To: J3                                                       J3/12-149
Subject: Interpretation concerning L-value functions
From: Van Snyder
Date: 2012 May 16

----------------------------------------------------------------------

NUMBER: TBD
TITLE: L-value functions in assignment
KEYWORDS: L-value functions, pointer, assignment, defined operators
DEFECT TYPE: Erratum
STATUS: J3 consideration in progress

QUESTION:

Consider the following

  module Funcs

    interface operator ( .op. )
      module procedure Unary, Binary
    end interface

  contains

    function Unary ( Arg )
      integer, intent(in) :: Arg
      integer, pointer :: Unary
    ...
    end function Unary

    function Binary ( Arg1, Arg2 )
      integer, intent(in) :: Arg1, Arg2
      integer, pointer :: Binary
    ...
    end function Binary

  end module Funcs

  program What

    use Funcs
    integer :: X, Y = 42

    10 .op. x = y

  end program What

Is the "10" in "10 .op. x = y" an operand, and therefore an argument of
a reference to Binary, or is .op. a reference to Unary and "10" a
statement label?

ANSWER:

It was an oversight that this ambiguous program conforms to the syntax
and constraints in two different ways.  Since the standard does not
establish an unique interpretation, it is not a conforming program.

EDITS:

none.

OR

Insert a constraint after C713 in subclause 7.2.1.1:

C713a (R732) In free form or in a statement that begins after a
      semicolon, if <variable> is <expr> and <expr> can be interpreted
      as both <int-literal-constant> <defined-operator> <expr> and
      <defined-operator> <expr>, in the case it can be interpreted as
      <int-literal-constant> <defined-operator> <expr> the statement
      shall have a label, and in the case it can be interpreted as
      <defined-operator> <expr> the statement shall not have a label.

OR

Insert a constraint after C713 in subclause 7.2.1.1:

C713a (R732) In free form or in a statement that begins after a
      semicolon, if the statement does not have a label, <variable> is
      <expr>, and <expr> is of the form <int-literal-constant>
      <defined-operator> <expr>, it shall not also be possible to
      interpret <expr> as <defined-operator> <expr>.

OR

Insert a constraint after C713 in subclause 7.2.1.1:

C713a (R732) In free form or in a statement that begins after a
      semicolon, if <variable> is <expr> and <expr> is of the form
      <int-literal-constant> <defined-operator> <expr> it shall not also
      be possible to interpret <expr> as <defined-operator> <expr>.

OR

Insert a constraint after C713 in subclause 7.2.1.1:

C713a (R732) In free form or in a statement that begins after a
      semicolon, if <variable> is <expr>, <expr> shall not be either
      <int-literal-constant> <defined-operator> <expr> or
      <defined-operator> <expr>.

No matter what choice for the above edits, add at the end of NOTE 7.35:

    "(10) .OP. X = Y

  an invalid assignment statement is

    10 .OP. X = Y

  because if there are both unary and binary interfaces for .OP. it is
  ambiguous whether 10 is a statement label."

ALTERNATIVE ANSWER:

It was an oversight that this ambiguous program conforms to the syntax
and constraints in two different ways.  Since the standard does not
establish an unique interpretation, it is not a conforming program.

That this program is ambiguous was caused by an oversight to allow a
<variable> to consist of <expr>, in which <expr> is a reference, as a
defined operation, to a function that has a data pointer result.

ALTERNATIVE EDITS:

Replace R602 and C602 in subclause 6.2 by

R602 <variable> <<is>> <designator>
                <<or>> <function-reference>

C602 (R602) <function-reference> shall be a reference to a function that
     has a data pointer result.

{Notice that this also precludes using a function referenced by way of a
defined operation in otherwise harmless contexts, such as an actual
argument that corresponds to a dummy argument that is not INTENT(in).}

SUBMITTED BY: Van Snyder

HISTORY: 12-149    m198  F08/xxxx submitted

----------------------------------------------------------------------