From: sgk@REMOVEtroutmask.apl.washington.edu   
      
   On Thu, 29 Aug 2024 17:59:04 +0000, Steven G. Kargl wrote:   
      
   > F2023:C8107 states   
   >   
   >    The namelist-group-name shall not be a name accessed by use association.   
   >   
   > Are following pieces of code standard conforming?   
   >   
   > ! Code 1   
   > module mod_nml1   
   >    implicit none   
   >    logical :: ldiag   
   >    namelist /nam_nml1/ldiag   
   > end module mod_nml1   
   >   
   > program ice_nml   
   >    use mod_nml1   
   >    implicit none   
   >    integer :: ilu   
   >    ldiag = .false.   
   >    write(*,nml=nam_nml1)  ! <-- Use assoc. of namelist-group-name   
   > end program ice_nml   
   >   
   > ! Code 2   
   > module mod_nml1   
   >    implicit none   
   >    logical :: ldiag   
   >    namelist /nam_nml1/ldiag   
   > end module mod_nml1   
   >   
   > program ice_nml   
   >    use mod_nml1   
   >    implicit none   
   >    integer :: ilu, j   
   >    namelist /nam_nml1/j   ! <-- Use assoc of namelist-group-name   
   >    ldiag = .false.   
   >    j = 42   
   >    write(*,nml=nam_nml1)  ! <-- Use assoc of namelist-group-name   
   > end program ice_nml   
   >   
   > Clarification of the interpretation of C8107 would be appreciated?   
      
   After a few minutes re-reading parts of Fortran 2023, I have concluded   
   that Code 1 is conforming and Code 2 is non-conforming.  My issue was   
   caused by gfortran accepting Code 2 without and error or warning under   
   -std=gnu option (default behavior).  If -std=f2023 is used, then gfortran   
   issues an expected error messages.   
      
   --   
   steve   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)