gprolog does not have an atom garbage collector. So it is important to control the number of created atoms. If this becomes an issue it is better to switch to strings (which are handled as Prolog list of char codes and thus are recovered at backtracking).
Daniel > Le 3 juil. 2017 à 07:44, David Logan <djlogan2...@gmail.com> a écrit : > > Well, it doesn’t seem to be sub_atom as far as it goes. I wrote my own left() > in order to do the same job, and the new code still fails with atom table > full. At this point, I could use some direction as to why this may be > happening. I do not see any nesting situation that would cause this. It > appears to me that as I backtrack to the repeat clause, it should remove the > atoms that are no longer in use. > > David Logan > > >> On Jul 2, 2017, at 3:07 PM, David Logan <djlogan2...@gmail.com >> <mailto:djlogan2...@gmail.com>> wrote: >> >> I am trying to use sub_atom to get a substring. It fails with “atom table >> full.” When I remove that single line and instead hard code some value to >> “NewPrefix”, it works fine (other than not writing the correct words of >> course.) Is this a bug, or does anyone know why this would be legitimately >> failing? >> >> I am not sure how this will arrive at users inboxes, so let me state that I >> am trying to attach two files: The test prolog file, and the input file I am >> using. >> >> Thank you, >> David Logan >> >> >> <test5.pl> >> <wordlist.txt> > > _______________________________________________ > Users-prolog mailing list > Users-prolog@gnu.org > https://lists.gnu.org/mailman/listinfo/users-prolog
_______________________________________________ Users-prolog mailing list Users-prolog@gnu.org https://lists.gnu.org/mailman/listinfo/users-prolog
