Underscores in big_str macro for floats

[mortenpi]

julia> 0.123_456
0.123456

julia> big"0.123_456"
ERROR: ArgumentError: invalid number format 0.123_456 for BigInt or BigFloat
Stacktrace:
 [1] top-level scope
   @ REPL[2]:1

It was mentioned in #24834 (comment), but I didn't find a dedicated issue. There used to be issue (#21092), but that was closed when the integer case was implemented.

[JeffreySarnoff]

yes absolutely (a thorn too long unpared)