Fix arc track hit test for arcs crossing 0 degrees

[DominikPalo]

pointWithinDistanceToArc (web/render.js), used by netHitScan to find the net under a click on arc tracks, normalizes the start and end angles to [0, 2π) and then checks whether the click angle falls inside the arc. Arc tracks are drawn with ctx.arc(center, r, start, end), i.e. going from start with increasing angle to end. When the normalized start is greater than the end, the arc wraps through 0 and covers angle >= start || angle <= end.

The wrap-around branch had the angles swapped:

if (angle1 > angle2)
  return (angle >= angle2 || angle <= angle1);   // true for every angle

Since angle2 < angle1, every angle satisfies one side, so for any arc crossing 0° a click anywhere on its full circle (within the track width) selected the arc's net. Arcs not crossing 0° were unaffected.

The fix swaps the comparison to angle >= angle1 || angle <= angle2.

Testing

I loaded pointWithinDistanceToArc from master and from this branch in Node and evaluated 24 points (every 15°) on each arc's circle against the range covered by ctx.arc(start, end):

Arc (start → end)	crosses 0°	wrong results on master	wrong results with this PR
-10° → 10°	yes	23 / 24	0
350° → 20°	yes	22 / 24	0
270° → 90°	yes	11 / 24	0
-90° → 0°	yes	17 / 24	0
30° → 120°	no	0	0
90° → 270°	no	0	0

The reported case, a click at 180° on a -10°..10° arc, returned true on master and returns false now. Points off the circle are rejected by both versions as before.

🤖 Generated with Claude Code