RFR: 8219585: [TESTBUG] sun/management/jmxremote/bootstrap/JMXInterfaceBindingTest.java passes trivially when it shouldn't

[Daniel Fuchs]

Hi Severin,

On 25/02/2019 15:31, Severin Gehwolf wrote:
> Hi Daniel,
> 
> Thanks for the review!
> 
> Latest webrev:
> http://cr.openjdk.java.net/~sgehwolf/webrevs/JDK-8219585/02/webrev/
> 
[...]
> I'm not sure we need a larger timeout. It's 20 seconds on my machine
> which seems more than enough. What makes you think we'd need to get
> this increased and it would actually do anything for test stability?

Meanwhile I launched your proposed changes through our
test system. The good news is that there was no failure
(even when repeating the test 25 times), and without
changing the 5 timeout - so that may be OK.
Maybe my machine has something that slows it down.
Or Maybe I should have used a timeout factor of 4.
It does pass if I use -timeout:4 when invoking it with jtreg.

The not so good news is that it seems none of our test
machine has a non-loopback address configured for local host.
So even with your changes, the test always trivially pass.

I see the message says "Ignoring manual test" - but the
test is not manual?

[...]

> Yes. I've tried to avoid those issues by allowing one non-loopback
> address and 127.0.0.1 explicitly. Is that a reasonable assumption?
> 

WRT using the loopback as an alternate configuration,
I wonder - should we add the loopback address before
checking size() < 1?
Not sure how stable the test will be in that case.
Also should we use InetAddress.getLoopbackAddress() instead
of using "127.0.0.1" directly?


> My bet is on it passing trivially with:
> 
> "Ignoring manual test since no more than one IPs are configured for 'localhost'"
> 

Well - yes - I believe that's what's happening.

best regards,

-- daniel