Using Marpa::R2::NAIF to clean up Context-Free Grammars

[Ruslan Shvedov]

Unproductive and unaccessible rules can be removed from a grammar easily using code like this:

# cleanup the grammar using Marpa NAIF
my $grammar = Marpa::R2::Grammar->new({

start => ...,

rules => ...,

unproductive_ok => 1,

inaccessible_ok => 1,

});

$grammar->precompute();

my $rules = $grammar->show_rules;

my @cleaned;

for my $rule (grep {!/unproductive|inaccessible/} split /\n/m, $rules){

my (undef, $lhs, @rhs) = split /^\d+:\s+|\s+->\s+|\s+/, $rule;

push @cleaned_pfg, [ $lhs, \@rhs ];

}

Working example can be seen here. Just a quick hack, but kind of cool and saved my day. I wonder whether such use is legitimate and, if so, can it be worth a special method, e.g. Marpa::R2::Grammar::cleanup().

[Christopher Layne]

Couldn't you just not use unproductive_ok and inaccessible_ok?

> --
> You received this message because you are subscribed to the Google Groups "marpa parser" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to marpa-parser...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Ruslan Shvedov]

On Sat, Sep 27, 2014 at 12:54 AM, Christopher Layne <cla...@anodized.com> wrote:

Couldn't you just not use unproductive_ok and inaccessible_ok?

Yes I could, and tried that at first, but that issued warnings I hated silencing.