Difference in parsing x.[y] to Expr

[c42f]

Likely due to #325 we have the following difference:

julia> dump(parsestmt(Expr, "x.[y]"))
Expr
  head: Symbol .
  args: Array{Any}((2,))
    1: Symbol x
    2: QuoteNode
      value: Expr
        head: Symbol vect
        args: Array{Any}((1,))
          1: Symbol y

julia> dump(JuliaSyntax.fl_parse(Expr, "x.[y]"))
Expr
  head: Symbol .
  args: Array{Any}((2,))
    1: Symbol x
    2: Expr
      head: Symbol quote
      args: Array{Any}((1,))
        1: Expr
          head: Symbol vect
          args: Array{Any}((1,))
            1: Symbol y

[c42f]

Ok, so this is related to the logic at

JuliaSyntax.jl/src/expr.jl

Lines 287 to 289 in acb609d

	if !@isexpr(a2, :quote) && !(a2 isa QuoteNode)
	args[2] = QuoteNode(a2)
	end

I think the subtle point here is that the right hand side of . should always be inert: that is, QuoteNode rather than Expr(:quote) in the AST. The flisp parser has a special case for this in parsing of x.$y and emits inert in that case. But that's technically incorrect and breaks expansion of quote expressions like the following

julia> y = :a
:a

julia> quote
           x.$y
       end
quote
    #= REPL[33]:2 =#
    x.a
end

julia> quote
           x.[$y]
       end
quote
    #= REPL[34]:2 =#
    x.:([$(Expr(:$, :y))])
end

See? In that second case the $y doesn't get expanded.

However! In current JuliaSyntax main branch we have the following which is correct

julia> JuliaSyntax.enable_in_core!(freeze_world_age=false)

julia> quote
           x.[$y]
       end
quote
    #= REPL[36]:2 =#
    x.:([a])
end

So while this is a regression, the new behavior is also technically correct, and a bugfix. So we need to decide what to do.

[c42f]

This syntax was found in the General registry in Accessors.jl and Setfield.jl (the predecessor package of accessors). But in these cases the syntax was only used in a test case for macro error handling - so nobody should be actually using this syntax with those packages.

See jw3126/Setfield.jl#83