Need reviewers for changes in java/net regression tests

[Pavel Tisnovsky]

Chris Hegarty wrote:
> On 07/26/10 13:05, Pavel Tisnovsky wrote:
>> Chris Hegarty wrote:
>>> Pavel,
>>>
>>> Great to see some of the test changes making their way back into 6
>>> Open. From an administration point of view this approach may be less
>>> than ideal, that is, we need a bug number for each push.
>>>
>>> Would it make sense to break down the networking tests to say
>>> java/net, sun/net and httpserver, then try to complete the backport of
>>> these areas? I would't try to add any new functionality to the tests,
>>> i.e. that was added for new bugs fixed in JDK7. Any tests with
>>> additional bugs numbers in the @bug tag probably should only have
>>> minimal changes to clean up resources, etc. Other than that it should
>>> be simply a matter of taking the JDK7 version of the file, right?
>>
>> I think I understand the problems with more pushes than is necessary. Is
>> it better to made webrev containing only "cosmetic" changes in all net
>> regression tests at first and after this webrev will be accepted send
>> another webrev(s) containing more changes (new bug numbers etc.)?
> 
> Sorry, I don't understand what you mean by 'cosmetic'. I think you can 
> do the complete backport of the tests in one go, but for convenience it 
> may be easier to do it section by section, java/net sun/net, etc.

Hi Chris,

I mean the difference between (for example) these changes:
http://cr.openjdk.java.net/~ptisnovs/JavaNetJTregTests/test/java/net/DatagramSocket/DatagramTimeout.java.cdiff.html
and
http://cr.openjdk.java.net/~ptisnovs/JavaNetJTregTests/test/java/net/Inet6Address/serialize/Serialize.java.cdiff.html

In the first case there's only minimal (I said "cosmetic") changes, not 
new test functionality, but in the case there's new functionality added 
to the test (test for regression for another bug).

> 
> -Chris.
> 
>>
>>>
>>> -Chris.
>>>
>>> On 07/23/10 12:20, Pavel Tisnovsky wrote:
>>>> Hi,
>>>>
>>>> I tried to backport more regression tests from OpenJDK7 to OpenJDK6.
>>>> Webrev with changes is available at:
>>>> http://cr.openjdk.java.net/~ptisnovs/JavaNetJTregTests/
>>>>
>>>> Can anyone please review the changes?
>>>>
>>>> There are still minor differences between OpenJDK6 and OpenJDK7
>>>> tests, mainly due to syntax enhancements in OpenJDK7.
>>>>
>>>> Please note that new test case (deserialization an Inet6 address with
>>>> a named interface) is included in
>>>> Inet6Address/serialize/Serialize.java regression test:
>>>> http://cr.openjdk.java.net/~ptisnovs/JavaNetJTregTests/test/java/net/Inet6Address/serialize/Serialize.java.cdiff.html 
>>>>
>>>>
>>>>
>>>> This issue which was fixed in OpenJDK7 but not in OpenJDK6 yet. Btw -
>>>> do anyone know, if this issue will be fixed in OpenJDK6?
>>>>
>>>> The changes made in these test are really minor and they are based (but
>>>> not equal) on changes done by Christopher Hegarthy in OpenJDK7:
>>>> http://hg.openjdk.java.net/jdk7/tl/jdk/rev/58f325ba3e27
>>>> (these changes are reviewed by Alan Bateman)
>>>>
>>>> Cheers
>>>> Pavel Tisnovsky
>>