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 2008-12-18

00:00:09MartynWhen I try using the RockBox Utility to install rockbox on my sansa e250R, I get a 404 error
00:00:20MartynIs there a problem with where files are stored for firmware?
00:00:45 Join moos [0] (i=moos@rockbox/staff/moos)
00:02:46 Quit XavierGr ()
00:03:51 Quit saratoga ("CGI:IRC (EOF)")
00:10:24kugelRockRabbit: Note that you need to re-login / start a new shell, else your PATH changes aren't active
00:10:53kugelRockRabbit: do "echo $PATH" to see if the compiler directory is in
00:14:58 Join hd [0] (
00:15:21 Quit HellDragon (Read error: 104 (Connection reset by peer))
00:15:36 Nick Martyn is now known as Martyn|away (
00:15:50 Quit hd (Remote closed the connection)
00:15:59 Join HellDragon [0] (
00:17:37 Nick Martyn|away is now known as Martyn (
00:19:23lucentheh, saratoga forgot that the maximum file size on FAT is 4.3GB
00:20:11Martyn*cough* on FAT32
00:20:22lucentright you are!
00:21:01MartynAnd it's one byte -smaller- than 4G
00:21:31MartynSo 3 gigs, 999 megs, 999kb, 999bytes
00:21:51 Quit bxcracer (Remote closed the connection)
00:23:45 Join mc2739 [0] (
00:23:51krazykitMartyn, i believe rockbox utility will not install the bootloader on the e200r, since you need to do some manual patching. the manual should have instructions on how to manually install
00:24:27BigBambiI think it just has a link to the wiki page
00:24:27BigBambiEither way, it is findable
00:25:13 Quit bmbl ("Woah!")
00:28:58 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111319]")
00:29:25 Join kugel [0] (n=chatzill@unaffiliated/kugel)
00:31:40 Quit Martyn ()
00:34:56Aurix_LexicoI don't know if this is entirely on topic, but I want to share it, it's a screen capture of a turret defense plugin I've been playing around with
00:35:10Aurix_Lexicothe boxes represent the path the enemies follow
00:35:19 Quit jgarvey ("Leaving")
00:35:19 Quit faemir (Remote closed the connection)
00:36:14Aurix_Lexicothe path finding clearly needs some improvement
00:39:54***Saving seen data "./dancer.seen"
00:41:11 Join faemir [0] (
00:41:35 Quit Zagor ("Clint excited")
00:44:06 Quit RockRabbit ("CGI:IRC")
00:45:11lucentAurix_Lexico: I'm looking :)
00:48:24 Join ThOr101 [0] (
00:48:53ThOr101Holy nutz, what an awesome piece of software. I've been using it less than 24 hours, and it just FLIES!
00:49:18 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
00:50:23ThOr101I do have one question though (of course). I couldn't find it in the iPod or Battery FAQ. In the documentation (ipod 5.5 80G) General Set -> Battery Set -> Charge During USB Connection. I don't see that on my build. Any clues?
00:50:26lucentThOr101: thanks for that, I think #rockbox-community might share your excitement
00:50:39lucentthis is kind of the developer channel ;)
00:51:06ThOr101Ahh, well, thanks then, and other than my Q, I'll keep the noise down. :-)
00:51:14ThOr101Or should I ask my Q there?
00:51:16lucentI don't own an iPod so I can't really comment on your other question
00:51:48krazykitlucent, no, this is a development AND support channel
00:52:17krazykitThOr101, are you using an official build?
00:52:26ThOr101Yessir. I used the auto installer
00:52:45ThOr101Which other than the well documented themes burped, worked flawlessly.
00:53:24krazykitThOr101, well, just wait around awhile and perhaps somebody with an ipod (or who knows the answer) will see your question
00:53:42ThOr101sounds good. Thanks :-)
00:55:14ThOr101I may have answered my own Q. When I went to RB info page to get the version, it has Battery: Charging (It is currently plugged in - held down menu to maintain RB state). So it does appear to be charging
00:58:08 Quit tyfoo ("Carpe diem")
01:02:20 Quit aneqrs ()
01:02:57 Join webguest26 [0] (n=4a39cdc8@gateway/web/cgi-irc/
01:03:04 Quit webguest26 (Client Quit)
01:13:17 Quit herrwaldo ("Konversation terminated!")
01:16:29 Quit faemir (Remote closed the connection)
01:19:35 Quit ThOr101 ("ThOr101 has no reason")
01:20:09 Join Thundercloud [0] (
01:25:06 Quit alexbobp ("leaving")
01:26:49 Join RockRabbit [0] (n=3aac9a01@gateway/web/cgi-irc/
01:30:00RockRabbitI'm getting errors using to build the development tools. Here is the output Anyone any idea what this might be? Lucent - this is not the error we got before - this is using my previous install of rockbox.
01:30:59 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
01:35:40lucentRockRabbit: it's a common error, quite google-able
01:38:04 Quit moos ("Rockbox rules the DAP world")
01:38:28lucentRockRabbit: i.e.
01:39:15lucentmy host compiler is gcc 4.3.2 btw
01:39:26lucentRockRabbit: what's your host compiler version? gcc −−version
01:40:39 Quit PaulJam (".")
01:42:38RockRabbitgcc (Ubuntu 4.3.2-1ubuntu11) 4.3.2
01:45:47RockRabbiti see theres a patch - but i have no idea how to apply it. Will the fix be backported to ubuntu as part of a normal update?
01:47:58kugelRockRabbit: I'm running ubuntu 8.10 x64 just fine did well
01:48:09 Quit Schmogel (Read error: 104 (Connection reset by peer))
01:49:02RockRabbitOdd. How is that possible, unless you applied the patch, or perhaps you have not applied all the updates to ubuntu?
01:49:31kugelI installed the toolchain 2 weeks ago or so. I had all updates at that time
01:50:00kugelmaybe something in svn broke it. You might try an older revision
01:50:24toffe82_I install it on ubuntu 8.10 and no problem too
01:50:46 Quit n1s ()
01:51:03kugelRockRabbit: did you make sure you have the build-essential package installed? if not, install it
01:51:12RockRabbitmmmm. im baffled. We both have the same OS, yours works and mine does'nt. I did get this all working a few weeks back on either Ubuntu or Fedora (can't remember which), so its even more puzzling.
01:51:14toffe82_with the last update of ubuntu last week
01:51:32 Quit lasser ("ChatZilla 0.9.84 [Iceweasel 3.0.4/2008112309]")
01:51:57kugelRockRabbit: as I said, it might be possible that is broken in svn. I recommend obtaining an older version to see if it works
01:52:19RockRabbityes i have build-essential. As far as I know I followed the exact same steps to install rockbox this time as i did a few weeks back, and then it worked without a hitch.
01:52:20kugeljust type "svn up -r19000" e.g. in the rockbox source folder, then try again
01:53:05RockRabbitsorry kugel i missed that previous comment. ok, ill give that a try.
01:55:10 Quit dfkt_dt (Read error: 54 (Connection reset by peer))
01:55:50 Part toffe82_
01:57:03RockRabbitcame back with "Skipped ."
01:57:38lucentRockRabbit: I'm inclined to say that either your hardware or your OS is hosed
01:58:12RockRabbitive got my fingers crossed its the os
02:00:17pixelmaI see "Skipped." when I accidentally try to update while in a wrong directory
02:01:58kugelRockRabbit: skipped is wrong
02:02:42kugelyou probably were in the wrong directory, as pixelma said
02:04:45RockRabbitis was in the rockbox-3.0 directory. Anyway, ive deleted the directory and am re copying from svn using "svn co .... -r19000"
02:04:54 Quit maddler (Read error: 110 (Connection timed out))
02:05:42kugelRockRabbit: oh, you're having the 3.0 source? so not from svn?
02:08:52 Quit Thundercloud (Remote closed the connection)
02:10:44RockRabbitYou have a point - im really not too sure how I got it. Ive been doing so much trial and error stuff over the past few weeks, and being new to Linux, Rockbox, svn, arm assembler, etc, the past few weeks have been a blur.
02:11:16RockRabbitive had about 8 different os's on the pc in the last few weeks too.
02:14:28RockRabbitthe memory just clicked into place. I got rockbox from a 7zip file - the release version, and i used svn to get the mkamsboot code.
02:15:30RockRabbitand the rockbox that i got up and running some weeks back was almost definately from svn
02:17:31kugelthe 7z archive *should* work, but you're definitely save(r) if you get the source from svn
02:17:47kugelthat's the recommended way anyway, it's much more comfortable too
02:18:04kugelRockRabbit: 8 different OSes sounds impressive :)
02:18:12lucentRockRabbit: don't worry too much, everyone messes up their first distro install or two, I mean *EVERYONE*
02:18:38lucentRockRabbit: but like, it's tough for someone to walk up to a hosed system and make it work again
02:18:42kugellucent: hey, I failed for now with the buttonlight
02:19:02lucentkugel: how do you disassemble the OF code anyhow?
02:19:04RockRabbitmy first distro! i think its no exaggeration to say I have installed Linux more than 300 times at least!
02:19:07kugelI'm not sure how the OF does it, but I can't get it to work parallel to microsd
02:19:25 Quit jhMikeS (Nick collision from services.)
02:19:31 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
02:19:32lucentkugel: can you make it light up (not paying attention to keeping microsd working) ?
02:20:02kugellucent: light you up in which way? How I disassembled?
02:20:19lucentkugel: I'm confusing two different questions
02:20:23 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
02:20:36lucentkugel: one question is "what tools do I need to disassemble Sansa AMS OF?"
02:20:58lucentthe other question is "are you able to trigger on/off the button light on Fuze at all?"
02:21:06kugelI used arm-elf-objdump (that should be installed with the toolchain)
02:21:42 Quit robin0800 (Read error: 54 (Connection reset by peer))
02:22:26kugellucent: yes I am able, but it's on GPIOD, and the microsd is basically taking the whole D port, so it isn't trivial to get microsd and buttonlight working simultaneously
02:23:14lucentkugel: microSD isn't working on my hardware anyways, I should probably try out the code and see if it works here also
02:24:09kugellucent: get it working is really easy (swap the lines in backlight-e200v2-fuze.c), as DIR needs to be set before the PIN
02:24:31kugelbut as soon as you use a microsd, you'll only have problems :)
02:25:16kugelwhat I disassembled today wasn't sufficient obvioisly, I guess the OF handles that stuff more in the SD driver
02:25:56kugelwhich I didn't/couldn't disassemble due to thumb/arm code switching, and objdump doesn't seem to be able to have both in one disassembly
02:27:01*lucent googles what is thumb code
02:27:28kugellucent: thumb code is code of reduced size. It uses 16bit instructions instead of 32bit ones.
02:28:14kugelso it's very small, but relatively slow, as you need to do more instructions to achieve the same as with arm code at times
02:28:36lucentah okay
02:28:49lucentI'm following this help okay, thanks
02:29:14lucentI only did a little assembly work for Z80 CPU on a graphing calculator
02:29:29kugelit's a start :)
02:29:29lucentthat's all I have known
02:29:40kugelthe disassembly I made today was the first I did on my own
02:36:33*kugel thinks he got it to work, somewhat
02:39:58***Saving seen data "./dancer.seen"
02:42:36 Nick fxb is now known as fxb__ (
02:51:08 Quit Lynx_ (Remote closed the connection)
03:04:20 Quit perrikwp (" ajax IRC Client")
03:15:04kugellucent: FS #9663
03:15:18kugelI go to bed
03:15:20 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111319]") worked fine with the r19000 rockbox source from svn
03:15:35 Quit XavierGr ()
03:24:40 Quit MethoS- (Remote closed the connection)
03:35:42 Nick Bensawsome is now known as Bensawsome-11111 (n=Bensawso@unaffiliated/bensawsome)
03:35:58 Nick Bensawsome-11111 is now known as Bensawsome (n=Bensawso@unaffiliated/bensawsome)
03:46:21 Join massiveH [0] (
04:05:17 Quit ch4os (Read error: 60 (Operation timed out))
04:05:34 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
04:05:36 Quit Acky (Read error: 104 (Connection reset by peer))
04:05:58 Quit massiveH ("Leaving")
04:06:33 Join alexbobp [0] (n=alex@
04:13:45 Quit Outerlimit ()
04:21:38 Join Ryco [0] (
04:24:06lucenthm, how did kugel get that disassembly in
04:24:26lucentarm-elf-objdump -D −−target binary -marm fuzea.bin-01.01.22-OF −−start-address=0x400 -Mforce-thumb
04:24:29lucentdoesn't match up
04:26:05Aurix_Lexicough, I'm using printf a lot, but whenever it reaches the bottom, it just starts again at the top and looks really ugly
04:26:09lucentit starts to match up at offset 0x49b0 though
04:26:11Aurix_Lexicois there any way to fix that?
04:26:29lucentAurix_Lexico: hm, logf maybe?
04:26:29Aurix_Lexicoin my little hacked up bootloader
04:26:37lucentview from the debug menu logf viewer
04:27:01lucentAurix_Lexico: look at this guy's code
04:27:11lucenthope I don't waste your time, I think it may be helpful though
04:29:02 Join blkhawk- [0] (
04:29:49Aurix_Lexicook, I guess something like that would work
04:30:06 Part Ryco
04:30:21 Quit blkhawk (Read error: 60 (Operation timed out))
04:30:59 Nick blkhawk- is now known as blkhawk (
04:35:58vertic23bah someone from U.S. here?
04:37:22advcomp2019vertic23, why?
04:37:58vertic23because I need an ipod video 80 GB 5.5 :>
04:39:15vertic23..and soon there is C25C3 maybe someone's coming around and get my ipod so I don't need to pay customs
04:39:59***Saving seen data "./dancer.seen"
04:40:26 Join Rob2223 [0] (
04:44:07 Join thegeek_ [0] (
04:44:24 Join mc2739_ [0] (
04:45:11 Quit jhMikeS (Nick collision from services.)
04:45:17 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
04:45:53 Quit mc2739 (Nick collision from services.)
04:45:54 Nick mc2739_ is now known as mc2739 (
04:47:22 Quit miepchen^schlaf (Read error: 60 (Operation timed out))
04:47:55 Quit Rob2222 (Read error: 104 (Connection reset by peer))
04:51:10 Quit RockRabbit ("CGI:IRC (EOF)")
04:51:13 Quit Rob2223 (Read error: 60 (Operation timed out))
04:55:29 Join Rob2222 [0] (
04:56:17 Join b0b [0] (
04:56:36 Quit b0b (Client Quit)
04:56:50 Join UnhelpfulWeb [0] (n=a49c99e6@gateway/web/cgi-irc/
05:02:48 Quit mc2739 ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
05:02:50 Quit thegeek (Read error: 110 (Connection timed out))
05:07:25 Quit esthar (Read error: 104 (Connection reset by peer))
05:14:13 Quit JdGordon (Remote closed the connection)
05:14:28 Join Rockrabbit [0] (n=3aac9a01@gateway/web/cgi-irc/
05:15:05RockrabbitAm I able to get write permissions for the twiki?
05:15:40RockrabbitI want to be able to update the Sansa c200 hardware mappings
05:16:10 Join JdGordon [0] (
05:19:37jhulstRockrabbit: I'd be happy to, but I'm assuming Rockrabbit isn't your real name. Did you read the instructions when you registered?
