intersection() infinite loop for valid MultiPolygons.

[erikh2000]

I tried to narrow this down as tight as I could. The initial polygons that generated the error had many more points, but I removed as many as possible to still produce the error. This call below will cause an infinite loop and crash the JS stack:

const multiPolygon1 = [
          [
            [
              [-89.7235275385363, 42.0439388757895],
              [-89.7237188969195, 42.0442244474719],
              [-89.7237975, 42.0456047],
              [-89.7235275385363, 42.0439388757895]
            ]
          ],
          [
            [
              [-89.7214720193067, 42.042492561329],
              [-89.7161784, 42.046186],
              [-89.7195447358245, 42.042911552283],
              [-89.7214720193067, 42.042492561329]
            ]
          ]
        ];
const multiPolygon2 = [
          [
            [
              [-89.7232165, 42.0413833],
              [-89.723708, 42.0440331],
              [-89.7237975, 42.0456047],
              [-89.7145748, 42.0462076],
              [-89.7144372, 42.0462097],
              [-89.7239301, 42.0397921],
              [-89.7232165, 42.0413833]
            ]
          ],
          [
            [
              [-89.7240178, 42.0405696],
              [-89.7239695, 42.0422763],
              [-89.7236804, 42.0423162],
              [-89.7240178, 42.0405696]
            ]
          ]
        ];
intersection(multiPolygon1, multiPolygon2);

I stepped through it in the debugger. connectEdges() gets caught in an endless loop on the i=6 iteration of the for(i = 0 loop. More specifically in the inner while(pos >= i) loop where nextPos() continuously returns 12 for the next pos value, and the same coordinate is added to resultEvents until the JS stack runs out of memory.

It's a little tough for me to understand how the code is meant to work with my current ignorance, or I would probably just fix the bug and send you the fix. I'll keep messing around with it. Insights, ideas, or of course, your own bug fix would be very welcome.

[erikh2000]

Confirmed the above bug is still there after a fresh git pull and rebuild just now.

[rowanwins]

I'll try and take a look this evening if pulls #58 and #60 help resolve this, I suspect #58 might. It's a one-liner fix so should be pretty easy to add to your code locally if you want to look at it straight away.

[rowanwins]

Yep I can confirm this is resolved by pull requests #58 and #60 thanks to @mfogel

[erikh2000]

I hand-merged #58 and #60 into my local code and YES!... the test case above passes with the changes. Very nice.

[w8r]

It's resolved in v0.4 https://codepen.io/w8r/pen/bMXPEy