J3/17-248
To: J3
From: Malcolm Cohen
Subject: Comment 22
Date: 2017 October 17


1. Introduction

Comment 22 notes that
    coarray[i]%allocatable%allocpoly
is not allowed (except possibly to intrinsic inquiries etc.)
but that
    coarray[i]%pointer%pointerpoly
is allowed.  The contention is that this should not be allowed.

Subgroup concurs.  However, the proposed fix in the comment would
also disallow the "good case" of alloc_poly_coarray[i], which is
perfectly cromulent because it has the same dynamic type on every
image on which it is established.

Therefore slightly more extensive surgery is required.


2. Edits to 10-007r2 (N2137)

[133:14] 9.4.2 Structure components, C917, delete
         "shall not be a polymorphic subobject of a coindexed object and",
         making the whole constraint read
   "C917 (R911) Except as an actual argument to an intrinsic inquiry
         function or as the designator in a type parameter inquiry, a
         data-ref shall not be a coindexed object that has a polymorphic
         allocatable potential subobject component."
{Delete problematic part of constraint.}

[133:15+] After C917, insert new constraint
   "C917a Except as an actual argument to an intrinsic inquiry function or
          as the designator in a type parameter inquiry, if the rightmost
          part-ref is polymorphic, no other part-ref shall be coindexed."
{Prohibit the problematic cases of polymorphic coindexed objects.}

===END===