05:19:56jhulstRockrabbit: Specifically the part about using your real name?
05:20:16 Quit Horscht (Nick collision from services.)
05:20:31 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
05:20:38 Join Horschti [0] (
05:20:52 Quit Horscht (Nick collision from services.)
05:21:04 Join Horscht [0] (
05:22:50 Quit Horscht (Client Quit)
05:23:24 Join Minthe [0] (
05:23:32 Quit Minthe (Client Quit)
05:23:44 Join Minthe [0] (
05:27:25 Quit Minthe (Client Quit)
05:27:38 Join Minthe [0] (
05:28:06RockrabbitI sed my real name when registering on the twiki just now. No I did not notice anything about having to use a real name on the IRC or the forums
05:28:50RockrabbitMy real name is phil taylor
05:29:23 Quit Horschti ("I am root. If you see me laughing, you better have a backup")
05:29:32jhulstRockrabbit: You don't need to use your real name on IRC or the forums, but it looks like you registered with RockRabbit as your wiki name, when it should be PhilTaylor
05:30:01Rockrabbitsorry i thought that was the usual nickname
05:30:17Rockrabbiti assumed it should match all my other nicks so youd know it was the same person
05:30:35Rockrabbitwhat should i do?
05:31:23jhulstRockrabbit: From the registration page, "Your WikiName must be your real first and last name concatenated"
05:31:35jhulstReregister with the correct name and I'll give you access
05:31:38UnhelpfulWebreregister with PhilTaylor as our WikiName, i'd say
05:33:04UnhelpfulWebas far as making sure that people know you on IRC, you can make an entry here for that:
05:33:51lucentvertic23: I'm in Jackson, WY USA
05:34:06lucentvertic23: I'd like to help if it makes a difference
05:34:16vertic23are you coming to the C25C3 in berlin? :>
05:36:03lucentvertic23: oh, no I don't think I am
05:36:43vertic23well, but I just realized that the ipod 5.5 with 80GB is almost as expensive here in .de as in .us
05:36:57vertic23hard to get them nowadays
05:37:15vertic23and I just need a rockbox player ...mine is broken
05:37:21lucentah, :/
05:38:10lucentthe only player I know in production which sort of runs Rockbox is like my Sandisk Sansa Clip, or Sandisk Sansa Fuze... but these are active development targets
05:38:11jhulstRockrabbit: You should be all set now
05:38:13scorchethere are much better rockbox targets than the 5.5th gen ipod
05:38:16lucentI wouldn't suggest them for everyday
05:38:43scorchelucent: and as such are not supported and shouldnt really be advertised as "working" lest people get the wring idea ;)
05:39:45Rockrabbitif i accidentally post some wrong info to the hardware mappings page, and someone else uses it and bricks their player because of my error, will people stop liking me?
05:40:24lucentRockrabbit: that's none of my business, except to revert the changes if I happen to know they are wrong
05:40:48RockrabbitDoes anyone know about addressing the GPIO ports?
05:40:59vertic23scorche: tell me about it
05:41:46vertic23I am in the "BuyersGuide" ...but I don't see any players which are still available
05:41:56scorchethat is what ebay is for ;)
05:41:58vertic23...and would fit my need
05:42:01Rockrabbitif i use the code "ldr r1, [r0, #0x8] " to read a pin, does the hex value 8 refer to pin number 5?
05:42:14vertic23well ...I looked - none of the ones I'd like are on ebay
05:42:25vertic23I looked for the M3L (iaudio) e.g.
05:43:21lucent#0x8 is probably 1<<4 if I'm not mistaken
05:43:39lucentwhich would be pin 4. but I really haven't any clue how to deal with hardware designs
05:43:40 Quit Aurix_Lexico (Operation timed out)
05:43:56 Join n9xvt [0] (n=187bed27@gateway/web/cgi-irc/
05:44:30n9xvtanyone built or building thier own hardware yet?
05:45:02lucentn9xvt: not that I know of with Rockbox in mind
05:45:18lucentn9xvt: would be a reasonable platform though
05:46:15n9xvtok,,i found a complete platform,,for the diy ,,thought you guys might wanna look,,
05:47:08n9xvtthis guy has built his own and many other items,,
05:47:25UnhelpfulWeblucent: wouldn't 0x8 be 1<<3?
05:49:25lucentUnhelpfulWeb: hm... 1 2 4 8 16 32 64 128
05:49:59lucent00000001 oh...
05:50:31 Quit wpyh (Read error: 110 (Connection timed out))
05:50:34lucentUnhelpfulWeb: yeah I bet you're right on that one
05:50:57scorchen9xvt: shame it is a HWcodec device...
05:51:12Rockrabbityes i agree - and that fits with the formula i was given that the pin is addressed by (1 << (x + 2)) where X is the pin number. Which would make 8 = pin 5. If thats true, im happy.
05:52:09vertic23scorche: ...still waitin' for your hints :>
05:52:25scorchevertic23: sorry, but i am quite busy at the moment...
05:52:46vertic23okay, I'm waiting then
05:52:49vertic23how long? :P
05:53:10scorcheif you wish to wait for me, likely about 4 hours
05:53:21vertic23will you think of me?
05:59:08 Quit n9xvt ("CGI:IRC (EOF)")
05:59:31 Quit amiconn (Read error: 110 (Connection timed out))
05:59:31 Quit pixelma (Read error: 110 (Connection timed out))
05:59:47 Join pixelma [0] (n=pixelma@rockbox/staff/pixelma)
06:00:33 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
06:02:30UnhelpfulWebgoing to start on converting pictureflow for scaled loads. thoughts on what should be the "right" size to use? i'd say the largest square to fit on LCD * (zoom%), but that would mean a cache rebuild whenever we change the zoom setting.
06:03:11UnhelpfulWeblogging load times on e200 suggests to me that at least the 80MHz PP targets are not fast enough to do away with the pfraw cache :/
06:12:00 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
06:13:23lucentUnhelpfulWeb: IMO a patch first that does what you'd like, and then hack it to be fast enough?
06:19:36UnhelpfulWeblucent: quality scaling is highly unlikely to be fast enough to drop the cache any time soon. i'm not sure just how fast PF will let you scroll, but it takes about .5s to scale down a 256x256 image to fit on the e200 screen
06:21:19UnhelpfulWebthe only way i can think of to get "rid" of the cache is to cache albums "near" the currently-viewed in the background, which of course has the caveat that if the user tries to scroll quickly right away, they will see lots of "no cover" images
06:29:32UnhelpfulWebThe question i'm asking is not so much "how do i use the scaled loader in PF" as it is "how big do i want the covers to be"... the wiki page for the plugin suggests using 100x100 cover arts, but clearly we don't want to just force a size in the source.
06:29:51 Quit sbhsu (Read error: 104 (Connection reset by peer))
06:30:02 Join sbhsu [0] (
06:31:16 Quit Minthe ("Leaving...")
06:33:58 Join lee321987 [0] (
06:35:02lucentUnhelpfulWeb: maybe half the largest dimension of the display, square, unless that doesn't fit on the display
06:35:10lee321987Is RB working on any DAPs that are currently in production?
06:35:46lucentlee321987: no supported release of rockbox runs on DAP currently in production.
06:36:04lucentlee321987: it's been that way for some years now I think
06:36:46lucentlee321987: need more programmers who are willing to tackle the SanDisk AMS AS3525 based players (e2x0 v2, Clip, Fuze)
06:37:20lee321987what are the c200 v2's based on?
06:37:48lucentoh, I don't know off my immediate memory
06:39:34lee321987thanks. TO ALL ROCKBOXERS - Have a merry Christmas/Hanukkah/ect... Peace
06:39:42 Quit lee321987 ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
06:39:46 Join Martyn [0] (
06:39:48UnhelpfulWeblucent: the source has a "PREFERRED_IMG_WIDTH" defined to 50 if LCD_HEIGHT < 100, 100 otherwise. perhaps i should just set that to MIN(LCD_WIDTH,LCD_HEIGHT)/2, and scale to that on load.
06:40:03***Saving seen data "./dancer.seen"
06:41:13lucentUnhelpfulWeb: that will make some people angry, but I would like to think that makes more sense
06:41:38lucenteven though that means it's still hard-coded
06:44:21UnhelpfulWeball of my album art is maximized, it never even occurred to me that i "should" be able to see other album cover except when it animated. like i said, we could base it on the zoom setting, but 1) the zoom setting is apparently supposed to be about changing the camera distance 2) that means cache rebuild on a setting change
06:46:49 Quit Martyn ()
06:47:10lucentI think I need a free reference introduction to ArM
06:47:16lucentARM*... suggestions?
06:48:08J-23ARM C programming?
06:48:10J-23or ARM hardware?
06:48:50UnhelpfulWeb might cover what you need
06:48:55lucentmy focus is to learn about assembly so I can disassemble code
06:50:14*lucent pokes download
06:51:00*J-23 downloaded it in four seconds
06:51:44UnhelpfulWeblucent: looking to disassemble some OF code? or more to read what the compiler turns rockbox source into, for optimization purposes?
06:52:06lucentUnhelpfulWeb: disassemble some OF code
06:52:41lucentSo, I need to brush up on ARM instructions and Thumb
06:53:06lucentJ-23: nice link, thanks. I'm on a slow link at the moment hence it was a bit of a wait for me
06:54:37lucentI need a couch or something to kick back on and read this 156 page guide
06:55:38Rockrabbitc200's are AMS5325
06:55:50Rockrabbitv2 that is
06:56:09lucentRockrabbit: huh? not ams as3525?
06:56:16UnhelpfulWebthis cheat-sheet might also be helpful:
06:57:10Rockrabbiti guess i got the numbers mixed up
07:01:04lucentUnhelpfulWeb: so um, BL is like jump with return?
07:01:20lucentnot familiar with the term "branch"
07:02:32UnhelpfulWeboh, geeze, it's been ages since i really knew anything about asm... "branch" instructions that i remember were more equivalent to goto
07:03:35lucentwell what I remember was for Z80 CPU on Texas Instruments TI-83 graphing calculator
07:03:56lucentwe had things like JP, JNZ, etc.
07:21:02UnhelpfulWeblucent: i think that what i'm thinking about doing with PREFERRED_IMG_WIDTH is pretty close to what was intended, actually - it looks for a cover.NxN.bmp for that size, and loads that if found. so the big change here would be using the preferred size regardless of whether the user has already provided a BMP scaled to that size, and defining the preferred size based on a ratio to screen size, rather than as a stepwise-constant function
07:23:18*lucent :)
07:24:16UnhelpfulWebalso, i don't get to say "stepwise-constant" very often.
07:25:01lucentUnhelpfulWeb: in where the hell does "buttonlight(bool on) [r0]" come from?
07:25:30lucentUnhelpfulWeb: is that something kugel (the human responsible for that disassembly snippet) added, or debug info from OF?
07:26:48 Join einhirn [0] (
07:26:48 Quit einhirn (Client Quit)
07:27:01 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
07:27:08saratogalucent: reading the logs, what did you think I forgot?
07:27:20UnhelpfulWebit could possibly be the latter. are you sure the snippet comes from the OF, and not some rockbox bit that he's debugging?
07:28:31lucentsaratoga: shoot, it was just some smart-alecky comment by me... you suggested to write a zero-data file to fill out the FAT filesystem
07:29:00lucentsaratoga: 'cept well, filesystem is 8GB FAT32, largest filesize is 1024*1024*1024-1
07:29:09lucentmy math is wrong
07:29:15lucentit is just shy of 4GB
07:29:18saratogalucent: I assumed you knew how to create multiple files . . .
07:29:25lucentsaratoga: I'm teasing :)
07:29:35saratogathis was worth asking me why exactly?
07:30:30lucentsaratoga: no worries!
07:31:03lucentsaratoga: I'm still confused though, I have a track preloaded on my Sansa Fuze, it is named "The Red Thread - Wax Museum"
07:31:03saratogaplease don't ping me in the logs if you dont' need me
07:31:37lucentafter deleting MTABLE.SYS (which contained the text of the name of the track) and filling up the FAT32 filesystem with zero-data files, and then unmounting
07:31:54lucentI grep through and there's still references to "Red Thread" and the song metadata
07:31:59lucentare there hidden files?
07:32:23lucentI'm not familiar with the Linux kernel vfat driver to know if hidden files show up or what is going on
07:32:28saratogadoes the file play?
07:33:51lucentsaratoga: the file doesn't show up when I mount the FAT32 filesystem
07:34:15saratogabut does it play?
07:34:15lucentsaratoga: I'm confused if it's just leftover filesystem metadata, or if there's hidden data in the FAT32 filesystem that the vfat driver doesn't show me
07:34:21lucentplay... where?
07:34:30saratogawhere ever you saw it
07:34:56lucentI saw the filename in a hexeditor of the dump of the /dev/sdb
07:35:11lucentI'm working on a local copy of that dump
07:35:16saratogahow should i know
07:35:30 Quit JdGordon (Read error: 104 (Connection reset by peer))
07:35:33lucentthe question is about FAT32 and vfat though, maybe you know?
07:35:49lucentI haven't had to interact with hidden files before via vfat
07:35:53UnhelpfulWebthis should probably move to -community?
07:36:44 Quit saratoga ("CGI:IRC (EOF)")
07:40:24 Join Acksaw [0] (
07:42:06mud-rbthis is probably a stupid question, but if my plugin spends all of its time waiting for button presses (ex. rb->button_get_w_tmo), then i don't have to worry about yielding, right?
07:48:07 Quit nplus (Read error: 110 (Connection timed out))
08:02:14 Quit __lifeless (Read error: 60 (Operation timed out))
08:05:35 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
08:09:04 Quit BHSPitLappy (Remote closed the connection)
08:10:19 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:11:53 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:12:43 Join Rob2223 [0] (
08:14:05 Join _lifeless [0] (n=lifeless@
08:16:26 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:17:25 Quit pondlife (Read error: 54 (Connection reset by peer))
08:27:05 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:27:44 Quit BigBambi (Read error: 148 (No route to host))
08:31:08 Quit Rob2222 (Read error: 110 (Connection timed out))
08:32:10 Join ender` [0] (
08:39:52 Join Hillshum_ [0] (n=4ba5f199@gateway/web/cgi-irc/
08:40:04***Saving seen data "./dancer.seen"
08:41:40 Quit Hillshum_ (Client Quit)
08:43:17 Join maddler [0] (n=maddler@
08:49:08 Join tyfoo [0] (
08:49:34 Quit tyfoo (Read error: 131 (Connection reset by peer))
08:51:03 Quit GodEater (" ajax IRC Client")
08:53:22 Join tvelocity[a] [0] (n=tony@
08:55:47 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
08:56:27 Join Hillshum_ [0] (
08:58:25lucentuhh... okay so I figured out the story on FAT32 filesystem of Sansa Fuze
08:58:50lucentthe "##MUSIC#" and so on directories are stored weirdly, I think they are hidden attribute or something
08:59:06lucenttheir attribute is 0x18 but most normal directories are 0x10
08:59:22lucentanyways, going in with a hex editor and changing 0x18 attribute to 0x10, they are now readable
08:59:32Hillshum_thats been in the thread for awhile
09:00:19*lucent looks sad
09:00:27lucentI couldn't find it, which URL please?
