Parsing floating point numbers with large or small absolute value

[exogba]

When parsing floating point numbers with very small or very large absolute values, the result of the parsing will be an error with error type 'value_out_of_range'.
auto result = scn::scan<float>("2e-1000", "{}"); // result will be an error
auto result = scn::scan<float>("2e1000", "{}"); // result will be an error
My question would be if this is the intended behaviour or not. Scanf for example would succeed in those cases and the resulting values would be 0/inf. And intuitively that is the behaviour I was expecting for scnlib as well.

[nicoheidtke]

Any opinions on this? @eliaskosunen
I too think a parsing error might not be the best solution.