If source code has certain accented letters [ 'ê', 'ô'], Gnu-Prolog (MinGW
distribution, gcc 3.4.5) fails to compile once in every five trials. It is
interesting that the compiler sometimes works fine, but sometimes claims that
it cannot unify_atom(Ý); as one can see, this atom is not among the accented
letters that I am trying to recognize. Work around: One can keep trying; the
compiler will accept the source code eventually; once compiled, the program
works perfectly well. Another work-around is avoidance of characters (use codes
instead).
% Culprit:
accents(['á', 'é', 'í','ú', 'â', 'ê', 'ô']).
%Error message:
/*
D:\nogui>gplc --no-top-level --min-size tokenizer
.pl -o tokenizer
line 195: atom expected
unify_atom(Ý),
Ý^ here
Translation aborted
compilation failed
*/
accented(W) :- accents(As), run_member(W, As).
run_member([X|_], As) :- member(X, As), !.
run_member([_|Xs], As) :- run_member(Xs, As).
alpha('-') :- !.
alpha(X) :- X @>= 'a', X @=< 'z', !.
alpha(X) :- accents(As), member(X, As), !.
check(X) :- alpha(X), !, write('Yes!'), nl.
check(X) :- write('No!'), nl.
isalphap :-
argument_counter(ACount), ACount=2,
argument_value(1, Symbol),
check(Symbol).
isalphap :- write('Usage: isalphap <char>'), nl.
:- initialization(isalphap).
_______________________________________________
Bug-prolog mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-prolog
