vsnyder@jato.jpl.nasa.gov (Van Snyder) (03/15/91)
In article <27452@uflorida.cis.ufl.EDU> peloquin@qtp.ufl.edu (Renee P. Peloquin) writes: >I thought the fortran standard allowed dimensioning an adjustable array to >ANYTHING. However, on the FPS-500, (strings on `which f77` reveals >$Header: crt0.c,v 1.10 88/09/20 14:11:01 grace Rel $ >4.6 (Berkeley) 3/30/83 >$Header: f77.c,v 1.148 90/06/06 09:54:38 steve Rel $ ) > >The following subroutine, of which I have included salient parts: > subroutine fij11 (fij, eom, neom, ns, irrep, ispin, iprd) > > integer iofft1ip, nt1ip, ipsytyp, iprpdpd, symvab, symvij, symvia, > $ nfab(2), nfij(2), nfia(1) > common /ipsym/ iofft1ip(maxrep,2), nt1ip(2), > $ ipsytyp(2,100), iprpdpd(maxrep,6), > $ symvab(maxrep,2), symvij(maxrep,2), symvia(maxrep,2), > $ nfab, nfij, nfia > > double precision fij(nfij), eom(neom,*) > >Gets the following error messages on compilation: >Error on line 49 of makfij.f: no vector operands allowed >Error on line 77 of makfij.f: no vector operands allowed > >(Where lines 49 and 77 are the first and last executable lines.) >Replacing the "double precision" line with: > double precision fij(*), eom(neom,*) > >Causes the subroutine to compile without complaint. > >Does anybody know why? Is this a bug or a feature? > > >-- >Renee P. Mattie >mattie@sun0.chem.upenn.edu It's a bug, but in your code: you can only use scalars for dimensioning parameters; you've used an array of length 2, viz. nfij. Regards, Van. -- vsnyder@jato.Jpl.Nasa.Gov ames!elroy!jato!vsnyder vsnyder@jato.uucp