00:00:00 --- log: started retro/12.12.16 03:51:32 --- join: carefish (~carefish@095-096-156-068.static.chello.nl) joined #retro 04:09:03 pfew, i slept like a baby :D 04:09:15 and almost done reading that pascal pdf 07:00:45 --- join: Mat2 (~Claude@91-65-144-133-dynip.superkabel.de) joined #retro 07:00:50 hello 07:40:21 morning 07:41:44 --- nick: tangentsleep -> tangentstorm 07:46:22 morning! 07:46:32 am now almost finished w/ the pascal pdf 07:47:03 what do you think? :) 07:48:06 liking what i see 07:48:19 --- join: intothev01d (~intothev0@unaffiliated/intothev01d) joined #retro 07:48:31 and i think i can do this, so i'll get on to it once i'm done w/ the pdf :P and then probably use it as a reference 07:50:47 sorry, wha PDF ? 07:51:10 http://freepascalanswers.files.wordpress.com/2012/09/startprogusingpascal.pdf 07:54:56 hello every one 07:55:06 'sup? 07:56:57 well, i brushed up on sdl last night and think i know how to get images on the screen 07:57:59 ah :) 07:58:05 that's rather easy :p 07:58:32 depending what you chose to use though, opengl/use sdl blit/flip 07:58:36 yeah, i actually had an old demo i made when i first started investigating pascal 07:58:47 ahm, I have reserved one port for direct screen access, port 17 07:58:59 (x,y,colour-) 07:59:47 for pixel-by-pixel drawing, right? 07:59:52 yes 08:00:13 and begun work on a blit routine before fell to sleep :) 08:00:24 this will be port 18 08:01:41 using sdl? 08:01:47 yes 08:02:18 so the VDP code is already able for screen plotting 08:02:24 oh nice 08:03:20 I'm back in one hour and upload the latest sources 08:04:03 carefish: read the romVDP.pas file, you can build your routines on that 08:04:11 okie doke. i'm gonna go take a shower and whatnot. seeya in a bit 08:05:25 hrm, must reboot 08:05:36 see you guys in one hour 08:05:39 --- quit: Mat2 (Quit: Verlassend) 08:06:04 where can i find that file? i don't think i have it 08:10:54 tangentstorm, so I'm working on converting over retro.py to retro.js still. I pushed my changes last night and the DivMod function seems to work correctly. I'm working on the run function now 08:12:09 https://portfolio.assembla.com/code/avm/git/nodes/302fd4bce8369949f4ed1c7c100d63e120789afd/metro/backport/romVDP.pas ? 08:12:55 so basically lists in python will need to be arrays in javascript I think 08:27:36 what do you mean by build my routines on that? 08:29:41 --- join: Mat2 (~Claude@91-65-144-133-dynip.superkabel.de) joined #retro 08:30:02 ok, I have upload the latest sources 08:30:26 carefish: what is the status of your work ? 08:32:02 just finished the pdf, written down how i'd like the arraystructure, and just started w/ a spriteclass 08:32:31 and i'm not sure if i can build my code around that romVDP 08:34:59 ok, there exist one procedure vdpPlotPixel (handle, adr, colour) 08:35:18 the handle is tScreen 08:35:36 adress calculation is x * cScnXRes + y 08:35:55 and colour is the pixel colour (0-255) 08:39:00 each pixel is handled like a sprite so you can overlay pixels over the character display without destroying it 08:40:44 it's like you have 800 x 600 single pixel sprites 08:41:51 ok, I see my sources are not megred to the git repro, uno momento 08:44:41 PROCEDURE plotPixel 08:44:41 (handle : tVDP; adr : LONGWORD; value : BYTE); INLINE; 08:44:41 BEGIN 08:44:41 pBitmap := handle.pBitmap; 08:44:41 rBitmap := handle.rVStart * cScnXRes + handle.rHStart + 08:44:42 pBitmap^.pixels + adr; 08:44:44 rBitmap^ := value; 08:44:46 END; 08:44:48 ^this is the procedure 08:45:48 handle.pBitmap: SDL surface 08:46:42 rBitamp: Reference pointer (needed because an SDL suface is an indirect C pointer, otherwise you get a core dump) 08:46:53 rBitmap of course 08:47:03 you see ploting to screen is quie simple 08:47:08 ^quite 08:48:17 you don't need to backup old pixel colour because this is handled automatical from the VDP code 08:48:50 and you don't need to flip the SDL surface, this is also already handled asynchron 08:50:50 you find the newest VDP code here: 08:51:32 https://www.assembla.com/code/avm/git/nodes/master/metro/backport 08:52:03 you will need romVDP.pas and ng.ports.pas 08:52:14 rename ng.ports.pas to ng_ports.pas 08:52:23 and retro will compile 08:53:15 --- nick: Mat2 -> Mat2-afk 09:04:13 Mat2-afk: can you please fork https://github.com/goldsmile/ngaro-pas or add it as a submodule in your repository? 