09:00:39lucentI'm not happy to re-invent the wheel on this
09:00:53Hillshum_some where in
09:02:03 Join GodEater__ [0] (i=c2cbc962@gateway/web/ajax/
09:02:04 Quit GodEater (" ajax IRC Client")
09:02:18 Join tvelocity[away] [0] (
09:02:58 Quit GodEater__ (Client Quit)
09:03:08 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
09:08:47 Quit tvelocity (Connection timed out)
09:10:50 Quit Rockrabbit ("CGI:IRC (EOF)")
09:14:12 Join petur [50] (n=petur@rockbox/developer/petur)
09:16:35lucentHillshum_: I see now, it's an abuse of ATTR_VOLUME_ID
09:16:58*Hillshum_ has no clue about that
09:18:20 Join Martyn_ [0] (
09:18:29 Quit tvelocity[a] (Read error: 113 (No route to host))
09:19:58Hillshum_will my flyspray message just show up in pidgin?
09:22:03 Quit Hillshum_ ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
09:22:31 Quit Martyn_ (Client Quit)
09:22:46 Join Martyn [0] (
09:26:24lucentmine haven't
09:26:37lucentI specified a Gtalk username though which may be quirky
09:28:20MartynSanDisk is about to release a WiFi enabled Sansa player
09:28:35MartynAs soon as that happens, I'm going to be a very, very happy RockBoxer
09:28:37B4gderisn't sansa connect that?
09:28:54MartynSort of, but the sansa connect has too small a screen, and too little capacity
09:29:11B4gderright but you said "a WiFi enabled Sansa player"
09:29:21MartynThis will have 16Gb flash, WiFi G networking, FM tuner (always a plus), and a nice big screen
09:29:40MartynB4gder : True, true.
09:29:48B4gderit'll be interesting to see if they still use AMS
09:30:02MartynSomehow, I doubt it.
09:35:55lucentthe connect ran linux, don't think it was AMS
09:36:24MartynYep, but they might move to VxWorks
09:36:32MartynSanDisk just made a deal with Wind River Systems
09:36:39*lucent lets out a hoot "yay!"
09:36:51scorcheas a note, speculation of this sort should go in #rockbox-community...
09:36:58lucentI'm hooting about something else though
09:37:34lucentHillshum: thanks for pointing me in the direction of that forum post, I made a similar patch to libfat code and got read access to that "hidden" stuff
09:39:11B4gderwell, wind river does a lot of linux too these days
09:39:18Martynscorche : Sure, but it's late at night :)
09:39:21 Join n1s [0] (n=nils@rockbox/developer/n1s)
09:40:01MartynIn any case, I'm working on a new module and theme but have run into a brick wall as far as widget control goes.
09:40:20scorcheMartyn: that doesnt mean that the channel guidelines go poof...also, not everyone in here is in the US
09:40:21MartynI want to bring a very "slick" interface to RockBox, and have the UI experience to do it.
09:40:30Martynscorche : Yes, I know :)
09:40:49B4gderslick interface in rockbox? yuck! B-]
09:40:59*B4gder hides
09:41:29LinusNwe have a reputation to live up to
09:41:43MartynYeah yeah ...
09:41:53B4gderbags of ... and all
09:42:01 Quit kachna|lappy (Read error: 113 (No route to host))
09:42:28MartynWhat I wanted to ask was "Is it better to develop for Stable, or for Daily?"
09:42:36LinusNour mission statement: "Develop the most complete bag of shit featurewise"
09:42:41MartynIf I'm working on low-level widgets, and such?
09:42:45B4gderMartyn: go against svn at once, I'd say
09:42:53B4gderit moves
09:43:06 Quit Hillshum (Read error: 110 (Connection timed out))
09:46:27lucentLinusN: are we changing it from "Release stable versions only when no hardware is in production that will run stable release" ?
09:46:29 Join Thundercloud [0] (
09:46:46LinusNlucent: :-)
09:47:14LinusN"Develop the most feature-filled bag of shit for outdated players"
09:47:24B4gdera mission statement!
09:47:41lucentI committed a wiki edit that links to the forum post mentioning "hidden" directories btw, on the SansaV2 page
09:48:03pixelmaslick is hopefully not the only concern... I'd hope it's still usable, seeing what webdesigners seem to find "cool" these days
09:48:16lucentI'm almost convinced it's in the wrong page, but I don't talk much to pre-AMS Sansa owners
09:50:32 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:53:14linuxstbMartyn: Which device(s) are you running Rockbox on?
09:55:03 Join nkf [0] (
09:55:55 Quit nkf (Client Quit)
09:55:56 Quit Galois (Read error: 104 (Connection reset by peer))
09:56:09lucentLinusN: I guess I could ask, if you would mind adding 'Sansa AMS' to the flyspray list of targets
09:56:32lucentLinusN: or Fuze + Clip + .... if it matters
09:56:37 Join Galois [0] (
09:57:22Martynlinuxstb : Right now? ipod, sansa e250
09:57:32MartynI figured I should start with the most limited devices first
09:57:39MartynI can get my hands on an archos and such later
09:57:55linuxstbWhich ipod?
09:58:17Martyn5.5 (video)
09:58:32Martyne250R by the way .. loading the bootloader was an absolute PAIN
09:58:46linuxstbGood ;) The 5.5g has a nice big LCD and relatively slow CPU.
09:58:51MartynThe "rockbox utility" app can't find the files it needs .. 404's all over the place
09:59:07Martynlinuxstb : Yep, that's why I picked it. Got it on eBay for $50
09:59:33Zagorlucent: we normally don't want bug reports on ports in progress
09:59:35MartynThe sansa e250R was free(ish) .. got it in a Woot Bag-of-Crap
10:02:04lucentnice score from WOOT
10:03:21MartynYep, got a couple music players in the last few BoC's
10:03:31MartynThe Sansa is the only one that will run RockBox
10:03:34MartynThe Fuze won't
10:03:54MartynAnd for whatever reason, I can't load it on the iRiver U series
10:04:26lucentwell, the Fuze doesn't run Rockbox very well, yet
10:04:46MartynWill the connect run it?
10:04:56lucentha, connect? I haven't any idea
10:05:05linuxstbMartyn: The website has a handy list of supported targets at the top of the front page...
10:05:18 Quit jhulst (Read error: 113 (No route to host))
10:05:48MartynI saw :)
10:10:22ZagorI think it would be nice if (at least) rockbox devs filled in their real name in their irc client. I keep having trouble connecting list mails to the person in irc.
10:12:19B4gderMartyn: there's been some work on the Connect and I think the wiki page on it has the details
10:21:07MartynFound it.
10:21:24MartynThat's why I use "Martyn" .. real names are easier
10:21:41MartynWell, one thing I can't complain about is the organization of the SVN code ...
10:21:57MartynIt's all rather well laid out, and will make updating the widgets quite easy for me.
10:23:45MartynI'm guessing the lack of certain optimizations in widget processing are due to having to be so cross-platform?
10:24:21B4gderMartyn: well, lots of things just never was done...
10:24:57B4gderbut yes, the many targets are somewhat problematic
10:25:20GodEaterZagor: isn't that what the irc nicks page in the wiki is for ;)
10:26:36ZagorGodEater: I was waiting for that :-) it's a hassle having to cross-reference that all the time, using the real-name field is a lot more helpful.
10:28:06 Join wpyh [0] (n=william@
10:29:21lucentwpyh: thanks for the suggestions earlier about Fuze target
10:29:57 Join kachna [0] (
10:30:51 Join beta2k_ [0] (
10:31:06 Join bodymind [0] (n=bodymind@
10:31:27 Quit Thundercloud (Remote closed the connection)
10:31:32wpyhlucent: was it useful?
10:33:00lucentwpyh: I'm still doing a lot of catch-up just on the little things, like I only now understand that the "##MUSIC#" folder is a directory on the internal FAT32 with a weird attribute so the Linux vfat driver omits it... and I'm going to keep trying best I can on other things ;)
10:33:12 Quit beta2k (Read error: 104 (Connection reset by peer))
10:33:17lucentsleep time for me now though
10:34:16 Quit Seed ("cu, Andre")
10:36:19 Join lasser [0] (
10:37:25*wpyh thinks the svn layout is a bit weird, but won't complain either
10:37:43MartynNot -that- wierd, I was able to find the code I need to edit readily enough.
10:38:05 Quit tvelocity[away] (Remote closed the connection)
10:38:30wpyhI guess it's not modular enough
10:40:02B4gdernot enough for what?
10:40:07***Saving seen data "./dancer.seen"
10:46:14wpyhfor an pluggable alternative GUI's to be written easily ;) (I guess someone will bash me for this)
10:46:45B4gderI would say having such a modular thing with no alternative GUI around is stupid
10:47:09B4gderwe normally don't do things "just in case"
10:47:17wpyhwell, alternative GUI's are a big no-no for some devs, so I understand
10:47:40GodEaterI've never heard that
10:47:44B4gderI'm not arguing against an alternative, just that preparing the code in advance for it isn't a good idea
10:47:48GodEateronly no-one has ever bothered to write one
10:47:57*wpyh checks the logs
10:48:24B4gderif the effect on the code isn't too big I certainly won't mind it
10:48:56 Join Rockrabbit [0] (n=3aac9a01@gateway/web/cgi-irc/
10:49:25*GodEater does care about the effect on the code much - only about the effect on the target / bin size
10:49:52wpyhwell, someone asked about the functional benefits of a big-icon menu system that I proposed
10:50:01RockrabbitCan arm-elf-objdump be used to dump the contents of a Sansa c200 v2 firmware file? If so how is it done, and if not what can I use? thanks
10:50:06wpyhIt's all here at 06:00
10:50:12Martynwpyh : No bashing. It's modular enough for me to write new widgets, which is the start
10:50:38B4gderwpyh: well I certainly could argue against the idea of such an ui, but not against allowing one to be made
10:50:40Martynwpyh : After checking out the code, it's only taken me about four hours to get a first-approximation implemenation of a spinner widget
10:50:59wpyhMartyn: well, it's modular enough for that :)
10:51:14Martynwpyh : Well, I do UI's for a living
10:51:16B4gderRockrabbit: yes it can
10:51:27MartynThere are four widgets I need, plus a few functions that aren't available.
10:51:37MartynOne of which is a alpha-channel aware compositor
10:51:46B4gderRockrabbit: arm-elf-objdump -D −−target binary -marm [file]
10:51:50wpyhB4gder: no idea / UI design is perfect ;)
10:52:17wpyhMartyn: in that case, I wish you success −− Rockbox would benefit from a better UI
10:52:21B4gderwpyh: no, but I was trying to clarify that the discussion is two-fold: the one about the particular UI and the one about allowing it to be "pluggable"
10:52:35ZagorMartyn: for us GUI challenged persons, what is a spinner widget?
10:52:37Martynwpyh : Agreed. Or, if not a "better" UI, a different one.
10:52:55MartynZagor : A spinner is a kind of select box, that shows you the previous and next value
10:52:55wpyhwell, a different one might not be so practical...
10:53:08*wpyh was thinking about something that spins
10:53:08MartynZagor : Commonly used in the iPhone, for example, to select dates
10:53:30*B4gder has seen an iphone on a picture...
10:53:48Martynbut also used in a lot of embedded devices to suggest that the list of items you're scrolling through is large
10:53:58Martynwithout actually showing all the values
10:54:07wpyhB4gder: the UI being pluggable means that new UI's can be easily added. However, new UI's are likely to meet opposition, so it's more like a chicken-egg problem
10:54:26B4gderwpyh: I disagree
10:54:27Zagor ?
10:54:42Martynwpyh : One of the nice things about Open Source software .. is that any additions I make start as patch-commits :)
10:54:48MartynI can always maintain the patches
10:54:49wpyhMartyn: hey, I can think of useful things to do with your spinner. like for example the volume settings in the settings menu
10:55:10MartynYes Zagor, that is a simple implemenation of a spinner
10:55:13wpyhB4gder: which one do you disagree with: the opposition, or the chicken-egg problem?
10:55:19B4gderhowever, I don't think doing a "pluggable UI" is an easy thing nor do I think the impact will be small
10:55:36Martynalthough that spinner shows only one value, and only suggests the previous and next ones .. in the case of the one you found, it implies "+1, -1"
10:55:40B4gderand there is where the obejctions will come
10:56:01Martynwpyh : There are a number of widgets missing
10:56:12wpyhI think the impact being not small is where the objections come from
10:56:17B4gderwpyh: I don't think there's a chicken-egg problem
10:56:23Martynwpyh : I'll just implement the most used ones first, then go from there
10:56:32wpyhMartyn: great :)
10:56:33ZagorI wouldn't mind big changes, if it is done well.
10:56:48 Join robin0800 [0] (
10:56:49MartynOne of the things I need to implement is a "canvas"
10:57:00MartynThat's how icon-driven interfaces are created
10:57:09n1sZagor: depend on what is changed too, no?
10:57:11MartynAnd once you have a canvas, you can do interesting things on it...
10:58:02Zagorn1s: well, no. generally anything that is useful and done well is welcome in my view.
10:58:19MartynOh −− good example where a spinner would be useful −−-
10:58:35Martyn"settings" "display" "peak meter"
10:58:43wpyhB4gder: well, there is just one UI, and new UI's are not easily accepted, then there would be no need to make it pluggable, so it would be more difficult to write new UI's, so there will be less new UI ideas, therefore the dependency loops
10:58:49wpyh(not strictly looping though)
10:58:54MartynThe minimum/maximum range should be in a spinner, not a list
10:59:31wpyhMartyn: you mean a canvas is where you draw your widgets?
10:59:32n1sZagor: isn't a changes usefull
10:59:33Zagorwpyh: the first person/team to create a new gui has to pull a heavy load, yes. that cannot be avoided.
10:59:37MartynSince the concept of a range -implies- that the values are all listed as integers from +-
10:59:43Martynwpyh : Correct
10:59:46B4gderwpyh: yes, I understand what a chicken-egg problem is, but again I don't agree that this is one
10:59:52n1sness depending on _what_ the change is?
10:59:53wpyhZagor: yes.
10:59:59Martynwpyh : Also objects (such as icons) are drawn on the canvas, and then you can do -very- interesting things
11:00:22Martyn(by shifting, moving, etc...)
11:00:35Zagorn1s: we have plenty of proposed changes (patches) that I don't consider useful
11:00:56Martynplus the canvas can be much larger than the actual physical display, allowing for interesting UI modes (such as having predrawn pages flick left and right)
11:02:02n1sZagor: yes, that's what i meant. Your "I wouldn't mind big changes, if it is done well." comment sounded like "Any change is fine as long as the code is clean" but maybe i'm just too tired
11:02:07wpyhB4gder: yes, on a second thought, this is not strictly a chicken-egg problem
11:02:35B4gderI would rather agree with Zagor that the first one going for a new GUI has some heavy lifting to do
11:02:45wpyhit just seems that (sorry) devs are mostly opposed to changes
11:02:54B4gderI disagree with that too
11:02:59wpyhI do agree with Zagor on that
11:03:07ZagorMartyn: what is the practical difference between a canvas and the frame buffer?
11:03:08*Martyn flexes digital muscles.
11:03:23MartynZagor : Frame buffers are stored in video memory, and represent an actual screen?
