardware...
10:41:58pondlifeBut that's a more general issue with lists in the sim.
10:42:08Bagdermarkun: well, that's the same situation as AVR32 then with nothing in rockbox optimized for it
10:42:09JdGordonthat can be fixed "easily"
10:42:25Casainho_so I must find a good ARM for this task
10:42:29 Quit foolsh ("Konversation terminated!")
10:42:33pondlifeWell, I put a yield in, but I may have put it in the wrong place...
10:42:47BagderCasainho_: a good that is cheap and not bga and...
10:42:52pondlifeAnyway, keyclick sounds good here on my targets.
10:42:57markunBagder: I know, but there are blackfin opts in speex and ffmpeg (not sure which codecs), maybe other projects as well
10:43:17JdGordonpondlife: im running off in a minute, ill be back in 2 hours or so.. if im still awake ill have a play
10:43:22Bagdermarkun: well, experience tells that the provided optimizes aren't good enough
10:43:53Bagderwe had ARM stuff "optimized" too, but that has been improved vastly by ourselves
10:44:13markunBagder: they are all done by the same guy who's working for analog devices, I hope he know's what he's doing
10:44:44Bagdermarkun: build one and prove it! ;-P
10:45:01jmspeexmarkun: Blackfin is indeed a cool platform, though the assembly is a bit "contrained"
10:45:05markunBagder: I could also buy one of those chinese gpl violations :)
10:45:18*amiconn thinks coldfire would be a way better choice than arm
10:45:32amiconnOf course, make sure to select one with data cache
10:45:54markunjmspeex: that's a pitty. Do you know what speedups you get with the speex opts?
10:46:19Bagderamiconn: yeah, but coldfire is a dying breed I think
10:46:32jmspeexmarkun: you mean how much I gain by using Blackfin assembly on Speex? Quite a bit actually. And there's still room for improvement
10:47:44jmspeexmarkun: Main problems for Speex on Blackfin atm are 1) gcc code quality and 2) instruction cache
10:48:00markunjmspeex: what about the instruction cache?
10:48:20LinusNblackfin is also quite expensive and only come in bga packaging as well...
10:48:40markunLinusN: why do you have to spoil everything :)
10:48:43GodEaterslightly OT, but possibly not - anyone interested in the OpenMoko project ? (
10:49:07 Join CaptainSquid83 [0] (
10:49:16LinusNwell, a few of them are in fact available in qfp, but not the ones that are particularly interesting to us advised me to go with the 525c when it's available to the public if I was interested in designing a DAP
10:50:17jmspeexmarkun: basically, the Blackfin (of at least the 535) has 16 kB instruction cache plus 32 kB instruction SRAM.
10:50:43markunjmspeex: what's the problem with that?
10:50:44amiconnI doubt one will find a decent cpu for an mp3 player nowadays that's not bga
10:50:47jmspeexUnless you explicitly put the code in I SRAM, you don't have much available.
10:51:04markunjmspeex: the S3C2440 in the Gigabeat also has 16kB
10:51:07LinusNmarkun: integrated codec...hmmm
10:51:11jmspeexIt's not a problem with Blackfin itself, just with the current Speex port, which doesn't do that
10:51:11*GodEater thinks "Rockbox as app" would be an awesome thing to add to the OpenMoko phone
10:51:25markunLinusN: I'd prefer an external one..
10:51:28LinusN"289-ball, 12x12 mm, 0.5 mm pitch mini-BGA" gah!
10:51:43jmspeexmarkun: The nice thing with Blackfin is that ADI is fully backing OSS
10:51:47markunLinusN: we can probably have them assembled in china ;)
10:51:59markunjmspeex: yes, that's also what I like about them
10:52:32 Join Xerion [0] (i=xerion@
10:52:39 Quit CaptainSquid (Remote closed the connection)
10:53:11 Join JazzBone [0] (
10:54:27LinusNamiconn: you are probably right
10:59:13amiconnThere's the MCF5372 (coldfire v3, 180MHz) which is QFP
10:59:17 Join miepchen^schlaf [0] (
10:59:47amiconnBut that doesn't have integrated USB; the MCF537x variants with integrated USB are MAPBGA (and go up to 240MHz)
11:03:20Bagdermarkun: manufacturing stuff in China isn't as easy as it sometimes sounds, and it does has a hefty initial cost too
11:04:17markunBagder: but you made the right contacts over there, right? :)
11:04:25Bagderhaha, oooh yes
11:04:32BagderI have my own secret mp3 factory now
11:04:51psycho_maniacBagder: may i join your seceret club? ;)
11:05:06Bagdersure, but you must promise to be very very secret
11:06:17psycho_maniacmums the word
11:07:50***Saving seen data "./dancer.seen"
11:10:07 Quit Xerion_ (Read error: 110 (Connection timed out))
11:11:13psycho_maniacdo you think that when a "rockbox player" is created there would be instructions on how to make it yourself and posibably some people willing to make it for the user?
11:12:28BagderCasainho_ plans for the first part, and I guess if there's demand there's room for the second too
11:15:46Casainho_psycho_maniac: It would be nice to buy one already assembly and in plastic case, and with RockBOc loaded :-)
11:16:10 Quit goffa_ (Read error: 110 (Connection timed out))
11:19:02pondlifeAny iPod users able to try ?
11:20:12psycho_maniacwould you like it tested with just that patch ?
11:20:35pondlifeYes please, if possible
11:20:48pondlifeThere's a good chance it will just make a horrible noise
11:21:07psycho_maniacheh. thanks for the warning.
11:21:21pondlife"brain-frying" was the term I think :)
11:21:39psycho_maniacwhat kinda patch is this? -px?
11:30:12GodEaterpondlife: building now for ipod video
11:30:23pondlifepsycho_maniac: -p0
11:30:33pondlifeGodEater: Thanks
11:30:52psycho_maniacdo you need 2 testers?
11:32:33*GodEater is a little peeved that for whatever reason "git apply" doesn't appear to work
11:35:01Bagder... no sansa view available in Sweden yet
11:36:35GodEaterer - pondlife, what's it supposed to do ?
11:36:43GodEaterI've set keyclick to "Strong"
11:36:51GodEaterand keyclick repeat to yes
11:36:58GodEaterand I can't hear anything
11:37:02pondlifeIt's meant to make a noise whenever you press a key
11:37:11GodEaterit's doing bugger all
11:37:24pondlifeWell, better than brain-frying :)
11:37:32GodEaterI guess so ;)
11:37:33pondlifeTry playing something
11:37:45pondlifeThen navigate during playback
11:37:46GodEaterah hang on
11:37:54GodEaterdoes it do the same as voice ?
11:37:58GodEateri.e. not work during pause ?
11:38:05pondlifeAll output stops during pause
11:38:08GodEaterah ok
11:38:11GodEaterthat works a bit better
11:38:12pondlifeBut that's a job for another fix
11:38:16GodEaterit's not, erm, pleasant though
11:38:26pondlifeIn what way?
11:38:31pondlifeShould just be a chk-chk
11:38:40GodEaterthe click isn't "clean"
11:38:47pondlifeIt's a very short beep
11:38:51GodEaterand occasionally I get a "beep"
11:39:11GodEaterokay, now I'm getting the brain frying thing
11:39:17pondlifeHmm, sounds like the beep timer isn't good on iPod Video too..
11:39:32pondlifeCan you describe the noise better?
11:39:45GodEatera beep that just doesn't stop
11:39:54GodEaterit just goes "eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee" a lot
11:40:00pondlifeThat's the one!
11:40:03GodEateryou have to start playback
11:40:06GodEaterto make it go away
11:40:18pondlifeCould you modify action.c and increase the click duration for me...just one number to change
11:40:37GodEatershould it be a click or a beep?
11:40:38pondlifeline 131, I think it was... see pcmbuf_beep(.., 2, ...)
11:40:40GodEaterI'm unclear
11:40:47pondlifeShould be a very short beep, sounds like a click here
11:41:09pondlifeBut it seems short lengths don't work well on some targets
11:41:27pondlifeIn action.c line 131, the 2 is the length
11:41:28GodEatera beep and a click are two distinct sounds to me
11:41:34GodEaterso I'm not clear on what you're telling me
11:42:04pondlifeI guess a click would be white noise, or a simple square on/off
11:42:15pondlifeFor now it should be a beep, but very very short
11:42:49pondlifeCould you just try increasing the beep length as an experiment
11:42:58GodEaterinitial key presses produce what I would describe as a click
11:43:13GodEaterif you turn off keyclick repeat, they stay that way
11:43:39GodEaterhowever, turning the repeat on and then scrolling the wheel produces lots of clicks, and then suddenly the beep that won't go away
11:44:13GodEaterany idea on what you want me to change the length to ?
11:44:18GodEater5? 10? 3?
11:47:01pondlifeEither it doesn't cope with short lengths or overlapping beeps.
11:47:32GodEaterin any case, the click isn't pleasant to listen to
11:47:38GodEaterimo anyway
11:47:56 Join Bagder_ [0] (
11:49:05 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
11:49:26 Join miepchen^schlaf [0] (
11:50:38GodEaterI'm getting a "proper" beep now with the length set to 100
11:50:44GodEaterbut it's still not "clean"
11:50:45pondlifeNo eeeeeeeeeeeeeeeeeee ?
11:50:49GodEaterthere's static at the end of the beep
11:50:58GodEaterhold on, I have to set repeat to on
11:51:09GodEaterno - I still get the eeeeeeeeeee
11:51:12 Quit weezerle ("...und tschüss!")
11:51:19GodEaterand this was before I even got to change the setting :(
11:51:43pondlifeOK, don't worry
11:51:44GodEaterit's also a completely different pitch to the keyclick beep
11:51:46GodEatermuch much higher
11:51:50 Join FOAD_ [0] (
11:52:35GodEaterit seems to be caused when two beeps are generated close together
11:52:43pondlifeSomething isn't happy in pcmbuf_beep with overlaps.
11:52:49GodEaterso it's easier to reproduce when the repeat is on
11:52:55GodEaterbut possible without it too
11:53:09pondlifeThat makes sense.
11:53:40pondlifeNot really a bug in the keyclick stuff, but an underlying issue which it reveals I suspect. Of course, keyclick is useless like this!
11:55:23 Join goffa [0] (n=goffa@
11:55:47GodEateryep :(
11:55:47pondlifeNot sure I can fix any further without a target that has the problem..
11:56:21amiconnHow long is one PCM buffer?
11:56:23pondlifeWas it easier to get eeeeeeeeeeeeeeeeee with a long or short duration?
11:56:33GodEatermade no difference
11:56:54pondlifeBut it works fine during playback
11:56:57GodEateris it possible to clear the pcm buffer before attempting to make a new click ?
11:57:06 Join agm3nt [0] (
11:57:11pondlifeThat would be bad
11:57:44pondlifeWe need to work towards PCM being a general output stage, running even when paused etc.
11:57:56amiconnThat's why I'm asking
11:59:15pondlifeamiconn: I'm not sure
11:59:21pondlifeI'm just relying on pcmbuf_beep working
11:59:37amiconnRight now 'Pause' halts PCM completely. It could just halt the codec thread feeding the pcm buffer, and then discard all unplayed pcm blocks
12:00:45amiconnIt requires that one pcm block is short enough that playing it up to the end isn't noticeable
12:01:10 Join TMM [0] (
12:01:28pondlifeI see why we get an eeeeeeeee and static...
12:01:29amiconnOn, and the codec needs to know how much to go back when resuming
12:01:41pondlifeIt decides to play 0 samples...
12:01:50pondlifeWhich is messy
12:03:54amiconnHmm, or the pcm api could allow to retain some block, while still allowing to play others
12:03:56 Quit J3TC- (".UPP.")
12:04:16*amiconn wonders how difficult it would be to implement late-mixing for voice, beeps etc
12:04:35 Quit Bagder (Read error: 110 (Connection timed out))
12:04:49 Nick Bagder_ is now known as Bagder (
12:05:04 Quit psycho_maniac (" HydraIRC -> <- Nine out of ten l33t h4x0rz prefer it")
12:05:55pondlifejhMikeS was looking into this a while back.
12:06:16jhMikeSamiconn: without preemtive threading it won't be very low latency....what's acceptable? now's the time to think anyway with this voice conversion.
12:07:17pondlifeGodEater: You able to try one more little mod? If so, pcmbuf.c line 1001...
12:07:54pondlifeBefore the call to pcm_play_data, insert if (samples != 0)
12:08:45 Part JazzBone
12:08:49 Quit FOAD (Read error: 110 (Connection timed out))
12:08:49 Nick FOAD_ is now known as FOAD (
12:09:02*GodEater runs make
12:09:19pondlifeHmm, why does is there a different square-wave generator for playing and not-playing...
12:09:24jhMikeSpondlife: oh, now you have to make this task a priority? I have to do work there. What I'm seeing is nightmarish especially re: what happens against the interrupt.
12:09:40pondlifeThat made little sense... (my sentence I mean)
12:09:44pondlifejhMikeS: ;)
12:09:56pondlifeBlame bascule for giving me a Gigabeat.
12:10:06pondlifeThat touch cross is really not so easy to use
12:10:11*GodEater just blames bascule generally
12:10:23jhMikeSfor a gigabeat why not implement the TIMER3 beep and forget the pcm buffer for beeping there?
12:10:42pondlifeI'd like this to be for all SWCODEC
12:10:59pondlifeIt was useful when driving and fumbling for my H340 too
12:11:11pondlifeBlind users may like it too.
12:11:13jhMikeSSure, but when the hardware is available...
12:11:21pondlifeBesides, pcmbuf_beep ought to work.
12:13:03jhMikeSit really can't too well. we need a mixer thread.
12:13:07GodEaterthat's much better pondlife
12:13:22*GodEater now goes to change the length back to 2
12:13:24*preglow freezes to death
12:13:33pondlifeSomething is wrong in pcm_play_data then... :(
12:13:51pondlifeOn iPod/Sansa
12:14:07pondlifepcm_play_data shouldn't do much if told to play 0 samples.
12:14:13pondlifeBut obviously it is.
12:14:17 Join J3TC- [0] ( should be fine
12:14:40pondlifeI guess it's starting DMA and outputting a very long block
12:15:02pondlifejhMikeS: Apparently if passed a 0 length and non-null pointer, it makes a brain-frying noise...
12:15:17pondlifeOn Sansa and iPod, but not on H300 or Gigabeat
12:15:31jhMikeS:D as it should ... dirty rascal thing to do.
12:16:01GodEaterdarn portalplayer
12:16:06pondlifeWhy Rockbox? None of the competition has "brain frying" built in !!
12:16:09GodEateryep - I like that now anyway pondlife
12:16:25pondlifeLength 2 sounds like a click
12:16:29 Nick parafin|away is now known as parafin (
12:16:31 Join weezerle [0] (
12:16:31GodEaterI'm still get beeps
12:16:37GodEaternot never ending ones now
12:16:49GodEaterjust when scrolling a long list every 10th "click" is a beerp
12:16:57 Join barrywardell [0] (
12:17:02pondlifemmm, beerp
12:17:13jhMikeSif it's passed 0, it will consult the callback. if the callback returns something, it will play it.
12:17:23GodEaternot quite every 10th
12:17:25GodEaterit's more random
12:17:37pondlifeIs this only when playing, or only when not playing, or both?
