In the function render_constellation we convert (to int) and round the coordinates with polar_to_win, and later pass these coordinates to draw_line_braille. I think in the middle of it we lose some precision, so it would be better to make the function draw_line_braille use "braille" coordinates by default, not ncurses coordinates. So we can do something like this:
if (config->braille)
{
int br_ya, br_xa, br_yb, br_xb;
polar_to_win(radius_a, theta_a, height * 4, width * 2, &br_ya, &br_xa);
polar_to_win(radius_b, theta_b, height * 4, width * 2, &br_yb, &br_xb);
draw_line_braille(win, br_ya, br_xa, br_yb, br_xb);
}And we can throw out a bunch of code from the function draw_line_braille :
void draw_line_braille(WINDOW *win, int ya, int xa, int yb, int xb)
{
// ncurses coordinates
int curs_xa = xa / 2;
int curs_ya = ya / 4;
int dx = abs(xb - xa);
int dy = abs(yb - ya);
// the rest of the function remains the same
}I tried it, and I think the "braille lines" are drawn more smoothly, but the stars of the constellations "jump around" too much, and it doesn't look too good. I believe somewhere the ncurses (!) coordinates of the stars might be calculated incorrectly, or it's a rounding error or something. The math is a bit complicated for me, so @da-luce if you have time, you can check out this problem.
In the function
render_constellationwe convert (to int) and round the coordinates withpolar_to_win, and later pass these coordinates todraw_line_braille. I think in the middle of it we lose some precision, so it would be better to make the functiondraw_line_brailleuse "braille" coordinates by default, not ncurses coordinates. So we can do something like this:And we can throw out a bunch of code from the function
draw_line_braille:I tried it, and I think the "braille lines" are drawn more smoothly, but the stars of the constellations "jump around" too much, and it doesn't look too good. I believe somewhere the ncurses (!) coordinates of the stars might be calculated incorrectly, or it's a rounding error or something. The math is a bit complicated for me, so @da-luce if you have time, you can check out this problem.