Inconsistent behavior for Arc

[ziyaointl]

Describe the bug
The behavior for Arc in p5py does match with the Processing Language

To Reproduce

from p5 import *

def draw():
    offset = 0
    x = 0
    while x < 2 * PI:
        arc((50.0 + offset, 50.0), 50.0, 50.0, x, 2 * PI)
        arc((50.0 + offset, 100), 50, 50, x, 2 * PI, 'OPEN')
        arc((50.0 + offset, 150), 50, 50, x, 2 * PI, 'CHORD')
        arc((50.0 + offset, 200), 50, 50, x, 2 * PI, 'PIE')
        x += QUARTER_PI
        offset += 50

if __name__ == '__main__':
    run()

void setup() {
  size(640, 360);
}

void draw() {
  int offset = 0;
  for (float x = 0; x < PI * 2; x += QUARTER_PI) {
     arc(50.0 + offset, 50.0, 50.0, 50.0, x, 2*PI);
     arc(50.0 + offset, 100, 50, 50, x, 2*PI, OPEN);
     arc(50.0 + offset, 150, 50, 50, x, 2*PI, CHORD);
     arc(50.0 + offset, 200, 50, 50, x, 2*PI, PIE);
     offset += 50;
  }
}

Expected behavior

Actual Behavior

System information:

• p5 release (version number or latest commit): 975f07e
• Python version: 3.8.3
• Operating system: MacOS 10.15.5

[ziyaointl]

Some progress on the tessellation branch:

[jeremydouglass]

@ziyaointl -- congratulations on this great work.

Could you comment on the remaining difference between the "expected" image above and the latest pipeline? It seems like there is a column of points and a line on the right edge that is missing. If that is the case, should those remaining differences in behavior be noted in an open issue -- even if they are not a high priority?

[ziyaointl]

@jeremydouglass Thanks! I think the column of points is actually a bug in Processing. The last column has start angle and end angle both set at 2*PI. Therefore, I think it's sensible for the behavior to be "not draw anything" but it could also be undefined. In p5.js, only PIE mode shows a small dot, and all other rows are empty.

[jeremydouglass]

Interesting! So would you suggest bug reports on the Processing repo / the p5.js repo?

It is an edge case, but nice if we can collectively minimize "undefined" as much as possible in visual outputs. arc() is already frustrating enough without also being inconsistent...

[ziyaointl]

Yeah, that's a good idea. I opened the respective issues here:
processing/processing#6063
processing/p5.js#4664