11:03:31MartynA canvas is an abstraction
11:03:44ZagorMartyn: in that case we already use a canvas :)
11:03:51MartynZagor : -sort- of
11:04:04wpyhwell, my observation is that devs like the status quo very much ;)
11:04:25B4gderwpyh: that's not my view of it
11:04:36MartynZagor : What's in the code is more like a Virtual Framebuffer than an actual canvas
11:04:37B4gdermy view is that devs are careful to not bloat or do harm
11:04:54ZagorMartyn: ok, what is the difference between those?
11:05:46wpyhB4gder: I think they are different views of the same situation
11:05:52MartynThe simplest canvas is a kind of container, that contains widgets. The widgets have properties (layer, Zpos, Xpos)
11:06:02 Quit jhMikeS ()
11:06:02 Quit Rockrabbit ("CGI:IRC (EOF)")
11:06:03Zagoryou don't have to give me a "GUI course 1a" if you don't want to :-) it's not strictly on topic either...
11:06:05pixelmathe "problem" is that the current GUI might not look "slick" but works well - on all targets and also for the blind or visually impaired
11:06:20wpyhSo, after drawing on the canvas, you would copy it over to the framebuffer for display?
11:06:25B4gderwpyh: yes, but with an important distinction: the devs are not opposed to change if the change is done fine with a good purpose
11:06:57linuxstbMy view is that Rockbox's current feature set is complete (for my needs). So the focus should be on making all existing features work perfectly on all hardware. But I accept that's not what most people enjoy doing, and Rockbox hacking is just people doing what they find fun.
11:07:14MartynZagor : Example by analogy. X-Windows can run on top of a framebufffer, but it doesn't provide widgets by itself. X manages various things like the mouse, keyboard, and rendering to the screen.
11:07:14wpyhpixelma: you've got a point there. maybe I'm looking for ways to make it look nice without making it bloat too much
11:07:22MartynA canvas is like a Window Manager
11:07:23pixelmaand these two things I would want even for a new GUI and so far haven't seen a good suggestion
11:07:29UnhelpfulWebi think if there's anywhere to "improve" the current GUI it would be in rearranging the menu heirarchy. lists of text trump icons every time, unless you have a desktop worth of space in which to present options.
11:07:37wpyhB4gder:I'll keep that in mind
11:07:47MartynIt's one level abstracted away from X, providing the widgets and placement that make the UI useable to developers
11:08:13wpyhpixelma: which two?
11:08:19linuxstbMartyn: Remember that Rockbox devices are constrained in terms of cpu and memory. You have no malloc for example.
11:08:34ZagorMartyn: aha, I get it. thanks.
11:08:38 Join gregzx [0] (
11:08:56Martynlinuxtsb : I've spent the last 10 years making UI's for cellular phones, set-top boxes, and kids toys. I know -constrained- :)
11:09:06pixelmawpyh: working on all targets (or screens) and usable even for visually impaired)
11:09:49Martynlinuxstb : However, since rockbox runs under a linux kernel, I can trust the kernel to deal with alloc() and all it's ilk
11:09:56linuxstbIt does?
11:10:04ZagorMartyn: it doesn't
11:10:06wpyhok, I guess I don't know anyone who is visually impaired. by that, do you mean people who must use the speech synthesizer?
11:10:29MartynZagor : I didn't mean on the device .. I meant when I want to simulate something without a device
11:10:30*dionoea wonders where the linux/rockbox confusion began... and why
11:10:41Martyn<−−- uses linux to develop
11:10:48ZagorMartyn: ah
11:11:10ZagorMartyn: it seems odd to use alloc in the sim though since that code then won't work on target
11:11:16scorchedionoea: it would be madness to write one's own everything when linux exists!
11:11:26dionoeascorche: hehe ;)
11:11:29pixelmawpyh: yes, people who can't see good enough to read the display or have a hard time to
11:11:39Martynmalloc() is expensive
11:12:30wpyhpixelma: well, I can't help in that area −− but I also don't think a different, alternative UI will make it harder for them to use RB.
11:13:35MartynFrankly, one of the reasons we always develop a new UI for most devices -is- because each device has different constraints
11:14:57Zagorwe use the same UI since 2002 ;-)
11:14:58MartynOne of the toys I worked on, basically a little IM messaging device (looks like a pager) had four revs of hardware. first one had a 160x160 greyscale screen .. second had a color screen of the same resolution .. third had a 320x320 screen, and the last one to come out had a 640x480 screen before the company went under.
11:15:40MartynZagor : That's not a good thing.
11:15:45MartynNot even a little.
11:16:05B4gderwell, now it's much fancier than 2002
11:16:18MartynAgain, that's not better either.
11:16:21B4gderI mean colors and icons and so
11:16:21n1swe've added colours and stuff
11:16:35MartynI'd argue that the original iPod UI was brilliant, compared to almost any other PMP available at the time.
11:16:50B4gderMartyn: it also depends on the audience
11:17:01MartynBecause it displayed all the information needed very clearly, and used a minimum number of physical controls to do it.
11:17:06B4gderwe don't have the typical ipod users in our audience
11:17:23wpyhMartyn: there goes −− please don't praise the iPod UI on this channel
11:17:38n1swpyh: why not?
11:17:53MartynThe design was quite good, for the time. Today, we'd laugh at it.
11:18:04MartynBut it met all the constraints of the device it was designed for.
11:18:16MartynAnd that, not any particular love of the device, is what I'm getting at.
11:18:22wpyhn1s: well, I got smacked once for doing that (IIRC it was on the community channel though)
11:18:26MartynAs I said, I'm a UI designer and programmer.
11:18:37n1sMartyn: a fair deal of though has gone in to the rockbox ui and it works very well for many people it sju may not look very fancy
11:18:43*gevaerts guesses that Martyn uses gnome or kde
11:18:46pixelmawpyh: there was an argument during the inline settings discussion (seems this spinner is a similar idea) - you have to know exactly if you are currently changing the setting or just "hovered" over the item
11:18:47n1seh s/sju/just/
11:19:12MartynIn the case of the current UI on the little Sansa e250R ... the icons are far too small, and the fonts almost unreadable in the current layout. That small display demands that the UI be paged.
11:19:40pixelmayou know that you can change fonts and icons?
11:19:40MartynOn a bigger device, or on a screen with more real estate (say, the 5.5 version of the iPod) the layout looks good, and is quite usable.
11:20:14Martynpixelma : Yes. That's not enough. The information must be able to be arranged and displayed in a way that best uses the screen. As I said, 'paged', not listed
11:20:29gevaertsMartyn: I strongly disagree there
11:20:31MartynGive me a couple weeks, and you'll start seeing it in the code and in functional mockups
11:20:41pixelmahow? And how will you navigate this?
11:20:50Martyngevaerts : That's certainly your right and opinion :)
11:20:52wpyhpixelma: ah, ok. maybe there could be a better way of presenting it to the person?
11:21:01*B4gder looks forward to some interesting discussions ;-)
11:21:14gevaertsMartyn: sure, and when I'm in user mode, that trumps any UI specialist :)
11:21:33 Join PaulJam [0] (
11:21:35wpyhlike, when he hovers on Volume, which is -30dB currently: "Volume, minus thirty decibels"
11:21:47Martynpixelma : well, once I have a canvas implemented, the navigation will likely be more or less the same as current. When you reach the end of a page, it flips to the next one.
11:22:30ZagorMartyn: I look forward to seeing your ideas. Welcome to the project!
11:22:33gevaertsHow is that better than scrolling? Surely that's a _huge_ step back?
11:22:56 Join tyfoo [0] (
11:23:03*UnhelpfulWeb does not see the improvement in scrolling the list by pages instead of lines
11:23:16wpyhthen if he clicks, "changing volume". if he scrolls, read the volume level. if he confirms the change, "new volume, minus thirty five decibels". if he cancels the change, "volume unchanged".
11:23:32pixelmathat's not much different to a list - and it doesn' let me see options in advance _regularly_ - at one point you'll see none ahead and next a lot (bot none of the previous)
11:23:34Martyngevaerts : Perhaps I should do a Flash mockup of some of the interface ideas. Only takes me a couple days, and I can easily simulate the way a few devices would work....
11:23:44*wpyh personally thinks scrolling by lines is better than by page
11:23:47alexbobpcancellable volume changing?
11:23:53gevaertsMartyn: you can, but I won't have a look at it then :)
11:23:58*gevaerts won't install Flash
11:24:10Martyngev : Then that's your own personal issue :)
11:24:15ZagorMartyn: that would be a good basis for discussion
11:24:20Martyngev : Borrow a friends machine then. -laugh-0
11:24:26gevaertsMartyn: Flash is a UI disaster
11:24:27MartynOr use a library computer that has it installed :)
11:24:33wpyhalexbobp: I'm talking about the volume setting in the sound settings menu
11:24:39pixelmaMartyn: also remember that our targets have a restricted amount of buttons and or weird layouts of them (and there are quite some differences)
11:24:55Martyngevaerts : I'm not going to get drawn into a "XXX sucks" debate. I just use the tools I use :)
11:25:03wpyhMartyn: how about doing it in the uisim?
11:25:05Zagorlet's not all kick him in the knees before he's even made an attempt, ok?
11:25:33Martynwpyh : uisim would require a full implementation first. Flash I can do a fast mockup of what I intend to implement, so people can get an idea of how it looks and feels.
11:25:56scorcheMartyn: are you aware of tab-completion for nicks?
11:26:11wpyhMartyn: I do prefer scrolling by lines, but I do like a new option that makes RB scroll by page
11:26:14alexbobpwpyh: oh. I didn't know there was a volume control in the settings.
11:26:24Martynpixelma : That's why I chose the Sansa e250R and the iPod gen 5.5 to start with .. easy enough to get, and one of them is pretty restricted in the amount of memory, buttons, etc
11:26:51Martynscorche : That's a per-client implementation. I've got it currently disabled in Linkinus
11:27:01alexbobpmaybe his tab complete feature adds the space ;)
11:27:32 Nick fxb__ is now known as fxb (
11:27:58*GodEater would look at the flash implementation. I'm not a flash-a-phobe like gevaerts
11:28:11n1sMartyn: i'm not sure if your'e aware, but we have an option called "paged scrolling" that makes lists scroll by page
11:28:28wpyhpixelma: talking about simplifying the UI, is it acceptable if the cabbiev2 ui is made simpler? I think the gradient background is an useless attempt at looking slick. (I'm aware of the other themes)
11:28:45MartynI should probably also pick up an iRiver H140 or something like it, as a good third device
11:28:52wpyhn1s:I'm not aware of that
11:28:54Martynn1s : That's not the kind of page I mean :)
11:29:13Martynalthough that kind of paging is -also- good :)
11:29:18linuxstbMartyn: Having a target with an LCD remote would probably be useful - the H140 fits that category.
11:29:31n1sthen I don't understand what kind of page you mean
11:29:45MartynI mean page as in "canvas page" or "new drawing"
11:30:48pixelmawpyh: I didn't invent this, maybe you could take it to the "default theme contest" thread in the forums and ask the creator about it?
11:30:49MartynIn terms of UI, think of it like opening a new "tab" in a browser .. somewhere you can present information in a different layout, appropriate to the information you're trying to input or display
11:31:09alexbobpMartyn: if you do it in flash, I promise I will look at it and relay the information to those who can't ;)
11:31:10wpyhpixelma: ok
11:31:30MartynLayouts on a canvas are generally considered to follow a "flow" ... this item is on top of another item, and next to a third item.
11:31:52MartynOr "These three items must be displayed evenly spaced, and centered"
11:32:19MartynThis is how, for example, things are laid out in GTK+ ( just as an example )
11:32:59MartynIn embedded devices, it's possible to make a very lightweight canvas implementation, if you keep the number and types of widgets restricted.
11:33:48MartynThe basic types are generally on the order of : "canvas" "textbox" "listbox" "scroll" "slider" "inputtextbox" "button" etc.. etc..
11:34:36alexbobpoh, I see what he's saying
11:35:03MartynFor a Canon camera I worked on, the entire UI (framebuffer, canvas, widgets, the whole works) fit in 220K of memory
11:35:05pixelmaseems I didn't understand what you meant with "paged" there but I guess I'm not the only one :)
11:35:29MartynThis is going to take a bit of typing .. but I'll try to be concise:
11:35:58MartynWell, PseudoDataStructure really
11:36:30Martyn[Canvas name=settings]
11:37:16pixelmawhole Rockbox is 250kB on the OndioFM (IIRC) - the firmware file that's really needed for basic functionality such as playing music (without fonts, plugins etc.)
11:37:33Martynpixelma : That one would then drop canvas support
11:37:42Martynand use the current UI
11:38:12MartynOr perhaps I could make it quite small .. but I'm bettingthat the Ondio has a text interface...
11:39:01MartynNo, looking at the screenshots, I'm -dead- sure I could still make the UI work
11:39:10MartynMonochrome, very small screen.
11:39:17pixelmatext interface? Well it has a bitmap LCD... but I think I stop here
11:39:57MartynYeah, I think the constraint would be purely how small the firmware file has to be. That's all hand tuning
11:40:05Martynand I'm quite good at making a small interface
11:40:12Martynbut back to the example I was trying for..
11:40:42MartynCanvas contains all the objects for a given display .. and a definition of a DisplayPort that defines the physical dimensions and capabilities of the display
11:40:47wpyhMartyn: I seriously think you should do a flash mockup of what you want to do, on charcell, monochrome and colour targets. that way, the devs might want to take a look at it. disclaimer: I'm not a dev
11:41:38MartynThen you create a "page" ( a given interface ) on the canvas. It's a logical datastructure that defines what widgets are on the display, and the callbacks that will update them
11:42:15GodEaterwpyh: will you please stop putting words in other people's mouths? People here are trying to understand Martyn's concept, not shoot it down.
11:43:21MartynA page, therefore can be thought of as an arrangement of widgets.
11:43:30MartynSound Settings, for example, wouldn't have to be a list of lists.
11:44:27MartynIt would be single page with four sliders ( Volume, Bass, Treble, Balance ) arranged on top of one another
11:44:51MartynYou select one of them, and then use whatever available keys you have to slide the widget left/right
11:45:19MartynOnce a value is selected, the page then activates "select an item" rather tan "configure this given widget"
11:46:01MartynChannel configuration would be on the same page, as a "selection scroll" with six values in it (stereo, mono, etc)
11:46:15MartynYou get the idea I think from here..
11:46:48MartynYou already have some similar concepts already in the code.
11:46:54MartynI found the "Graphical EQ"
11:47:25Martynand that's a good example of a multiwidget page
11:47:59MartynThe idea here is that those kinds of widgets would be available everywhere in the interface.
11:49:10MartynFor devices with more capability, more of the UI widget base becomes available.
11:49:47MartynThings like image sliders, object composition, perhaps even things like a video window
11:49:55Martynbut that's for far, far later.
11:50:19Martynfor now, I'll be happy to implement a scrollbox :)
11:51:24MartynCrap, power cut out.
11:51:36MartynStupid cold weather.
11:52:02MartynI'll continue this discussion later .. when we're all awake, and do a simple flash mockup of some of the interface ideas.
