test_spheroid_area fails

[strk]

This is on a 64bit linux system:

  Test: test_spheroid_area ...
sphere: 12360265021.4
spheroid: 12304814950.1
FAILED
    1. cu_geodetic.c:993  - CU_ASSERT_DOUBLE_EQUAL(a1,89.7127703297,0.0001)

[robe]

Debbie and Winnie are similarly unhappy.

Debbie: 64-bit Debian 6

sphere: 12360265021.4
spheroid: 12304814950.1
FAILED
    1. cu_geodetic.c:993  - CU_ASSERT_DOUBLE_EQUAL(a1,89.7127703297,0.0001)

Winnie - Mingw64 32-bit compile

 Test: test_lwgeom_area_sphere ...FAILED
    1. cu_geodetic.c:1173  - CU_ASSERT_DOUBLE_EQUAL(area,12360265021.3561,0.01)
  Test: test_signum ...passed
  Test: test_gbox_from_spherical_coordinates ...passed
  Test: test_gserialized_get_gbox_geocentric ...passed
  Test: test_clairaut ...passed
  Test: test_edge_intersection ...passed
  Test: test_edge_distance_to_point ...passed
  Test: test_edge_distance_to_edge ...passed
  Test: test_lwgeom_distance_sphere ...passed
  Test: test_lwgeom_check_geodetic ...passed
  Test: test_gserialized_from_lwgeom ...passed
  Test: test_spheroid_distance ...passed
  Test: test_spheroid_area ...
sphere: 12360265021.4
spheroid: 12304814950.1
FAILED
    1. cu_geodetic.c:993  - CU_ASSERT_DOUBLE_EQUAL(a1,89.7127703297,0.0001)

Winnie - 64-bit mingw64 compile

sphere: 12360265021.4
spheroid: 12304814950.1
FAILED
    1. cu_geodetic.c:993  - CU_ASSERT_DOUBLE_EQUAL(a1,89.7127703297,0.0001)

[pramsey]

It's better on debbie at r10420. Probably still a failure on winnie? It's a hard call, the new area approach is less numerically stable, so we're getting a wider range of results off of different platforms. On the other hand, it deals naturally with over-the-poles and near-the-dateline in a way the old approach did not. Hm.

[robe]

winnie is fine too as far as this is concerned even the win 32 CUnit seems okay.