--- Log for 12.05.104 Server: niven.freenode.net Channel: #rockbox --- Nick: logbot Version: Dancer V4.16p1 Started: 7 hours and 51 minutes ago 00.15.22 Join midk [0] (mk@ACC2AE13.ipt.aol.com) 00.23.34 *** Saving seen data "./dancer.seen" 00.29.20 Quit AciD (""Linux ? Ya moins bien, mais c'est plus cher"") 00.33.33 Quit friday101 (Read error: 110 (Connection timed out)) 00.42.09 Quit mecraw__ ("Trillian (http://www.ceruleanstudios.com)") 00.51.35 # Just so it's in the logs, this is me being concerned in case you (Badger) did not get why I laughed earlier ....The name "baGDer" made me smile, and I thought I'd share - no offense intended :) 00.52.03 # ... 00.52.09 # WiNS cracks me up 00.52.14 # LOLACS 00.53.31 # can't resist those little psychological tricks everywhere :) 00.53.51 # O WiNS 00.53.55 # LOLACSSSSSSSSSSSSS 01.03.20 # hmm 01.03.24 # *furious 01.18.22 # found lots of room for speed improvement in the line drawing algorithm 01.18.35 # and smaller code too! 01.18.53 # drawline? 01.18.57 # yeah 01.19.06 # OMFG PINCH ME 01.19.09 # jk 01.21.09 # hmm well i think i'm going to add a statusbar to breakout bc 01.21.14 # and if you are only drawing 90 and 45 degree lines - it's a killer 01.21.23 # we'll have to move down the top if i do 01.21.31 # we? 01.21.33 # lol 01.21.35 # i'll! 01.22.06 # if you coded it like i suggested, that will take a single line of code to ensure everything still works 01.22.15 # if not, then you have your work cut out 01.22.17 # oyes? 01.22.33 # *starts over 01.25.16 # the lot? 01.26.08 # pretty much 01.26.14 # i have to do a bit of work on it 01.26.21 # i have to move the whole overlay down a row 01.26.26 # and everything else up one 01.26.31 # and add the statusbar 01.26.49 # #define TopLine 10+4 01.27.11 # why 10+4? 01.27.21 # was at 10, now is 4 lower 01.27.40 # (10+4) actually 01.28.11 # 4 lower... it's going to be 8 high 01.28.18 # THAT TOO 01.28.20 # that too 01.28.28 # wait then where did the 4 come from? 01.28.34 # WHATEVER 01.29.06 # um 01.29.09 # yessir 01.29.19 # sir - lol 01.29.29 # yes sir funny sir 01.29.35 # midk assclown, sir 01.31.06 # so 10+8? 01.31.22 # no idea - only you knwo where from and where to 01.31.38 # i don't know where the 10 came from originally 01.31.41 # i think it should be just 10 01.31.45 # 8 for the text 01.31.48 # 2 for the line under it 01.31.53 # make that 11. 01.31.56 # as you wish 01.32.01 # TY TY 01.32.04 # or 10 01.32.09 # so it hits the line? 01.32.13 # you decide bc 01.32.17 # i need you to decide for me 01.32.35 # 11 so it bounces 1px away from the line, or 10 so it hits the line and bounces? 01.33.30 # dont worry - if its one line of code - you can correct it later if you want 01.33.40 # ok i'll say 10 01.34.06 # #define BALL_MIN_Y 10 /* farthest up the ball can go */ 01.34.08 # >_< 01.34.56 # #define Top 10 /* top of play area*/ 01.35.08 # whatever - the longer the variable, the more you have to type 01.35.13 # but BALL_MIN_Y ties in well with my other stuff 01.35.19 # #define BALL_MAX_Y 55 01.35.20 # #define BALL_MAX_TALL_Y 55 /* farthest down the ball can go with tall paddle */ 01.35.20 # #define BALL_MAX_SHORT_Y 56 /* farthest down the ball can go with short paddle */ 01.35.20 DBUG Enqueued KICK midk 01.35.20 # #define BALL_MIN_Y 10 /* farthest up the ball can go */ 01.35.20 # #define BALL_MAX_X 105 /* farthest right the ball can go */ 01.35.20 *** Alert Mode level 1 01.35.20 # #define BALL_MIN_X 1 /* farthest left the ball can go */ 01.35.58 # #define bXmin (1) 01.36.28 # hmm 01.36.37 # #define bXmax (LCD_WIDTH-1-ballW) 01.36.37 # good idea bc. 01.36.56 # may not need -1 01.37.31 # i like -1 OYES 01.37.37 # then you have 01.37.37 # int bX, bY; /* ball X,Y */ 01.37.46 # int ball_x = 34; /* ball start position (x) */ 01.37.47 # int ball_y = 34; /* ball start position (y) */ 01.37.56 # if (bX i'll make that bX and bY though 01.38.08 # whatver you like 01.38.12 # just a moment here 01.38.13 # i hate typing 01.38.17 # lol 01.38.20 # yessir sir 01.39.26 # love the way US call everyone "sir", outside of business it's quite the insult in the UK 01.39.35 # oh sorry 01.39.39 # np 01.39.44 # i think that's your way of saying "stop it" 01.39.45 # lol 01.40.01 # more of a "yeah right" or a "f you" 01.40.15 # i was going to say that i swear 01.42.17 # how many greyscales are really useable? 01.42.24 # ...in a game 01.42.30 # up to 33 01.42.51 # bluEchiP: I'm currently looking at your starfield demo code. 01.42.56 # OH I LOVE IT 01.43.05 # AHEMFLYTHRUSPACEAAEEMMMMMMM 01.43.05 # errrrr - it's just been vastly improved - but yes? 01.43.22 # *coughs to cover up that loud noise 01.44.18 # comment on the confusing use of "int t;" will cause me to cringe for the second time tonight - LOL 01.44.22 # I wonder why you are using so many I8 / UI8. 01.44.35 # i love ui32. 01.44.45 # This is actually slower most of the time and generates more asm instructions than using UI32 / I32 on a 32bit machine like the SH 01.45.02 # hmm 01.45.07 # did bc just get told? 01.45.10 # heh jk 01.45.16 # no keep ui32 01.45.17 # i like it 01.45.21 *** Alert Mode OFF 01.45.54 # that's handy to know - thanks 01.46.27 # now you've got me wondering what to do about sprite data 01.47.22 # Of course you have to use byte width for byte data like sprites, but I recommend using int for coordinates 01.47.47 # yes,thanks - that show() routine would be a fair chunk faster with that knowledge me thinks 01.47.57 # Second thing that puzzles me is your line algorithm - "else if (deltay)" with an comment saying /* horz */ 01.48.21 # OUCH! 01.48.21 # Shouldn't this read "else if (!deltay)" ? 01.48.32 # yes indeed - lol 01.48.42 # amiconn ooh 01.48.56 # do you like the use of "int t" in the show() routine? 01.49.31 # And the last thing - you call these routines sprite routines, but imo they aren't, because they don't save the background. 01.50.10 # bad boy bc 01.50.31 # so amiconn 01.50.37 # if i move the overlay down 1 row 01.50.48 # anything that uses a gray function will have to be moved up basically? 01.51.01 # three answers to that ... 01.51.10 # 1. they're now called "bc" routines because I am about to jack in using the core rand() function which is FAR too costly for me 01.51.17 # 2) they are very much in their infancy 01.51.55 # A final one - using rb->memset() in spr_CLS() would be very likely much faster than using a loop. 01.52.04 # 3) the way I have it all mapped in my head atm, saving the background will not be relevant, as the sprites will exist on different layers 01.52.23 # midk: yes 01.52.23 # blinding idea - THANK YOU 01.52.28 # bc lol 01.52.31 # ty ami 01.53.03 # i presume memset lives in iram 01.53.12 # gah this is NOT good. 01.53.20 # i may just use gray bitmaps for the numbers 01.53.36 # bluEchiP: Yes it does, and it is fully asm optimized :) 01.54.09 # midk: I really consider adding font support to the framework. 