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 2004-05-07

00:05:00 Quit pfavr ("ChatZilla 0.9.52B [Mozilla rv:1.6/1]")
00:20:57 Quit hiroshi ("blearg")
00:24:06 Join Zoggers [0] (
00:24:41DBUGEnqueued KICK Zoggers
00:24:41ZoggersA n y 0 n e w i l l i n g t o s e l l t h e i r a r c h 0 s ? ! you want one that works?
00:36:24 Join kaboofa [0] (
00:42:11 Join midknight2k3 [0] (
00:42:46 Quit midknight2k3 (Client Quit)
00:42:47 Join midknight2k3 [0] (
00:43:08 Quit midknight2k3 (Client Quit)
00:46:00 Join midknight2k3 [0] (
00:46:19midknight2k3hey amiconn?
00:46:27amiconnHi midknight2k3
00:46:35midknight2k3hey got grayscale working!
00:46:39midknight2k32 questions
00:46:46amiconnSounds really good.
00:47:43midknight2k31) does gray_position_display(0,1); set the *whole* grayscale overlay down a row?
00:48:30midknight2k3IE if i call gray_drawpixel(0,0); it will put a gray pixel at 0, 8?
00:48:42midknight2k3when the overlay is done a row that is
00:49:05amiconnYup. gray_position_display(0,2) would place it two rows down etc.
00:49:50amiconnBut: gray_drawpixel(0,0); will always put the pixel at (0,0) _in the grayscale buffer_
00:50:29amiconnThink of the grayscale buffer as a "window" that can be moved with gray_position_display()
00:51:53amiconnAll coordinates given to drawing functions are relative to this window.
00:53:32amiconnEven if moving the overlay down clips it at the bottom, this is only visual clipping, so if you move it up again, the clipped content will reappear.
00:53:47midknight2k3is there any way to have a bool option on that such sa "translate"
00:54:05midknight2k3anything outside of it wouldn't be drawn, as opposed to it all getting moved down?
00:54:27midknight2k3as is i'd have to move *all* of my drawing functions up 8 pixels
00:55:30amiconnThis is no translation at all: The grayscale buffer stays the same regardless of the gray_position_display() parameters.
00:56:17midknight2k3right but the problem is that now the window is down one rpw
00:56:20amiconngray_position_display() does only change the origin where the overlay is copied into the lcd hardware.
00:56:25midknight2k3so i have to move all of my drawing up 1 row?
00:56:34midknight2k3then wouldn
00:56:41midknight2k3t everything else move down too?
00:57:29 Join Nibbler [0] (
00:57:41amiconnUnfortunately I think that you will have to move your drawing one row up.
00:57:58midknight2k3correct, is there *any* way to keep everything the same but clip the top row off?
00:58:08midknight2k3i need it down one row because the score is written there
00:58:18midknight2k3and i need it basically where it's at because of all my drawing
01:00:20amiconnHmm, I think this is not easily possible, or feasible. You could change the drawing functions you use to subtract 8 from y before doing anything else, but then it would be hard to update the framework later.
01:02:12amiconnThere is (currently) no way to clip the overlay dynamically. I once thought about implementing that, but that would render the interrupt routine much more complex -> costs speed.
01:02:44 Join michaelconner197 [0] (
01:03:17michaelconner197hey all... got a question
01:03:36michaelconner197anybody know how to make rockbox (v2.2) save recording prefs? (levels, etc.)
01:06:02midknight2k3michaelconner197, it does
01:06:14midknight2k3just make sure you hit stop before shutting off
01:06:20midknight2k3try updating to a daily build
01:06:40midknight2k3amiconn, ok question 2: how can i use the "stride" parameter on a normal bitmap?
01:07:25michaelconner197midknight −− will try that, i'm using the first relase of 2.2.
01:07:31 Quit michaelconner197 ("Leaving")
01:08:12amiconnmidknight2k3: Hmm, have to take a look, that "stride" parameter was introduced by [IDC]Dragon.
01:09:07midknight2k3i spent about a half hour trying it out - no luck
01:09:08***Saving seen data "./dancer.seen"
01:09:28amiconnThat reminds me that I should really commit this (surrounded with #if 0 so that it does not compile for distribution)
01:10:23amiconnGrr, I don't have this "stride" within my internal version - have to merge.
01:12:32 Quit AciD ("Read error: 666 (Connection reset by Satan)")
01:12:52amiconnmidknight2k3: For showing the whole bitmap, you have to set stride = nx
01:17:35midknight2k3i have a 112x64 fullscreen bitmap i want to use
01:17:41midknight2k3how shall i call teh bitmap function?
01:18:12amiconnGrayscale bitmap or monochrome?
01:18:33Zoggersanyone want to sell their archos?
01:18:35 Join `vinman [0] (
01:19:19`vinmanany you guys have a site to get premade rvf files ?