12:18:56jhMikeSdata running out will NULL the callback as will pcm_play_stop so it shouldn't try playing anything
12:19:29pondlifejhMikeS: It sounds ok on some targets.. so the problem may be quite far "down"...
12:19:51preglowjhMikeS: i've got voice file building going now, so do tell if you need any
12:20:05jhMikeSpreglow: me up
12:20:12preglowaight, gimme a sec
12:20:16GodEaterpondlife: both
12:20:35*GodEater wonders how this patch will behave with 7738
12:20:42preglownext thing would be building rbspeexenc from the rockbox speex sources
12:21:08jhMikeSpondlife: I don't thinks it should be far down at all. pcm_play_data checks its params and the start must be non-NULL, size must be nonzero
12:22:06pondlifeHmm, the only thing GodEater changed was to NOT call pcm_play_data if the size was zero...
12:22:20pondlifeThat fixed both a long noise and static
12:22:38preglowjhMikeS: do you want a file or the sources, btw?
12:22:41pondlifeGodEater: I suspect you're hearing beeps colliding.
12:22:53GodEaterpondlife: I take it back
12:22:55pondlifeA long beep is just two clicks end-to-end
12:23:01GodEaterit's not possible to get beeps whilst playing
12:23:04GodEateronly when stopped
12:23:21pondlifeOK... I have an idea on that.
12:23:46GodEaterwhat is the "2" measured in ?
12:23:49jhMikeSpreglow: just give me a made file that I can just use
12:24:02pondlifeGodEater: ms I think
12:24:11GodEaterso two beeps colliding would be 4ms ?
12:24:21 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:24:24GodEaterthe beep is much longer than that
12:24:30pondlifesamples = NATIVE_FREQUENCY / 1000 * duration
12:24:38GodEaterI would say as long as 500ms
12:25:02preglowjhMikeS: target?
12:25:26jhMikeSpreglow: well, what's needed in order to build them anyway?
12:25:39jhMikeSI could want quite a few targets
12:25:42preglowjhMikeS: nothing much apart from a configure patch and rbspeexenc
12:27:01preglowrbspeexenc is one file which needs to be linked to svn libspeex
12:27:04jhMikeSpreglow: I've never made voice from the .lang files ever so something is needed to produce the WAV output
12:27:10pondlifeGodEater: Maybe comment out pcmbuf.c line 955...
12:27:16preglowjhMikeS: oh yeah, easpeak, flite, festival
12:27:22preglowjhMikeS: i had all of those installed by default, for some reason
12:27:30preglowjhMikeS: if you're on windows, you can use the native stuff too
12:27:31jhMikeSfree or fee?
12:27:35pondlifeGodEater: The line ending in pcmbuf_mix_buf[NATIVE_FREQUENCY * 4 / 8]
12:27:54pondlifeThen see if the beeps also occur during playback!
12:28:03jhMikeSpondlife: look in pcm.c and see why passing 0 and getting playback makes no sense at all
12:28:47preglowjhMikeS: i plan on having a proper patch requiring no user intervention for rbspeexenc going soon anyway, i'd be happy to cough up a couple of voice files for you in the meantime
12:29:09pondlifejhMikeS: Could pcm_play_lock() have a side-effect?
12:30:21pondlifeI see what you mean, but we have some strong empirical evidence... :/
12:30:30*pondlife must buy a PP target
12:30:38jhMikeSpondlife: I really doubt it but it could be pcm_play_dma_stop that does...hmmm
12:31:54jhMikeSpreglow: H10, e200, gigabeat, H100, x5 :P maybe that's too much?
12:32:04preglowgimme a sec
12:32:20pondlifejhMikeS: The comment in pcm-pp.c ...dual core support will require other measures... isn't relevant yet?
12:32:34preglowthere's no guarantee the voice files are right, btw, i think they should be, since i can playback what seems to be put in
12:32:42pondlifeI mean, disabling FIQ prevents a swap?
12:33:32jhMikeSpondlife: no, not relevant yet. and a swap?
12:34:08pondlifeWrong terminology, I meant something on the second core..
12:34:13GodEaterpondlife: sorry - got work to do now - won't be able to help much more today :(
12:34:15jhMikeSdisabling FIQ prevents the interrupt for asking for more data while the specific interrupt is masked
12:34:17pondlifeOK, thanks
12:34:42barrywardelljhMikeS: I'd like to have a look at improving the H10 scrollpad. I think you mentioned before that you had thought about it. Any suggestions?
12:35:46 Quit J3TC- (Read error: 110 (Connection timed out))
12:37:35preglowespeak's pronounciation is actually better than flites
12:37:45preglowtoo bad about that synthesizer...
12:37:47jhMikeSbarrywardell: on what level? particulars in the driver?
12:39:32jhMikeSI thought that basically it should work with time-based acceleration when pressed like buttons and like a wheel when wiped with the velocity-based acceleration.
12:40:29 Quit HellDragon (Connection timed out)
12:40:56markunpreglow: it can be used with the mbrola synthesizer
12:41:04jhMikeSpondlife: where is pcm_play_data being called with 0 anyway?
12:41:05 Join moos [0] (
12:41:11preglowjhMikeS: i'll just use different synthesizers for all of them
12:41:47preglowmarkun: how?
12:42:01markunpreglow: do you use linux?
12:42:55pondlifejhMikeS: Somehow in pcmbuf_beep...
12:43:19pondlifeI'm not sure how :/
12:43:24 Join FOAD_ [0] (
12:43:46jhMikeSsomehow? it should only attempt to start playback in that case if the callback is also not null
12:44:05pondlifeThe callback is definitely null.
12:44:29pondlifeSo, it basically locks and unlocks, that's all?
12:44:55markunpreglow: should sound a lot better
12:44:56jhMikeSyep...if playback was started, the interrupt is unmasked, if not it stays masked
12:45:11pondlifeThis is only a problem if playback wasn't started
12:45:27pondlifeIf playback was started, no issue.
12:45:56amiconnpondlife: You want button click on H300??
12:46:08*amiconn wonders why
12:46:21pondlifeBecause it makes unsighted operation easier
12:46:29amiconnThe big advantage of a real button target like the H300 is that you get tactile feedback
12:46:35pondlifeMaybe my NAVI button is duff
12:46:38pondlifeIt hardly moves
12:46:42jhMikeSpondlife: you mean never started or if it's playing currently?
12:46:42amiconnSo no click necessary for that
12:47:06amiconnI would like a click for the dreaded ipod wheels, and the H10 slider, but not for buttons
12:47:24pondlifejhMikeS: If it's never started I believe, but only GodEater and JdGordon have tried it.
12:47:43pondlifeCertainly after pressing STOP
12:48:09preglowjhMikeS: in case i'm wasting my time, i uploaded three of them for now, lemme know if they actually work and i'll supply more :)*
12:48:22pondlifeamiconn: It's certainly needed on the Gigabeat too.
12:48:48preglowamiconn: i've removed the entire codec choice thing in configure, it didn't work anyway
12:48:49jhMikeSpreglow: might be a bit before I can even test them
12:48:57preglowjhMikeS: sure
12:50:01 Quit Casainho_ ("ChatZilla 0.9.79 [Firefox]")
12:50:30jhMikeSpondlife: it's till making no sense...pcm_play_data should set the cb to NULL, in which case pcm_play_data_start should simply drop through to make sure things are stopped.
12:52:19pondlifeI agree
12:52:37pondlifeBut, based on GodEater's report...
12:52:43 Quit goffa (Remote closed the connection)
12:52:46pondlifeDo you have a PP target handy?
12:52:47 Join goffa [0] (n=goffa@
12:52:54jhMikeSit's an effect from something else I think
12:53:15pondlifeShould be easy to repro with keyclick_3.patch
12:53:21jhMikeSwhat if you beep, and then try another beep while beeping still? it will enter the wrong case
12:53:51pondlifeThe second beep is during PCM playback, right?
12:54:31jhMikeSif the beep is still playing, pcm_is_playing can be what happened to pcmbuf_read...maybe it's not null
12:54:59 Quit FOAD (Read error: 110 (Connection timed out))
12:55:00 Nick FOAD_ is now known as FOAD (
12:55:08 Nick fxb__ is now known as fxb (
12:55:14jhMikeSwhich FS is keyclick_3.patch?
12:55:40jhMikeSpreglow: thanks for the voice files btw
12:55:56preglowyou can build them yourself when i manage to get rbspeexenc building nicely :)
12:56:06pondlifejhMikeS: #7307
12:56:23amiconnpreglow: I presume voicefont.c didn't need adjustment?
12:56:27pondlifekeyclick_3 allegedly shows the problem, yet keyclick_4 doesn't
12:56:43preglowamiconn: i haven't even opened it in an editor yet, i just assumed it took what it got as long as it was named .mp3
12:57:11jhMikeSpreglow: how much data does speex_decode_int decode? one frame?
12:57:31preglowjhMikeS: exactly one frame of 160, 320 or 640 samples, depending on mode
12:57:50preglowand eats from the bitstream the bits it needs
12:58:40jhMikeSincremental on small packets like that is the way to go here
12:59:15jhMikeSat least then I can choose how many per loop
13:01:51 Join darkapostrophe [0] (
13:02:14preglowthat speex decoder i sent should be exactly the way to decode these speex clips
13:02:20preglowyou point at them with speex_bits_set_bit_buffer
13:02:34preglowthen just call speex_decode_int until no bytes (or just a couple) remain
13:02:58preglowno BITS
13:03:21 Join Casainho [0] (
13:03:29preglowbtw, i updated our speex.c yesterday with some new stuff you should consider too
13:03:42preglowone should always skip the first couple of samples coming out from the decoder
13:03:46jhMikeSI suppose whatever the mp3 callback returns...set the buffer to
13:03:49preglowoh, and yeah, remember to reset the encoder for each clip
13:04:14preglowyeah, just basically do a speex_decoder_init
13:04:32jhMikeSwhat about the ENH stuff?
13:04:48preglowenabled by default
13:04:58jhMikeSso I don't need to call that?
13:07:54***Saving seen data "./dancer.seen"
13:10:04amiconnpreglow: Yes it does. Of course we should change the extension for the speex clips
13:10:16amiconn.spx usually means an ogg containter?
13:11:36preglowamiconn: always
13:11:48preglowwe should not use .spx
13:11:53preglow.rspx or something
13:12:03amiconnOr .speex
13:12:12 Quit Rob222241 (Read error: 113 (No route to host))
13:12:26pondlifeDoes .ogg imply an ogg container, but not necessarily Vorbis?
13:12:32preglowi want it clear that these cannot be expected to play anywhere else
13:12:45 Join slomo12 [0] (
13:12:50preglowpondlife: that's what xiph wanted to happen, but even they have stopped believing it can be done now
13:12:56preglow.ogg today basically means vorbis
13:12:57pondlifepreglow: .talk?
13:13:18pondlifeIt would make more sense if the extension only related to the container.
13:13:20jhMikeSthe fracking pcm buffer design is the biggest block here or I'd be basically finished here
13:13:23slomo12howdy :)
13:13:36preglowpondlife: more sense, but how do you use separate programs for audio and video then?
13:14:19pondlifeLaunch the program first :)
13:14:23amiconnThe culprit is that windows associates extensions rather than file formats
13:14:33preglowuntil the os supports such distinctions, it's hopeless
13:14:41amiconn(and rockbox also does because it's the simplest method)
13:14:43preglowi believe linux is a lot closer to that than windows, which cares only about extensions
13:14:58slomo12I have an ipod, 3g and I believe it's the one with encryption, is it possible to wipe out apple OS and install ipodlinux and still be useable ?
13:15:00preglowamiconn: rockbox does support different things in an .ogg
13:15:03amiconnAmigaOS has a nice datatypes concept
13:15:27pondlifeShould be possible to open a video file in any audio app and just play with the audio streams...
13:15:42pondlifeAssuming there's audio, of course
13:15:47 Join goffa_ [0] (n=goffa@
13:17:33preglowpondlife: that's not usually whati want when i double click a video file...
13:17:50jhMikeSto heck with it, I'm not bending over backwards for that code's sake
13:18:54 Quit Dark_Apostrophe (Read error: 110 (Connection timed out))
13:18:54 Nick darkapostrophe is now known as Dark_Apostrophe (
13:19:32 Quit ___jake (Remote closed the connection)
13:21:13 Join Sedgewick [0] (n=Sedgewic@
13:22:15 Part slomo12
13:22:21 Join J3TC- [0] (
13:26:38 Quit goffa (Read error: 110 (Connection timed out))
13:30:27jhMikeSpreglow: will speex_bits_remaining continue to return the same value until more data is consumed?
13:33:26preglowwell, yeah
13:33:34preglowif no data is consumed, it'll say it's got the same amount of data
13:35:56jhMikeSjust making sure. it might need to process a message and go back to what it was doing before.
13:39:33 Join Isolinear [0] (
13:40:12LinusNspeaking of pcm playback, how do we really want it implemented?
13:40:39LinusNi'm thinking a mixer
13:41:11 Join goffa [0] (n=goffa@
13:41:12pondlifeYes. Voice, Playback, Beeps, Line In....
13:41:16LinusNwith at least 3 inputs; playback, voice and "aux" (for example beeps)
13:42:14LinusNseparate volume for each channel, and a master volume as well
13:42:41pondlifeWhich channels need EQ and other DSP? Only playback?
13:42:57jhMikeSmixer, mixer, mixer :)
13:43:00LinusNto keep it simple
13:43:04pondlifeVoice, beep can be mono too
13:43:23pondlifeNo need for send loops :)
13:43:26LinusNi also want to accomodate for sound in games
13:43:38LinusNbut that could probably be the playback channel
13:43:52pondlifeI'd think you'd want to mix music + game sound
13:44:00pondlifeSo make games use aux
13:44:05jhMikeSthis is pretty much what I had in mind but needs a thread to run it
13:44:24preglowand preemptive threading! \o/
13:44:25pondlifeCan't be DMA run?
13:44:26LinusNjhMikeS: a thread, or a periodic interrupt
13:44:42pondlifeI suppose you need to double-buffer..
13:44:52jhMikeScan't be DMA run
13:45:00LinusNthe thing is that we want very low latency as well
13:45:14jhMikeSpreemptive threads then :)
13:45:25*LinusN stops listening :-)
13:45:36LinusNwhy a thread?
13:45:44pondlifeA regular interrupt would do, no?
13:45:49jhMikeSat least that thread could also be preemted
13:46:12jhMikeSif it's in an interrupt, it holds everything down until finished
13:46:49pondlifeIt's fairly critical that it keeps running...
13:46:49LinusNbut it should still do the work in very small chunks
13:47:12pondlifeA bit of summing should be fairly predictable.
13:47:14LinusNto keep the latency at a minimum
13:49:25LinusNit *could* be run by the dma interrupt, but then we have a problem when there is no dma
13:49:53preglowwe'll also have major trouble if we use too much time in it
13:49:56LinusNwe don't want to play silence just to keep the dma interrupt going
13:50:04pondlifeWhy not?
13:50:08 Join JETC- [0] (
13:50:21*preglow has already experimented with hogging the cpu in the dma interrupt...
13:50:25 Join HellDragon [0] (i=jd@unaffiliated/helldragon)
13:50:28*jhMikeS too
13:50:53pondlifeI mean, why not keep DMA running all the time.
