RFR[8241072]: 'Reimplement Legacy DatagramSocket API'

[Alan Bateman]

On 06/05/2020 21:27, Patrick Concannon wrote:
>
> Hi Alan,
>
> With regards to the SetGetSendBufferSize.java test: yes, it was 
> created to increase the test coverage for the DatagramSocket class. 
> However, it was decided that it should be handled separately and was 
> pushed to the mainline in advance of this RFR - (JDK-8243488 
> <https://bugs.openjdk.java.net/browse/JDK-8243488>).
>
>
> Please find the updated webrev below:
>
> http://cr.openjdk.java.net/~pconcannon/8241072/webrevs/webrev.05/
>

I think this is a good shape and you (or Daniel) should be able to push 
soon after JEP 373 has been targeted and the code review is completed here.

I don't have any further comments on the implementation changes.

My only issue with the tests is that I don't think SendBufCheck should 
be pushed to jdk/jdk. Test SetGetSendBufferSize has 
testInitialSendBufferSize to test the initial size of the send buffer so 
you have test coverage for the initial size. My problem with 
SendBufCheck is that it is checks that bug JDK-8242885 exists in the old 
implementation. That just adds technical debt to the test suite. I'd 
prefer to see JDK-8242885 fixed and 
SetGetSendBufferSize/testInitialSendBufferSize updated to test a 
maximally sized IPv6 packet (we should probably have done that in 
advance of the JEP). At some point I think we should also make sure that 
we have a good set of tests that send large datagrams on the network 
(not the loopback) to ensure that we don't have any other issues.

-Alan.



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/net-dev/attachments/20200507/e4056179/attachment.htm>