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

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

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

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2005-12-15

00:00:21 Join nathanh [0] (
00:00:47 Quit San (Read error: 110 (Connection timed out))
00:01:22 Join Jungti1234 [0] (n=jungti12@
00:01:33miraklinuxstb: hum so with your make there is no static lib involved
00:02:13mirakthat's what I wanted to do linuxstb thanks
00:05:06 Join elinenbe [0] (
00:05:26linuxstb_mirak: No, I don't bother with that.
00:07:24linuxstb_One problem is that xvidcore is both an encoder and decoder - I've removed the obvious encoding parts of the library, but I'm sure that a lot of the functions still there are only used for encoding.
00:08:04linuxstb_And there are still some floats there, and some calls to libm functions
00:08:14linuxstb_(log10, sqrt etc)
00:09:35mirakI guess postprocessing is useless
00:10:56linuxstb_The example program, xvid_decraw contains some floats which are only used to display decoding time statistics - so they can be easily removed.
00:10:57 Join textchimp [0] (
00:11:12textchimpi'm having errors compiling the CVS rockbox
00:11:31Bagderfor that target?
00:11:43linuxstb_Are you using Linux, Cygwin or the Development Kit?
00:11:48 Join LinusN [0] (
00:11:59Bagderand what patches if any do you have applied
00:12:04textchimplinuxstb_: linux
00:12:08linuxstb_And what is the error? :)
00:12:12textchimpsorry, hang on, let me get it
00:12:38textchimpgcc -g rdf2binary.c -o rdf2binary
00:12:48textchimprdf2binary.c:1: internal compiler error: Segmentation fault
00:12:54 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:12:57Bagderbad gcc!
00:13:13textchimpi haven't changed anything but the cvs update
00:13:25xbshiftanyone here working on or interested in the iriver ifp7xx series rockbox port?
00:13:51Bagdertextchimp: but that source file has not changed in a good while iirc
00:14:01Bagderxbshift: we're all interested in having more ports! ;-)
00:14:35xbshiftwell so far a guy named tomal had ported it
00:14:39Bagdertextchimp: besides, an internal error in gcc is hardly Rockbox's fault
00:14:44xbshiftjust with minimal function though
00:15:10 Quit Rob2222 (Read error: 104 (Connection reset by peer))
00:15:15xbshiftwhat is required to start working on such a project? i mean knowledge wise.. asm, c programming ect
00:15:35linuxstb_You want to help tomal with the port?
00:15:59xbshiftsure, if i could
00:16:01Bagderthe best approach for this port is probably to ask tomal himself
00:16:20xbshiftfrom what i understand right now he has code injected from within a .wma file
00:16:20Bagderbut knowing C and basic asm will sure help
00:16:34xbshiftrunning the wma loads the rockbox
00:16:35 Join DT [0] (
00:16:42 Quit DT (Client Quit)
00:16:49xbshiftim familiar with C but i queeze at the thought of asm ;/
00:17:20linuxstb_The first thing would be to download his source code and try it out.
00:17:25dwihnoPeople writing asm for the targets, how do they test the code?
00:17:36Bagderdwihno: on the target! ;-)
00:18:16dwihnoI would probably brick my unit if I started doing that
00:18:37dwihnoHow hard is 68k assembly compared to x86?
00:18:46Bagderdwihno: just crashing the software doesn't mean bricking in most cases
00:19:00Bagderdwihno: 68k asm is way nicer than x86 imho
00:19:09miraklinuxstb: I checked, the function that use float are not used
00:19:18mirakthe only one is writetimer
00:19:29mirakbut it can be removed as well I think
00:20:40dwihnoBagder: I rather just leave the insane-o optimizing to the pro's ;)
00:20:52miraklinuxstb: writetimer is for loging
00:21:45mirakcan gcc build code with float for coldfire or is it just that it's done software ?
00:22:21Bagderit'll be done in software
00:22:29nathanhno hardware fpu, need soft fpu
00:24:05zeerothe colours in uisw32.h, are the all in B, G, R order?
00:24:08linuxstb_But I don't think we have libm
00:24:40Bagdersimple float operations are ok, but we don't have libm
00:25:33 Quit xmixahlx ("blah blah blah")
00:29:00 Quit Jungti1234 ("bye")
00:34:26textchimpso can anyone suggest what to do about that gcc error?
00:35:21Bagderupdate your gcc
00:35:30Bagderor use make -k for now
00:35:41Bagderor edit the makefile in tools to ignore the rdf2binary tool build
00:35:51Bagderbut your gcc is defect
00:36:18textchimpfunny, it worked fine on the cvs code from a few months ago
00:36:37linuxstb_Have you tried a clean cvs checkout?
00:37:03Bagdertextchimp: the rdf2binary code was last changed in june
00:37:28textchimplinuxstb_: what's the best way to do that? just check out into a new dir?
00:39:25textchimphm, it seems to be dumping the text contents of the cvs
00:39:45miraklinuxstb: I have checked floats sqr log10 double, they are not used by the decoder
00:43:11 Join stevenm [0] (
00:43:13 Quit linuxstb_ ("CGI:IRC")
00:43:17 Join linuxstb_ [0] (
00:43:19stevenmSo, how about that korean firmware ?
00:43:29linuxstb_mirak: You need to be able to remove the LDFLAGS=-lm line from the Makefile and still compile a working decoder
00:45:57 Quit mikearthur (Remote closed the connection)
00:46:06 Join mikearthur [0] (
00:47:11nathanhoh nifty, latest version of gnome recognises the H340 as a DAP and uses an appropriate icon
00:48:49 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
00:51:56stevenmperhaps whoever wrote that string was frustrated at something that he could not get to work
00:53:25 Join linuxstb_ [0] (
00:54:28miraklinuxstb: it's ok it works
00:54:39mirakmost of float functions where for post processing
00:55:23linuxstb_Have you tried running the test program to decode a file?
00:57:11mirakas a proof no
00:57:36 Join matsl [0] (
00:59:11 Join linuxstb__ [0] (
00:59:26 Quit linuxstb_ ("CGI:IRC")
00:59:31 Join linuxstb_ [0] (
00:59:59 Quit linuxstb (Read error: 110 (Connection timed out))
01:05:07miraklinuxstb_: well the standart exemple doesn't work
01:05:11mirakI mean the command line
01:08:10linuxstb_It works fine for me. What kind of input file are you using? It needs to be a raw xvid video stream
01:08:22mirakI use the cactus exemple
01:08:37mirakI did not tried the decompressor yet
01:08:45mirakI tried to compress the sample cactus
01:09:03mirakbut the commands of the doc doesn't match the exemple switchs
01:09:07linuxstb_Why are you testing the encoder?
01:09:22mirakI think cactus.pgm is not compressed
01:09:26linuxstb_Just use any xvid file you have already got, and extract the video stream from it.
01:09:37linuxstb_I used "mplayer" to do that.
01:09:47mirakmencoder or mplayer ?
01:10:14linuxstb_Anyway, the malloc situation looks good - all the mallocs happen at the very start of the decoding process, and it doesn't appear to call free
01:10:32linuxstb_So the simple malloc implementation used for the audio codecs will work nicely.
01:10:38mirakyes I think it just swaps the buffers
01:10:50mirakfor the frames
01:11:22mirakhem what command did you used for mplayer ? ...
01:11:31linuxstb_It mallocs about 4MB of memory for my 624x352 test file.
01:11:44linuxstb_mplayer -dumpvideo -dumpfile test.xvid myxvidfile.avi
01:11:58mirakok I have videos the size of the lcd already
01:12:02mirakin 10 fps
01:15:03***Saving seen data "./dancer.seen"
01:16:51miraklinuxstb_: it works for me
01:17:18mirakthere is a bunch of code that can be removed to reduce the binary size
01:18:04linuxstb_You should do that then, and also strip down the example program to the bare minimum - ready to make it a plugin in Rockbox.
01:18:20 Quit stevenm (Read error: 110 (Connection timed out))
01:18:30miraka plugin ok
01:18:46linuxstb_Yes, I think the first stage is a video-only plugin - ignoring audio for now.
01:19:05mirakwe will se for codec later
01:19:32 Quit Kohlrabi ("Leaving")
01:19:33linuxstb_We need to get the video decoding fast enough for there to still be enough CPU left to decode audio.
01:20:51miraklinuxstb_: with the plugins, can we use a kind of file chooser ?
01:21:03mirakto select a file to be read for exemple
01:21:37mirakI will see that tomorow
01:22:25mirakI am going to bed good night
01:22:45linuxstb_No - you write it as a "viewer plugin". So you just select a file in the file browser, and that will start the plugin.
01:25:18mirakI should buy a LCD screen
01:25:32mirakmy eyes are on fire
01:25:35linuxstb_Are you sure you can disable the post-processing?
01:25:51 Quit ender` (Read error: 113 (No route to host))
01:25:52mirakI removed the denoise and add noise
01:26:31mirakthe post processing is really post processing
01:26:44mirakit's called after the frame is decoded
01:27:12mirakit's called in the sample code
01:27:21mirakif I am not mistaken
01:27:33linuxstb_I guess we'll find out when we try to use it.
01:31:01mirakhum the extracted frame in pgm seems to be in YUV
01:31:56 Quit Moos ("Happy Birthday Rockbox")
01:34:23mirak./xvid_decraw -i test.xvid -d -c rgb24 -f tga
01:34:28mirakthis is working
01:34:57 Join stevenm [0] (
01:36:06linuxstb_It also has an rgb16 output format - perfect for the h300 (assuming it's rgb565)
01:36:27mirakthat's perfect for 16 bit color output yes
01:36:33miraknot for 18 ;)
01:37:09linuxstb_Given the speed the lcd works in 16-bit mode, I'm not sure 18-bit will happen.
01:37:13mirakwe will see how good iriver devs are
01:37:24mirakyou mean the latency ?
01:37:35webguest46grayscale output!
01:37:39linuxstb_No, the speed of the lcd_update() function
01:37:40mirakwell anyway on a screen of this size it's enough
01:37:49miraklinuxstb_: what's the problem
01:37:50mirakah ok
01:37:55mirakthe conversion
01:38:06 Quit tvelocity ("Leaving")
01:38:29mirakI was arguing with amiconn that maybe we could sacrifice some colors by just using shifts to gain speed
01:39:18 Join tvelocity [0] (n=tony@
01:39:22 Join San [0] (
01:39:24linuxstb_I know. But that's not related to this problem. We will need to write an optimised yuv2rgb565 function
01:39:46mirakin the xvid codec ?
01:39:56miraka bunch of stuff can be optimised
01:40:12mirakthough I don't know exactly what kind of assembly C code produce
01:40:39mirakI am not sure what is worth to put in assembly
01:41:18mirakI asked that the other day. I was wondering if system function like memcopy are optmised in asm for each target gcc can compile to
01:41:27mirakor if it's just a function that use C code
01:42:03linuxstb_It probably depends on the target. amiconn has implemented his own optimised memcpy function for both SH1 and Coldfire that outperforms the gcc version
01:42:04mirakthat kind of function could be optimised if not
01:42:18linuxstb_So we already have that.
01:42:55 Quit mirak ("Ex-Chat")
01:43:11 Quit PaulJ (".")
01:56:54 Quit matsl ("ERC Version 5.0 (CVS) $Revision: 1.776 $ (IRC client for Emacs)")
02:06:46 Quit dpassen1 ()
02:11:35LinusNok, i just committed fm radio support for h300
02:13:02Benacoolthanks Linus ^^
02:13:08Midgey34LinusN: you're my hero
02:13:22LinusNthe scan is slow as hell, thanks to the slow lcd driver
02:13:29LinusNbut it works
02:14:24markunLinusN: someone was asking about USB charging. Should it be possible to make it work just by defining HAVE_USB_POWER and setting a button to not enter USB mode?
02:14:46LinusNnot sure
02:15:00 Quit webguest46 ("CGI:IRC (Ping timeout)")
02:15:03LinusNwe might have to set a register in pcf50606 too
02:17:00RotAtoRany h3xx owner care to test out a new version of bejeweled for me?
02:17:30 Quit Febs ("CGI:IRC")
02:17:53 Quit stevenm ("Connection reset by beer")
02:18:32LinusNRotAtoR: total hang :-)
02:19:39LinusNi press on, and nothing happens
02:19:48RotAtoRoops, that may be a h100 build...
02:20:15RotAtoRi think i may have accidently rebuilt the sim when i thought i was in my build directory
02:20:23LinusNRotAtoR: i do prefer source code when i test stuff
02:21:04RotAtoRyes, but i can usually find more willing testers if i just distribute a .rock
02:21:09linuxstb_RotAtoR: Have you made bitmaps for the iPods?
02:21:22linuxstb_The Nano as well?
02:21:23RotAtoRand figured out what i think are good key bindings
02:21:38linuxstb_Cool. Let me know if you want me to test
02:22:13RotAtoRi didn't know the ipod port was able to run plugins already!
02:22:33linuxstb_Yes - plugins are working.
02:23:30RotAtoRnice, last time i looked at the status of the port the button driver was complete yet
02:23:42LinusNRotAtoR: any new .rock for me?
02:23:53RotAtoRworking on it as we speak ;)
02:23:53linuxstb_preglow's still working on it - it's working, but not perfect yet.
02:26:09RotAtoRok, updated at:
02:27:08RotAtoRso it works! :)
02:27:46LinusNslow as a dog, of course, but that's not your fault
02:28:09LinusNnext thing is to update the lcd driver with alpha masks
02:28:11linuxstb_Can you upload the .c file so I can test it on the ipod?
02:28:20RotAtoRok, it nice on the sims :)
02:28:40RotAtoRor i can just upload the .c too
02:28:49linuxstb_diff against cvs is fine
02:28:54RotAtoRbut the diff includes the plugins/SOURCES change
02:29:02linuxstb_That will be fine.
02:29:06 Quit linuxstb_ ("CGI:IRC")
02:29:59RotAtoRthe .c is now 1500 lines of bitmap definitions and only 1000 line of real code :S
02:31:21 Join markun_ [0] (
02:33:28 Quit markun (Read error: 113 (No route to host))
02:33:37linuxstb__RotAtoR: It works very nicely on my ipod.
02:33:57RotAtoRhow about the key config? any issues?
02:34:22linuxstb__No, using the scrollwheel like that works nicely.
02:34:32RotAtoRgood, i was hoping that would work well
02:34:41RotAtoRthat's one thing that's hard to tell with the sim
02:35:52linuxstb__preglow will be pleased.
02:37:01linuxstb__Somehow the status on the right side needs to be improved - the jewels are too nice now :)
02:37:13RotAtoRhehe, yeah
02:37:22RotAtoRyay, more bitmaps!
02:37:33RotAtoRone annoying thing though, the nano screen height isn't a multiple of 8, so there is a row of 4 unused pixels at the top
02:38:51linuxstb__How do I exit a game?
02:39:19linuxstb__Must be the button driver - that doesn't work.