13:51:00preglowand i got the most lovable little frozen player i've seen
13:51:16LinusNpondlife: maybe that's an option, but that would mean less cpu sleep
13:51:35preglowso, all in all preemptive threading with a realtime thread, then :>
13:51:39jhMikeSpondlife: all the time? silence might be useful to prevent a stop if DMA runs out and an insert if about to happen but other than that...
13:52:04jhMikeSwe could get a few ms latency with that
13:52:23pondlifeDoes DMA running itself use CPU...
13:52:29pondlifeThe mixing can stop..
13:52:33LinusNpondlife: no
13:52:43pondlifeJust keep sending the silent buffe
13:53:02jhMikeSmpegplayer does that to keep the clock running
13:53:02pondlifeMight be simpler, as long as battery life won't be reduced.
13:53:03LinusNwe still need to serve the dma interrupt to send a new buffer to the dma controller
13:53:21pondlifeThat could be very little work.
13:53:26 Join n1s [0] (
13:53:39LinusNthe benefit would be that the dma interrupt could do the mixing
13:53:57preglowwhat about dsp?
13:53:58pondlifeThe mixer should know what inputs are active and be able to optimise lots.
13:54:01jhMikeSbut won't work where you can't run the interrupt in parallel with a big transfer
13:54:02preglowit we only do mixing, the dsp will still be low-latency
13:54:03 Quit J3TC- (Read error: 110 (Connection timed out))
13:54:04 Nick JETC- is now known as J3TC- (
13:54:08 Quit goffa_ (Read error: 110 (Connection timed out))
13:54:24LinusNpreglow: yes
13:54:25pondlifedsp is outside of this, on the playback stream, in my model.
13:54:28jhMikeSyeah, and ll doubles memory needs...crossfade...ouch
13:54:44pondlifecrossfade is also outside of this
13:55:04pondlifeThis is a "hardware mixer emulator" only
13:55:06preglowpondlife: ok, so this does nothing to make dsp low-latency
13:55:14LinusNwould we want 2 playback channels and let the mixer do the xfade?
13:55:21preglowwe'll still have two second delays when adjusting eq bands
13:55:31jhMikeSeven so, pp can't do it
13:56:17LinusNas i see it, we have a few "problems" we want to solve:
13:56:27LinusN1) no sound when music is paused
13:56:39LinusN2) high latency in general
13:56:54pondlife3) Over complex PCM code?
13:56:58LinusN3) insufficient mixing between playback/voice/beep
13:57:27amiconnThe high latency cannot be solved completely without giving up our boosting concept
13:57:45 Join Thundercloud [0] (
13:57:45LinusNamiconn: probably not
13:57:59jhMikeSthat one too...another block I had in thinking about it
13:58:07amiconnWell, dsp latency could, if we put the dsp later into the chain
13:58:14pondlifeLinusN: Does (3) refer to the issue where fades affect voice/beeps too?
13:58:20amiconnBut then the pcm buffers would need to be 32 bit
13:58:27pondlifeOne step at a time, I reckon
13:58:32preglowamiconn: i don't see that as a problem...
13:58:43amiconnpreglow: low-mem?
13:58:48amiconniFO, logikdax...
13:58:57jhMikeSonly the main 2 codec outputs would need 32-bit
13:59:20amiconnwhy 2?
13:59:29 Join mf0102 [0] (n=michi@
13:59:59amiconnCouldn't crossfade be done *in* the pcm buffer?
14:00:12pondlifeCrossfade should be pre-dsp
14:00:27jhMikeSit most definitely shoudn't
14:00:44pondlifeDepends on the dsp, really.
14:00:46jhMikeSit should be the same enveloping mechanism for fade on stop/pause, crossfade, etc.
14:01:13jhMikeSsince you're doing the mixing, might as well apply volume settings too
14:01:17pondlifeWhy DSP(1)+DSP(2) when you can DSP(1+2)
14:01:56pondlifeEspecially if you include dynamic compression or similar..
14:02:10*amiconn would rather drop all audio fading, but not many users might agree with that
14:02:40pengoi listen to podcasts and fading is very annoying
14:02:41*jhMikeS hates the crossfade most of all
14:02:51pondlifeI love crossfade
14:03:07pengopondlife: because you want to hear every word being said before and after pausing.
14:03:09jhMikeSit's an annoying programming impediment that wastes tons of ram :p
14:03:41pengoi'd like a "podcast mode" with sane defaults
14:03:42pondlifeCrossfade on skip+shuffle is probably my top Rockbox feature.
14:03:55amiconnThe 'fade on stop/pause' was about the first option I ever disabled in rockbox
14:04:07amiconnThat was on archos recorder v1, rockbox 2.0 iirc
14:04:09pondlifepengo: The options are there..
14:04:27*JdGordon back
14:04:36 Join foolsh [0] (
14:04:51LinusNi believe the crossfade is a popular feature
14:04:58pengopondlife: is there an option to remember you spot when you accidently hit "next track" 30 minutes into a 1 hour track?
14:04:59jhMikeSwith a ll mixer, fade on stop pause would really be quite simple to dsp instead of volume fade. other channels would each play at their own volumes.
14:05:28pondlifepengo: Not as far as I know, but bookmarking isn't something I know much about...
14:05:29*jhMikeS says "nuke the crossfade" and put it under "things we'll never implement...again"
14:05:41*pondlife goes back to OF
14:05:45pengohaha jhMikeS
14:06:39 Quit foolsh (Remote closed the connection)
14:07:01LinusNbtw, the dsp block should be able to stretch the audio for tempo adjustment
14:07:23LinusNwhich in turn means DSP(1) + DSP(2)
14:07:23pondlifeIs someone drawing a nice block diagram ?
14:07:27JdGordonpondlife: latest patch doesnt fix the problem
14:07:46pondlifeThat means that GodEater must have done something else
14:08:12pondlifeJdGordon: So even with that pcmbuf_beep mod, you still fry brains?
14:08:53GodEaterpondlife: didn't - only did the two changes you asked me to make
14:09:04GodEaterinitially the beep length
14:09:09GodEaterand then the samples != 0
14:09:20pondlifeOK, but the samples != 0 fixed the eeeeeeeeeeeeeeeee , yes?
14:09:38*GodEater has to get food, but will be back in 15 mins
14:09:39 Part agm3nt
14:09:40pondlifeNot for JdGordon on Sansa, sadly
14:09:46*jhMikeS would much prefer to write "How to Add a 2nd DAC to You Player to Use Crossfade" hardware mod
14:10:21pengoi'd like a two-less-menu-buttons mod
14:11:05 Join Rob2222 [0] (
14:11:20*jhMikeS must investigate the real problem with this eeeeeeeeeeeeeeeeee thing soon
14:11:33preglowafter speex voice ui is done :>
14:11:57jhMikeSI guess that one-liner will suffice for now
14:12:38jhMikeSyes...and what to do with that pesky pcm buffer to make it clean :<
14:14:20jhMikeSwell, I'd rather put the "make it work for now" hacks in the old code anyway...I'll accept none in the new and I'll proceed as though pcmbuffer is the perfect mixer
14:14:55amiconnLinusN: Speaking about that - what happened to your on-the-fly mp3 stretching idea?
14:15:08LinusNamiconn: i kind of lost interest :-)
14:15:11amiconnAnd also, did you manage to get the oldplayer lcd working again?
14:15:20LinusNamiconn: haven't tried
14:15:49*pondlife has heard lots of time-stretching algorithms but none of them sound good enough for foreground audio...
14:16:09pondlifeNot as good as speed changing, certainly.
14:16:17LinusNi think it can be good enough for speech
14:16:31jhMikeSmp3 stretching as in time-streching compressed mp3 data? say what?
14:16:44pondlifeAlso time squishing always sounds better than stretching.
14:16:53pondlifei.e. throwing data away, not doubling.
14:17:03preglowbtw, i think maybe time stretching might be one of my next projects
14:17:14 Join atsea- [0] (i=atsea-@gateway/tor/x-37086b963b619613)
14:17:20jhMikeSstretching needs to magically create data that doesn't exist
14:17:23JdGordongot your flux capacitors in stock?
14:17:37LinusNjhMikeS: a poor man's stretch/compress which removes/doubles mp3 frames
14:17:45pengopreglow: please please try time stretching
14:17:51amiconnjhMikeS: Convert the mp3 stream to 320kbps cbr without bit reservoir, and then drop or replicate frames
14:18:15pondlifeI take it people want to time-squish rather than (or more than) time-stretch?
14:18:22pondlifei.e. speed up book reading
14:18:26*jhMikeS should have gotten the flux capacitors at low-low prices at that one wootoff
14:18:26pengopondlife: exactly
14:18:29amiconn(yet another bresenham application)
14:18:48LinusNpondlife: some people want to slow down music for rehearsal with their music instrument
14:18:56pondlifeAh, yes.
14:19:18pengoi've started to got used to listening to chipmunk voices though... after finally finding "pitch" under a menu other than "sound settings"
14:19:19pondlifeA-B repeat with time stretch
14:19:40amiconnAfaik, archos did audible seek that way on the hwcodec targets
14:19:42pengopondlife: might want to slow audio if learning a new language too
14:19:45LinusNpondlife: a killer app
14:20:06preglowLinusN: think i'll be using that lib you pointed me to for a quick attempt
14:20:11*pondlife can't remember the Archos OF any more :/
14:20:15LinusNpreglow: cool
14:20:15preglowit's decent enough quality and shouldn't be too hard to compute
14:20:32jhMikeSpengo: then the native speakers wonder about your weird accent
14:20:42LinusNi'm still thinking about psychoacoustic bass enhancement too
14:21:09LinusNshouldn't be that difficult
14:21:18preglowLinusN: i've had some attempts at that, all of them failures
14:21:19amiconnSome 'missing fundamental' stuff like the MAS does?
14:21:25LinusNamiconn: yes
14:21:27preglowLinusN: but then again, i haven't really ever heard a good one...
14:21:29LinusNpreglow: oh?
14:23:29pondlifeHmm, died
14:23:59pondlife uses LGPL
14:24:17 Join goffa_ [0] (n=goffa@
14:26:25LinusNtime stretch + crossfade would make a killer party jukebox
14:26:39pondlifeIncluding BPM detection
14:26:43LinusNcombined with beat matching
14:27:02pondlifeParty Mode just got good
14:28:07JdGordonwhats time streching?
14:28:16JdGordonslowing down the audio?
14:28:30pondlifeOr speeding it up
14:28:38pondlifeBut at the same pitch
14:29:22jhMikeSpreglow: I'm taking time out of my busy mpegplayer schedule to do this but another nice side effect will be that dsp access from plugins will be really simple. There's no "current DSP" notion but they're just assigned straight up at init and used with a pointer param.
14:29:36*JdGordon thought rockbox was full of audio puritsts... why do we want to fiddle with the music?
14:30:08pondlifeEach to their own....
14:30:42preglowjhMikeS: most excellent
14:30:47 Quit goffa_ (Remote closed the connection)
14:30:49pondlifeSeems like time stretching is a high-latency operation.
14:30:51preglowcode simplification > poor code :P
14:30:59preglowpondlife: not really "high"
14:31:05preglowthere's some buffering, but it's not high latency
14:31:19jhMikeSBeat detection too? Argh...we could just use backlight intensity now to serve as a peakmeter :)
14:31:37JdGordonthere is a patch for that
14:31:58jhMikeSay, dios mio
14:32:07pondlifeBacklght discolight
14:32:17 Quit Bagder (Remote closed the connection)
14:32:57jhMikeSoh...I got one...beat detection + bumping the screen image around with the bass
14:33:29jhMikeSI guess we'll need viewports for that
14:34:10*petur remembers gl.tter has a backlight beat patch
14:34:25pondlifeLGPL code can be included in Rockbox, right?
14:34:35 Quit Thundercloud (Remote closed the connection)
14:35:19 Quit goffa (Read error: 110 (Connection timed out))
14:35:34GodEateronly dumping the screen image? I think we could do better. Beat timed image distortions (in OpenGL of course) ;)
14:37:13 Join japc [0] (n=japc@
14:38:01LinusNpondlife: yes, i believe so
14:39:58GodEaterwhere's our code lawyer ?
14:40:29*Zagor dons the silly hat
14:40:33Zagorpondlife: yes
14:52:29jhMikeSpreglow: you said 2k stack?
14:53:13 Join Soap_ [0] (n=Soap@rockbox/staff/soap)
14:53:30preglowjhMikeS: that should do
14:54:45 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
14:55:16markunpreglow: did you have a chance to try mbrola, or busy with other things?
14:55:46preglowbusy with other things, among other assembling a table :P
14:56:56 Quit nicktastic (Client Quit)
14:57:57markunpreglow: your own table?
15:04:44 Nick bb_ is now known as bb (n=bb@unaffiliated/bb)
15:07:03 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
15:07:23 Quit animeloe ("Leaving")
15:07:55***Saving seen data "./dancer.seen"
15:10:30 Quit Soap__ (Read error: 110 (Connection timed out))
15:19:27 Join Guile`` [0] (n=Guile@
15:21:45 Join darkapostrophe [0] (
15:30:00 Join DerPapst [0] (
15:31:22 Quit JdGordon ("Konversation terminated!")
15:31:45 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
15:35:53preglowlinuxstb: looks like building from our svn libspeex should be doable
15:36:01preglowso i guess we should go for that
15:38:36 Quit pengo ()
15:38:40 Quit Dark_Apostrophe (Read error: 110 (Connection timed out))
15:38:41 Nick darkapostrophe is now known as Dark_Apostrophe (
15:40:03 Quit Guile` (Read error: 110 (Connection timed out))
15:40:10 Quit DerPapst (Nick collision from services.)
15:40:24 Join DerPapst [0] (
15:42:01 Join Thundercloud [0] (
15:42:07linuxstbpreglow: Sounds good.
15:45:53 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
15:46:40preglowlinuxstb: i guess we should build libspeexdsp in addition, somehow
15:46:43preglowall the sources are there
15:47:32 Quit nicktastic ("Leaving")
15:48:19 Join ackbahr [0] (
15:49:22barrywardelljhMikeS: sorry for running after asking my question earlier.basically, I think the H10 scrollpad should behave like it does in the OF
15:54:31pondlifelinusN: So might be useful.
15:54:55pondlifeAlthough it's C++.... :(
15:55:05GodEaterbooo hiss
15:56:19jhMikeSbarrywardell: me too...and that was a long way of saying it :)
15:56:40barrywardellI'm just not sure how is best to implement it
15:57:05barrywardellhow doe the acceleration work currently in other drivers?
15:57:42preglowpondlife: i'll be using that
15:58:44preglowor at least using it for ideas
15:58:54preglowi'll have to recode, but the core really isn't much code
16:01:39jhMikeSpreglow: not I get undefined refs to codec_calloc and codec_free. just stub them out?
16:01:58 Part LinusN
16:02:12preglowjhMikeS: you can either comment them away or stub them out in rockbox.h
16:02:25preglowand that you can do without checking for ROCKBOX_VOICE_CODEC
16:02:41jhMikeSI'd rather just comment them away :)
16:02:56preglowsure, go for it
16:03:11 Quit Casainho ("ChatZilla 0.9.79 [Firefox]")
16:04:31 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
16:04:33jhMikeSspeex_bits_init seems to need it
16:04:41preglowwe don't use speex_bits_init
16:04:59preglowwe use speex_bits_set_bit_buffer, and that just copies a pointer
16:05:11 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
16:05:12jhMikeSok, so just comment away that one
16:05:23jhMikeScan I just #if 0 speed_bits_init?
