WIP: Salmon Run 2600

[tschak909]

I am in the process of doing a port of Bill Williams' great game Salmon Run, originally done on the Atari 800. So far, I have been able to get smooth scrolling working (and have cleaned up a few of the little glitches that have been popping up in the kernel...) however, I have a bit of a problem....

 

as it stands, I need to be able to vary the scrolling of the game based on how long the joystick has been pushed upward... I can have this if I do all of the coarse and fine scrolling calculations inside the vertical blank. However, once the entire buffer has been scrolled, it resets the entire moving window to the start. I want the moving window to be able to wrap around.....

 

Nukey Shay suggested the calculations should be done in kernel, which I understand why, it allows for the next river line to be immediatelly pulled, without the jump caused by the entire window being calculated at the vertical blank..........however.........

 

it also makes varying the scroll speed ____MUCH____ more difficult, and I am trying to come up with a solution which will allow me to control the scrolling velocity...... but i seem to be STUCK..... does anyone have a better approach?

 

I am pasting a current copy of the kernel so you guys can see......

 

salmonrun_070312.zip

[SeaGtGruff]

However, once the entire buffer has been scrolled, it resets the entire moving window to the start. I want the moving window to be able to wrap around.....

You have two minor problems (I edited your code slightly-- changed tabs to spaces, removed the comments, and put the line labels on separate lines):

 

GameDisplay
	LDA #$FF
	STA PF0
	LDX CRSSCROLLCTR
	AND #$1F
TopRiver

 

In the above section, the AND doesn't have any effect on the X register, so you can take it out-- it isn't needed, anyway, since you take care of that later on when you increment CRSSCROLLCTR.

 

Here is where your real problem is:

 

MidRiver
	REPEAT 20
	LDA PF1DATA,X
	STA PF1
	LDA PF2DATA,X
	STA PF2
	STA WSYNC
	STA WSYNC
	STA WSYNC
	STA WSYNC
	STA WSYNC
	STA WSYNC
	STA WSYNC
	DEX
	STA WSYNC
	REPEND

 

When you decrement X, it can go from 0 to 255, so you need to do that AND here. But AND works on the accumulator, not on X, so you can do this...

 

	STA WSYNC
	DEX
	TXA				   ; <--- Add these three
	AND #$1F			  ; <--- lines right after
	TAX				   ; <--- you do DEX
	STA WSYNC

 

Michael

Edited March 12, 2007 by SeaGtGruff

[tschak909]

Thanks! how did I miss that? :-/

 

Okay, moving on to the joystick/velocity code.

 

-Thom

[SeaGtGruff]

(I edited your code slightly-- changed tabs to spaces, removed the comments, and put the line labels on separate lines):

By the way, sorry about that-- it was my stupidity, actually. Whenever I tried to compile your code, I would get weird compile errors right after the first copy of the MidRiver section (which is repeated 20 times), and I was trying every zany thing I could think of to fix it. The other night I actually removed the REPEAT and REPEND, and just copied and pasted those lines of code 20 times! So tonight, after I tried changing tabs to spaces, taking out the comments, moving the line labels to their own lines, etc., I *finally* figured out that it was because you're on Unix and I'm on Windows... so all I *really* needed to do was convert the file to CR/LF format! I should know better! But the weird thing was, DASM for DOS/Windows didn't care that the file wasn't in CR/LF, *except* when it came to the REPEAT and REPEND section, which got garbled after the first copy.

 

Michael

[vdub_bobby]

I am in the process of doing a port of Bill Williams' great game Salmon Run, originally done on the Atari 800. So far, I have been able to get smooth scrolling working (and have cleaned up a few of the little glitches that have been popping up in the kernel...) however, I have a bit of a problem....

 

as it stands, I need to be able to vary the scrolling of the game based on how long the joystick has been pushed upward... I can have this if I do all of the coarse and fine scrolling calculations inside the vertical blank. However, once the entire buffer has been scrolled, it resets the entire moving window to the start. I want the moving window to be able to wrap around.....

Couldn't you make the end of the buffer the same as the beginning...like this:

Now:

 

With end the same as beginning:

 

It uses more ROM but simplifies your kernel and you get smooth wraparound.

[tschak909]

actually the fix that SeaStGruff proposed fixed the problem (I was overflowing the index). I am currently working on the joystick support code.