02:39:32linuxstb__Guess I'll have to play all night...
02:39:36RotAtoRselect+play to save...
02:40:04linuxstb__Nope, I can't save either. As I said, I'll have to keep playing.
02:40:44 Nick linuxstb__ is now known as linuxstb (
02:40:59RotAtoRbut select+another key should eventually work, right?
02:45:27 Quit mikearthur (Read error: 104 (Connection reset by peer))
02:46:13 Join aliask [0] (
02:46:36aliaskRotAtoR: The new bejewled bitmaps are sexy.
02:46:45aliaskAre they the Gjewled ones?
02:46:54RotAtoRyes, they are
02:47:06aliaskWell good on you. It's a pleasure to play.
02:47:45 Join qwisp11 [0] (
02:47:54RotAtoRGweled has nice svg graphics, so it makes it easy to produce bitmaps which ever size I want
02:48:14linuxstbWhat tool did you use to convert them?
02:48:16aliaskWow, that's handy.
02:48:51RotAtoRSodipodi (, it's aparently what was used to create them
02:49:01aliaskHrm, maybe we should think about SVG graphics? It would help with having all the different screen resolutions.
02:49:13RotAtoRit has an absolutely dreadful user interface though
02:49:37RotAtoRat least on windows...
02:50:02Midgey34RotAtoR: I assume you know the jewel backgrounds overlap when swapping
02:50:32RotAtoRyep, and nothing can be done about it presently
02:50:41aliaskLinusN: Was the radio fix as simple as finding the new GPIO pin?
02:50:43Midgey34that's what I figured
02:50:45qwisp11I thought it looked nice like that
02:51:25linuxstbstarfield is very cool in black and white...
02:51:49LinusNaliask: no, there were a few bugs in the i2c driver to squash as well
02:52:12aliaskWell congrats on getting it up and running.
02:52:40linuxstbI'm assuming that a plugin should restore the original background and foreground colours if it changes it?
02:54:43aliaskFrom what I've seen in other plugins, people do, so I guess that's a yes.
02:55:50linuxstbI've just tried without, and the file browser comes back in the default blue and black. So maybe it isn't needed.
02:58:54LinusNtime to sleep
02:58:56LinusNnite all
02:59:15 Part Midgey34
02:59:20 Part LinusN
02:59:48linuxstbstarfield is now white stars on a black background on the h300 and ipod. Hope no-one objects to that.
03:01:25aliaskSounds good.
03:01:55aliasklinuxstb: Sorry, I got mixed up. They set the drawmode back to DRMODE_SOLID, not change the colours.
03:02:24 Join actionshrimp [0] (
03:02:37linuxstbaliask: OK.
03:05:31 Join Kyl3 [0] (
03:12:09 Quit Kyl3 ("( :: NoNameScript 3.81 :: )")
03:12:23 Join dropandho [0] (
03:13:12dropandhohey all!
03:13:22dropandhostill have no forum here- weiiird!
03:15:07***Saving seen data "./dancer.seen"
03:16:12lostlogic*scowl* getting spurious "no codec for file" messages on files that it will then decide to play fine moments later.
03:19:08qwisp11whats the difference between .wps and .rwps ??
03:20:28RotAtoRwps is for the main screen, rwps is for the remote screen
03:27:27 Join Rob2222 [0] (
03:28:34 Join joshn_454 [0] (
03:36:06 Quit actionshrimp ("a bird in the bush is worth two in your house")
03:36:56qwisp11what are the pixel dimensions of a h320 LCD?
03:38:37RotAtoRnp :)
03:43:56 Join amiconn_ [0] (
03:56:32qwisp11can I get a wps file to do a calculation and show the answer?
03:57:00qwisp11so instead of -20 for volume I get it to show 84 (plus the volume)
03:57:03qwisp11so 64
03:57:32qwisp11oh well
04:02:26 Quit amiconn (Read error: 110 (Connection timed out))
04:02:27 Nick amiconn_ is now known as amiconn (
04:02:50 Quit qwisp11 ()
04:10:30 Quit HCl ("Lost terminal")
04:17:50Benacoolwhat's the speed limit in starfield ? lol
04:18:48BenacoolI'm at +2000 mow ...
04:37:03 Join lush [0] (
04:37:10lushhola guys
04:37:29lushanyone home?
04:38:33lushanyone know where i can find a free wma to mp3 converter?
04:39:44lostlogictranscoding is for the devil. but if you must, just use the wav output plugin from winamp to write .wav files from the wmas and then encode them with lame
04:40:38lushdoesnt that do it in real time?
04:40:49lushaka forever?
04:41:27lostlogic... the wav output plugin should output as fast as it can decode... although I haven't used it in years...
04:43:30lushwell do you know of a less complicated solution?
04:44:09lostlogicnope. transcoding is still for the devil.
04:45:12lushwell thanks anyways
04:45:12 Join z35 [0] (
04:45:47lostlogicI don't think it's free, but I believe foobar2000 has transcoding support
04:49:13 Part lush
04:51:00saratogayou should probably use foobar for that task since it will conserve file tags
04:51:18saratogai don't believe winamp will do that if you write to wav first
04:51:43saratogaopps too late
04:51:47saratogaoh well
04:52:17 Quit saratoga ("CGI:IRC (EOF)")
04:56:33 Join DJDD_ [0] (
05:02:58 Join YouCeyE [0] (n=YouCeyE@unaffiliated/youceye)
05:03:20 Join Rob2222_ [0] (
05:15:11***Saving seen data "./dancer.seen"
05:19:13 Quit Rob2222 (Read error: 110 (Connection timed out))
05:31:19 Join bagawk [0] (n=lee@unaffiliated/bagawk)
05:47:51 Join ircfoo [0] (
05:49:42 Join Jungti1234 [0] (n=jungti12@
06:22:24lostlogicwow, this-ere coldfire is one neat lil' chip!
06:25:15 Quit RotAtoR ()
06:26:39zeso once rockbox is on the ipod, somebody needs to figure out how to reverse engineer the rio karma flashing :p
06:42:04 Quit YouCeyE ("Leaving")
06:44:06Maximethe karma.. until it dies :p
06:44:10aliaskHey wow, I just successfully did my first alteration of a program to fix a bug in C.
06:45:22bagawkaliask: w00t! Submit to the tracker
06:46:02aliaskIt's for one of the eli_sherer plugins which isn't in CVS
06:49:07Jungti1234nice rockbox
06:52:32 Quit ^BeN^ (Read error: 104 (Connection reset by peer))
06:53:07saa[b_r]iderwinword, or winpaint?
06:53:12saa[b_r]idersorry even
06:53:20saa[b_r]iderRockword or Rockpaint?
06:54:20saa[b_r]iderjust wanted to know which plug-in aliask fixed
06:54:50saa[b_r]iderwhat's the fix?
06:55:18aliaskWhen it displays a BMP file, it would turn the backlight off after the set time. And if you pressed a key it would quit the plugin, so the image would go away.
06:55:41aliaskI just made it so that only the stop button actually quit the plugin when displaying a BMP from a file.
06:56:06saa[b_r]idergood job :)
06:56:20saa[b_r]iderbut it's not uploaded anywhere yet I presume
06:56:42aliaskYeah, eli doesn't have an email address on his page either, so I guess I'll have to catch him on IRC
06:57:14saa[b_r]ideryou can also PM him at MR, just to make sure he gets the message
06:57:31 Join ashridah [0] (
06:57:31aliaskOh he's at MR is he? That makes things a bit easier.
06:58:02saa[b_r]ideryeah, he started the H300 color plug-ins thread for his work...
06:58:21 Nick ashridah is now known as Lost-notash (
06:58:31saa[b_r]idereither PM him, or mention the update in that thread, I'm sure he'll read it
07:00:05 Quit Lost-notash (Client Quit)
07:04:08 Join nathanh_ [0] (
07:07:09 Quit DrumRBoy|Away (Read error: 104 (Connection reset by peer))
07:10:19 Join Paul_The_Nerd [0] (
07:12:49 Part Sando
07:15:15***Saving seen data "./dancer.seen"
07:21:34 Quit pabs (Read error: 110 (Connection timed out))
07:24:09 Join webguest83 [0] (
07:25:30webguest83Hi does anyone know if Jorg's site is down? I am looking for firmware_rec.bin
07:27:34 Join lennyk [0] (
07:28:03 Join Lost-notash [0] (
07:29:06 Quit lennyk (Client Quit)
07:38:54Paul_The_NerdHowdy y'all.
07:46:32 Quit Rob2222_ ()
07:54:58 Quit Benacool ()
08:00:56Jungti1234morning amiconn
08:01:46 Quit ReKleSS (Read error: 110 (Connection timed out))
08:12:23 Join lamed [0] (
08:13:10 Quit lamed (Client Quit)
08:13:53 Join lamed [0] (
08:14:40 Quit lamed (Client Quit)
08:16:15 Join lamed [0] (
08:16:27*amiconn wonders why the iriver doesn't allow recording from the radio screen....
08:16:47 Quit lamed (Client Quit)
08:16:50amiconnThe archos allows this for long.
08:17:13Bgeramiconn they anounced that h100 cannot record from radio ("hardware limitations")
08:17:44amiconnYou _can_ record from the radio with rockbox, just it's cumbersome to do that as-is
08:17:56 Quit ircfoo ("CGI:IRC (EOF)")
08:18:27Bgeramiconn i know, i'm just answering to your question
08:18:40amiconnIt's just that the radio screen needs some further adjustments (and probably some code snippets as well) to allow this directly from the radio screen
08:18:51amiconnNo, I meant in rockbox
08:19:17Bgeraaa, aha
08:19:32amiconnI never use the original firmware
08:19:46Bgeri understood "iriver" as original fw
08:20:01amiconnI meant it as rockbox target
08:21:38Bgerthe new jewels are great:)
08:22:39webguest83I have a V1 Recorder and I want to flash. Jorg's site is down so I got Bluechip's dev kit and made rockbox.bin. Is this what I need to flash the bootloader?
08:23:03webguest83This is the first time to flash.
08:24:03amiconnwebguest83: No. You would need to build a full flash image (a multiple-step procedure) for a first time flash, but this way is *not* recommended
08:25:16webguest83Any idea where to get
08:26:57amiconnNow that we have bootbox (meaning the images are 100% rockbox), we can distribute them on the site. They are found on the BootBox wiki page
08:28:15 Join ender` [0] (i=ychat@
08:29:50webguest83Is Bootbox the recommended method now instead of the old two compressed image method?
08:37:43amiconnBootbox still uses two (optionallly compressed) images, only that the first one is no longer original archos
08:40:53 Join webguest43 [0] (
08:42:04 Part webguest43
08:46:56webguest83Thanks for your help. It worked. Now if I want to update I just browse to a more current .ucl and flash it right?
08:48:26 Quit nathanh_ ("CGI:IRC")
08:48:35amiconnYes, however, it is recommended (and if you prefer rombox.ucl, even required) to RoLo the new ajbrec.ajz first
08:53:14Jungti1234Is limitation in radio channel save?
08:57:51BgerJungti1234 say it in other way
08:58:35Jungti1234Is save limitation of radio channel?
08:59:16Jungti1234ah nothing
08:59:35Bgeri don't know
09:00:01amiconnYou mean how many presets you can store?
09:00:16*Paul_The_Nerd is still looking forward to being able to use multiple FMR files.
09:00:34Bgerthere was a patch from XavierGr iirc
09:01:37Paul_The_NerdI heard something like it's being worked on for being committed.
09:02:09amiconnJungti1234: 64 presets
09:02:11Paul_The_NerdI only really travel frequently between two cities, so it's really a minor thing anyway, but it is convenient.
09:02:52Bgerhehe the iriver limit was 20 iirc
09:03:54Jungti1234What is iirc?
09:04:13Bgerif i remember correctly
09:04:33Bgerwhat is -_-; ? :)
09:04:38Jungti1234It makes translation painfully.
09:04:54Jungti1234-_-; = oh no...
09:05:10Jungti1234hahaha maybe.
09:06:52 Quit DJDD_ (Read error: 110 (Connection timed out))
09:07:15 Quit webguest83 ("CGI:IRC")
09:09:42Bagderfeel free to add more terms
09:10:14Paul_The_NerdI suppose that'll become necessary, with popularity increasing.
09:10:29 Join bazz [0] (
09:10:33Bagderand its easier to refer to a URL that having to explain here
09:10:59Jungti1234thanks Bagder
09:11:20Jungti1234np = no problom
09:12:14Bgeradded :)
09:12:20bazzwhat does the data in lcd_framebuffer look like if you have LCD_DEPTH=16? is it just a 16 but rgb colour?
09:12:35 Join mirak [0] (
09:12:42bazzerr bit
09:13:15Bagderbazz: yes, but there are two versions of 16bit
09:13:35Bagderthe 16 bits are rgb565
09:13:57Bagderand there's a byteswapped version for ipod
09:14:14bazzBagder: nod, i noticed that in lcd-win32
09:15:18***Saving seen data "./dancer.seen"
09:16:19bazzthen in 1 bit colour you're either UI_LCD_BACKGROUND or black (0 and 1 respectively) after the bit manipulation right?
09:16:38bazzand in two you have a gradent between black, the background, and the background?
09:17:32 Join Rob2222 [0] (
09:17:45bazzwhat player has 2 bit?
09:17:53Bgeririver h100
09:18:22Bagderand older ipods
09:18:30Bagder(not that we support those... yet)
09:18:44bazzcool, thanks
09:19:35bazzBagder: someone mentioned you were thinking of doing the sdl port of the sim i think?
09:20:02BagderI am if no one else does it, but I don't think it'll happen during this year
09:20:22Paul_The_NerdWell, "this year" doesn't have too many days left anyway.
09:20:27Bgeronly 16 days left :P
09:20:33Bagder16 loooong days
09:20:37bazzBagder: i'm almost done with it :)
09:20:42Jungti1234Doesn't support unicode of preset?
09:20:49Bagderthat just rocks bazz!
09:20:57Bgerbazz really ?
09:21:08Bagderbazz: we'll love you for that
09:21:17Bgerso we can expect x11 sdl sim ?
09:21:28bazzBger: that's what i'm writing it on
09:21:46bazzactually, i built it on windows tonight and it works but not as well as on linux just yet
09:21:56Bagderwe should make sure the win version works too, and then from that day just have one single sim
09:22:19bazzyeah, i need to figure out what's broken with the windows one. if i move the window it locks up :)
09:22:26bazzotherwise, it works
09:22:38Bagdervery nice indeed
09:22:59bazzneed to add 16 bit support and port the thread stuff over, i'm still using pthread at the moment, but that shouldn't take long
09:29:07 Join YouCeyE [0] (
09:36:40linuxstbbazz: Do you support sound as well?
09:36:42 Quit Lynx_ (" reboot")
09:37:09 Join MrShlee [0] (
09:37:31Bagderoh right, sdl has sound too...