16:05:45 Quit J3TC- (Read error: 110 (Connection timed out))
16:05:58preglowbut we'll need it for rbspeexenc
16:06:02preglowso #if 0 isn't very nice
16:06:03jhMikeSand speex_bits_destroy?
16:06:16preglow#ifndef ROCKBOX_VOICE_ENCODER would be nicer
16:06:17GodEaterdid anyone ever identify the SoC used in the 2G Nano ?
16:06:44jhMikeSpreglow: you said "sure". :) and ROCKBOX_VOICE_ENCODER?
16:06:57 Quit Zagor ("Client exiting")
16:07:17preglowjhMikeS: "sure" as in we won't need it for voice, but we will need it if we're going to use our svn libspeex for building our own encoder
16:07:23preglowand in that case #if 0 isn't very helpful, heh
16:07:23DerPapstGodEater: not that i know of.
16:07:33linuxstbGodEater: I think it's an Apple branded chip with an Apple serial number....
16:07:39jhMikeSwhat about #ifndef ROCKBOX_VOICE_CODEC :D
16:07:45DerPapstGodEater: they still think it's some kind of custom samsung chip
16:07:55GodEaterDerPapst: that's what I though too
16:07:58preglowjhMikeS: well, then it would build for the voice decoder...
16:08:08preglowlinuxstb: samsung chip, isn't it?
16:08:18*GodEater has just noticed that the OpenMoko phone is using a Samsung SoC
16:08:28linuxstbpreglow: Yes, but "samsung chip" is all that's known I think...
16:08:31GodEaterwhich presumably has open documentation
16:08:57GodEaterI've just found some words on the Samsung site which are a close match with part of the serial no. on the Nano 2G's SoC
16:09:09DerPapst*jtag pops in mind*
16:09:52jhMikeSpreglow: argh, I'll let you setup your build #defines :)
16:10:02suplets get that 2g nano hacked
16:10:03sup: ]
16:10:24preglowjhMikeS: simple, there are two: ROCKBOX_VOICE_CODEC for voice decoder, ROCKBOX_VOICE_ENCODER for rbspeexenc, the last is currently only used in config-speex.h here
16:11:05jhMikeSouch, 521KB for sansa
16:11:09preglowto enable floating point
16:11:14preglowrockbox.ipod ?
16:11:32 Quit moos ("Rockbox rules the DAP world")
16:11:35jhMikeScheck the build table?
16:11:42amiconnDerPapst: If they glued the chips to the board to prevent measuring, they probably didn't put an open jtag interface on the board...
16:11:54DerPapstthough it would be very hard to jtag the 2nd gen nano. maybe the "soilder points" are between some PCB layers
16:12:02jhMikeSlast build I made is 471
16:12:08preglowjhMikeS: as far as i can see, 498kb
16:12:28linuxstbjhMikeS: Have you removed all the existing swapping code etc?
16:12:37preglowrockbox.ipod is 489kb here, i'd hardly call that a big leap in bin size
16:12:48preglowanyway, there are more stuff i can remove
16:12:53preglowis, even
16:13:07linuxstbBut haven't we saved that 512KB swap buffer?
16:13:13jhMikeSI can't wait for this build to finish...then crash...*crosses fingers*
16:13:22jhMikeSlinuxstb: yes
16:13:22 Join jpt9 [0] (
16:13:23GodEatersuspects are : or
16:13:26amiconnlinuxstb: Plus 48KB iram swap buffer
16:13:34jpt9i'm trying to install Rockbox on a Sansa E250R.
16:13:45jpt9I can't seem to get the Sansa to show up as a USB mass storage device.
16:13:51*DerPapst will do that soon too :D
16:13:53jhMikeSbut those buffers were only allocated before when voice was actually required
16:13:58jpt9(and what does show up doesn't seem to have a SYSTEM directory)
16:14:05linuxstbjpt9: I believe you need to switch it into "Rhapsody mode" in the original firmware.
16:14:32jpt9stupid question...
16:14:36jpt9I've done step 1.
16:14:48 Join Febs [0] (n=chatzill@
16:14:49jpt9will Rockbox Utility do step 2 properly if I have an R?
16:15:13linuxstbRockbox Utility will do the unzipping of for you, but that's all.
16:15:14 Join ATravelil [0] (
16:15:24 Quit CaptainSquid83 ("Miranda IM!")
16:16:42jhMikeSoh, _last_ build I made for e200 was 488KB
16:17:31GodEatersamsung's bloody site keeps crashing firefox :(
16:18:41DerPapstwhen ff crashes for me it's mostly because of stupid flash thingies
16:19:37amiconnFirefox crashing??
16:20:00GodEaterDerPapst: yes, I think that's what's causing it here too
16:20:16*linuxstb has similar problems with ff and flash (under Linux)
16:20:24DerPapstamiconn: yes. because of buggy ploojins
16:20:32GodEaterploojins ?
16:21:28DerPapstisn't that the way rockbox "pronounced" plugins a few moth back?
16:21:36DerPapst*month even
16:21:47DerPapstbah.. i give up :P
16:21:48jhMikeSwell, if I had designed this to lock up, it would be a job well done :)
16:21:56GodEatermust have missed it
16:21:59preglowjhMikeS: hmm? it's down to that size now?
16:22:22jhMikeSwhat do you mean?
16:22:38preglow16:16 < jhMikeS> oh, _last_ build I made for e200 was 488KB
16:22:45preglowjust wondering if that size is with speex voice
16:22:59jpt9i assume that now that I've Rockboxed it, I can use Rockbox utility to update it and install extras and the like?
16:23:01DerPapstGodEater: if you have windows ask MS Sam to say plugins for you ;)
16:23:45GodEaterWindows ?
16:23:49GodEaterMoi ?
16:23:54linuxstbjpt9: Yes - everything apart from "bootloader installation" should work. Is Rockbox working now?
16:23:56jhMikeS498908 (SVN) vs. 533504 (core voice)
16:23:59DerPapstthen flite or so
16:24:04 Quit ATravelingGeek (Nick collision from services.)
16:24:04 Nick ATravelil is now known as ATravelingGeek (
16:24:06DerPapstshould do the same :P
16:25:21*jhMikeS should probably debug on the sim first
16:25:44preglowjhMikeS: so we _lost_ binary size?
16:26:01*preglow reads again
16:26:01jhMikeSno? 1st since is taken from the build table
16:26:15preglowthat's good, now i can have fun optimizing for size :>
16:26:22 Quit animeloe ("Leaving")
16:26:47linuxstbHow much IRAM did it need in the end?
16:27:28preglowi haven't trimmed that yet
16:27:51preglowjhMikeS: increase the stack size above 2kb if you're getting bugs, it might not be completely accurate
16:28:06preglowi just know it uses 1.2kb in stack arrays, i adjusted that up to 2kb to be on the safe size
16:28:17preglowmight not have been safe enough
16:28:21 Join MethoS- [0] (
16:28:38 Quit MethoS- (Read error: 104 (Connection reset by peer))
16:29:55jhMikeSpreglow: well, can try indeed.
16:31:50jhMikeSpreglow: did you see how close 2048 was to full?
16:32:28preglowi could only check what speex allocates through its own stack allocation wrappers, and it uses that for all the big stuff
16:32:34preglowit used 1236 bytes of that
16:32:52preglowthe only other stuff that is supposed to be put on the stack are local variables
16:32:54jhMikeShow big is SpeexBits?
16:33:01preglownot big
16:33:25preglow36 bytes, or thereabouts
16:33:52 Join karashata [0] (
16:34:08 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
16:35:47jhMikeSjust to get the thread started was 9% of the 2k
16:36:36GodEaterIs it time for an AlbumArt section in the manual now that it's in svn ?
16:36:41jhMikeSI doubled the size - no change ... it's something else
16:38:12 Join MethoS- [0] (
16:38:53 Join PaulPosition [0] (
16:39:39 Nick fxb is now known as fxb__ (
16:39:44 Join jgarvey [0] (
16:39:54pondlifeGodEater: Yes, I should read it too :)
16:43:17n1sGodEater: would be very nice :) what do you think about putting it in the "advanced topics" chapter?
16:44:46jhMikeSpreglow: care to look at a pastebin and see if anything was horribly botched on my part?
16:45:12preglowjhMikeS: sure, go ahead
16:45:49preglowsim doesn't give you any useful info?
16:47:53jhMikeShere's the thing that is it the thing:p
16:48:27jhMikeSpreglow: am building that right now
16:48:36GodEatern1s: I'm not sure - I've been thinking about it
16:50:11 Quit hunz (Read error: 104 (Connection reset by peer))
16:50:59GodEaterIt probably needs it's own section, but also modifications to the WPS Tags pages
16:51:03 Join hunz [0] (
16:54:01preglowjhMikeS: you seem to be passing a pointer to a pointer to speex_decode_int
16:54:02preglowis that right?
16:54:02n1sGodeater: maybe it could fit in the "browsing and playing chapter"
16:54:28 Quit linuxstb (Read error: 104 (Connection reset by peer))
16:54:32preglowshouldn't it be: speex_decode_int(, &td.bits, (spx_int16_t *)td.src[0]);
16:54:54jhMikeSwhy are those pointer void * anyway? shouldn't they be typed to catch this stuff?
16:55:03preglowbecause he passes along different types of structs
16:55:14preglowbecause of the hierarchial way the modes are built on top of each other
16:55:31preglowmight be a better way of doing it, dunno
16:55:39jhMikeSthere's better ways and typesafe ways
16:55:48preglowprobably :)
16:55:54 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
16:55:58jhMikeSsome kind of union declaration (forget the attribute)
16:56:13preglowis that gcc only, then?
16:56:24jhMikeSI guess
16:56:25preglowjmspeex wants libspeex to build well on all kinds of compilers, including c++ compilers :>
16:56:38jhMikeSour version?
16:56:49preglowsure, we can deal with it as we wish
16:56:54preglowbut i can't be bothered, that's for sure
16:57:14preglowwe won't touch this code too often anyway :>
16:57:15jhMikeSwell, guess these little mistakes only have to be caught once
16:57:19preglowyeah, exactly
16:59:09jhMikeSheh...working perfectly :p
16:59:31preglowbugs 0, thom 1
16:59:55jhMikeSloading a plugin bit the dust however :)
17:00:46preglowhaha, it can't all be good
17:01:18jhMikeSthis is litterally the first time the code ran with the dumb error, so I'm suprised I can whiz around the menus without any trouble
17:01:33linuxstbSo you're listening to speex voice?
17:01:41linuxstbThen \o/
17:01:50preglowwoot \o_
17:02:53preglownice crowd effect!
17:03:04markunpreglow: do we save only space or also CPU time?
17:03:15markun... by using speex instead of mp3
17:03:17*GodEater should have lined his up
17:03:52preglowmarkun: don't know about cpu time, for some reason, mp3 is bloody fast
17:04:08preglowbut speex still has optimization potential
17:04:46linuxstbBut we save all that codec swapping...
17:05:04jhMikeSradio screen voices nicely
17:06:15GodEateris this for voice in general, or just .talk ?
17:06:19 Quit ackbahr (Remote closed the connection)
17:06:28jhMikeSall SWCODEC voice
17:06:42*GodEater realises that was a dumb question
17:07:03markununtil someone writes a speex decoder for MAS
17:07:19jhMikeSI had some clips come out wrong though
17:07:28jhMikeSNot sure if that's the voice file or not
17:07:40markunjhMikeS: wrong in what way?
17:07:41*jhMikeS runs with gdb
17:07:51jhMikeSwrong string voiced in system
17:07:57***Saving seen data "./dancer.seen"
17:09:11 Quit atsea- (Read error: 104 (Connection reset by peer))
17:09:58 Quit nicktastic ("Leaving")
17:10:31 Join toffe82 [0] (
17:11:54 Join martii [0] (i=c24bfb32@gateway/web/cgi-irc/
17:12:28n1sjhMikeS: which player?
17:12:41 Join foolsh [0] (
17:12:46martiiguys I have ipod 4th gen and itrip
17:12:52jhMikeSe200 sim
17:13:17linuxstbmartii: Seen this page?
17:13:25 Quit MethoS- (Remote closed the connection)
17:13:25 Quit spiorf (Read error: 110 (Connection timed out))
17:13:41 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
17:14:55jhMikeSI'll try on target to be sure it's not a sim thing
17:16:02linuxstbWouldn't the features.txt be different in the sim - needing a different voicefile?
17:16:43n1slinuxstb: I fixed that right after the lang cleanup so sims and targets could use the same voicefiles but it's been broken :(
17:16:51n1sI see the problem for e200
17:16:53linuxstbHow did you fix it?
17:17:06jhMikeSshould it fault or cause problem if the clip is wrong or nonexistent though?
17:17:57n1senable the features in config*.h files for sims too, most of them didn't cause any trouble and add a couple of stubs for the ones that did
17:18:28preglowjhMikeS: 've you got a patch for all this? i'd like to try my hand at some size optimizing
17:19:03jhMikeSI will but it's dodgey atm if you want the dodgeyness
17:19:09preglowi'm cool with dodgey
17:19:21n1sjhMikeS: the problem is that voice ids are matched with ids in the lng file so if a string is missing all clips will match the wrong string after that
17:20:11jhMikeSright. I had a slight format change that would forever fix that in the most final way but...
17:20:14scorche|wDerPapst: PM me your address when you get this
17:20:54*jpt9 loves rockbox.
17:20:58jpt9it plays doom!
17:21:07GodEatermusic too
17:21:15n1sjhMikeS:, preglow: there's another thing that should be fixed, we don't actually check the version of the loaded language file atm.
17:21:38n1sI have a patch and will test it as soon as i can
17:22:10linuxstbn1s: What's the e200 lang problem? I can't spot anything odd in the config-e200.h file...
17:22:53jpt9sansae kick ass.
17:22:53n1slinuxstb: HAVE_MULTIVOLUME is inside the #ifndef SIMULATOR block
17:23:34jhMikeSHAVE_MULTIVOLUME shouldn't be on the sim, eh?
17:23:43linuxstbAh, and there are now voice strings related to multivolume?
17:24:24jhMikeSexactly why voice stings should have unique global IDs
17:24:25martiilooks like I'm not lucky my itrip is PAV4026
17:24:38martiilinuxstb: how can I help to make it working?
17:24:48n1sjhMikeS: ondio has HAVE_MULTIVOLUME defined for sims and that works fine
17:25:14linuxstbmartii: Write a low-level serial driver for Rockbox to talk to the itrip, then implement support for the Apple Accessory Protocol in Rockbox...
17:25:17jhMikeSstill, unique global IDs for strings would make any voice file work on any target
17:25:29jhMikeSsave and mp3/speex difference of course
17:25:51linuxstbHow would you generate those IDs?
17:25:59preglowmd5? :>
17:26:05martiilinuxstb: I've been thinking more about debugging testing ready code
17:26:06n1sjhMikeS: sure, making the format more robust is great
17:26:19jhMikeSThey aren't. They're added from the last one used when a new string is added and never, ever change
17:26:23martiilinuxstb: I'm not skilled enough to do what you're asking for
17:26:29n1smartii: ther's no ready code for that
17:26:57jhMikeSid = next_id++...always
17:27:06martiin1s: where to start from?
