Rethink about tokens' data types

[Roadagain]

@Iruyan-Zak has implemented parsers but token data is returned as string now. i and he talk on slack we should return in another way.

so rethink about tokens' data types.

[Iruyan-Zak]

OK. Then I change the definition of Parser to polymorphic. And then I define a type named Token, which will be the part of the return type of naturalNumber and more.

[Iruyan-Zak]

Probably you think that, some parsers should return a expression. So I'm planning some data types.

• [Expression] a Term or a binary tree of ones whose node is a binary operator which includes multiply operator *.
• [Term] a XXX and a list of unary operators optionally.
• [XXX] an identifier, a literal, or a parenthesized Expression. This means "can become Expression without other tokens."

So, my questions are

1. Does my plan look cool?
2. If so, how should I name XXX?

[Roadagain]

it looks cool and great.

In "Progamming in Haskell", XXX is called as factor.

if it is needed, i will show you published sources in "Programming in Haskell".

[Roadagain]

i've thought '+' should be in prefixOperator cuz tiny should operate "+2" as 2.

i need your opinion.

[Iruyan-Zak]

Mathematically, it is better to add +, but I think it is unnecessary. Have you ever used unary + operator in any programming languages?

[Roadagain]

never... it is really unnecessary lol

[Iruyan-Zak]

In JavaScript, +true has a meaning to cast boolean to int as a kind of hack. But it is better than being ready for hacking to support such needs with honesty way.

[Iruyan-Zak]

Can I create PR?

[Roadagain]

yes, i will merge it.

[Roadagain]

close via dfc3313