This looks a lot cleaner. It helps if I don't process off the end of a span list.
While there are presumably DSP issues causing portions of the tank to either not show up or draw incorrectly, I'm not sure if it's that my processing of the list and the associated framebuffers is wrong (i.e. ending early), or if it's that the DSP isn't actually issuing enough data to draw the full tank. I do see a few "pulled" triangles every so often, so there might still be DSP issues lurking.
But at this point I believe the DSP is in better shape than I'd expected, and the real problem is figuring out how the rasterizer hardware works: How does it get hidden, how does it get shown, how are the 2 layers combined, that sort of thing. If anyone is willing to take a stab at it, I'm willing to commit my changes to a branch.
It warms my heart to see Moogly doing an amazing job picking up the piles of junk I left lying around 10 years ago. This likely never would have gotten done without that guy. He's been doing some great MAME stuff in recent years.
And let this be a reminder to all y'all that there is still some rad arcade hardware out there whose emulation would benefit from time and effort (maybe yours?).
> It warms my heart to see Moogly doing an amazing job picking up the piles of junk I > left lying around 10 years ago. This likely never would have gotten done without that > guy. He's been doing some great MAME stuff in recent years. > > And let this be a reminder to all y'all that there is still some rad arcade hardware > out there whose emulation would benefit from time and effort (maybe yours?).
You're making a really good point, but let me underscore it:
There's nothing special about what I'm doing. I knew nothing at all about the DSP56156 before I decided "Hey, I want to look into this."
It took a lot - this this the culmination of 10 days of me angrily stepping through the DSP code, looking each instruction up in a manual I found online, and jotting down what it "should" be doing in C-like code over in notepad, and investigating each time the results seemed to diverge from what the manual said they should do.
But it's also not anything special. It's something any coder could do with enough gumption.
Turns out that some debug code from when I was first getting the span draw-er up and running, which was only accepting spans that didn't exceed the presumed clipping range of the framebuffer viewport, was biting me in the ass.
Obviously still some DSP bugs lurking, but with Friday being a holiday and the potential for me taking Thursday off, I'm in a good position to get this whole thing done and dusted for the release.
Index 
Threaded 
MAME WIP: Polygonet Commanders (Update)
Reply
[Re: 
