------- Comment #4 from irar at il dot ibm dot com 2008-06-26 18:33 ------- (In reply to comment #3) > Can't you simply handle a NULL return > from evolution_part_in_loop_num in the vectorizer?
The problem is that this happens during the transformation (for the code created by the vectorizer). We assume that all the checks were done during the analysis phase, and don't expect anything to go wrong during the transformation. This is why there is an assert that the evolution is not NULL. But, anyway, shouldn't evolution_part_in_loop_num return a valid value for (short int) {(short unsigned int) i_44, +, 1}_1? Ira > Richard. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36630