01.54.27 # i certainly can't move this up. 01.54.38 # i suppose i'll use bitmaps then until font support is added 01.54.38 # mk: feel free to nick the starfield stuff 01.54.39 # thanks though 01.54.45 # heh ty 01.55.17 # Font support demands to change another thing in the framework first, namely gray_drawbitmap() 01.56.06 # my god - it's in assembler - wow - cool 01.56.36 # gray_drawbitmap() will be changed to use the same bitmap format as lcd_bitmap(), so you can use bmp2rb to make bitmaps for it as well. 01.56.56 # what is that format do you know? 01.56.58 # calm down bc 01.57.01 # do you know assembly? 01.57.08 # lol - just a wee bit 01.57.13 # ami: i'll use the drawgraymap 01.57.19 # probably keep it actually 01.57.24 # THATS A REALLY GOOD IDEA! 01.57.29 # *keeps idea for self 01.57.38 # it would match my splash screen font. 01.58.10 Join silencer_ [0] (~silencer@adsl.via.ecp.fr) 01.58.20 # midk: gray_drawgraymap() will stay with it's current format, only gray_drawbitmap will change. 01.58.32 # i'm going to use drawgraymap 01.58.35 # easier and better 01.58.41 # maybe i'll grayshade my font 01.58.44 # oyah oyah oyah. 01.59.20 # ami - do you think that 33 shades are really visible in a game? 02.00.30 # Imo the maximum of 33 shades is too slow for a game, I would only use 8..16. 33 shades is for image viewers or such. 02.00.41 # i use 8 02.00.55 # 8 and 16, not 9 and 17? 02.01.04 # dahaha. 02.01.19 # amiconn: did you say you were creating a jpg or bmp reader? 02.01.54 Quit silencer (Read error: 54 (Connection reset by peer)) 02.02.00 # if you are, you may be delighted to know that ID3 specifies the standard image size as 64x64 (although supports anything) 02.02.05 # midk: [IDC]Dragon is developing a jpeg viewer (which basically already works as he told me) 02.02.30 # i want to see. 02.02.40 # Next thing I'm thinking of is (1) using draw modes to shorten the code a bit and (2) optimized drawing routines. 02.02.50 # there's an album art id3 info tag 02.04.18 # ami: how many "register int x"'s am i likely to get away with, and does it assign them from the top of the declaration list down ....do you know? 02.04.18 # midk: The jpeg vewer cannot run in parallel to music playback, as it has to use the playback buffer 02.04.37 # aw 02.04.53 # blimey - i knew it was big - but i didn't know it was THAT bit 02.05.12 Join silencer [0] (~silencer@138.195.149.157) 02.05.29 # bluEchiP: I don't know about the assignment sequence - the gcc optimizer sometimes does very odd things. 02.06.43 # Most of the time you don't really need to explicitly require "register" 02.07.06 # As long as you don't reference a variable's address and the scope of that variable is not too large, it will be put in a register. 02.07.38 # cool - i must say, i am mighty impressed with what our compiler does 02.08.01 # If you want to know for sure which variable is registered and which is not, you have to disassemble the object module. 02.08.16 # yes, i must look into the assembly at some point 02.08.38 # i've not looked at it AT ALL yet :( 02.09.21 # The odd thing with the gcc optimizer is - while being very clever at some spots, it does not do obvious optimizations in other spots. 02.10.05 # I suppose the guy who wrote it had to stop somewhere - but it is often annoying to find glaring omissions 02.11.15 # The best example for that was the recording data transfer loop (before I optimized it). There I could see both behaviours close together. 02.11.36 Join george_4_man [0] (Joosa@ool-44c4728b.dyn.optonline.net) 02.12.02 # can anyone help me with cygwin please? 