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

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

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

#rockbox log for 2005-11-11

00:00:36preglowman, amiconn could have done wonders for this mandelpod app
00:00:56amiconnIs it slow? ;)
00:01:00preglowog yes
00:01:07preglowoh yes, that is
00:01:19amiconnhehe, mandelbrot.rock is rather fast even on archos :-P
00:01:43preglowthe coldfire one is heaps faster than this one
00:01:46preglowbut this is colour, of course
00:02:47amiconnDoesn't matter at all
00:02:52amiconn(unless they do cpm)
00:03:09 Quit korpse (Remote closed the connection)
00:03:12 Join korpse [0] (
00:08:16 Quit korpse (Remote closed the connection)
00:08:31 Join korpse [0] (
00:09:17preglowoh well
00:09:21preglowenough of fun and games
00:12:42 Join webguest06 [0] (
00:13:34 Quit korpse (Remote closed the connection)
00:13:55 Quit webguest06 (Client Quit)
00:15:25 Part len0x
00:16:19 Join Gibbed [0] (
00:16:20 Quit Rick (Nick collision from services.)
00:16:28 Nick Gibbed is now known as Rick (
00:18:58 Join korpse [0] (
00:19:12preglowlinuxstb: not only the printf triggers it
00:19:16preglowlinuxstb: other things can as well
00:21:57linuxstbIt's very annoying. I can't understand the reason for it.
00:22:32 Quit muesli- (Read error: 110 (Connection timed out))
00:22:40linuxstbThe startup code seems to be working fine - otherwise I am sure we would get lots of other strange behaviour.
00:24:02 Quit korpse (Remote closed the connection)
00:24:19linuxstbI don't think it's a hardware initialisation bug - because a particular build either works or it doesn't - consistently.
00:24:36preglowit must be overflow
00:24:40preglowi'm almost convinced
00:24:45preglowthe graphics survives
00:24:48preglowbut the font doesn't
00:24:56preglowbut how, i don't know
00:25:13preglowthen again, it can't be
00:25:22preglowyou write tons of long strings to bug via snprintf
00:25:26preglowone more shouldn't matter
00:33:10linuxstbI wonder if it's my choice of gcc options causing a problem.
00:33:32linuxstbThey are in tools/configure - search for arm
00:34:16 Join korpse [0] (
00:34:22preglowfomit pointer and mcpu?
00:34:25preglowseriously doubt it
00:34:41 Quit |joshn| ("KVIrc 3.2.0 'Realia'")
00:34:57preglowbtw, shouldn't you use −−march?
00:35:06preglow−−mcpu makes code that is compatible with lower models
00:35:27preglowat least that's the way it is for x86
00:36:36linuxstbNo, I don't think so - I've just read the gcc man page again.
00:37:35preglowright, it's different on arm
00:39:19 Quit korpse (Remote closed the connection)
00:39:31 Join korpse [0] (
00:41:45 Quit Kohlriba ("Leaving")
00:44:30*preglow thinks it's time to split up crt0.S ...
00:44:35 Quit korpse (Remote closed the connection)
00:44:45 Join korpse [0] (
00:45:52linuxstbpreglow: I did start that way, but then decided to merge them. But I don't feel strongly either way.
00:46:38preglowstrongly no, but i think it's more practical that way
00:47:12preglowno matter
00:48:51 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:49:49 Quit korpse (Remote closed the connection)
00:51:14 Join korpse [0] (
00:56:52***Saving seen data "./dancer.seen"
01:02:37 Quit andy ()
01:16:15preglownow there's a surprise
01:16:20preglowa successful ubuntu upgrade
01:20:01linuxstbhehe. Fancy testing if the ATA driver can write to your flash? I haven't tested the writing part of the driver yet :)
01:21:53preglowjust need a sec to finish a mail here
01:22:10preglowyou haven't tested writing on your unit either? :)
01:23:31linuxstbBut maybe we should be sensible - we don't need to write at the moment.
01:26:10preglowwell, i doubt it can harm
01:26:11preglowand besides
01:26:15preglowit needs testing sooner than later
01:26:19preglow_OR_ later
01:26:43preglowi'll just finish an email, then i'll test it
01:27:15 Join webguest75 [0] (
01:27:41linuxstbMaybe it could be useful - e.g. writing debugging information to a file.
01:32:14 Join matsl [0] (
01:33:25 Quit Moos ("Glory to Rockbox")
01:33:40 Quit webguest75 ("CGI:IRC")
01:36:49preglowok, let's go
01:37:10preglowany specific tests?
01:37:17preglowjust open and close, or write a bit as well?
01:37:49linuxstbI guess just open and close to start with. That will do some writing.
01:39:08 Quit matsl ("ERC Version 5.0 (CVS) $Revision: 1.776 $ (IRC client for Emacs)")
01:39:32 Part linuxstb
01:41:20 Join linuxstb [0] (
01:41:36preglowdoesn't look good
01:42:06preglowcan any of the swedes here tell me what 'fena' means? :)
01:42:10linuxstbOh. I won't try it as well then.
01:42:19preglowlinuxstb: i'll give it another go
01:44:05linuxstbI've just tried it. Nothing broke. Lets see if the file was created.
01:44:33preglowlinuxstb: works :)
01:44:39preglowi'll try to write some data
01:45:00linuxstbNo, nothing... I forgot the / at the start of the filename.
01:46:15preglowfirst string written on ipod by rockbox:
01:46:57preglowso looks to me like your driver works just perfectly
01:47:45linuxstbWell it's mostly the existing Rockbox driver. I didn't need to make many changes at all.
