Infomation from SyntaxTree

[Iruyan-Zak]

Now, SyntaxTree discards the information of Token kind e.g. Literal. Semantics parsing will use it.

[Roadagain]

i think splitting Leaf into Literal and Identifier.
if it feels good for you, i will fix as this.

[Iruyan-Zak]

I don't like it by the semantic reason, but it is handy.
So, I agree to you.