https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89365
--- Comment #3 from Bader at lrz dot de <Bader at lrz dot de> --- I agree with Harald's assessment. The test case as delivered by me is indeed incorrectly written for the POINTER and ALLOCATABLE cases, in both of which I believe the bounds should be taken from the actual argument (versus remapping them). This has historical reasons (ifort has the POINTER and ALLOCATABLE case wrong), and the F2018 drafts did pick up some additional words in the last phase of corrections (see 8.5.8.7 para 1: "The bounds and shape of an assumed-rank entity with the ALLOCATABLE or POINTER attribute are determined as specified in 8.5.8.4." Regards Reinhold