01:49:05 Quit yngwi ("Chatzilla 0.9.68a [Firefox 1.0.7/20050919]")
01:49:20linuxstbIt worked for me as well. I wonder where Rockbox got the file timestamp from.
01:50:37preglowgod knows
01:50:50preglowbut i need to go to beed soon
01:51:01preglowi'll vanish for a couple of days, but will return for more hacking soon
01:51:39preglowfirst: check if ubuntu can still boot after upgrading, brb
01:54:57preglowlike i thought......
01:55:15preglowi love nvidia
02:00:20linuxstbTime for bed now. See you later.
02:05:04 Quit ashridah ("haircut")
02:07:56amiconnlinuxstb: On units without an rtc, rockbox bases the time stamp for new files on the build date
02:10:11preglowipods do have rtcs :)
02:10:37preglowwell, my ipod's got a freakin' clock
02:12:08amiconnI guess this clock is wrong after experimenting with ipl or rockbox...
02:15:25preglowoh, i doubt he's enabled it yet
02:16:49preglowlots and lots of code left to be written
02:17:22amiconnHmm, iiuc the ipod does have a clock, but no rtc chip?
02:17:40preglowwell, an ipod never truly switches off
02:18:17preglowi've switched mine on and off, written to it extensively, run alternatie firmwares, reset the entire flash
02:18:21preglowand the clock is still correct
02:18:40preglowthe settings, like 24-hour clock, is forgotten
02:18:43preglowbut the time is still right
02:18:55pregloweven the date is right
02:19:18amiconnI would expect that with a software clock, if you run ipl or rockbox, the clock would stop working
02:19:25preglowyes, me too
02:19:31amiconnMaybe the ipod does have an rtc chip after all?
02:19:41preglowthere are a couple of regs that end with _RTC
02:20:09preglowthey might of course just be called so rather arbitrarily
02:20:11amiconnHmm, perhaps is part of the pp?
02:20:19preglowbut it does look like it's got an rtc
02:20:20 Join DJDD [0] (
02:20:23preglowyes, i'd expect that, actually
02:20:25amiconn*perhaps the rtc is...
02:20:54 Quit ehntoo (Read error: 104 (Connection reset by peer))
02:25:06preglowtime to schlafen
02:31:00 Join hd [0] (i=hd@
02:47:48 Quit goa (Read error: 110 (Connection timed out))
02:47:49 Nick hd is now known as goa (i=hd@
02:56:55***Saving seen data "./dancer.seen"
03:07:23 Quit BirdFish ("system reboot")
03:21:38 Join BirdFish [0] (n=bradbox8@
03:24:05 Quit DJDD (Read error: 110 (Connection timed out))
03:38:01 Join XavierGr [0] (
03:51:33 Quit dpassen1 ()
03:52:08 Nick BirdFish is now known as BirdFish[a] (n=bradbox8@
03:55:54 Join ashridah [0] (
03:58:07 Join StrathAFK [0] (
04:09:20 Join DJDD [0] (
04:14:32 Quit Strath (Read error: 110 (Connection timed out))
04:30:00 Join solexx [0] (
04:32:00 Quit hardeep ("BitchX: coming soon to a theatre near you!")
04:41:51 Quit solexx_ (Read error: 110 (Connection timed out))
04:56:59***Saving seen data "./dancer.seen"
05:07:06 Quit DJDD (Read error: 110 (Connection timed out))
05:26:20 Join ehntoo [0] (
05:31:46 Nick BirdFish[a] is now known as BirdFish (n=bradbox8@
05:42:49 Quit BirdFish ()
05:58:29 Join Mxm`Pas`Bien [0] (
05:58:30 Quit Maxime (Read error: 104 (Connection reset by peer))
06:01:48 Join BirdFish [0] (n=bradbox8@
06:02:04 Quit RotAtoR ()
06:03:29 Join |joshn| [0] (
06:13:08 Join Life_ [0] (
06:13:22 Part Life_
06:42:23 Join chiller_ [0] (
06:43:21 Quit chiller (Read error: 104 (Connection reset by peer))
06:43:22 Nick chiller_ is now known as chiller (
06:57:03***Saving seen data "./dancer.seen"
07:36:24 Join ender` [0] (i=ychat@
08:12:04 Join Gibbed [0] (
08:12:04 Quit Rick (Nick collision from services.)
08:12:12 Nick Gibbed is now known as Rick (
08:13:15 Join lefright [0] (
08:18:42 Part lefright
08:26:33 Join B4gder [0] (n=daniel@
08:28:31 Join einhirn [0] (
08:45:35 Join Zagor [0] (
08:45:36 Join _FireFly_ [0] (
08:57:06***Saving seen data "./dancer.seen"
09:33:45 Join LinusN [0] (
09:36:16 Quit phaedrus961 ("Leaving")
09:37:55 Quit arkascha (Read error: 110 (Connection timed out))
09:59:47 Quit ashridah (Read error: 110 (Connection timed out))
10:07:34 Join DJDD [0] (
10:11:36 Join phaedrus961 [0] (
10:44:05preglowswedes: what does 'fena' mean? :)
10:44:18LinusNa fin
10:44:23LinusNon a fish
10:44:45LinusNbut you can also say "att vara en fena på något"
10:44:51preglowahh, yes
10:44:58preglowthat's what he means, probably
10:45:03LinusNwhich means to be very good at something
10:45:09preglowyes, kind of figured that
10:45:25LinusNnorwegian isn't that much different from swedish
10:45:59preglowyou don't say
10:46:10preglowthis one word i hadn't picked up, anyhow
10:50:32 Join wubbla [0] (
10:50:35wubblahoi! :)
10:53:55 Join Jungti1234 [0] (n=jungti12@
10:57:09***Saving seen data "./dancer.seen"
10:59:44wubblais there an opensource win32 cvs (commandline) client?