09:38:38bazzlinuxstb: yeah
09:38:57MrShleeError accessing playlist control file.. that means the playlist is just empty right? Seems to be looping :|
09:39:30Bgerhaha in the internet slang FM means fsck me....
09:40:24 Join Lynx_ [0] (
09:41:54 Join LinusN [0] (
09:42:47Paul_The_NerdI think we'll stick with the slightly older definition.
09:43:26Bgeryep :)
09:45:15linuxstbAm I right in thinking the SH1 doesn't have an instruction cache?
09:45:34 Join _FireFly_ [0] (
09:45:45 Join B4gder [0] (
09:47:24Bgerheh is this moin = morning ? :)
09:47:38Bgerhaha ok :P
09:49:16_FireFly_i have a little quistion :) why does nobody have an interrest about my remote-settings-patch ?? nobody has at least replayed to my rockbox-dev-ml thread
09:49:33Bgerhm, _FireFly_ i'll give it a try
09:50:00_FireFly_maybe the patch is broken against the latest cvs
09:50:11_FireFly_which is on tracker
09:50:45_FireFly_i don*t want that it gets commited now but at least i whish there where some comments about it :)
09:51:16Paul_The_NerdWhat does it do?
09:51:22Bger_FireFly_ patch No ?
09:51:33Bgeri guess it adds scroll speed etc for the remote
09:51:57linuxstbpreglow: Are you around?
09:52:28Bger_FireFly_ = Stephan ?
09:53:22Bger_FireFly_, please, add yourself in
09:54:05_FireFly_Bger yepp
09:56:37Bgerlooking at it now
09:56:43_FireFly_Bger the link is in the ml
09:57:16Bgeri think i've found it, patch No 1362248
09:58:51Bgerhm, what's the reason for changing (HZ/10) with (1000/HZ) in lcd_scroll_delay ?
09:59:54 Quit Vladoman (Read error: 110 (Connection timed out))
10:00:01B4gderthat sounds plain wrong to me
10:00:12B4gdereven though the result is the same if HZ is 100
10:00:14_FireFly_preglow or amiconn has mentioned it that HZ/10 is wrong
10:00:53 Join Vladoman [0] (
10:01:04linuxstbIf you want to calculate the number of ticks in one tenth of a second, then HZ/10 is correct. Unless I need more coffee...
10:01:16linuxstbI'm not sure what 1000/HZ represents
10:01:46Bgeramiconn ?
10:04:02 Join Gibbed [0] (
10:04:02LinusN17:14, sorry
10:04:04 Quit Rick (Nick collision from services.)
10:04:14 Nick Gibbed is now known as Rick (
10:04:21linuxstbI remember now, 1000/HZ converts from ticks to ms.
10:04:28linuxstbThanks LinusN.
10:04:48linuxstbLinusN: Am I right in saying there is no instruction cache on the SH1?
10:04:48 Quit Lost-notash ("CGI:IRC (EOF)")
10:04:53 Quit San (Read error: 110 (Connection timed out))
10:05:49LinusNno caches whatsoever
10:06:36Bgerhm, it's logical... 1hz = 1/sec
10:06:52_FireFly_how can i escape the '_' so that twiki doesn't write my NICK italic ??
10:07:26linuxstbLinusN: Thanks. That explains the lack of an invalidate_icache() function.
10:07:31Bgeruse <verbatim>_FireFly_</verbatim> maybe ...
10:08:06LinusNor perhaps <nop>_FireFly_ ?
10:08:20Bgeri thought about this too...
10:08:48Bgerbut wasn't <nop> only for not interpreting the next word as wikiword ?
10:09:18_FireFly_verbatim works
10:09:48 Quit amar ("CGI:IRC")
10:12:14_FireFly_<nop> also
10:18:49 Join hshah [0] (
10:19:04 Join mikearthur [0] (
10:20:22Bgerhello, mirak
10:24:11preglowlinuxstb: around now
10:26:36linuxstbJust the man. I think we need to implement the invalidate_icache() function. Do you know how to?
10:27:11linuxstbI'm assuming the cache is enabled at the moment (I'm running your patch)
10:27:22preglowit is
10:27:28preglowwhy do you need to invalidate it?
10:27:44preglowa quick read run over a big range of memory should do
10:27:45preglowtried that?
10:27:48linuxstbI'm getting crashes when running plugins for the second time.
10:27:50preglowjust read 64k om memory or something
10:28:06B4gderinstruction cache
10:28:35linuxstbThe ipodlinux wiki mentions address 0xf000c000 as the "cache invalidate base". But a quick grep doesn't show that address being used by IPL.
10:30:42 Quit mikearthur (Read error: 104 (Connection reset by peer))
10:31:02preglowlinuxstb: ipl don't use it
10:31:49preglowlinuxstb: i have no idea how cache works, and i doubt they do either
10:35:50saa[b_r]ideris it possible to reverse the compiling process?
10:36:22 Quit bazz ("Leaving")
10:37:40preglowB4gder: ipod's got a unified cache
10:37:46preglowB4gder: i assume it just buffers all reads
10:38:45linuxstbThen this probably isn't the reason for the plugin crashes. An lcd_update() will remove plugin data/code from the cache.
10:39:22linuxstbOr do cache any particular areas of memory?
10:40:06preglowi have no ide
10:40:18preglowi ripped both the cpu clock and cache parts right out of ipl
10:40:27preglowand i don't think they're too clear on how it works themselves
10:41:31Bgersaa[b_r]ider what do you mean ?
10:41:50linuxstbpreglow: OK, I'll forget that issue then.
10:42:27linuxstbDid you read in the logs that Bejeweled now has those Gnome SVG graphics, and that it works on both ipods?
10:43:26saa[b_r]iderBger: well there's a modified .rock plug-in, and I'd like to see the source...
10:43:56linuxstbA screenshot is in the bejeweled thread on
10:44:27linuxstbRotAtoR also changed the key mappings so the click wheel moves the cursor horizontally with wrapping. It works very nicely.
10:44:55B4gdersaa[b_r]ider: you can get a dissassembly from it, but that's not what you want ;-)
10:45:29linuxstbsaa[b_r]ider: Just ask the author for the source :)
10:45:32amiconnBagder, Bger: The scroll_delay argument is in ms, but the internal delays are measured in ticks. So the correct conversion is delay_ticks = delay_ms / (1000/HZ)
10:46:01saa[b_r]iderlinuxstb: I did, but no replies :(
10:46:35Paul_The_NerdWow, color bejeweled almost makes me wish I had something other than an H120.
10:46:39saa[b_r]iderpreglow: I think this is what bejeweled looks like now:
10:46:56preglowdamn, that's slick
10:47:27saa[b_r]iderH300 plug-ins are getting lots of colors :)
10:47:31preglowas they should
10:48:59saa[b_r]iderI'm trying to learn how sanitarium added color to pong and snake.... I don't think he uploaded the moddified plug-in to RB
10:49:21 Join mikearthur [0] (
10:49:48 Join aegray__ [0] (
10:50:49saa[b_r]idercomparing the different versions of bejeweled.c in CVS didn't help me, cause bejeweled has lots of colors, and it's not easy for a person who doesn't understand code to understand :|
10:51:22nathanhlook at the starfield patch on the rockbox patches page
10:51:23LinusNi think most of the game plugins would benefit from using rgb bitmaps instead of just changing the bg/fg colors
10:51:28nathanhvery simple addition of color
10:52:53Bgerwhere's the new bejeweled ?
10:52:56saa[b_r]idernathanh: good idea! thanks!
10:53:36saa[b_r]iderlinusN: making my way learning the code one step at a time :) maybe one day I can help
10:53:43*saa[b_r]ider crosses his fingers
10:54:11saa[b_r]iderBger: at MR it said it was gonna happen soon, don't know what they posted at the RB forum
10:54:30Bgeraha, k
10:55:24 Quit YouCeyE ("Leaving")
10:55:27MrShleesaa[b_r]ider Read a book :|
10:55:34MrShleeJumping into C isn't the way to learn
10:57:09B4gderI learned it that way
10:57:36nathanhi learnt by jumping in
10:57:38MrShleeok, its a possible option but its good to learn the basics
10:57:43preglowit has come gradually no matter how you do it anyway
10:57:47preglowi just started trying to code in it
10:57:51preglowworked out well in the end
10:58:12linuxstbnathanh: Did you see that I committed a patch to starfield that simply gives white stars on a black background?
10:58:15B4gderof course it also depends on what other languages you know already
10:58:21nathanhlinuxstb: yes, saw that
10:58:30linuxstbnathanh: What does your patch do?
10:58:31nathanhlinuxstb: mine gives a color to each star, i think it looks kinda funky
10:59:07preglowoption, please
10:59:13linuxstbnathanh: funky yes, but not realistic :)
10:59:20nathanhstars are all different colours
10:59:25preglowdo you think bejeweled is realistic? :P
10:59:28nathanhthey're rarely white
10:59:44linuxstbstarfield now looks like Elite - that's my reference for how space looks.
11:00:04nathanhmy starfield looks like the one on the old ami500
11:00:13nathanhwhich is my reference for when computer graphics stopped sucking :-)
11:00:50preglowcomputer graphics stopped sucking the moment people started drawing ansis
11:01:29*LinusN sees four yorkshiremen on the horizon
11:01:36BgerLinusN could you add #define CONFIG_LED LED_VIRTUAL to firmware/export/config-h300.h ?
11:01:47nathanhback when i was a lad, we didnt have no computer graphics, we had to make do with postit notes stuck to the fridge
11:02:04linuxstbYou were lucky to have a fridge....
11:04:17*preglow can see what the logical progression of this might be...
11:04:21miraknathanh: atari !!!
11:05:10Paul_The_Nerdpreglow: Does it involve 307200 people standing in bleachers, each holding a red, green, and blue sheet of paper and folding them to represent varying brightnesses?
11:05:20Paul_The_Nerd'cuz it would rock if it did.
11:05:42mirakthe doubles aren't floats ?
11:05:51LinusNPaul_The_Nerd: no 10fps video with that one...
11:06:21preglowmirak: double are more precise than floats
11:06:23linuxstbmirak: Yes, doubles are floats.
11:06:45linuxstb(in terms of them both being bad in Rockbox)
11:06:56Paul_The_NerdLinusN: Maybe if they were replaced by really well trained monkeys...
11:07:30LinusNBger: done
11:08:02BgerLinusN eh, at last:)
11:08:23Bgeri was annoying nearly everyone here with commit access about this
11:08:50nathanhwhats the led virtual for... hdd access?
11:08:54Bgerok, very bad english
11:09:11Bgernathanh it's like the hdd led of the PCs... but virtual
11:09:27Bgerbecause h3x0 and ondio (iirc) doesn't have real led
11:12:07 Join XavierGr [0] (
11:12:15Bgermorning, XavierGr ;)
11:12:35nathanhlinuxstb: your patch doesnt do white stars on black background for grayscale devices
11:13:08linuxstbnathanh: No.
11:13:13XavierGrmorning all
11:13:38linuxstbnathanh: greyscale devices don't have white
11:14:21BgerXavierGr where's your battery tsr plugin ?
11:14:52Paul_The_NerdIs that greenish color you see behind so many grayscale lcds there because of technical requirement, or did someone at some point think it was an attractive color?
11:14:56XavierGri the tracker :)
11:15:19***Saving seen data "./dancer.seen"
11:15:34BgerPaul_The_Nerd i've seen whitish grayscale lcds...
11:15:58XavierGrBger it adds 2-3 functions to the plygin api so you will have to compile it.
11:16:16BgerXavierGr i'll do it
11:16:29B4gderPaul_The_Nerd: on the Archos at least, the greenish is due to the leds that do the backlight
11:17:09Bgeri suspect that the green leds are between the cheapest ones ...
11:17:23preglowi suspect it's a contrast issue
11:17:34preglowgreen leds are the cheapest
11:17:36preglowgreen and red
11:17:39XavierGrBger: Please tell me your opinion and/or comments. (If it works for the H300) :)
11:17:51BgerXavierGr hehe oki
11:17:54B4gderI don't think it is a contrast issue, I find white leds better on the Archos
11:18:30XavierGrSomw guy in MR had changed his screen's led to red....
11:18:34Paul_The_NerdI was just looking at an old HP41CX calculator and thinking "Y'know, black and white displays seem to have kept pretty much the same background color for who knows how long, and to me at least it's just not really that attractive"
11:18:35BgerXavierGr btw, u worked on something about the fm radio, or i'm wrong?
11:18:43*Paul_The_Nerd is talking about the led-off color.
11:18:57XavierGrthis is still on the tracker and waits for some dev to llok at it.
11:19:04XavierGrmulti fmr preset list support
11:20:09XavierGrI thin it is the right time for someone to look on this now that H300 has working FM radio.
11:20:22 Join mirak_ [0] (
11:20:23Paul_The_NerdXavierGr: I compiled the FM presets in the wiki into an FMR file for each city, and posted a .zip there. Figure it'd go well with the patch.
11:20:46nathanhdo the devs take patches off sf? or do you need to get on the mailing list and annoy the devs to commit them
11:21:16XavierGrnathan if you annoy them in the ml naybe it will get commited faster, I am not sure yet :)
11:21:18mirak_if I want the pointer that reference X bytes of contigous data I do &char[X]
11:21:24mirak_I don't remember
11:21:33nathanhmirak: sort of, thats equivalent to char+X
11:22:02nathanhmirak: its a pointer to the X-th element in the array
11:22:05BgerXavierGr what do you mean with this "marginal error" in the plugin's description
11:22:40XavierGrPaul_The_Nerd: It will work perfectly!
11:23:05XavierGrBger. There is a possibility of a slight error. But it will be less than 15 minutes.
11:23:14XavierGr(normal testing)
11:23:39Bgerslight error like ?
11:23:52XavierGrBger: The plugin writed the log whenver the HD is accesed by Rockbox.
11:24:10XavierGrIf on the final write there is no succesfull write the log will not be updated.
11:24:28Bgerah, you mean the situation when the plugin can't append to the log on disk
11:24:31Bgeroh ok
11:32:48Bgerhm, is there any problem in starting a tsr plugin in apps/main.c (using AUTOROCK) ?
11:33:07*dwihno thinks the H300 FM radio support deserves a news item on the web front page
11:33:13XavierGrdid it crashed?
11:33:13dwihno(Yay for progress, btw!) \o/
11:33:24BgerXavierGr noo
11:33:28Bgeri mean other thing
11:33:32XavierGrah ok :)
11:33:34 Quit mirak (Read error: 110 (Connection timed out))
11:33:39XavierGrthe did it worked? :D
11:33:40Bgerlook in apps/main.c
11:33:46Bgeri haven't tested it yet
11:33:49amiconnLinusN: Is there a reason why the fmradio_i2c now does two DELAY calls in sequence?
