Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!ames!pasteur!cory.Berkeley.EDU!koster
From: koster@cory.Berkeley.EDU (Anonymous user)
Newsgroups: comp.sys.amiga.tech
Subject: Re: glitching sprites while doing ScrollVPort
Message-ID: <6084@pasteur.Berkeley.EDU>
Date: 1 Oct 88 18:10:17 GMT
References: <2716@sugar.uu.net>
Sender: news@pasteur.Berkeley.EDU
Reply-To: koster@cory.Berkeley.EDU.UUCP (Anonymous user)
Organization: University of California, Berkeley
Lines: 22

In article <2716@sugar.uu.net> karl@sugar.uu.net (Karl Lehenbauer) writes:
>I'm having some problems with a little technology demonstrator I'm cooking
>up.  It is a lunar lander that vertically and horizontally scrolls a
>320X200 screen over a 640X400 IFF ILBM using ScrollVPort.  The problem is
>that when the screen is scrolling, the sprites glitch, show as vertical
>crud for the whole height of the display and such.

I think for scrolling the data fetch for each scan line must be one word
earlier, which is where sprites 5,6 and 7 have their DMA channels. This is
discussed in the hardware manual. The solution may be to only use sprites
0-4, or knock off one word at the beginning of each line, making it a
304 x 200 screen, or perhaps just shifting everything to the right by
one word, and moving the H.POSITION knob on your monitor.

The symptoms you describe (tall vertical lines, garbage sprites that come
in and out) are exactly the ones I found when someone had his preferences
display position way to the left, and was running a program that used
sprite 5. That was REALLY fun to debug! I just assumed it was my fault, and
just couldn't find the bug.

David Ashley
koster@cory.berkeley.edu