NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/alt/sys/amiga/demos/1016

Internet Message Format | 1992-07-25 | 1.9 KB

Xref: sparky alt.sys.amiga.demos:1016 comp.sys.amiga.programmer:11634
Path: sparky!uunet!olivea!mintaka.lcs.mit.edu!hal.gnu.ai.mit.edu!psteffn
From: [e-mail] (Paul Steffen)
Newsgroups: alt.sys.amiga.demos,comp.sys.amiga.programmer
Subject: Re: HELP with filled vector vertex "fixes?"
Message-ID: <[e-mail]>
Date: 25 Jul 92 12:13:20 GMT
References: <[e-mail]>
Sender: [e-mail]
Organization: /etc/organization
Lines: 30

In article <[e-mail]> [e-mail] writes:
>
>    I was looking at some filled-vector code and I was wondering: Why do
>certain vertices need to be "fixed" (with BCHG's, etc), while others do not?
>
>    Toying with some of my own routines, I get lots of streaks and glitches
>when I don't "fix" the vertices.  What, exactly, are the conditions under which
>this needs to be done?
>
>    Thanks!
>
>Kevin

Well, the way the blitters fill mode is quite simple.  Imagine the blitter
as being fed a string of bits.  Depending on current state of the blitter,
it will output a string of 0's or 1's.  Anytime, there is a set bit on the
blitter's input, the state is reversed.  Therefore;
source = 0010010001000001001001   becomes,
destin = 0011110001111111001111

Generally, what involves polygon filled vectors is plotting the lines first
in a line mode where the blitter leaves one pixel per scan line.
As the blitter really only fills between the points, this is usually quite
suffice.  The purpose of checking corners is merely to make sure that
the blitter doesn't miss a pixel as it's filling which causes the 'bleed'
off to the left.  Obviously, if there is only 1 bit set on the line,
the blitters logic will remain SET to the end of line so you need to make
sure that the filling remains within the boundries of the polygon.

STratoHAK/TPPi