coordinate_system='confocalellipsoidal' does not get lg from coframe.

[rtoy]

Imported from SourceForge on 2024-07-09 19:41:37
Created by richardgobeli on 2021-10-12 01:57:59
Original: https://sourceforge.net/p/maxima/bugs/3876

---

When running this test file the fri tensor does not produce the lg after running cmetric(true);
Making each factor a separate sqrt function forces Maxima not to try canceling factors by changing sign of the factors and causing all off diagonal terms not to cancel.

This fri definition works.

  fri:matrix([-sqrt(e^2-v)*sqrt(e^2-w)/sqrt(f^2-e^2)/sqrt(g^2-e^2)/sqrt(e^2-u)/2,
              -sqrt(e^2-u)*sqrt(e^2-w)/sqrt(f^2-e^2)/sqrt(g^2-e^2)/sqrt(e^2-v)/2,
              -sqrt(e^2-u)*sqrt(e^2-v)/sqrt(f^2-e^2)/sqrt(g^2-e^2)/sqrt(e^2-w)/2],

             [-sqrt(f^2-v)*sqrt(f^2-w)/sqrt(e^2-f^2)/sqrt(g^2-f^2)/sqrt(f^2-u)/2,
              -sqrt(f^2-u)*sqrt(f^2-w)/sqrt(e^2-f^2)/sqrt(g^2-f^2)/sqrt(f^2-v)/2,
              -sqrt(f^2-u)*sqrt(f^2-v)/sqrt(e^2-f^2)/sqrt(g^2-f^2)/sqrt(f^2-w)/2],

             [-sqrt(g^2-v)*sqrt(g^2-w)/sqrt(e^2-g^2)/sqrt(f^2-g^2)/sqrt(g^2-u)/2,
              -sqrt(g^2-u)*sqrt(g^2-w)/sqrt(e^2-g^2)/sqrt(f^2-g^2)/sqrt(g^2-v)/2,
              -sqrt(g^2-u)*sqrt(g^2-v)/sqrt(e^2-g^2)/sqrt(f^2-g^2)/sqrt(g^2-w)/2]),

Attachments:

• https://sourceforge.net/p/maxima/bugs/3876/attachment/testing%20confocalellipsoidal.wxm

[rtoy]

Imported from SourceForge on 2024-07-09 19:41:38
Created by vttoth on 2021-10-12 04:01:06
Original: https://sourceforge.net/p/maxima/bugs/3876/#f89b

---

• status: open --> pending
• assigned_to: Viktor Toth