compilation error in test file

[doudou]

fcl/test/narrowphase/detail/convexity_based_algorithm/test_gjk_libccd-inl_gjk_doSimplex2.cpp:181:67: note:   deduced conflicting types for parameter ‘_Tp’ (‘double’ and ‘float’)
     const ccd_real_t eps = std::max({1., p_OA.norm(), p_OB.norm()}) * kEps;

Changing the 1. to 1.0f does fix the error, but I'm guessing that the double/float error might be a configuration thing (maybe ?).

[sherm1]

Whoops! Should be T(1) (where T is the template argument) to avoid committing to float or double prematurely (assuming we're still supporting float precision).

This suggests that CI doesn't include float tests?

[SeanCurtis-TRI]

Actually, in this case, it should be ccd_real_t(1). This test isn't expressed in terms of T, but in terms of ccd_real_t. And, clearly, this test is running against a float-based libccd. I'll patch the test.

[SeanCurtis-TRI]

The PR for this has just been posted. I've confirmed the reported compilation error and that this fixes it (on my machine). @doudou, if you'd like to confirm as well, that'd be appreciated. Otherwise, we'll just go straight through PR review and merge it.

[doudou]

@SeanCurtis-TRI yep, #373 does the trick