01:22:06scott666checking the mailing list archives
01:25:27 Quit Zoggers ("Leaving")
01:27:14 Join TheDude2 [0] (
01:30:03 Quit `vinman (
01:30:03 Quit silencer (
01:30:03 Quit kaboofa (
01:30:03 Quit ricflair (
01:30:03 Quit jkerman (
01:30:03 Quit Nibbler (
01:30:03 Quit matsl (
01:30:03 Quit c0utta (
01:30:03 Quit Ka_ (
01:31:11 Join Threshold [0] (
01:31:11NJoinsilencer [0] (
01:31:59NJoinc0utta [0] (
01:33:29 Join abinitio [0] (
01:35:24midknight2k3sorry, back?
01:35:28midknight2k3sorry, back.
01:35:40amiconn[01:18:15] <amiconn> Grayscale bitmap or monochrome?
01:35:54midknight2k3monochrome at the moment
01:36:28midknight2k3which will lead me to question 3 after this one is answered :D
01:36:29NJoinkaboofa [0] (
01:37:30abinitioI just bought the jukebox recorder and have installed the rockbox 2.2. Wonderful really
01:37:46midknight2k3daily builds have improvements
01:37:56midknight2k3but happy you like it :)
01:38:46abinitioI use a mac and am wondering if there is a way to build playlists with an interface on the desktop
01:38:56amiconnmidknight2k3: So you would have to use gray_drawbitmap(bitmap, 0, 0, 112, 64, 112, true, <fg_value>, <bg_value>);
01:39:44midknight2k3amiconn: so for b&w, fg value would be 0 and bg would be 255?
01:40:14amiconn- if you want it opaque, that is. If you want a transparent background, use "false" as the 7th parameter, <bg_value> doesn't matter then.
01:40:35midknight2k3fg is the bitmap itself
01:40:41midknight2k3and bg is whitespace around it?
01:40:47 Quit abinitio ("Leaving")
01:41:48amiconnfg is used for all pixels whose bits are 1, while bg is used for all pixels whose bits are 0 (if draw_bg == true)
01:43:17midknight2k3and one more
01:43:30midknight2k3is there going to be a converter from gray or rgb -> drawgraymap format?
01:44:11amiconnI did not yet write one
01:44:48midknight2k3i meant will you/somebody else?
01:45:58amiconnBeware that the format needed by gray_drawbitmap is different from the format produced by bmp2rb as well. But I'm just about to change that (it doesn't really make sense to have 2 different internal bitmap formats)
01:46:55amiconnIf I find the time, I may write a converter based on bmp2rb.
01:47:10midknight2k3i love the new drawgraymap bitmap styles
01:47:24midknight2k3i was considering creating something like it for lcd_bitmap
01:47:31midknight2k3ie { 1,0,0,1,
01:47:42midknight2k3. .
01:47:43midknight2k3 ..
01:48:16midknight2k3i even created my graymaps from scratch - not something i enjoy doing with bitmaps :)
01:48:54amiconnYes, creating bitmaps for grayscale is much easier because it uses one byte per pixel.
01:49:49amiconnAnyway, if we would use that for b&w bitmaps as well, this would waste much space (requires 8 times the memory)
01:50:03midknight2k3hmm yes...
01:50:23midknight2k3or maybe just add in a lcd_drawmap function
01:50:29midknight2k3so if you need to use it you can
01:50:33midknight2k3for grayscale icons or whatever
01:51:06amiconnIt _may_ be possible to do that with bitfields (much easier to read for humans, same memory requirements for the machine)
01:51:55amiconnHowever, this would require that lcd_bitmap() has to be converted to use the format of gray_drawbitmap() and not vice versa.
01:52:06midknight2k3i meant a new function
01:52:14midknight2k3that would use the gray format
01:52:22midknight2k3and keep lcd_Bitmap for the current bitmaps
01:52:47amiconnNo, that new function would have the drawback of wasting memory. I meant _defining_ the bitmap with bitfields.
01:53:02NJoinKa_ [0] (
01:53:32amiconnThis would make the definitions almost as easy to read as these for gray_drawgraymap() while not needing more memory.
01:54:33midknight2k3i don't get it, but ok. :D
01:55:19amiconnI will have a closer look at bit fields, then my next version of the grayscale framework will contain an example of this.
01:55:35midknight2k3thanks for all your work
01:55:49midknight2k3breakout i think will be 10x better using your functions :)
01:55:55amiconnI'm currently in the process of polishing the framework for committing it.
01:57:01amiconnBtw: ALthough not yet shown on the rockbox home page, I have committed a speedup patch for the lcd routines (+8% on recorder).
01:58:07amiconnLocally, I've already done a major rework of the lcd driver, which yields another +10% on the recorder
01:58:30midknight2k3wait, will we now be rb->drawgraymapping?