02.12.20 Join scott666_ [0] (~scott666@24.245.59.203) 02.12.37 # ami: I've got a whole bunch of new optimisations for existing code ...I should submit them to you - you can claim the work and they will get included :) 02.13.10 # ami: you might be interested to compare my mods to the bressenham algorithm 02.13.42 # george_4_man: yes, but for the next 3 minutes I am making tea - queue up some questions while I am gone :) 02.14.02 # okay 02.14.23 # i tried to do a make, but sh-elf-gcc doesn't exist, where may i find that? 02.14.26 # bluEchiP: Yes, that is one thing I want to try (at first within the grayscale framework). 02.15.31 # george_4_man: See http://rockbox.haxx.se/docs/cygwin_sdk.html , especially step 3 02.15.48 # yeah, i'm on that right now, i did everything 02.16.39 # Hmm. You did install the sh-gcc and sh-binutils packages? 02.16.43 # yeah 02.16.57 # could it have something to do with not being able to download the setup.ini? 02.17.10 # Perhaps you should look where in the cygwin dir structure the executables are located. 02.17.22 # for what? 02.17.24 Quit silencer_ (Read error: 104 (Connection reset by peer)) 02.18.21 # To know (1) if they are really there and (2) for knowing their path. Perhaps they are in a path that isn't included in the search path of the shell by default. 02.18.47 # it's not in there 02.19.29 Join silencer_ [0] (~silencer@nino.via.ecp.fr) 02.19.29 Quit silencer (Read error: 54 (Connection reset by peer)) 02.19.48 # back in 30 or less 02.20.07 # ah, i found it, thanks 02.20.10 Quit midk (Read error: 54 (Connection reset by peer)) 02.20.30 # george_4_man: where? 02.20.53 # opt/sh1 02.21.23 # So next thing is to check your search path: 02.21.31 # i just copied it to bin 02.21.38 # is that a bad thing? 02.22.05 # george_4_man: i can heartily reccomend the dev kit on my home page 02.22.25 # That's another alternative, but I prefer to not put things that don't belong to genuine cygwin into /bin 02.23.04 # so how do i change the search path? 02.23.35 *** Saving seen data "./dancer.seen" 02.24.49 Join midk [0] (mk@AC87045A.ipt.aol.com) 02.25.20 # george_4_man: One moment, have to take a look. 02.25.43 # k thanks 02.39.46 Part george_4_man 02.42.05 # Grr, now he is gone. It took me quite a while to find it; there are definitely too many ways to do something in unix. 02.42.28 # heh 02.42.34 Nick scott666_ is now known as scott666 (~scott666@24.245.59.203) 02.43.00 # maybe he got kicked and plans to return :) 02.43.46 # well, *I'VE* implemented all your suggestions, so you ARE appreciated - actually, what is more appreicated is your proof reading my code for me :) 02.44.29 # Did you update starfield.c on the web already? 02.44.45 # take me 10 seconds.... 02.44.58 # actually, let me just check it still compiles 02.45.29 # done 02.49.40 # (1) You can optimize the coordinate bounds checks by doing "if ( (unsigned) x >= LCD_WIDTH || (unsigned) y >= LCD_HEIGHT)" 02.49.45 # (That reminds me that I wanted to use that as well..) 02.50.14 # which line? 02.50.48 # E.g. in bc_plot() 02.50.53 # oh yes, very cool 02.50.59 # (found it as you said) 02.51.49 Join silencer [0] (~silencer@adsl.via.ecp.fr) 02.52.05 # that's a real hoopy bit of thinking there :) 02.52.23 # (whoopy?) 02.53.06 # (as in "saz that hoopy Ford Prefect there's a cool frood who really knows where his towel is") 02.53.24 Quit midk (Read error: 60 (Operation timed out)) 02.54.10 # bluEchiP: Dito in bc_line(), only that the condition(s) is/are reversed. 