[+atari2600land]

What's it about for people who don't know (myself included).

[tschak909]

Salmon Run is a game in which you are a salmon which is swimming upstream. The object is to swim upstream as fast as possible, jumping against waves, while avoiding bears, fishermen with nets, and birds, which will stop at nothing to snatch you from the water. If you swim far enough, you will reach the lake, at which point you will meet up with a girl salmon, and.... :-) well you know.... so... anyway, you get an extra life. :-)

 

I am attaching both a binary and a screenshot.

 

<-- atari 800 salmon run

 

Salmon_Run.zip

[tschak909]

Just giving a quick status update, I am rewriting the kernel to display an asymmetrical playfield for the river display. After a few false starts, I have been able to simplify to a list of page offsets to up to 256 possible combinations of river, biased by the difficulty switch and the current level.

 

-Thom

[tschak909]

Just giving a quick status update, I am rewriting the kernel to display an asymmetrical playfield for the river display. After a few false starts, I have been able to simplify to a list of page offsets to up to 256 possible combinations of river, biased by the difficulty switch and the current level.

 

-Thom

 

still working on it, i had to slow down considerably due to work obligations, but things are stabilising again. I have calculated 1024 bytes of playfield data, and am in the process of entering it all (I am throwing together a quick script to take the playfield i drew manually and provide the PF1/PF2 data.) Then I can do a pass on the new kernel. I've been studying a lot of disassemblies, and learning VERY quickly....)

 

-Thom

[tschak909]

I am resuming work on Salmon Run, I had to calculate 255 possible random combinations for playfield state, I did it all inside a text file, separated into columns in which I first drew the binary patterns, and then calculated the PF1, PF2, PF3, and PF4 data accordingly....

 

A snippet of the graphic notepad file:

 

RIVER1	 RIVER2	  RIVER3   RIVER4
PF1	 PF2	  PF2	   PF1	
Normal	 Reversed Normal   Reversed
---------------------------------------------
00	11000000 00000000 00000000 00000011	192	0	0	192
01	11110000 00000000 00000000 00001111	240	0	0	240
02	11111100 00000000 00000000 00111111	252	0	0	252
03	11111111 00000000 00000000 11111111	255	0	0	255
04	11111111 11000000 00000011 11111111	255	3	3	255
05	  11111111 11110000 00001111 11111111	255	15	15	255
06	11111111 11111100 00111111 11111111	255	63	63	255
07	  11000000 00000000 00000011 11111111	192	0	3	255
08	11110000 00000000 00001111 11111111	240	0	15	255

 

I have attached the code here:

 

riverdata.zip

 

I also developed a quick tool to output the necessary .byte statements using Ruby:

 

# salmon run graphic dump dev tool
# t.cherryhomes 2007-04-05


pf1data = Array.new
pf2data = Array.new
pf3data = Array.new
pf4data = Array.new

pf1binary = Array.new
pf2binary = Array.new
pf3binary = Array.new
pf4binary = Array.new

l = 0

File.open("riverdata.txt") do |f|
	while line = f.gets

			if (l < 4) then

					# do nothing.

			else
					currentLine = line.split(/\t/);
					currentBinary = currentLine[1].split

					pf1data.push(currentLine[2]);
					pf2data.push(currentLine[3]);
					pf3data.push(currentLine[4]);
					pf4data.push(currentLine[5]);

					pf1binary.push(currentBinary[0]);
					pf2binary.push(currentBinary[1]);
					pf3binary.push(currentBinary[2]);
					pf4binary.push(currentBinary[3]);
			end

	l+=1;

	end
end

File.open("riverdata.asm","w") do |f|

	f.puts "PF1Data"
	pf1data.each do |i|
			f.puts "\t.byte " << i
	end

	f.puts "PF2Data"
	pf2data.each do |i|
			f.puts "\t.byte " << i
	end

	f.puts "PF3Data"
	pf3data.each do |i|
			f.puts "\t.byte " << i
	end

	f.puts "PF4Data"
	pf3data.each do |i|
			f.puts "\t.byte " << i
	end

end

puts "Done.\n"

 

and it works just fine. :-)

 

Now, I am going to hammer out the asym kernel, will keep you guys posted :-)

 

-Thom