11:34:00Bgerthere is a define AUTOROCK
11:34:21stevebim just wondering - i saw on the website about the rockbox bootloader allowing you to boot in to USB mode even if the hard drive doesnt work for some reason
11:34:26stevebis this the case with the h340 as well?
11:34:43Bgerif you set it and put autostart.rock in the plugins dir it'll run it
11:34:50Bgersteveb no, it doesn't work yet
11:35:03Bgerit's planned for the next bootloader
11:35:09Bgeramiconn ?
11:35:09nathanhwoo next bootloader
11:35:13XavierGrBger: it will run the plugin that you renamed autorock.rock?
11:35:15wubblawhich arguments for lame.exe do you use?
11:35:18 Join perplexity [0] (n=joust@
11:35:21Bgerto autostart.rock
11:35:22nathanhcant believe radio works already
11:35:27nathanhits like christmas came 2 weeks early
11:35:31Bgerhehehe :)
11:35:39stevebhah. at the rate things seem to be going for the H3xx the next boot loader will be out by christmas
11:35:39XavierGrBger: every plugin you want?
11:35:50BgerXavierGr it looks like so
11:36:04Bgerbut i don't know what's the situation regarding tsr plugins
11:36:12XavierGryou could try.
11:36:16Bgeramiconn could you answer my question ?
11:36:21XavierGrI don't have a compiled version right now.
11:36:30Bgeri'll try soon
11:36:55 Part LinusN
11:37:33Bgerbut if it's ok this will be the best way to test the battery...
11:37:38amiconnBger: What question?
11:37:57Bgeramiconn is there any problem in using the AUTOROCK feature with tsr plugin
11:38:27 Quit mikearthur (Read error: 110 (Connection timed out))
11:39:46Jungti1234What is Virtual LED?
11:40:09XavierGrA LED where is not presetn physically
11:40:26XavierGrIn our situation a LED marker rendered on screen.
11:40:32BgerJungti1234 it's on the statusbar
11:40:33Paul_The_NerdIt flashes an icon in the status bar to represent the an LED flashing.
11:40:35amiconnBger: No there shouldn't. The autorock feature was actually designed for using tsr plugins
11:40:43Bgeramiconn ah, ok, 10x :)
11:40:51Bgerso, XavierGr, i'll try it
11:41:08XavierGramiconn: Do you any time to compile the battery benchmark plugin to see if it works for archos targets?
11:41:25XavierGrI haven't seen anyother here that owns an archos.
11:41:37amiconn[IDC]Dragon implemented it for using alpine_cdc.rock
11:41:39XavierGrAlso it would be interesting to see if it will work on an Ondio.
11:41:57Jungti1234What updated?
11:43:36 Join San [0] (n=sanitari@
11:45:21amiconnXavierGr: Had no time yet :(
11:45:46XavierGrOk just mentioning. If you find the time please do and tell me your comments.
11:51:58 Quit edx ()
11:53:56linuxstbDoes the Virtual LED indicate disk access or the fact that the disk is spinning?
11:54:33_FireFly_linuxstb afaik disl-access
11:54:38Bgerdisk access
11:54:41Bgernot disk spinning
11:54:52Bgerjust like the hdd led on the PC
11:55:03_FireFly_or the led on the h1xx :)
11:55:08_FireFly_the red one
11:55:31 Join edx [0] (
11:55:34 Quit `3nergy (Connection reset by peer)
11:57:08 Join `3nergy [0] (
11:57:42 Join muesli_- [0] (
11:58:19mirak_if I declare a static buffer in a programm, if I go outside the boundary of that memory space, en outside of the program memory what happens ?
11:58:33mirak_does the OS bitch ?
11:58:40nathanhmirak: the devil takes your soul
11:58:56mirak_I guess it's protected in linux
11:59:48mirak_ok anyway don't know why I ask that, there is no reason it happens since static declaration is the same size of the malloc
12:00:33 Join BalanceST [0] (
12:00:36merbananmirak_: most likely you will get a segfault
12:01:01 Join Mongey| [0] (
12:01:45B4gderit makes a buffer overflow
12:01:56B4gderand god do I know about those ;-)
12:02:07Bgerhahaha, B4gder
12:02:36BgerB4gder only if u try to access the program's memory
12:02:38nathanhif a static pixmap shares the same horizontal row as a dynamic text, the static pixmap disappears
12:02:46Bgerthe own's program memory
12:03:27B4gderwell a static buffer is within the programs memory
12:04:06_FireFly_nathanh where ??
12:04:06 Quit BalanceST (Client Quit)
12:04:20nathanhfirefly: ok, im working on the iamp.wps
12:04:39nathanhthe top row has a %pv (the volume as a string) as well as a bunch of dynamic pixmaps (like the %bp and play status)
12:04:42nathanhthose are fine
12:04:48 Join Moos [0] (
12:05:02nathanhi moved the ogg/mp3/mp2 icons into the top row too... %fc
12:05:04nathanhand that disappears
12:05:14nathanhif i change %fc to a dynamic type, it stays on screen
12:05:26nathanhif i remove the %pv and leave %fc static, it stays on screen too
12:05:52_FireFly_can i see the code??
12:05:55nathanhi can fix it
12:06:36nathanhi want to grok the gwps code so i can add color and compositing
12:07:08_FireFly_nathanh i mean the wps-file
12:07:37nathanhsure, here are the lines that matter
12:07:41nathanh %pv
12:07:58nathanhand a bunch more like that, i wont list them all
12:08:19_FireFly_does this work ??
12:09:14_FireFly_ehm btw what do you mean with to leave the %fc to as static
12:09:46nathanhi mean i edit gwps-common.c and dont chagne WPS_REFRESH_STATIC
12:10:09nathanhif i change %fc to WPS_REFRESH_DYNAMIC it redraws once per frame and all is well
12:10:27nathanhbut obviously thats not the real fix, the real fix is to figure out why %pv is clearing a rectangle bigger than it needs to be
12:15:53nathanhfound it, /* clear the line first */
12:15:53nathanh display->set_drawmode(DRMODE_SOLID|DRMODE_INVERSEVID);
12:15:53nathanh display->fillrect(0, ypos, display->width, string_height);
12:16:16nathanhsilly thing clears the whole line
12:18:02_FireFly_why not
12:18:18nathanhbecause it means you cant put a WPS_REFRESH_STATIC on the same horizontal row
12:18:23nathanhit will get erased
12:18:29_FireFly_because the fn can*t know how long was the old string
12:18:40nathanhaye, im looking at it now, wondering how to fix that
12:19:28nathanhi can make everything WPS_REFRESH_DYNAMIC, though thats ugly
12:19:39nathanhi can record the bounding box of every draw operation, then only clear those in the next frame
12:20:54preglowlinuxstb: slowcoder confirms that it's just a matter of reading 8kb of data to flush the cache
12:21:33_FireFly_nathanh could you please test this patch ??
12:22:15nathanhive already got that applied
12:22:24nathanhbut thats nothing to do with what im talking about
12:23:31_FireFly_nathanh the %fc should be on the same line as the %pv am i right ?
12:23:50nathanhill explain it again, i dont think ive made it clear
12:24:04nathanhfirst line of the wps is %pv, thats a dynamic string, contains the volume lke -10 or -24
12:24:14nathanhso its got WPS_REFRESH_DYNAMIC
12:24:30nathanhive got a bitmap that im sticking at (1,1), its a bitmap so its WPS_REFRESH_STATIC
12:24:32mirak_what a moron, I was wondering why &variable was always the same
12:24:50_FireFly_nathanh try this %pv %?fc<%xdh|%xdi|%xdj|%xdx|%xdl|%xdf|%xdk|%xdH|%xdy|>
12:24:52nathanhnow the problem is that WPS_REFRESH_DYNAMIC redraws the %pv every now and then
12:25:13nathanhand that erases the entire row, including the bitmap, and because the bitmap is static it doesnt redraw
12:26:41nathanhok that works, is that because the WPS_REFRESH_DYNAMIC flag applies to everything on the line?
12:27:01mirak_how can I get the size of memory used by a variable ?
12:27:07mirak_ah sizeof !
12:27:18mirak_I remember slowly
12:27:22 Join Lost-notash [0] (
12:27:28_FireFly_nathanh no
12:27:44 Quit Lost-notash (Client Quit)
12:28:06_FireFly_but now the code and the images are on the same line
12:28:35nathanhright... so?
12:28:36 Join Lost-notash [0] (
12:28:39mirak_hem sizeof works for known types
12:28:51_FireFly_but you are right it shouldn't matter where the %xd is located
12:28:53mirak_I can I now how space was reserved for a particular pointer ?
12:29:01mirak_how much
12:29:06 Join Delazon [0] (
12:29:06nathanhmirak: you cant
12:29:07 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
12:29:10nathanhmirak: welcome to C
12:29:21DelazonThanx for the RADIO!!! :D
12:29:36nathanhhrm, let me explain how i understand wps, see if im grokking it
12:29:41Bger10x goes to Linus & petur
12:29:43preglowmirak_: you know because you are the one that allocated the memory
12:29:52_FireFly_mirak_ try sizeof(*pointer)
12:30:07mirak__FireFly_: this returns 1 and I doubt it's 1
12:30:15Bgerbut that will give you the size of the type the pointer points to
12:30:16preglowchar pointer?
12:30:20mirak_& return 4 it's the memory older adress
12:30:26_FireFly_Bger ah right
12:30:28Lost-notashIsmo: sizeof (*pointer) would be useless
12:30:37Lost-notash_firefly even
12:30:40preglowmirak_: fact remains, you allocated the buffer yourself, you should know how big the memory is
12:30:43mirak_preglow: it's a pointer that point to the output frame data from xvid
12:30:52mirak_preglow: it's done by the codec
12:30:52preglowwell, then xvid should tell you how large it is
12:31:01mirak_yes it tells the image size
12:31:07preglowwell, there you are
12:31:08mirak_maybe there is a stat
12:32:38 Join edx__ [0] (
12:32:42 Join Petur [0] (
12:33:07Lost-notashaah, the joys of HTTP, that it allows a connection to go on forever, and thus, web-based irc was born :)
12:33:34Peturyeah ;)
12:34:14Lost-notash(which, depending on how many braindead NAT/firewalls there are between you and the webserver, should allow for 2 hours of channel inactivity before booting you :) )
12:36:14 Quit nathanh (Read error: 113 (No route to host))
12:36:43Bgerhi, Petur :)
12:36:47Bgercongrats for radio
12:36:48 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
12:37:06PeturLinus did most
12:37:34_FireFly_Petur and maybe i have asked the right questions to point to the right way :)
12:37:34 Quit Petur (Client Quit)
12:37:56 Join nathanh [0] (
12:38:00nathanhfirefox: you still here?
12:38:05_FireFly_in conjuction with the soft-I2C
12:38:09 Join Petur [0] (
12:38:11_FireFly_there is no firefox
12:38:23nathanhblah, force of habit
12:39:15Peturwtf... I was too kind to the irc webclient - got kicked
12:40:26PeturI just started on the radio and got stuck when the I2C didn't cooperate. Haven't checked what Linus did yet (no time atm)
12:40:45 Quit textchimp ()
12:45:05Mongey|Enabled the virtual ATA LED on the H300<<< wats that do
12:45:48 Quit Jungti1234 ("bye")
12:47:05Lost-notashnathanh: either the ATA led is software enabled, or it's talking about turning on an icon/image on the LCD.
12:47:37nathanhLost-notash, why you telling me?
12:47:58 Join DrumRBoy320|away [0] (
12:48:24 Quit muesli_- (Read error: 110 (Connection timed out))
12:48:25Lost-notashnathanh: because pressing tab in firefox when i'm using the webclient jumps from 'Mongey' to 'nathanh' instead of completing 'mongey
12:48:30Lost-notashbloody thing
12:48:36nathanhwebclient = poo compared to xchat :-)
12:48:57Lost-notashnathanh: normally i'd be using a shell, but that's down because the server's being relocated
12:49:07Lost-notashshould have been back up by now
12:49:16 Quit edx (Read error: 110 (Connection timed out))
12:50:12Lost-notash(of course DNS changes take time when not managed properly, and when the hosting is free, or freeish, it's not a priority
12:53:06 Join muesli_- [0] (
12:59:25Mongey|so is it anything signifacant
12:59:51Lost-notashit's just a minor feedback path, probably
13:00:21*Lost-notash doesn't know for sure, since the ATA LED on the H1xx is hardware controlled, from memory
13:01:52BgerMongey| yes, it's not significant
13:01:56 Quit Lost-notash ("CGI:IRC (EOF)")
13:01:57 Join Lost-notash [0] (
13:02:05Lost-notashbloody web irc client
13:02:14Bgerjust gives u feedback when rockbox reads/writes to disk
13:04:12 Quit Mongey| ()
13:04:34 Nick edx__ is now known as edx (
13:06:27Peturhmmm... I don't see a reason why recording shouldn't work on H300. Too bad I can't build RB where I am now :-\
13:06:42PeturAnd no time tonight :(
13:08:41BgerPetur someone said it doesn't work or ?
13:08:50Bgerit's just not enabled
13:08:56Peturit's not enabled
13:09:06Bgerany hints how to enable it ?
13:09:23PeturHAVE_RECORDING in the H300 config file
13:09:29Bgerhm, i'll try now
13:10:02Peturaargh ... a reason to stay on this irc... should be working ;)
13:10:28B4gdertoo much work makes you a dull boy! ;-)
13:11:22Peturno work tonight! Going to a concert of my favorite rockband
13:11:33markun_B4gder: I emailed the webmaster of a japanese gigabeat wiki if he can help me find gigabeat owning programmers
13:12:30PeturBger, keep that reset pin with you :)
13:12:39Bgerhahaha ok
13:12:40 Join Zak1392 [0] (
13:12:55PeturUsed quite a lot lately ;)
13:13:09 Join muesli- [0] (
13:13:22Zak1392hey guys
13:13:23aliaskFrom memory enabling recording on H300s freezes them.
13:13:32Bgeri use a toothpick
13:13:36 Nick markun_ is now known as markun (
13:13:43aliaskHey Zak1392
13:13:58Bgeraliask when did u test this last time ?
13:14:22BgerPetur we need some button defines also, going to correct this
13:14:36aliaskLinus did, a few days ago. I'll check the logs
13:14:37 Join webguest05 [0] (
13:14:50DelazonI like that the radio keeps playing even when you manage your files with the USB interface :D
13:14:51PeturThe audio mux was wrong, so that might be the reason
13:14:51 Quit Lost-notash ("sleep")
13:15:20Zak1392Would it be possible to have a wps sort of thing for the FM radio?
13:15:22***Saving seen data "./dancer.seen"
13:15:41Peturunified GUI
13:15:51Peturlots of work
13:16:16 Quit webguest05 (Client Quit)
13:16:18Delazonis there any GUI in color like the iRiver GUI planed for Rockbox?