17:27:06pondlifeIt would be great for developers to have global voice IDs.
17:27:35n1smartii: <linuxstb> martii: Write a low-level serial driver for Rockbox to talk to the itrip, then implement support for the Apple Accessory Protocol in Rockbox...
17:27:48 Join J3TC- [0] (
17:27:55martiin1s: OK any tutorials how to do it?
17:28:09pondlifeIn fact, just get the serial driver part done :)
17:28:19n1smartii: no
17:28:27martiin1s: any specs?
17:28:39jhMikeSunless we get too hungry for new settings I think 16 bits would be enough. I guess the initial IDs would just be done automatically by index.
17:28:41n1sthat's the "fun" part :)
17:28:50 Quit idnar (Nick collision from services.)
17:28:52 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
17:29:01linuxstbjhMikeS: Ah, so you mean they'll just be numbers added manually into the lang file?
17:29:06jhMikeSstring could also be simply deleted
17:29:31jhMikeSlinuxstb: yes...add the ID, increment to next ID
17:29:59markunmartii, n1s: what about this?
17:30:01jhMikeSthe voice building tool would of course spit out errors on a duplicate
17:30:48 Join lee-qid [0] (
17:30:53n1smartii: you need a serial driver, which we don' thave to implement that
17:30:56jhMikeSwe'd still have indexes that are target specific of course and are used internally, but those indexes would be reverse lookup from the IDs in the lang file. each slot would have the ID associated with it.
17:31:25markunn1s: talking to me?
17:31:44n1smarkun: ah, yes too similar nicks :)
17:32:15n1smarkun: afaik ipl only got their serial driver to work on 3g ipods
17:34:50linuxstbWould be fun to add support for Apple accessories to the non-iPod targets...
17:36:07 Quit hunz ("Lost terminal")
17:36:36GodEaterlinuxstb: wouldn't that just be adding insult to injury ?
17:36:51linuxstbIn what way?
17:37:48GodEaterwell, if we got accessories working on targets other than ipods
17:37:57GodEaterit would REALLY upset the ipod owners
17:38:13pondlifeLet's do it then!
17:38:25linuxstbServes them right for buying ipods...
17:38:34GodEaterthat applies to you and me then...
17:38:50linuxstbI don't own any accessories though....
17:38:59GodEaternor me
17:39:34 Quit nicktastic ("!@^#&$*(")
17:41:05jhMikeSpreglow: will have a patch shortly
17:44:44jhMikeSI wouldn't even yet. It's only running in the sim a little bit.
17:46:41amiconnjhMikeS: If we want global unique IDs, we need a way to load the now non-contiguous IDs *without* wasting memory
17:47:19jhMikeSno gaps...that's part of the spec
17:47:20amiconnThat's the hard part
17:47:33amiconnThere will be gaps for a target
17:47:41amiconnNo target will need all IDs
17:47:42jhMikeSno, there won't be
17:47:58jhMikeSjust sort them and binary search at load time
17:48:14amiconnHow would that work?
17:48:56jhMikeSif the non-continuous IDs are sorted, the inverse lookup from ID to index is trivial. Strings in program code will still be accessed by index.
17:50:07amiconnHow would the core know which IDs a target needs?
17:50:09markunlinuxstb: unfortunately the UART of the Gigabeat is not connected to the dock connector
17:50:10jhMikeStable[INDEX] = { "string", UNIQUE_ID }
17:50:27 Quit Nico_P (Remote closed the connection)
17:50:38amiconnI mean when it loads the .voice/.lng file
17:51:32jhMikeSit can iterate then table, and look for the voice id in the file _or_ iterate the voice file table and see if that id exists in the target's table.
17:51:48amiconnWhat target's table?
17:52:07amiconnRight now the target has no table - and I'd rather not want to waste memory for such a table...
17:52:14jhMikeSany targets table...the target will have the global ids in association with index
17:52:22krazykitamiconn, your name is in the do you know what license the WPSs in SVN fall under?
17:52:45jhMikeSamiconn: I didn't expect you would frankly
17:52:52krazykitno one else seems to have any idea :-/
17:53:52jhMikeShowver, a parallel array for the IDs means each element could be 16 bits and not 32
17:54:27*linuxstb counts 740 id: lines in english.lang...
17:54:59*GodEater bets linuxstb didn't count them, and that he used some combination of grep and wc
17:55:40*linuxstb bets GodEater was spot on
17:55:49 Join Domonoky [0] (
17:56:50 Quit foolsh ("Konversation terminated!") 64k enough for a string table?
17:59:08preglowanother jhMikeS size patch :P
17:59:12 Quit jpt9 (Read error: 110 (Connection timed out))
17:59:25 Quit linuxstb ("Client Exiting")
18:00:09jhMikeSI just thought of a way to get the robustness for no size cost in the string table. I also though of a way to have threadsafe current viewports with no extra parameter :)
18:00:48 Join foolsh [0] (
18:00:53 Quit foolsh (Read error: 104 (Connection reset by peer))
18:01:03preglowskipping tracks around track boundaries is still really buggy
18:01:50*jhMikeS hopes he's not opening can of racing worms in other code with this.
18:01:55preglowwasn't that one of the things mob was supposed to fix?
18:02:31*jhMikeS thought it was supposed to abstract things too and make the code cleaner
18:02:43preglowit hasn't? :>
18:02:49 Join atsea- [0] (i=atsea-@gateway/tor/x-c707381b7def75ce)
18:04:21preglowjhMikeS: ugh, patch starts being a bastard around memswap128, says it detects a reversed patch
18:04:22jhMikeSalso, if I change the audio thread priority...well, poof. that should be a high priority thread, not low. I _can_ get it to operate faster than the original code without any sort of special code yielding too.
18:04:37jhMikeSmemswap128 is no more of course
18:04:42preglowperhaps that is why, then
18:05:07jhMikeSit should still be on your system I take it?
18:06:40jhMikeSthat's the last stuff in there so just take the removals out I guess
18:06:49preglowit's still here, yes
18:07:09preglowwhich makes it weird that the patch doesn't apply cleanly
18:10:43*jhMikeS tries an H100 build since that has the system memory guard :\
18:11:08 Join nanok [0] (n=nanok@
18:12:17 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
18:12:43 Join Xerion_ [0] (
18:16:28 Quit petur ("work->home")
18:18:08 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
18:18:26jhMikeSrrrr...region IRAM is full :\
18:18:36preglowjhMikeS: all the exc_* tables can be taken out of iram
18:19:04preglowi was planning on introducing some new iram define there that doesn't include most of them
18:19:13jhMikeSwhere are those?
18:19:44preglowa bit surprising anyway, really, i thought we'd have enough iram after shrinking the stack
18:20:10jhMikeSthere's an extra output buffer don't forget...perhaps that should be stacked's friggin I really need IRAM for that?
18:22:50 Quit martii ("CGI:IRC (EOF)")
18:24:50jhMikeStaking those table out did the deed
18:25:14 Quit karashata ("I will *SO* make you regret that... Later... *is a lazy dragon*")
18:27:33 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
18:29:05preglowextra output buffer?
18:29:35jhMikeSfor the voice samples...that used to be inside the mpa codec
18:30:18 Quit Xerion (Read error: 110 (Connection timed out))
18:30:19 Nick Xerion_ is now known as Xerion (
18:30:46 Join MethoS- [0] (
18:31:05 Quit MethoS- (Remote closed the connection)
18:33:08 Join MethoS- [0] (
18:38:55 Join Frazz [0] (
18:40:01 Quit mf0102 (Read error: 110 (Connection timed out))
18:47:49 Join foolsh [0] (
18:49:13 Quit foolsh (Remote closed the connection)
18:49:52 Part pondlife ("Gone")
18:50:07 Quit barrywardell (Remote closed the connection)
18:50:32 Join bertrik [0] (
18:59:08 Quit J3TC- (Read error: 110 (Connection timed out))
19:02:53 Quit MethoS- (Remote closed the connection)
19:03:04 Join MethoS- [0] (
19:06:32 Join OlivierBorowski [0] (
19:06:49 Quit Febs (Read error: 104 (Connection reset by peer))
19:07:21 Join eigma [0] (n=cat@
19:07:59***Saving seen data "./dancer.seen"
19:16:15 Join Siku [0] (
19:17:08 Join pepie34 [0] (
19:21:36 Join mud [0] (
19:21:45 Nick mud is now known as mud-rb (
19:22:33 Nick fxb__ is now known as fxb (
19:22:58jhMikeSaha! it's working now in the sim at least (no target check yet) :)
19:23:54eigmajhMikeS: last night, kkurbjun was basically trying to implement a ringbuffer for the m:robe's UART.. any suggestions for the best way of doing that using rockbox's synchronization primitives?
19:24:45jhMikeSeigma: should we even need them? you can have a perfectly non-blocking queue without any
19:25:14eigmaare you talking about the uart's hardware fifo?
19:25:18jhMikeShowever, sempaphores are available if you don't want polling and would rather block threads to save cycles
19:26:22jhMikeSI don't know the details of what's being attempted here or I'd have more ideas.
19:27:18eigmato be honest, I don't either. my best understanding is that we want to nonblockingle queue "packets" (that may be longer than the hardware fifo) to be sent across the uart
19:27:29 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
19:27:36 Join farhank [0] (n=furk@
19:29:24jhMikeSit sounded like he just wanted to prevent an interrupt from changing things in the background, but set_irq_level would be enough for that.
19:30:01eigmaat a low-level, yeah (wouldn't disabling the particular UART interrupt be more efficient?). I meant at higher level what we needed to achieve, in case there was a better solution
19:30:20jhMikeSnot if it's just a sub-uS event
19:30:34eigmalooking at the code, looks like we want an asynch recv queue
19:30:40farhankHello, I have issues with my h10 that i'm not sure I can fix as I'm not a programmer. When I installed rockbox, and installed a voice file all the menu entries aren't where there supposed to be, the system menu doesn't speak correctly with the default festival english.voice. I tried compiling my own voices but still nothing. I have a feeling it is a problem with a lang file in the h10 target.
19:31:03jhMikeSqueues are easy to make async if only one reads and only one writes
19:32:22eigmaI don't even know why we're using a queue for receives. any single recv packet won't be larger than about 6 bytes, so the hardware 32-byte FIFO should be plenty
19:32:57eigma*a [software] queue
19:33:15jhMikeSdoesn't this thing have DMA requests?
19:33:28eigmawe don't use them, and i don't think it's worth it for the UART
19:33:34n1sfarhank: is your build current?
19:33:36jhMikeSworth it?
19:33:39farhankI'm stucking using a build before 8-05-07. before the voices got target spicific.
19:33:48farhankits not currently current no.
19:33:55eigmathe added complexity, for one 6-byte transfer every .1 seconds?
19:34:01jhMikeSI think explioting hardware features is always worth it
19:34:08n1sfarhank: why are you stuck with that?
19:34:11farhankI had a current build yesterday, and that's when I realized that all my voice files were a bit messed.
19:34:13eigmaI hear you..
19:34:20jhMikeSit will alleviate complexity in the software...just need some driver code.
19:34:43n1sfarhank: if you install a current build you _need_ a current voicefile
19:34:44jhMikeSI guess the pcm driver should be what does this
19:34:55farhankbecause when I use the newer builds with the newer voice files something is messed up in my system menu. menu entries are all moved around.
19:35:23mud-rbfarhank: try on a fresh install? when i updated some time ago my menu entries got all messed up as well but it was fixed by a clean install
19:35:31farhankI tried the precompiled festival voice, and I also compiled a voice of my own there are miner differences but nothing...
19:35:40farhankit doesn't help, I did.
19:35:49amiconnWee :)
19:36:08amiconnc200 sees an even bigger speedup from the bridge configuration that I expected :)
19:36:16n1sfarhank: do you have a rockbox.something file in the root of the player?
19:36:29farhanka folder?
19:36:30jhMikeSeigma: I'm really not in touch with this. I don't doubt I'd exploit everything in hardware and have many more ideas with a device handy. :)
19:36:37farhanki have all that
19:36:41amiconnkkurbjun: around?
19:36:49eigmajhMikeS: yeah.. thanks.
19:37:07n1sfarhank: what "all that" do you have the _file_ in the _root_ ?
19:37:11mud-rbfarhank: i believe he means a firmware file, like rockbox.mi4 on the sansa, i think it's different extension for others
19:37:31mud-rbnot the .rockbox folder
19:37:37farhankthe only issue i'm having with is when I use the current build with the current voices. the system menu and the context menu is all different, i'm blind so I don't know if the entries are there or just changed I haven't got someone to look at it.
19:37:44jhMikeSgood thing is the speex seems to run well enough I feel like committing it now...hrm better wait for the tools eh? :)
19:38:23n1sfarhank: some of those menus have changed a lot
19:38:32nanokfarhank: might this be because of the remaping of the buttons?
19:38:36farhankits interesting because the festival voice says different things then my precompiled voice.
19:38:55farhankI think its a lang problem...
19:39:09n1sfarhank: are you building your own voice with a current svn checkout?
19:39:13nanokfarhank: maybe you get a different menu than what you expect, because of the recent remapping
19:39:27Domonokyfarhank: please check if there is a rockbox.mi4 / rockbox.h10 file in the root of your player, ifso. delete it..
19:40:43farhankthere is not..
19:41:19farhankyou said the root? not in the rockbox folder itself right?
19:42:08Domonokyyes... the rockbox.xx file is now in .rockbox, but sometime ago it was in the root, and if you have both it causes problems..
19:42:21n1sfarhank: can you check if the menu entries work correctly even if they are not where you expect them to be?
19:42:33farhanknope. no one around
19:42:45 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:43:14farhankif I go in to the system menu, or the settings menu, it says random things like party mode and just not what's supposed to be there.
19:43:20 Join bluey [0] (
19:46:26mud-rbfarhank: and this has been happening for quite a few versions you said, right? or is it very recent?
19:46:40jhMikeSpreglow: A 2k voice stack gives 78% usage on e200
19:46:51farhankfor example, when I go in to the system menu and check the battery it says q then I check the battery and it says q last and then gives me the battery status.
19:47:23farhankits happened on all the versions from oct 15 to now... i just got my h10 yesterday.
19:48:38farhankalso if I go to settings apparently under general settings if format... but that isn't right.
19:51:08 Join Llorean [0] (
19:52:38 Join animeloe_ [0] (n=animeloe@unaffiliated/animeloe)
19:53:20mud-rbfarhank: i think you mentioned compiling your own voice files, this may be a stupid question but you tried the precompiled ones, correct?
19:53:47farhankyeah I did at least the festival one.
19:53:52farhankits broken also.
19:54:29 Join MethoS-- [0] (
19:54:38 Quit OlivierBorowski (Remote closed the connection)
19:55:22farhankthis is with the current build, apparently the voice is telling this is what is in my settings menu. sound settings, general settings, format and clear preset list.
19:55:33farhanksomethign is definitely broken.
19:55:39n1sfarhank: can you rename your .rockbox folder to something else and install a new current build and a current voicefile from the site and see if it still happens?
19:55:54farhankthis is what I just did.
