[teststabilization] RFR: 8226825: Replace wildcard address with loopback or local host in tests - part 19

[Chris Hegarty]

Julia,

> On 26 Jun 2019, at 18:33, Julia Boes <julia.boes at oracle.com> wrote:
> 
> Hi,
> 
> Please find below a patch for:
> 
> 8226825: Replace wildcard address with loopback or local host in tests - part 19
> 
> https://bugs.openjdk.java.net/browse/JDK-8226825
> 
> 
> webrev:
> 
> http://cr.openjdk.java.net/~dfuchs/jboes/webrev_8226825/webrev/

This is good work, and interesting what one finds when digging into some
of these old tests. I can see that AsyncDisconnect does not operated as
expected ( the currently checked in version passes on my machine,
without the test executing correctly )!

Rather than catching and re-throwing non-test sensitive exceptions, it
is simpler to just let them flow up the stack ( the methods that are
currently catching can just declare `throws Exception` ). Since jtreg
will report a test failure if there is an uncaught exception in the main
thread.

Additionally, any threads that are started by a test, AsyncDisconnect is
one such example, should wait (join) the threads that it has started.
This is good practice, since jtreg considers a test complete when its
main method returns ( could be non-daemon test threads left behind ).

-Chris.