Multisprite kernel scanline issues

[+Al_Nafuur]

I am having multiple scannline issues with 1942 and the multisprite kernel.

 

• scanline jumps up to 278 or 293 lines if all 5 virtual sprites are parked at the same Y value above the visible screen ( Y > 88)
• scanline jumps to 263 lines if a virtual sprite has an X position > 160
• scanline jumps to 290 lines for one frame if an virtual sprite moves from 0 to 255

 

Even if I avoid all situations metion above, I still have some frames with 263 scanlines!

 

Attached my modified ex_multisprite_9_objects.bas binary where the virtual sprites can be move offscreen:

ex_multisprite_9_objects.bas.bin

[+Karl G]

9 hours ago, Al_Nafuur said:

scanline jumps up to 278 or 293 lines if all 5 virtual sprites are parked at the same Y value above the visible screen ( Y > 88)

This one I wasn't able to reproduce. At what Y value does this happen to you? You might want to experiment with different values for "offscreen" to see if there's a value that doesn't trigger that.

 

9 hours ago, Al_Nafuur said:

scanline jumps to 263 lines if a virtual sprite has an X position > 160

That I see too, but there's also no reason to let X exceed 160. After 159, you can reset the X position to 0, and not have to deal with that issue. I can't think of a case where the X value would need to be 160 or greater.

 

9 hours ago, Al_Nafuur said:

scanline jumps to 290 lines for one frame if an virtual sprite moves from 0 to 255

Both of these values would be offscreen, so I think you should be able to work around this by not dropping below 0, or jumping from 0 to a safe offscreen value.

 

As you have noticed, the Multisprite kernel has a few "quirks", and is a bit less polished than the other kernels.

 

[+Gemintronic]

Al_Nafuur mentioned in the 1942 topic the scanline issue was fixed.

 

What was the fix?  I'm making a lot of multi sprite kernel games and would like to avoid such issues.

[+Al_Nafuur]

52 minutes ago, Gemintronic said:

Al_Nafuur mentioned in the 1942 topic the scanline issue was fixed.

Not the multikernel scanline issue was fix. Just a bug in the 1942 code was fixed that caused one of the situations mentioned above.

 

 

52 minutes ago, Gemintronic said:

What was the fix?

basically a "dim" was used where "const" should have been used

https://github.com/Al-Nafuur/PlusROM-Hacks/commit/271ae18637e27b59d155d5e140fea511e7e60d63

 

52 minutes ago, Gemintronic said:

 

  I'm making a lot of multi sprite kernel games and would like to avoid such issues.

You have to avoid the situations mentioned in my first post.