RFR (S) 8016752: Test case for 8005857/8016740 (assert in GC_locker from PSOldGen::expand with -XX:+PrintGCDetails and Verbose)

Dmitry Fazunenko dmitry.fazunenko at oracle.com
Fri Nov 20 11:16:12 UTC 2015


Hi Thomas,

Thanks a lot for looking at the fix.  I totally agree with you: CR 
should be well descriptive. Fixed.

A question: This test is for Parallel only, because it's a regression 
test for a problem in Parallel.
It could be easily made applicable for all collectors to check that 
PrintGCDetails + Verbose doesn't crash when OOM happens.
What is your opinion, does it make any sense or that case is very unlikely?

Thanks,
Dima


On 20.11.2015 13:46, Thomas Schatzl wrote:
> Hi,
>
> On Mon, 2015-11-16 at 20:46 +0300, Dmitry Fazunenko wrote:
>> Hello,
>>
>> may I have a couple of reviews for a fix made Andrey Zakharov, please:
>>
>> https://bugs.openjdk.java.net/browse/JDK-8016752
>> http://cr.openjdk.java.net/~dfazunen/8016752/webrev.00/
>>
>> Comments: during the fix TEST.ROOT is modified: java.version is
>> registered as a property that can be used in the @requires tag.
>> It allows to declare a test as fastdebug specific:
>>
>> @requires java.version ~= ".*fastdebug"
>    looks okay. Is it possible to update the CR to include a description
> of the problem (+ change the summary), not just referencing other bugs.
>
> Thanks,
>    Thomas
>
>
>



More information about the hotspot-gc-dev mailing list