https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110993
Bug ID: 110993
Summary: Possibly bogus diagnostic on renamed interface import
Product: gcc
Version: 14.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: rimvydas.jas at gmail dot com
Target Milestone: ---
Spotted while checking why someone was manually mangling names with bind(c).
$ cat foo.f90
module intfb_pack
interface
! subroutine bar(x) bind(c, name="bar_") ! does not warn
! subroutine bar(x) bind(c, name="bar") ! Error: More actual than formal
arguments in procedure call
subroutine bar(x) ! Warning: Shape mismatch in argument 'x'
use iso_c_binding, only : c_float
implicit none
real(c_float) :: x(45)
end subroutine
end interface
end module
subroutine foo(x)
use intfb_pack, notthisone => bar
implicit none
real :: x(3)
call bar(x)
end subroutine
$ gfortran -c foo.f90
foo.f90:17:13:
14 | use intfb_pack, notthisone => bar
| 2
......
17 | call bar(x)
| 1
Warning: Shape mismatch in argument 'x' between (1) and (2)
The diagnostic is a bit surprising given that selective import with "use
intfb_pack, only:" does not emit diagnostics. Moreover, the "bind(c,
name="bar")" case hard error seems very wrong.