01:58:36midknight2k3or still using framework in plugins?
01:59:41amiconnGrayscale is still a plugin-only thing, but it uses the lcd routines of rockbox (namely rb->lcd_blit()), so it profits from these speedups
02:01:36 Join BlauChip [0] (
02:03:24midknight2k3lol bc
02:03:28midknight2k3new nick every time
02:03:37amiconnHallo BlauChip
02:03:51BlauChipHallo Ami ;)
02:04:19BlauChipPaying respect to my many German friends today :)
02:04:49midknight2k3umm bc i'm not german.
02:05:32BlauChiphmmm, Do the US's have an interesting spelling of the colOR blue?
02:06:19BlauChipbut it is a color and not a colour?
02:07:33BlauChipdyslexia lures
02:07:56amiconnBlauChip: I recommend azul (this is Spanish)
02:08:25BlauChipkinda loses the BC'ness though :(
02:10:09BlauChipanyone here play Othello - and is any good at it?
02:12:02midknight2k3well i'm okay
02:12:08midknight2k3not a pro
02:12:31BlauChipjust wrapping up v2 -and fancy a new AI bot - but i'm not good enough to know what it should do
02:12:44midknight2k3mm don't ask me. :)
02:18:54amiconnmidknight2k3: Ooooops, sorry I was wrong with the stride parameter!
02:19:43amiconnThis is counted in whole bytes, so it is not 112 for full width, but 14
02:20:39midknight2k3i tried 8, 13, 112, 111, 5, a whole multitude of possibilites
02:21:18amiconnI just found that while integrating it into my local version.
02:22:36 Quit midknight2k3 ("Going... BYE!")
02:22:47 Join midknight2k3 [0] (
02:27:01BlauChipanyone have a calculator that is able to realise that (1/3)*3=1 ?
02:27:29amiconnMy brain ;)
02:28:04BlauChipand is also able to produce the root of 2 accurately and quickly
02:28:52midknight2k3OH WAIT
02:28:59midknight2k3in school someone was trying to prove that
02:28:59amiconnThis cannot be done exactly with numeric method, since sqrt(2) is an irrational number.
02:29:12 Quit Threshold ("L8r I am outta here!")
02:29:14midknight2k3oen guy was saying 0.3333.... * 3 = 0.999999.....
02:29:18BlauChipi posted the method on the mailing list the other day
02:29:39midknight2k3what are bc's thoughts on the subject?
02:29:55BlauChipon root 2, or rounding errors?
02:30:19midknight2k3does 0.3333333333333333 * 3 = 0.99999999999999999999 or 1?
02:30:34BlauChipdepends on how you evaluate it
02:30:41midknight2k3.3 with a line over it.
02:30:44midknight2k3* 3
02:30:50midknight2k3.9 with line over it
02:30:51midknight2k3or 1?
02:30:59DBUGEnqueued KICK BlauChip
02:30:59BlauChipdepends on how you evaluate it
02:31:37BlauChipif nothing else it is proof that you should learn to deal wtith fractions
02:32:10BlauChip1 over 3 times 3 is 1
02:32:24midknight2k3i know that was me + wes' point
02:33:30BlauChip.3 recurring time 3 is .9 recurring ....correct, but generally unhelpful
02:35:59BlauChipthe "best" solution is to evaluate the result to one decimal place further than you need and round
02:36:17midknight2k3just btw, nm == never mind :)
02:37:00BlauChipoh! I thought it meant "no" mind ...I shall stop nodding :9
02:37:32amiconnmidknight2k3: Already tested gray_drawbitmap with the correct <stride> value?
02:38:04BlauChipI didn't realise that when you said you read that link, that you were trying to imply that you didn't understand it
02:39:44midknight2k3amiconn: nope, can't right now
02:39:47midknight2k3tomorrow i shall
02:40:03midknight2k3bc: it didn't help is all
02:40:06midknight2k3not really
02:40:23BlauChipstride is the width of the data which holds the sprite
02:40:55midknight2k3i know but it was not working
02:40:59midknight2k3does *that* make sense?
02:41:10CtcpIgnored 4 channel CTCP requests in 20 hours and 56 minutes at the last flood
02:41:10*BlauChip nods feverently
02:41:20amiconnBlauChip: Yup. I did not integrate this extension (made by Jörg) into my local copy until now.
02:42:09amiconnNow it is integrated, as I'm preparing it for commit.
02:42:24BlauChipdo you have a similar thing on the Y coord?
02:42:34midknight2k3no need
02:42:54midknight2k3if you tell it the x stride the y stride must only be 1 value
02:43:04BlauChipso no windows then
02:43:06amiconnCompilation will be disabled by default (via #if 0) to prevent having it in the distribution.
02:43:40BlauChipwindows -> view-ports
02:43:42amiconnThe thing is that you don't need the stride parameter for partial display in y direction.
