**** BEGIN LOGGING AT Sat Apr 27 18:02:19 2002 --> adiamas (~adiamas@as5300-9.216-194-23-95.nyc.ny.metconnect.net) has joined #rockbox --- Topic for #rockbox is Open Source Jukebox Firmware - http://bjorn.haxx.se/rockbox/ --- Topic for #rockbox set by Zagor at Fri Apr 12 09:45:52 okay... logging again :) i mailed the log from the 26th out to the list? sorry.. not mailed... cvs'd ah did you update the .t file too? nope.. ill get that.. done I tried to add a cron job, but got some permission problems so it's still manual lids related? I don't know. I'll find the error mail Subject: Cron (cd $HOME/bjorn_html/rockbox; cvs update; make) rcmd: socket: Permission denied cvs [update aborted]: end of file from server (consult above messages if any) ah it's not having CVS_RSH=ssh but it shouldn't you should update anonymously ah, yes. I *should* :-) aaaaah say no more ;-) trying again GOTCHA cross-sector filenames now work too it ain't pretty, but it works :) heh, I hope you smack on some comments explaining the non-prettiness i'm trying... it's the sector caching that makes a mess i'll check it in so you can take a look sure! it's in at the bottom, fat_getnext() one sec its not that bad really i found an unnecessary line: ptr = ent->cached_buf on 963? no, 971 ah right since it's already being done on 963 removing that, I can collapse the ifs a bit i'd say this moves the fat code into GREEN status * Bagder horrays what a difference a day makes :) * Bagder presses reload on the rockbox page like crazy hehe now plopp 7 green ones 9 yellow to go yeah we should write a little converter for pics to our internal pixmap format did you write one for gifs to some odd format? thinking about the rockbox logo now no that was hand-pixeled by thomas yes, but I recall something about you doing something like that on Siemens? a converter that is ah, yes umm, i can't remember the name of the format :) in fact just loading a software showing the logo will make a *great* foto yup PPM i think the format is called portable pixmap ah yes I'll hack up a little app embryo right now the simulator should be able to do this, right? do what? run this app, showing the logo yes if you run it, include the ata code and run ata_init(). that should prevent it locking the disk. ah, the cron job works now fine but I'm not gonna try running it on target yet I can understand that :) we should fix an unlock patch for the usb 2.0 driver too. that would be really good I'll make it possible to start tetris from the app's menu :-) :) hm, doesn't work what's the problem? I can't manage to get the "cursor" to show up cursor? I have three lines in a "menu" ah and I want to move a "marker" along the side of them i see hm I need to force a screen update somehow I guess of course I do silly me yes :) yes, and clear the marker when I move away... :-) ok, time for file.c roxxers hehe, now I can start tetris from the menu ;-) nice! gotta clear the screen first though hmm, both the man pages for remove() and unlink() claim POSIX conformance remove() is undoubtedly a better name it is well.. depends on how you look at it... you aren't really 'removing' the file from the system... you are more so 'unlink'ing it from the system you're removing it from view :) but your also unlinking it from the file system ;) hey you can't drop the block in tetris! no, only speed it up * Bagder made the x11 lcd display flicker free now try the new uisim just committed lots of updates :) how do you select an entry? right ah but you can't leave tetris ;-) I better fix that still, our first menu! I like that big font hehe now should I move the X11 stuff into a subdir? or should I put the app stuff in a subdir? put X11 in a subdir it makes sense then we put all apps in the root and just link to win32 or x11 of course, the app stuff should probably be in the firmware dir... yeah, but I expect we'll have lots of little "toy apps" mainly for the simulator oh yes! do we have a sleep() yet? well we have it but it's not working right on target, since we haven't set up the timer interrupt yet I was just thinking how I should proceed to slow my app down ;-) sleep(ticks) why not sleep(ms) prevents ms/MS_PER_TICK all over ;-) use HZ/freq instead no reason really, it's just the simplest implementation doesn't really matter, we can have both ;-) yup now, which key do we stop tetris with? a function key or OFF I'll use off for now oh what fun ;-) :) add a big "Tetris" text too for the fun of it ´ed bed time --- Bagder is now known as Bagderzzz <-- Zagor has quit ("bye") ummm.. what happened to the makefile for uisim? nevermind <-- Bagderzzz has quit (carter.openprojects.net irc.openprojects.net) --> Bagderzzz (~daniel@as3-3-2.ras.s.bonet.se) has joined #rockbox * adiamas screams at this stupid person my god.. i wasn't happy totally with this guys id3 editor.. so i started digging into it... heh it has to be -the- worst design and execution of code i have ever seen i did the same thing with grip lets have a struct of data, that also has bools in it that tell you if that data field needs to be edited. why? so you can pass it around your functions. but lets be_really_ lazy and use getopt and rely on the externs * adiamas screams softly into the night who hacked out the tetris file again? --- edx|sleeping is now known as edx|studying * edx|studying thinks he's sick sleeping more than 10 hours the last two nights :? --- Bagderzzz is now known as Bagder hi bagder :) hey I moved the x11 sim stuff to its own dir yesterday yea i just updated :) so now we both could use the app.c and tetris.c in the uisimulator root good idea. nods who wrote the tetris originally? "Mattis Wadman" there is another problem with the tetris code. that give me a macro redefinition: #define HAVE_RECORDER_KEYPAD what do i do about it? ahh hehhe (you need this define there?) i compile with ARCHOS_RECORDER defined the this is defined automatically. uh, no it shouldn't be there ok ill delete it im trying to understand wtf is happening with the block_dat data another thing... unistd.h is not for windows same #define is in the x11 should it be there? so im gonna do that: #ifndef _WIN32 #include #endif right edx also, for sleep, kernel.h must be included by tetris.c (or app.c) well, there is no target version of sleep() yet there is. kernel.c i think i have seen it.. yes but it doesn't work Bagder in tetris.c could you tell me what 'frame' is with respect to block_data? eeeh I don't know really k... ill figure it out one way or another.. Bagder: how do yo uselect a menu entry? :) * adiamas begins to think that the numbness in his left pinky and ring finger is _not_ a good thing go right but tetris is the only one you can select ;-) ok :) ill upload updated files that are compatible with win32 simulator. ok? do that I'll make sure the x11 runs fine with em yep go ahead ;) edx.. do you understand the block struct? done, and modified the app.c and tetris.c slightly no more unistd.h :-) ok neither for linux nor for windows :) have you commited them yet? yes ok you can press 'off' to abort tetris now (which is the numerical keypad enter in my case) yea - i use the same keycodes ok great :) the menu still moves a little fast :) (tetris too) how is it on linux? that sleep is 1 second for mee which makes everything a bit too slow ;-) oh lol sleep 1 is defined Sleep (1000/HZ * 1) for me which is 1/10 second ;) heh the target sleep() takes 'ticks' as argument and we don't know yet how fast that is so I haven't bothered yea bjorn said it was deifined with HZ (but maybe that value is not accurate yet) its 1/100 second btw.. not 1/10 edx... do you understand the block_data structure? no matter.. thats the least problem. adiamas: no i didnt really read the code edx|studying: oh, right, haven't seen that k... adi: wait a sec hmm strange... does it already contain rotoation... ? lcd_drawpixel(start_x+x+block_data[block][frame][0][i], looking at that.. the first array parameter contains the block #.. the second one the frame (?) - maybe rotation? the third parameter is either 0 or 1.. but i dont know why nods nods? "To lower and raise the head quickly, as in agreement or acknowledgment." ahh * edx|studying nods too was just irritated because you didnt write it as a /me msg ;) i guess the four integers are the directions.. (block_data is seperated into 4-int blocks) i just don't see the blocks there... * adiamas shrugs the first block is the square thingie.. im playing around with it now aha.. uhm.. so: {{0,1,0,1},{0,0,1,1}} is a square? i got it! k.. shoot *snip!