[clang-tidy][request] Detect floating-point casts that cause loss of precision

[chrchr-github]

double f(double a, double b, float c) {
    return a + (float)b + c;
}

The result of the cast is used as double, so it is unlikely to be intentional. There should be no warning when the cast affects overload resolution though (if that is even possible).

[firewave]

Looks like this is already coved by -Wdouble-promotion.

<source>:2:27: warning: implicit conversion increases floating-point precision: 'float' to 'double' [-Wdouble-promotion]
    2 |     return a + (float)b + c;
      |                         ~ ^
<source>:2:16: warning: implicit conversion increases floating-point precision: 'float' to 'double' [-Wdouble-promotion]
    2 |     return a + (float)b + c;
      |              ~ ^~~~~~~~

https://godbolt.org/z/edfMs737P

[chrchr-github]

-Wdouble-promotion warns about the implicit conversion to double, but not about the nonsensical explicit cast to float.