[antlr-interest] translating a tree to another tree

[Philippe Faes]

Dear all,

I would like to translate an AST, generated with ANTLR to another AST. This
is usefull for having layers of intermediate representations, each new layer
more abstract than the previous.
I tried using a treewalker and setting output=AST, but this gives two
problems.
First, I need to set   ASTLabelType=Token so that the TreeAdaptor can take
Tokens as input. This is not what I want, since the tree being walked
constists of Tree nodes.
Second, the generated code tries to read retval.stop on objects of
TreeRuleReturnScope, but these objects only have a "start" field. The "stop"
field is found in ParserRuleReturnScope, but the tree walker uses
TreeRuleReturnScope instead.

Am I doing something wrong? Is the generation of new trees from old trees
not supported in ANTLR? If not: Why? It seems like a natural thing to try.

kind regards
Philippe
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.antlr.org/pipermail/antlr-interest/attachments/20080210/3edd39a1/attachment.html