Re: [ft] FT_FloorFix and FT_CeilFix

[Werner LEMBERG]

Hello Wijciech!


Sorry for the very late response :-)

> FT_FloorFix does not round towards minus infinity and FT_CeilFix
> towards plus infinity for fractional negative values.  This is
> inconsistent with expected semantics as well as the docs, according
> to which these functions work as expected.  The problem with the
> docs is, that FT_RoundFix is semantically correct, yet the
> description is incorrect w.r.t. to its semantics.  Werner pointed me
> out these are solitary (legacy?) functions not used in FT anywhere,
> added back in 2001.
>
> Example for FT_FloorFix: http://coliru.stacked-crooked.com/a/5f2b10846110c3d7

I suggest fixing both the functions and the descriptions to do the
right thing.  It seems to me that they were never tested because they
aren't used.

Can you provide a patch?


    Werner