(j3.2006) (SC22WG5.4569) WG5 letter ballot 2 on Fortran 2008interpretations

Malcolm Cohen malcolm
Tue Oct 18 23:51:42 EDT 2011

Reinhold Bader further writes

F08/0047  The discussion of the interp (which was rather shortened down
          compared to the original paper) says that
          "The PUBLIC statement at A contradicts the PRIVATE attribute at
           B which violates the standard."

          If this refers to the constraint C514 I conclude that my vote
          must be "NO", since we're talking about two entities here, one
          being the type definition and the other one being the generic.

          It appears that the situation that use-name and generic-spec
          in R525 can be the same has not been adequately addressed.

You can vote no, but the answer is still that you have specified both PUBLIC FOO
and PRIVATE FOO and that is not allowed.  FOO the identifier is a single entity
no matter how many other entities can be accessed by FOO.

Don't just stop reading when you reach the end of C514, keep going and read
5.3.2p2 - it is explicit stated that accessibility applies to identifiers.

We have had identifiers being usable to access multiple entities since F90, this
is not a new issue and is completely covered by the existing text in the

................................Malcolm Cohen, Nihon NAG, Tokyo.

................................Malcolm Cohen, Nihon NAG, Tokyo. 

More information about the J3 mailing list