[antlr-interest] A few suggestions re. ANTLRWorks

Mark Volkmann r.mark.volkmann at gmail.com
Tue Feb 5 05:00:12 PST 2008


On Feb 4, 2008 10:45 PM, Jim Idle <jimi at temporal-wave.com> wrote:
>
> > -----Original Message-----
> > From: Olivier Lefevre [mailto:lefevrol at yahoo.com]
> > Sent: Monday, February 04, 2008 4:48 PM
> > To: antlr-interest at antlr.org
> > Subject: [antlr-interest] A few suggestions re. ANTLRWorks
> >
> > Next comes the Input Text dialog. The Start Rule was set to a
> > (seemingly randomly chosen) rule in the grammar describing a single
> > line, so that the debugger terminated after the first line of input.
> > This drove me _nuts_ and it took me a while to realize (by chance)
> what
> > was going on.
> > It seems to me that the only sensible default for the Start Rule is
> the
> > topmost rule of the grammar, whatever that is, and that it should be
> > set to _that_ the first time.
>
> There is no real reason that the first rule is likely to be more correct
> as a target than any other. Choose your rule and it remembers your
> choice. There are many grammars where the 'start' rule depends on what
> you are testing. It might be more likely than not that the first rule
> listed is the start rule, but more likely that a rule ending with the
> EOF token is the start rule.

I agree that the best first guess is to find the rule that ends in
EOF. However, if none do then guessing that the first rule in the
grammar is the start rule seems much better than the current,
seemingly random guess. Is it random? I haven't been able to detect a
pattern.

-- 
R. Mark Volkmann
Object Computing, Inc.


More information about the antlr-interest mailing list