[GEOMETRY-99] Inconsistent equals and hashCode

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Labels: beta1
• Assignee: Unassigned
• Reporter: Matt Juntunen
• Created: 04/Jul/20 20:23
• Updated: 21/Aug/21 03:45
• Resolved: 05/Jul/20 01:28

Description

The following classes have inconsistent behavior between equals and hashCode in regard to +0.0 and -0.0:

• EpsilonDoublePrecisionContext
• Vector1D
• Vector2D
• Vector2D
• SphericalCoordinates
• PolarCoordinates
• AffineTransformMatrix2D
• AffineTransformMatrix3D
• AxisAngleSequence

This inconsistency comes from the use of == or Precision.equals to compare doubles in the equals method (which consider +0.0 and -0.0 to be equal) and Double.hashCode in the hashCode method (which returns different hash codes for +0.0 and -0.0). Hence, instances could be considered equal but have different hash codes.

We should standardize on the use of Double.compare in equals and Double.hashCode in hashCode.

Issue Links

• links to: GitHub Pull Request #93