From <@iec.co.uk:malcolm@num-alg-grp.co.uk> Thu Jan 25 11:35:27 1996 Received: from sun2.nsfnet-relay.ac.uk (sun2.nsfnet-relay.ac.uk [128.86.8.45]) by dkuug.dk (8.6.12/8.6.12) with SMTP id LAA20170; Thu, 25 Jan 1996 11:35:13 +0100 Via: uk.co.iec; Thu, 25 Jan 1996 10:33:31 +0000 Received: from mars.nag.co.uk by nags2.nag.co.uk (4.1/UK-2.1) id AA04831; Tue, 23 Jan 96 12:16:56 GMT From: Malcolm Cohen Message-Id: <3099.199601231201@mars.nag.co.uk> Received: by mars.nag.co.uk (5.65c/UK-2.1) id AA03099; Tue, 23 Jan 1996 12:01:54 GMT Subject: Re: (SC22WG5.1006) Suggested edit to X3J3/95-007R2 To: ErikW.Kruyt@dkuug.dk (+31 71 5276804) Date: Tue, 23 Jan 96 12:01:47 MET Cc: SC22WG5@dkuug.dk In-Reply-To: <199601230926.KAA11451@dkuug.dk>; from "+31 71 5276804" at Jan 23, 96 9:00 am > page 60, line 7 and line 18 [ALLOCATABLE and POINTER statements] > "the array-spec shall be a deferred-shape-spec-list" > should read: > "the deferred-shape-spec-list shall be present" > > Argument: we want to exclude the sole array name, which is not an array-spec > (R513). No, we do not want to exclude the sole array name. Because DIMENSION X(:) ALLOCATABLE X is perfectly legal Fortran. An ALLOCATABLE array must be an array, which means that it has its arrayness specified either in an in a DIMENSION attribute or DIMENSION statement, or a in an ALLOCATABLE statement. So there is no problem. Also, scalar POINTERs are allowed, and it would be quite inappropriate for them to have a attached (because they would then be arrays!). Cheers, -- ...........................Malcolm Cohen, NAG Ltd., Oxford, U.K. (malcolm@nag.co.uk)