[metaas-dev] ASFieldAccessExpression

[Matthew J Tretter]

> Do you have corresponding changes to ExpressionBuilder too?

I don't think I made any.  You already had ASTASFieldAccessExpression (and 
therefore the ExpressionBuilder logic), I believe.. it just didn't have a 
corresponding interface (or any functionality).

> I guess the Eclipse model of allowing
>
> A.B
>
> to be represented as either a field-access or just a qualified-name
> (depending on how much semantic informations is available) is ok, but I
> wonder what happens if modifications to the tree mean that the
> representation should change?

I was thinking about this as I was working on ASFieldExpression.  Wouldn't 
variable bindings first need to be in place for us to do this kind of 
analysis (a la Eclipse's Name.resolveBinding)?  Which would in turn force a 
change in the getName functions (can't return Strings anymore).  I guess you 
could sidestep the issue by inspecting the AST, but it seems like something 
you're going to need anyway (and already could use in AutoImporter).

matthew