11:00:46wubblareally? :)
11:00:50wubblawhere to get it?
11:01:41B4gderthere are several being one
11:02:24B4gderbut since we are rockbox here, windows devel means cygwin
11:02:31B4gderand cygwin comes with a command line cvs
11:03:10wubblaand isn't it possible to build it on linux too?
11:03:34B4gderbut you asked for win32
11:03:55 Join webguest64 [0] (
11:04:12wubblahehe, you're right :)
11:04:41wubblawouldn't mingw32 be sufficient for win32?
11:04:48wubblabecause i already have that...
11:05:01B4gderunless you feel like making it work on your ownw
11:05:58preglowmingw will probably do for sim compiles
11:06:11B4gderyes it might
11:06:15preglowbut for target compiles, you'll need a cross compiles as well, which i'm pretty sure are easiest to build with cygwin
11:06:23B4gderbut you'll need perl, make, sh and things setup too
11:06:55wubblahmm... my original intention was to experiment with the simulator
11:07:03amiconnBtw, I once tried using microsoft sfu. No go, the make command doesn't even know the -C option
11:07:12wubblaas the h300 port isn't yet in a usable state :)
11:07:39B4gderwubbla: still, going with something else than cygwin on windows WILL cause you grief and work
11:07:44markunwubbla: Just install Linux, makes it all easyer :)
11:07:54 Join Sandking [0] (
11:07:58B4gderamen to that
11:08:06wubblamarkun: i have got linux
11:08:21markunso there is no problem
11:08:32wubblamarkun: but i've just got shell access to these machines...
11:08:42wubblamarkun: so, no X
11:09:01markunAre there X servers for windows?
11:09:07wubbla...and therefore no simulator (I assume...)
11:09:12B4gdermarkun: yes cygwin ;-)
11:09:30markunyou can run putty with X forwarding I think.
11:11:12wubbla already contains cygwin, right? :)
11:11:44B4gderthat's a stripped down cygwin
11:11:51B4gderwith the cross compiler included
11:11:57 Join amiconn_ [0] (
11:12:26wubblanow that's something i'd call "user-friendly" :)
11:12:53 Join ashridah [0] (
11:18:00 Quit Jungti1234 ("Bye~ -")
11:18:52B4gderI'm not sure I agree
11:19:24 Join Jungti1234 [0] (n=jungti12@
11:20:25wubblahmm... afaik rockbox relies on C...
11:20:45wubblawouldn't it be possible to use C++?
11:20:55B4gderyes it would, but we don't
11:20:59wubblaor would this be too much "high-level"?
11:21:34*ashridah shudders at the code-size explosion c++ could bring in the hands of the wrong compiler
11:21:58B4gderwubbla: we just don't need C++
11:22:27preglowmmm... c+++
11:22:30preglowbut no
11:22:35preglowc is better for our needs
11:22:59solexxc#, c#!
11:23:04solexx(just kidding)
11:23:10wubblaat least for me, it's much easier to write c++ code that c...
11:23:24wubblaespecially when using the STL :)
11:23:58Zagorwubbla: that's why you don't write code for tight embedded systems :-)
11:23:59*ashridah flees
11:24:05wubblahehe :)
11:24:46*ashridah shudders at the thought of getting 2-page error messages from rockbox's compilation process
11:25:32B4gder"I can see you're not in the softwar business" B-]
11:25:42wubblai definitely agree that the kernel as well as the drivers need to be written in C/asm... but what about the plugins?
11:26:07B4gderyou should be able to write plugins in C++
11:26:26B4gderno one's been that crazy yet ;-)
11:26:27ashridahwubbla: i doubt there's much stopping you using it right now. it's just that the plugin interface will still be C :)
11:27:11wubblahmmm... a C++ wrapper for the plugin-interface would be nice
11:27:45B4gderyes, and a python one
11:27:51wubblahehe :)
11:28:23*ashridah watches wubbla stumble headfirst into a sore spot
11:29:13 Quit amiconn (Read error: 110 (Connection timed out))
11:29:14 Nick amiconn_ is now known as amiconn (
11:31:09wubblahow many people are actively contributing (developing) to rockbox?
11:31:14 Join leftright [0] (
11:32:12B4gderthe credits list is >100 names by now
11:32:19B4gderwe're some 30 committers or so
11:32:30B4gderthe dev list is >100 people
11:32:56wubblathis is insane :)
11:33:48leftrightfor info, Replay gain values are not displayed in ID3v2 info if Predixis Musicmagic Mixer software is used on the track
11:33:48B4gderdevotion I'd say
11:34:20wubblaargh...i can't wait for the h300 port :)
11:35:09markunaccording to Jungti1234 there are over 20 000 H3xx users in Korea alone.
11:35:59markunI wonder how long it will take before the #rockbox toppic is in Hangul :)
11:36:57B4gderyou mean in "#rockboxkr" ? ;-)
11:38:04markunLike that is now completely taken over by Brazilians
11:40:49wubblain apps/plugin.h are all functions which plugins have access to?
