Should you be able to initialize a float with a char?

[user1234]

On Thursday, 19 May 2022 at 00:27:24 UTC, Walter Bright wrote:
> People routinely manipulate chars as integer types, for 
> example, in converting case. Making them not integer types 
> means lots of casting will become necessary, and overall that's 
> a step backwards.

it is indeed but let's be honest, having builtin char, wchar, and 
dchar, is only usefull for overload resolution and string 
literals.

     ubyte c = 's'; // OK
     ubyte[] a = "s".dup; // NG

without the string literal problem, there's only the overload 
resolution one and for this one builtin character types could be 
library types, e.g struct wrapping ubyte, ushort, uint.