09:04:38 when you just copy the files, i have no way to merge them back in except to manually run a diff/merge tool 09:05:01 or work directly on the repository if you want :) 09:05:19 ng_ports.pas(10,3) Fatal: Syntax error, "UNIT" expected but "FUNCTION" found 09:05:22 intothev01d : yep py list => js array 09:05:38 function vm.handle_syncport( msg : int32 ): int32; this is the line 09:07:12 carefish: you can't "use ng_ports" 09:07:34 if you look in ng.pas, you'll see there's a bunch of {$i ng_xxxx.pas } comments 09:07:40 those are include directives 09:07:50 conceptually it's just one giant module called "ng" 09:08:22 {$ifdef nested_units} isn't a real thing :) 09:08:56 the unit line at the top of those files is just to trick emacs, which has a really clunky pascal mode 09:11:25 well damn I'm stuck with no idea at this point 09:12:19 for some reason process.argv is undefined no matter what in my retro.js. however if I create just another .js file and put console.log(process.argv); it works just fine to display the arguments. I don't know what the problem could be 09:12:41 hmm 09:13:11 can you post the latest sources somewhere? 09:13:15 sure 09:13:16 let me push 09:14:52 here are my changes https://github.com/intothev01d/ngaro-js/commit/2eaddc36f29398f0cbdde3fd01700d9af1584d4d 09:15:31 running node test.js will print out the args just fine, however you get undefined if you run node retro.js 09:15:36 !! 09:16:06 http://pastebin.com/VF9syKa0 do i need to do more than just that? can i know use Direction[NN].sprite 09:16:08 i don't know if it has something to do with the rest of the content of retro.js 09:16:21 but it seems like it wouldn't especially since I'm calling it first thing 09:16:39 var spriteDirections : Direction; ? 09:17:37 one min carefish 09:20:34 maybe all these functions have to be vars? 09:20:52 i am baffled by this behavior :) 09:21:17 me too 09:21:29 it just makes no sense really 09:21:29 so... what i would do is take everything out except that one line, and then add each thing in one by one to see what causes it 09:21:39 yea guess that's what I'll do 09:23:22 i am no longer baffled 09:23:30 you have a function called process :) 09:23:41 oh son of a !!! 09:23:45 lol 09:23:47 it parses the file before it executes it 09:23:49 tangentstorm: I have sync my sources with https://github.com/goldsmile/ngaro-pas ... no error so far (paranoia mode on, I mean it's GIT) 09:23:49 yea that was dumb 09:24:04 Mat2-afk: thanks :) 09:24:08 I would have gotten there eventually after putting all the stuff back in :D 09:24:11 thanks for catching that 09:24:14 np 09:24:21 okay... carefish.. looking 09:24:52 yea working now 09:24:53 lol 09:25:34 carefish: this works too, but i think the enumerated type approach has some advantanges over this 09:25:53 with this version, you'd use whatever.NN[5] or whatever 09:26:05 i actually don't think it's what you wanted to express originally 09:26:20 you wanted an open array (growable array) of sprites for each direction 09:26:27 this is fixed at 8 entries per direction 09:30:45 ah, ok :o 09:31:32 so what would have to be changed? 09:32:08 because i got syntax errors on Direction = record ( nn/etc/etc ); 09:32:23 carefish: you want http://pastebin.com/Ha0cGD04 09:32:53 whoops.. i meant "here" instead of "scene" on line 28 09:35:07 thanks! nice and clear, so all i'd have to do now was make a Pixel class that contains a draw that draws to plotPixel? 