Name: elR10090 Date: 02/23/2001



Java HotSpot VM (build 1.3.1-beta-b16) hangs up while running the test

    nsk/regression/b4311710

This failure was observed in server and client java_g VMs in -Xmixed
mode and server and client java VMs in -Xmixed, -Xint, -Xcomp modes
under WinNT (Win NT 4.0 Server P-II with two processors, 350MHz,
512Mb RAM).

This test passes fine on other platforms (Linux, Solaris Sparc,
Solaris Intel) on ladybird-b16 and works less then 2 minutes.

Also, the test passes on merlin-b50 on all four platforms (Linux,
Solaris Sparc, Solaris Intel, Windows) in -Xmixed, -Xint, -Xcomp
modes on both server and client VMs.

This failure looks like a regression of the known bug

    4311710: HotSpot VM: Process.getInputStream().read() method hangs for destroyed Process
    
This failure also probably explains bug

    4417524: VM hangs up while running nsk/stress/network/network003 test
    
Both tests try to read from the Process stdout. It looks like the read() method
hangs if Process is destroyed.

The test belongs to the testbase_nsk testbase located at:

    /net/sqesvr.eng/export/vsn/VM/ladybird/testbase/testbase_nsk
    
To reproduce the failure please use 'doit.sh' script found
in the directory:

    /net/sqesvr.eng/export/vsn/VM/GammaBase/Bugs/<this bug number>

Use:
    sh doit.sh $JAVA_HOME

======================================================================

Name: elR10090 Date: 02/26/2001



This failure also looks like a regression of the known bug

    4311711: Classic VM: Process.getInputStream().read() method hangs for destroyed Process
    
======================================================================