02.55.16 # i've also noticed that "abs(expression)" causes the expression to encoded thrice and calculated twice 02.55.48 Quit silencer_ (Read error: 104 (Connection reset by peer)) 02.56.14 Join friday101 [0] (~why@pcp06796835pcs.bowie01.md.comcast.net) 02.56.19 Join silencer_ [0] (~silencer@nino.via.ecp.fr) 02.56.40 # SOMEONE PLEASE FIX THE HEBREW FONT PROBLEM! 02.56.45 # gooday! 02.56.48 Quit friday101 (Client Quit) 02.57.01 # lol 02.57.11 # lol 02.57.34 # i'm sure there's a good jewish joke in that somewhere 02.57.40 # drive-by feature requests 02.57.44 # lol 02.58.28 # i thought jews were stereo-typically supposed to be frugal with money not manners 02.58.44 # (best i could come up with at short notice) 02.59.13 Quit arspy87 ("Leaving") 02.59.50 # yeah, i was trying to think of one 03.00.34 Quit silencer (Read error: 54 (Connection reset by peer)) 03.01.34 Join midk [0] (mk@AC838038.ipt.aol.com) 03.01.42 Join silencer [0] (~silencer@nino.via.ecp.fr) 03.02.37 # bluEchiP: Apart from 3 locations in bc_line() where the optimized bounds check would be applicable, I didn't find more things. 03.02.58 # fantastic - thanks for your time :) 03.03.14 # I'm wondering if the abs can be optimised 03.03.58 # abs. 03.04.02 # hi bc and ami 03.04.17 # bluEchiP: I did not find an abs() in your code ?! 03.04.27 # in bc_line() 03.04.37 # deltax and deltay 03.05.15 Quit silencer_ (Read error: 104 (Connection reset by peer)) 03.06.17 # Ah oops, you already tried to replace it. 03.06.22 # deltax = abs(x2-x1) 03.06.22 # becomes 03.06.22 # deltax = x2-x1; if (deltax<0) deltax = (~deltax)+1; 03.06.22 DBUG Enqueued KICK bluEchiP 03.06.22 # ?? 03.06.52 # sorry - what is there atm is wrong 03.07.03 # Unfortunately, your "&= 0x7FFFF" can't work. 03.07.05 # i just stripped the sign - I didn't 2's it 03.07.34 # ho to you too. 03.07.55 # toot 03.08.17 # what do you think to the above idea? 03.09.33 # About that: deltax = x2-x1; if (deltax<0) deltax = (~deltax)+1 ? 03.09.38 # yes 03.09.52 # is the rule, no c++ style comments? 03.10.00 # yes 03.10.11 # handy for debugging though 03.10.11 # dah. 03.10.31 Join silencer_ [0] (~silencer@adsl.via.ecp.fr) 03.10.32 # I think this will generate almost the same code as simply using deltax = abs(x2-x1); 03.11.05 # it's that "almost" I'm counting on 03.11.50 # Since this is only called twice in your function, it should have a neglectible impact on performance. 03.12.04 # but it IS in drawline() 03.12.14 # which is quite a popular graphical call 03.12.43 # If you really want it fast, you'd have to use asm. 03.13.14 # amiconn: as bc was saying, usually the gray in my breakout paddle is sort of flickery -- any way to sort of smooth it out? 03.13.45 # yes, I supopse as have no plans to submit the code to Rockbox, I do not need to conform to their coding standards - hmmm 03.14.38 # You can code your "if(deltax < 0) deltax = (~deltax)+1" with only 2 asm instructions... 03.15.01 # Oops, 3 instructions of course 03.15.05 # lol 03.15.52 # midk: As I did not see it yet, I can't tell for sure, maybe not. 03.16.31 Join Mylacc [0] (~Mylacc@63-174-114-63.uterr.blacksburg.ntc-com.net) 03.19.24 Quit silencer (Read error: 60 (Operation timed out)) 03.21.57 # bluEchiP: deltax = x2 - x1; asm ("cmp/pz %0 \n bt 1f \n neg %0,%0 \n 1: \n" : "+r"(deltax) ); 03.22.39 Ctcp Ignored 3 channel CTCP requests in 5 hours and 45 minutes at the last flood 03.22.39 # * bluEchiP stares at code until it makes sense 03.22.46 # Of