09:36:34 i'm not sure what you're doing yet :) 09:36:43 you could make Sprite a class and add a draw method 09:37:30 so then it'd be for sprite in here.viewport[looking] do sprite.draw? 09:37:39 argh mat2 you broke my {$IFDEF GRAPHICAL} stuff :) 09:37:55 sprite.draw(plotpixel stuff); 09:38:23 carefish: yes on the for loop 09:39:11 Mat2-afk: https://github.com/goldsmile/ngaro-pas/commit/2ce039fefd4e8ff1b27c7475345bd9042fdf89e7 ... i'm trying to preserve the ability to run this over ssh in a non-graphical terminal 09:39:52 when i hand-merged yesterday, i added all those {$IFDEF ..}s to allow either way of doing things 09:40:41 awesome, so it's all wired up to check for arguments, so far works with passing --dump. now I just need to figure out how to handle the image file, all these memory and stack arrays, and work on the processor function to process them 09:42:18 in python, [0] * 12 would create 12 list elements correct? 09:42:21 intothev01d: sounds like you have your bearings 09:42:23 yep 09:42:30 then inputs.append(0) does what? 09:42:38 add another one initalized to 0 to it? 09:42:41 yes 09:42:45 ok 09:42:53 that's an odd thing to do 09:43:05 that's what is in the retro.py 09:43:13 inputs = [0] * 12 09:43:13 inputs.append(0) 09:43:58 in the run() function 09:44:46 tangentstorm I don't know how much further I'll get other than just directly converting the if else stuff in process. I'm not to familiar with how it's all supposed to work and I don't want to mess up the logic on it but I can handle all the parts I basically understand 09:45:47 --- nick: Mat2-afk -> Mat2 09:45:55 tangentstorm: sorry 09:47:39 Mat2: no problem. that's why i wanted us both using the same repo. i'll re-add that stuff in a minute 09:48:09 intothev01d: okay... so the python version doesn't need to do that inputs=[0]*12 line 09:48:17 or the append? 09:48:21 that's for the "--with" 09:48:29 oh ok 09:48:36 nope, it doesn't need that append(0) either 09:48:43 it does need a list though 09:49:19 ok, the VDP code is finished I hink so far 09:49:24 the idea is that you say --with whatever ( or in retro, you say: needs whatever' ) and it loads that file from disk 09:49:37 ^think 09:49:38 oh ok 09:50:01 im now testing it out and start debugging 09:50:06 but if *that* file includes another file, then you have to push the current file onto a stack ( list.append() in python ) and process the new file before you finish the first one 09:50:12 --- nick: Mat2 -> Mat2-debugging 09:51:09 intothev01d: since c isn't great with dynamically sized structures like that, you'd preallocate an array of file handles like this... i'm guessing this was just ported from the c. it's unnecessary in python 09:51:29 and also in javascript, I'll just create an array and append to it 09:51:33 correct 09:52:07 intothev01d: implementing --with is more than i asked you to do, and to do it right you'll need to implement the file input device 09:52:36 I'll just drop the list array and argument, then if someone wants to actually implement it later will work 09:52:44 cool :) 09:53:23 now i get to work with fs i think in node 09:53:26 for file opening 09:53:42 loading the image should be a much simpler case: you just need to somehow load an array of 32-bit integers from disk 09:53:55 --- join: FlashingYoshi_ (ad4965cd@gateway/web/freenode/ip.173.73.101.205) joined #retro 09:54:33 --- part: FlashingYoshi_ left #retro 09:54:53 Mat2-debugging: i'll wait until you commit your changes, then add the ifdefs back 09:55:07 meanwhile, i will start on a game loop in retro :) 09:55:32 oh but first i will patch retro.py :) 09:57:08 tangentstorm: that should be the curent code 09:57:12 ^current 09:58:01 I had already pushed my changes 09:58:27 pot 17: check , actual code 09:58:34 ok 09:59:09 cool. :) 10:00:10 so to assign something to use for romvdp 10:00:25 pixeliwanttodraw : tvdpdata; ? 10:00:35 and then initialize the values somewhere else? 10:04:50 not sure. need to read the new code first 10:08:18 crc: i pushed some changes to retro.py and added a .bzrignore for the junk *.img files produced by the ngaro tests. r422 in my branch. 