The std.math functions 'log', 'log10', 'log1p', 'log2' and 'logb' only have versions that accept a real argument and return a real value. Many other std.math functions, e.g., 'sin', 'cos', etc., also have versions that accept and return double as well as float. It would be preferable if the former group were consistent with the rest.
I believe this can be closed with version 2.102.0.
Not entirely, it seems the additions of `log1p` overloads for float and double have messed with the accuracy of those functions (beyond what is to be expected as a result of not doing intermediate calculations with reals). See: https://issues.dlang.org/show_bug.cgi?id=23750
THIS ISSUE HAS BEEN MOVED TO GITHUB https://github.com/dlang/phobos/issues/10356 DO NOT COMMENT HERE ANYMORE, NOBODY WILL SEE IT, THIS ISSUE HAS BEEN MOVED TO GITHUB