course this is machine dependent, so to be surrounded with #ifndef SIMULATOR .. #endif 03.23.40 # What does Dir Buffer is full mean? I know it means that it wont display the directories but with is the max amount or how do i fix this? 03.23.48 # in the menu 03.24.53 # but i'll be blown if I can find it 03.25.42 # general->system->limits 03.26.14 # ok, found that 03.26.27 # first is max files in dir 03.26.34 # and the second is max playlist size 03.26.46 # my problem is max dirs 03.26.47 # i think 03.27.03 # because there are only 3000 mp3s and my max files in dir is set to 4000 03.27.11 # so that shouldnt be the problem 03.27.21 # hm - what version do you run? 03.27.42 # 2.2 03.27.46 # v2 recorder 03.28.12 # ami: can i replace the "\n"'s with real newlines for readable? - sorry, I never did get my ahead around this presentation of assembler 03.28.25 # Mylacc: 2.2 - release or daily build? 03.28.30 # release 03.28.50 # try a daily build - cos it may have been found and fixed 03.29.06 # ok, thanks, will do 03.29.07 # Mylacc: The default for max files in dir is 400. This shouldn't be set too high; it's sufficient to set this just above the maximum number of mp3s that you have _in one dir_ 03.29.09 # problem is all the devs use daily, so help on release stuff is difficult to obtain 03.29.46 # ami: is there a doc somewhere that explains this asm format? 03.30.16 # bluEchiP: You can split the asm code string into several lines, but the \n's inside the string must remain. 03.30.24 # ok - cheers 03.30.54 # Yes, see http://gcc.gnu.org/onlinedocs/gcc-3.0.4/gcc_5.html#SEC68 , section 5.36 03.31.14 # bcbccc 03.31.30 # This is how I write longer asm blocks in that format: 03.31.35 # asm( 03.31.58 # "cmp/pz %0 \n" 03.32.24 # "bt 1f \n" 03.32.37 # "neg %0,%0 \n" 03.32.48 # "1: \n" 03.33.01 # : /* outputs */ 03.33.45 # ./* %0, in & out */ "+r"(deltax) 03.33.50 # ); 03.34.21 # (sorry the dot at the start of the second last line shouldn't be there) 03.34.37 # tanks 03.34.38 # thanks 03.34.54 # what is the "1:" ? 03.35.21 # sorry, this might be a dumb question but did the file format chance? there is no .rockbox directory, only a rockbox-daily-20040511 directory 03.35.40 # did you download source code or build? 03.35.55 # The 1: is a local label (target for the bt instruction) 03.36.05 # ha 03.36.08 # good call, thanks 03.36.23 # oh! - so what's the "f" in "bt 1f" then ...forward reference? 03.36.49 # i had guessed that 1f was a hex value specifying the branch distance 03.37.01 # yup, 1f means "jump to the next "1" label you find in "f"orward direction. 03.37.07 # Mylacc: lol ;) 03.38.14 # so "bt heref" would match with "here:" ??? 03.38.32 # sorry, i must be sounding really thick right now 03.39.21 # no, because "here" wouldn't be a local label. To match "here:", you would have to use "bt here". 03.40.11 # The asm above is explained quickly: 03.40.18 Join silencer [0] (~silencer@nino.via.ecp.fr) 03.41.04 # should i presume asm( "code" : %0 : %1 : %2 : %3 ); 03.41.09 # "+r"(deltax) specifies that deltax is both an input and output parameter, and it should be put in a register. 