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

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

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

#rockbox log for 2009-01-12

00:00:50 Join akur [0] (
00:00:54 Part akur
00:01:07amiconnHmm, but when not buffering, it's also kept nearly full...
00:01:58 Quit bertrik (Remote closed the connection)
00:02:06amiconnVarying between 11 and 13 while playing a -c2000 ape file
00:02:30kugelZagor: ha, the hissing noise is gone here too with a full/clean rebuild
00:02:40Zagorkugel: excellent!
00:03:29*amiconn thinks that this proves that building on cygwin can actually be helpful ;)
00:03:31kugelhmm...I wonder why the clip creates corrupted config.cfgs
00:03:52Zagorkugel: it does? how are they corrupted?
00:04:07Zagoramiconn: haha
00:04:08kugelno idea
00:04:19kugelI can't open them in a text editor
00:04:26kugeland they prevent booting
00:04:42amiconnOpen them in a hex editor and check..
00:05:22 Join MOHDMACH [0] (
00:05:53 Join chronon [0] (
00:06:25MOHDMACHI finally figured out a .wps file with a volume tag and a conditional working the way I want it to.
00:06:39MOHDMACHExcept the text does not display at all.
00:07:40Bagderkugel: all zeroes...
00:08:08MOHDMACHI tried moving the album, aritist, track info to a different location on the .wps file but that did not work.
00:08:09kugelreally? weird
00:08:42kugelI don't have a hex editor handy, can you recommend one for linux?
00:09:17kugeloh I didn't know about that one
00:09:28Bagderhexdump -C is what I usually use
00:10:18kugelit's not all zeroes here
00:10:20MOHDMACHThe viewport is set in size and position which I have checked multiple times and I have the font colored black and the WP screen is white so the font color is not the problem.
00:10:38kugelbut it looks weird, for sure
00:11:14PaulJamMOHDMACH: could you post your wps code somewhere (pastebin or similar).
00:11:18kugelah wait, it's all zeroes
00:11:29Bagderright, hexdump hides repeated lines
00:11:50 Join barf_ [0] (
00:11:58kugelyea, plus: I didn't notice the first column being the offset ;)
00:13:14 Quit advlaptop2019 ("Leaving")
00:13:54 Quit chronon (Remote closed the connection)
00:14:09 Join chronon [0] (
00:15:54MOHDMACHI should not have to remove the viewport and the tags associated with the volume but I have and now the text is back.
00:18:08 Quit Virindi ("Leaving")
00:18:36 Quit solexx (Read error: 110 (Connection timed out))
00:20:18 Quit agaffney (Read error: 104 (Connection reset by peer))
00:20:32 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
00:20:39 Quit robin0800 (Remote closed the connection)
00:20:40PaulJamMOHDMACH: which text is not displayed?
00:20:42kugelZagor: ha, playback majorly improved I guess
00:20:44 Join robin0800 [0] (
00:21:06MOHDMACHTrack, Artist, Album information.
00:21:22Zagorkugel: you mean music is better than hiss? ;)
00:21:49MOHDMACHAnywhere I could post a screendump (besides photobucket).
00:21:56kugelmp3 seems still problematic, seems to crash the clip after some seconds
00:22:59Zagor128 runs for several minutes for me. 320 fails pretty quick.
00:23:22kugelmy files are somewhere between that
00:23:31kugel~200 vbr
00:24:11PaulJamMOHDMACH: you could try
00:25:49 Join BigBambi [0] (n=alex@
00:27:18TheSkunkManI'm looking for information for a possible port, so where could I find endian the player uses
00:27:27 Quit Llorean (Read error: 104 (Connection reset by peer))
00:27:40TheSkunkManwouldit be located with the cpu documentation
00:27:48BagderTheSkunkMan: I guess reverse engineering will tell
00:27:54 Join Llorean [0] (
00:27:55Bagderthe cpu can be either way
00:28:13Bagderthe MIPS one I presume
00:29:48 Quit barf_ ()
00:30:08 Quit robin0800 ("Leaving")
00:30:15 Join robin0800 [0] (
00:30:46 Quit gevaerts (Nick collision from services.)
00:30:58 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
00:31:01PaulJamMOHDMACH: from the CustomWPS wiki page: If any viewports are defined, you cannot display text in the default viewport (use it for images only)
00:32:23MOHDMACHThats exactly what is supposed to happen(It is what is happening at least in the volume bar image case) but it is not supposed to cover the whole screen.
00:32:50MOHDMACHIt is only supposed cover the progress bar while the volume is being adjusted.
00:33:42PaulJamthe point is, that as soon as you use one viewport in the WPS, then you have to set up a viewport for _all_ the text elements in the WPS.
00:35:17MOHDMACHWhat is the default viewport?
00:35:49PaulJamthe whole screen (minus the statusbar height)
00:36:54MOHDMACHSo if I set up only one viewport I have to setup one for all the text also.
00:37:51MOHDMACHIt's back to the text editor for me then. Thanks for the help.
00:40:47PaulJambtw, there might be another problem with your WPS: the volume images, the progress bar and the conditional viewport are overlapping. (they all have 22x174 as origin)
00:43:31MOHDMACHThat was on purpose. While the volume is being adjusted the volume bar covers the progress bar just like in the original ipod firmware or similar at least.
00:44:47MOHDMACHI find it strange that when the viewport is not present the important text is not displayed.
00:44:52 Quit gregorovius ()
00:45:54MOHDMACHThe size of it as you seen is just supposed to cover the progress bar not the whole screen.
00:47:00MOHDMACHI guess from what you said I have to set up a viewport for all that text but I just find it strange is all.
00:47:10 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:47:32PaulJamMOHDMACH: i would recommend to put the volume stuff into its own conditional viewport so basically the %mv line would be: '%?mv.5<%Vda|%Vdv>' (with the conditional viewport with the descriptor "a" containing the volume bar)
00:50:51MOHDMACHSo in my case since the volume bar is separate bitmap images I would do the following %?mv.5<%Vda|%xda|%xdb|%xdc....|%Vdv>
00:51:54 Quit Zagor ("Clint excited")
00:52:26PaulJamno, load the images inside the conditional viewport (with the coordinates 0|0 ) and also the line '%?pv<%xdm|%xdn|%xdo|%xdp|%xdq|%xdr|%xds|%xdt|%xdu|%xdv|%xdw>'
00:53:26***Saving seen data "./dancer.seen"
00:53:54 Quit ender` (" I find television very educating. Every time somebody turns on the set, I go into the other room and read a book. -- Grouc")
00:55:10 Quit parafin (Read error: 60 (Operation timed out))
00:55:33 Quit blkhawk (Read error: 104 (Connection reset by peer))
00:55:37 Join blkhawk [0] (
00:55:39MOHDMACHHow would I load the images inside the conditional viewport? Would it be like so: say if I have an image xl then the conditional viewport
00:55:56MOHDMACHNevermind that comment is messed up.
00:57:11 Quit bmbl ("Woah!")
00:59:42 Join parafin [0] (
01:00:11MOHDMACHWould it be like so: say if I have an image %xl|a|volume1.bmp|22|174|width|height| so I would create a viewport: %Vl|a|0|0|width|height|0|-|-|
01:01:32PaulJamthe %xl tag doesn't have width or height parameters
01:02:35MOHDMACHSorry I forgot :)
01:03:19PaulJamthe code for the conditional volume viewport should look like that:
01:05:24MOHDMACHSo all the images below a %Vl tag go into that viewport
01:06:37PaulJamyes all lines following a viewport definition (%V or %Vl) until there is anoter viewport defined (or the file ends)
01:08:36 Join Oliver__ [0] (
01:09:25MOHDMACHYou should write that in the CustomWPS cause that is useful right there.
01:09:32 Quit robin0800 (Remote closed the connection)
01:10:56Oliver__hello! i have question about rockbox on my iriver h320: when i browse through a playlist (e.g. holding down the "down" button its incredibly slow and after a few secs playback is interrupted until i stop browsing. i dont have this issue on my iriverh140. any idea?
01:11:24PaulJamwell, the wiki says: All lines before the first viewport declaration are drawn in the default (full-screen) viewport, and lines after a viewport declaration are drawn within that viewport.
01:12:10MOHDMACHYeah that's why I said nevermind.
01:13:32Oliver__does anyone have an idea?
01:13:45PaulJamdo you use the same setting on you H300 and H100?
01:13:57Oliver__i tzhink so, yes
01:14:37Oliver__what kinda setting could have influence on what i described?
01:14:39MOHDMACHBut if I use the coordinates 0,0 for the volume viewport it would cover the play mode status bitmaps instead of covering the progress bar. Or were you just talking about the viewports that need to made for the text?
01:14:52Oliver__or is ist becaUse of the
01:15:06Oliver__or is it because of the color grqaphnics?
01:15:11PaulJamMOHDMACH: the coordinates are relative to the viewport
01:18:55 Join rockboxn00b [0] (n=4765a2f1@gateway/web/cgi-irc/
01:19:52PaulJamwell, on my h300 scrolling in the playlist isn't really fast (~10 entrys / second) but i don't experience audio skipping.
01:20:20rockboxn00bdoes rockbox support only 5th gen video ipods?
01:20:29 Join webguest74 [0] (n=45b765c5@gateway/web/cgi-irc/
01:20:35 Quit webguest74 (Client Quit)
01:20:44MOHDMACHI have an ipod 5g and it works fine on there.
01:20:47PaulJamrockboxn00b: the list of supported targets is on the front page
01:20:59rockboxn00bno hope for playing FLAC on 3rd gen then?
01:21:24rockboxn00bI saw the list but was asking if if it were too buggy
01:21:25linuxstbAre you talking about the 3rd gen Nano?
01:21:25Oliver__its more slow on mine... ill check it, PaulJam
01:22:35Oliver__i have about 5 entries/sec. with skipping
01:22:42Oliver__strange, isnt it?
01:22:54 Quit rockboxn00b (Client Quit)
01:23:00PaulJamOliver__: what file formats are you plaing?
01:23:12Oliver__in this case those are oggs
01:23:21Oliver__have the same with flac
01:24:25 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
01:24:32MOHDMACHI'll play around with it and see if if corrects the problem. Thank you very much for the help.
01:24:58 Part MOHDMACH
01:25:42 Quit DerDome ("Leaving.")
01:25:57Oliver__PaulJam: same with mp3, although the playback breaks seem 2b smaller
01:26:51PaulJamhmm, i don't really have an idea. maybe you could check if this happens with default settings too.
01:29:47Oliver__yes. wait a moment
01:31:52Oliver__ok, it works with default settings... hm
01:33:05PaulJamthen enable your usual setting one after the other and check each time in order to find the setting that causes it.
01:37:01Oliver__got it
01:37:04 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
01:37:21Oliver__it occurs as soon i change the font, PaulJam
01:41:39PaulJamok, when using a small font i get the skipping too.
01:41:52 Quit moos ("Rockbox rules the DAP world")
01:43:04Oliver__but the default font is a small one (12)
01:43:23Oliver__i just chose another 12pt font
01:43:57PaulJamwhich one?
01:44:55Oliver__terminus. interesting: its working only with the 2 helvetica-12pt ones
01:47:27PaulJamindeed, with teminus the perfomance is very bad. could you file a bugreport on the tracker?
01:48:29Oliver__i will
01:51:32 Quit PaulJam (".")
01:53:24 Quit Oliver__ ()
02:03:46 Join dfkt_dt [0] (i=dfkt@unaffiliated/dfkt)
02:05:02 Join soap_ [50] (n=soap@rockbox/staff/soap)
02:05:11 Join toffe82_ [0] (
02:06:07 Quit blkhawk (
02:06:07 Quit miepchen^schlaf (
02:06:07 Quit gromit` (
02:06:07 Quit toffe82 (
02:06:07 Quit fdinel (
02:06:07 Quit dfkt (
02:06:07 Quit kugel (
02:06:07 Quit HBK- (
02:06:07 Quit tvelocity (
02:06:07 Quit timc`` (
02:06:07 Quit Rob2223 (
02:06:07 Quit TheSkunkMan (
02:06:08 Quit linuxstb (
02:06:08 Quit merbzt (
02:06:08 Quit feisar (
02:06:08 Quit idshark (
02:06:08 Quit GodEater (
02:06:09 Nick toffe82_ is now known as toffe82 (
02:06:20 Quit Thundercloud (Remote closed the connection)
02:06:22 Quit toffe82 (Killed by (Nick collision))
02:06:23NJoinblkhawk [0] (
02:06:23NJoinmiepchen^schlaf [0] (
02:06:23NJoingromit` [0] (
02:06:23NJointoffe82 [0] (n=chatzill@
02:06:23NJoinfdinel [0] (
02:06:23NJoindfkt [0] (i=dfkt@unaffiliated/dfkt)
02:06:23NJoinHBK- [0] (
02:06:23NJointvelocity [0] (
02:06:23NJointimc`` [0] (n=aoeu@
02:06:23NJoinRob2223 [0] (
02:06:23NJoinTheSkunkMan [0] (
02:06:23NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
02:06:23NJoinGodEater [0] (n=ge@rockbox/staff/GodEater)
02:06:23NJoinidshark [0] (
02:06:23NJoinmerbzt [0] (n=benlar@
02:06:23NJoinfeisar [0] (
02:07:00 Join toffe82_ [0] (n=chatzill@
02:07:34 Nick fxb is now known as fxb__ (
02:07:53 Nick soap_ is now known as soap (n=soap@rockbox/staff/soap)
02:08:38 Quit dfkt (Nick collision from services.)
