Re: DCG restrictions on the left-hand side [Novacore]

I think its worth adopting this behaviour:
/* Scryer Prolog */
?- a-->b.
    error(existence_error(procedure,(-->)/2),(-->)/2).
?- assertz((a-->b)).
    error(permission_error(modify,static_procedure,(-->)/2),assertz/1).
?- a-->b.
    error(existence_error(procedure,(-->)/2),(-->)/2).
Mild Shock schrieb:

Although SWI-Prolog implements a lot of checks.
It still allows this one:
 ?- a-->b.
ERROR: Unknown procedure: (-->)/2
 ?- assertz((a-->b)).
true.
 ?- a-->b.
true.
 Scryer Prolog managed to prevent this, in that
the assertz/1 throws an error.
 Think will finally adopt this behaviour to
prevent the end-user from accidentially
 fiddling with (-->)/2.
 Mild Shock schrieb:

>
This was discussed here:
>
Definite Clause Grammars Harmonization Proposal
Last edit jschimpf? October 20, 2016, at 12:42 AM
https://eclipseclp.org/wiki/Prolog/DCG
>
Not sure why Scryer Prolog doesn't implement it.
>
Mild Shock schrieb:

>
In Novacore we are lucky, its a side effect of
phrase translating the head and a multi-file check:
>
?- [user].
p;q --> r.
Fehler: Kann Prädikat ;/2 nicht umdefinieren, nicht als Mehrdateien markiert.
>
But might work out differently if a DCG is expanded
and asserted. Ok, its actually not so bad:
>
?- expand_term((p;q --> r), X), assertz(X).
Fehler: Kann Prädikat ;/2 nicht aktualisieren.
     user auf 1
>
>
Mild Shock schrieb:

Guess how many newer Prolog systems allow this here:
>
?- [user].
p;q --> r.
    true.
>
On the other hand SWI-Prolog disallows it:
>
?- [user].
p;q --> r.
>
ERROR: user://1:29:
ERROR:    No permission to define dcg_nonterminal `p;q'

>

>