Fix #3. Use { and } for explicit grouping

[stasm]

This only changes the EBNF grammar and doesn't change the ASDL. The grouping is supposed to help the parser in potentially ambiguous situations. Once the parser figures out the parse tree, the final AST will be unambiguous.

[zbraniecki]

why placeable?

key = { { $foo } ->
  [one] Value
}

is that intended to work?

[stasm]

Yes. It's not what the serializer would normally serialize to, but I don't think we need to artificially limit the syntax here. It doesn't change the meaning compared to just using $foo, just like (1) === 1 in JavaScript.

[zbraniecki]

So, what's the use case?

[stasm]

The use case is tightly related to #4. In particular, we need a way to remove the ambiguity from call-expressions with multiple arguments, some of them being select-expressions:

key = { LIST(foo, { bar, baz -> 
        [bar, baz] Bar and Baz
    }) }

[zbraniecki]

r+

[stasm]

Thanks for the review. After talking to @Pike last Friday I partly changed my mind about ASDL. The placeable type shouldn't be part of the ASDL, still, but I think it makes sense to separate the following two concepts:

• expressions which can be selectors in a select-expression,
• expressions which can be inside of placeables.

I updated the PR to reflect that. It should map closely to the data structures we'll have in the Rust implementation. @zbraniecki, this probably doesn't need another review from you, but I wanted to let you know before I merge.