[antlr-interest] More DOT Questions

Randall R Schulz rschulz at sonic.net
Fri Jun 29 15:23:40 PDT 2007


On Friday 29 June 2007 15:07, Terence Parr wrote:
> On Jun 29, 2007, at 3:04 PM, Randall R Schulz wrote:
> > On Friday 29 June 2007 14:50, Terence Parr wrote:
> >> On Jun 29, 2007, at 2:43 PM, Randall R Schulz wrote:
> >>> ...
> >>>
> >>> What I want to know is what dot configuration and other
> >>> processing or transformation ANTLRWorks clearly does behind the
> >>> scenes to get the results it does.
> >>
> >> I'm pretty sure it calls ANTLR's DOTGenerator.  Just check the
> >> source I'd say.
> >
> > Clearly that's not so, since when I take the output of DOTGenerator
> > and
> > pass it to dot, the results are very different than those produced
> > within ANTLRWorks itself.
>
> For DFA?  Just the squiggling lines.  All other viz stuff is done
> directly by Jean in AW.

That's too bad. DOT can't do what's required?

Anyway, in the long run (after the grammar is developed, stable and 
being used) I'm more interested in visualizing ASTs. I really liked 
your example on page 191 of TDAR and included it in the command-line 
driver I wrote for the parser I'm working on now.

But as I mentioned, the result it produces (when you render the 
resulting DOT file) is quite different from that produced when 
visualizing the AST in the ANTLRWorks debugger. I'd like to be able to 
replicate the latter.

Of course, I can play around with DOT, but I'd assumed that the graph 
visualization in ANTLRWorks was being done by dot—I figured that was 
why there was a configuration preference for the location of the dot 
executable.


> ...
>
> If I remember correctly, it's converting my DOT to DOT it can parse
> to get the splines out and display in Swing.

I see...


> Ter


Randall Schulz


More information about the antlr-interest mailing list