Xref: utzoo comp.os.minix:7676 comp.sys.ibm.pc:36948
Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!iuvax!purdue!bu-cs!xylogics!world!madd
From: madd@world.std.com (jim frost)
Newsgroups: comp.os.minix,comp.sys.ibm.pc
Subject: Re: dosread.c again
Message-ID: <1989Oct25.165923.28395@world.std.com>
Date: 25 Oct 89 16:59:23 GMT
References: <3717@ast.cs.vu.nl> <3a18.2536ede8@ibmpcug.co.uk> <3721@ast.cs.vu.nl> <2501@optilink.UUCP> <1989Oct20.170447.19573@utzoo.uucp> <5182@uhccux.uhcc.hawaii.edu> <6627@ficc.uu.net> <11039@phoenix.Princeton.EDU> <6660@ficc.uu.net>
Reply-To: madd@world.UUCP (jim frost)
Followup-To: comp.sys.ibm.pc
Organization: Software Tool & Die
Lines: 20

In article <6660@ficc.uu.net> peter@ficc.uu.net (Peter da Silva) writes:
|And how about standard library calls like "popen"?
|How do you do that on an IBM?

I did an off-the-cuff implementation a couple of years ago.  If it's a
write pipe, open a tmp file on popen, dump the stuff to it, and spawn
the other process on pclose.  If it's a read pipe, spawn the process
immediately with output to a tmp file and start reading on its
completion.  Perfect semantics is a little more difficult but it does
work fairly well.  Someone actually turned my implementation into a
real thing, so I guess there was demand for it :-).

Personally I'd just put UNIX on the machine and solve the whole
problem.

Followups redirected to comp.sys.ibm.pc.

jim frost
software tool & die
madd@std.com