13:16:56B4gderyes please
13:17:04Zak1392color bitmaps in wps's would be a start
13:17:06nathanhthats what im coding
13:17:15MrShleeThe FM doesn't get cancelled when USB mode is activated
13:17:33MrShleesure its good to have some music when your transfering files but would be worth fixing.
13:17:34aliaskI noticed that as well today
13:17:35Zak1392nathanh: color bitmap support?
13:17:40nathanhZak1392, both
13:18:14nathanhbut first i need to understand the lines code, and how it renders wps files
13:18:25nathanhthere is ... oddness
13:18:27aliasknathanh: I think you could use some of eli_sherer's code for that
13:18:34nathanhaliask, link?
13:19:04aliaskUnder the rockpaint plugin, there is a c file dedicated to reading colour bmp files.
13:19:36B4gderthere was also an incomplete patch posted to the dev list with a color bmp loader
13:19:46nathanhloading the bitmaps isnt too hard, its incorporating it into gwps without breaking the rest of it
13:20:15aliaskBger: I can't seem to find where LinusN actually enabled recording. Maybe it wasn't him at all.
13:21:02Bgertime for test ... :)
13:21:08aliaskGood luck! :)
13:21:31Bgeroh, settings reset :)
13:21:44Peturthat's normal
13:21:50Bgeryes, ok
13:23:21XavierGrit crashed?
13:23:39Peturtoo bad - would have been too easy
13:23:42Bgergoing to recording screen, that is
13:23:42aliaskOh well, at least the results are consistent with my memory. I'm not going crazy yet!
13:24:00Bgerhaha at least radio works :P
13:24:12Peturalright, gotta go
13:24:19Bgerhm, strange
13:24:28Bgerit's not dead
13:24:42Bgerthe backlight thread is ok and button reading :P
13:25:07 Quit Petur ("CGI:IRC 0.5.7 (2005/06/19)")
13:25:37Bgerthe hard shutdown worked too
13:25:40aliaskI suppose you _could_ send S.O.S. signals with the backlight if you were desperate.
13:26:14Bgeri prefer not having the need of doing it :P
13:26:40XavierGraliask I think it is quite easy to implement that on a plugin.
13:26:53aliaskThe S.O.S thing?
13:27:27aliaskI'll get on it XavierGr
13:28:37Bgerso i tried entering in the recording screen and it nearly hung
13:31:27Bgermaybe i should try some logf ? :)
13:31:45 Quit muesli_- (Read error: 110 (Connection timed out))
13:33:24Bgerdsp.c: In function `apply_crossfeed':
13:33:26Bgerdsp.c:476: error: can't find a register in class `ADDR_REGS' while reloading `asm
13:33:45 Join DrMoos [0] (
13:34:22preglowyes yes ye syes
13:34:23preglowdebug build
13:34:31preglowyou need to enable fomit-frame-pointer
13:34:40Bgerah, yes
13:34:50Bgersorry, but how do i do this ? :)
13:35:01preglowis there a debug flag set when doing debug builds?
13:35:02Bgerthere are not enough free regs, yep ?
13:35:05preglowi just wanna disable that entire routine
13:35:28 Quit Moos (Read error: 104 (Connection reset by peer))
13:36:33 Nick DrMoos is now known as Moos (
13:37:21Bgerpreglow : according to configure ...
13:37:22Bger [Dd])
13:37:23Bger echo "define DEBUG"
13:37:23DBUGEnqueued KICK Bger
13:37:23Bger debug="-DDEBUG"
13:37:23Bger GCCOPTS="$GCCOPTS -g -DDEBUG"
13:37:34preglowi'll just disable that opt for debug
13:37:49Bgerpreglow wanna test it ?
13:38:51Bger#if defined(CPU_COLDFIRE) && !defined(SIMULATOR) && !defined(DEBUG)
13:39:02preglowyes, i know
13:39:25Bgerno, i mean, should i do it this way
13:39:32preglowi'll commit it now
13:39:39preglowand yes, that's what i do
13:41:24Bgerit seems i lied you
13:42:18 Join DJDD_ [0] (
13:42:40Bgerabout -DDEBUG
13:42:52Bgerthis is true only if you choose Developer->Debug build
13:43:15_FireFly_but not when logf-build i guess :)
13:43:25preglowdebug right
13:43:49Bgersorry :(
13:43:53preglowwhy the hell is fomit-frame-pointer not in debug builds anyway?
13:44:00preglowwhy don't need it
13:44:15 Quit muesli- (Read error: 104 (Connection reset by peer))
13:44:17DelazonOne question?!
13:44:34BgerDelazon just ask
13:45:00DelazonDo you have any idea about videoplayback? Is it possible with rockbox and if so, how far away is it?
13:45:00Slasheripreglow: Hmm, if i am correct, that will disable crossfeed also from logf builds.. not good
13:45:23BgerSlasheri no, it wont :)
13:45:30Slasheriah, then it's ok
13:45:42Bgerbut in fact that's the problem now :)
13:45:57Bgeri'm trying to make a logf build
13:46:10Slasherioh, then you just should fomit the frame pointer..
13:46:25preglowbah, fuck it
13:46:36preglowi'll remove it again, and someone who cares can fix the issue with fomit-frame-pointer
13:47:04BgerSlasheri : just added && 0 to the opt for now :)
13:47:07SlasheriHmm, maybe i could do that.. But currently my configure script has flash stuff etc
13:47:33Bgerbecause u flash every day ? ;)
13:47:54SlasheriBger: just add GCCOPTS="$GCCOPTS $GCCOPTIMIZE" to the logf section
13:48:08Slasherino, but that is needed to compile the flash image and flash bootloader :)
13:48:23CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:48:23*Bger hates iriver fw
13:49:06Bgerespecially the fact that it reads the whole disk on usb plug-in
13:49:12 Quit dropandho ()
13:49:22preglowBger: so do we, if you use dircache
13:49:36Bgeryes, but u don't have to wait
13:50:55Slasheriand btw, tagcache is progressing good also :) And loading it to ram at boot up should not cause any noticeable delay.. Much better than with iriver fw with 20-60s delay :)
13:51:34Zak1392has anybody tried this patch:
13:51:42Zak1392i'm compiling now
13:52:03_FireFly_Zak1392 the unsimplified version is already in cvs
13:52:18_FireFly_this is only an simplified version
13:52:46Bgerok, the logf doesn't show anything more
13:57:23BgerXavierGr, are you here ?
14:04:58aliaskOh well, I'm out for the night. G'night everyone.
14:05:14 Quit aliask ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
14:08:23preglowhow do you control bejeweled on an ipod right now?
14:09:31linuxstbpreglow: Thanks for the cache info. Because there are called to lcd_update() between the loading of different plugins, we can be sure that the plugin ram is not in the cache. So that's not the problem.
14:09:56 Quit linuxstb ("Leaving")
14:10:00 Join saa[b_r]ider [0] (n=saab_rid@
14:13:22 Quit merbanan (Read error: 104 (Connection reset by peer))
14:23:02Bgercan anyone explain me why's that
14:23:04Bger stack = rb->plugin_get_buffer(&stacksize);
14:23:04Bger stack = (void*)(((unsigned int)stack + 100) & ~3);
14:23:04Bger stacksize = (stacksize - 100) & ~3;
14:23:36Bgeri mean why's that +100
14:24:07preglowshould the new latex docs be in a separate cvs checkout?
14:24:12preglowlike cvs co manual
14:24:24B4gderyes I think so
14:24:45preglowme too
14:24:52preglowit's looking ripe for commiting about now
14:25:37BgerB4gder ?
14:27:10preglowwoop, meting
14:27:37mirak_preglow: have you noticed that there is two xvid_malloc methods ?
14:27:55mirak_with a different prototype. I though it wasn't possible in C
14:30:28Bgerit is, mirak_
14:30:53Bgerthe compiler distinguish them by parameters' types
14:31:05B4gderno, not i C
14:31:10B4gderonly in C++
14:31:15Bgerreally ?
14:31:25markunmirak_: how's is coming along?
14:32:04markunmirak_: almost ready to compile xvid and to start on a plugin?
14:32:11mirak_markun: I removed mallocs in the exemple file
14:32:25mirak_no we need to remove mallocs in xvid codec
14:32:28mirak_I am on that
14:32:42markuncan't you use the codec malloc for now?
14:32:57mirak_markun: no I don't think it's possible
14:33:16mirak_don't know
14:33:51mirak_I am brainstorming
14:33:54markun01.10.14 # <linuxstb_> Anyway, the malloc situation looks good - all the mallocs happen at the very start of the decoding process, and it doesn't appear to call free
14:33:57markun01.10.32 # <linuxstb_> So the simple malloc implementation used for the audio codecs will work nicely.
14:34:46mirak_yes but at decoder creation there is a resize
14:35:06mirak_because at first you don't know the size of the output frame
14:35:49mirak_for now I am trying to make it work without malloc and for PC
14:36:02markunBut you will know the size if you only allow for frames that are equal or smaller than the screen
14:36:03mirak_the rockbox malloc might trouble my weak C skills
14:37:08mirak_markun: yes, it will be hardcoded with a #define
14:38:03mirak_what I want to do is a struct with char[sizeofimage] around 4 times
14:38:16mirak_then when it asks to create an image, I want to give him the reference to that struct
14:38:22mirak_instead of calling the malloc
14:39:02mirak_markun: maybe you could look at the code and find a good solution
14:39:06B4gderthen make one of a fixed size fo the worst case
14:39:17markunI don't have time now :(
14:39:44BgerXavierGr ?
14:39:49markunCan you decode directly into the frame buffer?
14:40:54mirak_markun: I am not thinking about that yet
14:41:06mirak_having it working static for PC would be fine first
14:41:13markunyes, sure
14:42:28 Quit Zak1392 ()
14:46:19 Join mikearthur [0] (
14:51:23 Quit nathanh ("Leaving")
14:53:31 Join Kohlrabi [0] (
14:54:08 Quit _FireFly_ ("Did anyone see my lost carrier?")
15:01:34XavierGrBger: What?
15:04:17XavierGrhm you want to ask about the stack definition? Truth is that back then I din't know how to use plugin_get_buffer so I copied the creation of the new thread from alpine_cdc.
15:04:27Bgerjust saw it
15:04:32XavierGrNow that I see that I don't know why it is used.
15:04:45 Quit DJDD_ ("Trillian (")
15:04:50XavierGrDoes the plugin works on H300?
15:05:03Bgerbut if i'm right, there is an buffer overflow in your implementation
15:05:33Bgerhm, there is not
15:05:42Bgerbut the check is not correct
15:05:56XavierGrit works or not?
15:06:00Bgerworks, yes :)
15:06:06Bgerjust to note something
15:06:32Bgerlook at your thread()
15:06:41XavierGrI just don't know stack = (void*)(((unsigned int)stack + 100) & ~3); means
15:07:04Bgeru have there 2000 bytes (and more) on the stack
15:07:19BgerDEFAULT_STACK_SIZE is 0x400 = 1024bytes
15:07:54XavierGrah you mean that all variables in the thread are stored in the stack?
15:08:06Bgerif i'm right, u should change if (stacksize < DEFAULT_STACK_SIZE) to if (stacksize < SIZE)
15:08:13XavierGrso it will overflow when it reaches 1024?
15:08:17Bgerno no no
15:08:44Bgeru get very big stack from the remaining plugin memor
15:08:55Bgeron iriver it's 768k (iirc)
15:09:03Bgerand your plugin is ... 10 k maybe ?
15:09:12XavierGrSIZE is just a definition for the char array that holds the temporary changes between a disk activity
15:09:28Bgeryes, but... ok, let me explain u
15:09:45Bgerstack = (void*)(((unsigned int)stack + 100) & ~3);
15:10:02Bgerstack = rb->plugin_get_buffer(&stacksize);
15:10:02Bger stack = (void*)(((unsigned int)stack + 100) & ~3);
15:10:33XavierGryes it sets the adress after the plugin into the plugin buffer.
15:10:38Bgerthe first line sets "stack" pointer to the remaining memory of the 768k
15:10:40XavierGrthough I dont understand the second line.
15:10:52Bgerthe second line is nearly simple :)
15:10:55XavierGryuo and the 2nd?
15:11:26Bgeri think Jorg adds 100 bytes just to be sure and safe
15:11:38 Join actionshrimp [0] (
15:11:48XavierGrthat is clear from the + 100 but why is there a ~3?
15:12:00Bgerand & ~3 part alligns the address of the "stack" to 8 bytes
15:12:02XavierGrand what "~" does?
15:12:16XavierGrbitwise operation?
15:12:17Bger7 = 0000 0111 binary
15:12:26Bgerand ~7 = 1111 1000
15:12:51XavierGrthe NOT operation?
15:12:56Bgeri think one thing and write other
15:13:05Bger3 = 0000 0011
15:13:14Bger~3 = 1111 1100
15:13:33XavierGrinversed, right
15:13:55XavierGrI must learn bitwise operation.
15:14:03Bgerthe whole idea behind this is that on sh1 u cannot do memory unaligned accesses
15:14:35Bgerheh, i learned this ~ 2 days ago or so :)
15:14:44lostlogiccan the virtual led be put in WPSs?
15:15:02Bgerlostlogic not yet :)
15:15:23***Saving seen data "./dancer.seen"
15:15:32BgerXavierGr all i wanted to say is that the if (stacksize <DEFAULT_STACK_SIZE) check is not correct
15:15:49Bgerbecause your thread needs more than 2k stack
15:16:05XavierGrDEFAULT_STACK_SIZE is defined elsewhere
15:16:05XavierGrah yes
15:16:15XavierGrthat's what I meant befire.
15:16:32XavierGrso my thread uses more than 1024 bytes of memory.
15:16:36Bgerotherwise the plugin looks pretty good ;)
15:17:13Bgerhm, this +100 (and -100 to the stacksize) is very probably for local variables
15:17:15XavierGrthen either the we must change the comparson and create a thread with a greater stack
15:17:28XavierGror reduece the SIZE to 1024 (remove it completely) and use the DEFAULT_STACK_SIZE.
15:17:58Bgeri don't think this plugin will ever write "out of memory" :)
15:18:20XavierGrI should put 100 bytes for the jpeg viewer too. Because I use the plugin_get_buffer.
15:18:21Bgerjust change it to SIZE
15:18:34Bgermaybe more
15:18:36XavierGryeah I will do that.
15:18:49XavierGrThanks, I learned something new!
15:18:50Bgerand, one more note ... use more descriptive defines ...
15:18:55BgerSIZE of what
15:19:00BgerBATTERY_OFF ...
15:19:45XavierGrFact is that I don't know when the char array will get full, because I am note sure how many changes in the batter_voltage can occur between 2 disc accesses.
15:19:49Bgerhm, the gui is not exported to plugins yet, isn't it?
15:19:55Bgerthe new gui i mean
15:20:17BgerXavierGr but you know the size of the array
15:20:18XavierGrI am not sure.
15:20:23Bgerand you can check it
15:20:42XavierGrbut maybe it will be full and lost some changes.
