fix result of String.toFloat "." (#336)

[pisys]

which is now a Err instead of Ok NaN.
Also added tests for String conversion.

[evancz]

Thanks!

I wonder if this is how we should be implementing this in the first place. I don't really remember why I chose this route. Maybe it would be better to use a regex like [-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)? or to just put parseFloat in a try block and trust that it'll do the right thing?

I also wonder why parseFloat('.') == NaN in JavaScript. Can you try to find out why that is true?

[evancz]

Thanks!

I wonder if this is how we should be implementing this in the first place. I don't really remember why I chose this route. Maybe it would be better to use a regex like [-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)? or to just put parseFloat in a try block and trust that it'll do the right thing?

I also wonder why parseFloat('.') == NaN in JavaScript. Can you try to find out why that is true?

[pisys]

parseFloat is less strict than the current implementation. It allows whitespaces, exponents, and only demands a prefix of the input string to be a float. Since "." is not a Float parseFloat returns NaN!

However, parseFloat can also parse "Infinity", which is a Number in JavaScript but not in Elm! So I discourage the mere wrapping of parseFloat until there is a explicit Infinity thing in Elm, such as x = Infinity (instead of x = 0/0).

[pisys]

parseFloat is less strict than the current implementation. It allows whitespaces, exponents, and only demands a prefix of the input string to be a float. Since "." is not a Float parseFloat returns NaN!

However, parseFloat can also parse "Infinity", which is a Number in JavaScript but not in Elm! So I discourage the mere wrapping of parseFloat until there is a explicit Infinity thing in Elm, such as x = Infinity (instead of x = 0/0).

[jvoigtlaender]

How is this pull request related to https://github.com/elm-lang/core/pull/342?

[jvoigtlaender]

How is this pull request related to https://github.com/elm-lang/core/pull/342?

[jvoigtlaender]

I'm closing this for now, since it appears to be overridden by https://github.com/elm-lang/core/pull/342. Please re-open with comment if you disagree.

[jvoigtlaender]

I'm closing this for now, since it appears to be overridden by https://github.com/elm-lang/core/pull/342. Please re-open with comment if you disagree.