11:52:13*GodEater looks forward to seeing it
11:52:16MartynMeanwhile, I'll keep editing code to get a few widgets I'd like to use :)
11:52:21gevaertsMartyn: Just be careful not to take options away from us :)
11:52:48MartynWell, considering I don't have commit access ... I don't think that's a problem gev
11:53:17gevaertsMartyn: I meant more that "dumbing down" the interface is a good way to get any proposal shot down here
11:54:29 Quit Martyn ()
12:00:59 Quit gregzx (Read error: 54 (Connection reset by peer))
12:01:04 Quit linuxstb (Read error: 60 (Operation timed out))
12:06:46 Join gregzx [0] (
12:16:04 Join tvelocity [0] (
12:18:14 Quit robin0800 (Remote closed the connection)
12:19:00 Quit ender` (Read error: 131 (Connection reset by peer))
12:19:14 Join ender` [0] (
12:20:29soapLlorean, when you wake up: Can you tell who put any given post in the trash? is not a terribly useful post, but it is not far astray from the rest of that thread and I think it was bad PR to remove it. I wouldn't mind your opinion on it and your movement of said post back to the thread (I'll be at work) if you concur.
12:22:03scorchesoap: he did
12:22:05GodEaterI thought that too soap
12:22:22soapGodEater's on my side?
12:22:27soapback it goes!
12:22:58GodEaterI don't think I've deleted a post for a while
12:24:10 Join moos [0] (i=moos@rockbox/staff/moos)
12:24:35GodEaterI find dropping a valium before I read through the forums has chilled me out enormously
12:25:00scorcheGodEater: the unlogged channel is over there−−> ;)
12:25:37GodEaterI don't mind that staying for posterity ;)
12:25:55soapLlorean, IMHO there are plenty of "Offerings of Opinion" in that thread and I see no reason (outside a bit of lazy spelling) why this one was singled out. If you had a previous discourse with said poster regarding their lazy spelling or some other reason for shit-canning the message I'm curious. I did move it back - not looking to start a fight - just feel the trash-binning was bad PR IMHO.
12:31:18stu8ballHi all.
12:31:34stu8ballYou know the way Rockbox remembers between reboots what song it was playing, and what its position in the song was?
12:31:56Zagorstu8ball: we call it "resume"
12:31:57stu8ballWhen I mount my DAP as a USB disk on the computer, what file can I find that info in?
12:34:27Zagorstu8ball: /.rockbox/config.cfg
12:35:27stu8ballk, thanks
12:35:34PaulJami thought nvram.bin and .plalyst_control
12:35:35GodEateris it ?
12:35:43*GodEater agrees with PaulJam
12:36:01Zagoroh dear. /me reads the code poorly then.
12:38:27 Join planetbeing [0] (
12:39:29 Join webguest46 [0] (n=4f8386a7@gateway/web/cgi-irc/
12:40:11***Saving seen data "./dancer.seen"
12:40:30 Quit webguest46 (Client Quit)
12:42:16planetbeingGodEater_: I was randomly poking in the irc logs and noticed you looking at xpwn source, trying to see if a similar DFU mode can be used on the second generation nano.
12:42:45planetbeingAny luck with that?
12:43:25planetbeingBe aware that DFU mode code on the iPhone and iPod touch is probably written by Samsung, not Apple.
12:44:26 Quit moos ("Rockbox rules the DAP world")
12:44:33 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:45:11planetbeingAlso, the xpwn dfu-util is basically openmoko's version, but adding a really random security checksum with some sort of hardcoded salt. The code is in iTunesMobileDevice.dll, which is probably iPhone/iPod touch specific.
12:47:00planetbeingLastly, DFU mode on the iPhone loads code into RAM, applies a RSA signature check on it and then executes it (this takes place on DFU manifest as described by the standard USB DFU specs)
12:48:02planetbeingWe defeat the signature check using a stack overflow in the DER certificate parsing code.
12:48:47 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:49:50 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
12:50:56planetbeingI've noticed that the iPod efforts are always said to be stymied by "encryption". But if there's only that and no codesigning, then one kernel bug in any firmware revision and you win forever, so your path is easier than ours was.
12:53:00GodEaterplanetbeing: a fellow rockboxer found DFU mode on the 2nd Gen Nano, but only by trashing the firmware partition first
12:53:13n1sI don't think it's known that there
12:53:17GodEaterI was searching for the same on the 3rd gen nano - but that's proved impossible so far
12:53:17n1s's no signing
12:53:41GodEatersince the UMS on the iPod's side doesn't export the firmware partition (if there even is one)
12:54:17planetbeingInteresting that theres no buttom sequence.
12:54:20GodEatersome of our other hackers have modified the openmoko DFU tool to work with other players (meizu)
12:54:31GodEaterplanetbeing: well - not one that we've found anyway ;)
12:55:02planetbeingBut youve got a DFU mode endpoint out of the firmware trashin?
12:55:09GodEaterI came to the iphone scene a little late - was the DFU key sequence on that documented somewhere by apple, or found by someone else ?
12:55:18GodEaterplanetbeing: only on the 2G yes
12:55:24planetbeingtheres a chance code in the clear could be sent
12:55:38planetbeingdoes itunes offer to do anything with it?
12:55:41GodEaterthat was the theory yes - no idea what was tried after that
12:55:57GodEaterplanetbeing: not sure, as I say, I don't have a 2G, so it wasn't me that was working on it
12:56:07planetbeingall right.
12:56:09GodEaterLamdaCalculus37 is our 2G owner
12:56:30planetbeinghas there been any effort with fuzzing the firmware?
12:56:45GodEaterI think there's basically been zero effort so far
12:56:58planetbeingim not sure what the situation is with the ipod firmware. does everything work in kernel mode?
12:56:59GodEaternone of the people here with the smarts owned a latter model ipod until recently
12:57:09GodEaterand everyone else that turned up interested in a port lacked the skill to do anything
12:57:17planetbeingI see.
12:57:54*GodEater wonders why whenever planetbeing turns up, the rest of the rockbox hacker community seem to disappear, and he's left answering all the questions on his own :(
12:58:26planetbeingwho else is interested in the port in this community?
12:58:48GodEaterthe previous gen hackers on the ipod include linuxstb, amiconn and jhMikes
12:59:13GodEaterthere are probably others too, but they're the main experts as far as I'm aware
12:59:35GodEaterlinuxstb is UK based like me, amiconn is in Germany, and jhMikes is in the US I think
12:59:50GodEaterI was trying to persuade amiconn to go to your CCC presentation
12:59:57GodEaterbut he's been grumpy and said no :)
12:59:58ZagorI'd say there are plenty of us interested in a port, but since the prospects have looked grim many people have avoided getting an ipod.
13:00:46Zagortargets with many users have a special appeal
13:01:29GodEateresp. if we get lucky like we did with the early ipods, and find that porting to later generations is fairly trivial
13:01:38planetbeingWell, at least you guys havent found any certs.
13:01:44GodEateri.e. a port to 2G is then easy to move to 3G and then 4G Nanos
13:02:23planetbeingThat lessens the chance of codesigning. Even if there was stuff in the encrypted payload, that still means they dont sign the headers.
13:03:22planetbeingRSA is also computationally intensive if you dont have specialized hardware.
13:03:39GodEaterI don't even have "my" 3G to play with at the moment. Mrs GodEater has it, and she's the other side of the planet to me at the moment.
13:04:08planetbeingI take it RC5 was used instead of AES?
13:04:16planetbeingin previous gens?
13:04:36GodEaterI'm not sure I recall
13:04:42GodEaterwe nicked a lot of stuff from the ipodlinux guys
13:04:57*GodEater wonders if they've successfully resurrected their website yet
13:05:17 Quit UnhelpfulWeb ("CGI:IRC (EOF)")
13:05:25planetbeingi think they have a few ideas themselves.
13:05:30 Quit reacocard (Remote closed the connection)
13:05:37planetbeingapparently no equipment to execute it?
13:06:01GodEaterallegedly not - they came up with that plan a long time ago
13:06:05GodEaterand it never went anywhere
13:06:12scorchefor various definitions of "ideas"...we have talked a bit about different methods, but not going anywhere stongly
13:06:14GodEaterI think the "linux4nano" project got a *bit* further
13:06:36planetbeingwhat do you mean by "further"?
13:06:43scorchethe last idea we were talking about involved delayering a chip and reading the ROM from a SEM, so....
13:06:47GodEaterthey managed to get a dump of code off the hardware
13:06:55planetbeingscorche: hahaha
13:07:06planetbeingReally? CODE?
13:07:40GodEaterthat was back in May though
13:07:47GodEaterI've not heard any more since then
13:08:25GodEaterdoes anywhere here have a copy of that dump? I don't think I do...
13:08:28planetbeingoh, the flash dump
13:08:37planetbeingapparently it was mostly encrypted?
13:09:00planetbeingwith one apparently cleartext jump.
13:09:26GodEaterthat might well be it :)
13:09:28planetbeingthough if they got that, they can just write their own code into te cleartext section.
13:09:50planetbeingsince its a standard flash chip, thats easy
13:10:22*GodEater has to run off to a meeting :(
13:10:28planetbeingno need to muck with trying to read ROM. o.O
13:11:01planetbeingah. have fun
13:13:27planetbeingi'll be around. i can be found on in #xpwn, if you guys need a hand reversing something. Other than that, I don't think there's sufficient similarities between s5l8900 and whatever soc they use in the latter gens to actually KNOW anything useful. ;)
13:16:57n1skugel: FS #9662 is a duplicate of FS #9551, could you add your findings in a comment there instead?
13:19:31kugeln1s: it's not a duplicate
13:20:11kugelit's about that the stop button goes up one dir, instead of stopping/additional to stopping
13:20:39 Join skipper [0] (
13:20:46n1sits' about stop behaving inconsistently on a single player, and the older one is about play and stop behaving inconsistently on all players, why is it not a duplicate?
13:21:17 Quit AndyI (Read error: 60 (Operation timed out))
13:21:41kugelbecause here the stop goes up one dir. that's a simply a bug, not inconsistent
13:22:05kugelalso, Llorean told me to file a bug report, as he wanted to have a look at it
13:23:24 Join AndyI [0] (i=AndyI@
13:23:36kugelthe stop button should already be "global" on the e200, so it doesn't need to be globalized, but the bugs need to be fixed
13:27:18 Join reacocard [0] (
13:29:46 Quit planetbeing ("Lost terminal")
13:32:05 Join johnnyboy [0] (
13:44:27 Join fredddy [0] (
13:45:10 Quit kachna (Read error: 110 (Connection timed out))
13:46:25 Join faemir [0] (
13:48:28 Join Aurix_Lexico [0] (
13:49:58 Quit fredddy (Remote closed the connection)
13:50:34 Join fredddy [0] (
13:52:16kugellucent: I've been using the 1.01.15 OF version for my disassembly
13:54:45 Join stRR [0] (
13:58:57 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:03:47 Join Kitti [0] (n=himka_co@
14:03:51 Part Kitti
14:04:28 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
14:14:30 Join TomKazimiers [0] (
14:16:05TomKazimiersHi there, I am here to ask for wiki write permissions - Currently I am playing around with using a compact flash card with my iriver H300 and would like to add another working card to the wiki :)
14:17:20ZagorTomKazimiers: you're added now
14:18:56TomKazimiersthank you very much :)
14:27:37 Quit XavierGr ()
14:34:00 Quit johnnyboy ("Leaving.")
14:36:12 Quit wpyh (Read error: 110 (Connection timed out))
14:36:51 Join wpyh [0] (n=william@
14:40:14***Saving seen data "./dancer.seen"
14:41:23 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
14:41:50 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
14:47:47 Quit faemir (Remote closed the connection)
14:51:27 Part TomKazimiers
14:52:19 Join faemir [0] (
14:53:12 Quit Acksaw (Read error: 104 (Connection reset by peer))
14:53:28 Join Acksaw [0] (
14:59:24 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
15:13:48 Quit tvelocity (Read error: 113 (No route to host))
15:13:50 Quit faemir (Remote closed the connection)
15:15:05 Join tvelocity [0] (
15:17:53 Quit Llorean ("Leaving.")
15:18:13 Join Llorean [0] (
15:20:06 Join samurailink3 [0] (
15:20:15 Part samurailink3
15:25:01 Join nplus [0] (n=nplus@141.25.Globcom.Net)
15:25:56 Join Schmogel [0] (
15:26:43 Quit Llorean (Read error: 54 (Connection reset by peer))
15:27:13 Join kachna [0] (
15:34:25 Quit tchan ("WeeChat 0.2.7-dev")
15:34:36 Join tchan [0] (
15:34:46 Join Llorean [0] (
15:35:44 Join obo [0] (n=obo@rockbox/developer/obo)
15:36:34 Join faemir [0] (
16:01:39 Join {phoenix} [0] (
16:18:42 Quit PaulJam (".")
16:19:33kugelmoos: hey, how's your e200v2?
16:20:09moosHi kugel: didn't find the time to try yet. Hopefully today...
16:21:23kugelmoos: I've put another patch up, which you might try (when you found the time that is)
16:21:43moosSure I will thanks, and of course let you know
16:22:23 Join obo_ [0] (n=obo@rockbox/developer/obo)
16:23:21stRRdoes anyone know if RB works on e280?
16:24:00 Quit obo (Read error: 110 (Connection timed out))
16:28:20advcomp2019stRR, what version.. if you have v1, yes
16:28:47kugelstRR: the website knows...
16:29:32stRRadvcomp2019: well, i don't know for sure. it's just says e280 on the backside
16:30:09advcomp2019stRR, look at the firmware version then
16:30:11 Quit obo_ (Remote closed the connection)
16:31:01stRRadvcomp2019: where can i find it?
16:31:39stRRwell 01.02.18E
16:31:56B4gdercongratulations, rockbox works on your device!
16:32:05advcomp2019the wiki and/or the manual should tell you and looks like it can be
16:32:44Zagorisn't "make install" supposed to install .rockbox/fonts/08-Rockfont.fnt too?
16:33:54Zagorand what's wrong when SDL_OpenAudio() fails without giving any error?
16:34:53 Join toffe82 [0] (
16:35:31 Part LinusN
16:35:57pixelmait should install fonts that are used in any of the included themes
16:36:48ZagorI get 10 themes but no fonts
16:37:45pixelmaRockfont is used in cabbiev2 for Archos (and other displays with 64 pixels LCD height I think)
16:38:13 Join massiveH [0] (
16:38:50 Quit massiveH (Client Quit)
16:40:15***Saving seen data "./dancer.seen"
16:42:17wpyhGodEater: I don't understand what you mean by me "putting words in other people's mouths".
16:43:07B4gderI can do the branching right now
16:43:57*gevaerts has no objections
16:44:09LloreanSounds good.
16:44:11moosNo more defaults seeting changes?
16:45:57LloreanNo sense delaying the branch for something like that, at least.
16:46:21moosis the final release date fixed?
16:46:55gevaertsmoos: some people like 26 december, as in "Rock-Boxing Day"
16:47:09B4gder"svn co svn:// rockbox-3.1"
16:47:21*B4gder is too lazy to write something in a wiki page
16:47:31moosgevaerts: ok :)
16:48:11moosle's hope that will attract people "massively" again
16:48:39mooslet's even
16:48:58kugelB4gder: so freeze is over?
16:49:19B4gderfreeze is officially over, but of course we still prefer bug fixes - in both branches!
