J3/16-116r1
To:         J3 Members
From:       Bill Long & Malcolm Cohen
Subject:    Clause 7 edits
Date: 2016 February 08
References: 16-007

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

In 7.1.4 Evaluation of operations, "<forall-assignment-stmt>" should
be in the obsolescent font. Forall is obsolescent in F2015.

In the syntax rule for "<primary>", R701, we switched, between F95 and
F2003 from "variable" to "designator". This change seems to have ben
made in other, related places, except in the first sentence of 7.1.9.2
Type, type parameters, and shape of a primary, where we repeat the
list of options from R701, except for using "variable" in the place of
"designator". This seems inconsistent.

Also, <designator> includes named constants, which are already included in
<primary> via <constant>.  The BNF <constant> is defined as being
<named-constant> (already included in <designator> via <object-name>) or
<literal-constant>.  Thus changing "<constant>" to "<literal-constant>"
in "<primary>" (R701) will avoid the ambiguous parse without omitting
anything.


Edits to 16-007:
-----

[137:17] 7.1.2.2 Primary, R701,
         "<constant>" -> "<literal-constant>".
{Avoid ambiguous parse.}

[143:5] 7.1.4 Evaluation of operands, para 2,
  put "<forall-assignment-stmt>" in obsolescent font.

[152:24-28] 7.1.9.2 Type, type parameters, and shape of a primary, para 1,
  change "variable" to "designator" thrice
   (once in the first sentence, and twice in the sentence beginning
    "If it is a variable..."),
  and change "constant" to "literal constant" thrice.