11:45:20wubblaso i will never ever need to include another header when just writing plugins? :)
11:45:46preglowwubbla: stl is pretty much what makes c++ generated code bloated
11:46:23wubblapreglow: yeah, i know
11:46:32B4gderwubbla: there are quite a few existing plugins you can check
11:46:38wubblapreglow: but it makes developement a lot easier
11:46:45preglowwubbla: sure, i love stl myself
11:46:56wubblapreglow: because you don't need to care about memory management
11:47:02preglowtoo bad it's the only part of the c++ standard "library" that's usable
11:47:30preglowwubbla: couple stl together with boost::shared_ptr and you truly never have to worry about memory management
11:48:00wubblapreglow: i'd like to see boost too in rockbox, but that would surely be overkill :)
11:48:44wubblayeah, shared_ptr's are extremely powerful
11:49:06leftrightLear: any chance of looking into this Predixis MusicMagic tag issue, replay gain doesn't work if MusicMagic has added hashed info to the tag
11:49:27wubblai'm rarely using raw-ptr's in my source-code since i discovered shared_ptr as well as scoped_ptr :)
11:51:55preglowusing raw ptrs in c++ is a sin these days
11:52:59leftrightbut Foobar and tag editors see the replay gain info regardless
11:53:29leftrightjusr rockbox which doesn't see/apply the info
12:00:21 Quit markun ("leaving")
12:03:29 Quit linuxstb (Read error: 110 (Connection timed out))
12:04:34 Join linuxstb [0] (
12:05:18 Join Moos [0] (
12:09:59 Quit Sandking ()
12:15:43 Join nodak [0] (
12:15:52 Quit linuxstb (Remote closed the connection)
12:16:22nodakcould U help me with my olympus mROBE 500i??
12:19:20 Quit nodak (Client Quit)
12:37:56dwihnoI was thinking... Isn't it some kind of requirement to "stop" USB devices, such as flash card readers? Just umounting the device sounds a bit dangerous...
12:38:04dwihno(before unplugging, that is)
12:40:03LinusNdoesn't sound dangerous to me
12:40:12LinusNusb is a hot-plug architecture
12:43:14preglowdwihno: there's always some oses using write caching
12:43:22preglowdwihno: but unounting should flush it
12:43:28preglowunmounting, even
12:44:00*preglow feels an urge to play some iDoom
12:45:55dwihnoHow about those computers with flash card readers...
12:46:21dwihnoAre the cards merely acting like floppy devices, or does it depend on the type of card reader?
12:46:49preglowsome of them are usb
12:47:07dwihnoUSB ones should be stopped then?
12:47:24preglowi think the driver takes care not to enable write caching
12:47:34preglowon those devices, you don't even need to unmount
12:51:22dwihnobut as long as devices are unmounted, it's safe to unplug?
12:51:49preglowcan't ask for much more
12:52:15dwihnoPerhaps I'm just being scared breaking something
12:52:32preglowwhat os do you use?
12:53:04preglowin linux you should just unmount, and you'll be fine
12:53:07dwihnowindows 2000, xp and freebsd
12:53:28preglowwindows automatically disables write caching for all usb devices i have, so there you can just yank the cable out when you're done without unmounting anything
12:54:18_FireFly_that's not true at all
12:54:39preglowno, it's true
12:54:50dwihnoYou want the truth? You can't handle the truth ;)
12:54:54_FireFly_i had a data-loose after i had unplugged my "usb-stick" without "eject" it
12:54:58preglownotice the usage of the words 'all usb devices i have' there
12:55:21_FireFly_then it's only for you but
12:55:37preglowyes, but it should serve as an indication
12:55:54preglowduring all my h120 development, i've never clicked the nagging unplag-button thing
12:56:00 Join kurzhaarrocker [0] (
12:56:00preglowand it still lives, no data corrupted
12:56:10 Part kurzhaarrocker
12:56:13_FireFly_me either ;) because i'm not useing it under windows ;=)
12:56:17preglowthe same goes for ipod and camera
12:57:11***Saving seen data "./dancer.seen"
13:04:15 Join DangerousDan [0] (
13:17:17 Quit XavierGr ("Trillian (")
13:27:20preglowback in a couple of days
13:27:24 Part preglow
13:37:33 Join ep0ch [0] (n=ep0ch@
13:51:42B4gderwe should make WPSes capable of loading bmps in "current directory"
13:51:45 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
13:51:54B4gderthen we can start bundling WPSes in builds
13:52:54amiconnThere are problems with that (assuming you mean the wps should load images from the dir where the .wps is located
13:53:16B4gderenlighten me!
13:53:24amiconnCurrently a .wps has to reside directly in /.rockbox in order to be remembered across reboots /.rockbox may become extremely cluttered
13:54:22amiconnI have two alternative suggestions:
13:55:18amiconn(1) My old idea, only allow one single .bmp per wps that contains all imagery used in the .wps. Many skinning gui systems work this way
13:57:31B4gderthat's actually not a problem with loading bmps in the same dir as the WPS
13:57:32amiconnThe name mapping could even be implicit, i.e. example.wps would load example.bmp
13:57:54B4gderbut I agree we should address it anyway
13:58:12amiconn(2) The images associated with a wps could reside in a sub-dir of where the .wps file is located
13:58:33B4gderand why is that a problem?