02:08:42 Nick dfkt_dt is now known as dfkt (i=dfkt@unaffiliated/dfkt)
02:20:35 Quit toffe82 (Connection timed out)
02:34:25 Join kkurbjun [0] (
02:51:08 Quit t0mas (Read error: 110 (Connection timed out))
02:53:29***Saving seen data "./dancer.seen"
02:59:34 Join sbhsu [0] (
03:12:26 Quit culture (Read error: 110 (Connection timed out))
03:16:35 Quit timc`` (Connection timed out)
03:21:25 Quit axionix_ (Read error: 60 (Operation timed out))
03:23:45JdGordonumm... whaa?~ someone found the datasheet for the chipset in the iriver e100 :)
03:24:52 Join timc`` [0] (n=aoeu@
03:25:19 Join hd [0] (
03:25:52jhMikeSwell, then spread the wealth :)
03:28:07JdGordonintegrated jped coproccessor!
03:36:31 Quit HellDragon (Read error: 110 (Connection timed out))
03:38:42 Join webguest76 [0] (n=6394f511@gateway/web/cgi-irc/
03:39:56webguest76i need some help with my rockbox
03:41:59 Quit webguest76 (Client Quit)
03:47:58jhMikeSjped = Joint Feet?
03:48:48JdGordonbah.. jpeg
03:49:24toffe82_jhMikeS:are you beginning to work on the sa9200 ;)
03:49:31Unhelpful*thanks*, i thought we might be developing an exciting new feature! :P
03:49:38 Join japc [0] (
03:50:25jhMikeStoffe82_: I did get firmware on it so, I guess so. It was just bugging me sitting there. The focus is still the beast at this point.
03:50:41JdGordonis a hardware idct decoder exciting?
03:50:52UnhelpfulJdGordon: meh! ;)
03:50:57jhMikeSJdGordon: It doesn't look like it's much of a reference manual though.
03:51:37toffe82_I found a X60 and try to get but I think that the bid are going to be high
03:51:44JdGordon;filename=MIPS-4KEcDataSheet.pdf us the mips core which might be more useful?
03:51:50toffe82_it is already 80$
03:51:52Unhelpfulshould i assume that rockbox is running entirely in priveleged mode? i wanted to try something that might be useful for benchmarking ARM code... although only, to my knowledge, on beast.
03:52:06jhMikeSUnhelpful: Yes
03:52:47Unhelpfulsweet. the freescale in beast is arm11, yes? i want to play around w/ some example ASM i found for arm11 performance counter registers :D
03:53:17jhMikeSUnhelpful: Yes.
03:53:42JdGordonso, the first 15 bytes of the iriver firmware is 11 22 33 44 55 66 ... ff any crazy ides on how to decrypt the firmware from that? :p
03:54:57Unhelpfulwe should be able to make some inline ASM to set up / capture the counter values, then... they can count cycles, cycles/64, and a pile of various interesting performance things, stalls waiting for memory and such.
03:55:35jhMikeSARM 1136jf-s
03:57:29*jhMikeS get remindsd he needs a MIPS target
03:58:27Unhelpfulheh... mine are both arm, e200 and beast
03:58:27 Quit MarcGuay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")">
03:59:22 Quit hd (Client Quit)
03:59:44Unhelpfulwould timing data from beast be of any use for tuning code for other ARM CPUs? or do they tend to perform too differently?\
04:00:37jhMikeSI think probably not so useful (given the branch prediction and other features)
04:04:50 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
04:05:31 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
04:12:07*JdGordon wonders if our iriver fwpatcher can decrypt the e100 firmware file
04:13:20LloreanI thought someone said it could.
04:13:32LloreanOr rather, I'm not sure if it's fwpatcher
04:13:37LloreanI think it was the tool for the iFP port.
04:13:40TheSkunkManI am new to this sort of stuff, but typically what would it take to decrypt an iriver firmware file?
04:13:58TheSkunkManIf you could set me in the right direction
04:14:20JdGordonhey TheSkunkMan
04:14:43_Auron_time, patience, and loss of sanity
04:14:50jhMikeStoffe82_: If we want the sa9200 to really be viable, Rockbox UMS really must get into shape and work in the bootloader.
04:15:32 Join HellDragon [0] (
04:17:12JdGordonTheSkunkMan: if we are really lucky, iriver didnt change the encryption, just the key so our current tools might be able to decrypt it
04:18:51toffe82_jhMikeS: I didn't notice yesterday when I was updating my sa9200, but booting with usb attached, rockbox failed to launch. Is this what you are talking about ?
04:20:05toffe82_If you boot rockbox and attached the usb after, there is no problem
04:21:51jhMikeStoffe82_: I think it's a problem with the USB thread. The beast will soft deadlock too.
04:22:47toffe82_there is a problem with the usb on the hdd1630 too when you connect it , it reboot until you disconnect
04:23:04jhMikeSThere are problems when using the stack and all threads don't ack.
04:25:03JdGordonLlorean: do you know which tool the ifp uses?
04:25:21LloreanJdGordon: Not a clue.
04:26:01LloreanI almost think it's just scramble
04:28:24*jhMikeS thinks he'll just simplify the usb thread code (it's quite messy)
04:31:35 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:33:52unstableIs there a page on the wiki showing what works and what doesn't on e280 v2? I know there is a lot of work to go, but I wanted to know if I'm able to dual boot with rockbox, and if rockbox plays music, what issues there are, what features are implemented etc. Does a wiki page summarizing this stuff exist?
04:36:01amdgoonunstable: ther, I'm not sure if it's fwpatcher
04:36:26amdgoonunstable: ther, I'm not sure if it's fwpatcher
04:36:55amdgoonsorry c'n p problem here
04:37:17unstableSo I won't be able to put it on my e280 v2 no matter what?
04:37:44amdgoonit is not ready from what I read today.
04:38:17amdgoonyou can check that link to see what is working.
04:39:02amdgoonsome buttons work, some don't
04:44:11 Join Barahir_ [0] (
04:46:16 Quit amdgoon ("Eat right, exercise regularly, die anyway.")
04:46:38 Quit toffe82_ (Read error: 104 (Connection reset by peer))
04:50:24jhMikeSI think the ata exclusive stuff needs to be reversed so ata is made exclusive after all acks and exclusive is release before acking disconnect. it's a race condition otherwise.
04:51:54 Join Makuseru [0] (
04:52:31jhMikeSI'd better just quiet down until I've really examined things though (no need to sound foolish by posting every thought) :P
04:53:02Unhelpfulwhat would i do if you took my job? ;)
04:53:33***Saving seen data "./dancer.seen"
04:54:01jhMikeSloaf around and drink alot [more] beer?
04:54:51Unhelpfulyou must mean somebody else, i don't like beer. heck, probably anybody else.
04:55:23Unhelpfulbut, hey sometimes you think you're sounding foolish when somebody else decides it was a *great* idea. :)
04:55:49 Quit Seed ("cu, Andre")
04:57:18 Join blkhawk- [0] (
04:59:05 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
05:01:31 Quit Barahir (Read error: 113 (No route to host))
05:02:14 Quit aurix_lexico ("Leaving.")
05:02:29 Quit Makuseru (Connection reset by peer)
05:07:21 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
05:08:54JdGordonifp_decode cant decode the e100 .hex :(
05:11:50 Join Makuseru [0] (
05:13:27 Quit blkhawk (Read error: 110 (Connection timed out))
05:14:16 Nick blkhawk- is now known as blkhawk (
05:24:28 Quit Acksaw (Read error: 104 (Connection reset by peer))
05:24:45 Join Acksaw [0] (
05:36:27 Quit LambdaCalculus37 ("Ka-chunka")
05:38:07 Quit Makuseru (Read error: 104 (Connection reset by peer))
05:39:26 Quit saratoga ("CGI:IRC (EOF)")
05:40:58 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
05:45:22 Join itcheg [0] (i=62db4767@gateway/web/ajax/
05:45:53TheSkunkManafter running it says to "append /usr/local/mipsel-elf/bin to PATH" , could someone explain this to me
05:59:50 Quit TheSkunkMan ("Leaving")
06:00:32 Quit amiconn (Nick collision from services.)
06:00:33 Join pixelma_ [0] (n=pixelma@rockbox/staff/pixelma)
06:00:34 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
06:02:54 Quit pixelma (Read error: 60 (Operation timed out))
06:05:38 Join nibbler__ [0] (
06:05:51 Join undertakingyou [0] (
06:07:34 Quit nibbler_ (Read error: 60 (Operation timed out))
06:08:59undertakingyouIs there any planned support for Creative's line of Zen players?
06:09:15scorchewe dont "plan support" of any device
06:09:50undertakingyouHow would I encourage support of these devices?
06:09:55scorcheports happen when people come forward with the capability, time, the device itself, and the motivation to make a port happen
06:10:46undertakingyouWell, I have the device but no other knowledge or experience. Is there a way that I could help?
06:11:30Unhelpfulnot right now, now... but you could learn :)
06:12:08scorcheyou could find people who are like-minded as you, but who might have skills, however the only sure way for it to happen is if you yourself learn how and begin ;0
06:12:29undertakingyouWhat language is rockbox written in?
06:13:08scorchemostly C, with some ASM bits here and there
06:13:39undertakingyouThis may not be the forum for this question, but how do you just test a device and/or capture the devices needs to change the code?
06:14:06Unhelpfulthere's nothing to test until somebody's writing for it
06:14:43 Quit BlakeJohnson86 (Remote closed the connection)
06:21:49 Join BlakeJohnson86 [0] (
06:27:05 Quit BHSPitLappy (Read error: 110 (Connection timed out))
06:29:37 Quit scorche (Read error: 131 (Connection reset by peer))
06:30:44 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
06:38:46 Join FlynDice [0] (
06:41:06 Quit scorche (Nick collision from services.)
06:41:29undertakingyouis there any list of Real Names to irc nicks in the wiki?
06:41:59 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
06:42:07unstableDo you work for Apple Inc's Law Deptartment, patent division.. why do you need real name?
06:42:27 Quit itcheg (" ajax IRC Client")
06:42:38krazykitundertakingyou, there is. it's linked from the
06:42:49undertakingyouunstable: It seems that in the Wiki real names are all that are used. I have found someone working on a port for my device, I wanted to chat with them.
06:42:53undertakingyoukrazykit: thank you.
06:44:14 Quit bobsalad (Connection timed out)
06:46:17 Nick HBK- is now known as HBK (
06:48:00 Join bobsalad [0] (
06:49:40 Quit jhulst (Read error: 104 (Connection reset by peer))
06:53:35***Saving seen data "./dancer.seen"
06:54:32 Join nuonguy [0] (
06:56:32 Quit scorche (Nick collision from services.)
06:57:25 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
07:01:48 Quit scorche (Nick collision from services.)
07:02:41 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
07:31:51lucentundertakingyou: WikiName should be linked from irc name, but not all of them are on the ports
07:35:31FlynDiceWould someone be able to tell me where my button-e200v2.c fits into the big picture. ie what actually calls or uses it or steer me in the right direction to investigate?
07:37:05JdGordonbutton_get (in firmware/dirvers/button.c) calls button_Read_target() (in that button-e200v2.c file)
07:39:57FlynDiceIs there a super secret way I can find that out on my own next time?
07:42:21JdGordonFlynDice: you havnt done much C coding/debugging before?