10:08:53 ok. time to make a game! :) 10:13:23 yeah, seems to work 10:13:48 I can change the foreground and background colour on the fly at vsync 10:13:53 in retro 10:14:19 now testing access to attribute and charmap 10:20:31 tangent, I have it to where it can check for the image file now 10:20:37 tangentstorm * 10:20:49 :) awesome 10:21:24 I don't have a clue though about what does on after those cli arg checks 10:21:30 goes* 10:21:46 http://pastebin.com/1uKbTXKh 10:21:55 line 443 and on ( i know I can ignore the set_termios for now) 10:22:45 to me it looks like it's opening the image and loading it into the memory thing 10:24:36 lunch time 10:34:22 Mat2-debugging: okay, i re-applied the ifdefs stuff 10:34:54 if you do a git pull in ngaro-pas we should be on the same version now :) 10:36:50 carefish: for SetBitmap, you can use a property, which allows the syntax to be sprite.bitmap := xxx and it'll call your setbitmap behind the scenes 10:37:00 ok 10:38:26 carefish: give me a moment, and i'll check in a new program file in the root directory that loads and shows Mat2's terminal 10:38:39 then we'll set it up so you can call your routines interactively from the retro prompt 10:45:39 --- join: kumul (~kumul@cvx-ppp-66-50-141-36.coqui.net) joined #retro 10:54:11 carefish: okay, if you do a git pull, there's a lazarus project in the top level directory now that just loads retro 10:54:28 you probably need to copy retroImage from lib/retro into that directory for the time being 10:55:38 carefish: play a bit with retro to get firm 10:56:02 20 16 6 out 10:56:40 for example sets the global foreground colour to radio-active green 10:58:18 (beware you have 256 colours) 11:00:44 hmm. nothing's happening 11:04:05 I have just pushed a fix 11:04:41 git was out of sync 11:06:22 now I'm in sync these error occur: 11:06:27 Error: Forward declaration not solved "vm.clear;" 11:07:21 i guess i missed an IFDEF. add -dGRAPHICAL to your compiler options and it should work 11:10:57 thanks 11:11:40 now I ge these error: 11:11:43 ng.pas(20,36) Error: Specialization is only supported for generic types 11:12:33 you're working off your old repository 11:12:59 you need the latest version of xpl 11:13:12 stacks.pas is now generic 11:13:30 where can i find that sample project tangentstorm? 11:14:08 https://github.com/goldsmile/rpld25 11:14:31 or just go to the top level directory, and run "git pull; git submodule update" 11:14:47 * tangentstorm punches the makefile for not working on windows :/ 11:18:42 ah, so that's how to init the retro thingy? 11:18:52 and with plotpixel i actually draw inside retro? 11:19:13 with vdpPlotPixel !!! 11:19:22 in theory, if we can get it working 11:19:47 nice :), i'll fiddle around with it 11:20:13 Mat2-debugging: i see a cursor but it's not in the right place and no matter what i do, the text is always green on purple 11:20:35 do you want a screenshot? 11:21:11 http://img.is1337.eu/f791e2eb.png 11:21:25 yep. that's it 11:21:45 Mat2-debugging: his screenshot looks like what i see too 11:23:19 that's because of the 11.5 image file 11:23:40 oh. do you have a new image we should be using? 11:23:49 I use the 11.4 image 11:23:59 ok 11:25:05 carefish: what did you use to make and post a screenshot so fast? 11:25:39 tangentstorm: remember, you added some CRT stuff which broke he screen handling in 11.5 11:25:52 yeah but crc took it out :) 11:26:53 these red boxes are escape sequences 11:29:45 http://imgur.com/oZAKO 11:30:18 tangentstorm: a tool a friend of mine made 11:31:03 i'm going to make it in C++ ( or pascal, not sure yet) in the xmas holidays 11:31:15 make what? 11:31:25 the tool to make screenshots really fast 11:31:29 carefish: can you change the forground colour ? 11:31:37 how do i do that Mat2-debugging? 11:32:00 and how do i call vdpPlotPixel? 11:32:40 vm.vdpPlotpixel didn't exist, and vdpPlotpixel didn't either, (in the pascal sample from git) 11:32:56 take a look at vdpROM.pas 11:33:02 sorry romVDP.pas 11:33:50 try 16 6 out 11:34:04 on my linux system it works 11:34:44 where would it put the pixel? 