* one block always consists of four pixels! {0,1,0,1},{0,0,1,1} coordinate of first pixel: (0,0) of second pixel: (1,0) ; of third pix (0,1) of fourth pixel (1,1) can you see how i do it nth pixel coord = (data[0][n],data[1][n]) ahhhh!!! got ya... can you do me a favor.. write that in a comment above it? * adiamas likes comments that make sense. heheh sure :) * adiamas is toast... is it: x | | ----y or is it y| | | sencond one ------x ? nod // // block_data is built up the following way // // first array index specifies the block number // second array index specifies the rotation of the block // third array index specifies: // 0: x-coordinates of pixels // 1: y-coordinates of pixels // fourth array index specifies the coordinate of a pixel // c comments mate ;) like that? no c|| * edx|studying is not a very good formulator c++ even ok /* block_data is built up the following way first array index specifies the block number second array index specifies the rotation of the block third array index specifies: 0: x-coordinates of pixels 1: y-coordinates of pixels fourth array index specifies the coordinate of a pixel */ is that understandable... :/ looks fine to me --> edx|disconnected (edx@pD4B9E91F.dip.t-dialin.net) has joined #rockbox --- edx|disconnected is now known as edx int block_data[7][4][2][4] = { { {{0,1,0,1},{0,0,1,1}} }, is that translated to: int block_data[7][4][2][4] = { { {{0,1,0,1},{0,0,1,1}}, {{0,1,0,1},{0,0,1,1}}, {{0,1,0,1},{0,0,1,1}}, {{0,1,0,1},{0,0,1,1}} }, ? because the block needs rotation (but its the same in any direction) i would assume so. * edx commited tetris.c yeah, seems every block pivits on (1,1) or there in abouts --- ChanServ gives channel operator status to edx <-- edx has kicked edx|studying from #rockbox (Stupid irc corpse!) --- edx removes channel operator status from edx --- edx is now known as edx|studying --> Zagor (~bjst@as9-5-6.k.s.bonet.se) has joined #rockbox hi :) hi edx|studying: that tetris.c in the win32 dir should be removed now shouldn't it? aehm - yea wait a sec ok it is :-) is there any way to make 'sed' not use / for its' regex? yes * edx|studying doesnt understand a word Zagor: just use it after the s sxmooxpoo ok. strange that the man page doesn't say that what's 'sed' one of those secret tricks ;-) heh give me the man page ;) sed is a dead cool unix command line tool what does it do? uh, replaces regexes in a stream ;-) what are regexes (sorry for asking so much) regex is a pattern, similar to wildcard but a lot more advanced and powerful ah ok that is like *.[mp3|m3u] or something like that? ".*(mp3|m3u)$" would be good regex for that aha that $ indicates that the ( and ) do not belong to the name? no, $ matches end of string so that the mp3 or m3u must end it ahh ok - really advanced ;) right, and this is only scratching on the surface hehe what does sxmooxpoo do? lol it replaces all occurrances of 'moo' with 'poo' :-) so sed can also rename files? no, it replaces text in a stream cat file | sed -e s/replace/this > newfile ok couldnt one use it for filenames as well.. i have needed something like that a couple of times ;) sure you can in unix you just write up a quick script to do it hehe its easier in bash than with sed unless you wanna do things that bash can't do heh * Bagder jumps, PsycoXul is alive ;-) like 'for f in *;do mv $f `echo $f | sed -e s/bleh/blah/g`;done' or 'for f in *;do mv $f ${f//bleh/blah};done' heh or whatever :p Bagder: yeah that's precisely what I was doing, only "bleh" contained a full path with / characters dependency generation now fixed for 'firmware' * edx|studying shakes awestruck * adiamas cocks his head to the side and stares at edx what awestrikes you? "like 'for f in *;do mv $f `echo $f | sed -e s/bleh/blah/g`;done' or 'for f in *;do mv $f ${f//bleh/blah};done'" unix in a nutshell actually, it's only: for d in $(DEPDIRS); do { if [ ! -d $$d ]; then mkdir $$d; fi; }; done but then what the heck das that do? $(CC) -MM $(CFLAGS) $< |sed '\''s|$$*$\.o[ :]*|\1.o $(<:%.c=%.d) : |g'\'' > $@; [ -s $@ ] || rm -f $@ * Bagder falls off his chair laughing it creates the directories listed in $(DEPDIRS) if they don't already exists * edx|studying cries out despairing what does "fi;" do there? it's the end of "if" ahh lol think: if(true){ } hmm what does the second thing do? it runs the dependency generator, with, umm, correct parameters :)