Re: Still more on Fortran I/O

Article View: aus.stats.s

Article #16136

#16136

From: Allen Mcintosh
Date: Fri, 17 Jan 1992 03:57

11 lines
1213 bytes

Charlie Geyer is correct.  The lI77/lF77 order is vendor dependent.  One vendor
(I have forgotten which one) required one of the libraries be specified
twice - a consequence of not having a Fortran main program.

On systems whose Fortran compilers/libraries are based on the
Feldman/Weinberger compiler, or systems where all the Fortran has been
run through f2c (which is based on the Feldman/Weinberger compiler), you can
usually get away with bracketing the call to do_S in main.c with calls to
f_init() and f_exit().  f_init on the SPARCs seems to have arguments; I have
no idea what they might be for.
D

Technical Details

Message-ID: <9201170338.AA02524@larch.bellcore.com>

Path: arc1-ca.pugleaf.net!sparky!uunet!munnari.oz.au!uniwa!dmsperth!dmssyd.syd.dms.CSIRO.AU!news

Show Full Headers

Complete Article Headers

X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: fa931,172bcefc8e3d9200
X-Google-Attributes: gidfa931,public
X-Google-ArrivalTime: 1992-01-16 20:21:45 PST
Path: arc1-ca.pugleaf.net!sparky!uunet!munnari.oz.au!uniwa!dmsperth!dmssyd.syd.dms.CSIRO.AU!news
From: mcintosh@larch.bellcore.com (Allen Mcintosh)
Newsgroups: aus.stats.s
Subject: Re: Still more on Fortran I/O
Message-ID: <9201170338.AA02524@larch.bellcore.com>
Date: 17 Jan 92 03:57:12 GMT
Sender: daemon@syd.dms.CSIRO.AU
Organization: CSIRO Div. Maths & Stats, CSIRO Div. Applied Physics
Lines: 10
X-Gateway: dmssyd.syd.dms.CSIRO.AU