19:56:51farhanki just downloaed a new rockbox build from the current build section, and am using a voice compiled yesterday.
19:57:19 Quit pepie34 ("Ex-Chat")
19:58:34n1sfarhank: a voice from yesterday is _not_ recent enough in this case since there were some string order braking changes late yesterday
19:58:57amiconnYes, and H10 got alarm wakeup yesterday...
19:59:02n1svoice works just fine on h10 sim here at least
19:59:11 Quit XavierGr (Read error: 104 (Connection reset by peer))
19:59:34n1samiconn: that's what I meant by string order braking changes, people really should note such things in commit messages...
20:00:33PaulPositionalarm works great, btw! :D
20:01:33 Quit eigma ()
20:01:39n1sbtw I wonder if we should start bumping voice version when it breaks just like plugins or if it's better to say some things wrong...
20:01:39 Join spiorf [0] (
20:02:01 Quit animeloe (Connection timed out)
20:02:49PaulPositionOh... a question : I've been trying to find any sort of information on how to get the best out of recording through a device's internal mic. −− Now I know Rockboxers are mostly real audiophiles and wouldn't touch one onboard mic with a ten foot pole, but surely there must be some settings better suited to it? Can't find much in the wiki or forums, but maybe I missed something?
20:03:06amiconnIt's combined .lng/.voice version. I think we should bump it - now that we're providing automated voice builds, it's not a big problem
20:03:42amiconnBut such a commit should point out that voice / will break
20:03:49amiconn- /
20:04:46n1samiconn: I am about to commit a patch that makes the voice file loader actually check the version, where is a good place to have that version define as voicefont needs it too?
20:05:04LloreanPaulPosition: Are there really that many recording settings that you feel confused as to what each might do?
20:05:11amiconnGood question...
20:05:41amiconnn1s: Btw, there is a .lng version definition - and .lng/.voice are coupled
20:06:07farhankright let me reboot and see if this is fixed.
20:06:28 Join XavierGr [0] (
20:07:47n1samiconn: right, do you think it should use the same? also the voice version is defined in language.h but genlang doesn't read that...
20:09:02farhankhmm interesting.. it seems to be working with the festival voice. i'll have to check out an snv update and see what I can come up with...
20:09:13 Quit animeloe_ (Connection timed out)
20:10:04n1smaybe that should wait until the speex voice stuff so we only brake it once
20:10:25 Join robin0800 [0] (
20:11:21PaulPositionllorean - You're right in that there's not so many settings. I was looking for some tips like there are for line-in stuff, but I guess if there isn't any, well there isn't any and that's that. It's okay, too.
20:12:05 Quit MethoS- (Read error: 110 (Connection timed out))
20:12:40 Quit japc (Read error: 110 (Connection timed out))
20:13:37 Quit MethoS-- (Connection timed out)
20:13:44 Join MethoS-- [0] (
20:15:01 Join thegeek [0] (
20:16:18LloreanPaulPosition: Really with the internal mic, all you need to do is find a gain level that feels good for the situation you're in, I think.
20:18:33PaulPositionLlorean - Might be the case, indeed.. Anyway, I'll test stuff, I shouldn't be so lazy. :p
20:18:36farhankyeah. still broken with a compiled voice.
20:19:25farhankbut its not broken with the festival voice. hmmm strange
20:22:09 Part bluey ("Leaving")
20:22:21 Join bluey [0] (
20:26:55preglowjhMikeS: then it can be shrunk further
20:27:03preglowjhMikeS: stack usage should never spike any more than that
20:33:14jhMikeSbtw, I need no IRAM output buffer. it's fine in RAM.
20:34:12jhMikeSit gets copied anyway for the DSP conversion
20:35:17preglowcaught any new bugs?
20:35:59jhMikeSJust that I wasn't resetting the pointer I was giving to the DSP each time starting a fill, so that was the problem
20:37:14 Join petur [0] (n=petur@rockbox/developer/petur)
20:38:49 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
20:38:51jhMikeSoh, on H100 the stack is running at about 84%
20:39:17preglowdo we want to ride it as close as possible?
20:39:30 Quit ender` (" I find television very educating. Every time somebody turns on the set, I go into the other room and read a book. -- Grou")
20:39:44preglowlike i said, speex stack usage should be completely static for a given mode
20:40:05preglowso if we don't use uwb, it shouldn't run at more than 85%
20:40:44jhMikeSsure...cut it close
20:41:09preglowbut anyway, is that code commitable as it is?
20:41:17preglowi guess we should give users some advance warning on this
20:41:41preglowalso, bagder needs to fix up the build system a wee bit
20:41:42jhMikeSNot quite there yet.
20:42:08jhMikeSwhen I'm happy I can't do anything to bug it no matter what
20:42:41 Quit bluey ("Leaving")
20:42:52farhanki'm going to try compiling my own rockbox build from svn again and see how that goes with this new voice...
20:43:47 Join ender` [0] (
20:44:40jhMikeSI'll have a patch in FS once I've ironed out some stuff
20:47:17preglowcool, should you include my stuff in your patch as well so people get a complete package?
20:47:50farhankwhat's the deal when you stop a wma file and then go back to resume playing of it later it starts from the begining?
20:48:01preglowfarhank: wma doesn't support seeking
20:48:10preglowfarhank: that automatically means it also doesn't support resuming
20:48:12jhMikeSpreglow: hand it over...better not get conflicts applying it! :)
20:48:23preglowjhMikeS: i've only touched configure and
20:48:58farhankhmmm wait, but I can fastforward and it works fine.
20:49:41 Join PaulJam [0] (
20:49:43preglowfarhank: no, no you can't
20:49:45preglownot in wma
20:50:03preglowyou need to stop drinking right now if you think you can :)
20:54:23farhankreally? because I know I skipped through a file a bit ago. it was an unprotected not drm file.
20:55:57preglowfarhank: really
20:56:11preglowi've got a patch for wma seeking sitting on my hard disk right here, and i know for sure that it is not in rockbox yet
20:58:00PaulJamwhen you fast forward in a wma rockbox pretends to scan throught the file, but as soom as you release the button, it jumps back to the old position
20:58:33 Join mirak [0] (
20:58:42 Quit petur ("switching")
20:58:51 Join petur [0] (
21:00:39 Join agm3nt [0] (
21:00:45jhMikeShey, what's with this, the sansa from woot that has Rhapsody fw and backing :ppp
21:01:18mud-rbnot just a 200r?
21:02:01jhMikeSIt says e250R but wasn't advertised as such
21:02:25 Join stevenm [0] (n=stevenm@
21:02:27advcomp2019i have heard other people with that too
21:02:47stevenmn1s, Hi. I think the logic rearrangement in synth.c broke playback
21:03:04nanoki had a smilar experience, but fortunately the reverse: i was sure i was buying an r, but somehow i got the vanilla one
21:03:06n1sstevenm: really, I tested it quite a lot
21:03:12nanoki was very happy :)
21:03:20n1sstevenm: do you have an example file?
21:03:23 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
21:03:24stevenmn1s, Things sound differently, with looped instruments
21:03:51stevenmn1s, I will go grab fresh sources and double-check, but I was trying out some pitch bend stuff and noticed something strange
21:03:53n1sstevenm: oh, probably my fault then
21:04:03stevenmn1s, I think we need state_looping after all
21:04:52 Join hannesd [0] (
21:05:29farhanki'm now compiling a voice and have just compiled a new rockbox build after deleting my svn directory.
21:05:36farhanki hope this works.
21:07:27*ender` yawns
21:07:45n1sstevenm: hmm, yes I think I understand, now both the loop tests _could_be true at the same time but only if start_loop > end_loop, i thought that never happend but might be wrong...
21:08:00***Saving seen data "./dancer.seen"
21:08:13stevenmHmmm.. That is strange. shouldn't happen
21:08:28 Quit animeloe (Connection timed out)
21:09:23stevenmFinally built. bug occurs in 15569.. building 15561 now
21:10:32 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
21:10:49jhMikeSit has a big nasty "REFURB" on the back in chrome letters too xP
21:10:52 Join alienbiker99 [0] (
21:11:52nanokjhMikeS: that's a feature, not a bug ;)
21:11:57stevenmn1s, 15561 works right
21:12:12jhMikeSraised letters no less...tactile interface
21:12:20n1sstevenm: ok, I'll look into it, thanks for spotting
21:12:27stevenmn1s, np. want a file?
21:12:38n1sstevenm: yes that would be great
21:13:06stevenmn1s, stevenm/weirdbend.mid">
21:13:19 Join Daniel_S [0] (i=57b0e798@gateway/web/cgi-irc/
21:13:38n1summm 404
21:13:40stevenmn1s, you notice it approx 5-10 sec into the file. there are a lot of looped instruments in the background that are playing in the earilier version but not in the later one
21:13:56 Quit Siku (Nick collision from services.)
21:13:59 Join Siku [0] (
21:14:04stevenmn1s, eep! no space left on device. You log into the account on a windows machine and the stupid roaming profile eats up all your space.
21:14:08stevenmn1s, hang on....
21:14:31jhMikeSpreglow: is this going to interfere if I reconfig now?
21:15:57jhMikeSguess not
21:16:32stevenmDangit!! Anyone know the equivalent of rm -rf on a sun machine? the -f does nothing!
21:16:44stevenmn1s, try again now
21:17:04*stevenm tries to clean 20 megs of realplayer settings out of his unix account
21:17:04n1sstevenm: got it, thanks
21:18:23stevenmn1s, yay!
21:18:33preglowjhMikeS: this'll only kick in if you try to make a voice file
21:19:49stevenmn1s, I don't remember exactly, but I think the looping logic checks if we have overshot the loop in either end
21:19:58jhMikeSno wonder I was getting lockups, I was calling queue_empty, then skipping pulling the message out :P
21:20:04stevenmn1s, first check is something like, if we are looping and position is before start, reverse direction
21:20:12nanokstevenm: they are stored as bmp screenshots of the configuration windows and dialogues. when you need to restore them, the simple operation of parsing the bitmaps is performed, to figure out the configuration
21:20:17nanoknice and ellegant
21:20:38 Quit spiorf (Remote closed the connection)
21:20:41preglowguess i should start workin making rbspeexenc compile from our libspeex....
21:20:47stevenmn1s, second check is, position is past loop end, and if the patch is looped, turn looping on, and either go back to start, or reverse direction
21:21:07stevenmn1s, some patches loop start->end and back to start, and some go start->end and then end->start, backwards
21:21:08*jhMikeS should have all these stupiditities cleared up shortly
21:21:34stevenm(eith appropriate delta offsets added past the start, or you get honkiness)
21:21:47 Join animeloe_ [0] (n=animeloe@unaffiliated/animeloe)
21:22:04n1sstevenm: ok, got it, will revert my checkout now and see what broke it
21:23:00stevenmn1s, thanks! (I was talking about the ifs on lines 319 and 334 of synth
21:23:24stevenmn1s, but I must run along now. we've got an embedded class. 20 people and one (two?) functioning boards for the whole class. better get it done early!
21:23:55 Quit stevenm ("Connection reset by beer")
21:27:19farhankhmm i've just finished compiling my voice now to see if it works.
21:30:20 Quit animeloe (Read error: 110 (Connection timed out))
21:31:44farhankwow. i got it to work.
21:32:41jhMikeSpreglow: should any .voice files be plopped up there?
21:33:23 Quit scorche|w ("CGI:IRC (Ping timeout)")
21:34:36 Quit amiconn (Nick collision from services.)
21:34:45 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
21:34:59preglowjhMikeS: up where?
21:35:48jhMikeSin a FS task.
21:35:50 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
21:36:55jhMikeSbtw: there's one problem here, the clips are so small that voice thinks it's stopped before it even sounds often times. that needs addressing for sure but it's much snappier
21:37:16preglowwell, that i can do little about :>
21:37:23farhankahh. the talk mp3 clips? I wondered why it was doing that.
21:37:27jhMikeSI'll figure something out to signal it
21:37:30 Join Zagor [0] (
21:37:49farhankI thought it was my talk clips messing up but I guess it isn't me after all.
21:37:52jhMikeSit's things like the "Shutting Down"'ll often not hear them
21:38:35Lloreanfarhank: They're discussing a new voice system, it's not the same thing.
21:38:40farhankwell I compiled my rockbox build for my player after deleting my svn checked out directory, and made a voice and it seems to be working now.
21:38:49jhMikeSvoice plays pcm to the end of the clip unless stopped. who knows, this might fix the truncation problem.
21:39:27jhMikeSI can guarantee no race condition in the threading so noone look there for further oddness :)
21:40:33jhMikeSok...FS up in a few minutes
21:40:41 Join foolsh [0] (
21:40:53preglowi'd better get rbspeexenc ready, then.....
21:41:09preglowmy oh my, how i am bored by build systems
21:42:17 Quit homielowe (Remote closed the connection)
21:42:18jhMikeShaha...I got to do the fun part...well not all of it. you got to work on the codec.
21:42:30preglowwell, i think the codec was the fun part :>
21:42:47preglowit really just required familiarizing and some optimizing
21:43:41 Quit Daniel_S ("CGI:IRC")
21:44:16jhMikeS403 lines for the whole voice thread ain't bad
21:44:42 Quit nicktastic ("Leaving")
21:45:10preglowi'd say it's excellent
21:47:02 Quit farhank ()
21:47:42 Quit foolsh (Remote closed the connection)
21:49:00preglowi need to compile another libspeex, this time in tools/
21:49:16 Join mokkurkalve_ [0] (
21:49:16preglowthree libspeex, this is kinda fun :>
21:51:14 Quit Siku ()
21:52:45preglowshit, i haven't even looked at talk clip generation
21:53:36n1sjhMikeS: I hope you bump the voice version in voicefont and talk.h when that is committed
21:53:37jhMikeSbah! feel free to contribute then ! :P
21:54:04*jhMikeS doesn't know much about the voice file stuff...just the threading
21:54:55 Join Arathis [0] (
21:55:07n1slet's hope someone does it then :) I have no idea what speex will do if we feed it mp3 data...
21:55:14 Quit desowin ("use linux")
21:55:30jhMikeScould test that
21:55:51jhMikeSit should just error and not play it
21:56:06 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
21:57:17jhMikeSpreglow: wait, speex_decode_int returns < 0 on error? I forgot to add a check.
21:57:51preglowjhMikeS: doesn't hurt, i guess :)
21:57:58ZagorBOTD: Fuller's ESB
21:57:58 Join jpt9 [0] (
21:58:06 Quit jpt9 (Client Quit)
21:59:00BagderZagor: you need to spell out "beer" to get the triggers to trigger ;-)
21:59:13Zagorapparently :)
21:59:37markun... and still petur doesn't show up
21:59:41preglowZagor: that is a nice one...
21:59:43preglowclassic esb
22:00:03preglowi want fuller's london porter, haven't tasted that for years :/
22:00:05*petur gets triggered
22:00:22preglowBagder: btw, you might have to change the script that calls configure and builds voice files
22:00:29jhMikeSpreglow: in that case, should I reset the decoder?
22:00:40preglowBagder: since i've removed the encoder selection step
22:00:50preglowjhMikeS: you should always reset it before doing a clip, that should be enough
22:00:58preglowjhMikeS: i assume we don't want to retry on errors?