16:49:27kugelof course
16:49:34ZagorI see nobody changed "backlight_timeout_plugged" as was mentioned yesterday :)
16:50:00Zagordo we have any serious known bugs?
16:50:39kugelwhen is the release scheduled? I guess 23rd isn't actual anymore?
16:50:56kugelthough I think releasing before christmas would actually be nice
16:51:19 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
16:51:48ZagorI don't think 23rd is ruled out. we didn't delay the branch for a reason, just laziness
16:51:58kugelah nice
16:52:15moos16:46 gevaerts moos: some people like 26 december, as in "Rock-Boxing Day"
16:52:16LloreanZagor: Well, we delayed the branch for a reason.
16:52:21B4gderI'll check out a 3.1 branch and make a set of binaries
16:52:23ZagorLlorean: we did?
16:52:30B4gder"release candidates"
16:52:33LloreanZagor: The freeze was delayed a couple days due to the DDoS
16:52:41LloreanTo keep the freeze length the same the branch was delayed.
16:53:02LloreanOr at least, I *thought* we delayed the branch because the freeze started late.
16:53:06kugelbut we could still release in 5 days, don't we?
16:53:12LloreanProbably, yes.
16:53:35B4gderI don't forsee a whole lot of frantic fixing in the 3.1 branch
16:53:51Lloreankugel: I have the Flyspray RSS subscribed, so I saw.
16:54:07kugelok, but you may not have noticed that I added a fix
16:54:39Zagordo we have an SDL guru?
16:54:51kugelbut I'm not sure if the fix should be committed, there's some quirks with the time&date screen
16:55:22 Join obo [0] (
16:55:30kugelLlorean: I could fix those too. Is resolving that desired for 3.1? I'd think so
16:55:54Lloreankugel: I think it should wait until after 3.1
16:56:04LloreanThe current keymap "works". It's a little unexpected apparently, but at least it functions.
16:56:18kugelyea, that's true too
16:56:25LloreanBetter to play it safe with keymaps, and fix it for 3.2 where there's time in case any new unexpected behaviour shows up
16:57:15kugelok, that gives us at least the possiblity to do it right, as there might be other changes involved to the keymap
16:58:51 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
17:01:48*B4gder builds 3.1RC binaries
17:02:52LloreanI was considering asking if we were doing those.
17:03:59B4gder"under construction" ;-)
17:04:48stsquadIs there a paramter I can pass to make to see the actual CC invocation used?
17:05:10B4gderI'm not sure if Zagor broke it, but it used to be make V=1
17:05:10Zagorstsquad: V=1
17:05:39wpyhI noticed that some bmp files inside wps/cabbiev2 have the executable bit set. Is this intentional?
17:05:42stsquadthat works, thanks
17:05:49 Quit Zagor ("Client exiting")
17:06:04LambdaCalculus37wpyh: Which files?
17:06:25wpyhthe files ending with -128x64x1.bmp
17:10:27pixelmacurrent is able to distinguish different remote WPS based on the size of the main screen (e.g. there is iCatcher.128x64x1.160x128x2.wps (in use for the H100) and iCatcher.128x64x1.220x76x16.wps for the H300). I tried using this system with cabbiev2 to have different WPS for Iriver remotes and the Cli
17:10:38*stsquad realises why he can't use normal libc for simulator debug fprintfs
17:11:53stsquadwhats the quickest way to do quick and dirty prints to stdio for debugging?
17:13:04mud-rbstsquad: DEBUGF should be what you're looking for. (normal printf formatting works in it too)
17:14:55pixelmaso I had three files - the two cabbiev2.128x64x1.mainscreenresolution.wps and the simple cabbiev2.128x64x1.wps ... and while the latter existed only this one was packaged (also in H300/H100 builds) and the ohers were ignored. Could this matching be somehow done differently?
17:17:58stsquadmud-rb: Did you see my mail? DEBUGF seems to be multiply defined in many places
17:20:45mud-rbstsquad: ah, i see it. i'm not too sure about all of that, i just do plugins.
17:21:45 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:21:46stsquadnps - went for the obvious debug.h :-)
17:22:34 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:23:38 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
17:24:17 Join PaulJam [0] (
17:24:55 Quit neddy (Read error: 104 (Connection reset by peer))
17:26:13stsquadSo how am I meant to pass paramters to menu item strings that contain %s's?
17:26:27stsquad"Save playlist as %s?"
17:32:25B4gderthe 3.1RC builds are now completely built
17:40:35 Quit Horscht ("")
17:48:48 Part B4gder
17:49:09 Join stoffel_ [0] (
17:49:12wpyhLambdaCalculus37: do you see the files I mentioned as having the executable bit set?
17:50:06LambdaCalculus37wpyh: I did. I can't change them right now, though.
17:53:38 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
18:02:11 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
18:04:58 Quit faemir (Remote closed the connection)
18:05:45 Join faemir [0] (
18:08:23 Quit jhulst (Read error: 60 (Operation timed out))
18:11:54 Nick fxb is now known as fxb__ (
18:13:00 Quit faemir (Remote closed the connection)
18:15:09 Quit petur ("work->home")
18:18:41 Join karashata [0] (n=karashat@
18:19:47 Join faemir [0] (
18:19:50 Join japc [0] (n=japc@
18:21:24 Join itcheg_ [0] (i=41d59de2@gateway/web/ajax/
18:22:36 Quit stoffel_ (Read error: 113 (No route to host))
18:23:12 Quit itcheg (Nick collision from services.)
18:23:21 Nick itcheg_ is now known as itcheg (i=41d59de2@gateway/web/ajax/
18:30:53 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:31:42 Quit bodymind ("Ex-Chat")
18:36:07 Join tyfoo2 [0] (
18:37:18 Join toffe82_ [0] (
18:39:52 Join elinenbe [0] (
18:40:19***Saving seen data "./dancer.seen"
18:40:52 Join bertrik [0] (
18:41:07elinenbelooks like I will need to upgrade my iPod and Toshiba in a couple of months!
18:43:41 Join Jaykay [0] (
18:44:17rasherHm, why don't we display a splash when trying to load a broken wps?
18:45:21rasherAlso, did the WPS syntax change since Rockbox 3.0?
18:46:01LloreanNothing incompatible.
18:46:14LloreanWe have new tags, but 3.0 WPSes should be 100% compatible as far as I know.
18:46:23pixelmathere were some additions (%mv and the settings thing...) but as Llorean says
18:47:12*rasher discovers %mv
18:47:22 Join MethoS- [0] (
18:47:31rasherThat might be nice to use to get a giant volume meter when changing
18:48:17rasher* Make %t stricter by aborting if a value is not given.
18:48:23pixelmawanted to suggest using it in cabbiev2 (not for a "meter" but changing to numerical display while adjusting, like the plain statisbar does)
18:48:25rasherMight this not break some WPSes?
18:48:31pixelmastatus too
18:49:57 Quit Xerion (" ")
18:50:17pixelmahmm... it could. Wasn't aware of that %t change. I'm not even sure if some shipped WPSs used it without parameter, has that been checked?
18:50:25 Quit tyfoo (No route to host)
18:51:02*rasher fires up grep
18:51:39rashergrep "%t[^0-9]" wps/*wps yields nothing
18:52:31 Quit tvelocity (Read error: 110 (Connection timed out))
18:52:37*pixelma is not fluent in regexp
18:53:03rasherMeans "%t followed by anything other than a number"
18:53:09 Join tvelocity [0] (
18:54:44pixelmaaha, thanks for the explanation. So it looks ok
18:55:14rasherSeems so. For the shipped WPSs anyway
18:55:55 Quit toffe82 (Read error: 110 (Connection timed out))
18:59:05 Join stoffel_ [0] (
18:59:10 Join herrwaldo [0] (
19:04:06 Join BigBambi [0] (
19:05:01 Join tvelocity[a] [0] (
19:09:07 Quit skipper (Remote closed the connection)
19:10:30rasheramiconn: does drawmode not effect bitmaps on colour targets?
19:12:36amiconnDrawmode affects mono bitmaps, but not native bitmaps, on all targets
19:12:57amiconnOn targets where this is contradictory (i.e. mono targets) all bitmaps are currently affected
19:13:15*rasher wonders why draw_album_art even bothers setting the drawmode before drawing, then
19:13:35amiconnBecause someone didn't read GraphicsAPI ?
19:14:06*amiconn wonders whether 3.1 will have the dreaded shuffled playlist resume bug fixed
19:15:01*rasher was hoping inverting the albumart would be as simple as setting the drawmode
19:15:06 Join aneqrs [0] (
19:15:30rasherAs in inverting the colours (negative)
19:15:33amiconnLike photographic negative?
19:15:48rasherYeah, when the screen mode is set to negative, to get the "real" album art displayed
19:16:15amiconnWhich colour target has screen inversion?
19:16:25rasherThe e200 does at least
19:16:37*amiconn thinks that's a strange feature for a colour target, and should be removed
19:17:03amiconnImo it only makes sense on mono targets, where it's not possible to set bg/fg colours
19:17:10rasherWhy? It's perfectly useful - if you have a light theme you like, but want dark. I rather like cabbiev2 inverted.
19:17:22rasherDoesn't work for all themes of course
19:17:40rasherJust that albumart comes out looking weird
19:17:51amiconnBut it is possible, requiring 2 steps: (1) Draw the albumart bitmap as usual. (2) Set DRMODE_COMPLEMENT, and draw a filled rectangle across the whole bitmap
19:18:08rasherooh, neat
19:18:38amiconnKeep in mind that filling something with DRMODE_COMPLEMENT is significantly slower than filling with a solid colour or the background
19:19:08elinenbeare there any patches or plans to read album art from id3 tags? (including jpg art?)
19:19:18amiconnSolid colour uses memset16(), background uses memcpy(). DRMODE_COMPLEMENT xor's each pixel with all-1-bits
19:19:33rasheramiconn: Yeah, figures
19:19:52rasherooh, that indeed works beautifully
19:21:06 Quit tvelocity (Read error: 110 (Connection timed out))
19:21:25n1samiconn: is " the dreaded shuffled playlist resume bug" in FS?
19:31:01 Quit stoffel_ ("leaving")
19:31:37rasheramiconn: Most targets in fact seem to define HAVE_LCD_INVERT
19:32:09 Join Seed [0] (
19:32:12kugelbug: an empty .wps isn't empty
19:32:57rasherWorkaround: a text with nothing but "%wd" or "%we" in it
19:33:51pixelmakugel: why bug?
19:33:52 Quit japc (Read error: 110 (Connection timed out))
19:34:26kugelpixelma: if I load a wps containing nothing I want nothing :)
19:34:47rasherWhat does it contain?
19:35:10pixelmaturning off the statusbar globally also helps and that's why %wd and %we exist in case someone wants something different in menu and WPS
19:35:24kugelthe old rockbox default one, that one with peakmeters
19:37:49 Join miepchen^schlaf [0] (
19:40:40kugeljust a little tease: runtime with cabbiev2: 18:03, runtime with showing the main menu:19:51
19:40:43 Join t0mas [0] (n=tomas@rockbox/developer/t0mas)
19:41:43kugelwhich is 11% more runtime
19:43:09pixelmaok, so completely empty wps are rejected. Just played around a bit, adding a newline already helps. I don't think I would have expected this but in a normal use case no-one would have a completely empty WPS
19:43:47kugelthat's a major showstopper imho
19:44:01ameyerwow.. 11%
19:44:08ameyerthat's quite a hit
19:46:07kugeland that's only cabbiev2, expect more complex wpses to be even more significant
19:47:03ameyerwhat about the old default wps?
19:47:22ameyerI mean, more or less complex than cabbiev2
19:47:30kugelthere will be a test with a complete empty wps, let's see what that yields
19:47:42rasherThe old default uses peakmeters - that's probably even worse
19:47:54kugelcredits go to Jaykay btw for doing the benches
19:47:57ameyeris there a copy of the default wps somewhere?
19:48:12kugelameyer: in your wps folder
19:48:20ameyerjust a placeholder
19:48:26lucentkugel: I tested FS #9663 on 8gb Fuze, it behaves as you state in the task (need to have SD activity or press buttons when booting).
19:48:47kugelweird isn't it?
19:49:01kugelbut it's better than what's in svn for sure
19:49:55lucentI played a whole FLAC encoded album last night when I went to sleep, it was nice and I didn't notice any error
19:49:56ameyerliterally, just has comments saying something along the lines of "this is a dummy file that allows rockbox to reset to the default wps"
19:50:15kugelameyer: and that doesn't work?
19:50:44lucentkugel: I did find a bug, can you tell me why this is? When backlight is off, and I press a button, the backlight turns on but the screen does not update until an additional button press
19:51:07lucentkugel: so it appears as if I haven't pressed a button, but it's just that the screen has not updated
19:51:25rasherIt's a bit odd that rockbox_default doesn't show the default (cabbiev2)
19:52:09pixelmarockbox_default is rockbox_builtin
19:52:20LloreanIt should probably be renamed.
19:52:32Lloreanameyer: I believe the actual text of that .wps is on the Archos WPS gallery.
19:52:43kugellucent: works for me, can you give me a reproduce recipe?
19:52:48pixelmaameyer: there used to be a copy in the Archos WpsGallery and it must be somewhere in the code
19:52:55*pixelma too slow
19:54:16lucentkugel: on my player, 1 ) turn on player 2) wait for backlight to turn off 3) press "down" button with your simplified keymap patch 4) observe display 5) press "down" and observe menu selection jumping twice down
19:54:27ameyereh, probably should just DIY anyway
19:55:50lucentkugel: i.e. first press of "down" and the display still shows selected Files menu item, second button press and it jumps down to Resume Playback
19:56:06lucentkugel: but I'm pretty sure it's not jumping, it's just screen update lagging behind
19:56:37kugellucent: oh yea. it's not the display though
19:56:51kugeltry in wps, it works there
19:57:17lucentoh okay, I have to find out what wps is, hang on
19:57:27kugelwhile playling scren
19:59:27lucentkugel: confirming what you say, |>Settings>Theme Settings>Browse Themes>cabbiev2 *select*
19:59:58kugellucent: yes. wps refers to the screen you are led to when you play music
20:00:22kugelthat one is completely configurable using .wps files
20:00:25lucentoh okay I'm talking only about the Rockbox menu
20:00:49kugelI noticed. but the wps updates properly, so it's not the display
20:01:18lucenthm. So it is not the display code, that is what you say?
20:01:30lucentwhy might this be happening on the main rockbox menu?
20:01:45kugelcan someone try to reproduce that on a supported target?
20:01:51lucentgood idea
20:02:05kugelI fail to on my e200, but that might be due to the scrollwheel which always works a bit differently
20:02:38PaulJamlucent: maybe i misunderstand your problem, but could it be that you have the "first keiypress enables backlight only" setting enabled?
20:03:51PaulJamah, forget what i said.
20:04:01lucentPaulJam: the trouble is, the screen is not updated but the button does activate its action
20:04:46kugelhappens on e200 too (I just used my reduced keymap)
20:05:11lucentmaybe a bug in rockbox code elsewhere, maybe a bug in reduced keymap
20:05:13kugelPaulJam: yes he does. it's the default setting so he might not know about it
20:05:23kugellucent: unlikely
20:05:29kugelthat it's the keymap i mean
20:05:38lucentPaulJam: however, setting "first keypress enables backlight only" to YES is a good workaround :)
20:06:56kugelPaulJam: can you try to reproduce?
