RFR(S): URG: JDI test com/sun/jdi/ProcessAttachTest.sh and other 3 jdi tests failed in nightly

[Daniel D. Daugherty]

On 5/22/14 10:29 AM, Dmitry Samersoff wrote:
> Dan,
>
> Thank you for review!
>
> Please, see also below.
>
> On 2014-05-22 19:41, Daniel D. Daugherty wrote:
>> http://cr.openjdk.java.net/~dsamersoff/JDK-8043716/webrev.01/
>>
>> General:
>>      - I presume you've grep'ed for all uses of InetAddress.getLocalHost()
>>        in the JDI code to make sure that you've gotten all the uses.
> Yes.
>
>>      - You need to run an Aurora Adhoc job with all of the
>>        Serviceability stack from JDI and up.
> I created a bad network configuration for Linux, Windows and MACOS X in
> my home lab and run all com/jdi/* tests.
>
> Is it sufficient - I would like to commit changes ASAP?

The NSK/VM testbase has an nsk.jdi subsuite. You should also run the
nsk.jdb subsuite. I don't think vm.heapdump exercises JDI.

Dan



>
>> src/share/classes/com/sun/tools/jdi/SocketAttachingConnector.java
>>      No comments.
>>
>> src/share/classes/com/sun/tools/jdi/SocketTransportService.java
>>      No comments.
>>
>> src/share/transport/socket/socketTransport.c
>>      line 206: memset((void *)&hints, 0, sizeof(struct addrinfo));
>>          You might consider:
>>
>>               206: memset((void *)&hints, 0, sizeof(hints));
>>
>>          I think it's a little cleaner pattern since you don't
>>          have to know the type of 'hints'.
> Changed.
> -Dmitry
>
>> test/com/sun/jdi/BadHandshakeTest.java
>>      No comments.
>>
>> test/com/sun/jdi/RunToExit.java
>>      No comments.
>>
>> Dan
>>
>>
>> On 5/22/14 6:54 AM, Dmitry Samersoff wrote:
>>> Hi Everybody,
>>>
>>> Please review:
>>>
>>>    http://cr.openjdk.java.net/~dsamersoff/JDK-8043716/webrev.01/
>>>
>>> Problem summary:
>>>
>>>     InetAddress.getLocalHost().getHostName() returns configured host name
>>> ever if "localhost" is known.
>>>
>>>     If configured hostname resolves to IP address different from
>>> "localhost" one, a test would fail because after JDK-8041435 the server
>>> binds to localhost only if no address specified.
>>>
>>> Fix summary:
>>>
>>>       Calls to InetAddress.getLocalHost().getHostName() replaced to
>>> explicit "localhost" but "*" as a hostname could be used to restore old
>>> behavior.
>>>
>>>       Also, fixed windows compiler warning and added better
>>> diagnostic to the test.
>>>
>>>
>>> CR Link:
>>>      https://bugs.openjdk.java.net/browse/JDK-8043716
>>>
>