07:43:02FlynDiceer no... 20 years ago I played a bit
07:44:23FlynDiceI like puzzles though ;)
07:44:47JdGordonpuzzless are easier sometimes :p
07:46:44 Join JdGordon|uni [0] (i=7bf38c1f@rockbox/developer/JdGordon)
07:49:01 Nick JdGordon|uni is now known as JdGordon_ (i=7bf38c1f@rockbox/developer/JdGordon)
07:55:05FlynDiceI can grep but do I need to grep the firmware directory for every function in my button driver file?
07:56:37JdGordonjust btton_read_device....
07:56:49JdGordonit should only be caled in 1 file
07:57:45kadobanFlynDice: if you need to know what calls every function, then maybe yes. (unless of course JdGordon or someone who knows what they're doing is around :P)
07:59:07FlynDiceI was hoping there was more efficient new fangled technology I could learn about....
07:59:45kadobanif you use an IDE, often they'll have something prettier, but i imagine that it generally boils down to the same thing
08:00:26 Quit JdGordon (Remote closed the connection)
08:15:23 Quit GodEater_ (" ajax IRC Client")
08:16:50 Join GodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
08:17:49JdGordon_tv out on the ipod video!
08:22:45JdGordon_patch just uploaded... basic bmp only
08:23:47*GodEater_ 's flabber is gasted
08:26:31 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
08:26:35GodEater_does TV out go via the broadcom ?
08:26:37 Join Rob2222 [0] (
08:31:02 Join Zagor [0] (n=bjorn@
08:34:08 Quit BigBambi (Read error: 113 (No route to host))
08:40:28 Join ender` [0] (
08:41:15 Join jayde_drag0n [0] (
08:41:45jayde_drag0nhey i had a couple of questions that aren't answered in the FAQ if anyone has a minute to assist me
08:43:18 Join axionix [0] (
08:44:13jayde_drag0nthe manual says rockbox has some sort of issue of seeing my music if i upload my stuff thru itunes.. i'm not entirely sure what this means.. but after a rockbox installation will i still be able to scroll thru my music via genre.. and will it also see the DISC number.. this is very important for me as i listen to a LOT of audiobooks and some are organized that way.. otherwise they will play out of order and render the book useless
08:45:07 Quit Rob2223 (Read error: 110 (Connection timed out))
08:45:58lucentjayde_drag0n: not sure about that, I don't think any OF loaded music will be visible
08:46:22jayde_drag0nwhat does OF mean?
08:46:28lucentOF = original firmware
08:46:44lucentor alternately OEM Firmware ;)
08:46:57jayde_drag0nwell i don't mind doin a restore before install.. BUT afterwards can i still sync via itunes?
08:47:08jayde_drag0nor do i need to use a different program?
08:47:25lucentI can't answer conclusively, it's been years since I tried rockbox on an iPod
08:47:58scorchejayde_drag0n: rockbox doesnt have an issue with itunes-loaded files...itunes just places the files in odd named folders and changes their long as you have your tags intact, you can still use the database
08:48:08lucentthe way it *usually* works is that you install the rockbox bootloader code to your player device, and this allows an option to choose the original firmware at boot time
08:48:36scorcheand by "option", he means "you can hold down a button for booting the OF"
08:48:43lucentso if it works like that, then you should be able to boot the OF and sync with whatever software you used before
08:48:50lucentyup, thanks scorche
08:49:03jayde_drag0nokay.. i can't find enough screenshots to answer my other questions.. can i still scroll thru stuff via genre like i currently do .. or will it more look like a winamp screen?
08:49:30lucentthere's a file-directory based browse, and a database browse
08:49:38lucentthe database browse is initialized from tag information
08:49:41scorcheit can look however you wish it can browse by genre if your files have the genre tag field properly filled out
08:50:11jayde_drag0nlol oh yes.. i'm obsessive about orgianization on my files ESPECIALLY since i use it for audiobooks
08:50:29jayde_drag0neverything is tagged.. disc numbers filled out.. track numbers
08:50:52scorcheif you use itunes, browsing by the actual disk structure will be useless
08:51:05jayde_drag0nbut i know from experience in linux.. that there was no way to organize dic numbe first.. then track number.. and it rendered my audiobooks useless
08:51:27jayde_drag0ni had to use itunes for it to do it right
08:51:40scorcheyou dont have to here
08:51:45jayde_drag0notherwise my screen on my ipod went like this track 1 track 1 track 1 etc
08:51:48lucentjayde_drag0n: at this point I think we've provided enough information for you to do experiments on your time and then go further
08:51:56*lucent :)
08:52:09amiconnJdGordon_: Status icons on Player don't work at all anymore :(
08:52:11*lucent points to the serious nature of #rockbox
08:52:32scorchedatabase can provide you with a large amount of customization that it sounds like you will want...see
08:53:08 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:53:20 Quit FlynDice ("Konversation terminated!")
08:53:37***Saving seen data "./dancer.seen"
08:54:00jayde_drag0nlucent: i wan't being unserious.. i had questions.. this was the technical fourm.. and there was no rule stating admins only or no newbie questions allowed else i would have kindly stayed away i apologize if i have offended you
08:54:03 Join petur [50] (n=petur@rockbox/developer/petur)
08:54:10*Bagderr spots TV-out for ipod video
08:54:13 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:55:04Zagoryeah, that's a nice little present
08:55:22jayde_drag0nscorche: thank you for answering my questions i appreciate it
08:59:52 Quit nibbler__ (Read error: 113 (No route to host))
09:02:47 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
09:04:35 Nick Barahir_ is now known as Barahir (
09:04:38XavierGrBagder: And it's not the only target with that goodie, also the beast comes with it. It would be great to see it implemented at some point.
09:05:16B4gderof course, this is of course still very target-specific
09:06:06XavierGrBagder: Let's hope that this feature doesn't use the broadcom chip on ipod video
09:06:42B4gderit doesn't, at least not what it looks like
09:09:17amiconnjhMikeS: The usb disconnect-reconnect cycle bug seems to be gone, and my beast charges from usb, but the charging display seems a little weird
09:09:28 Quit jayde_drag0n ("Leaving.")
09:10:14amiconnI started charging from low battery (so low that I needed to plug the charger to trick the builtin loader into starting), but after just a few minutes connected to usb in rockbox, the battery icon shows nearly full
09:11:01 Quit thegeek (Read error: 54 (Connection reset by peer))
09:11:30 Join thegeek [0] (
09:14:48JdGordon_does anyone have any ideas/pointers on trying to decrypt the iriver e100/lplayer firmwares?
09:15:14JdGordon_ifp_decrpt couldnt do it and neither could fwpatcher (although i may not have fixed it properly to try)
09:17:01amiconnJdGordon_: Do you have an idea regarding the Player status icons? I'm quite sure it has to do with the statusbar stuff
09:17:46JdGordon_umm... I probably had an idea yesterday which is why I asked... I'll have a look in a while, in windows atm
09:24:32 Quit JdGordon_ (" ajax IRC Client")
09:24:56 Join JdGordon|uni [0] (i=7bf38c1f@rockbox/developer/JdGordon)
09:25:12 Nick JdGordon|uni is now known as JdGordon (i=7bf38c1f@rockbox/developer/JdGordon)
09:29:38 Join t0mas [0] (n=tomas@rockbox/developer/t0mas)
09:32:55lucentI'm frustrated with the issue I have in the Sansa AMS Fuze code for SD handling
09:33:08lucentwith lots of delays and stuff, external card insertion works great
09:33:14lucentwithout delays, there's a panic
09:33:24lucentbut the delay I put in is really huge
09:34:14kadobanlucent: try smaller delays until it starts exploding?
09:34:24lucentI put delays in weird places
09:34:28lucentI will try that of course
09:35:00lucentkadoban: I'm concerned that our command to initialize the SD controller is not reliable to know when it is completed
09:35:01 Quit tarbo (Read error: 60 (Operation timed out))
09:35:09lucentand maybe it is returning too early before the hardware is ready
09:36:22 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:38:17 Quit chronon ()
09:38:23 Join tarbo [0] (n=me@unaffiliated/tarbo)
09:42:03 Join JdGordon_ [0] (n=jonno@rockbox/developer/JdGordon)
09:43:05 Join Thundercloud [0] (
09:43:33JdGordon_amiconn: can the statusbar be disabled on the player?
09:43:58JdGordon_it seems global_settings.statusbar is compiled but possibly not set for charcell
09:44:01B4gderit doesn't have one, it has fixed icons
09:44:30B4gderfixed as in set in the actual lcd
09:44:33 Quit JdGordon (" ajax IRC Client")
09:47:16 Nick JdGordon_ is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
09:47:36JdGordonyeah, I know.. but the setting it still there for some reason
09:50:07JdGordonyep, easy fix.... should I polute the source with some more #ifdef HAVE_LCD_BITMAPS and make sure this doesnt happen again? or just quick fix where its needed?
09:56:17 Join markun_ [50] (n=markun@rockbox/developer/markun)
09:56:53 Quit markun (Read error: 110 (Connection timed out))
09:57:46 Nick markun_ is now known as markun (n=markun@rockbox/developer/markun)
09:59:05JdGordonmarkun: you see the datasheet for the iriver e100 showd up?
10:00:49B4gderI found it interesting that one "non-chinese" player went MIPS based
10:01:36t0maswhy? is it so china specific?
10:01:46B4gderuntil now, yes
10:02:08B4gderthe Actions (ATJ) chips have own appeared in those chinese players
10:02:10t0masm68k is mips?
10:02:16B4gdert0mas: no
10:02:22B4gderm68k is m68k ;-)
10:02:30t0masmight have been some form of ;-)
10:03:25lucentI just got microSD insert to work on 8gb fuze again
10:03:44lucentI need an expert though to walk me through *why* what I did fixes this
10:03:53lucentwho should I ask?
10:06:07JdGordonamiconn: fixed
10:07:13JdGordonactually... can we kill the damn charcell already? :D
10:10:49ZagorJdGordon: let's go get our pitch forks
10:17:20 Join kugel [0] (n=chatzill@unaffiliated/kugel)
10:17:55kugellucent: try to ask me :)
10:18:12lucentkugel: okay so my patch I issued before with the 2 additional mci_delay()'s was incomplete
10:18:33lucentthere's a third mci_delay() I discovered tonight that needs to be present for stable microSD operation on 8gb fuze
10:18:43lucentit's just after the -7 return check thingy
10:19:06lucentI'm just guessing you know the part of the code that I am describing, that's not a very clean description
10:19:22kugelyea, I thinki I do
10:19:41lucentrevised post to pastebin, just one second so I can do that
10:22:05lucentkugel: between that patch and the read 8 bytes at a time work-around for bad pixel values, the "ghost" scroll wheel problem has not appeared
10:22:21lucentI am testing now to see which one affects the scroll wheel for my usage
10:22:31 Quit kachna (Read error: 60 (Operation timed out))
10:22:46kugellucent: I don't think #ifdef SANSA_FUZE is theh right way
10:22:54lucentI understand
10:23:13kugelis the delay noticeable? Don't forget that there's always a ~300ms delay at inserting the card
10:23:40lucentkugel: the mci_delay is imperceptable
10:24:01 Quit ajb (Read error: 104 (Connection reset by peer))
10:24:11kugellucent: actually, can you try putting 1 mci_delay into send_cmd?
10:24:22lucentoh, at the end?
10:24:31 Join ajb [0] (
10:24:40kugelI suspect it shouldn't hurt do have a sane delay between each command
10:25:22kugellucent: there's this "while(MCI_STATUS(drive) & MCI_CMD_ACTIVE);"
10:25:36kugeltry doing while(MCI_STATUS(drive) & MCI_CMD_ACTIVE) mci_delay();
10:25:41lucentokay I will try
10:27:36lucentkugel: on both? that particular code snippet appears twice in that routine
10:28:25kugelI thought of the second one, but you can also try both
10:29:20lucentokay I'll try the last one
10:30:30 Quit Thundercloud (Remote closed the connection)
10:30:31ajbUnhelpful: Sorry I disapeared yesterday. wrt FS #9758 won't it get confusing reporting a SVN id if someone has modified their tree beyond it?
10:33:12kugelif that works, we can think of reducing the delay
10:33:43 Join alexbobp_ [0] (n=alex@
10:33:48 Quit alexbobp (Read error: 60 (Operation timed out))
10:34:34lucentkugel: doesn't boot.
10:34:36 Nick alexbobp_ is now known as alexbobp (n=alex@
10:34:50kugeldoesn't boot?
10:35:01lucentkugel: nope. turn on player, sits at rockbox bootloader screen
10:35:10lucentdoes respond to button presses. does not boot any further
10:35:17kugeltry deleting config.cfg
10:35:31lucentI deleted .rockbox/* and unzipped fresh
10:35:35lucentit should work right?
10:36:10lucentmaybe my build is not good, I'll try a fresh rebuild
10:36:22kugelcan can you try pressing a button, to make it verbose?
10:36:30kugelwhile booting that is
10:37:19kugellucent: also, when changing lcd, button and sd stuff, the bootloader should be updated too
10:38:02lucentoh okay
10:38:45ajbhmmm, is there a lastbot or /last command to see when people where last active on this channel?
10:38:54lucentverbose bootloader is okay, it gets past and then loads the rockbox firmware I guess, but just sits at the loader screen (not rockbox background) as I described
10:39:08lucentkugel: adding mci_delay() as you suggest appears to break boot
10:39:26kugelapparently, yes
10:40:31kugelmaybe having it in the while isn't such a good idea
10:40:42kugelbut I do think that it should be in send_cmd
10:40:56 Quit alexbobp (Read error: 60 (Operation timed out))
10:41:39lucentkugel: I don't have mci_delay() included in the while(...) loop, it is after the loop
10:41:51lucentI will try now inside of the while(...) loop
10:43:55lucentkugel: same result
10:43:59lucentno boot
10:46:29 Quit nuonguy ("This computer has gone to sleep")
10:46:39lucentkugel: including mci_delay() inside the first while(...) loop is successful to boot
10:48:24kugelis the microsd good?
10:48:36lucentno, microsd insert -> panic
10:49:16kugellucent: try making it to a do { mci_delay } while.. loop
10:49:24kugelso that mci_delay can't be skipped
10:49:29lucentah yes, okay
10:49:58lucentkugel: no boot
10:53:39***Saving seen data "./dancer.seen"
10:56:50 Join sunny [0] (n=sunny_vi@
11:01:34 Join alexbobp [0] (n=alex@
11:02:01lucentkugel: why, does this work okay for your player?
11:02:31 Join culture [0] (
11:06:19kugellucent: well, I wouldn't say it works ok for my player, but for my microsd
11:06:33kugelapparently the majority of the microsds
11:06:47kugelbut if it works with that few delays it's just fine
11:07:33lucentkugel: sorry, I meant to ask if adding a mci_delay() in the while(...) loop we discussed is working for your player to boot okay
11:07:44lucenton my player it does not boot with that delay in the looop
11:08:08kugelno, not fine
11:09:07lucentI am buying more microSD cards and can do some more tests in 1 week
11:09:15 Quit alexbobp (Read error: 60 (Operation timed out))
11:09:30lucentI need 2GB size microSD cards for a different project, I can test with those cards
11:09:35kugelZagor: btw, I wonder why you didn't think of dependency problems/codec api version when I reported the problems first :/
11:09:54Zagorkugel: limited brain capacity? :)
11:10:29B4gderI bet it was insufficient amounts of coffee!
11:10:38kugelI didn't think of it at all, so my brain is more limited
11:11:57kugelmp3 needs 0% boost on ams it seems. it only boosts at buffer refilling, for a split second (refilling 370KB of audio buffer doesn't take long)
11:12:16kugelat least with the test files that work
11:12:40 Join alexbobp [0] (n=alex@
11:13:09 Quit kugel (Remote closed the connection)
11:13:11Zagorwe could even skip the buffering boost on flash targets
11:13:15 Join kugel_ [0] (n=chatzill@
11:13:26 Nick kugel_ is now known as kugel (n=chatzill@
11:13:58kugelI thought of it too, but I wasn't sure if it applies for all targets/codecs
11:14:08 Quit shodanX ("leaving")
11:14:26 Join shodanX [0] (
11:15:07Zagoras I remember, buffering boost is done to keep the disk spinning for a short a time as possible.
11:16:32kugelbut what if a codec runs only a bit more than real time. refilling buffer paralell to decoding may require boosting again in such a case
11:18:03Zagorflash buffering is very quick. I think it would be difficult to find a case where boosting makes a real difference.
11:18:28Zagoron the other hand buffer boost doesn't cost much either, so perhaps there's not much point adding more #ifdefs
11:18:29kugelit takes quite a while on the e200
11:19:00*kugel is back later
11:19:06 Quit kugel (Remote closed the connection)
11:30:02 Quit alexbobp (Read error: 110 (Connection timed out))
11:31:43 Join alexbobp [0] (n=alex@
11:31:47lucentsomething else is wrong, that I don't know
11:35:47 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
11:42:20ZagorLinusN: do you know what happens if we would ignore the bypass bit requirement and just "hot switch" CPUDIV on coldfire?
11:43:32Zagorwrong answer :-)
11:44:36 Join kachna [0] (
11:47:57Zagorkugel: if we don't boost buffering, and that causes the codec to boost instead, we haven't lost anything. boost is boost.
11:49:00 Nick fxb__ is now known as fxb (
11:50:11t0masZagor: CPUDIV controls the cpu clock right? so that might have a really weird effect on other things relaying on the cpuclock?
11:51:02Zagort0mas: I don't see how waiting for PLL resync changes that
11:51:59t0mashm.. is that the only thing the bit accomplishes?
11:52:43t0mas(have you tried if it is possible to change CPUDIV without that? or is that a really stupid question? :-))
11:53:05Zagorthat is basically what I asked linusn
11:54:09ZagorI might do some experiments tonight
11:54:35t0mashmhm.. I ment.. "isn't the change just prohibited"
11:54:42t0masor.. not processed..
11:55:34Zagorit's prohibited in the docs. or at least that's not the proper way to do it.
11:55:45LinusNi would think that changing the frequency on the fly could cause temporary illegal timing
11:55:52amiconnZagor: There is still a dependency problem for codecs. Even after your api bump, a build made with plain 'make' still uses the old ci->configure arguments and hence doesn't work
11:56:25amiconnThis time it looks like codec fail because of the api version mismatch
11:56:31ZagorLinusN: what does "illegal" mean in this context?
11:56:38LinusNabove spec
11:56:40Zagoramiconn: that's bad
11:57:11t0masZagor: setting the bypass bit makes the cpu-clock be based on the crystal instead of crystal + PLL right?
11:57:11LinusNdepending on how the pll works and when you do the change
11:57:18LinusNt0mas: yes
11:57:19t0masso indeed.. it might be that the PLL emits some to high frequency while changing from one to another?
11:57:21Zagort0mas: yes
11:57:29LinusNt0mas: that is my guess
11:57:39Zagorwell we
11:58:02Zagorwe're not changing that mode at all (in my experiment)
11:58:12t0mas(why would you want to skip it? to keep the cpu from dropping in frequency steeply while switching to or from boost?)
11:58:27Zagorbut then again I've never made a PLL so I don't know where it could fail
11:58:36Zagort0mas: yes, the pll wait takes a long time
11:58:54t0masok, and all that time the cpu runs at a very low frequency?
11:59:02Zagoron other targets we have dividers we can set without waiting, making boost near-instantaneous
11:59:08t0masso boosting and un-boosting is essentially an expensive operation now?
11:59:11LinusNi wouldn't want to risk anything by skipping the bypass, the errors resulting from it could be subtle and cause all kinds of weirdness
11:59:12Zagort0mas: yes, and we wait for PLL and do nothing else
11:59:26Zagort0mas: on coldfire, yes
12:00:09amiconnLinusN: I would expect lockups because the clock change is probably not glitch free
12:00:14t0mashm.. but I fear they didn't put that in the docs for no reason
12:00:41ZagorLinusN: on the other hand the docs doesn't say it's a bad idea. it just sais "to change PLL frequency you must use the bypass bit". but we don't want to change the PLL frequency.
12:00:50t0mas(is it legal in sweden to check how to original FW changes frequencies?)
12:01:14Zagort0mas: it's legal pretty much everywhere
12:01:42LinusNZagor: afaik, boosting is all about changing the pll frequency
12:02:18Zagorno, it's about changing the cpu frequency.
12:03:37t0masehm.. but to change the cpu frequency you change the output frequency of the PLL?
12:04:46ZagorPLL = Phase-Locked Loop. CPUDIV is after this loop.
12:05:06Zagorwe already do this on other targets, it's only coldfire that must wait for pll
12:05:46 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
12:06:13pondlifeAnyone else getting odd playback with current SVN?
12:06:24pondlifeAs in, select track n, it plays track n+21
12:06:27pondlifen+1, even
12:06:31LinusNZagor: i must say that the docs are a bit unclear
12:06:34LinusN"The PLL Configuration Register must always be programmed to Bypass mode before it is
12:06:34LinusNreprogrammed to change any clock frequency"
12:06:54ZagorLinusN: yes, that's why I'm nursing this small hope :)
12:07:05 Quit kadoban (Read error: 113 (No route to host))
12:07:06pondlifeAlso Follow Playlist is not working - it follows the dir, but selects track 1
12:07:47JdGordonit does? :(
12:07:55t0maswell.. the line LinusN has here is more like "never change anything without setting it to bypass first"
12:08:35 Join _rossy_ [0] (
12:08:48Zagort0mas: that's one way to interpret it. but it's not explictly said so.
12:08:55pondlifeJdGordon: This only happens after natural playback end, I think
12:09:07t0mas(I'm reading about how a PLL actually works)
12:09:11pondlifei.e. play a directory, with repeat and auto-change off.
12:09:12LinusNit all depends on how the pll circuitry uses the register values
12:09:45pondlifeWhen playback completes, the browser (reasonably) has track 1 selected. Then try to play track n (where n>1)
12:10:00LinusNi mean, does the hardware care about if the values are actually changed, or does any write to the register cause a pll relock?
12:10:07pondlifeI've only played one album so far, will see if it's reproducible shortly
12:10:12JdGordonanyone got a few short mp3 clips to test this with?
12:11:40*ajb suggests makes a good library of test clips encoded for various codecs
12:12:25JdGordonpondlife: when playback ends, track 1 is selected!?
12:12:34 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:13:15pondlifeThat's reasonable, though. I guess you weren't expecting a track change event?
12:13:23JdGordonthat doesnt sound right
12:13:33JdGordonIts selecting the last track in the folder here
12:13:50t0masZagor: I guess the problem with this is that we don't know if it is doing weird things when you skip the bypass bit?
12:14:27t0masit could do weird things to the cpu.. (running it at a to high frequency for a very shot period of time?) that might just be enough to lock it up sometimes?
12:14:54t0mas(sometimes possibly being 1/1000000 times? which would create a "bug" we can't reproduce?)
12:16:34LinusNthat's what i fear
12:18:28t0mas"When the clock circuit is switched from “bypass” to “normal operation”, the switch-over is delayed until the PLL is locked."
12:18:37t0masit's the switch from bypass to normal they care about?
12:19:55 Join havien [0] (
12:20:59t0masZagor: how do we wait for PLL lock? by switching from bypass to normal? or by checking the LOCK bit in PLLCR? (bit 31)
12:21:54JdGordonseems the WPS doesnt like lots of small tracks... 5s tracks are too fast for it to keep up (the playlist index)
12:22:28rasherJdGordon: Could that be related to issues when skipping quickly?
12:22:38JdGordonno idea
12:23:17t0masZagor: have you read this: "If this bit is written 0, the PLL is not used, and the crystal clock is sent directly to the CPU. Write this bit 0 before changing any other bit in this register."
12:23:38JdGordonpondlife: I cant reproduce your problm... im 99% sure its playing the track selected, and selecitng the last file in the dir when it finishes
12:24:01t0masZagor: especially the "Write this bit 0 before changing any other bit in this register." part? as CPUDIV is in the same register
12:24:35t0masand "After writing 1 to this bit, new setting will become active after a hardware controlled delay."
12:24:54t0masso I guess if you don't set it to 0 first.. the change just won't have any effect
12:24:57pondlifeJdGordon: I may have started playback from the database, then gone into the file browser during the album.
12:25:10pondlifeLeave it with me, I'll work out a proper repro recipe
12:25:33pondlifeAh, I had also done some dirskip earlier on - might be relevant
12:25:47JdGordonhmm... this is wierd.... the wps is showing the playlist number way too early for the next track
12:26:32JdGordonand by way too early, I mean with <3s left on the previous track
12:26:46pondlifeJdGordon: There are 2 different track transitions - one when the codec moves over, a second when the UI does.
12:27:06pondlifeThe UI should all be using the second one
12:27:23 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:27:49pondlifeNo idea which event is at which transition though.
