On Thu, Jul 09, 2009 at 09:28:14PM -0700, Chip Salzenberg wrote: > I favor simply making it a compile-time error to use split in scalar > context. No problem with silently changing semantics and no more surprising > behavior. Everybody's happy. (right?) ... and, of course, a runtime error if the split is in undetermined context at compile time and in scalar context at runtime. -- Chip SalzenbergThread Previous | Thread Next