From j.l.schonfelder@liverpool.ac.uk Fri Jul 4 15:28:43 1997 Received: from pcmail.liv.ac.uk (pp@pcmail.liv.ac.uk [138.253.252.13]) by dkuug.dk (8.6.12/8.6.12) with SMTP id PAA15339 for ; Fri, 4 Jul 1997 15:28:42 +0200 Received: from jlspcnt.liv.ac.uk [138.253.102.118] by pcmail.liv.ac.uk with smtp (Exim 1.61 #3) id 0wk8Pf-0006Ea-00; Fri, 4 Jul 1997 14:28:39 +0100 From: Lawrie Schonfelder Sender: J.L.Schonfelder@liverpool.ac.uk Reply-To: j.l.schonfelder@liverpool.ac.uk To: sc22wg5@dkuug.dk Subject: Re: (SC22WG5.1407) Varying string standard In-Reply-To: <199707040503.HAA09554@dkuug.dk> Message-ID: Date: Fri, 4 Jul 1997 14:28:37 +0100 (British Summer Time) Priority: NORMAL X-Mailer: Simeon for Win32 Version 4.1.1 Build (17) X-Authentication: IMSP MIME-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII On Fri, 4 Jul 97 14:07:47 +0900 Minoru Tanaka wrote: > My colleague tried compilation of the module using our Fortran 95 compiler > alpha version for Win95/NT (currently its name is Fortran 90 compiler yet :-)). > At that time, we had the following compilation errors: > > Fortran 90 diagnostic messages: program name(ISO_VARYING_STRING) > Module subprogram name(s_to_c) > jwd2063i-s "iso_vst.f90", line 308: The dummy argument of the elemental procedure shall not appear in a > specification expression.(name:string) > Module subprogram name(s_to_fix_c) > jwd2063i-s "iso_vst.f90", line 320: The dummy argument of the elemental procedure shall not appear in a > specification expression.(name:length) In the latest version these are not elemental, merely pure. > Module subprogram name(s_ass_s) > jwd2063i-s "iso_vst.f90", line 339: The dummy argument of the elemental procedure shall not appear in a > specification expression.(name:expr) This is definitely meant to be ELEMENTAL. The code is ELEMENTAL SUBROUTINE s_ass_s(var,expr) type(VARYING_STRING),INTENT(OUT) :: var type(VARYING_STRING),INTENT(IN) :: expr ! assign a string value to a string variable overriding default assignement ! reallocates string variable to size of string value and copies characters CHARACTER :: tmpexpr(1:LEN(expr)) tmpexpr = expr%chars IF(ASSOCIATED(var%chars))DEALLOCATE(var%chars) ALLOCATE(var%chars(1:SIZE(tmpexpr))) var%chars = tmpexpr ENDSUBROUTINE s_ass_s The line CHARACTER :: tmpexpr(1:LEN(expr)) appears to offend a constraint that I had not noticed until now, Page 214, lines 3-5, which restricts the use of dummy args in spec-expr to appear as args to intrinsic inquiry functions. Thus removing one of the important capabilities of PURE functions introduced in F95. In this particular case I can work round the problem within the module since the internal structure of the type is visible so I can replace the offending declaration by. CHARACTER :: tmpexpr(1:LEN(expr%chars)) instead of the call to the overload for strings defined elsewhere in the module. A user of the module cant do this because the internal structure is rightly private. The major advance of F95 in supporting semantic extension/data abstraction and moving towards Object orientation is therebye seriously damaged yet again by an unnecessary restriction in this area. Why was this constraint not phrased in terms of specification functions, in fact why is this constraint needed at all? > Module subprogram name(replace_ssc) > jwd1727i-s "iso_vst.f90", line 2345: The name cannot be specified because it is already used as a module > function name.(name:replace_sss) This is a typo and has been fixed in latest version > > > --- > Minoru Tanaka > Email: m-tanaka@lp.nm.fujitsu.co.jp > > > Check the URL for latest versions. http://ns7.liv.ac.uk/jls/ -- Lawrie Schonfelder Director, Computing Services Dept. The University of Liverpool, UK, L69 7ZF Phone: 44(151)794 3716, Fax: 44(151)794 3759