Re: Empty rule spoiling error reporting

[Tim Van Holder]

hz kto wrote:
> > > My grammar contains a lot of optional rules like this
> > >
> > > rule1 : token3 {...}
> > >         | token2 {...}
> > >         | opt_something1 token1 {...}
> > >
> > >
> > > When in certain context parser does not see token1,2,3 nor anything
> > > non-empty from opt_something1, it matches an empty rule and then, when it
> > > starts searching for tokens that could follow next, it only finds token1
> > > since it already in the state corresponding to opt_something1.
>
> I just want to report more possible tokens. As it stands now I report only one 
> instead of several in these cases.

Well then I'm afraid you'll have to refactor your rules:

rule1
: token1            { ... }
| something1 token1 { ... }
| token2            { ... }
| token3            { ... }
;

Now the parser should be able to list more possible tokens
if a bad token is seen at the start of rule1.

Note that this requires a lot of work if you have all your opt_
rules in the form:

opt_foo
: /* empty */
| foo_format_1
| foo_format_2
| foo_format_3
;

I strongly recommend using only trivial opt_ rules from the get go:

opt_foo
: /* empty */
| foo
;

foo
: foo_format_1
| foo_format_2
| foo_format_3
;

This uses more rules but makes it much easier to work with.  Similarly
for mul_ an plus_ rules (to handle */+ repetition):

mul_foo
: /* empty */
| plus_foo
;

plus_foo
: foo
| plus_foo foo
;