To: J3                                                     J3/22-118r1
From: Bill Long
Subject: Reply to J3/22-118
Date: 2022-March-04



Discussion:
-----------

Paper 22-118 proposes to change the interpretation of the SHAPE
argument of C_F_POINTER in the ISO_C_BINDING module to make the
numerical entries in the SHAPE argument be interpreted as the upper
bounds for the corresponding dimensions of the FPTR pointer instead of
its shape. Thus, a call to C_F_POINTER of the form

  CALL C_F_POINTER (CPTR, FPTR, SHAPE=[3,3], LOWER=[-1,0])

would result in a Fortran pointer FPTR that has an actual shape of
[5,4], inconsistent with the current standard and programmer
expectations. As noted in paper 22-118, a previous proposal to
explicitly specify upper bounds (instead of the shape) in C_F_POINTER
was not accepted.

Irrespective of the merits of the proposal, now that the CD for F202x
has been completed, the window for new features is closed. The
proposal in 22-118 is judged to be a sufficiently significant change
to constitute a new feature. Therefore, subgroup recommends No Action
for this paper at m226 of PL22.3.  Feature proposals deemed as too
late for F202x can be submitted for consideration in a future
revision.