------- Comment #3 from jvdelisle at gcc dot gnu dot org 2005-10-30 07:07 ------- The patch in the original post is wrong. The patch in comment #2 is correct and is very similar to the patch done for pr16080 for which the test case read_null_string.f90 was made.
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24584