13:58:42amiconn(again, implicit mapping is possible, example.wps would use .bmp files in example/
13:59:07B4gderI don't understand
13:59:39B4gderwhat's "implicit mapping" in this context
13:59:49LinusNdir name is <wps name>
13:59:50amiconnWell, e.g. Christi has a .wps that uses 50 .bmp files. If loading would be implicit from the directory where the.wps is located, all 50 .bmps would be in /.rockbox
14:00:07B4gderagain, you think of setting it as default
14:00:11B4gderI think of loading it
14:00:21B4gderthey are two separate problems
14:00:25amiconnYes, and the default is loaded from .rockbox
14:00:29B4gdernow, yes
14:00:41LinusNi think having a dir with the same name as the .wps is a good idea
14:00:44B4gderthat would need to be addressed for this to be really good
14:01:10B4gderLinusN: yes, but we'd need to allow the wps to be in a sub dir or other dir than .rockbox
14:01:34B4gderbut yes, having the dir named like the wps is a fine idea
14:02:11B4gder".rockbox/wps/coolness.wps" loads pictures from ".rockbox/wps/coolness/"
14:02:45LinusNsen lunch
14:02:50B4gderand as a first shot, we'd require wpses to be in .rockbox/wps to survive a reboot
14:10:25ashridahB4gder: that would in theory mean you could extend it to include support for an archive like format that contains the lot
14:10:57 Join ripnetuk [0] (
14:11:12 Quit ripnetuk (Client Quit)
14:11:13 Join ripnetuk [0] (
14:11:48ripnetukfirefly - hows the remote wps coming along :)
14:13:07 Quit ripnetuk (Client Quit)
14:13:35 Quit actionshrimp (Read error: 110 (Connection timed out))
14:14:50*ashridah suspects the daisy audio format won't be possible for that blind archos user :(
14:16:25 Join ripnetuk [0] (
14:16:36ep0chdaisy is like c
14:16:36ep0chdaisy is like a cuesheet?
14:17:16ashridahit seems to be more complicated than that
14:17:41ep0chripnetuk: i believe firefly put a patch in the tracker
14:18:11 Quit ashridah ("shower. sleep")
14:19:16ripnetukcool... i will go and find it
14:20:11 Join Sandking [0] (
14:20:53 Join muesli- [0] (n=muesli_t@
14:21:22ripnetukare we going down the route of a seperate wps for remote, or having extra lines in the current .wps file?
14:21:45B4gderI think separate is gonna be easier to deal with
14:21:53ripnetuki agree
14:21:56B4gderit also gives you better flexibility
14:22:04ripnetukyes, more combos
14:22:54ripnetukdamn, cant right click and copy shortcut on tracker page, then use wget (i know, i know, use curl!)
14:23:19B4gderyou can't?
14:23:36ripnetuki think its a sourceforge thing
14:24:25B4gderthey might block wget's user-agent or something
14:24:48ripnetukyeah, wget gives me a html file with a silly filename (xxx.php)
14:25:10 Quit _FireFly_ ("If you can't laugh at yourself, make fun of other people.")
14:25:24B4gderthat's because wget doesn't care about the Content-Disposition: header
14:26:02*B4gder checks
14:26:19ripnetukthats uncool (for wget)
14:26:39B4gderI usually do "curl -o outfile [URL]" to copy things from the sf trackers
14:26:53ripnetuki will try that next time then... cheers
14:27:08B4gderof course wget has such an option too...
14:27:32ripnetukthe follow redirects?
14:27:47B4gderthat's no redirect
14:28:09ripnetukanyway, i got it using firefox :)
14:28:28ripnetuknow to build uber-rockbox with WPS support :) :) :)
14:28:43B4gderI had a weird experience just a few hours ago
14:28:54B4gdermy main unit's stop didn't stop the music
14:29:48ripnetuki love the way on linux once you work out how to do something you can always script it for next time... on windows that is not the case
14:34:22*ripnetuk plays guess the -p option for patch :)
14:34:41B4gdermy favourite game
14:34:52ripnetukims till losing :( will have to fire up vi
14:34:54ripnetukto look at patch
14:35:22ripnetukaha it was a trick one -p 0
14:35:50LinusNi played bejeweled while listening to music the other day, and i realized that it would be nice to still have music control on the remote while running a plugin
14:35:54 Join actionshrimp [0] (
14:36:54ripnetukLinus - indeed, although it would have to be up to the plugin in question if it supported it... how about a plugin api function passThroughKeypress
14:37:08ripnetukthen the plugin could handle the keys, and pass through any its not interested in
14:37:35LinusNripnetuk: it would be enough to just call the default handler
14:37:58ripnetukof course
14:38:15LinusNshould be pretty straightforward
14:38:57 Join DJDD_ [0] (
14:40:10ripnetukCan the chipset in the iRiver pass through the analog in mixed into the analog out? i want to plug my smartphone into my iRiver so I dont have to change headphones when switching between watching video on phopne, and mp3s on rockbox (phone has silly sized jack plug, and adapter is small, and easily lost)
14:42:33 Join XavierGr [0] (
14:43:49Jungti1234Should like to play a game as quickly as possible in iriver h300. :)
14:44:11 Quit Mxm`Pas`Bien (Read error: 104 (Connection reset by peer))
14:44:16LinusNripnetuk: yes
14:44:27ripnetukwps looks good - thanks firefly (and of course timid) - i think today is the last day I run iriver firmware
14:44:58 Join Maxime [0] (
14:45:25LinusNripnetuk: i can't even remember that day anymore
14:45:41ripnetukyeah, but you dont rely on the remote then
14:45:56LinusNno i don't, but i use it a lot nowadays
14:46:26ripnetuki have a friend who bought a ih140 and has never even plugeed in the remote /me doesnt understand
14:46:42ripnetukits the only real feature of the iriver over the archos IMHO
14:47:43B4gderand wma playback
14:48:29XavierGrLinusN: Does rockbox fdprintf supports %f handler?