20:07:50kugelPaulJam: or is that even known already? you seem to have heard about it
20:08:04kugeloops /me missread
20:08:33 Quit miepchen^schlaf ()
20:11:02 Join Thundercloud [0] (
20:11:27 Join miepchen^schlaf [0] (
20:14:21Aurix_Lexicohow many bits is a int in ARM?
20:15:51lucent32 I thought, if ARM is a 32-bit processor
20:16:22kugela word is 4bytes
20:18:07 Join enginerd3 [0] (i=407c962a@gateway/web/ajax/
20:18:44 Quit miepchen^schlaf ()
20:18:56 Quit tyfoo2 ("Carpe diem")
20:20:00 Join miepchen^schla [0] (
20:21:10 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
20:21:15saratogaI like all this talk of experimenting with new GUI stuff
20:21:34 Join nk [0] (n=4157b702@gateway/web/cgi-irc/
20:26:41PaulJamkugel: sorry, i was on the phone. i just tried on my h300 with default settings, and i can not reproduce it here.
20:27:48 Join Richlv [0] (n=rich@
20:28:11kugelPaulJam: weird. I can't really believe that it's due to my keymap file
20:29:07 Quit enginerd3 (" ajax IRC Client")
20:29:26PaulJamlucent: did you use the default settings?
20:29:47Lloreankugel: Do you leave the scrollwheel stuff defined? Is it possibly an artifact of using keys possibly with scrollwheel list acceleration?
20:30:04lucentPaulJam: yes, default settings
20:31:07kugelLlorean: I removed everything with scrollwheel from the keymap, and #if 0'd the HAVE_SCROLLWHEEL and two acceleration defines (else it didn't work anyway)
20:31:35lucentI don't know how to write the code for it, but a test for the screen being at fault or not is to enter some code which draws to the screen anytime the backlight is turned on
20:31:46Lloreankugel: Well, any apps-level problem (which it'd more or less have to be if it'd show up cross-target) would show up in the sim.
20:31:52LloreanSo you could do testing for it yourself, too.
20:32:20pixelmabutton action conflicts are quite easy to get, especially if you use less buttons ("unprotected" release and repeat events for short and long keypresses etc.)
20:32:25Richlvwhich company from the list at is considered to be the most opensource friendly lately ?
20:32:46n1snone of them?
20:32:52lucentnone of them, I'd say
20:33:00 Join enginerd3 [0] (i=407c962a@gateway/web/ajax/
20:33:15Richlvis there such a manufacturer at all ? :)
20:33:16LloreanIn terms of "toward Rockbox", at least Sandisk donated a couple pieces of hardware, once, long ago.
20:33:19lucentSanDisk has been developer friendly, as "someone" sent RockBox team a cabbage patch of players to hack on
20:33:30lucentbut no specs
20:33:36kugelLlorean: well, not in this case I suppose, the sim has no backlight
20:33:36Lloreanlucent: They haven't really been developer friendly. More like "publicity friendly" if anything.
20:33:45Lloreankugel: The sim simulates a backlight...
20:33:54lucentRichlv: Neuros is opensource friendly, but I don't know about Rockbox targets made by them
20:33:59kugelit does?
20:34:10Lloreankugel: The background changes color for it, at least on the mono targets.
20:34:13LloreanI don't know what it does on color ones.
20:34:21Lloreanlucent: There are no Neuros Rockbox targets.
20:34:37Lloreanlucent: The list of supported targets is short, and on the front page of the site.
20:34:39pixelmaLlorean: the output "backlight on or off" was removed from the default output... but should still be there I think
20:34:46Richlvargh, neuros isn't available here
20:35:00kugelLlorean: I have not noticed such a behavior on a e200 sim. But yea you're right, I seem to remember seeing that on the h300 sim on the remote
20:35:58Richlvi see apacer, nash, canyon, easy, aurum, trekstor, emtec, creative, acme
20:36:06pixelmakugel: there should be a "backlight on" and "backlight off" somewhere you could enable as debug output to the console
20:36:15kugelwill try, thanks
20:36:22LloreanRichlv: The only Rockbox targets are the ones listed on the front page of the site.
20:37:19LloreanRichlv: There's other players being worked on, which you can read about in the forum or wiki, but there's no guarantee any of those will ever work fully, or when.
20:37:50RichlvLlorean, i see. does any of the companies i listed seems more opensource friendly than others ?
20:38:00lucentkugel: brief question about OF disassembly, did you label the buttonlight routine or is there debug info in the OF?
20:38:02LloreanRichlv: That question really has nothing to do with Rockbox.
20:38:13LloreanRichlv: Please, if you have discussion that's not about Rockbox, take it to our off-topic channel
20:38:36 Quit enginerd3 (" ajax IRC Client")
20:39:42kugelpixelma: I cannot find a debug output (only debugaudio and debugwps)
20:40:17kugelbut there's backlight, I notice when I change the "first button press only turns backlight on" to yes
20:40:24***Saving seen data "./dancer.seen"
20:43:27 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
20:44:17kugeldomonoky: ping
20:46:40mooswrong windows :)
20:47:23domonokykugel: pong
20:47:42*gevaerts spots lots of rbutil people
20:47:54 Join Xerion [0] (
20:47:59kugeldomonoky: there's an issue with your button patch I heard
20:48:15kugelprobably due to disabling irqs
20:48:21bluebrotherhmm, adding flyspray ids to fixed issues in the RockboxUtility changelog looks like a good idea.
20:48:34gevaertsI've been wondering for a while if it wouldn't be better if rbutil installed the latest release on full install insteal of the current build
20:48:38domonokykugel: what issue ? and i thought disabling interrupts was not needed ?
20:48:47bluebrotherBagder: is it possible that twiki automagically links FS #1234 text to the tracker?
20:48:54kugeldomonoky: but your patch still does, as you haven't updated it
20:49:24domonokykugel: yes, i didnt find time todo anything for ams-sansas last days.
20:49:31kugeldomonoky: apparently, if you hold the down button the lcd inverts
20:50:19kugelgood to know that it works just fine without disabling irq (even better given that issue)
20:50:45CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
20:53:20kugeldomonoky: have you seen my backlight fading & buttonlight patch already?
20:54:13bluebrothergevaerts: might be a good move if we have releases regularly.
20:54:16*n1s agrees with gevaerts
20:54:55bluebrotherotoh, I'm also wondering if it would make sense to replace the "are you sure" dialog with a selection like "release, current, abort" dialog
20:56:15gevaertsI noticed a few times that people think they have 3.0, which may or may not be true. Anything that makes this more clear is fine by me
20:56:17lucentbluebrother: replace "are you sure?" with "Which release?"
20:56:44 Quit balou (
20:56:44 Quit lostlogic (
20:56:44 Quit maraz (
20:56:44 Quit liiwi (
20:56:44 Quit blippe (
20:56:44 Quit BlakeJohnson86 (
20:56:54domonokykugel: no. will take a look, when i find time. maybe this weekend.
20:56:56 Quit stsquad (Remote closed the connection)
20:57:30bluebrothergevaerts: even after the note was added for 3.0?
20:57:38 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:57:52gevaertsbluebrother: good question. I don't remember...
20:58:28gevaertsAnyway I think that we want new users to do the full install, and I think new users are better off with the release
20:58:39 Join stsquad [0] (
20:59:24pixelmakugel: revert and you'll "see"
21:00:19NJoinblippe [0] (n=none_of_@
21:00:19NJoinBlakeJohnson86 [0] (
21:00:19NJoinbalou [0] (
21:00:19NJoinliiwi [0] (
21:00:19NJoinmaraz [0] (
21:00:19NJoinlostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
21:00:44kugelpixelma: ah, thanks. no wonder I didn't know about backlight
21:04:38kugelLlorean: so, I cannot reproduce that problem in a e200 sim using my reduced keymap
21:09:32bluebrothergevaerts: something like this: ?
21:11:08kugelbluebrother: I think rbutil should always install a release, unless you checked "download developement builds" somewhere in the settings
21:12:06gevaertsbluebrother: I'd change the wording a bit, maybe "to install the most recent *development* build"
21:12:48lucentI *hate* Yes/No prompts
21:13:00lucentthey are confusing for everyone
21:13:13bluebrotherwell, if automated installation installs the latest release then the manual install should use the most recent build
21:13:16lucentmuch prefer to have verbs which say what you're doing ;)
21:13:41gevaertsOK/Cancel would work just as well for this one
21:13:43Lloreankugel: If it doesn't happen with the full keymap on targets, and it doesn't happen with the reduced one on sim, it's almost certainly somewhere in the drivers, right?
21:14:02LloreanAnd since it happens on v1, it may have something to do with the scrollwheel that you've somehow missed.
21:14:35gevaertsAlso maybe expand "This will install Rockbox 3.0" a bit to (briefly) explain again what "small" means
21:14:52bluebrothergood point.
21:14:52 Quit nk ("CGI:IRC (EOF)")
21:14:57*gevaerts thinks about good wording for this
21:15:23kugelLlorean: that would mean the e200v1 driver is faulty too
21:15:39 Join vertic39 [0] (
21:15:56Lloreankugel: It doesn't necessarily mean there's a fault anywhere.
21:16:23LloreanIt could mean that somewhere there's a check if the e200 keymap is defined, rather than a check for having a scrollwheel, probably.
21:16:40LloreanSomething that doesn't cause an error because it doesn't actually depend on a scrollwheel, just changes a behaviour when one's around.
21:16:44*pixelma wonders why Linus removed that line of code completely and not just commented out
21:17:00pixelmaseems like it proofs useful for debugging
21:17:43kugelLlorean: hm, it happens on my fuze too though
21:17:48pixelmaeh... proves
21:19:00 Join ZincAlloy [0] (n=d9eef7c0@gateway/web/cgi-irc/
21:19:02Lloreankugel: Does the fuze define the e200 pad, maybe? To make plugins work easier, perhaps?
21:20:48gevaertsbluebrother: s/ nstall/ install/
21:22:00bluebrotherfixed :)
21:22:04 Quit BigBambi (Remote closed the connection)
21:22:28kugelLlorean: no, SANSA_FUZE_PAD. plugins aren't build yet
21:22:37bluebrothernow I just need to figure why automated installation doesn't work anymore for me. Seems the version check is a bit problematic
21:23:33Lloreankugel: Well then, I've no further guesses. But if it's not reproduceable in the sim or on other targets, it's not likely going to be in apps code, then.
21:25:17 Join freddy__ [0] (
21:25:19gevaertsbluebrother: I'm not entirely happy about the text yet. It seems a bit overloaded now (I know, my fault for suggesting more explanations...). Maybe an extra linebreak between the two paragraphs?
21:26:03lucentbluebrother: I disagree with the use of "OK" though. I suggest "Install" "Cancel" buttons
21:26:50kugelyea, I'm not liking the all the text too
21:26:59kugelI found it pretty self-explaining the way it was
21:27:22 Quit fredddy (Read error: 110 (Connection timed out))
21:27:50gevaertsMaybe a compromise? Keep the "will only contain... to run Rockbox" and drop the "To get additional..." bit?
21:28:21 Join japc [0] (
21:29:10kugeljust name those "basic installation", "fully featured installation", if the previous ones aren't self-explaining enough imho
21:29:49LloreanOr "Minimum install" to go with "Full Install"?
21:29:56kugelprobably with a "find out more about installation types" link somewhere near
21:30:33 Quit japc (Client Quit)
21:30:54 Quit Thundercloud (Remote closed the connection)
21:31:45bluebrotherlucent: you have a point, though I won't change that immediately −− that dialog uses pre-defined buttons, and "Install" is none of them. Thus this change will need a bit more work than the switch from yes / no
21:33:54 Quit vertic23 (Read error: 110 (Connection timed out))
21:34:45lucentbluebrother: it's not just your problem, I realize, it's the way Win32 API does things
21:37:11bluebrotherwin32 api? This is Qt4.
21:37:31 Join tessarakt [0] (
21:38:28 Join BigBambi [0] (
21:39:48*bluebrother spots an error with bootloader install on mr100
21:40:04lucentyeah, I think Trolltech had Win32 API on their brains when they did the message box API
21:40:25kugelor they just decided to keep it simple
21:40:54kugelit's not like you're not able to make other buttons
21:42:42 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
21:44:06 Quit karashata (Remote closed the connection)
21:44:56 Join karashata [0] (n=karashat@
21:51:35 Quit Lear ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
21:52:43 Quit freddy__ (Read error: 54 (Connection reset by peer))
21:53:01 Quit kharo ("Leaving.")
21:57:40 Join Tangent1 [0] (
21:57:57 Quit jhulst (Read error: 60 (Operation timed out))
21:58:57 Quit LambdaCalculus37 (" ajax IRC Client")
21:59:45Tangent1music files should be transferred right into the root directory of my rockboxed ipod, yes? Will the formatting get mixed up if I put them in /music/artist/filename.extension format to keep it tidy?
22:00:28LloreanTangent1: You can put them anywhere you like.
22:00:48LloreanDoes one of our documents seem to suggest putting them in the root?
22:00:56 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
22:01:22Tangent1the manual says that they should be in /music/artist/filename.extension format
22:01:45Tangent1i wanted to be sure if the first slash meant root or just the folder you put the files in
22:01:57LloreanTangent1: What part of the manual says that?
22:02:18LloreanBy "the manual" you mean the PDF file? If so, could you direct me to the section?
22:02:26Tangent13.1.4 in the pdf
22:03:42LloreanI don't see a mention of a 'music' folder there.
22:03:50Tangent1that's why I came to ask
22:04:11Tangent1to see if the leading slash meant to designate the root folder specifically
22:04:20Tangent1okay, thank you much
22:04:35gevaertsTangent1: for the WPS missing tag fallback it doesn't matter (i.e. it can be as deep as you like). For album art I have no idea
22:05:25*Llorean thinks that should probably be explained more clearly.
22:05:28pixelmareminds me that I wanted to commit the patch in
22:05:46Llorean"When given a file, some parts of Rockbox assume the parent folder name is the Album name, and its parent folder will be the Artist name."
22:06:15pixelmasome parts of Rockbox?
22:06:30LloreanThe default WPS (and Album Art possibly or apparently?)
22:06:56gevaertsI'd still add the FS #9134 thing as well
22:07:30LloreanYeah, note it's not required.
22:07:36*Llorean has no ability to fix it from here, right now.
22:07:58 Join jgarvey [0] (
22:08:47pixelmait all depends on how the WPS is made - maybe cabbiev2 does... album art also looks one level up (only for cover.bmp I believe but not sure. have to look it up on the AlbumArt wiki page)
22:09:23 Quit _lifeless (Read error: 60 (Operation timed out))
22:09:41 Join _lifeless [0] (n=lifeless@
22:10:04ameyeris there some technical reason why bookmarking is really screwy for AAC? Like, set a bookmark 1:00 in and it'll resume more like 1:15 in
22:10:47pixelmaah, albumtitle.bmp also works one level up (at least wiki says so)
22:10:59 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:11:27ameyer(CFmodded iPod mini running r19441-081214 here)
22:11:37ameyererm, mini2g
22:12:26 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
22:12:53BigBambiameyer: Wouldn't it be resuming in general, not just bookmarking?
22:13:16ameyerand starting playback is laggy
22:13:17BigBambii.e. if you stop to clear the buffer then resume, does it do it?
22:13:29ameyerBigBambi: good question
22:14:02BigBambiA further question would then be how well and accurately does AAC seek?
22:14:41 Quit GodEater_ (Remote closed the connection)
22:14:49ameyerok, stopping at 13:45.
22:14:55vertic39bam I just read that the 5.5 ipod video has -50% battery life than the original firmware :/ that's not good - why is it like that?
22:14:58LloreanBigBambi: I think the accuracy depends on the file length.
22:15:02ameyerresumed at 14:00
22:15:18Lloreanvertic39: It was like that a _long_ time ago. It shouldn't be any more.
22:15:24 Quit ZincAlloy ("CGI:IRC (Ping timeout)")
22:15:27 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
22:15:32ameyerstopped at 14:10, resumed at 14:25
22:15:34BigBambiLlorean: Yes, I thought I remembered there was something slightly off with it
22:15:39*ameyer senses a pattern
22:16:10BigBambivertic39: Read where?
22:16:48ameyerseeking seems ok
22:18:57*ameyer wonders if this is just something screwy about the way this file is encoded
22:19:55pixelmaLlorean: if someone helps making these sentences more clear, I could commit it (lacking ideas how to put it in clear enough way but so that it doesn't sound too much like a "must")
22:21:35 Quit bmbl ("Woah!")
22:22:12saratogathe AAC file parser scales numbers by a factor of 10 instead of a power of 2
22:22:16pixelmaand btw. this seems to be the only place album art is currently mentioned in the manual, at least a search for "album art" in a c200.pdf didn't turn up something else... :\
22:22:31saratogawhich seems wasteful given that divides are a bit slower then shifts on most targets . . .
22:23:58 Quit t0mas ("Leaving")
22:24:04saratogaactually i guess it amkes no difference here
22:24:12Lloreanpixelma: Maybe something like
22:24:37 Join Hillshum [0] (
22:24:43ameyersaratoga: could thaat explain resume being off by like 15 seconds?
22:24:58ameyerand parsing .m4a files seems really, really slow
22:28:33 Join t0mas [0] (n=tomas@rockbox/developer/t0mas)
22:28:50 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
22:29:33 Quit {phoenix} (Remote closed the connection)
22:29:47saratogaameyer: no
22:30:03bluebrotherhmm. Am I the only one that feels the "Time and Date" menu misplaced below System? Wouldn't it be better below Settings?
22:31:15bertrikI think it used to be under system settings
22:31:38 Part Tangent1
22:32:38 Quit BigBambi (Remote closed the connection)
22:32:56gevaertsI think it can be argued that time is not a setting
22:33:16*Hillshum can see it either way
22:33:50gevaertsi.e. you don't really have a choice when setting it, in contrast to all other settables
22:34:12 Quit kachna (Remote closed the connection)
22:35:07amiconnbluebrother: You're not the only ony...
22:35:54*gevaerts would expect loud screams if the settings got changed without the user asking for that. Time is slightly different in that respect
22:36:54 Join kachna [0] (
22:36:55kugelrasher: may I ask you why the translations of the yesno screen contain the button?
22:37:30kugelhas there been desire to have the buttons translated?
22:37:37bertrikI didn't really see the point when moving the date/time to the system menu
22:38:58bluebrotherwe now have a setting below the System menu. That's kinda weird (and I dislike that "Rockbox Info" isn't the topmost entry anymore)
22:39:58pixelmaLlorean: thanks, I like it. Maybe it could be a tad bit simpler if you have an idea how to do that without losing information. If not I'm going to commit your version (also make the mentioned AlbumArt wiki page a link)
22:40:02kugeltime isn't a real setting imho. It's not in the cfg
22:40:27***Saving seen data "./dancer.seen"
22:40:32kugelthe sleep timer wasn't in the settings before either but under system, and no-one argued about that
22:41:09pixelmakugel: some lang files also translate the button name. I think that if it is common in that language to do so, why not
22:41:18Lloreanpixelma: I personally don't think it could be simpler without reintroducing confusion.
22:42:00bluebrotherthe sleep timer wasn't placed well below System either
22:42:01pixelmamost likely
22:42:07kugelpixelma: I haven't looked at all langs, but the ones where I looked at didn't translate the button. But if translating buttons is ok, then I take my question back
22:42:17bluebrotherplus, the time _is_ a setting.
22:42:52 Quit moos (Read error: 104 (Connection reset by peer))
22:43:39Hillshumwhy does my VM seem to be taking longer to compile than Cygwin?
22:43:47bluebrotherby that means it is even a persistent setting that doesn't need to be saved in a cfg. It only advances automatically :)
22:44:31kugelwell, it changes without one doing anything (as in he's not changing a setting)
22:44:50kugelit cannot be compared to settings which are in the settings menu imo
22:45:02bertrikit can IMO
22:45:53bluebrotherwhy can't it be compared? You are _setting_ it, thus it's a setting.
22:46:33kugelyou set it once, and then it changes from alone. A setting is something you set and it never changes again after until you re-set it
22:46:48 Quit karashata ("G'bye everyone!")
22:47:02bluebrotherwho says that a setting cannot change by itself?
22:47:23bluebrotherthe point is that you are _setting_ it. Not that it can't be changed otherwise.
22:47:31*pixelma notices that it's probably needed to check out the release branch too for those manual fixes to be in the release too...
22:47:36 Join BigBambi [0] (
22:47:40 Quit BigBambi (Remote closed the connection)
22:47:41kugeland I didn't deny that it's a setting
22:48:20kugelI say it's a setting which is differnet from the ones in the settings menu, and thus it's arguable if it has to be in there
22:48:44bluebrothermay I quote you from 2 minutes ago?
22:48:54 Join Zagor [242] (
22:49:06kugeli said it's not a "real" setting, yes
22:50:34 Join BigBambi [0] (
22:50:52bluebrotherby saying a setting never changes from outside you argue that time is not a setting. Now you tell that you didn't deny it's a setting. That's the opposite.
22:52:08kugeluhm yea, get picky. Anyway, I've stated my opinion
22:52:16bertrikIMO, the person who moves a setting away from the settings menu has something to explain to other people, not the other way around
22:52:34kugelit was announced on the -dev ml, you had a chance to object the placement before it was committed
22:52:44bluebrotheryou stated your two contradictory opinions. Yes.
22:53:02kugelbertrik: I think he did
22:53:50bluebrotherso what? Is it now forbidden to ask for reasoning about changes that have been made? Do I need to accept everything just because it was done some time back?
22:53:58*gevaerts thinks that selecting a playlist can be seen as setting the current playing track ;)
22:54:21amiconnYeah, and it changes itself ;)
22:54:35bluebrothergevaerts: move it to the System menu!
22:55:46 Join petur [50] (n=petur@rockbox/developer/petur)
22:57:12 Quit domonoky (Read error: 54 (Connection reset by peer))
22:58:05 Join GodEater_ [0] (n=ge@rockbox/staff/GodEater)
22:58:15bertrikI can't find the discussion about the time screen on the -dev ml
23:00:18pixelma"patch discussion - FS #9173" I believe
23:03:29bluebrotheryep, dated Nov 25
23:03:46bluebrotherI even stated my opinion that it should go below the settings menu ...
23:03:47bertrikhmm, I missed that, the title could have been a lot clearer
23:04:12*bluebrother agrees with bertrik
23:17:05 Quit bmbl ("Woah!")
23:18:40 Join ZincAlloy [0] (n=d9eec122@gateway/web/cgi-irc/
23:20:08Zagoryay, clip mp3 works flawlessly
23:20:12Zagorin the sim...
23:22:03LloreanZagor: As in, you've resolve the small-buffer issue?
23:22:52Zagorno. as in, I had a small hope of getting a crash in the sim too, but didn't.
23:23:06ZagorI'll give it a go with valgrind though
23:25:35 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
23:28:10kugelZagor: the "high mem" targets cannot play mp3 either
23:29:11kugelI suspect the clocks are set up wrong, so that we're just too slow for decoding (flac and wav play fine)
23:29:30kugelI have no idea why it crashes/reboots though instead of just skipping
23:29:33pixelmaLlorean: have a small problem with the sentence... I think it's better to have a second one that does not mention album art for targets without album art
23:29:53kugelZagor: high mem as in e200v2/fuze with their 8MB
23:29:58Lloreanpixelma: Ah, makes sense.
23:30:04Zagorkugel: if we were too slow, surely it would boost a lot more?
23:30:10kugelit boosts
23:30:19pixelmaalthough I realise that it mentioned it before but just with one word
23:30:21Zagoryeah but only very briefly
23:30:25kugelbut I haven't noticed that boosting doesn't do anything for decoding
23:30:52kugelI have forced 248MHz, and it behaves exactly the same
23:30:57Zagorif mp3 decoding was so slow it couldn't keep up, it would be boosting most of the time
23:31:23Zagorkugel: and yet you think it's clock related?
23:31:24Lloreanpixelma: It should be possible to just drop mention of Album Art and the "and" before the second part of the sentence?
23:31:25kugelas far as I understood, clocking the main cpu high isn't enough for decoding, other clocks might need to boost too
23:31:53Zagorsounds far-fetched to me. but not impossible.
23:32:08kugelthere's several controllers attached to the CGU unit
23:32:12 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
23:32:39bertrikhmm, 31 MHz seems enough to play ogg without boosting, yet 248 MHz would not be enought for mp3?
23:32:42kugelZagor: how do you else explain that flac/wav play fine, and forcing 248MHz don't help?
23:33:03Zagorkugel: codec bug?
23:33:11kugelbertrik: I suspect that 31 and 248 is equally fast for decoding
23:33:24kugelI don't think the codecs are bugged
23:33:26ameyerI haven't seen the clip actually play flac
23:33:50amiconnWell, since other targets are able to play mp3 at significantly lower cpu speed, I'd suspect a bug in the ams sansa pcm driver somewhere, which is triggered by the mpa codec
23:33:57ZagorI think performance problems would manifest in many other ways than crashing
23:34:09kugelamiconn: ogg doesn't play fine either
23:34:12ameyerI'm guessing slow/buggy SD driver, myself
23:34:19kugeljust add crossfeed and it's mostly unplayable too
23:34:32Zagorameyer: then wav would be the most buggy
23:34:34kugelameyer: if it was the sd driver flac was worse than mp3
23:34:41BigBambiameyer: flac and wav would stress the SD much more than mp3/ogg
23:34:42amiconnameyer: That would break high-bitrate stuff like wav, aiff, and lossless formats
23:34:53ameyerI can't get FLAC to play, period
23:35:01kugelI think he got it :)
23:35:36ameyerusing the buffer audio data from flash patch, ogg seems ok-ish
23:35:43amiconnIsn't there a problem with the dma not being capable of transferring a whole pcm block at once? Maybe there's a bug in the chunk handling
23:36:02ameyerFLAC dies instantly with some sort of SD-related error message
23:36:11kugelbertrik: have you tried adding some dsp in ogg? with default settings ogg plays for me too, but as soon as I add dsp stuff, it's just about as bad as mp3
23:36:17amiconnAfaik the various codecs output different sized pcm blocks
23:36:52pixelmaLlorean: yes, just getting rid of all album art related things in that sentence works without any problems
23:37:05amiconnIf you want to rule out sd driver problems, maybe try ramdisk? Hmm, probably requires working usb...
23:37:20bertrikamiconn, indeed I remember the DMA length to be quite small, so there is a mechanism to break up pcm transfers in multiple dma transfers
23:37:26ameyerto be more specific, *PANIC* (newline) SD : DATA TIMEOUT,
23:37:55Zagorwe have a ramdisk implementation?
23:38:05bertrikkugel, ogg still plays fine here at 31 MHz with crossfeed (using the flash_buffering patch)
23:38:19LloreanZagor: gevaerts developed it for USB testing.
23:38:40kugelbertrik: for me it sometimes play, but stops after 4-5 seconds mostly
23:38:52kugelbtw: I have also tried aac, it doesn't work well too
23:38:58*amiconn would suggest working bottom-up, and not introducing to many potential problems at once
23:39:11ameyerI mostly agree with kugel about ogg with the buffering patch. I wouldn't exactly call it "fine", but it's close to stable
23:39:35ameyerFLAC and mp3, not so much
23:39:39amiconnI.e. try to stabilize sd first, using plain read/write torture. When it works, proceed to pcm etc
23:40:12ameyerI *think* the flash buffering sd driver's just too slow right now
23:40:20amiconnThe flash buffering patch is something I'm very sceptical about
23:40:32kugelamiconn: as far as I see the sd driver works fine, I did various file operations without problems. Those panics only come upon playing audio.
23:40:35 Quit jgarvey ("Leaving")
23:40:41Zagoramiconn: me too
23:40:48bertrikit produces lots of nasty compiler warnings in the current version
23:40:51 Quit jhulst (Remote closed the connection)
23:41:47ameyerI agree with amiconn, fix the sd driver and buffering first
23:42:04amiconnkugel: Tried test_disk speed and write&verify tests, both boosted and unboosted?
23:42:15kugelno, not yet
23:43:52kugelbut I'm skeptical that the sd driver is broken. I rather think dma (or clocks as I already said) are causing our problems
23:45:42ameyerI haven't seen a non-SD-related error message on the clip
23:46:10ZagorI have seen some plain data access panics
23:46:31Zagorhmm there is no way to make the sim exit by itself, is there?
23:46:36ameyersomething vaguely related to the sd driver is screwy
23:46:56bertrikthis is mostly speculation
23:47:23ameyerkugel: I'd call "doesn't work correctly for flash sizes > 2 GB" broken
23:47:30ameyeror is that fixed?
23:48:17amiconnZagor: No, except on for Player and RecV1 sims, since the long-hold power button poweroff isn't simulated. Just close the window...
23:48:39*ameyer should really shut up since it's not like he can actually fix anything
23:49:19 Join moos [0] (i=Mustapha@rockbox/staff/moos)
23:49:25 Quit bertrik ("Leaving")
23:49:54Zagorvalgrind found nothing :-(
23:50:39amiconnIs there any hope for a gdb stub for any of the newer targets?
23:51:18ZagorI was thinking about that earlier today. Someone should look into that...
23:51:57*amiconn didn't use it often on SH, but the few times he did, it was quite useful
23:52:28kugelameyer: that's not the SD drivers fault. That's SanDisk's fault to use some scary bank-switching system instead of just use SDHC for the internal memory
23:53:54kugelmy microsdhc works just fine
23:54:38kugelamiconn, Zagor: what I also find weird is, that when I play music from my microsd, I still get the panics for "Internal memory"
23:55:23amiconnThat's not weird at all
23:55:53amiconnPlayback regularly updates the resume position, which means writing to the internal flash
23:56:49amiconnThis is delayed to the next "spinup" with normal buffering, but the flash buffering hits the disk very frequently, so the resume position is also written frequently
23:57:14kugelI haven't applied flash buffering
23:57:20kugelif you mean the patch
23:57:45amiconnOkay, but you'll still have rather frequent rebuffering, due to the low mem size
23:58:38kugelbut then again it works fine for flac which rebuffers fine regulary (and more often than mp3), which mp3s don't even start to play

Previous day | Next day