Found a UDP regression when exercising UDP from solaris to win32 on
JDK1.1.7A build. What happens is I get data corruption on some datagram
packets going from solaris to win32 when I run UDP in stress mode. This
is a new problem. I never seen data corruption before. The problem only
occurs from solaris to win32. The problem happens with and without the
JIT.
As part of the attachment I have the UDP Test Suite. To reproduce this
problem on JDK1.1.7A do the following:
% uncompress udp.tar.Z
% tar xvf udp.tar
% cd udp
% make
On win95 type:
win95> cd udp
win95> run_udp_server
On solaris type:
solaris> cd udp
solaris> client1 win95; client2 win95; client3 win95
The results are stored in the results directory. Grep for the failures
using:
solaris> grep ERROR results/*
JDK1.1.7A build. What happens is I get data corruption on some datagram
packets going from solaris to win32 when I run UDP in stress mode. This
is a new problem. I never seen data corruption before. The problem only
occurs from solaris to win32. The problem happens with and without the
JIT.
As part of the attachment I have the UDP Test Suite. To reproduce this
problem on JDK1.1.7A do the following:
% uncompress udp.tar.Z
% tar xvf udp.tar
% cd udp
% make
On win95 type:
win95> cd udp
win95> run_udp_server
On solaris type:
solaris> cd udp
solaris> client1 win95; client2 win95; client3 win95
The results are stored in the results directory. Grep for the failures
using:
solaris> grep ERROR results/*