14:48:53B4gderwe have no float support
14:49:05XavierGrand how can i print to a file an integer divided by 100
14:49:18XavierGrp.x I want to print ventivolts
14:49:41XavierGrrockbox uses integer e.g 330 = 3.33
14:49:41B4gderinteger/100 and integer % 100
14:49:45B4gderfor example
14:50:05LinusNfprintf(fd, "%d.%02d", val/100, val%100);
14:50:37LinusNor something like that
14:50:57XavierGrok thanks
14:54:25 Quit DJDD (Read error: 110 (Connection timed out))
14:54:34XavierGrI just made a tsr plugin that logs in a txt file battery information like: on Exact second: Battery level percent, centivolts, estimated remaining time,
14:54:54XavierGrit will be good for battery benchmarks
14:54:55 Quit ep0ch (Read error: 110 (Connection timed out))
14:55:22XavierGrevery time the disk spins up for other reasons the plug_in will write the info to the file.
14:55:41XavierGrThis way the log will catch up until the last succesfull HD write.
14:56:09XavierGrthen using a graph we can see how the battery did on the test.
14:56:55B4gderanyone sits on a wps with bmps I could do some tests with?
14:57:13***Saving seen data "./dancer.seen"
15:04:55amiconnBtw, speaking about disk spinup and saving - I had an idea for an api for delayed file saving
15:05:28amiconnToday we save the config into the special sector, mainly because we want to delay the saving
15:05:53amiconnWe could save the configuration into a .cfg, but then we need delayed file saving
15:06:39amiconnThe ata driver could allow to register a callback for delayed file saving.
15:07:28Zagorsounds good
15:07:51LinusNiirc, the main reason for the config sector was not to delay the saving, but to have an atomic write on the ATA level
15:09:01LinusNand it was convenient, at the time when the filesystem code wasn't all that stable
15:09:03ZagorI like the config sector because it's the KISS solution. however I agree it's a bit inelegant for the user.
15:09:15Zagorand also not very expandable
15:09:41B4gderwe do suffer slightly because of it
15:09:54LinusNi'm all for .cfg file for defaults
15:10:17Zagorme too
15:10:48LinusNwe should perhaps still keep the resume info in a separate file
15:11:11B4gderwe should put the very-often-updated info separately
15:11:53LinusN.rockbox/default.cfg and .rockbox/resume.cfg (or something)
15:12:55Jungti1234hi Linus
15:13:28Jungti1234Take the trouble.
15:13:43solexxB4gder: what sort of wps do you need?
15:13:59B4gdernever mind, I got one off the gallery
15:14:17B4gderthanks anyway
15:17:53Zagorare we saving the settings on each spinup on iriver too?
15:19:25B4gderyes, afaik
15:20:10Zagorstrictly speaking we don't have to do that on units with soft shutdown
15:22:12Jungti1234I will go out.
15:22:26Jungti1234No one sees off. :(
15:22:34B4gder1 - assumes WPSes in .rockbox/wps
15:22:44 Quit Jungti1234 ("Bye~ -")
15:22:45B4gder2 - assumes BMPs in .rockbox/wps/[wps name]/
15:23:41B4gderseems to work
15:26:06LinusNB4gder: go go go
15:27:51B4gderyou mean a commit and run? ;-)
15:37:44B4gderI'll do it later tonight
15:40:18 Quit StrathAFK (Read error: 104 (Connection reset by peer))
15:46:00 Join Strath [0] (
15:47:10amiconnLinusN: Iiuc the .playlist_control file stores all information about the current playlist?
15:47:37amiconnIf so, it wouldn't it make sense to store resume info there as well?
15:48:57LinusNamiconn: fair enough, but it doesn't contain the time, and it can be quite large
15:49:25LinusNso i suggest the time is stored somewhere else
15:51:58amiconnHmm, I have never seen a large .playlist_control, but then I don't know exactly how it works
15:52:21LinusNif you insert a lot of files or dirs, it may grow quite big
15:52:26amiconnOf course it doesn't make sense to store the resume info in a large file, at least not when it's ascii
15:53:29amiconnI'm not really sure whether it's a good idea to store the settings as .cfg (ascii)
15:53:32 Quit B4gder ("time to say moo")
15:53:40LinusNi think it is
15:54:01amiconnThe advantage would be that it's stored in a file, is human readable, and we only have one format to deal with
15:54:46amiconnOtoh, storing frequently updated info (resume, runtime) there costs more processing time
15:55:27amiconnWe could even get rid of the bitfield array, but only if we find a solution for the rtc ram part of the settings
15:56:53amiconnThe Archos Ondio firmware stores settings in a binary file (ondiost.bin in the root)
15:59:46LinusNi suggest we do like this in the first step:
16:00:06LinusN1) save "regular" settings in .rockbox/default.cfg
16:00:27LinusN2) save runtime and resume info in a binary file like the ondio
16:00:37LinusNhow about it
16:01:24LinusNwe will still need the rtc for some things, though, for the charging info
16:18:25 Quit muesli- ("ich will Kühe!!!")
16:19:42amiconnLinusN: Totally different question - how do I use curl to upload sth using http PUT (no authentication)?
16:20:14LinusNcurl -d name=val http://blabla/bla.cgi
16:20:38LinusNamiconn: listen to Zagor and ignore me
16:22:38amiconnZagor: Hmm, that's what I tried, however, nothing happens
16:22:49amiconnNo error, but also no file on the server
16:24:26*Zagor reads the manual
16:28:13amiconnHmm, the problem is probably on the server side
16:28:19 Quit paugh ("bbiab")
16:32:34amiconnNm, the problem *was* on the server side
16:34:53leftrightwould anyone like the file which has the Predixis MusicMagic info written to tags which renders the Replay Gain unreadable ?
16:35:39leftrightjust thought you might be interested as this Predixis software appears to be catching on
16:36:50LinusNleftright: is it big?