02:43:53midknight2k3bc there is no such thing as a y stride
02:44:17midknight2k3an x stride tells it the width −− thus it can simply make new y columns until it runs out of data
02:44:56BlauChipHow do you display a specific section of a 500x500 sprite in the top right 25x25 corner of the screen?
02:45:14midknight2k3you say that the width is 25 and the height is 25
02:45:22midknight2k3and the stride would be 500 still
02:45:47BlauChipspecific section
02:45:50midknight2k3*hopes sincerely that his bluffing is convincing
02:45:59BlauChipsadly not
02:46:00midknight2k3change the x stride
02:46:10amiconnBlauChip: Just let your data poiter point to the pixel you want top left, set x, y, nx, ny as you want it. stride will still be 500
02:47:12BlauChipcan you point midway through a byte?
02:47:45amiconn?? Of course not. I thought grayscale here, where 1 pixel == 1 byte.
02:48:57amiconnFor b&w this would be much more difficult to implement.
02:49:59midknight2k3bad bc bad
02:50:14BlauChipstill contemplating a full graphics engine
02:50:25midknight2k3ooh like 3 + d === #D
02:50:28BlauChip1bpp would be expensive
02:50:33midknight2k3*can do basic math
02:50:52midknight2k3what do capital numbers look like?
02:51:16BlauChipnot really relevant until you get to advanced differential calculus
02:51:21midknight2k3hey here's a good joke
02:51:40midknight2k3"can rockbox handle bools?" "only if they return TRUE!" HAHHAHAHAHAHAHHAHAHHAHAHAHA
03:07:01midknight2k3did nobody get the joke?
03:09:12***Saving seen data "./dancer.seen"
03:14:44midknight2k3umm fine then
03:17:25midknight2k3rebooting brb
03:17:50 Quit midknight2k3 (Read error: 54 (Connection reset by peer))
03:19:32 Join midknight2k3 [0] (
03:20:02amiconnGuess I should get some sleep now, already past 3 a.m. here.
03:20:13BlauChipwhere are you?
03:20:29amiconnBerlin, Germany.
03:20:58BlauChipright, I keep forgetting about the weird BST thing
03:21:19amiconnNite all
03:21:23 Part amiconn
03:22:35 Nick midknight2k3 is now known as midk|FOODY (
03:22:40midk|FOODYdid you like my joke bc
03:22:55BlauChipnot sure i understood it
03:25:35 Join Nibbler [0] (
03:25:53BlauChip(kettle) brb
03:45:32midk|FOODYback in a bit
03:45:36 Quit midk|FOODY ("Going... BYE!")
03:45:56 Join infamouse [0] (
04:10:58 Join midk [0] (
04:10:58 Quit Nibbler (Read error: 54 (Connection reset by peer))
04:28:33 Quit infamouse (Read error: 110 (Connection timed out))
04:45:15BlauChipnite all
04:45:24 Part BlauChip
04:50:56 Quit TheDude2 (Read error: 54 (Connection reset by peer))
04:55:33 Join jkerman [0] (
05:07:03 Join infamouse [0] (~Aaron@
05:07:38infamousehow can I tell what kind of battery I have?
05:08:41infamouseit's set in rockbox to 2200 mah
05:08:52infamousebut I don't think that is what comes stock
05:08:54scott666are you using a v2/fmr?
05:08:55infamouserecorder v2
05:09:15***Saving seen data "./dancer.seen"
05:09:18scott666you have a 2200 mah, non-replaceable, Li-Ion battery
05:09:45midkgee ty scott
05:09:49midki can't handle things
05:09:51midkplease take over.
05:09:59scott666way ahead of you midk
05:10:16infamousehmmm, really....I thought the stock batteries were 1500 mah ones
05:10:28scott666thats on the recorder v1
05:10:37scott666those are NiMH AAs too
05:11:03infamouseso there is no way to replace the ones I have in my recorder now/
05:12:05scott666well it is possible to replace them, its just that archos doesnt sell the battey
05:12:22infamouseoh, as opposed to on the v1, they did?
05:12:45scott666archos didnt, but you can get NiMH AAs at most electronics stores
05:14:02infamouseoh, but if my batteries were to ever fail and I needed to replace them, I could go to an electronics store and get some, correct?
05:14:08midkwrong there
05:14:11midkyou CAN replace the batteries
05:14:25midksells replacements for about $20
05:14:32midkthey're like 2800mah
05:14:41DBUGEnqueued KICK scott666
05:14:41scott666:13 PM> <scott666> well it is possible to replace them, its just that archos doesnt sell the battey
05:15:14infamouseok, I'm just trying to make sure that it is possible to replace my battery in the event that I need to
05:15:14midkyou didnt say WHERE
05:15:21midkyes you can infamouse
05:15:24scott666he didnt ask
05:15:39scott666you shouldnt have to replace the batteries anyway
05:15:56infamouseanyway, speaking of batteries...I can't seem to find trick charge or deep discharge in any of the settings
05:17:05infamouseare those settings available for the v2?
