Newsgroups: comp.parallel.pvm
Path: ukc!uknet!nessie!doc.ic.ac.uk!pipex!howland.reston.ans.net!europa.eng.gtefsd.com!MathWorks.Com!news.kei.com!ssd.intel.com!joel
From: joel@ssd.intel.com (Joel Clark)
Subject: Re: I miss a timeout mechanism
Message-ID: <Cn1HoD.HAr@SSD.intel.com>
Sender: usenet@SSD.intel.com
Nntp-Posting-Host: sail
Organization: Supercomputer Systems Division (SSD), Intel
References: <2ml3c5$je9@hpsystem1.informatik.tu-muenchen.de>
Date: Tue, 22 Mar 1994 00:07:24 GMT
Lines: 19

In article <2ml3c5$je9@hpsystem1.informatik.tu-muenchen.de> schmiher@Informatik.TU-Muenchen.DE (Herman The German) writes:
>Hello,
>it would sometimes be convenient to have a timeout mechanism,
>notably in the pvm_receive() function (the blocking one).
>It can happen that an addressed sender is dead and the waiting task
>hangs forever. I check the status of the addressed task in cases I like
>the apllication to behave safely. 
>A timeout would be more convenient, though.
>
>I'm surely not the first one to mention this point, so are there any
>solutions or ideas out there?
>
>Hermann.
>


alarm() and SIGALRM?????

joel