16:37:07leftrightabout 6 megs
16:37:57leftrightunless you install this predixis software and test it on one of your files
16:38:05LinusNdcc it to me, and i'll put it on my server
16:38:50dwihnoDoes this predixis stuff add extra data to the stream?
16:39:20leftrightit adds hash files to the tags, but I'm no expert
16:40:11leftrightfoobar and tags redaers show the info, foobar still displays and plays the file correctly
16:40:52leftrightLinus; dcc ?, sorry dont understand
16:41:26LinusNdcc is the irc file transfer protocol
16:43:43*leftright looks for this ftp
16:44:00Slasherileftright: you can't use dcc from webchat
16:44:50leftrightthanks, saved me a square search
16:45:03LinusNleftright: i /msg'd you
16:47:41 Join Kohlrabi [0] (
16:57:15***Saving seen data "./dancer.seen"
17:00:40 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:14:40 Quit TiMiD (Read error: 104 (Connection reset by peer))
17:20:11 Quit Sandking ()
17:22:34 Join Febs [0] (
17:41:22 Quit Febs ("CGI:IRC (EOF)")
17:46:26 Join Febs [0] (
17:47:26amiconnLinusN: back to the settings files - what charging info do you mean?
17:48:24amiconnAlso, using .cfg files for ordinary settings means that some 'ordinary' settings which are today stored in the rtc ram will be stored on disk
17:48:30amiconn(like volume, bass etc)
17:49:39LinusNi mean that the charging algorithm needs the battery capacity etc, which is in rtc ram
17:49:58LinusNcar adapter mode is another
17:50:21amiconnHmm, but these are also ordinary settings...
17:50:45amiconnWe need to find a way for handling them both in rtc (binary) and on disk
17:54:39amiconnHmm, how does car adapter mode work on the player? There is no rtc...
17:54:54LinusNit doesn't have a charging screen
17:55:21LinusNso the car adapter mode setting isn't important in the boot
17:55:38amiconnThe player does have a charging screen
17:56:05amiconnOnly the very old models without ata power controls don't show it
17:56:16LinusNnot the flashed players
17:56:47LinusNcar adapter mode only works on players flashed with rockbox
17:56:59amiconnMy player is flashed, and it shows the charging screen when I plug the power adapter in off state
17:57:25amiconnIt was me who coded the player battery animation for the charging screen...
17:57:49LinusNthen i guess car adapter mode has never worked on the player
17:59:45amiconnHmm, probably. Car adapter mode should work (even for non-flashed rockbox), just not from off state
17:59:55LinusNoldplayers don't start when the adapter is plugged in, so there's little use for a charging screen :-)
18:00:23amiconnThey don't? Interesting...
18:00:24Maximewhat does "car adapter mode" is? :x
18:00:43LinusNMaxime: the archos players boot when the charger is attached
18:01:18LinusNso the car adapter mode resumes the playback when it senses external power
18:01:27LinusNi.e when you start the car
18:01:32amiconnLinusN: Do you know when this hardware feature was added? Was that with adding ata power control, or together with the new lcd?
18:01:37LinusNand it pauses when the power disappears
18:01:48LinusNamiconn: i don't know
18:05:45 Join arkascha [0] (
18:30:58 Quit wacky ("ahuh")
18:34:09 Part LinusN
18:37:53 Join karim [0] (
18:45:54 Quit karim (Remote closed the connection)
18:47:20 Quit webguest64 ("CGI:IRC (EOF)")
18:57:19***Saving seen data "./dancer.seen"
19:02:05 Quit arkascha (Read error: 110 (Connection timed out))
19:04:10 Join linuxstb [0] (
19:15:40 Part linuxstb
19:22:11 Part XavierGr
19:38:06 Join dpassen1 [0] (
20:06:15 Quit |joshn| ("KVIrc 3.2.0 'Realia'")
20:32:21 Join linuxstb [0] (
20:36:42 Join cwbowron_ [0] (
20:38:00 Part cwbowron_
20:47:01linuxstbCan someone with patch-tracker rights close patch #1352575 (the Shorten codec) ? I've just committed it.
20:52:57 Join RotAtoR [0] (
20:52:58BirdFishhas anyone here used dasm68 before?
20:54:46linuxstbamiconn: Thanks.
20:55:34 Join _FireFly_ [0] (
20:57:21***Saving seen data "./dancer.seen"
21:19:57 Quit _FireFly_ ("Leaving")
21:28:34 Part leftright
21:29:32 Join XavierGr [0] (
21:33:15 Quit Moos ("Glory to Rockbox")
21:44:01 Join xmixahlx [0] (n=xmixahlx@
21:47:33xmixahlxi notice shorten codec is added to CVS - what is the status of this? anyone know?
21:51:04linuxstbxmixahlx: It works very well. No seeking support yet (or maybe ever), but it's working.
21:51:32solexxmost SHN files I have seen had no seek tables anyway
21:52:22xmixahlxmmm all newer shorten files have them (>3.0 iirc)
21:53:00solexxit's been some time since I tried the last time. Now I am using FLAC...
21:53:56linuxstbShorten is very inefficient at decoding. Some live music traders like to keep them, but personally I convert everything to FLAC - and I think most people are the same when they get a Shorten file.
21:54:19linuxstbI meant to say inefficient at encoding - in terms of compression ratio.
21:54:30xmixahlxcuz decoding is blazing fast
21:55:38solexxbtw, is Ogg decoding significantly slower than MP3?
21:55:58linuxstbI don't know, I never use either.