05:17:24scott666all charging is done with hardware on the v2
05:17:40midkVIA is better than WITH
05:18:39 Quit scott666 (Excess Flood)
05:18:55 Join scott666 [0] (
05:19:21infamouseahh, so is there any harm in charging my v2 when it still has half it's charge left? is that going to decrease the life of the battery?
05:19:28scott666infamouse: did you get that pm?
05:20:04scott666booted for flooding
05:20:08scott666ill dcc it
05:21:13infamousebooted for flooding in a PM?
05:22:55infamousescott666: I like it
05:23:09infamousewould have answered most of my questions
05:23:17scott666any more questions just ask and ill add them to it
05:23:20infamousebut is there any harm in charging my v2 when it still has half it's charge left? is that going to decrease the life of the battery?
05:25:30scott666li-ion batteries have no memory effect, youll be fine\
05:25:36scott666(added to the faq)
05:26:12infamousewell don't li-ion have a very SMALL memory effect? but one I won't probably notice, correct?
05:26:44scott666in the faq i worded it 'virtually no memory effect.'
05:27:41infamousegood deal
05:27:49infamousewhat exactly is a UCL file?
05:30:02infamouseI thought it was a flash file, but that flash page got me a little confused
05:31:05scott666ucl is a form of compression thats used by rockbox to fit into the flashrom
05:31:21scott666until you flash, you dont need to worry about them
05:31:30scott666though i do recommend flashing
05:31:37infamouseI'm thinking about it...I hate waiting for it to boot up
05:50:02 Join Tenaenae [0] (
05:50:02 Quit Tenaenae (Remote closed the connection)
05:53:15 Join Nibbler [0] (
05:55:38 Quit infamouse (Read error: 60 (Operation timed out))
06:02:56 Quit scott666 ("i'll be back...eventually...")
06:11:53 Join infamouse [0] (
06:11:59infamousewhat's the point of upside down mode?
06:12:33midkto use it upside down?
06:12:54infamouseI didn't know if there were certain situations where you would need it
06:14:12midknot sure heh
06:14:15midkfor fun i gues
06:15:51infamouseis the ajbrec.ajz in the flash zip file, version 2.2?
06:16:29midki don't know
06:20:23infamousewhat if I were to play the ucl file
06:20:26infamousewhat would happen
06:20:43infamouseand why isn't the ucl file included in the flash zip if it has to do with flashing?
06:34:16infamousewell, anyway, I just flashed and this is awesome!
06:34:54midkinfamouse: get the ucl from a daily build
06:35:30infamousewon't that make it so it flashes with the daily build?
06:36:35midkgot to go
06:36:40midkthat's what you want
06:36:47 Quit midk ("yo yo yo cya later YO YO YO wasa wasa!")
06:36:49infamousebut the daily builds aren't the most stable....
07:09:17***Saving seen data "./dancer.seen"
07:45:06 Quit Nibbler (Read error: 54 (Connection reset by peer))
07:51:12 Quit infamouse (Read error: 60 (Operation timed out))
07:53:38 Quit dwihno (Read error: 60 (Operation timed out))
08:01:03 Join dwihno [0] (~dw@
08:29:31 Join amiconn [0] (
09:09:05 Join Nibbler [0] (
09:09:21***Saving seen data "./dancer.seen"
09:11:40 Join methangas [0] (
09:18:43 Join AciD [0] (
10:38:03 Join jj2dogs [0] (
10:38:56jj2dogsHi all... I must say I love the rockbox..
10:46:39jj2dogsAbout the new 2.2 there is a plugn flash_firmware.. How do I flash it with rockbox??
10:48:14jj2dogsAnd why it isn`t possible to use it as a usb disk for pda??
10:53:06 Join [IDC]Dragon [0] (
10:53:16amiconnHi Jörg
10:53:48[IDC]Dragonjj2dogs: read for instructions on flashing
10:54:10[IDC]DragonJens or script-bot-autogreet?
10:55:19amiconnI will repeat myself: There is _no_ bot involved, only that my irc client informs _me_ by beeping/flashing if some special people logon.
10:56:58amiconnBtw, I have a question. It might be that only Linus could answer it, though.
10:57:45amiconnYesterday I told you that I want to change the lcd driver to also use the unrolled transfer loop.
10:58:22[IDC]Dragonyes, please go ahead
10:58:29amiconnI have already done that locally, and the speed gain even exceeds my expectations: it gives another +10% for the recorder.
10:58:47[IDC]Dragonthe 3 byte command are so slow?
10:59:09jj2dogs[IDC]Dragon: So I can`t flash a player??
10:59:24[IDC]Dragonjj2dogs: currently no.