15:20:51MrShleeBtw.. Do we know if the Battery chargers on the Rockbox firmware.. H3XX?
15:21:03BgerMrShlee yes, it charges
15:21:10XavierGrI mean that maybe it will fill the array. When that happens update will be ingored until a disk access is triggered.
15:21:23Bgerbut rockbox doesn't yet know whether it's fully charged
15:21:53MrShleeI dont think my Iriver knows that.. my battery is boned
15:22:01MrShleelike 2 hour max life.
15:22:31BgerMrShlee iriver fw or ?
15:23:32MrShleeIriver fw never seems to charge full.
15:27:36BgerMrShlee are you sure that u've left the unit connected to the charger enough time ?
15:27:42Bgerin rockbox i mean
15:28:35MrShleefor a long while.. yeah
15:29:59Bgerhm, strange
15:30:26MrShleefull charge in iriver fw is 78% (8hours) in rockbox.
15:30:33MrShleebut it will only last a limited time.
15:35:24mirak_preglow: ok I accidently pasted the xvid_malloc method over another one name. I don't kno what happened ...
15:36:45 Quit MrShlee (Read error: 104 (Connection reset by peer))
16:01:17 Quit Delazon ("Chatzilla [Firefox 1.5/undefined]")
16:19:14lostlogichave 6 line patch for hdd led in wps support, will patch tracker it later.
16:19:23 Join b0br [0] (
16:29:59 Quit B4gder ("Time to say moo")
16:32:50 Quit Zagor ("Client exiting")
16:35:39 Quit z35 (Read error: 110 (Connection timed out))
16:38:25 Join Nixsos [0] (
16:38:28 Join z35 [0] (
16:39:00markunpreglow: Isn't -fomit-frame-pointer a good option to add anyway?
16:39:15 Join lamed [0] (
16:40:00lamedpreglow: around?
16:40:10markungood day lamed
16:40:24lamedgday to you too!
16:40:47lamedI see you stopped fiddeling with unicode for the last couple of days
16:41:54markunyes, busy to finish some things for school before I leave on holliday (only a week)
16:43:04markunThe font caching code has be made faster.
16:43:48markunWith a big font (for small fonts no caching is used) and a lot of long filenames the treebrowser lags like hell
16:44:18lamedyeah! thank you! i was wondering if anyone spotted it...
16:45:27lamedI see preglow hadn't had time to go through my pitch.patch patch..
16:51:54lamedok it's at sourceforge now...
16:52:04lamedtake a look
16:54:45 Join qwisp11 [0] (
16:55:27markunDoes anyone here speak Japanese well?
16:56:54mirak_what is the adress bu size of coldfire ?
16:57:01mirak_markun: sayonara
16:57:24 Join _FireFly_ [0] (
17:01:16markunmirak_: people in are all the time nagging for video support. If you finish the xvid plugin it might be a motivation for them to help with the porting :)
17:02:06mirak_that's not what while motivate me
17:02:12markunAlthough I don't think that an mp3 player that can only be used with a computer running windows will attract many hackers.
17:02:22mirak_they should start to help now if they want it fast
17:02:36mirak_I don't know gigabeats
17:03:34markunjust another player.
17:04:42mirak_ if I declare a struct with a char[3] and a int, will the int be right at &char[4] ?
17:04:45markun300MHz ARM and a 320x240 color screen, but they made a very bad firmware. You would expect it to play videos with there specs..
17:05:15mirak_yes definitely
17:05:24mirak_I play divx on a mac G3 400mhz
17:05:40mirak_they might need to use slower resolutions
17:05:44markunxvid doesn't use any floating point, does it?
17:06:02mirak_that implementation avoid it
17:06:08mirak_there is float for postprocessing
17:06:18mirak_but we will never use that
17:06:49markunEven if it decodes at 0.1 fps at first it will be very cool.
17:07:08mirak_that's optimistic
17:07:14markunThe codecs were so slow in the beginning..
17:07:32Bgerthe mp3 was under 10% realtime iirc
17:07:52markunI thought it would be difficult to get them real time, but linuxstb and preglow did it with only a few optimisations.
17:07:55*Bger is happy with the FM tuner
17:08:10mirak_Bger: on H300 ?
17:08:17mirak_the tuner
17:08:18 Join NicoFR [0] (
17:08:25mirak_I will try that
17:08:55Bgerbtw, because the cpu runs @ 11MHz while in tuner and the lcd_update is awfully slow....
17:09:06mirak_So I was asking just before, if data type inside a struct are automatically aligned to the bus size
17:09:13markunvideo, wma and mp3 encoding are they only advantages the original H300 firmware has over rockbox right?
17:09:35markunmirak_: I'm not sure. You would think they are.
17:09:35mirak_ if I declare a struct with a char[3] and a int, will the int be right at &char[4] or will it skip some space to be on the good base adress ?
17:09:37Bgerand runtime ...
17:09:51markunand usb charging?
17:09:59Bgerand USB host
17:10:10mirak_why 11mhz ?
17:10:13markunwell, quite some things on the todo list then :)
17:10:24Bgeri want to participate in the last
17:10:39Bgermirak_ more battery life
17:11:26qwisp11USB host was a bit rubish when if you had any sort of computer near by you could p&p to get USB 2 but USB host was only USB 1
17:11:35markunSo many people working on useful code for the gigabeat firmware. I should just wait more and it will practically be finished :)
17:11:52Bgerhahaha markun
17:12:20Bgerdon't wait, otherwise gigabeat will follow the destiny of other rb-ed players ...
17:13:07Bgerout of production before rb being ported ....
17:13:19markunthe nice thing is that the laters generation Gigabeat players (the X models) use exactly the same components as the F series.
17:13:33Bgeronly the display is different ?
17:13:55markunyes, and they made it more compact by rearanging everything.
17:14:12Bgerhehe that's goot
17:14:27markunBut the F series seem to use 3 different LCD's anyway, so that's a problem.
17:14:29mirak_coldfire bus size is 32bits I guess
17:14:40BgerXavierGr i edited a little battery_bench
17:14:53Bgermirak_ iirc, it has 16bit bus to the SDRAM
17:15:25***Saving seen data "./dancer.seen"
17:15:36markunI don't know, but I would have guessed 32
17:16:09markunmirak_: it's probably in the wiki and in the datasheet
17:16:23BgerXavierGr i'll give it to u tomorrow
17:25:12 Quit Rob2222 ()
17:27:03 Join RotAtoR [0] (
17:27:57mirak_size_t is a C type ?
17:33:55 Quit b0br ("CGI:IRC")
17:36:42preglowlamed: i've been busy
17:37:13preglowand i'll continue to be so for some hours
17:38:17lostlogicxsourceforge has been down more than up lately.
17:39:06preglowwell shit, surprise
17:39:15 Quit NicoFR ()
17:41:02mirak_I want to change the malloc function with my malloc, like it's done for the codec, but I use a different one for now, but will change after.
17:41:23mirak_so I created a buffer
17:41:37mirak_static char static_malloc_base[4*1024*1024];
17:41:37mirak_static int static_malloc_offset=0;
17:41:51mirak_I used static, I am not sure if that's what I must use
17:42:21mirak_I want this variable to be kind of global
17:42:33mirak_is it ok to do that ?
17:44:09mirak_seems it works
17:45:44preglowlamed: and yeah, i didn't understand what you wanted with pitch.patch
17:46:00preglowlamed: now that i see your description on sf, it's good i didn't commit it, that's wanted behaviour
17:46:23_FireFly_ah the nice filename :)
17:47:09 Join muesli_- [0] (
17:48:37 Quit Strath (Read error: 104 (Connection reset by peer))
17:49:01 Quit qwisp11 ()
17:50:51mirak_preglow: seems my stuff is working, you where right, just a fake malloc is good.
17:51:17mirak_it would crash if there is video resizing of the stream since xvid supports that
17:51:35 Join Acksaw [0] (
17:51:40mirak_and with a 220*176 stream it only takes 1184800 of ram
17:51:43 Quit Acksaw (Client Quit)
17:51:57mirak_by malloc at least
17:52:43 Quit Lynx_ (Nick collision from services.)
17:53:13 Join Lynx_ [0] (
17:55:38muesli_-kölle alaaf :D
17:55:41 Join Strath [0] (
17:56:07_FireFly_muesli_-: na und when interressiert das schon;)
17:56:21muesli_-wayne schon ;)
17:56:54Lynx__FireFly_: mich auch! ;)
17:57:25_FireFly_moep noch so einer
17:58:15Lynx_preglow: yes sir, sorry sir! :)
17:58:30*Lynx_ salutes
17:58:57muesli_-fisher's fritze fishes fresh preglow ;)
17:59:19Lynx_so, there was some talk about xvid earlier here, what's that about?
18:01:30 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
18:02:18preglowmirak is porting xvidcore to rockbox, i think
18:04:35 Quit San (Read error: 104 (Connection reset by peer))
18:05:15Lynx_mirak_: ZMOG, teh cool!! when is it done, when is it done?!?!?!? ;)
18:09:33 Quit _FireFly_ ("Leaving")
18:14:42 Join Lynx0 [0] (
18:20:14saa[b_r]iderguys, is the RB forum down again?
18:20:41muesli_-i am just browsing in it
18:21:34saa[b_r]iderthanks muesli... won't load for me for some reason
18:22:30 Quit Lynx_ (Nick collision from services.)
18:22:50 Nick Lynx0 is now known as Lynx_ (
18:23:02muesli_-some have reported the same while others could browse at the same time
18:39:43*preglow hugs RotAtoR
18:39:56preglowso, how do you control bejeweled on an ipod?
18:40:05lamedtried my patch?
18:41:15preglowthe pitch patch?
18:41:34lamedyeah... lol
18:41:35preglowno, i haven't tried it, but like linus says, it breaks behaviour that should be there
18:41:40RotAtoRpreglow: :)
18:41:42RotAtoRPLAY to start
18:41:42RotAtoRSELECT+PLAY to save/resume
18:41:42RotAtoRSELECT+MENU to exit
18:41:42DBUGEnqueued KICK RotAtoR
18:41:42RotAtoRSELECT to select and show high scores
18:41:42RotAtoRScroll to move
18:41:43***Alert Mode level 1
18:41:43RotAtoRDirections to swap
18:42:00lamedpreglow... what behaviour is that?
18:42:11preglowRotAtoR: so you have to use the wheel to scroll horizontally, with wraparound?
18:42:23preglowlamed: when you press right or left, the setting is supposed to snap back when you release it
18:42:34preglowRotAtoR: sounds a bit cumbersome
18:42:54preglowRotAtoR: can't we use up/down/right/left as well? i don't think those keys are needed for anything else
18:43:16lamedhuh. I was sure it's a mistake. what's the purpose of having it to snap back?! I thought it's ment for faster changing the pitch value, wasn't it?
18:43:22preglowlamed: no
18:43:23RotAtoRhmm, or maybe use directions and scrolling?
18:43:32preglowlamed: it's meant for beatmathcinh
18:43:38preglowbeatmatching, arghh
18:43:39preglowcold fingers
18:44:02preglowlamed: when you try to sync two songs, you sometimes want to pitch up just for a second
18:44:10preglowthat's what those buttons are for
18:45:02preglowi don't know if you've seen someone do it for proper turntables before, but very often you will see them just push a record slightly along
18:45:10preglowthat's exactly what left/right is for in the pitch screen
18:47:29lamedkindof silly imho. I _am_ mixing & syncing songs. you're pretty right of what you're saying, but they have control over the _beat speed_... that means, at first you're supposed to coardinate the two song's beats, and then you'll use quick changes to make the beat sound at the same time.
18:48:12preglowbut i don't see how you can think it's silly then
18:48:56lamedwell, have you ever used a deck?
18:51:44***Alert Mode OFF
18:54:43lamedI'll agree with you that it's Kinda simillar, at least if when you keep holding left/right, it speeds up and not only changing by a constant value of two, which isn't always enough. then, when you'll leave the ON button it will go back to the latter speed. even if so, I'll still think that because of the time it takes RB to actually change the playback speed, it's not quite ideal to use that setting for minor changes at all. I gtg but i'll go through
18:55:16preglowlamed: i agree, but it wont always have this slow response time
18:55:22preglowlamed: we'll fix it so the change comes at once
18:55:46preglowbut it just needs some work
18:56:44mirak_does rockbox support printf or other system calls ?
18:56:51preglowmirak_: snprintf
18:56:58preglowwhich writes to a buffer
18:57:02preglowthen put it on the screen with lcd_puts
18:57:34lamedpreglow: btw by any change will we have seperate control over the current and next playing song pitch?
18:57:39 Join webguest51 [0] (
18:58:14preglowlamed: well, it's possible, of course, i just don't see the use
19:00:02lameda pretty far out there thought. there are some bpm recognization softwares, haven't investigated into that, but we might have RB to auto sync songs
19:01:02mirak_preglow: can I still build as if even with printf or will it fail to build ?
19:01:19preglowlamed: hmm, yes, it's possible, i guess
19:01:29preglowlamed: but i've never seen _good_ bpm detection
19:01:35mirak_I am trying to have something that built even if not functional
19:01:43pregloweven traktor, which plenty of people use with final scratch, is really glitchy
19:01:59preglowmirak_: it will build, but it wont link
19:02:38mirak_I need some help to do the make etcetera
19:02:47mirak_some path
19:03:32 Quit webguest51 ("CGI:IRC")
19:03:57lamed(btw, if you _do_ have very fast control over the pitch, you can fade it to a full stop pretty neatly. that's a nifty effect dj's use from time to time, and I think that the hardware will allow slower then half of the normal playing speed.) preglow: if we could only get two seperate output channels from the player... a deck from a box. :D
19:04:16lamedok gtg 4 real cya
19:04:25preglowlamed: i'd like to be able to rewind as well :PP
19:04:33preglowlamed: but that'd require a _REALLY_ fast decoder
19:05:08preglowrewind as in spinning the record backwards really fucking fast, then starting over again
19:05:16preglowi don't know how common that is outside of drumnbass, heh
19:06:03 Nick RotAtoR is now known as RotAtoR|Away (
19:06:03DBUGEnqueued KICK RotAtoR|Away
19:07:57 Quit tvelocity ("Leaving")
19:08:49hshahthe forums still not working?
19:09:25mirak_in the make file, what SRC variable is for ?
19:10:41mirak_sorry but I am newbie for that
19:10:58 Join Lear [0] (
19:11:37preglowmirak_: i'm not exactly a makefile expert
19:12:32 Join Petur_ [0] (
19:13:35 Nick Lynx_ is now known as Lynx_awy (
19:13:45mirak_ok it's used to create the OBJ list
19:15:26 Quit Petur_ (Client Quit)
19:15:27***Saving seen data "./dancer.seen"
19:15:44LearHm... To play an AAC on Rockbox, it _has_ to be encoded with iTunes?
