home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!strath-cs!cen.ex.ac.uk!JRowe
- From: JRowe@cen.ex.ac.uk (J.Rowe)
- Newsgroups: comp.lang.fortran
- Subject: Re: flushing output in f77
- Message-ID: <BxnoG5.LnF@cen.ex.ac.uk>
- Date: 13 Nov 92 13:07:16 GMT
- References: <141324@lll-winken.LLNL.GOV>
- Sender: JRowe@cen.ex.ac.uk (J.Rowe)
- Organization: Computer Unit. - University of Exeter. UK
- Lines: 21
- In-Reply-To: sterne@dublin.llnl.gov's message of 12 Nov 92 22:10:39 GMT
-
- In article <141324@lll-winken.LLNL.GOV> sterne@dublin.llnl.gov (Philip Sterne) writes:
-
- > Is there a clean standard-conforming way of emptying the write
- > buffer for a unit in f77?
-
- Standard conforming? No. But what we do here in the Physics Department at
- Exeter is to have a tiny c routine:
-
- #include <stdio.h>
- void flushout()
- {
- fflush(NULL);
- }
-
- which we can compile to an object file and then link to our fortran.
- It seems to work fine under SunOS, ConvexOS, AIX and HP-UX. Some OSs
- demand an underscore to be added to the routine name in the c source,
- I can't remember whether it goes on the beginning of the name or the
- end.
-
- John.
-