How to parse large files with DCG?

Hello,

With DCG in gnu-prolog, do you know a way to directly parse a file from 
a stream instead of first processing the file into a list of tokens?

I know the trick of redefining the 'C'/3 to get a token from a stream, 
as explained in the Eclipse manual for example, but gnu-prolog does not 
seem to offer this predicate to the user.

Do you have any idea or trick to share?

Thanks a lot in advance,
Best regards,
Erick.