NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/lang/fortran/5165

Text File | 1993-01-25 | 1.9 KB | 35 lines

Newsgroups: comp.lang.fortran
Path: sparky!uunet!utcsri!helios.physics.utoronto.ca!alchemy.chem.utoronto.ca!system
From: [e-mail] (System Admin (Mike Peterson))
Subject: Re: Help with UNIX f77 file buffer system
Message-ID: <[e-mail]>
Organization: University of Toronto Chemistry Department
References: <[e-mail]>
Date: Mon, 25 Jan 1993 16:17:32 GMT
Lines: 24

In article <[e-mail]> [e-mail] (Petros Dafniotis) writes:
>Hello netters,
>
>  I know very little about UNIX (in fact I am using VMS for almost 9yrs now).
>A friend of mine had the following problem:
>"His program (all written in FORTRAN by 1-2 people years ago) takes quite
>some time to finish. He is writing info produced at
>run time to 2-3 output files. However, 90% of the runs are proved to be
>useless quite soon; so he uses CTRL-C to abort execution. But then he finds
>no info in the output files. I am guessing that the UNIX file buffer system
>still keeps these numbers in the buffer and what he needs is to flush the
>buffers after each single write. Could someone tell me how can this be
>accomplished in UNIX? (he is using HP-RISC and SUN SPARC-I workstations)."

Try the 'flush' routine from the BSD F77 library (it should be in HP-UX
9.0, but is not in HP-UX 8.07); Sun has had that library for ages.
If you are stuck at 8.07, write a C routine that calls fflush
after using FSTREAM (see p. 9-30 of the FORTRAN manual) to get the
C "FILE *" pointer. This approach will not work on stdin/stdout/stderr
at 8.07, but works fine on other unit numbers.
--
core error - bus dumped    -*-    Mike Peterson, SysAdmin, U/Toronto Chemistry
 *******   As usual, I speak only for me, myself and I; nobody else   *******
E-mail: [e-mail]  Tel: [phone]  Fax: [phone]