12:28:08 Quit sunny (Read error: 104 (Connection reset by peer))
12:29:35pondlifeReally, the codec-transition should be entirely internal to playback.c, not in playlist.c. But IIRC, that's not (yet) the case.
12:29:58JdGordonwern't we going to do something about this hwole playback mess in the new year? :)
12:31:10JdGordonit looks like it might be plaback doing the nautghy here
12:31:19JdGordontelling everyone its changed tracks too early
12:31:49kugelwell, it tells when the codec begins decoding the next track
12:33:02kugelwhich isn't entirely wrong I think. you can't really break the two track change transitions if you don't want the codec to wait on the UI
12:33:27JdGordonno.. I'm wrong.. the wps is updating too early... follow playlist isnt getting the next track untill its actually changed tracks
12:34:09kugelwhat might work is to send the event not when the codec changes track, but when the pcm buffer is empty for that track, but that needs some kind of seperating the pcm buffer
12:34:49kugel"pcm buffer is empty for that track" as in no parts of the previous track are in the pcm buffer anymore
12:35:21JdGordonI'm not sure how it works.. but the PLAYBACK_EVENT_TRACK_CHANGE event is sent at the correct time...
12:35:30JdGordonthat is what the wps should be using
12:35:51JdGordonassuming everyone is OK with forcing everything to use events more...
12:36:55pixelmaIIRC, this WPS shows track change too early is an old bug (I believe introduced with MoB?) which NicoP tried to fix last year but that fix introduced all kinds of skip tracks issues we had in the first half of last year (until around DevCon) and then he reverted that fix
12:37:20JdGordonwhat sort of skip track issues?
12:38:41pixelmasporadically - sometimes skip 3rd track in a playlist or sometimes a few tracks (probably as much as fit into the buffer) or if you skipped tracks manually you could "confuse" it even more
12:39:05JdGordonso bassically its the problem that playback is tied into the ui dangerously?
12:39:49pixelmaI had that skip 3rd track quite reliable in certain directories, never figured out what makes those directories so special compared to other ones where it didn't happem
12:39:53pixelmahappen too
12:41:13pixelmaI think NicoP would know best but didn't see him active often lately
12:41:22JdGordonyeah :(
12:42:12 Join fyrestorm [0] (
12:43:19 Quit ze (Read error: 113 (No route to host))
12:45:23 Part B4gder
12:47:29 Quit _rossy_ ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
12:48:17pondlifeJdGordon: I think events would be good. Certainly decouple the playback code from it's clients.
12:48:27pondlifeThe problem is where does playlist.c fit
12:49:15JdGordonSomewhere next to plaback.c...
12:49:59JdGordonI imagine say for a track change wps->playback(next track)->playlist<->(get track filename)
12:50:04JdGordonif that makes any sense
12:50:14pondlifeplaylist_() functions are called and used for the UI, so I think they should sit on the UI side of the timing..
12:50:28JdGordoncuesheets seem to be a problem becuase they are only handled by the wps atm by the looks of it
12:50:53pondlifeCuesheets should be handled by playlist.c, and transparent to the UI code ideally..
12:51:24 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
12:51:25JdGordonexcept with some way to know if it is a cue or not so the wps can put markers in
12:51:33 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
12:51:33*pondlife hasn't used cuesheets
12:51:39pondlifeDon't they appear as a playlist?
12:51:47pondlifeOr as a long track with markers?
12:51:54kugelthe latter iirc
12:51:55JdGordonlong track with markers i think
12:51:56pondlifeThe former is what I'd expect
12:52:08pondlifeBut I'm probably not the target audience
12:52:32JdGordonI think playlist and playback are completly seperate, except playback needs to get the filenames from playlist, no other interaction
12:53:00pondlifeSo playlist has no concept of track transition/playlist index?
12:53:38JdGordonplaylist is simply a list of filenames...
12:53:40***Saving seen data "./dancer.seen"
12:53:56pondlifeWhen playback goes from track 1 to track 2, does playlist.c get involved?
12:54:24JdGordonnot really, it needs to be quiried to get the filename, but thats pre-change
12:54:48pondlifeSo where does the playlist index live? In playback.c ?
12:55:03*pondlife apologises for not having the source handy
12:55:10pondlifeAh, ok
12:55:21JdGordonI dont know how it works now.. im saying how it shold work
12:55:25pondlifeAh, ok
12:55:25JdGordonprobably very different :p
12:55:34 Join nplus [0] (n=nplus@141.25.Globcom.Net)
12:55:35pondlifeWell, I agree with you, at least :)
13:00:24 Quit kachna (Read error: 113 (No route to host))
13:03:01JdGordonchanging topics slightly... a possible way to fix the no voice while paused problem... could we dump pcmbuf to disk if that happens, do the talk then on resume just read that ddump back in and continue?
13:03:27pondlifeSounds overcomplicated
13:03:44pondlifeBetter to just not pause the PCM, but the playback
13:03:53 Join DerDome [0] (
13:04:00JdGordonyes, but it would work untill a proper pcm mixer is done
13:04:02pondlifeIf that's too slow, then only do it that way when voice is in use
13:05:24 Quit kugel (Remote closed the connection)
13:05:51 Join kugel [0] (n=chatzill@unaffiliated/kugel)
13:13:14JdGordonwe dont support internal cuesheets yet right? if we ever do, is it reasonable to assume that tracks with an internal cue would never have an external .cue file?
13:15:05JdGordonI'm thinking of a way to make .cue 99% transparent to the ui by moving it into playback, so a cue is checked for just like AA (if its enabled)
13:18:46 Join Severian [0] (
13:19:06 Join Seed [0] (
13:21:51 Quit agaffney (Read error: 104 (Connection reset by peer))
13:21:59 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
13:23:09ZagorJdGordon: is voice on pause a mixing problem?
13:23:43*ajb uploads v3 of the gitsvn version patch
13:23:45linuxstbJdGordon: No, we don't support them currently. Why is it important that a file with an internal cue doesn't have an external one? We would need to decide which to prioritise in such cases, but I think that choice should just be internal to the metadata parsing code.
13:24:05Zagorcouldn't we just send voicebuf to dma directly if playback is paused?
13:25:48JdGordonZagor: iiuc its a problem that the audio pcm gets mixed in too early
13:27:00JdGordonlinuxstb: thinking aloud.. currently the id3 sturct has an int to say if there is a cue (and if its internal or not), I think I'd like to change that to a pointer to a cuesheet struct.. nothing really cares if its internal or not once its loaded...
13:27:16pixelmaJdGordon: if playback is responsible for cuesheets, wouldn't you need two different versions then? Not that I'd insist on cuesheet support on hwcodec but I imagine more differences would make everything harder to maintain
13:28:00t0masZagor: spotted the last quotes from the manual?
13:28:07amiconnZagor: It is partly a mixing problem and partly a problem how pausing works - it pauses the entire pcm
13:28:27t0masin the manual: "Write this bit 0 before changing any other bit in this register." and CPUDIV is in that register
13:28:33Zagoramiconn: exactly, so why don't we just send voicebuf if paused, unmixed?
13:29:01 Join AndyI [0] (i=AndyI@
13:29:04amiconnDon't ask me... I don't really understand how our pcm playback works
13:29:10JdGordonpixelma: CRAP! forgot about hwcodec again... I guess it shouldnt be too difficult to tack cuesheet searching onto it though
13:29:13t0masand "After writing 1 to this bit, new setting will become active after a hardware controlled delay." so it just won't activate when you don't write 1 to the bypass bit as I read it
13:29:19Zagort0mas: which quotes?
13:29:24t0masthose 2
13:29:31Zagorah :)
13:29:55t0masthose are a bit less debatable
13:29:57Zagorwell manuals are one thing, reality sometimes another. I still think it's worth experimenting with.
13:29:58JdGordonZagor: we still need to trash pcmbuf somehow... the codec needs somewhere to dump pcm before it can be sent to dma
13:30:18ZagorJdGordon: pcmbuf and voicebuf are different buffers
13:30:53Zagorthey get mixed afterwards
13:31:19Zagoror rather, audiobuf and voicebuf are different buffers. they are both in pcmbuf...
13:31:31Zagorit is rather confusing
13:34:00linuxstbJdGordon: So you're planning on loading the cuesheet at the time the track is buffered?
13:34:11JdGordonI think so
13:34:21JdGordonor at least put it on the buffer like AA
13:34:56linuxstbPutting it on the buffer wouldn't make any sense unless it was done at buffering time...
13:36:00 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
13:36:43 Quit kugel (Remote closed the connection)
13:36:58JdGordonyeah, the .cue would get loaded into the buffer but not parsed untill the track is actually started, only 1 .cue would ever be open at a time so 1 cuesheet struct will sit in static ram, the tracks id3 struct will either point to it or be NULL if its a regular track
13:37:13linuxstbWhy not parsed?
13:38:15JdGordonwell.. it could be, wastes less filebuf doing that..
13:39:03linuxstbShouldn't a parsed structure take less RAM than the text file?
13:39:14JdGordonyeah, thats what I meant
13:39:31*JdGordon is getting sleepy
13:40:02linuxstbAnd that also makes it easier to store different kind of sub-track information - the metadata parser for that format could fill the cue struct with anything.
13:40:24 Quit AndyIL (Read error: 110 (Connection timed out))
13:41:26 Nick fxb is now known as fxb__ (
13:42:09JdGordoncuesheet was done before MoB right? I assume thats why it doesnt work like this already?
13:43:12 Quit bmbl ("Woah!")
13:43:33 Join PaulJam [0] (
13:43:44 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
13:47:01 Quit HellDragon (
13:47:01 Quit kkurbjun (
13:47:01 Quit advcomp2019 (
13:47:01 Quit neddy (
13:47:01 Quit amigan (
13:47:01 Quit Tristan (
13:47:04 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
13:47:08 Join lazka [0] (
13:48:19NJoinamigan [0] (
13:48:22NJoinTristan [0] (
13:48:51 Join _lifeless [0] (n=lifeless@
13:49:16lazkashould SansaV1 charging work now? Mine does endless reboots when I plug it in...
13:49:55dionoeaworks fine here (if i plug it in after booting and hold select)
13:50:31dionoeaI have an e260v1
13:52:11pixelmaJdGordon: MoB doesn't exist on hwcodec... and weren't you one that helped coding the original cuesheet support?
13:53:08pixelmaNicoP was the main dev though IIRC
13:54:11 Join gregzx [0] (
13:56:59JdGordonyes and yes
13:57:29JdGordonnico took over pretty much once the parser was done... So all the loading is new to me
13:59:48 Quit culture (Read error: 60 (Operation timed out))
14:02:13NJoinkkurbjun [0] (
14:02:19NJoinHellDragon [0] (
14:02:19pixelmaI see
14:02:30 Join neddy [0] (n=john@nat/sun/x-7da3769213bcece2)
14:07:06 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
14:07:35 Join TheSphinX^ [0] (n=cold@
14:08:01 Join itcheg [0] (i=62db4767@gateway/web/ajax/
14:10:25JdGordonno more errors! time to see if it actually works!
14:11:37*JdGordon doesnt have any .cue files :p
14:15:36JdGordonholy crap! I think it works first go!
14:15:49B4gderwe don't believe you! ;-P
14:15:57JdGordonI know :'(
14:16:57JdGordonid3 spoofing is still done in the wps, but loading and skipping works
14:18:00lazkadionoea, do i have to hold select?
14:18:16dionoealazka: no. That's just to prevent it from entering USB mode
14:18:27dionoeas/USB/MSC or whatever it's called
14:21:03 Quit __lifeless (Connection timed out)
14:21:17lazkaI'll try that... If it doesn't work it's maybe because of the player because e.g. yesterday the bootloader crashed and after that both firmware images had bad CRCs and all audio files hat bad data, most don't play anymore.
14:22:10dionoeahum, you might have a corrupted file system or something
14:23:12 Quit itcheg (" ajax IRC Client")
14:24:14lazkadionoea, select works, thanks.
14:26:32 Join Schmogel [0] (
14:29:14dionoealazka: I guess that it should also work in normal USB mode. You'll have to ask jhMikeS about that.
14:31:44lazkaIs it still a bug if the filesystem is corrupt? (I'm pretty shure it is)
14:44:22lazkaFS 9790
14:44:26 Quit mcuelenaere ()
14:46:55 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
14:47:55JdGordondid someone break its not on the frontpage?
14:50:44 Join josch [0] (
14:50:51joschhello there!
