RFR[8241072]: 'Reimplement Legacy DatagramSocket API'

[Patrick Concannon]

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/

Kind regards,

Patrick



On 17/04/2020 10:35, Alan Bateman wrote:
> On 15/04/2020 20:22, Patrick Concannon wrote:
>> :
>>
>> WRT the PDSI issue, I've created a bug to track this and have 
>> assigned it to myself. You can view the issue here: 
>> https://bugs.openjdk.java.net/browse/JDK-8242885
>>
>>
> Thanks. That should be trivial to fix and should allow SendBufCheck to 
> be updated to avoid testing that the older implementation has the 
> wrong limit. The JEP hasn't been proposed to target yet so I think 
> there is time to get this sorted out in advance.
>
> Is SetGetSendBufferSize.java about extending test coverage? If so then 
> I think the range of input needs to be expanded a bit, e.g. attempting 
> to set to 0 should throw IAE, it should at least be possible to set it 
> to 64. I'm curious why the test checks getSoTimeout after the socket 
> has been closed. Do you know if the test coverage for the receive 
> buffer size is sufficient?
>
> -Alan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/net-dev/attachments/20200506/5f699732/attachment.htm>