22:01:15 Join AceNik [0] (n=AceNik@
22:01:24jhMikeSI can't stop voice though, so I have to reset and continue getting queued data
22:01:29 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
22:01:36jhMikeSit might not be the same clip afterall
22:01:37 Part pondlife ("Gone")
22:01:56jhMikeSwill speex bits remaining be 0 after a reset?
22:02:00 Join merbanan [0] (n=banan@
22:03:06 Quit animeloe_ (No route to host)
22:04:05 Join mud_ [0] (
22:04:18preglowmerbanan: being a codec dude, would you know if x264 has a reputation of creating nice, valid bitstreams?
22:05:26merbananpreglow: it should
22:05:27 Part AceNik
22:07:03preglowyeah, i guess so too
22:07:19saratogapreglow: i've never heard of trouble with x264
22:07:20preglowmost things about it seems to look good quality, so i guess it does
22:11:15 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
22:11:44 Join animeloe_ [0] (n=animeloe@unaffiliated/animeloe)
22:13:26 Quit TMM ("Ex-Chat")
22:14:15 Join stewball [0] (n=WTFOMGBB@
22:15:04jhMikeSok, it doesn't give a care about mp3 data. just silence.
22:16:39n1sjhMikeS: that's good but we should really bump the version anyway :)
22:17:21jhMikeStrue but it's should also handle corrupt data gracefully
22:17:31 Quit mud-rb (Read error: 104 (Connection reset by peer))
22:18:32 Join J3TC- [0] (
22:20:24jhMikeSshould the version change for HWCODEC too?
22:20:41preglowargh, i'm close
22:20:59 Quit animeloe (Read error: 110 (Connection timed out))
22:21:06jhMikeSI'll just change it...whatever
22:21:18 Nick parafin is now known as parafin|away (
22:21:51n1sjhMikeS: IMHO it should to avoid having several versions to keep track of (and even be unified with the lng version as they will be at 4 when voice is bumped and very much related)
22:25:41 Quit MethoS-- (Read error: 110 (Connection timed out))
22:28:00 Join Arathis2 [0] (
22:28:15 Join Domonoky_ [0] (
22:28:15 Quit Domonoky (Read error: 104 (Connection reset by peer))
22:30:53 Quit mirak ("Ex-Chat")
22:33:05 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
22:40:16preglowmaking libspeex build in system context is very fun, apparently
22:40:20preglowhow i hate build systems
22:42:08 Join salty-horse [0] (n=ori@pdpc/supporter/active/salty-horse)
22:43:03 Quit animeloe_ (Read error: 113 (No route to host))
22:43:17salty-horsehi. using a build from a few days ago, I notice that when playing a directory and playing the last file, the "next track" in the WPS sometimes displays that last file (instead of nothing)
22:45:31 Quit Arathis (Read error: 110 (Connection timed out))
22:45:33jhMikeShaha, the sound menus work again with the pcmbuf low latency just perfect
22:48:03 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
22:48:43 Quit nicktastic (Read error: 104 (Connection reset by peer))
22:52:01*freqmod_nx hates buildsystems too, I didn't manage to make speex build the Right Way (TM) when I added it
22:53:52 Quit merbanan (
22:53:52 Quit PaulJam (
22:53:52 Quit ATravelingGeek (
22:53:52 Quit weezerle (
22:53:52 Quit tedrock (
22:53:52 Quit davina (
22:53:52 Quit z35 (
22:53:52 Quit Langly (
22:53:52 Quit agm3nt (
22:53:52 Quit robin0800 (
22:53:52 Quit Isolinear (
22:53:52 Quit Soap (
22:53:52 Quit grndslm (
22:53:52 Quit sup (
22:53:52 Quit maddler (
22:53:52 Quit marcosource (
22:53:52 Quit gtkspert (
22:53:52 Quit sslashes (
22:53:52 Quit courtc (
22:53:52 Quit [mbm] (
22:53:52 Quit petur (
22:53:52 Quit ender` (
22:53:52 Quit pixelma (
22:53:52 Quit jgarvey (
22:53:52 Quit Rob2222 (
22:53:52 Quit n1s (
22:53:52 Quit male (
22:53:52 Quit advcomp2019 (
22:53:53 Quit tkooda (
22:53:53 Quit jhMikeS (
22:53:53 Quit Rick (
22:53:53 Quit jumijoze (
22:53:53 Quit kclaf (
22:53:53 Quit joshin (
22:53:53 Quit DraX (
22:53:53 Quit BHSPitMonkey (
22:53:53 Quit newbyx86 (
22:53:53 Quit jmspeex (
22:53:53 Quit XavierGr (
22:53:53 Quit Xerion (
22:53:53 Quit miepchen^schlaf (
22:53:53 Quit billenium (
22:53:53 Quit GodEater_ (
22:53:53 Quit Weiss (
22:53:53 Quit annulus_ (
22:53:53 Quit lodesi (
22:53:53 Quit crashd (
22:53:53 Quit DogBoy (
22:53:53 Quit safetydan (
22:53:53 Quit Zagor (
22:53:53 Quit amiconn (
22:53:53 Quit Lear (
22:53:53 Quit Bagder (
22:53:53 Quit Dark_Apostrophe (
22:53:53 Quit Guile`` (
22:53:53 Quit HellDragon (
22:53:53 Quit iamben (
22:53:53 Quit jhulst (
22:53:53 Quit RaRe (
22:53:54 Quit ivan` (
22:53:54 Quit SirFunk (
22:53:54 Quit martin_ (
22:53:54 Quit GodEater (
22:53:54 Quit FunkyELF (
22:53:54 Quit jepler (
22:53:54 Quit fxb (
22:53:54 Quit TTThomas (
22:53:54 Quit guyzmo (
22:53:54 Quit J (
22:53:54 Quit tuplanolla (
22:53:54 Quit shodanX (
22:53:55 Quit jurrie (
22:53:55 Quit maxkelley (
22:53:55 Quit blithe (
22:53:55 Quit scorche (
22:53:55 Quit DiDjCodt (
22:53:55 Quit [omni] (
22:53:55 Quit Toxicity999 (
22:53:55 Quit preglow (
22:53:55 Quit saratoga (
22:53:55 Quit Llorean (
22:53:55 Quit PaulPosition (
22:53:55 Quit FOAD (
22:53:55 Quit micols_ (
22:53:55 Quit n17ikh|Lappy (
22:53:55 Quit Ave (
22:53:55 Quit Kohlrabi (
22:53:55 Quit pabs (
22:53:55 Quit BrianHV (
22:53:55 Quit feisar (
22:53:55 Quit Domonoky_ (
22:53:55 Quit J3TC- (
22:53:55 Quit alienbiker99 (
22:53:55 Quit hannesd (
22:53:55 Quit thegeek (
22:53:55 Quit atsea- (
22:53:55 Quit Echelon (
22:53:55 Quit solatis (
22:53:55 Quit andrewg867 (
22:53:55 Quit rvvs89 (
22:53:55 Quit lostlogic (
22:53:55 Quit Shaid (
22:53:55 Quit dionoea (
22:53:55 Quit qwm (
22:53:55 Quit Jon-Kha (
22:53:55 Quit Slasheri (
22:53:55 Quit salty-horse (
22:53:55 Quit mud_ (
22:53:55 Quit scorche|w (
22:53:55 Quit nanok (
22:53:55 Quit lee-qid (
22:53:55 Quit idnar (
22:53:55 Quit toffe82 (
22:53:55 Quit DerPapst (
22:53:55 Quit Sedgewick (
22:53:55 Quit daurnimator (
22:53:57 Quit kkurbjun (
22:53:57 Quit lids (
22:53:57 Quit gromit` (
22:53:57 Quit Seed (
22:53:57 Quit DataGhost (
22:53:57 Quit rasher (
22:53:57 Quit midkay (
22:53:57 Quit ryanakca (
22:53:57 Quit sbeh (
22:53:57 Quit lastebil (
22:53:57 Quit parafin|away (
22:53:57 Quit krazykit (
22:53:57 Quit sneakums (
22:53:57 Quit BjoernErik (
22:53:57 Quit bnakiddmj (
22:53:57 Quit Hadaka (
22:54:03 Quit Arathis2 ("Bye, bye")
22:57:02NJoinpabs [0] (
22:58:38 Quit bertrik ("Reconnecting...")
22:59:49 Join Doomed_ [0] (
22:59:49NJoindionoea [0] (
22:59:49 Join andrewg877 [0] (
22:59:49 Join ord [0] (
22:59:49 Join Shaid` [0] (
22:59:49NJoinsafetydan [0] (n=safetyda@rockbox/developer/safetydan)
22:59:49NJoinDomonoky_ [0] (
22:59:49NJoinJ3TC- [0] (
22:59:49 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
22:59:49NJoinamiconn [0] (n=jens@rockbox/developer/amiconn)
22:59:49NJoinalienbiker99 [0] (
22:59:49NJoinhannesd [0] (
22:59:49NJoinsaratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
22:59:49NJoinagm3nt [0] (
22:59:49NJoinender` [0] (
22:59:49NJointhegeek [0] (
22:59:49NJoinrobin0800 [0] (
22:59:49NJoinXavierGr [0] (
22:59:49 Join Llorean [0] (n=llorean@rockbox/administrator/Llorean)
22:59:49NJoinBagder [0] (n=daniel@rockbox/developer/bagder)
22:59:49NJoinXerion [0] (
22:59:49NJoinpixelma [0] (i=pixelma@rockbox/staff/pixelma)
22:59:49NJoinatsea- [0] (i=atsea-@gateway/tor/x-c707381b7def75ce)
22:59:49NJoinjgarvey [0] (
22:59:49NJoinPaulPosition [0] (
22:59:49 Join Dark_Apostrophe [0] (n=darkapos@supporter/monthlybyte/DarkApostrophe)
22:59:49NJoinGuile`` [0] (n=Guile@
22:59:49NJoinRob2222 [0] (
22:59:49NJoinn1s [0] (
22:59:49NJoinHellDragon [0] (i=jd@unaffiliated/helldragon)
22:59:49NJoinIsolinear [0] (
22:59:49NJoinFOAD [0] (
22:59:49NJoinmiepchen^schlaf [0] (
22:59:49NJoiniamben [0] (
22:59:49NJoinjhulst [0] (n=jhulst@unaffiliated/jhulst)
22:59:49NJoinRaRe [0] (
22:59:49NJoinbillenium [0] (
22:59:49 Join ivan` [0] (n=ivan`@unaffiliated/ivan/x-000001)
22:59:49NJoinSoap [0] (n=Soap@rockbox/staff/soap)
22:59:49NJoinmale [0] (
22:59:49NJoinadvcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
22:59:49NJoinBrianHV [0] (
22:59:49NJoingrndslm [0] (
22:59:49NJoinsup [0] (
22:59:49NJoinEchelon [0] (
22:59:49NJoinDraX [0] (n=alex@xmms2/developer/DraX)
22:59:49NJointkooda [0] (i=goaway@
22:59:49NJoinjhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
22:59:49 Join Rick [0] (i=rick@unaffiliated/rick)
22:59:49NJoinFunkyELF [0] (
22:59:49NJoinSirFunk [0] (
22:59:49NJoinmartin_ [0] (
22:59:49NJoinGodEater_ [0] (n=bryan@rockbox/staff/GodEater)
22:59:49NJoinjumijoze [0] (n=jumijoze@unaffiliated/mike-zed)
22:59:49NJoinsolatis [0] (
22:59:49NJoinkclaf [0] (n=kclaf@
22:59:49 Join joshin [0] (n=joshin@unaffiliated/joshin)
22:59:49NJoinWeiss [0] (
22:59:49NJoinGodEater [0] (n=bryan@rockbox/staff/GodEater)
22:59:49NJoinannulus_ [0] (
22:59:49NJoinmicols_ [0] (
22:59:49NJoinjepler [0] (n=jepler@emc/developer/jepler)
22:59:49NJoinBHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
22:59:49NJoinfxb [0] (
22:59:49NJoinlodesi [0] (
22:59:49NJoinmaddler [0] (
22:59:49NJoinn17ikh|Lappy [0] (
22:59:49NJoinTTThomas [0] (
22:59:49NJoinnewbyx86 [0] (
22:59:49NJoinguyzmo [0] (
22:59:49 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
22:59:49NJoincourtc [0] (n=court@unaffiliated/courtc)
22:59:49NJoin[mbm] [0] (i=mbm@openwrt/developer/mbm)
22:59:49NJoingtkspert [0] (
22:59:49NJoinsslashes [0] (i=sslashes@
22:59:49NJoinmarcosource [0] (
22:59:49NJoinShaid [0] (
22:59:49NJoin[omni] [0] (
22:59:49NJoincrashd [0] (
22:59:49NJoinscorche [0] (i=Blah@rockbox/administrator/scorche)
22:59:49NJoinjmspeex [0] (
22:59:49NJoinDogBoy [0] (n=john@unaffiliated/dogboy)
22:59:49NJoinqwm [0] (
22:59:49NJoinfeisar [0] (
22:59:49NJoinlostlogic [0] (n=lostlogi@rockbox/developer/lostlogic)
22:59:49NJoinKohlrabi [0] (
22:59:49NJoinmaxkelley [0] (
22:59:49NJoinrvvs89 [0] (n=rvvs89@pdpc/supporter/active/rvvs89)
22:59:49NJoinDiDjCodt [0] (
22:59:49NJoinjurrie [0] (
22:59:49NJoinblithe [0] (
22:59:49NJoinshodanX [0] (
22:59:49NJointuplanolla [0] (
22:59:49NJoinJ [0] (
22:59:49NJoinpreglow [0] (n=thomj@rockbox/developer/preglow)
22:59:49NJoinAve [0] (
22:59:49NJoinandrewg867 [0] (
22:59:50***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
22:59:50***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
22:59:56***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
22:59:59***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
23:00:09 Quit solatis (Killed by (Nick collision))
23:00:11 Join solatis [0] (
23:00:11NJoinmerbanan [0] (n=banan@
23:00:11NJoinPaulJam [0] (
23:00:11 Join ATravelingGeek [0] (n=ATG@pdpc/supporter/student/ATravelingGeek)
23:00:11NJoinweezerle [0] (
23:00:11NJointedrock [0] (
23:00:11NJoindavina [0] (
23:00:11NJoinz35 [0] (n=z@
23:00:11NJoinLangly [0] (
23:00:13***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
23:00:14 Join n17ikh [0] (
23:00:15 Join bertrik [0] (
23:00:15NJoinsalty-horse [0] (n=ori@pdpc/supporter/active/salty-horse)
23:00:15NJoinmud_ [0] (
23:00:15NJoinnanok [0] (n=nanok@
23:00:15NJoinlee-qid [0] (
23:00:15NJoinidnar [0] (i=mithrand@unaffiliated/idnar)
23:00:15NJointoffe82 [0] (
23:00:15NJoinDerPapst [0] (
23:00:15NJoindaurnimator [0] (n=daurnima@unaffiliated/daurnimator)
23:00:15NJoinkkurbjun [0] (
23:00:15NJoinlids [0] (
23:00:15NJoingromit` [0] (
23:00:15NJoinJon-Kha [0] (
23:00:15NJoinSeed [0] (
23:00:15NJoinSlasheri [0] (i=miipekk@rockbox/developer/Slasheri)
23:00:15NJoinDataGhost [0] (
23:00:15NJoinryanakca [0] (n=ryan@ubuntu/member/ryanakca)
23:00:16NJoinrasher [0] (n=rasher@rockbox/developer/rasher)
23:00:16NJoinmidkay [0] (n=midkay@rockbox/developer/midkay)
23:00:16NJoinsneakums [0] (
23:00:16NJoinkrazykit [0] (
23:00:16NJoinparafin|away [0] (
23:00:16NJoinbnakiddmj [0] (i=0@
23:00:16NJoinBjoernErik [0] (
23:00:16NJoinlastebil [0] (
23:00:16NJoinsbeh [0] (n=sbeh@
23:00:16NJoinHadaka [0] (
23:00:16 Quit alienbiker99 (Read error: 104 (Connection reset by peer))
23:00:18 Quit DerPapst (Excess Flood)
23:00:18*jhMikeS takes his dramamine
23:00:19 Join DerPapst [0] (
23:00:31DerPapstonly 27 survived ^^
23:00:40 Join scorche|w [0] (
23:00:41***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
23:00:54 Quit solatis (K-lined)
23:03:45 Join animeloe_ [0] (n=animeloe@unaffiliated/animeloe)
23:04:16CtcpVersion from freenode-connect!freenode@freenode/bot/connect
23:05:57PaulJamsalty-horse: this is already known, see FS #8148
23:07:44 Quit robin0800 (" Try HydraIRC -> <-")
23:08:04***Saving seen data "./dancer.seen"
23:08:34salty-horseeven though it's still "unconfirmed"
23:08:48 Quit andrewg867 (Read error: 110 (Connection timed out))
23:08:52 Quit n17ikh|Lappy (Connection timed out)
23:09:41 Join kugel [0] (i=kugel@unaffiliated/kugel)
23:10:16 Quit Domonoky_ (Read error: 110 (Connection timed out))
23:10:43 Quit Shaid (Connection timed out)
23:10:44 Nick Shaid` is now known as Shaid (
23:11:50 Quit Soap_ (Read error: 104 (Connection reset by peer))
23:11:56 Join Soap__ [0] (
23:12:18 Quit qwm (Connection timed out)
23:13:11preglowgod, i'm going to need hard liquor to maintain interest soon
23:14:19jhMikeSwhat else needs doing?