10:59:26amiconnHowever, some time ago I talked about this with Linus, and he asked me to stay with a C file. I wanted to convert it to an asm (.S) file.
10:59:59jj2dogs[IDC]Dragon: Dam... The rock box is much better that the original firmware..
11:00:42[IDC]Dragonjj2dogs: many players also have a not flashable chip. did you check yours?
11:00:53jj2dogsAnd is it possible to aply some patch to player??
11:01:07jj2dogs[IDC]Dragon: Where do I find it??
11:01:30amiconn[IDC]Dragon: I _have_ done it in C, but it looks very strange indeed, and definitely _is_ compiler dependent.
11:01:44jj2dogsI red the patch.html but I am not wise from it...
11:01:47[IDC]Dragoninfo -> debug-> h/w info
11:02:14amiconnIt would look much more straightforward in pure asm, be much more logical at the same time, since there is no single C instruction left.
11:02:28amiconnIt would also save some more bytes of precious IRAM.
11:03:35amiconnThe only reason why Linus asked me to stay with a C file is that if I change it to .S, the connection to the history of lcd.c in cvs would not be obvious.
11:04:05amiconnHowever, you have done this for memcpy.S and memset.S some time ago...
11:04:33[IDC]Dragonmaybe that's the reason Linus didn't like it...
11:04:58[IDC]DragonI guess you'd have no problem sticking with C
11:05:18[IDC]Dragonit's not the most pretty, but will work.
11:06:14amiconnI still think switching to .S would be the far better way. If you want to have a look at what I have, look at">
11:09:25***No seen item changed, no save performed.
11:12:40[IDC]Dragonlooks a bit stange, your asm subroutines outside of any function
11:13:31[IDC]Dragonand what is this SN macro?
11:16:53 Quit jj2dogs ("Leaving")
11:17:31 Join limbus [0] (
11:26:08amiconn(sorry for being afk)
11:27:01amiconnThis SN macro is used to convert numeric constants like LCD_DS etc. into strings, since the asm() pseudo-function needs it that way.
11:33:20[IDC]Dragonlots to learn from your code, great
11:33:39[IDC]DragonI think I was passing it as an argument to overcome this
11:34:45amiconnThe problem is that you can't use the extended asm() syntax (the one with :::) outside of functions, so I had to figure out this.
11:36:31[IDC]Dragonwhy do you want it outside of a function?
11:38:05amiconnBecause I want to have pure asm subroutines. I could put this into a dummy function (this is what Linus did in the gdb stubs), but this would generate additional instructions that would never be used.
11:39:22amiconnFor information about how the SN macro works, see and
11:44:01[IDC]Dragonin general, I know about # and ## of the preprocessor
11:45:30[IDC]Dragonwhat prevents you from committing your LCD code?
11:46:24amiconnAs I said before, this looks very strange and would look much more straightforward in .S form.
11:46:54[IDC]Dragonwe can still do that.
11:47:09amiconnFurthermore, I wanted to wait a bit to see if my previous optimization breaks something on the player.
11:47:34[IDC]Dragonour player user just left...
11:47:39amiconnWhile I tested the player routine by using it for the recorder, I don't have a real player to test.
11:48:02[IDC]DragonI broke the one i have
11:48:38amiconnWhat is broken? Can it be fixed?
11:49:09[IDC]Dragonif I knew, I'd fixed it
11:52:46 Quit methangas (" HydraIRC -> <- irc client ownage!")
11:53:14amiconnIs this the one that broke while testing the flash procedure for players?
11:56:22 Join methangas [0] (
11:56:52[IDC]DragonI did the serial and uart boot mod
11:57:06[IDC]Dragonthen it stayed blank :(
11:57:18[IDC]Dragoneven after I undid the mods
11:57:45[IDC]Dragonbut the box is alive, I can control it via uart boot
11:58:13amiconnDoes that mean that "only" the display does not work?
11:59:21[IDC]Dragonno, the startup somehow hangs
12:00:25amiconnme too.
12:17:32 Quit [IDC]Dragon ("no fate but what we make (EOF)")
12:37:27 Join Strath_ [0] (
12:51:02 Join [IDC]Dragon [0] (
12:51:11amiconnre Jörg
12:52:18[IDC]Dragonhi again
12:54:26amiconnIiuc, your dead player does boot correctly via uart, but not from ROM?
12:55:10 Join Quelsaruk [0] (~jirc@
12:55:29[IDC]Dragonit doesn't boot into rockbox, but I can control it to some extend
12:55:59[IDC]Dragon(blink the LED, toggle the LCD post lines, read and program the flash)
12:56:11 Quit Strath (Read error: 110 (Connection timed out))
12:57:06[IDC]DragonRockbox can't be cold-started on a player, that's what I originally wanted to debug
12:57:14amiconnI know.
12:58:06amiconnWhat happens if you write the original firmware back to the rom?