11:35:12 direct to the retro screen you see 11:35:51 the handle variable is tScreen, initiated in ng.pas 11:35:58 colour being specified how? 11:36:09 yeah, i saw that 11:36:13 choose values from 0 o 255 11:38:33 Mat2-debugging: can you give us the exact code you're running in retro? something we could put into the image and test? 11:39:13 start retro 11:39:18 then just write 11:39:24 65 16 6 out 11:39:43 this whould change the foreground colour 11:39:50 I make a screenshoot 11:41:25 http://i.imgur.com/5h2jD.png 11:41:45 hmm, what is the adr : LONGWORD value : BYTE plotpixel wants? 11:42:13 probably an offset into the display's ram 11:42:28 yes 11:42:29 Mat2-debugging: i don't think we're running the same code 11:43:04 i see message 16 is vdpRenderDisplay(tScreen) in ng_ports.pas 11:43:28 oh, that's an old version 11:43:29 which doesn't seem to take any arguments off th estack 11:44:01 and the explanation :) 11:44:18 well, how do i setup longword/byte? (first time i toy around with eh.. a console/pixel plotter/..how do you call this?) 11:45:04 you can take a look at vdpRenderChar, where plotPixel is used, it's the same routine as vdpPlotPixel 11:45:37 adr is a offset into the bitmap 11:45:43 ^an 11:46:03 offset into bitmap being? i tell it what pixel to plot from the bitmap i'm providing? 11:46:03 this is frustrating. i guess i don't understand something about how git works 11:46:21 tangentstorm: ok this is a git problem, were both out of sync 11:47:18 so to set up pixel 20, adr := 20 11:48:38 https://github.com/goldsmile/ngaro-pas/blob/master/src/ng_ports.pas 11:49:04 16: begin 11:49:04 data.pop1 (h); 11:49:04 vdpWriteFgReg (tScreen, h); 11:49:04 end; 11:49:22 hmm, the code is up to date 11:49:31 at least in the repro 11:51:16 Mat2-debugging: okay yeah i have that now on my local machine 11:51:23 all that routine does is set a variable though 11:51:46 that variable is the internal register for the forground colour 11:52:02 http://pastebin.com/D6UZxvqE something like this? (and where do i see it on the screen? 11:52:36 Mat2-debugging: okay, i'm with you in theory, but nothing changes on screen no matter what i set it to 11:54:45 uno momento 11:55:28 I need to compare all files because it can be romVDP is not up to date 11:55:49 http://img.is1337.eu/8779b932.png 11:57:14 tangentstorm: I have rebuild retro and on my system it works 11:57:29 (rebuild from the git repro) 11:58:24 --- join: erider (~chatzilla@unaffiliated/erider) joined #retro 11:58:39 okay i think lazarus may have been caching something 11:58:51 i deleted everything and am just running build.bat 11:58:58 hi all 11:59:07 hi erider 11:59:18 nothing :/ 11:59:24 hey erider 11:59:27 oh, that's interesting 11:59:37 what triggers the redisplay? 12:00:25 the retro listener 12:00:40 son of a... :/ it's leaving the color on the stack again :( 12:01:14 how is the game going 12:01:23 I think to undertand the problem 12:01:49 wait, I upload a tar archive 12:02:02 and now it's not. i'm going to save it as a function in the retro image so i know i run the exact thing every time 12:02:36 please extract the archive in one directory and build retro with fpc 12:02:45 ok 12:06:43 you will find backport.tar in the src directory of ngaro-pascal 12:06:58 try compiling this version 12:07:09 with -dGRAPHICAL 12:07:31 carefish: take a look at ng.pas 12:07:55 tangentstorm, what is the byte size of the retro-11.5 img supposed to be? 12:08:15 {terminal emulation} 12:08:15 cx,cy : longword; 12:08:15 count : longword; 12:08:15 tScreen : tVDP; 12:08:15 refresh : longword; 12:08:42 ^-- tScreen: that's the handle to the terminal bitmap 12:08:44 intothev01d: no idea. you can use the one from retroImage 12:09:24 oh ok, it's correct 12:09:30 just making sure I was doing it right 12:10:16 Mat2-debugging: running that version, i get the same thing 12:10:33 always green on purple 12:10:57 so I'm at the part where I read in the binary file to memory I think 12:11:03 but I have no clue what it's really doing 12:11:12 f = open( imgpath, 'rb' ) 12:11:13 memory = list(struct.unpack( cells * 'i', f.read() )) 12:11:27 I assume open the file in binary for read, then don't know what unpack and all is 12:11:46 intothev01d: that's python's way of breaking the file into 32-bit cells 12:11:59 tangentstorm: I don't have a clue why the same code works here and not for you 12:12:04 ohhh i see 12:12:56 cells = int(os.path.getsize(imgpath) / 4) doesn't this just get the whole size of the retroImage file / 4? 