[palito-dev] dashcircle

[Michael Devey]

ok moving dashes and boundary conditions are hideous and I give up for 
tonight.  I'll crack a clean version of dashed line this week some time, 
but to start getting patrols to look better here's dashcircle.

maybe I'll do semidashcircle as well but I feel it's too much effort for 
nothing.

subnote for barrett - previously all my complaining about wrong 
directions was infact using atan2(x,y) instead of atan2(y,x) (wrong 
index in vectors...) grrrr. Also I was using a completely unnecessary 
manipulation on the initial angle - I should have just got it's 
adjustment from the manipulated radius (As I have now) moo!

void flush2d_dashcircle(float x, float y, float r, float moves, int num)
{
   /*    
   flush lod?
   3 lines (4 vertices) per small arc until lod.
   */
       #define MAGICMYKE_pi 3.141592654
       #define MAGICMYKE_dash_vertices 4
       int vertices  = MAGICMYKE_dash_vertices;
       float v[2], v1[2] = { x,  y },  v2[2] = { r, 0 };    
       float radius, startR, sweepR;
       int *ip, i, j;
       
       if(!dynbuf((void **) &ipoints, &isize, 
(vertices*num)*sizeof(int)*2))
           return;
       ip = ipoints;
       sweepR = MAGICMYKE_pi / (vertices-1) / num;
       afim2_mult_vector2(v, sp, v1);
       matrix2_mult_vector2(v1, sp, v2);
       startR = atan2 (v1[1], v1[0]) + moves;
       radius = sqrt(v1[0]*v1[0] + v1[1]*v1[1]);
       for(j=0; j < num ; j++){
           for(i = 0; i < vertices; i++, startR += sweepR){
               *ip = (int) ( v[0] + radius * cos(startR) );  ip++;
               *ip = (int) ( v[1] + radius * sin(startR) );  ip++;
           }
           flush_polyline( ip -(vertices*2) , vertices);
           startR += sweepR*(vertices-2);
       }
}