12:58:36[IDC]Dragonthe LCD lines don't move when I start the firmware, so I suspect it crashes before.
13:00:41amiconnido you have the original rom image of that very box? Perhaps somehow the contents of the file did change a bit, did you try another rom image?
13:01:48[IDC]Dragonsure I do have the ROM
13:02:08[IDC]DragonI'm pretty confident I restored that, too
13:05:16amiconnWhat about the hardware mask?
13:05:57amiconnAnd: is it possible to build a debug version of the bootloader?
13:09:26***Saving seen data "./dancer.seen"
13:12:17[IDC]Dragonplayers don't have a hardware mask
13:12:30[IDC]Dragonit's all done by the version number
13:12:50[IDC]Dragondebug version of boot loader? what for?
13:38:23Quelsaruklunch time, cu
13:38:26 Quit Quelsaruk ("Leaving")
13:43:08 Join CGI047 [0] (
13:43:14 Nick CGI047 is now known as mattzz (
13:51:52amiconn[IDC]Dragon: If you could flash a debug version of the boot loader, you may be able to detect the exact spot where it crashes.
13:52:26[IDC]Dragonthe boot loader doesn't crash
13:52:35[IDC]Dragonthe Archos firmware does.
13:55:33amiconnIs there a disassembly of the Archos firmware for the player? If so, and you have a LA available, it might be possible to find out where it crashes.
13:57:33[IDC]Dragonphew, that sounds like a lot of work
13:58:05[IDC]Dragonthen I'd better buy one from ebay
14:10:10limbusHI all, a coworker showed my his iPod yesterday. well, not much to say about that, apart from the fact that I saw a game called "Music-Quiz" that plays parts from music files and then, of course, the player has to guess where it comes from.
14:10:34limbusSomebody knows if there are any license / patents problems rebuilding that game for rockbox ?
14:10:58limbusit would be one more game possible for the players.
14:11:42mattzzlimbus: did you check google about apple, ipod, music-quiz?
14:12:12amiconnlimbus: Imo, if you re-implement it without knowing the original source and name it differently (in case the name itself is patented) there should be no problem
14:14:06limbusok, I'll think abou that
14:17:28limbusdid I mention that the "buttons" of the iPod are REALLY cool ?
14:17:33limbusno moving parts
14:21:08 Quit methangas (" HydraIRC -> <- s0 d4Mn l33t |t'z 5c4rY!")
14:58:16 Quit AciD (Read error: 54 (Connection reset by peer))
15:03:30 Join pfavr [0] (
15:09:28***Saving seen data "./dancer.seen"
15:38:08 Join AciD [0] (
15:54:29mattzz[IDC]Dragon: r u listening?
16:15:16 Quit AciD (Read error: 54 (Connection reset by peer))
16:18:15 Quit mattzz ("CGI:IRC")
16:28:43 Join methangas [0] (
16:31:45 Quit methangas (Client Quit)
17:03:24 Join AciD [0] (
17:09:33***Saving seen data "./dancer.seen"
17:27:42 Join Ka__ [0] (~tkirk@
17:42:03 Quit pfavr ("ChatZilla 0.9.52B [Mozilla rv:1.6/1]")
17:45:42 Quit Ka___ (Read error: 110 (Connection timed out))
17:46:59 Join deadite66 [0] (
18:13:40 Part limbus
18:32:24 Quit deadite66 (Read error: 110 (Connection timed out))
18:58:14 Quit [IDC]Dragon ("no fate but what we make (Session timeout)")
18:58:14 Quit Nibbler (Read error: 104 (Connection reset by peer))
18:59:09 Join estesoyyo [0] (
18:59:30 Quit Urgo (Read error: 60 (Operation timed out))
19:03:23 Join Urgo [0] (
19:09:36***Saving seen data "./dancer.seen"
19:12:17 Part estesoyyo ("Usa Jabber, es libre")
19:17:19 Join Strath__ [0] (
19:35:43 Quit Strath_ (Read error: 110 (Connection timed out))
19:37:33 Join wake [0] (
19:42:34 Quit AciD (Read error: 60 (Operation timed out))
20:12:02 Quit wake ("leaving")
20:38:35 Join Nibbler [0] (
21:09:40***Saving seen data "./dancer.seen"
21:12:44 Join BLuECHip [0] (
21:17:52BLuECHipEvenin' all
21:25:40 Join wake [0] (
21:31:13 Nick Strath__ is now known as Strath (
21:34:11 Join AciD [0] (
21:36:12Strathhrm.... * is the third more active viewer of my website :)
21:41:46 Quit wake ("leaving")
21:52:07 Join scott666 [0] (
22:07:09 Join amiconn_ [0] (
22:14:20 Quit amiconn (Nick collision from services.)