14:53:44***Saving seen data "./dancer.seen"
14:54:06ZagorJdGordon: fixed
14:55:05joschi wanted to buy a new rockbox player and the toshiba gigabeat X-series seams profitable but is appearently not sold anymore - do you guys know any other sources for used stuff than ebay?
14:57:16linuxstbThe X has always been hard to find. The Gigabeat F is much easier, and Rockbox runs well on the S series now (although the install procedure isn't straightforward on the S).
14:59:10joschoh the S is missing in the BuyersGuide - thx for the info!
14:59:39linuxstbIt's listed amongst the "in development" targets - see the TargetStatus page for others, and their current status.
15:00:57linuxstbBut I think the S is the only "in development" target that's actually usable. It's only not listed as "supported" due to the installation process/documentation not being ready for the general user.
15:01:10 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:01:48joschoh no prob for me! great!
15:03:14 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
15:13:34 Quit CaptainKwel (Remote closed the connection)
15:13:52 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
15:20:17 Quit itcheg (" ajax IRC Client")
15:21:17 Join kachna [0] (
15:26:22 Part LinusN
15:27:21mcuelenaeregevaerts: but such a large difference, is that normal?
15:27:37gevaertsDepends on what you're doing
15:27:47mcuelenaere(I have to admit, I didn't put attention when I monitored the OF)
15:28:20mcuelenaeredmesg gives something like: new SCSI device detected, and then 4 usb resets
15:28:22mcuelenaereand that's it
15:28:36mcuelenaerewhile graphicusb is still indexing..
15:28:51mcuelenaerecould a large number of frames influence this indexing?
15:29:45gevaertsAnyway I'd really look at a working device first
15:30:03 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
15:33:57 Join BXCracer [0] (n=bxcracer@
15:34:49 Quit Slack_ (Read error: 60 (Operation timed out))
15:35:27 Quit _lifeless (Remote closed the connection)
15:35:36 Join _lifeless [0] (n=lifeless@
15:36:34 Join Slack_ [0] (
15:37:34 Quit Kopfgeldjaeger ("Serverwechsel")
15:37:57 Join obo [0] (
15:38:20 Join grndslm [0] (
15:39:20 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
15:39:30 Nick fxb__ is now known as fxb (
15:39:50mcuelenaeregevaerts: is it normal that graphicusb takes up ~800MB of RAM?
15:40:22 Join Kopfgeldjaeger [0] (
15:40:34gevaertsmcuelenaere: I never looked
15:50:04mcuelenaeregevaerts: I filtered all that can get filtered and now I get a lot less data and the program is more responsive
15:50:18mcuelenaereI've looked at the data before every reset, but AFAICS it looks ok
15:51:57Unhelpfulajb: i'm inclined to agree with you... what i was thinking was r<svn rev>-<build date>-<extra>, where extra is git commit ID or an indicator that the repository has local changes for svn, but maybe it's best to take the discussion to the mail list. more devs will see it there, and it works a little better for talking about things with devs who don't happen to be there when you are.
15:52:00mcuelenaerehowever, it seems like the device adds 13 bytes of 00's
15:52:13mcuelenaereafter a CBW: INQUIRY
15:52:22mcuelenaerein response to a INQUIRY*
15:53:29gevaertsThat could well be your problem
15:54:10mcuelenaereright, now how to fix it? :)
15:55:56gevaertsmcuelenaere: maybe compare with another device (like a flash thing) to make sure you understand the packets correctly
15:56:15mcuelenaereI'll compare with the OF
15:56:52mcuelenaerewhat wonders me is that graphicusb reports these extra 13 bytes as vendor-specific, and doesn't alert on it
15:56:59mcuelenaerewhile the linux (and windows) driver does
15:58:41gevaertsmcuelenaere: there are some traces of the rockbox stack on a sansa at gevaerts/usb/">
16:04:20*mcuelenaere thinks he's on to the bug
16:04:40 Quit Seed ("cu, Andre")
16:05:10 Join Seed [0] (
16:06:03mcuelenaereit adds the 13 bytes of the next packet on to the first
16:06:12mcuelenaereso you were right gevaerts, it merges the two
16:08:39gevaertsSo you now just have to find how to force a packet to end
16:08:58 Join {phoenix} [0] (
16:09:07 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:10:40 Part B4gder
16:12:11mcuelenaereI think I just got a bit further :)
16:12:21mcuelenaerehmm still one reset
16:15:53 Quit japc ("Ex-Chat")
16:16:17 Quit Zagor ("Client exiting")
16:16:59 Join robin0800 [0] (
16:21:28 Join japc [0] (
16:24:01 Part pondlife
16:29:05 Part loswillios
16:29:59*mcuelenaere believes USB is working on the Onda VX747 \o/
16:36:02mcuelenaeremd5 check matches! :)
16:37:00gevaertsNow the zvm ;)
16:37:37mcuelenaereI'll have to look where my ZVM is actually
16:37:45mcuelenaereI've not seen in since ... months I think
16:39:48mcuelenaeregevaerts: there are still some issues when I try to read past the 1MB RAM disk (USB resets), but I suppose that's normal behaviour?
16:40:46gevaertsThat could mean a bug in the ramdisk size handling. I wouldn't worry too much about that though
16:42:40mcuelenaeredoes the usb stack handle resume/suspend?
16:44:58LambdaCalculus37Regarding the Gigabeat F manual, I was thinking about adding a small bit of description about the Gigabeat X into it as well, since those two targets use the same firmware and have the same keymaps.
16:46:07 Join BXCracer_ [0] (n=bxcracer@
16:49:59 Quit Seed ("cu, Andre")
16:53:04 Quit BXCracer (Read error: 110 (Connection timed out))
16:53:48***Saving seen data "./dancer.seen"
16:58:10 Quit TheSphinX^ ("XChat@Linux")
17:01:38 Join flydutch [0] (
17:03:04 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:05:16 Quit BXCracer_ (Read error: 60 (Operation timed out))
17:13:02 Quit robin0800 (Read error: 110 (Connection timed out))
17:14:01 Join robin0800 [0] (
17:16:51gevaertsmcuelenaere: no
17:17:54ajbUnhelpful: OK, I shall msg the list....
17:18:08linuxstbLambdaCalculus37: That sounds sensible, as it's a shared manual...
17:19:53LambdaCalculus37linuxstb: Okay, I'll add some mention of the X to the manual, then.
17:20:42 Join OoOFreemanOoO [0] (n=52ea00c9@gateway/web/cgi-irc/
17:23:10 Join BXCracer_ [0] (n=bxcracer@
17:26:09 Quit OoOFreemanOoO ("CGI:IRC (Ping timeout)")
17:26:51 Join OoOFreemanOoO [0] (n=52ea00c9@gateway/web/cgi-irc/
17:27:20 Quit OoOFreemanOoO (Client Quit)
17:27:43 Join toffe82 [0] (
17:29:19 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:29:22 Join nuonguy [0] (
17:35:12 Join kugel [0] (n=chatzill@unaffiliated/kugel)
17:58:26 Quit grndslm ("Leaving")
17:58:32undertakingyoumcuelenaere: ping
18:01:48 Quit petur ("work->home")
18:02:01 Join saratoga [0] (n=9803c264@gateway/web/cgi-irc/
18:04:46 Quit jhMikeS (Nick collision from services.)
18:04:52 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
18:04:57 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
18:05:30 Join kadoban [0] (
18:09:00mcuelenaereundertakingyou: pong
18:10:09saratogais it ok to file bugs for the website on the bug tracker?
18:11:05Lloreanwww is in SVN, right? :)
18:11:06 Join bertrik [0] (
18:11:15LloreanBut yeah, I think so. I thought there was a category in the tracker for it already?
18:11:44undertakingyoumcuelenaere: I understand that you are working on the creative zvm port, is that right?
18:12:01mcuelenaerewell, actually more were
18:12:31undertakingyouDo you know if anyone has taken that up?
18:12:49undertakingyouI have a zvm, but no C knowledge, and would like to help.
18:13:09kugelI think you need both, in this case
18:13:18mcuelenaerenope, nobody has done anything that really helped the port
18:13:27mcuelenaerethere were several people that investigated it
18:13:35undertakingyoukugel: you are probably right.
18:13:36mcuelenaerebut they didn't really got far or stopped or ..
18:13:56kugelthe problem was this homebrewed FS, wasn't it?
18:14:05mcuelenaereundertakingyou: apart from no C knowledge, do you have any experience with embedded stuff?
18:14:12mcuelenaerekugel: FS?
18:14:18undertakingyouNope, none whatsoever.
18:14:29mcuelenaerekugel: yes, well that isn't really a problem anymore
18:14:37mcuelenaeremost of it has been resolved
18:14:41undertakingyouI have dabbled in C# and I do webscripting but nothing on embeded systems.
18:14:44mcuelenaereit just needs someone to do the last bit
18:14:58 Join gregzx [0] (
18:14:59mcuelenaereundertakingyou: then it'll require you to do quite some learning I guess
18:15:38undertakingyouThat is what it looks like. So with a zvm can it be recovered if the firmware doesn't work or does it just brick?
18:15:50mcuelenaerekugel: currently the main problem(s) are the PIC (which controls the buttons and other input/output stuff) and USB I guess
18:16:04mcuelenaereand completing CFS
18:16:17mcuelenaereCreative File System
18:16:27kugelah, the homebrewed one ;)
18:16:30mcuelenaereyep :)
18:16:35kugelwhat are the problems with the PIC?
18:16:40mcuelenaerebut, AFAIR it was almost completely figured out
18:16:42 Join grndslm [0] (
18:17:08mcuelenaerekugel: it runs Creative code, so it isn't any standardised/documented protocol
18:17:16mcuelenaere(the main CPU communicates with the PIC through I˛C)
18:17:26kugeluhh, terrible
18:17:31mcuelenaereso one needs to RE the PIC's code
18:17:54mcuelenaereand I really don't like to do that
18:18:12kugelI have to say: "No, thanks" too ;)
18:18:29mcuelenaerewell, I remember there was a simulator around
18:19:03mcuelenaerebut still, the PIC has all sorts of low level abilities (which I don't have knowledge of); so I think simulating it will need a really good similar environment to work
18:24:13 Join webguest00 [0] (n=5b84f61a@gateway/web/cgi-irc/
18:28:37 Quit webguest00 (Client Quit)
18:28:42 Join webguest00 [0] (n=5b84f61a@gateway/web/cgi-irc/
18:33:39 Quit webguest00 (Client Quit)
18:43:13mcuelenaeregevaerts: I get scsi unknown cmd 2F when trying to format the RAM disk on Windows
18:43:47 Quit blahrus (Remote closed the connection)
18:47:10 Quit DerDome (Read error: 104 (Connection reset by peer))
18:53:44kadobani'm writing a plugin that reads/writes files occasionally. do i need to be worrying about the storage_* functions for spinning up and down, or is that just if i want to override the default behavior? (in other words, if i don't use those, is the disc going to be spinning constantly?) i don't have a player with a non-flash hard drive, so it's pretty impossible for me to tell...
18:53:52***Saving seen data "./dancer.seen"
18:54:38 Join BigBambi [0] (
18:55:51 Quit BigBambi (Client Quit)
18:56:05 Join BigBambi [0] (
18:59:16 Quit Zambezi (
18:59:16 Quit preglow (
18:59:16 Quit DaCapn (
18:59:16 Quit tim__b (
18:59:16 Quit Tuplanolla (
18:59:16 Quit Dieterbe (
18:59:47NJoinZambezi [0] (i=stolgfor@
18:59:47NJoinDieterbe [0] (
18:59:47NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
18:59:47NJointim__b [0] (
18:59:47NJoinDaCapn [0] (
18:59:47NJoinpreglow [0] (
19:00:26 Quit preglow (
19:00:26 Quit DaCapn (
19:00:26 Quit tim__b (
19:00:26 Quit Dieterbe (
19:00:26 Quit Zambezi (
19:00:26 Quit Tuplanolla (
19:00:29NJointim__b [0] (
19:00:33NJoinDaCapn [0] (
19:00:37NJoinDieterbe [0] (
19:00:38NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
19:03:27 Quit kugel (Nick collision from services.)
19:03:31 Join kugel_ [0] (
19:03:42 Nick kugel_ is now known as kugel (
19:04:01gevaertsmcuelenaere: According to the SCSI spec that's VERIFY. From my reading it's optional. Does the operation continue afterwards?
