[teststabilization] RFR: 8227539: Replace wildcard address with loopback or local host in tests - part 20

[Michael McMahon]

Hi Daniel,

I think I prefer the pattern of URL creation where you use URIBuilder
and then toURL() as the final step, rather than toString() and then new 
URL(string).
But, it's not a big deal.

NoLoopbackPackets looks like it could be simplified with try with 
resources (line 166)
unless you expect the close() to sometimes throw an exception that needs 
to be ignored .?

Can you use the single-arg version of Collectors.joining() in B6210227.java
as the prefix and suffix are empty?
What about using InetAddress.isAnyLocalAddress() instead of testing for 
zero bytes
in isWildcard()?

Looks fine otherwise.

Thanks,
Michael.

On 10/07/2019, 17:17, Daniel Fuchs wrote:
> Hi,
>
> Please find a patch that fixes some of the network test
> to stop the using the wildcard issue whenever possible.
>
> 8227539: Replace wildcard address with loopback or local host in
>          tests - part 20
> https://bugs.openjdk.java.net/browse/JDK-8227539
>
> webrev:
> http://cr.openjdk.java.net/~dfuchs/webrev_8227539/webrev.00/
>
>
> test/jdk/java/net/MulticastSocket/NoLoopbackPackets.java is the outlier
> here. It's been observed failing because it was receiving packets
> it didn't expect. The test is modified to skip packets that were
> not sent by the test. This is achieved by making the payload
> pseudo-unique.
>
> best regards,
>
> -- daniel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/net-dev/attachments/20190710/058218d0/attachment.html>