12:14:36 I'll push what I have so far 12:16:55 tangentstorm: try 65 16 6 out 12:17:10 tangentstorm, when you get a chance it's here https://github.com/intothev01d/ngaro-js/blob/master/src/retro.js 12:20:46 tangentstorm: I have uploaded a picture of it: retro.jpg where you see what's looking here 12:24:38 I must prepare for work, see you guys later 12:25:10 --- quit: Mat2-debugging (Quit: Verlassend) 12:26:56 intothev01d: yes it does... an integer is 4 bytes so the /4 gets the number of cells 12:27:13 oh i get it, 1,000,000 cells then 12:27:31 yeah 12:28:14 ok so I have that part to determine the number of cells, now I guess I need to put the binary file into an array with that many cells? 12:28:48 yeah. maybe there's some way to load the entire thing from disk as a typedarray? 12:29:11 not really sure. i've never used the node filesystem api 12:29:13 this is where I'm lost really. I'll take a look and see what node does with binary files 12:30:02 i can't seem to figure out how to draw a pixel on the retro screen :( 12:30:15 or could that be because i don't have the proper retroImage ? 12:30:39 carefish: no, it's because something is broken on windows 12:30:51 that's what mat2 and i have been talking about 12:31:17 so i could upload these pascal files to my vps and run it in there? 12:31:26 tangentstorm: I merged in your changes 12:31:32 crc: cool :) 12:36:59 wow this channel is busy now :-) 12:39:24 --- join: Mat2 (~Claude@91-65-144-133-dynip.superkabel.de) joined #retro 12:40:12 tangentstorm: short, have tested the code on another machine, where it works 12:40:25 hmm 12:40:35 on a win32 machine? 12:40:52 carefish: did you install the jedi-sdl headers, or are you just using what came with fpc? 12:41:07 no an old P4 wih ubuntu 12:41:24 tangentstorm: i'm using lazarus, that had sdl with it 12:41:39 yeah, i've tried both 12:42:00 carefish: do you know how to make sdl work in opengl mode instead of directx or whatever it's defaulting to? 12:42:14 or vice versa? 12:42:28 i know how to make opengl work with opengl yes 12:42:46 not that difficult 12:43:23 the only possible error left I can see: The colour information is no writen correctly ino video mem 12:43:34 for some graphic chips 12:43:37 SDL_Init(SDL_INIT_EVERYTHING); screen = SDL_SetVideoMode(640, 480, 32, SDL_OPENGL /* | SDL_NOFRAME */); 12:43:52 then a whole bunch of commands to init opengl 12:44:20 I try to compensate though RGB composing and see if it works the same 12:44:42 ok, later, ciao 12:44:46 --- quit: Mat2 (Quit: Verlassend) 12:45:14 SDL_HWPALETTE 12:45:16 Give SDL exclusive palette access. Without this flag you may not always get the colors you request with SDL_SetColors or SDL_SetPalette. 12:45:20 carefish: thanks 12:45:33 http://pastebin.com/fu2eXCKw 12:47:48 ^ the init stuff for opengl 12:48:39 carefish: can we still use the sdl bitmap stuff then? or do we have to use the opengl api's if we go that way? 12:49:38 you can't use the sdl bitmap stuff that way 12:49:44 you'd need a seperate screen for that 12:49:51 init'd w/ other params 12:50:00 SDL_OPENGLBLIT 12:50:01 Create an OpenGL rendering context, like above, but allow normal blitting operations. 12:50:10 oh.. :o 12:50:19 hmm, good to know, i didn't know about that :) 12:50:44 http://sdl.beuc.net/sdl.wiki/SDL_SetVideoMode 12:51:26 last time i've used the SDL wiki was about 4 months back :P 12:53:28 hmm 12:53:38 if i manually change the colors, it does change 12:53:45