https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88685
kargl at gcc dot gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P3 |P4
Status|UNCONFIRMED |NEW
Last reconfirmed| |2019-01-04
CC| |kargl at gcc dot gnu.org
Known to work| |7.4.1
Ever confirmed|0 |1
Known to fail| |8.2.1, 9.0
--- Comment #1 from kargl at gcc dot gnu.org ---
Interesting bug. Changing doit() to
subroutine doit(cls)
class(tarr), pointer :: cls(:,:)
cls(1,1)%cl = 1
print *, 'cl = ', cls(1,1)%cl
print *, 'cl(i)= ', cls(1,1)%cl(1:3)
print *, 'cl(i)= ', cls(1,1)%cl(1:1),cls(1,1)%cl(2:2),cls(1,1)%cl(3:3)
print *, 'cl(i)= ', cls(1,1)%cl(1),cls(1,1)%cl(2),cls(1,1)%cl(3)
end subroutine doit
shows that the issue involves the reference individual elements. With
gfortran 7.4.1, get
% gfc7 -o z a.f90 && ./z
cl = 1 1 1
cl(i)= 1 1 1
cl(i)= 1 1 1
cl(i)= 1 1 1
with 8.2.1 and 9.0 I get
% gfc8 -o z a.f90 && ./z
cl = 1 1 1
cl(i)= 1 1 1
cl(i)= 1 1 1
cl(i)= 1 -1 -1
