[antlr-interest] "flat" trees

[Gerald B. Rosenberg]

At 12:37 PM 3/13/2008, Richard Clark wrote:

>I recently wrote a translator that used ASTs pretty extensively,
>including a custom map/reduce-style library to do transformations. It
>seems to me that the AST trees were perfectly sufficient -- there are
>nodes, nodes have multiple ordered children, you can do
>search/insert/remove operations on nodes, etc.

Did you build anything in the way of a generalized library to do the 
AST access?  How do you handle breaking AST structure changes?

In any case, would be very interested in seeing at least the 
interface API/method signatures that you settled on for your access operations.

Thanks,
Gerald