23:15:27preglowjust compiling a version of libspeex for linking with rbspeexenc
23:15:36preglowbut i can't bloody make makefiles behave
23:15:44*DerPapst donnates preglow some nice wodka with 40% water
23:15:54*Dark_Apostrophe drools
23:16:04scorche|wnot vodka....wodka
23:16:09Dark_ApostropheMmm... vodka
23:16:10 Quit animeloe (No route to host)
23:16:26*preglow wants whisky
23:16:42Dark_ApostropheWhiskey's got too strong of a taste IMHO.
23:16:51scorche|whow about a nice scotch?
23:16:55Dark_ApostropheI like more neutral things, like rum or vodka
23:17:04markunpreglow: shivas regal ok?
23:17:26markunoops :)
23:17:45 Join animeloe__ [0] (n=animeloe@unaffiliated/animeloe)
23:19:49 Join jac0b [0] (
23:19:56preglowmarkun: sure!
23:19:59 Quit J3TC- (Read error: 104 (Connection reset by peer))
23:20:15jac0bthis task is now closed
23:20:57Zagorwhat is the friendly way to borrow the codec/plugin iram space?
23:21:20 Quit merbanan (Remote closed the connection)
23:23:08jac0bis there another patch that is similer to the scrolling margins patch?
23:24:02scorche|wjac0b: it was closed because a similar patch was committed
23:24:04PaulJamjac0b: a somilar functionality mate its way into the official build (but it is only possible to set the left margin)
23:24:07jhMikeSpreglow: linuxstb usually straightens me out on those pretty quick
23:25:10 Quit saratoga ("CGI:IRC (EOF)")
23:25:23BagderZagor: iirc, plugins and codecs have a set amount of iram
23:25:50Bagderand we have no way to dynamically change that
23:26:04preglowZagor: what do you want it for? :>
23:26:19ZagorBagder: that's fine, I just want to steal it for a while in usb (for testing). but all i've found so far are #defines in playback.c
23:26:40jac0bokay thanks
23:26:52PaulJami thought plugins have no own iram at all (and need to steal it from playback if they need it)
23:26:55Zagorpreglow: testing the theory that my bulk problems are due to slow dram
23:27:35 Join J3TC- [0] (
23:28:42Zagorlooking at it from another angle, can I simply use it during usb? will playback/codecs go nuts and/or will it write stuff to it?
23:28:46 Quit Thundercloud (Remote closed the connection)
23:29:04jhMikeSZagor: bulk problems?
23:29:30ZagorjhMikeS: yes, that's what holding back usb currently. I can only do small bulk transfers, not large ones.
23:29:38jhMikeSwhat happens?
23:29:43Zagorbit errors
23:30:18Zagoror at least I think that's what happens. the host rejects the transfers with various error codes that I've interpreted as bit errors
23:31:25 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
23:31:28 Quit Frazz (Read error: 104 (Connection reset by peer))
23:31:43 Quit SirFunk (Remote closed the connection)
23:32:05 Quit animeloe_ (Read error: 110 (Connection timed out))
23:32:36Zagorbtw, are there any windows-using developers that can help me compile a win32 binary of my testing program? I don't have a win32 compiler.
23:32:50*jhMikeS has one
23:33:04amiconnZagor: cygwin...
23:33:05Zagordo you have a few minutes to spare?
23:34:27Zagoramiconn: good point
23:35:08preglowor mingw
23:35:11preglowwhich is even faster to set up
23:35:16*preglow never liked cygwin
23:35:27Bagderand mingw is fine as cross-compiler on debian...
23:36:00Bagderthat came out wrong, I bet it works fine on all distros
23:36:01 Join colin_ [0] (
23:36:03n1sZagor: there's audio_iram_steal() in playback.c, not really sure when that's needed though...
23:37:00Zagorn1s: yeah I saw that too, but I didn't quite see when/how it is used
23:37:02jhMikeSthere won't be
23:37:24n1sZagor: only used by plugins that use iram afaict
23:37:46jhMikeSyou'll just have to stop playback after core voice
23:38:01linuxstbZagor: apt-get install mingw32 ?
23:38:06kugelLlorean: can you read what I wrote in #rockbox-community and answer me here (aka I forgot that it's off-topic there and I am to lazy to type it here again)
23:38:11n1sjhMikeS: nice that the voice works while in ll mode now :)
23:38:13*linuxstb is too late...
23:38:50 Quit animeloe__ (No route to host)
23:39:11preglowwaaaha :/
23:39:41 Join BoD[] [0] (
23:39:52 Quit davina ("xchat on Ubuntu 7.04")
23:40:12BoD[]I'm trying to use RBUtil on a video ipod (60gb)
23:40:58BoD[]when I install the bootloader it says "it's already up to date"
23:41:35jhMikeSn1s: It can breakup at time but usually not and will get the clip out. Has to do with timing in the pcm buffer.
23:41:41n1sBoD[]: so what happens if you reset your ipod (holding MENU+SELECT?)
23:42:08Lloreankugel: I didn't change the text editor keymap
23:42:22n1sjhMikeS: currently it gets about 1/2 second of the first clip out so it's a big improvement :)
23:42:45BoD[]oh that's right I have to reboot it in a special mode?
23:42:58jhMikeSon h100 and e200 they finish but might break and complete
23:43:24 Join JdGordon [0] (
23:43:26n1sBoD[]: MENU+SELECT is the only way to really reboot it otherwise the apple firmware will just sleep
23:43:27kugelBut your commit still changed the line options menu from power button to down button
23:43:45n1sreally I'm certain the manual mentions this stuff
23:44:05BoD[]n1s: well I just did it
23:44:21BoD[]but it can't autodetect
23:44:29amiconnFiddling with the PP mono lcd bridge has a nice effect
23:44:35BoD[]and so I select it manually
23:44:45BoD[]but then it says "No Ipods Found"
23:44:47*amiconn can now use mpegplayer on mini G2 fullscreen without the audio stuttering :)
23:44:57n1sBoD[]: you restarted the ipod like I told you, did it display anything while booting?
23:45:02Lloreankugel: If so, it's inappropriately using either the WPS or the List context.
23:45:34BoD[]n1s: well it started the apple firmware
23:45:36Lloreankugel: Are you 100% sure it's my change that did it?
23:45:36amiconnNow I just need to implement my all-new greyscale lib idea for grain-free display of moving content
23:45:47kugellook into text_editor.c, it doesn't use any context. It uses the actions directly
23:46:05n1sBoD[]: did you install rockbox? or just the bootloader?
23:46:08kugeli want to say, it's using action_std
23:46:45BoD[]n1s: I didn't install anything yet
23:46:58kugeland you changed action_std_cancel to button_down, so it changed in the text editor too
23:46:59BoD[]n1s: it tried to install the bootloader but didn't work
23:47:04preglowif you do a makefile target that has no deps, it'll always be executed, yes?
23:47:27BoD[]wasn't there a special mode that you could set when booting the ipod?
23:47:30amiconn256fps at 80MHz (!)
23:47:36n1sBoD[]: did it tell you you had a bootloader on the very _first_ attempt and you are _certain_ you don't?
23:47:37jhMikeSamiconn: but H120 the graylib has such a nice silver-halide look like photos from the 1800's :)
23:47:38 Join jpt9 [0] (
23:47:39Lloreankugel: So the power button is now cancel?
23:47:51LloreanWhat _exactly_ is the problem? For example, I have no idea what you mean by "overmapped"
23:47:53 Quit colin_ ()
23:47:55kugelthat was just an example
23:47:59LloreanWell, the power button is action_std_cancel..
23:48:06kugelcancel is on button_left, as before
23:48:20amiconnjhMikeS: That look won't change; the H1x0 doesn't suffer from the graininess in moving content because the excessive slowness of the panel
23:48:34LloreanOh, no, Left is cancel
23:48:44LloreanSTOP is power
23:48:51jhMikeSamiconn: it does have the wave that moves down the display though
23:48:54LloreanAnyway, how is a button "overmapped" using the standard context?
23:48:56kugelthe problem is, that the text editor used action_std_cancel for the list options
23:49:01BoD[]n1s: well actually my ipod had rockbox on it, then I "reinitialized" with itunes. I thought it was like starting with a "fresh" ipod. Am I correct?
23:49:02amiconnThat wave can't be removed
23:49:14kugelbut the left button is actually cancelling the editor
23:49:18Llorean"for the list options" doesn't mean anything to me...
23:49:23LloreanTo cancel the list options?
23:49:24jhMikeSof course just goes with the look of old film
23:49:31LloreanFor is a very non-descriptive verb.
23:49:35amiconnBut I will probably make a 'greylib tuning' plugin (based on test_scanrate, plus gamma adjustment)
23:49:46kugelthe text editor has two menues
23:49:46n1sBoD[]: I have never used itunes on an ipod so idk what it does but if you are told you have a bootloader I'd assume you do so install rockbox :)
23:49:53jhMikeSLlorean: :) /me is Foring today?
23:49:54 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
23:49:56amiconnThat'll be for advanced users who want the best possible greyscale quality
23:50:03BoD[]n1s: I'll try
23:50:41kugelone is like a file menu. it's on the left key, and only opens when you've made changes. otherwise the button will just quit the editor
23:50:41Lloreankugel: That still doesn't tell me what you mean. To bring up the menu?
23:50:52Lloreanscorche|w: The point was that "for" ISN'T a verb, and the whole thing was lacking one.
23:51:12linuxstbBoD[]: I don't think rbutil will know that you've deleted a bootloader by restoring with itunes. So reinstall it anyway.
23:51:13*amiconn is looking for a 4th gen grayscale or mini 1st gen owner to test some things
23:51:13BoD[]by the way this utility is very cool :)
23:51:17Lloreankugel: That's not a menu. That's "Confirmation of saving on exit"
23:51:20kugelthe other menu contains insert above and concat to above etc
23:51:32linuxstbBoD[]: Forget what I just said...
23:51:42BoD[]linuxstb: uh ok :)
23:51:44kugelOk, then it's not a menu
23:51:52kugelbut you know what I'm talking about
23:51:55Lloreankugel: Now I do.
23:51:57BoD[]linuxstb: well I'm installing rockbox right now, we'll see if the bootloader works
23:52:00Lloreankugel: Now, what _exactly_ is the problem?
23:52:02linuxstbBoD[]: But it doesn't hurt to install the bootloader multiple times...
23:52:03jhMikeSSome things are more for than others
23:52:11 Quit salty-horse ("Leaving")
23:52:14BoD[]the bootloader will just start the apple firmware if rockbox isn't there, is that it?
23:52:19LloreanAs I didn't change anything about action_std_cancel, I don't see how I could've caused a problem with it.
23:52:31n1sBoD[]: right
23:52:36kugelthe line options menu was on power button before your commit
23:52:43kugelnow it's on button_down
23:52:48 Quit jac0b ()
23:52:52BoD[](installing themes...)
23:52:57kugeland power button is unmapped
23:53:01*jhMikeS is getting used to the new keymap and will probably start fighting to not go back to the old way soon :)
23:53:08Lloreankugel: This still doesn't tell me what's wrong.
23:53:15LloreanAll it tells me is "the keys are different"
23:53:24LloreanDoes something else use button_down directly?
23:53:26preglowlinuxstb: you didn't happen to do that libspeex native build thing yesterday, did you? :>
23:53:38kugelremoving a line was on button down
23:53:41Lloreankugel: Then that's a bug in the text editor.
23:53:53LloreanA plugin should DEFINITELY not mix button actions with direct button references
23:54:20kugeland, canceling(quiting/"Confirmation of saving on exit") is not button_left, which doesn't make sense to me
23:54:25kugelnot = on
23:54:25LloreanIt should either always use ACTION_context_actualaction notation, or BUTTON_button notation, but not a mix
23:54:56Lloreankugel: I told you, I didn't change action_std_cancel, so I don't see how if it uses that to determine cancelling I could've been at fault.
23:55:22kugelyou changed action_std_cancel to button_down
23:55:35kugelno you didn't
23:55:38LloreanNo, I didn't.
23:56:35kugelbut afaik, the text editor uses action_std_cancel which is mapped to button_down in that context (and was power button previously)
23:57:12LloreanWhatever caused this problem, if it relates to a change in that function, has nothing to do with my patch.
23:57:14BoD[]bootleader doesn't seem to be there :( Apple firmware loaded after menu+select
23:57:39kugelIt's a general problem with the button map in test editor
23:57:50LloreanThen file a bug report, and leave my patch out of it.
23:58:05linuxstbBoD[]: Which version of rbutil are you using? (it should say in the About menu)
23:58:26jpt9the rockbox site says that it doesn't have any USB support on the Sansa...
23:58:31BoD[]linuxstb: revision 15190 (m1.0.2)
23:58:31jpt9what does that mean in terms of charging?
23:58:51linuxstbBoD[]: Then download the latest - 1.0.2 had a bug which stopped it installing the ipod bootloader...
23:58:56kugelin text_editor.c is written something like if (action_std_cancel) line options menu
23:58:59jpt9it has a little plug icon next to the battery meter; is it actually charging, or is it just running off AC power?

Previous day | Next day