Feature request ensure_loaded/1 [ISO 7.4.2.8] about gprolog HOT 4 OPEN

ensure_loaded/1 is not yet implemented (it has some issues for a native compiler). gprolog proposes the ensure_linked/1 directive which could help you.

from gprolog.

Currently the work around is to use include/1. But include/1
is not smart, it cannot deal with an import graph. The main
benefit of ensure_loaded/1 would be to avoid duplicate

include/1, which is not part of the requirement of include/1.
So one workaround would be if one could define user
directives, like:

:- dynamic(visited/1).

ensure_loaded(Path) :- 
     visited(Path), !.
ensure_loaded(Path) ;-
     assertz(visited(Path)),
     include(Path).

from gprolog.

ensure_loaded/1 is not yet implemented (it has some issues for a native compiler). gprolog proposes the ensure_linked/1 directive which could help you.

As a side note, Logtalk provides the functionality of the ensure_loaded/1 directive. See:
https://logtalk.org/manuals/userman/programming.html?highlight=reload#flag-reload

from gprolog.

Logtalk is not ISO core standard compliant in the way it deals
with Prolog texts. The minimum Logtalk would need to provide would
be include/1 and ensure_loaded/1. But it begs the question whether

any useful semantic can be made, Logtalk compilation units have
a name braket. On the other hand the ISO core standard indeed specifies
ensure_loaded/1 based on include/1. It mentions include explicitly:

But there is a small gap, the standard doesn't elaborate when
Prolog text designators denote the same Prolog text. So I have
a new proposal, assuming a predicate absolute_file_name/2 that

then is responsible for the sameness:

ensure_loaded(Path) :- 
    absolute_file_name(Path, AbsPath),
    \+ visited(AbsPath), !,
    assertz(visited(AbsPath)),
    include(Path).
ensure_loaded(_).

But a challenge can be to have visited/1 automatically flushed
sometimes, so as to regain re-consult semantics.

from gprolog.