19:16:52 Join petur [0] (
19:17:09preglowLear: ask the faad people
19:18:00peturBger: what's your conclusion regarding recording?
19:18:14petur(on H300)
19:18:26peturhe was trying to enable it
19:19:01Learpreglow: well, I did look at why a Nero-encoded file crashed, and that was due to Rockbox alright...
19:19:17preglowcontainer parsin?
19:19:25LearThe metadata reader didn't get some essential information, and promptly crashes with a divide by zero... :)
19:19:41preglowyou'd have to ask dave about that, he wrote it
19:20:00LearIs that linuxstb?
19:20:40Slasheriinteresting.. i hope that metadata thing wouldn't crash very easily because with tagcache, all files on disk are passed through that code :)
19:21:27preglowwell, it's just another reason to make it rock solid
19:23:34amiconnThere's another place that crashes with a divide-by-zero in the sim: skipping tracks with crossfade disabled
19:23:59LearIsn't there patch on the tracker for that?
19:24:32preglowahh, sim
19:25:23 Quit lamed ("CGI:IRC (EOF)")
19:26:33Slasheriamiconn: Hmm, interesting. Do you have a backtrace?
19:26:58Slasheri(can't test sim right now..)
19:28:24LearSlasheri: or look here:
19:29:06peturon H1xx, is microfone internal or external?
19:29:22SlasheriLear: oh, that should be committed
19:30:06preglowLear: i've probably asked this a ton of times: did you check out lostlogic's tremor patch?
19:30:10 Join Benacool [0] (
19:30:28peturhehe, that's my first patch :D
19:30:51amiconnStrange patch that is...
19:30:57peturI know
19:31:11SlasheriHmm, indeed.. that fix seems a bit strange
19:31:13Benacoolis it normal that when i plug my h320 with USB that the Radio stay on but the signal is weeker ?!
19:31:17peturit should be fixed another way
19:31:25peturit did solve my problem
19:31:42Slasherifade_insert shouldn't be called at all if crossfade is not enabled
19:31:51peturthat's what I mean
19:32:15Slasheripetur: but feel free to fix the crossfade, that would be great :)
19:32:40peturlooking into recording now, and don't have much time tonight
19:33:13 Join linuxstb [0] (
19:33:18LearBenacool, sounds similar to my H140. When I connected it to my work laptop, I could hear very faint radio in the headphones. :)
19:33:24peturI couldn't fix it better because I don't understand the framework (what is called where)
19:33:59Benacoolyea it's weird xD
19:34:03linuxstbLear: As you've noticed, the mp4/m4a parser needs a lot more work.
19:34:40linuxstbBut the parser itself shouldn't crash - so get_metadata() and tagcache should be fine.
19:34:45Slasheripetur: ah :) did you check the backtrace with debugger? fade_insert should be called by crossfade(..) only and that should be called if crossfade is enabled when inserting data into pcm buffer
19:34:50LearThere seems to be no metadata tagging standard. Seen three different versions (iTunes, Nero, Foobar2000)...
19:34:53linuxstbIt just fails to correctly parse some files.
19:35:26Benacoolwhen the hard-drive spinning the radio signal is weeker ...
19:35:33peturI shall make a note of it ;)
19:36:34Slasheripetur: ah, in fact crossfade is called always when skipping tracks manually.. either with crossfade mode or flush mode
19:36:43peturnow is 'mic' on H1xx internal or external?
19:37:46Slasheripetur: but thanks trying to fix it, i found now the real problem :)
19:37:48preglowgod, we need to fix this bmp mess with the wpses
19:38:06Learwhat mess?
19:38:15preglowa jillion bmp files per wps
19:38:59LearThere's a patch for that too... :)
19:39:05preglowdisk cache crashes the nano
19:39:26preglowok, so now i need to reset the settings somehow
19:39:35preglowlinuxstb: help
19:39:44linuxstb:) I did that yesterday...
19:39:50preglowi just needed to verify...
19:40:09CtcpIgnored 2 channel CTCP requests in 24 minutes and 21 seconds at the last flood
19:40:09*petur runs off to cook some dinner
19:40:19preglowok, any easy ways of reseting settings in windows?
19:40:41linuxstbThe way I fixed it was to "dd" the first 64 sectors from /dev/sdd, and then use a hex editor to clear the config sector values, and then dd those sectors back.
19:40:48preglowi don't think i'll try that
19:40:57preglowthink i'll just bump the config sector version in a custom build
19:41:10linuxstbThat will work better.
19:41:13preglowthis pc doesn't have cygwin anyway
19:42:36preglowwhere is the config sector version number?
19:42:47linuxstbprobably apps/settings.[ch]
19:42:53preglowfigured that much
19:43:37linuxstbNear the top of settings.c - CONFIG_BLOCK_VERSION
19:44:00preglowi wonder why it does this
19:44:05preglowprobably something funky with the ata driver
19:44:28linuxstbDid you try disabling the sleep?
19:44:34preglownah, but will
19:45:08preglowi've just got to finish this button driver today
19:45:27linuxstbIIRC, you just need to comment out the only call to create_thread() in drivers/ata.c, and also comment out the contents of the ata_sleep() function.
19:45:59linuxstbIn fact, now I've got your button source, I can try it myself.
19:46:49preglowdone it now
19:47:24preglowhmm, what the hell
19:47:33preglowthe button driver is really glitchy
19:48:01preglowloading bejeweled now....
19:48:09preglowand still...
19:48:13preglowok, it's hanged
19:48:38preglowok, there
19:48:51preglowit's real purdy
19:49:04linuxstbDid it load?
19:49:13preglowyes, second time around
19:49:17preglowloading is sporadic at best
19:49:21linuxstbI've experienced the same
19:49:30linuxstbSo I don't think it's specific to the Nano.
19:49:55preglowsome graphical glitches, so rotator's got some work left
19:51:18 Quit aegray__ (Remote closed the connection)
19:51:21preglowit's just a bitmap bug in some tiles, it seems
19:51:26linuxstbThe SELECT+MENU combination doesn't seem to work though - should it?
19:51:35preglowi have no idea
19:52:08preglowgreen tiles seem to have a glitch in lower right corner
19:52:25linuxstbYes, I get the same
19:53:04preglowok, so how does one exit? :)
19:53:14linuxstbReboot :)
19:53:18preglowno, i did it
19:53:28linuxstbYou're better than me then.
19:53:29preglowhmm, now for how
19:53:33linuxstbHave you tried starfield?
19:53:38preglowmenu + select did it
19:53:53preglowoh, but how do i exit the menu?=
19:53:57preglowwill try starfield now
19:54:20linuxstbThe clickwheel is used to control the speed (positive and negative) of the stars
19:54:36linuxstbBut be sensitive...
19:55:05preglowlooks damned cool
19:55:22Benacoolis there a limit of speed in starfield ??
19:55:38Benacooli was at more that 2000 yesterday ...
19:56:48preglowsewcond time plugin load crashes a lot
19:56:53preglowthat's the bug you thought was cache related?
19:57:04preglowcan't see how it can be
19:57:07preglowi loaded the same plugin agin
19:57:43linuxstbYes, it's now obvious it's not the cache.
19:58:33preglowshould snow work?
19:59:25linuxstbI think so.
20:00:00linuxstbBut it seems to freeze for me.
20:00:16preglowdon't look like it
20:00:20preglowfreezes here
20:00:28 Join muesli- [0] (
20:00:29linuxstbWell, it's snow...
20:00:34 Quit muesli_- (Read error: 104 (Connection reset by peer))
20:00:41preglowperhaps it does white snow on a white background
20:00:56linuxstbThat's a cheap plugin.
20:01:03preglowsnake2 doesn't exactly fly either
20:02:20preglownow the green jewels look just fine...
20:02:47 Join San [0] (
20:04:39preglowlinuxstb: did you fix the apple boot bug?
20:04:48preglowlinuxstb: ahh right, sorry, that was the cop sleep tyhing
20:05:31preglowhah! i managed to exit bejeweled completely
20:21:39 Join xmixahlx [0] (n=xmixahlx@
20:24:03mirak_anyone can explain the viewersconfig file ?
20:26:42 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
20:29:39markunhi t0mas
20:30:37markunmirak_: extension,plugin,icon (in rb bitmap format)
20:31:07 Join akaidiota [0] (
20:31:15 Join Rob2222 [0] (
20:33:11 Join JazzBone [0] (
20:36:38mirak_markun: how do I bind a plugin into the code ?
20:37:35mirak_I just need to put the C file into the plugin rep and make will find it ?
20:38:03mirak_maybe SOURCES
20:38:17markunyes, put it in SOURCES
20:39:23markunHow does it run on the PC now?
20:39:30mirak_what do you mean ?
20:39:40markunDid you get rid of the mallocs?
20:40:07markunnice. What does the exmaple do? Output the first frame as a picture?
20:40:09mirak_that was easy to do but I wasted time using a different approach than using a fake maloc
20:40:29mirak_markun: the exemple have a decoder and an encoder
20:40:40mirak_plus a stat outputer
20:40:47mirak_we used only the decoder
20:41:04mirak_actually it allows to read a raw xvid file
20:41:13mirak_and extract the image
20:41:44markun1st image?
20:41:49markunOr all of them?
20:42:21markunWow, sounds like the plugin isn't far away..
20:42:34mirak_if you help me :)
20:42:54mirak_can I put a path to the c file like ./thefile.c
20:42:58lostlogicxpatch 1381872 added to patch tracker for virtual LED support in WPS.
20:43:00mirak_in SOURCES ?
20:43:23markunjust look at the SOURCES file.
20:43:30markunDo it just like all the other plugins.
20:48:33 Quit akaidiot (Read error: 110 (Connection timed out))
20:50:22 Join Maxime` [0] (
20:50:44 Join ^BeN^ [0] (
20:51:04^BeN^mmm i have a large bmp
20:51:17^BeN^and i want to converte it with bmp2rb
20:51:45^BeN^and the code dont join to all the windows
20:51:51^BeN^what can i do?
20:58:15 Quit linuxstb (Read error: 110 (Connection timed out))
20:58:28mirak_markun: I have a question :D
20:58:48mirak_since there is not just one .c file but also a full folder
20:58:59 Join linuxstb [0] (
20:59:01mirak_I don't know how to invoke a local make
20:59:18mirak_I guess the folder must have is own make like other codecs
20:59:25mirak_rockboy for exemple
20:59:46peturadd it to SOURCES and do a normal make
21:00:26markunmirak_: look at the Makefile in apps/plugins. It also calls make for all the subfolder.
21:02:17markunAh no, you have to add your folder to SUBDIRS in the Makefile
21:03:42markunjust copy the Makefile from rockboy and edit it for xvidcore
21:04:41 Quit Maxime (Read error: 110 (Connection timed out))
21:04:48 Join qwisp11 [0] (
21:06:02 Join DangerousDan [0] (
21:09:18 Join solexx_ [0] (
21:10:16qwisp11Is there any way I can help with rockbox if I can't really program. It's wonderful and I'd like to give something back.
21:10:20mirak_arghhh does rockbox can include time.h ?
21:10:38preglowmirak_: if it's not in firmware/export or apps/
21:10:39preglowthen no
21:15:08 Join Mongey| [0] (
21:15:31***Saving seen data "./dancer.seen"
21:16:39t0masanybody seen this: ?
21:18:55mirak_preglow: help
21:19:22mirak_so this file time.h I take it from /usr/include/ ?
21:20:29preglowhell no
21:20:43 Join linuxstb_ [0] (
21:20:43preglowif rockbox doesn't have time.h in firmware/export or apps/, it's not there and you can't use it
21:20:56 Join mymomthelush [0] (
21:20:58preglowyou can't just use system header files
21:21:41 Quit solexx__ (Read error: 110 (Connection timed out))
21:21:43mirak_ok so I must remove every reference
21:22:22linuxstb_mirak_: The timer code will be very useful for profiling the decoder in Rockbox, so it's best to try and keep it.
21:22:24preglowno one told you it'd be easy :-)
21:22:47mirak_linuxstb: profiling ?
21:22:48linuxstb_But as preglow said, you can't use system libraries that don't have Rockbox equivalents - unless you implement the missing functions.
21:23:19linuxstb_mirak_: Yes, xvidcore uses utils/timer.c to time how long it spends in each part of the decoding process.
21:23:43mymomthelushhey how do i install the daily builds on the player?
21:23:45mirak_linuxstb: yes but of what use is it ?
21:23:52qwisp11t0mas> thats great :p
21:24:09linuxstb_mirak_: It could tell us which parts of the library need optimising.
21:24:12preglowmirak_: so we can see where we need to optimise
21:24:20preglowmirak_: this decoder isn't just going to fly out of the box
21:24:25preglowit'll need work, work and then more work to be usable
21:24:33wubblawouldn't it be easier to use not that sophisticated video codecs as MPEG-4, but something like MPEG-2?!
21:24:34linuxstb_This decoder will need a taxi to leave the box
21:24:42mirak_preglow: if we optimise everything we will not have to need it
21:25:09preglowmirak_: yes, but that's what we're saying, those timer functions are nice to see _what_ needs optimising
21:25:14mirak_wubbla: mpeg2 consumes more cpu I think
21:25:17preglowi sure as hell won't optimise everything
21:25:24wubblamirak_: huh?!
21:25:30preglowmirak_: improbable...
21:25:35wubblamirak_: more than MPEG-4?!
21:25:37mirak_wubbla: at least on my mac G3 400 I can't play a dvd
21:25:47linuxstb_The timers are also useful to tell you how much you've improved the decoding _after_ you've done an optimisation.
21:25:55mirak_I think the bitrate needed would to high
21:25:55preglowwhich reminds me
21:25:55qwisp11mymomthelush: then select your player from under the bleeding edge sub heading
21:26:02preglowsomeone needs to look at lostlogic's profiling work
21:26:23wubblamirak_: then you've got a bad/ineffective mpeg2-codec...
21:26:32mymomthelushi love this stuff
21:26:39mymomthelushso much better than the iriver fw
21:26:44preglowdamn straight it is
21:26:53preglowthat firmware makes me want to kill myself
21:26:55mirak_wubbla: on linux with mplayer vlc or xine gxine it's the same
21:27:08mymomthelushyeah only task left is converting my wma files to flac
21:27:18mirak_wubbla: I can play a divx fine but the DVD just lag
21:27:24preglowmymomthelush: wma will come sooner or later
21:27:26mirak_or VOBS
21:27:37preglowwell, ok
21:27:46preglowperhaps because of the larger amount of data
21:27:55mymomthelushwell i finally found a converter for it so im happpy
21:28:02preglowbut the decoding itself should be more intensive
21:28:09wubblamirak_: hey come on! don't try to tell me that MPEG4 is easier (from the CPU's point of view ;-) ) to decode than MPEG2
21:28:40mirak_wubbla: well my poor mac just can't on linux
21:28:44mymomthelushdoes anyone know how to register your nick?
21:28:49peturBger, still not around?
