Skip to content

tests: don't require strict equality for floats

On some architectures (e.g. i386 in Debian), obtaining strictly equal floating-point values is more complex than it looks, leading to test (and therefore build failures):

Expected: 1.400000
     Got: 1.400000
FAIL read_dng: Exposure time (../tests/check_dng.c:177)

Instead of requiring equality, we can work around this problem by adding a tolerance on the comparison.

1e-7 is an arbitrary tolerance selected as the lowest working value on the author's test setup.

Merge request reports

Loading