RFR: 8146728: TestPrintGCDetailsVerbose is never run by jtreg

Stefan Johansson stefan.johansson at oracle.com
Tue Jan 12 19:56:13 UTC 2016



Kim Barrett <kim.barrett at oracle.com> skrev: (12 januari 2016 19:53:21 CET)
>On Jan 12, 2016, at 1:46 PM, Kim Barrett <kim.barrett at oracle.com>
>wrote:
>> 
>>> On Jan 12, 2016, at 11:02 AM, Stefan Johansson
><stefan.johansson at oracle.com> wrote:
>>>>> CR:
>>>>> https://bugs.openjdk.java.net/browse/JDK-8146728
>>>>> 
>>>>> Webrev:
>>>>> http://cr.openjdk.java.net/~kbarrett/8146728/webrev.00/
>>>> The change looks good if we want to continue running the test with
>debug-builds only, but I wonder if it really makes sense to do so. I
>suspect that the reason we had the restriction was that Verbose was
>only available in debug-builds. But now when using -Xlog:gc*=trace the
>test should work with product builds as well, or am I missing
>something?
>>> 
>>> The test was created to verify that some assertions were no longer
>triggered.  Not product => no assertions to be triggered.
>> 
>> That said, there probably is some benefit to beating on (GC) logging
>in a product build.  So sure, I’ll make that change.
>
>In which case, should I also remove the jdk.debug requires property? 
>This is the only test presently using this feature.  It can always be
>added back if some test needs it in the future.
>
I'm all for removing it, as you say, it's easy to add back in the future when needed. 

>I think this is also the only test using ~= (regex match) in a
>@requires.

Is this documented somewhere? Just so the functionality isn't forgotten when the only use is removed. 

Stefan 



More information about the hotspot-gc-dev mailing list