From owner-freebsd-hackers  Wed Oct 15 05:19:36 1997
Return-Path: <owner-freebsd-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id FAA05997
          for hackers-outgoing; Wed, 15 Oct 1997 05:19:36 -0700 (PDT)
          (envelope-from owner-freebsd-hackers)
Received: from word.smith.net.au (ppp20.portal.net.au [202.12.71.120])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id FAA05951
          for <hackers@FreeBSD.ORG>; Wed, 15 Oct 1997 05:19:20 -0700 (PDT)
          (envelope-from mike@word.smith.net.au)
Received: from word.smith.net.au (localhost [127.0.0.1])
	by word.smith.net.au (8.8.7/8.8.5) with ESMTP id VAA00326;
	Wed, 15 Oct 1997 21:44:42 +0930 (CST)
Message-Id: <199710151214.VAA00326@word.smith.net.au>
X-Mailer: exmh version 2.0zeta 7/24/97
To: stesin@gu.net
cc: Mike Smith <mike@smith.net.au>, hackers@FreeBSD.ORG
Subject: Re: Call for Fortran assistance. 
In-reply-to: Your message of "Wed, 15 Oct 1997 12:37:25 +0300."
             <Pine.BSF.3.96.971015121206.1265F-100000@trifork.gu.net> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Wed, 15 Oct 1997 21:44:40 +0930
From: Mike Smith <mike@smith.net.au>
Sender: owner-freebsd-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

> On Wed, 15 Oct 1997, Mike Smith wrote:
> 
> > Our second design threw the data through a FIFO, but Fortran has funny 
> > formatting requirements for sequential-access data that have stymied us.
> 
> 	What was so strange there with the formatting?

To quote from /usr/src/lib/libI77/README:

Unformatted sequential records consist of a length of record
contents, the record contents themselves, and the length of
record contents again (for backspace).  Prior to 17 Oct. 1991,
the length was of type int; now it is of type long, but you
can change it back to int by inserting

> > We tried converting to ASCII and then parsing it back in, but that's 
> > too slow.
> 
> 	As far as I recall, Fortran is able to handle "binary"
> 	input records (in a way like one do read(2)/write(2)
> 	of a memory buffer containig a C struct foo { ... };
> 	or some kind of an array)

You can do this with an unformatted file open for direct access, but 
this fails (blocks forever) on a FIFO.  Worse, you can't vary the size 
of the entity read.  8(

>	Anyway you will be
> 	able to fill a Fortran array with bytes from an input
> 	stream without any conversion and use the buffer' content
> 	in any way you like then, using Fortran' equivalent
> 	of C union xxx { ... };  that is a COMMON construct.

This would be OK if it worked (forgive any syntax, I'm not a Fortran 
programmer):

	character*1	foo

	open(unit=16,file='/tmp/fifo',form='unformatted',access='direct',
	$ recl=1,status='old')

	do irec = 0, 10000
		read(unit=16,rec=irec,end=200)foo
		...
	enddo

but given a FIFO; eg. in a test case:

# mkfifo /tmp/fifo
# cat /kernel > /tmp/fifo &
# ./testprogram 

it just sits there doing nothing.

mike