RFR: 8062744 jdk.net.Sockets.setOption/getOption does not support IP_TOS

[Michael McMahon]

On 04/11/14 14:45, Michael McMahon wrote:
> On 04/11/14 12:49, Chris Hegarty wrote:
>> On 4 Nov 2014, at 12:05, Michael McMahon 
>> <michael.x.mcmahon at oracle.com> wrote:
>>
>>> On 04/11/14 11:18, Chris Hegarty wrote:
>>>> On 4 Nov 2014, at 11:15, Michael McMahon 
>>>> <michael.x.mcmahon at oracle.com> wrote:
>>>>
>>>>> On 04/11/14 11:10, Alan Bateman wrote:
>>>>>> On 04/11/2014 11:00, Michael McMahon wrote:
>>>>>>> Thanks Alan. How about I just split the test and check the 
>>>>>>> option setting behavior in OptionsTest.java
>>>>>>> which doesn't have any reference to the jdk.net.Sockets APIs and 
>>>>>>> then check the Sockets.supportedOptions()
>>>>>>> method in the new test?
>>>>>> That sounds okay too, I assume it's possible to come up with a 
>>>>>> good name for the new test too.
>>>>>>
>>>>>> -Alan
>>>>> SupportedOptions.java maybe.
>>>> That would be fine, or
>>>>    SupportedOptionsTest.java, or
>>>>    StandardSocketOptionsTest.java
>>>>
>>>> -Chris.
>>> Ok. I've just realised that OptionsTest.java is testing all of the 
>>> legacy option set/get calls
>>> together with the new setOption(), getOption() calls. Problem with 
>>> this:
>>>
>>> 1) There is no legacy API for setting IP_TOS on ServerSocket
>>>
>>> 2) The test is not in 8u-dev (because the setOption/getOption API is 
>>> 9 only)
>>>
>>> So, I'd like to stick with the original test (renamed).
>> OK, I see the issue now.
>>
>> Just another comment when I looked into this further; Should 
>> java/net/SocketImpl.java also add IP_TOS to it’s list of server 
>> socket options?
>
> Right. Good catch. I've updated the webrev including the test to catch 
> the above.
>
> http://cr.openjdk.java.net/~michaelm/8062744/webrev.2/

here is the JDK 8 version. Almost the same, but only the changes 
relating to jdk.net.Sockets
apply in this version.

http://cr.openjdk.java.net/~michaelm/8062744/webrev.1.jdk8/

Thanks
Michael