Regress on winnie cu_print

[robe]

Winnie's CUnit hasn't been running for some time since th pkg-config changes in 3.0, so not sure how long this issue has been happening. Now CUnit is running again after [d3d372eb89c10821654007f1a8872916aa9ffc1b/git] -

Though I don't recall seeing this error on my desktop which should be running much the same chain as winnie except for version of GEOS / GDAL / Proj

Suite: ptarray
  Test: test_ptarray_append_point ...passed
  Test: test_ptarray_append_ptarray ...passed
  Test: test_ptarray_locate_point ...passed
  Test: test_ptarray_isccw ...passed
  Test: test_ptarray_signed_area ...passed
  Test: test_ptarray_insert_point ...passed
  Test: test_ptarray_contains_point ...passed
  Test: test_ptarrayarc_contains_point ...passed
  Test: test_ptarray_scale ...passed
Suite: printing
  Test: test_lwpoint_to_latlon_default_format ...passed
  Test: test_lwpoint_to_latlon_format_orders ...passed
  Test: test_lwpoint_to_latlon_optional_format ...passed
  Test: test_lwpoint_to_latlon_oddball_formats ...passed
  Test: test_lwpoint_to_latlon_bad_formats ...passed
  Test: test_lwprint ...passed
  Test: test_lwprint_roundtrip ...[cu_print.c:553]
 Expected: 1.#INF
 Obtained: 0
[cu_print.c:554]
 Expected: -1.#INF
 Obtained: 0
FAILED
    1. cu_print.c:553  - CU_ASSERT_EQUAL(atof(s),(double)__builtin_inff())
    2. cu_print.c:554  - CU_ASSERT_EQUAL(atof(s),(double)-__builtin_inff())

[Algunenano]

Fricking Windows. Those tests are checking that if you pass infinity, print it and read it back you get the same number.

Though I don't recall seeing this error on my desktop which should be running much the same chain as winnie except for version of GEOS / GDAL / Proj

These are pure Postgis / C code. Are you testing current HEAD in your desktop too? Might be a mingw difference or something like this.

Let me know if you can check it, if not I'll disable those 2 tests since they are not super important.

[robe]

Yes testing head in both cases. The good news is that it fails on my desktop too. I had been suffering from not running Cunit tests too.

  Test: test_lwprint_roundtrip ...[cu_print.c:553]
 Expected: 1.#INF
 Obtained: 0
[cu_print.c:554]
 Expected: -1.#INF
 Obtained: 0
FAILED
    1. cu_print.c:553  - CU_ASSERT_EQUAL(atof(s),(double)__builtin_inff())
    2. cu_print.c:554  - CU_ASSERT_EQUAL(atof(s),(double)-__builtin_inff())

Anyway I stillneed to check stable-3.0 to see if it's an issue there too.

[Algunenano]

I introduced the behaviour and the tests recently, so it will only affect master. I'm away from a PC until next week so feel free to either comment those 2 tests out or I'll do it when I'm back.

[Raúl Marín <git@…>]

In c0d947e/git:

Disable INFINITY rountrip tests

They fail in Windows / MinGW for some reason
Closes #4735