Uploaded image for project: 'Qt'
  1. Qt
  2. QTBUG-29158

QTriangle3D::intersection(const QRay3D &ray) const

    XMLWordPrintable

Details

    • Bug
    • Resolution: Out of scope
    • P2: Important
    • None
    • 5.0.0
    • Qt3D
    • None

    Description

      float QTriangle3D::intersection(const QRay3D &ray) const

      seems to deliver wrong results.

      The point calculated on the planes surface does not always return true when tested with contains.

      It seems to be a problem with the test against the triangles normal which is the first test done in contains.

      the normal is not normalized and can be a large value which leads to qFuzzyIsNull returning false for the intersection point calculated before.

      example:
      ray with origin (0, 0, 0) and direction (0.995806098, 0.017942428, 0.08971226215)
      triangle with (70, -15, 6), (70, 15, 6), (40, 15, 6)

      possible solution:
      normalize the normal used to test for a component perpendicular to the planes surface.

      hints:
      other 3D geometry classes should be checked, wheter they may use similar checks which may lead to this error (normals should always be normalized?!?)

      More Info:
      In this particular case its a problem in contains().
      The calculated normal should be normalized.
      But it is a general problem with the 3D classes.
      The normals (from plane and triangle) are not normalized but they are used to test for contains() in a qFuzzyIsNull() comparison.
      For large normals this may lead to Problems.
      At least the call to QTriangle3D::intersection(const QRay3D &ray) should be consistent!
      You calculate a point on the triangles plane which is then checked in a contains() call on the triangle and returns false.
      I think you need to have a look at all the basic geometry classes and check the normalization of normals and how they are used.

      Attachments

        For Gerrit Dashboard: QTBUG-29158
        # Subject Branch Project Status CR V

        Activity

          People

            seanharmer Sean Harmer
            apfelstruhdl David Schug
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Gerrit Reviews

                There are no open Gerrit changes