On 25 December 2017 20:45:40 CET, Steve Kargl <s...@troutmask.apl.washington.edu> wrote: >The attach patch fixes a problem when a LOGICAL subprogram >appears as the first element in an array constructor, which >is interpreted as a LOGICAL type-spec, which fails because >the argument is of type LOGICAL instead of INTEGER. > >Regression tested on i686-*-freebsd and x86_64-*-freebsd. > >OK to commit?
+ p = [real(kind=4) :: x, y] + q = [real(kind=8) :: x, y] + if (any(p .ne. r2)) call abort + if (any(q .ne. r3)) call aborts +end program foo aborts? abort? Thanks,