22:14:20 Nick amiconn_ is now known as amiconn (
22:37:57 Join [IDC]Dragon [0] (
22:38:19amiconnre Jörg
22:38:23[IDC]DragonHi Jens
22:38:33[IDC]Dragondo you have the ne cvs set up?
22:39:55amiconnNot yet. I have mailed Bagder, but did not yet receive an answer.
22:40:07[IDC]Dragonneither did I
22:40:29[IDC]Dragondunno if he will answer, or silently set it up
22:40:45amiconnIn the meantime, I have converted my new lcd code to an asm file. Looks much cleaner and saves 20 more bytes of IRAM.
22:41:13amiconnIf you want to have a look,">
22:41:36[IDC]Dragonso I can plut it in my private build
22:44:05amiconnYesterday I brushed up my grayscale.c to prepare it for commit (squash that <tab> characters, integrate your <stride> parameter). Now I find that I cannot commit at all.
22:44:44[IDC]Dragondo you use the time API?
22:46:10[IDC]Dragonanonymous cvs works (read only)
22:48:29amiconnI have optimized the clipping at the right border for gray_drawgraymap while integrating <stride> support.
22:48:30[IDC]Dragonmy user account not .-(
22:49:07amiconnNeither does mine: "rejected access to /cvsroot/rockbox for user ..."
22:49:07[IDC]Dragonwhatwas happening there?
22:49:31amiconnWith the clipping?
22:51:04[IDC]Dragonthe new cvs is pretty fast
22:52:04amiconnWithout <stride> support, it was necessary to let the x-loop always run all the way to the right, even if some pixels are not drawn due to clipping, otherwise the position of the data pointer would not be correct for the next row.
22:53:05amiconnWith stride support, the data pointer for the current row is discarded at the end, so it is possible to quit the loop when reaching the right border.
22:57:19[IDC]Dragonyes, this is common practice.
22:59:08 Join timd0ug [0] (~timdoug@
23:01:54amiconn[IDC]Dragon: I still don't understand why the Rockbox crew switched to an own cvs.
23:02:06amiconnApart from sometimes being a bit slow, I never had problems accessing sourceforge cvs since I was allowed to commit.
23:02:12amiconnAuthentication with ssh public key was much more convenient and secure as well.
23:02:33amiconnAt least I hope this will be integrated into the cvs as well.
23:02:52[IDC]Dragoni haven't managed that
23:03:07[IDC]Dragontried, but didn't work
23:03:20[IDC]Dragonfor another project it did
23:04:19amiconnI have done it by the manual, worked the first time I tried it. I took me a bit of research to get the ssh daemon running (with cygwin). Still not perfect, but it works.
23:05:26[IDC]Dragonahh, my gdb works again
23:06:41amiconnHmm, gdb is one thing I have never tried using yet. I suppose I would need a serial connection for that?
23:07:31[IDC]Dragonbut under windows, it's very unpleasant (=unstable)
23:07:38amiconnThis is too bad, since I don't want to mod my box. Apart from that, my laptop does not feature a searial port.
23:08:10[IDC]Dragonmaybe a USB converter would do
23:09:43***Saving seen data "./dancer.seen"
23:14:07 Join top_bloke [0] (
23:33:04 Join cjnr11 [0] (
23:33:09 Part cjnr11
23:44:32amiconn[IDC]Dragon: I had a look into the cvs manual. Renaming a file _does_ link the history.
23:44:33 Quit top_bloke (Read error: 54 (Connection reset by peer))
23:44:50amiconnIt is even possible to make the revision number start with the value of the old file + 1.
23:45:26[IDC]Dragonah, ok. I didn't rename memcpy/memset
23:45:43amiconnMaybe I can convince Linus that it would be better to switch to lcd.S
23:45:55[IDC]Dragonit seems to bother you...
23:46:27CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:46:27*[IDC]Dragon can't reproduce the voice problem with gdb
23:46:39amiconnIf you did compare my lcd.c with lcd.S, you should know why.
23:47:23amiconnRenaming a file in cvs is simple (basically, it is a "linked" commit)
23:47:53amiconn(1) Rename (or replace) the file locally
23:48:01amiconn(2) cvs remove old
23:48:07amiconn(3) cvs add new
23:48:23[IDC]Dragoncan we post-mortem rename memset.c into memset.S ?
23:48:34amiconn(4) cvs commit -m "Replaced old with new" old new
23:49:06[IDC]Dragonyou do commant line cvs?
23:50:29amiconnOf course I do. Since I did have to install cygwin anyway, I use many of the *nix tools under windows.
23:51:03[IDC]DragonI like tortoise cvs
23:51:15[IDC]Dragonand a graphical merge tool
23:51:45amiconnI did never have a look at gui cvs tools.
23:53:08amiconnI use(d) openssh from cammand line for the ssh authentication as well.
23:54:30 Quit timd0ug ("My damn controlling terminal disappeared!")

Previous day | Next day