19:04:26mcuelenaereyes, it doesn't stop
19:04:36mcuelenaereit repeats and repeats
19:04:44mcuelenaereI manually stopped the operation
19:04:57mcuelenaereshould I retest and wait whether it times out eventually?
19:05:01gevaertsOK. Maybe we return the wrong error code for "Not implemented" then
19:06:35 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:10:05gevaertsI'll have a look at the SCSI spec later today. Now I'm going home
19:10:27*mcuelenaere has the feeling his NAND driver isn't really fast
19:12:23 Join karashata [0] (n=karashat@
19:12:32 Quit jfc (Read error: 54 (Connection reset by peer))
19:12:41NJoinZambezi [0] (i=stolgfor@
19:15:35NJoinpreglow [0] (
19:24:14LambdaCalculus37This is a strange error:
19:24:23LambdaCalculus37How can this be when Rockblox has a Clip keymap?
19:24:34LambdaCalculus37(That is if there's one...)
19:25:09kugelLambdaCalculus37: you're playing with the clip keymap?
19:26:29domonokyLambdaCalculus37: the cause is of course the error around line 130 or before..
19:29:21LambdaCalculus37kugel: No, just building normally.
19:30:48 Quit tvelocity (Connection timed out)
19:31:37 Join tvelocity [0] (
19:32:54*pixelma wonders about the fixme comments in rockblox.c
19:34:39 Quit mcuelenaere (Read error: 113 (No route to host))
19:35:30pixelmaLambdaCalculus37: do you say that SVN doesn't build for the Clip currently?
19:36:13kugelpixelma: I added them, I plan to resolve soon
19:36:41pixelmakugel: but why?
19:37:16LambdaCalculus37pixelma: Yes, did you see the pastebin?
19:37:27kugelwhy not? I think it should be looked at
19:37:36pixelmakugel: What's the problem? I mean Rockblox is working correctly for scrollwheel targets for a while...
19:38:06pixelmaand targets without scrollwheel
19:38:09kugeli see, those which I added at e200 and fuze are apparently wrong
19:38:35kugelbut they don't define this SCROLL_WHEEL as ipods do, which was weird, so I added a comment
19:39:19kugelwell, not wrong. it's just about why they don't define that
19:39:44domonokyLambdaCalculus37: but current clean svn builds for clip, see buildtable.
19:40:05 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
19:40:30 Join miepchen^schlaf [0] (
19:41:08 Join Thundercloud [0] (
19:41:13kugel"working correctly" doesn't mean mysterious spots can't be reviewed
19:44:46pixelmaaha, didn't see that rockblox had some special scroll wheel handling
19:46:38 Quit miepchen^schlaf ()
19:46:47 Join aptzero [0] (n=penovich@
19:46:54kugelyes, and I wondered why that apparently only applies for ipods (which might very well just have been overlooked when creating the e200 keymap), so I planned to look at that again, if noone beats me to it
19:49:24 Join miepchen^schlaf [0] (
19:51:08 Quit jhulst (Read error: 104 (Connection reset by peer))
19:51:49LambdaCalculus37kugel: It's all yours.
19:52:53kugelLambdaCalculus37: ?
19:54:14 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:54:32kugelmcuelenaere: nice!
19:54:45mcuelenaerekugel: ?
19:55:15LambdaCalculus37kugel: I mean you can take a look at the scroll wheel handling.
19:55:47kugelmcuelenaere: your commit :)
19:55:47 Quit havien (Read error: 104 (Connection reset by peer))
19:55:57mcuelenaerekugel: ahh :)
19:56:11mcuelenaerethank the usb tracer ;)
19:56:13kugelI followd your talk with gevaerts a bit, and it seems like it was a tough job, so I find it nice that it works :)
19:56:28 Join tyfoo [0] (
19:57:06mcuelenaerewell, with the usb tracer it wasn't that hard anymore :)
19:57:22mcuelenaerewithout, I would still be reading code and datasheets :)
20:00:48 Quit axionix (Remote closed the connection)
20:02:03 Join jgarvey [0] (
20:03:10kugel:) Good job in any case
20:09:06 Quit miepchen^schlaf ()
20:11:49 Join Zagor [242] (
20:12:53Zagormcuelenaere: excellent work! nice to see the usb tracer being helpful once again
20:13:07mcuelenaerethanks :)
20:14:23kugelmcuelenaere: beware, if you keep committing working USB stuff, you might end up as our new usb pro, and you don't want that :)
20:14:51mcuelenaereheh, I only have 2 devices here so :)
20:15:14gevaertsI guess for this one software tracing could have been sufficient, but hardware tracers usually have much better software
20:15:24*kugel points to the AMS sansas that need USB sometimes too ;)
20:15:44*gevaerts points out that we don't have anything resembling a datasheet for that one yet
20:16:04ZagorjhMikeS: did you get any further about that?
20:16:10*kugel isn't in the mood for bad news today :/
20:16:55 Join ze0 [0] (i=ze@
20:30:19 Nick ze0 is now known as ze (i=ze@
20:30:23*mcuelenaere thinks he found the SD power enable pin
20:32:27 Quit nuonguy ("This computer has gone to sleep")
20:33:43kugelyay, good news again :D
20:35:27 Join archivator [0] (i=archivat@
20:39:22 Join jfc [0] (
20:41:25gevaertsmcuelenaere: did you trace that unknown cmd 2F thing? Does the host call SCSI_REQUEST_SENSE (0x03) after that?
20:41:49mcuelenaerenope, should I retry?
20:42:04gevaertsIf you can, that would be helpful
20:44:09mcuelenaerehmm it gets stuck on test_unit_read 0 and 1, that means it storage_read() returns -1, doesn't it?
20:45:29gevaerts"gets stuck" as in the host keeps retrying that?
20:45:59mcuelenaerebut I don't get it, I have defined USE_RAMDISK
20:46:26mcuelenaereI'll see what Linux gives
20:46:26 Quit Thundercloud (Read error: 104 (Connection reset by peer))
20:47:02 Join miepchen^schlaf [0] (
20:47:47mcuelenaerehmm Windows detects at least on 'disk'
20:48:16 Quit ze ("leaving")
20:48:29gevaertsTEST_UNIT_READY depends mostly on usb_exclusive_ata()
20:48:35mcuelenaereah ok
20:48:37 Join tessarakt [0] (
20:48:41mcuelenaerewell, now it sees both of them
20:49:37*gevaerts tries to find out from the scsi specs what the correct response is for an unimplemented optional command
20:49:50 Quit nplus (Remote closed the connection)
20:50:02mcuelenaereok I get unk cmd 2F now when formatting
20:50:17mcuelenaerethen it resets usb and sets config again
20:50:20mcuelenaereand retries
20:50:47gevaertsSo it doesn't retry the command or skip it
20:50:59 Join fml [0] (n=4fd3c0dd@gateway/web/cgi-irc/
20:51:13mcuelenaereyou mean cmd 2F?
20:51:27gevaertsWe could solve this by implementing VERIFY(10), but then we'd have the same problem at the next optional command that windows tries
20:51:38mcuelenaereit does retry and doesn't skip it
20:51:42bertrikgevaerts, a sense of 0x052000 perhaps?
20:51:53 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
20:52:00fmlHello. Is the FM region setting reachable through the "normal" settings menu? Or only from the FM screen?
20:53:21gevaertsmcuelenaere: it retries the command by itself, or it resets the device first?
20:53:33bertrikfml, the context button on the fm radio menu option allows you to set fm settings
20:53:48 Join Hillshum [0] (
20:53:55***Saving seen data "./dancer.seen"
20:54:34mcuelenaeregevaerts: it's something like this:
20:54:51fmlbertrik: yes, I know. But is it the only way? Is it somewhere in the "normal" settings (not context ones)? I couldn't fint it.
20:55:13mcuelenaereoh I forgot the read10 1
20:55:18mcuelenaereit sometimes give that too
20:56:22gevaertsbertrik: what would 0x052000 mean?
20:56:39*gevaerts tries to get that from the spec, but it's not always very clear
20:56:53mcuelenaere-{ 0x052000, "Invalid command operation code" },
20:56:57bertrikfml, I don't know, but I know those two. I think there are no other ways to reach the setting.
20:57:11at0m fml, can't find another way either, i'd guess it's the only way doing it from that menu yes
20:57:16mcuelenaerefound on
20:57:57fmlOk, thanks. I thought I was blind!
20:59:20fmlI was assuming that every setting is in the normal menu and, as a bonus, sometimes in a context one.
20:59:56bertrikgevaerts, what mcuelenaere said
20:59:58mcuelenaerehmm Windows seems to have given up, but no sign of that on the host
21:00:03*gevaerts looks things up
21:00:44gevaertsok. That should work
21:02:24 Quit fml ("CGI:IRC 0.5.9 (2006/06/06)")
21:03:05mcuelenaeregevaerts: could a bad ep_stall() implementation cause this?
21:05:04gevaertsmcuelenaere: I'm not actually sure if the ep_stall() there is entirely correct. I've seen other devices do it, but...
21:05:24*gevaerts prepares a patch
21:07:33 Join Seed [0] (
21:07:49mcuelenaeregevaerts: about usb_drv_send() (not non-blocking), what's the code supposed to do when it's waiting till finish? yield() ?
21:08:12mcuelenaere(and couldn't this be done in usb_core?)
21:09:26Unhelpfulkugel: i think there are good arguments in favor - removing duplicate API pointers, removing init functions that only copy pointers
21:09:44Unhelpfuland i like any patch that "adds" something while having more deletions than inserts :)
21:09:48kugelthat's why I hope you didn't gave up on this ;)
21:10:06 Quit jhulst (Read error: 104 (Connection reset by peer))
21:10:35gevaertsmcuelenaere: the ARC code just busy-waits. I don't really see how this could be done in usb_core.c, but I'm open to ideas
21:10:43gevaertsmcuelenaere: can you try ?
21:10:59*kugel assumes Zagor thought "Why am I doing this shit?" at times doing the make overhaul
21:11:09kugelthat's part of the business ;)
21:11:26Unhelpfulrecursive make considered harmful. you need another reason? ;)
21:11:47gevaertsYes. Dependency handling broken :)
21:12:13kugelchanging a running system? ;)
21:12:34Zagorkugel: actually I quite enjoyed doing it. a thorough cleanup feels great to do well.
21:12:47kugelyea, that's true
21:13:11kugelbut come on, at least once, you thought that, admit it :)
21:13:23mcuelenaerehmm the added code seems to crash the device?
21:13:25 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:13:35 Join igwinnimma [0] (
21:13:52Unhelpfulkugel: bitmap corruption? ;)
21:14:35gevaertsmcuelenaere: maybe the bits in usb_storage_init_connection() duplicate something you already do. Those are not actually part of what I'd like to see tested (but I forgot about them)
21:14:47Zagorkugel: nope, sorry :-) I was a bit worried about the lack of response to the patch though, but that turned to be standard Fear Of Make.
21:15:09Zagorturned out*
21:15:38*gevaerts sets Zagor to work on FS #9545 :)
21:15:58kugelnow guess why there's so little hieroglyphics specialists
21:16:21Zagorgevaerts: hehe. unfortunately my plate is rather full at the moment.
21:16:46Zagorkugel: the complexity of make is highly overrated. all it does is compare file dates...
21:17:29kugelbut it's black magic for most people
21:17:58 Quit tyfoo ("Carpe diem")
21:18:56gevaertsI should split that patch into a HAVE_MULTIDRIVE/HAVE_MULTIVOLUME patch and a multi-driver patch
21:19:48kugelgevaerts: HAVE_MULTIDRIVE/HAVE_MULTIVOLUME is supposed to be the same?
21:19:58kugelat least at above a certain abstraction level
21:20:37gevaertskugel: depends on how you look at it. At lower levels they certainly need to be split
21:21:24gevaertsThe only level where they are somewhat related is in disk_mount_all() I think. Below that you have drives, and above that you have filesystems
21:21:28kugelsure, with "above a certain abstraction level" I meant, the filebrowser doesn't seperate them
21:22:12*kugel decides to keep away from that
21:22:46gevaertskugel: if you read my comments, you'll see that this "unification" is one of the unsolved problems :)
21:23:52 Quit Seed ("cu, Andre")
21:24:04Zagorah! /me found the codec dependency bug
21:24:40Zagorplugins have the same bug
21:26:08 Quit grndslm ("Leaving")
21:26:42amiconnWhat is it?
