Fix incorrect case sensitivity in floating-point number parsing methods

[PopSlime]

Summary

The Single.Parse, Single.TryParse, Double.Parse, and Double.TryParse methods are case-insensitive in .NET Core 3.0 and later, when parsing string representations of PositiveInfinity, NegativeInfinity, and NaN.

References

• dotnet/coreclr@b30280d
• Fixing up the Double/Single parsing code to be correct coreclr#20707 (comment)

[gewarren]

I left a suggestion to try and make it more clear that all of the symbols are compared case insensitively.

[tannergooding]

Changes overall LGTM, but it looks like the suggestions from @gewarren still need to be taken and the latter suggestion should be applied to all the similar places. That is, ensuring they use a separate sentence of the same form rather than putting the information in parentheses.