Fixed incorrect representation of array index parsing

[Lut99]

This PR fixes the incorrect grammar rule that array indices can only follow array literals. Instead, it is modelled as a special kind of unary operator which follows its operand.

Fixes #29

[DanielVoogsgerd]

I am missing precedence rules for unary operators, I think. Assuming left to right as a base rule. This would mean that !array[0] would fail for example

[Lut99]

I’d have to double-check, but I’m quite sure that, due an implementation quirk of the pratt parser we have, it actually matches any prefixed unary expressions first, and then it does binary & postfix unary. Which should result in the correct binding, in your example. But please double-check by finding out 😇

Not sure how to encode that in the grammar, though. Aren’t they included in the association & precedence tables?