RE: int/FILE* bug in Iperf 1.6.4

To: James Devenish <j-devenish --at-- users.sourceforge.net>
Subject: RE: int/FILE* bug in Iperf 1.6.4
From: Kevin Gibbs <kgibbs --at-- ncsa.uiuc.edu>
Date: Sun, 5 Jan 2003 19:42:48 -0600 (CST)
Cc: iperf-users --at-- dast.nlanr.net
Content-type: TEXT/PLAIN; charset=US-ASCII
In-reply-to: <3.0.5.32.20030105095217.01009080 --at-- pop.ncsa.uiuc.edu>
Reply-to: Kevin Gibbs <kgibbs --at-- ncsa.uiuc.edu>
Sender: owner-iperf-users --at-- dast.nlanr.net

> >Date: Sun, 5 Jan 2003 20:33:32 +0800
> >From: James Devenish <j-devenish --at-- users.sourceforge.net>
> >To: dast --at-- nlanr.net
> >Subject: int/FILE* bug in Iperf 1.6.4
> >
> >There appears to be some interchange of (int) and (FILE*)
> >in Extractor.cpp and Listener.cpp and PerfSocket.cpp
> >(and that makes iperf crash of LP64 platforms, for instance).
> >This includes instances of close((int)fd). Looks like you
> >could just redefine one of Extractor's constructors to take
> >FILE* instead of int. I'd send you a patch but I have to
> >catch a bus now.

Thanks for the notice. This is a rarely used feature of representative 
streams and thus rarely if ever causes any crashes, but we will fix it in 
future releases.

Kevin

Next by Date: 1.6.5?
Next by thread: 1.6.5?

Other Mailing lists | Author Index | Date Index | Subject Index | Thread Index