21:27:00amiconnIt doesn't seem to affect all codecs
21:27:54Zagorthe problem is that codelib/ isn't searched for include files
21:28:06amiconnZagor: I think the problem most people (including me) have with 'make' is that processing is highly non-linear
21:30:04mcuelenaeregevaerts: now Windows seems to wait a bit, and then does the same loop again
21:30:11Zagorso the codecs that don't include codecs.h themselves (but instead rely on codeclib.h to do it) don't get that dependency
21:30:19mcuelenaereso it waits 2 seconds before sending 2F again
21:30:26gevaertsmcuelenaere: maybe try removing the ep_stall()
21:31:28amiconnPlugins usually do include plugin.h though
21:31:33bertrikmaybe the unstall after the stall doesn't work (just guessing)
21:32:12mcuelenaeregevaerts: wow, now everything is more responsive and it immediately gives 'cannot format'
21:35:07gevaertsok. I checked the UMS spec, and while I'm not sure if that stall is really wrong, it's at least optional
21:35:42*gevaerts wonders about the "cannot format" though
21:35:53mcuelenaerewell perhaps my usb driver implemented stall not entirely according to the rockbox norms?
21:36:11mcuelenaereand I can edit the RAM disk directly through 010 editor, so there's no problem with accessing it
21:36:38mcuelenaeregevaerts: if I choose Quick Format, it works
21:37:06 Join aurix_lexico [0] (
21:37:06gevaertsah ok. Maybe it tries one of the SCSI format commands in the non-quick case
21:37:26*gevaerts doesn't see the need to pretend to low-level format the device
21:37:49 Join TheSkunkMan [0] (
21:47:33 Part aptzero
21:48:39 Join tyfoo [0] (
21:49:08 Join Thundercloud [0] (
21:56:57 Quit at0m ("CGI:IRC (EOF)")
21:58:56 Join tyfoo2 [0] (
21:59:04 Quit LambdaCalculus37 (" ajax IRC Client")
22:03:45 Quit aurix_lexico ("Leaving.")
22:11:58 Join akur [0] (
22:12:05 Part akur
22:12:09 Join akur1 [0] (
22:12:12 Part akur1
22:13:20 Quit tyfoo (No route to host)
22:13:20 Nick tyfoo2 is now known as tyfoo (
22:13:48 Quit Hillshum (Remote closed the connection)
22:14:26Unhelpfulwait, pluginlib/codeclib files don't get checked for deps?
22:19:16Zagorfixing that now
22:20:48Unhelpfulif #9770 goes in, there will be pretty much no reason for core wrappers to add new content - it was only being done before to add init functions that copy rb.
22:22:53 Join akur [0] (
22:23:07 Part akur
22:23:59ZagorUnhelpful: I'm afraid I have to rename lib/bmp.h
22:24:13Zagorwill pluginlib_bmp.h offend you greatly? :)
22:25:00Unhelpfulcompletely fine with me... is it to get rid of the conflict between the two bmp.h files?
22:25:11Unhelpfuland the relative include that the conflict leads to?
22:25:31ZagorI need to make sure the dependency generator uses the right file
22:25:47Unhelpfulgo ahead and name resize.h whatever you want, as well, then. hopefully it's being deleted soon. ;)
22:26:22 Quit lazka (Remote closed the connection)
22:26:27 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
22:29:40 Quit {phoenix} (Remote closed the connection)
22:32:20evilnickUnhelpful: I meant to ask you, what's the eventual reason for moving bmpresive out of the core? (I'd give a r. number but can't get onto the Commits since 4 weeks for some reason.
22:32:50Unhelpfulevilnick: not moving, exactly, but offering in pluginlib on targets where it's not useful in core.
22:33:43evilnickSo would that speed up the usage? Or reduce the binsize? (I'm not an expert but would like to learn more than I currently know)
22:33:44Unhelpfulresizing to mono produces bad results, and besides, we don't, to my knowledge, offer albumart on anything mono - so the resizer is disabled, saving about 1KB or so for the resizer itself, and maybe another .5KB in support code in the loader.
22:34:00Unhelpfulso you're volunteering to maintain my work? ;)
22:35:36evilnickChrist no! I see my role as more of hanging out in #rockbox-community and being unfunny than really contributing much. I'm another of those users who is not prepared to learn coding. (Veering off-topic so will now shut up)
22:39:52 Join Aurix_Lexico [0] (
22:41:45 Quit BigBambi (Read error: 60 (Operation timed out))
22:43:17kugeldomonoky: ping
22:44:08amiconnmultiple .h files can lead to all sorts of problems. That includes name clashing with system includes
22:44:45amiconnErr, multiple .h files with the same name
22:50:41 Quit XavierGr (Nick collision from services.)
22:50:52 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:53:59***Saving seen data "./dancer.seen"
22:58:57 Part unstable
22:59:52 Quit bmbl ("Woah!")
23:03:17Unhelpfulamiconn: i already ran into that, long before importing resize into pluginlib, just by trying to *use* it in pluginlib
23:04:05*amiconn also ran into this when trying to fix mysterious codec compilation problems
23:04:32UnhelpfulZagor: well... if you feel better about uintptr_t, now, what do you think of the ALIGN_BUF in here?
23:04:33amiconnA number of codecs comes with a config.h, which collides with rockbox' own config.h ...
23:04:53 Quit jhulst (Read error: 110 (Connection timed out))
23:05:47Zagoramiconn: I see none
23:06:15amiconnNot anymore... I renamed them when I found that cause
23:06:20 Quit karashata ("G'bye everyone!")
23:07:02 Join akur [0] (
23:07:20 Part akur
23:07:28Unhelpfulthe uintptr_t is avoidable, though, one could convert to unsigned int, calculate addr % align, and subtract that from the original pointer, to align down
23:08:25amiconnUsing uintptr_t is the clean way, imo
23:09:05Unhelpfulthe braces seem to be effective at hiding the temps from the enclosing scope, as well.
23:10:07 Quit BXCracer_ (Remote closed the connection)
23:10:18kugelUnhelpful: variables are only valid in the block they are declared in (and in inner blocks, but they may be overwritten), but never out blocks
23:11:05Unhelpfulright. as i expected, but i sometime think in terms of python scoping, which is... different.
23:14:32 Quit saratoga ("CGI:IRC (EOF)")
23:15:35 Join Anges1 [0] (
23:21:12ZagorUnhelpful: I'm fine with that
23:21:54Unhelpfulok... it's in resize.c in that patch, which is currently the only use in core of align_buffer
23:25:59Unhelpfulwould it most likely belong with ALIGN_UP/ALIGN_DOWN, in firmware/export/system.h?
23:27:31ZagorI'd say so, yes
23:28:46 Quit igwinnimma (Read error: 110 (Connection timed out))
23:28:53kugelbtw: what's left using smooth/simple resizing?
23:30:01 Quit Anges (Read error: 111 (Connection refused))
23:30:03Unhelpfula few things, i don't think any that do it with actual bitmaps, anymore.
23:31:12Unhelpfulwith bitmap files, i mean. i think the ppm viewer plugin supports resizing
23:31:20 Quit evilnick (" byeeeee")
23:32:29kugelah, test_resize and ppm viewer
23:33:05kugeldo we consider dropping it?
23:33:11Unhelpfulsolvable by deleting test_resize? ;)
23:33:42Unhelpfulppm viewer can possibly be converted - just add recalc_dimension and resize_on_load to pluginlib
23:34:00 Quit itcheg (" ajax IRC Client")
23:35:18kugelUnhelpful: test_resize could test the core resizer
23:35:18Unhelpfuli'd really like to propose a new name for at least the latter of those functions, preferably *before* it gets more than one caller.
23:36:11Unhelpfulkugel: maybe... maybe merge it with the greylib resizer test? test the greylib plugin on greyscale targets, and test "native" scaler output on color ones?
23:36:17kugelor does the core only support resize on load, i.e. not resizing an already loaded one?
23:36:34Unhelpfulkugel: i don't like "resize_on_load" as a name, it's not really a very good description of what it does.
23:37:02kugelwell, I didn't think it can only resize on load
23:37:23Unhelpfulwhat it does is resize any image that can be fetched in chunks via a callback with the right signature. it needn't be while loading from disk, and it't needn't be bitmap.
23:37:36domonokykugel: pong ?
23:37:38kugelI used test_resize resize once, and I know you can resize it "on the fly" using the scrollwheel for example
23:38:09kugelUnhelpful: ah, that sounds more like what I thought about it :)
23:38:30kugeldomonoky: did I understand correctly, you are able to read the REC button with the code?
23:38:40kugelthe fuze's button code I mean
23:39:09domonokykugel: i dont know if i can read rec, the reading are just too unstable.
23:39:41Unhelpfulif you added an output plugin for rgb24, and a scaler frontend for it, you could load a 24-bit bitmap into memory, and scale it to various sizes, producing new native bitmaps... for example. resize_on_load seems like the kind of name that will keep devs from trying to do something like that with it.
23:39:50 Quit jhMikeS (Read error: 54 (Connection reset by peer))
23:40:19kugeldomonoky: I thought you printed the dbop values, it read like you actually could read rec
23:40:29 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
23:40:44kugelother than that, I have no idea why this code is so unstable on e200v2
23:41:26kugelUnhelpful: is it named "resize_on_load" anywhere in the code?
23:41:35kugelif yes, that would be bad I suppose
23:41:46Unhelpfulthat's the current name of the scaler function.
23:43:17kugeldomonoky: have you looked at the commit where my patch got in? maybe the guy missed something essential
23:43:17domonokykugel: i printed the values, but most of the time i only got all bits higher then 11 set to 0, so i can not say for sure if rec is in the bits > 12 i very seldom see them. Something is still wrong on e200v2.
23:43:36Unhelpfulmaybe scale_image_callback?
23:43:59kugelI guess, resize() is too short?
23:44:02 Quit flydutch ("/* empty */")
23:44:14 Quit Thundercloud (Remote closed the connection)
23:44:35kugelmaybe prefix it with bitmap_, but that's all needed imo
23:44:44 Join Thundercloud [0] (
23:45:33Unhelpfulbitmap_ or image_... i just wonder if bitmap_ would make people think too much of lcd-native bitmaps, or the bitmap file format.
23:46:13kugelwell, it can essentially only resize bitmaps (either those on the disk, or native ones in the ram), so it's correct
23:46:59kugelI think of bitmap as in uncompressed rgb565 (or whatever bit depth) image data
23:47:19Unhelpfulit can resize whatever you can feed it. the ppm viewer could quite easily include a scaler frontend, if the backend were exposed in the plugin API
23:47:21kugelupto 24bit, thatis
23:47:37kugelit cannot resize jpeg, can it?
23:48:17Unhelpfulonly if we add a jpeg loader that provides the right callback ;)
23:48:20kugel"whatever you feed it" sounds like you can mp3's or text files to it
23:49:10kugeljust to clear things up: what are the requirements to the data passed to (leaving the callback alone)
23:49:18 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
23:49:56Unhelpfulbasically, you give it an output bitmap structure, an input size, and a callback. the callback will provide 8-bit-per-channel data, either RGB or greyscale, depending on whether the target has a color LCD
23:50:33kugelso the callback is actually resizing?
23:51:25Unhelpfulno, the scaler resizes, but you don't pass it any actual image data when you call it. it calls the callback to get the source image data.
23:51:38 Join culture [0] (
23:51:43 Join ShadowGazer [0] (n=443da770@gateway/web/cgi-irc/
23:52:11 Quit domonoky (Read error: 104 (Connection reset by peer))
23:52:12kugelyes, and that feeds the scaler with uncompressed image data?
23:53:02Unhelpfulyes, but the point is that there can be more than one implementation of that callback
23:53:37kugelwell, of course
23:53:45kugelthat's the point of callbacks, isn't it? ;)
23:54:06ShadowGazerCan anyone here help me install rockbox on my sansa e280? It's version 1.0 I made sure when I bought it. The rockbox util won't recognize it. And when I try to install it manually, it'll reboot like its installing but goes right back to the main screen like nothing was installed
23:54:30kugelso, I got this all right. the scaler needs uncompressed image data (which is bitmap, to me)
23:54:32Unhelpfulso the scaler does only handle 8-bit-per-channel uncompressed data itself, yes. i suppose bitmap_scale or bitmap_resize would be ok.
23:55:02 Quit ShadowGazer (Client Quit)
23:56:30 Quit bertrik (Remote closed the connection)
23:57:26*Unhelpful wonders if mpegplayer relies at all on the current align_buffer's aligning the buffer end as well...
23:58:55 Join BigBambi [0] (

Previous day | Next day