One common gotcha of PEG grammars is having two alternatives in a rule in which one that appears first is contained in one that appears last. In this scenario, the second one will never match as if there is input text that matches it, it will also match the first and that comes first.

To avoid this problem, this PR create an initial version of a small utility that detects these cases and raises to alert the user.

We don't need to detect all cases, only the ones that is easy enough to implement. The current form is just an initial prototype that does the matching based in the string representation. To make this better we need to improve the matching algorithm into a visitor to allow checking rules that contain options. For example:

some_rule:
    | NAME '+' NAME 'blech'?
    | NAME '+' NAME ';'

One "straightforward enough" way to do this is to generate all possible string representations: with and without optional and doing substring matching, but at this point a visitor that visits both alternatives at the same time and advances accordingly is probably better.

We could also support some rule expansion so we also detect something like this:

some_rule:
    | NAME '+' NAME
    | NAME some_other_rule
some_other_rule:
   '+' NAME ';'