short description:
What steps will reproduce the problem?
1. use grammar and CLLRS sem/ directory in attached files
2. srec([anybody]) shows intended output in ASCII
3. rec([anybody]) fails with GraleJ saying
Exception: java.lang.ArrayIndexOutOfBoundsException: 2
-----------------------------------------------------------------
more elaborate description:
GraleJ currently only supports Gerald and Michael's preliminary
implementation hack for positive contribution constraints. It fails
on the parallel negative contribution constraints. The difference
is minimal:
1) positive contribution (on basis of artificial lex entry 'alks'
(sic!), just to illustrate the point):
srec([alks]) output:
^lambda(b,A:[exists(a,B:[{walk(b,a,C)}])])/([*])
This is displayed nicely in GraleJ. Lists of [*] are also handled
nicely by GraleJ, as I saw in other examples that I tried for
Michael.
What GraleJ can't deal with is the counterpart,
2) negative contribution constraints, again illustrated with a
lexical entry:
srec([anybody]) output:
neg(A:[^exists(a,B:[{person(b,a)}],C:[a])])/~([*])
GraleJ fails to disply this, saying:
Exception: java.lang.ArrayIndexOutOfBoundsException: 2
Could be that you didn't extend the preliminary star syntax to
this negative case.
Original issue reported on code.google.com by frankmri...@gmail.com on 26 Apr 2010 at 5:04
Original issue reported on code.google.com by
frankmri...@gmail.comon 26 Apr 2010 at 5:04Attachments: