Vlad Dumitrescu <>
Fri Jun 2 14:53:09 CEST 2006

On 6/2/06, Richard Carlsson <> wrote:
> I suggest that you use the functions
> that I mentioned to map a tree to a simple Canonical Form:
> something like {NodeType, [Attributes], [SubTrees]}. This
> translation would make your problems with handling the old
> erl_parse representation simply go away, and you are still
> free to add your annotations etc. Then send this canonical
> form as a term to the Java side.

I thought the tree node representation was {tree, NodeType,
[Attributes], [SubTrees]} and thus didn't understand your
suggestion... but after checking again, I suppose I forgot everything
about the current hairy representation [*] right after implementing it

[*] the representation being {tree, #attributes,

Or since we already have to hack the parser (in order to gather the
required annotation information), we can build this simplified
representation directly...

Thanks again!

More information about the erlang-questions mailing list