[antlr-interest] Re: SQL grammar tree construction problem
Thomas Brandon
tom at psy.unsw.edu.au
Thu Nov 13 16:37:26 PST 2003
Sorry, that was poorly stated, yes a tree must have a root, but it
need not have a unique root, it can have a root with siblings not
just children, but JTree doesn't like this.
So ROOT SIB1 SIB2 is a valid tree (i.e. ROOT with siblings SIB1 and
SIB2. You don't need to have #(ROOT SIB1 SIB2) (i.e. a unique root
with 2 children). But JTree only renders the first top-level node,
i.e. for #(ROOT CHILD1 CHILD2) SIB1 SIB2 (i.e. root with two
children and two siblings) it will only show #(ROOT CHILD1 CHILD2),
SIB1 and SIB2 are not shown.
Tom.
--- In antlr-interest at yahoogroups.com, "micheal_jor"
<open.zone at v...> wrote:
> --- In antlr-interest at yahoogroups.com, "Thomas Brandon" <tom at p...>
wrote:
> > Not sure that this is the problem but...
> > I don't think Antlr requires rooted trees.
>
> As Ter pointed out in a different thread a coupla days ago, ANTLR
> doesn't generate concrete syntax tree (CSTs) by default.
>
> AFAICT, this ensures ANTLR doesn't generate a rooted subtree for
each
> rule by default. It is up to the grammar developer to specify the
> exact structure of the generated AST. Of course a tree has to have
a root.
>
> Cheers,
>
> Micheal
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
More information about the antlr-interest
mailing list