21:28:58mirak_wubbla: on OS X it can use the ATI hardware extraction
21:29:02markunI would prefer a MPEG-4 codec over MPEG-2. The iriver firmware also can do it, so it's not completely impossible.
21:29:29qwisp11someone rang us earlier and asked how using xp was for us. :p what a wierdo
21:29:50mirak_markun: if we can do mpeg4 we can do mpeg2
21:29:59preglowi am an idiot
21:30:31 Quit petur ("here today, gone tomorrow")
21:30:34 Part mymomthelush
21:31:32mirak_markun: if I can do it anybody can do it. if one of you tried it it would probably be done already
21:31:33 Quit qwisp11 (Read error: 104 (Connection reset by peer))
21:32:36mirak_preglow: rockbox have timer functions ?
21:32:51 Join b0br [0] (
21:33:39markunmirak_: maybe you can use rb->current_tick to do some timing
21:34:40 Part b0br
21:34:43 Part Mongey|
21:38:30 Quit hshah (Read error: 104 (Connection reset by peer))
21:40:08mirak_well I just comment this methods for now
21:40:16 Quit joshn_454 ("KVIrc 3.2.0 'Realia'")
21:44:05 Join Mmmm [0] (
21:45:27 Quit z35 (Read error: 110 (Connection timed out))
21:46:28 Join z35 [0] (
21:46:37 Join amar [0] (
21:47:40 Quit amar (Client Quit)
21:54:53 Quit muesli- (Read error: 110 (Connection timed out))
22:10:02 Quit San (Read error: 110 (Connection timed out))
22:10:03 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
22:10:04preglowlinuxstb_: have you enabled fiqs and stuff yet?
22:10:26 Join DreamTactix291 [0] (
22:10:45 Join muesli_- [0] (
22:13:40linuxstb_preglow: No, not yet.
22:14:57linuxstb_I've got to the stage of initialising the audio codec, and am about to start on the tricky stuff.
22:15:07 Join NicoFR [0] (
22:16:54linuxstb_preglow: I also haven't found the bootloader problem yet. All I know is that even the old versions fail to load the apple firmware if the cop is asleep.
22:23:38 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
22:23:43preglowthere's some strangeness in the wheel handler code i don't quite get
22:24:08mirak_if I have a printf("blabla") how do I convert it to rock box api ? int (*snprintf)(char *buf, size_t size, const char *fmt, ...); there is to many parmaetrs on that oe
22:24:45 Quit xbshift (Read error: 104 (Connection reset by peer))
22:24:57lostlogicxmirak_: You need logf support and send it to logf probably unless you want to have your own logfile and use fdprintf
22:25:10lostlogicxmirak_: there's not a 'console' to send things to per-se in rockbox
22:25:36mirak_I don't want to display something for now
22:25:44mirak_text would be fine
22:25:46mirak_on screen
22:25:50preglowmirak_: if you're doing debugging with printf, do what lostlogic says, if you're trying to display something to the screen for other reasons, use snprintf, then lcd_puts
22:26:09preglowmirak_: snprintf writes to a buffer, you then need to write that buffer on the screen
22:26:44mirak_what is size_t ?
22:27:54linuxstb_It's an integer - the type you use to represent the size of a memory buffer
22:28:10mirak_yeah but in that function
22:28:14mirak_what's the usage ?
22:29:04linuxstb_The size of the buffer you are "printing" into.
22:29:40linuxstb_i.e. snprintf(buf,sizeof(buf),"Normal string %d",i);
22:29:53lostlogicxthe maximum number of characters to print. prevents buffer overflwos
22:29:55mirak_that's an ansi method
22:30:05mirak_I mean I am short on docs ...
22:30:11lostlogicxsprintf w/o the n is the way many buffer overflow security holes are created.
22:30:23linuxstb_If you run Linux, just type "man snprintf"
22:30:31mirak_I did that I don't have it
22:30:44lostlogicxman 3 snprintf?
22:30:55linuxstb_mirak: What Linux distro do you have?
22:31:24preglowwonderful, i'm doing tech support in #ipodlinux now
22:31:26linuxstb_apt-get install manpages-dev
22:31:26mirak_what distro do you have ?
22:32:06 Quit NicoFR ()
22:34:08 Join qwisp11 [0] (
22:35:50preglowthis clickwheel driver is... strange...
22:35:51mirak_it's not possible to not use api->snprintf ?
22:35:57preglowmirak_: yes it is
22:36:11preglownot use?
22:36:17preglowif you want a ready made printf, that's what you've got
22:36:31mirak_I can't just do snprintf
22:36:40amiconnpreglow: The clickwheel is strange by itself ;)
22:36:41mirak_I need the api stuff before ?
22:36:58preglowi just find it kind of weird that this drivers works as well as it does in ipl
22:37:16preglowwhen i just apply it naively here, it's real jittery
22:39:25linuxstb_Have you tried disabling the RTC? i.e. remove CONFIG_RTC from your config-ipodnano.h file. It's possible the RTC reading could interfere with the button driver.
22:41:48preglowwe'll see
22:41:53preglowi'll just try rewriting it a bit first
22:42:02preglowno, i'm pretty sure that's not the issue
22:43:53preglowwe have access to the position of the wheel
22:44:08mirak_ok just a question. I have seen in the api there is a strmcp. That's a rockbox optimised function right ?
22:44:20preglowi think i'll just rip out their way of dealing with this, and finding my own way of doing stuff
22:44:26mirak_I wonder generally why some methods can work and why some can't
22:45:12mirak_if I use strcmp without api->strcmp , does it mean it uses gcc method or does it means it find itself the rockbox one ?
22:45:36preglowlinuxstb_: got any ideas of the best way to deal with the clickwheel? just settinga constant threshold would beok, i guess, but i also want some acceleration
22:45:53preglowlinuxstb_: so you'll move quicker through lists if you move your finger real fast
22:48:20 Join idanm [0] (i=idan@
22:49:01idanmhello dudes! anyone knows how to compile wps with the bleeding edge files?
22:49:18idanmI get: "No wps module present, can't do the WPS magic!"
22:51:54^BeN^what do you think? :
22:52:05^BeN^solitaire for h3xx
22:52:51preglowidanm: cvs co wps
22:53:10preglowidanm: you need to check out wps module
22:53:26preglowor are you using a bleeding edge source tarball?
22:53:27idanmdo I have to use csv in order to compile wps?
22:53:28preglowif so, you're out of luck
22:53:59preglowidanm: ok, i don't think you can get wpses without using cvs thus far
22:54:07 Quit Nixsos ()
22:54:30idanmis it complicated to use csv?
22:55:07mirak_it can be
22:55:25idanmis there a simple guide ?
22:55:41Mmmmcvs is easy... just look at the simple guide to compiling in the wiki
22:55:54preglowcvs is very easy
22:55:56preglowjust two lines
22:59:49 Join webguest44 [0] (
23:00:16preglowlinuxstb_: in a position to test a rockbox.ipod for me?
23:01:16preglowif so,
23:01:25preglowthe button driver should be lower sensitivy now
23:02:14preglowstill it's a bit awkward to press a button after using the wheel, but that can be fixed quite easily
23:03:10 Quit Rob2222 ()
23:03:10 Quit qwisp11 (Read error: 104 (Connection reset by peer))
23:04:10idanmwhere do I get the csv program? it seems that the link in the site is outdated
23:05:00idanmwhat's the prefered client name?
23:05:31preglowif you're using windows, tortoisecvs is pretty nice
23:05:39preglowit's gui based
23:05:54 Join Rob2222 [0] (
23:06:53 Join webguest24 [0] (
23:08:56Mmmmidanm: I just use bluechip's devkit which has everything you need...
23:10:24idanmwhere can I find it/
23:11:22MmmmIt's all in the simple guide! :)
23:11:29idanmI have that
23:11:43Mmmmthen you have CVS installed!
23:11:58idanmhmm.. cool
23:14:38mirak_ passing arg 1 of pointer to function from incompatible pointer type , I do that api->snprintf(&screen_buffer,sizeof(screen_buffer),"xv
23:15:10preglowmirak_: what type of variable is screen_buffer?
23:15:13mirak_screenbuffer is char screen_buffer[52]
23:15:19preglowthen screen_buffer is enough
23:15:21preglowno ampersand
23:15:34***Saving seen data "./dancer.seen"
23:15:54idanmit even works!
23:16:33MmmmGood innit?
23:16:47mirak_warning: function declaration isn't a prototype
23:17:03mirak_don't know why it says that
23:17:18mirak_the function and prototype are in th same file
23:19:07*preglow produlates linuxstb_
23:19:19mirak_so it must be in a .h to not have this warning
23:20:16mirak_it must have parameters ?
23:23:42webguest24Show us the prototype
23:23:47webguest24I think
23:24:20 Join pinkutank [0] (i=pinkutan@
23:24:38pinkutankwoo arent we crowded today
23:24:43pinkutankhow are you guys
23:25:14linuxstb_preglow: I've just tried your new build - it works better if you can fix the problem with pressing a button after using the wheel.
23:25:30pinkutankIw anted to ask if we lsoe drm if we use us firmware to bootload the h300
23:25:34linuxstb_And do you know why using the scrollwheel doesn't turn the backlight on?
23:25:40preglowlinuxstb_: it should...
23:25:47preglowif not, update the build again
23:26:37pinkutankis linusn around
23:26:43linuxstb_Also, I think we should stop Rockbox wrapping round when scrolling in a list. It's too easy to overshoot when using the wheel. Is there already an option for that?
23:26:47webguest24pinkutank: I think you do
23:26:56pinkutankI also wnat to know if usb charging works in rockbox
23:26:59webguest24But I'm not sure.
23:27:19amiconnlinuxstb: The wrap-around is a wanted feature. It doesn't wrap with repeat
23:27:26preglowlinuxstb_: scrollwheel wasn't turned on because i don't go through button_tick for the wheel, i insert events directly
23:27:31amiconnMaybe it's a problem with the wheel
23:27:37preglowlinuxstb_: but the newest build should turn on backlight
23:27:49 Quit xmixahlx ("blah blah blah")
23:27:59preglowamiconn: it will be because the wheel wont work by using button repeat
23:28:05preglowamiconn: that'd be far too slow and defeat the entire purpose
23:28:16 Nick RotAtoR|Away is now known as RotAtoR (
23:28:16DBUGEnqueued KICK RotAtoR
23:28:17amiconnYes I know
23:28:34amiconnDid I tell that I don't like the wheel? ;)
23:28:59 Join Jungti1234 [0] (n=jungti12@
23:30:15 Join qwisp11 [0] (
23:30:18preglowlinuxstb_: and i don't quite know why the next button problem is there, it shouldn't be
23:30:22preglowbut i'll figure it out
23:30:28preglowi can live with the wraparound, btw
23:31:13amiconnYou could define that the wheel always sets BUTTON_REPEAT in its events as a workaround
23:31:48amiconnOr perhaps, only do that if the interval between 2 events is short
23:31:58preglowwe'll figure it out
23:32:00preglowbut now i gotta run
23:32:09preglowsee you guys tomorrow
23:33:12 Quit JazzBone ("Leaving")
23:33:18qwisp11does rockbox play movies, if so what extension does it need?
23:33:54pinkutankI also wnat to know if usb charging works in rockbox with h300, does anyone know
23:34:07 Quit muesli_- (Read error: 110 (Connection timed out))
23:34:16qwisp11I dont think it charges
23:34:27qwisp11but it connects damn quickly
23:35:26pinkutankbecause my regualr charging port is blown
23:36:39webguest24qwisp11: no movies yet
23:37:42amiconnRockbox does play movies, but only on archos so far
23:37:54CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:37:54*amiconn hates imprecise questions
23:41:50 Quit webguest24 ("CGI:IRC (Ping timeout)")
23:42:02mirak_sprintf(filename, "%svolhdr.m4v", filepath); I need to convert that function. is there something simpler than using strcat ?
23:42:38mirak_c suck
23:44:13linuxstb_Just use snprintf(filename,sizeof(filename),.....
23:44:26linuxstb_But why do you need to do that anyway? What is that filename used for?
23:45:41mirak_it's to generate the images
23:45:46mirak_not very usefull
23:45:51mirak_but well
23:46:29mirak_ok sorry I mistaken what was doing the function
23:46:29 Quit qwisp11 (Read error: 104 (Connection reset by peer))
23:47:59idanmhow do I get the uisimulator from csv? co rockbox didn't gave it.
23:48:14Mmmmcvs co uisimulator
23:48:31 Join Kyl3 [0] (
23:48:46idanmwhat's exactly is csv co?
23:48:46 Join San [0] (
23:49:58mirak_man cvs ?
23:50:28Mmmmcvs not csv! :) Concurrent Version System apparently!
23:51:19Kyl3I'm trying to apply a patch that modifies several files. I but the patch in the /build dir where I compiled it, then after I compiled it I did patch −−binary -p1 < remote_type.patch and then it says "file to patch" so I type in the path (ex /apps/gui/gwps-common.c) and it says it doesnt exsist even though its there, what should I do?
23:53:41Mmmmwhat is the path in the patch?
23:54:34Kyl3the patch is in the root directory I compile in (C:\Rockbox\home\guest\build)
23:54:46RotAtoRif it's in the same directory you made it, I think you should use -p0
23:54:46 Join webguest63 [0] (
23:55:10 Quit saa[b_r]ider ()
23:55:11Kyl3I'll try that
23:55:15RotAtoR-p tells patch how many directory levels to strip from the path
23:56:09Kyl3well -p0 didnt work.
23:56:48Mmmmopen your patch in a text the first few lines what does the path say?
23:56:50 Quit webguest63 (Client Quit)
23:56:58Kyl3When I patched before (patches that modify one file) I used to put the patch in the directory if the file to be modified
23:57:04RotAtoRthen I'd try different -p values, its seems that would be the most likely issue if you're getting "file not found" errors
23:57:12Kyl3diff -ru rockbox-devel/apps/gui/gwps-common.c rockbox/apps/gui/gwps-common.c
23:57:14Kyl3−−- rockbox-devel/apps/gui/gwps-common.c2005-12-13 20:15:21.996000000 +0900
23:57:14Kyl3+++ rockbox/apps/gui/gwps-common.c2005-12-13 18:42:37.385472000 +0900
23:57:17Kyl3then more files
23:57:28Kyl3but I dont wanna flood the rool
23:57:35Mmmmlooks like -p1 should work then!
23:57:43Kyl3diff -ru rockbox-devel/apps/gui/gwps.c rockbox/apps/gui/gwps.c
23:57:45Kyl3−−- rockbox-devel/apps/gui/gwps.c2005-12-09 10:11:14.000000000 +0900
23:57:45Kyl3+++ rockbox/apps/gui/gwps.c2005-12-13 18:45:37.324211200 +0900
23:58:20Kyl3and when it says "file to patch" I put: /apps/gui/gwps.c
23:58:30Kyl3and it says not found

Previous day | Next day