To: J3 J3/18-107 From: Bill Long Subject: Dummy arguments of assumed type and size Date: 2018 January 31 Discussion ---------- Paper 13-283, passed at J3 meeting 201, somehow failed to get entered into the F2018 draft. This paper is aimed at repairing that oversight and differs only in that the edit instructions have been updated to refer to N2146. The Introduction/Discussion is unchanged from 13-283 and is not reproduced here. Edits to N2146 -------------- [Page 316] 15.5.2.4 Ordinary Dummy Variables, paragraph 13 rewrite: - split the first sentence into a separate paragraph - replace the rest with a new paragraph "If the actual argument is a noncoindexed scalar, the corresponding dummy argument shall be scalar unless \begin{itemise} \item the actual argument is default character, of type character with the C character kind (15.2.2), or is an element or substring of an element of an array that is not an assumed-shape, pointer, or polymorphic array, \item the dummy argument has assumed rank, or \item the dummy argument is an assumed-type assumed-size array. \end{itemise} {Allow scalar to associate with an assumptive dummy.} [Page 320] 15.5.2.11 Sequence association, paragraph 1, first sentence: after "represents an element sequence if" change "it" to "the dummy argument is not assumed-type and the actual argument". {If the dummy argument is assumed-type, we don't have traditional sequence association - especially in the case of character, otherwise we get a divide by NaN trying to calculate the number of elements.} [Page 320] 15.5.2.11 Sequence association, following paragraph 4, insert a new paragraph: "If an assumed-type dummy argument is an assumed-size array and the actual argument is scalar, the dummy array has exactly one element. Otherwise, an assumed-type assumed-size dummy array has the same number of elements as the actual argument." {New thingo semantics in "sequence" association.}