03.41.39 # As this is the first parameter, that register will be referenced with %0 within the asm. 03.41.57 # understood 03.42.51 Quit silencer_ (Read error: 60 (Operation timed out)) 03.42.57 # The first instruction "cmp/pz %0" is "compare if %0 is _p_ositive or _z_ero. 03.43.01 # I have over 125 folder in one directory, is this causing the problem cause it's still happening 03.43.52 # If yes, it sets the t condition flag, otherwise it resets it. 03.44.21 # yep - i'm pretty cool with the assembler, just not the inline thing 03.44.43 # Ok then. 03.44.46 # cheers 03.45.26 # I must pull down the SH1 programmers docs and learn them all - but I used to write BIOSes for a living, so asm as a concept is well understood by a stray corner of my grey matter 03.45.44 # The syntax within the asm() string is the same as for pure asm, except of the %0 thing. 03.46.16 # that's cool - there seem to be lots of weird things to learn about what starts happening after the colons 03.46.23 # SH1 asm is not really hard to learn - since it is a risc machine, there aren't that many instruction. 03.46.42 # risc - cool - my last disasm was for a risc chip 03.46.52 # I learned it in a few days. 03.47.19 # If you really want to optimize, there are some pitfalls, though. 03.47.23 # wrote a documenting disassmebler for a series of RISC based CPUs 03.47.39 # MIPS-X core 03.48.03 # pitfalls? 03.48.25 # arguments with gcc? 03.49.19 # No, within the asm itself I mean. There are situations where an instruction sequence would cause a pipeline stall, which can be resolved by re-ordering. 03.50.02 # you wanna try dealing with "delayed branching" - that's a mind blower 03.50.38 # You have to - unconditional branches are always delayed branches with the sh. 03.51.06 # how many delay slots? 03.51.12 # One. 03.51.27 # that's not too bad 03.51.51 # two slots really forces the abusive use of NOP 03.52.10 # Another thing: Don't use a register you load from memory in one instruction within the next instruction. 03.52.58 # Next: Keep memory load/store instructions at longword boundaries (if executing from 32 bit memory, i.e. IRAM) 03.53.17 # There are a few others. 03.53.18 # fully and completely understood 03.53.28 # (at a silicon level) 03.53.39 # thanks for the heads up 03.53.52 # you can write some GREAT code if you try to though 03.54.11 # it's great for getting reference addresses to variables 03.55.27 Quit silencer (Read error: 104 (Connection reset by peer)) 03.55.44 # Have a look at lcd.S as well as copy_read_data() and copy_write_data() in ata.c for some really tricky optimizations. 03.56.37 # ld r1,100[r2] 03.56.38 # ld r3,r1 03.56.38 # ...r3 = 100+r2 03.56.41 # (all 3 end up in IRAM, so rule #2 above applies. However, plugin code does not.) 03.57.13 Join silencer [0] (~silencer@138.195.149.157) 03.58.28 # Ooops, copy_read_sectors() and copy_write_sectors() of course. 03.58.35 # OK, i reordered the folders to letters in the alphabet so i am assuming 125+ folder was the problem 03.59.06 # can you report that to the mailing list pleae so it does not get forgotten about 03.59.16 # ^^ @ Mylacc 03.59.30 # I should really get some sleep now. 03.59.42 # bc 03.59.43 # pm 03.59.44 # ? 03.59.46 # amiconn: thank you for all your help - it is greatly appreciated 03.59.51 # nite ami 03.59.59 # sleep good my friend 04.00.13 # nite 04.00.19 Part amiconn 04.00.39 # ok, will do 04.00.54 # subscribing first 04.00.54 # lol 04.03.57 # ok, once i get the address i will send it out, prolly tomorrow since i did a daily mailing 04.04.13 # thanks for your help 04.04.14 # nice one dude 04.04.18 # you're welcome 04.23.38 *** Saving seen data "./dancer.seen" 04.24.12 Join Strath__ [0] (aopen@216.165.135.240) 04.24.25 Nick Strath__ is now known as Strath (aopen@216.165.135.240) 04.24.29 # strathy. 04.24.49 # hey midi 04.25.01 # what's up? 04.25.28 #