21:56:07solexxI have some -q8 Ogg files and have more than 50% boost ratio
21:56:28solexxbut I don't have MP3 files of the same quality
21:56:57solexxlinuxstb: do you only use lossless encoding? even on your mobile player?
22:00:29 Quit Kohlrabi (Nick collision from services.)
22:00:31 Join linuxstb_ [0] (
22:00:33 Join Kohlriba [0] (
22:01:11linuxstb_solexx: Yes. My main reason is simply that I want one copy of my music that I can play anywhere. I have a H140 and a 60GB iPod - so 100GB of mobile storage.
22:02:07linuxstb_But I also have a lot of MP2 files captured from digital radio.
22:02:59xmixahlxcan anyone update
22:03:10xmixahlxit doesn't reflect >2.5 release
22:03:29linuxstb_Yes - anyone can.
22:08:26xmixahlxthanx for the help - later
22:08:43 Part xmixahlx ("blah blah blah")
22:11:49 Quit Febs ("CGI:IRC (EOF)")
22:25:16solexxhas anybody developed a good concept to manage a music collection in different formats in parallel?
22:25:42solexxi have a big pile of FLACS and another pile of MP3/Oggs
22:26:59solexxnevermind, I should be thinking silently
22:31:09amiconnlinuxstb: The shorten decoder addition moved actual code into a .h file. This is considered dirty coding style...
22:37:37 Join webguest68 [0] (
22:40:50linuxstb_amiconn: Are you talking about ff_log2_tab or the actual functions? I agree the table shouldn't be there - but the code is inline.
22:41:02amiconnI mean the table
22:41:14linuxstb_I agree - I'll move it.
22:41:45amiconnSome remarks concerning lcd-16bit.c
22:42:25amiconn(1) lcd_nopixel() should always do nothing, that's what it's for
22:42:56amiconn(2) lcd_invertpixel() should invert the pixel, that means complement brightness and colour on a colour lcd
22:43:32amiconnOf course I mean nopixel() and flippixel()
22:43:33webguest68are you's talkin bout the iriver h300
22:44:07amiconn(3) The block functions are most likely unnecessary for LCD_DEPTH >= 8
22:44:40amiconnThey *are* unnecessary for sure.
22:44:43linuxstb_webguest68: lcd-16-bit will be used by both the iPod and h300 ports - at least initially.
22:45:28linuxstb_amiconn: Thanks for the comments. I know there is still a lot of work to do in that file.
22:45:44amiconnPerhaps you'll get an idea about my lcd driver intentions for >=8 bit lcds when you check apps/plugins/lib/gray_draw.c
22:48:12 Quit webguest68 ("CGI:IRC (EOF)")
22:48:48amiconnOne difference of gray_draw.c and a core lcd driver is that the pixel function arguments are different, but that's just because the grayscale lib supports variable resolution
22:52:07linuxstb_Am I right in saying the framebuffer is one byte per pixel?
22:53:33amiconnYes it is (if the grayscale lib is operating in buffered mode)
22:53:39linuxstb_If so, it seems a lot of the drawing functions can be copied from there - but using a short* instead of a char*
22:53:50amiconnyes, should be possible
22:54:23amiconnIn fact some functions can be simplified, since they don't have to deal with varying level mapping
22:56:02amiconnThe grayscale lib supports 2..33 shades of grey; since it would be very cumbersome to let each plugin handle varying depth itself, the pixel values are normalised to be 0..255 regardless of internal depth
22:57:26***Saving seen data "./dancer.seen"
23:24:11 Quit Maxime (Read error: 110 (Connection timed out))
23:25:13 Join ]RowaN[ [0] (
23:26:06]RowaN[hey guys whats shorten like compared to flac?
23:27:27]RowaN[in lossless mode that is
23:38:10]RowaN[no replaygain/tag support, doh! but slightly higher compression, woo hoo!
23:38:11linuxstb_"No hardware support" now needs updating in that description of Shorten :)
23:38:43RotAtoR]RowaN[: that's slightly worse compression
23:39:20RotAtoR63.5% of original size as compared to 55.5% for monkey, for example
23:39:24]RowaN[i see
23:39:37]RowaN[i did wonder why the table cell was colored shitty brown
23:39:45Bagderhere's another =>
23:39:50]RowaN[its hardly a celebration color, compared to bright green
23:40:40]RowaN[does rockbox support monkeys audio?
23:41:42linuxstb_Two reasons - no GPL license and it's apparently very computationally hard to decode.
23:41:58linuxstb_s/GPL/GPL compatible/
23:42:45linuxstb_So it will probably not happen for both legal and technical reasons.
23:43:09 Join Maxime [0] (
23:43:34]RowaN[maybe i should switch all my flac stuff to wavpack.. better compression and all the features i use (replaygain, tags)
23:45:42linuxstb_I don't think there are any strong reasons for (or against) either. FLAC decodes faster, but wavpack is slightly smaller.
23:46:43linuxstb_Wavpack has a hybrid mode, but oggenc can encode directly from flac files.
23:56:05]RowaN[should faster decoding excite me? does it mean more battery life?
23:56:32Bagderyes it does
23:56:46linuxstb_In theory. But the smaller filesize will help wavpack. So it's a close contest.
23:57:03linuxstb_But FLAC will probably play for longer
23:57:06Bagderyes its a time vs time ;-)
23:57:23*amiconn wonders whether this monster will still work when restructuring is done :/
23:57:28Bagderdecoding time vs time between spinups
23:58:54linuxstb_FLAC plays for about 12-13 hours. I would expect wavpack to be quite close to that - so it depends if that's an issue for you.

Previous day | Next day