RFR(xxs): 8199518: test/hotspot/jtreg/runtime/SelectionResolution tests take a lot longer to run with fastdebug after JDK-8198423

coleen.phillimore at oracle.com coleen.phillimore at oracle.com
Tue Mar 13 18:50:39 UTC 2018

This looks good and like a trivial patch.  I think you should push it.

On 3/13/18 2:12 PM, Thomas Stüfe wrote:
> Hi,
> Bug: https://bugs.openjdk.java.net/browse/JDK-8199518
> webrev: 
> http://cr.openjdk.java.net/~stuefe/webrevs/8199518-SelectionResolution-tests-time-out-after-8198423/webrev.00/webrev/ 
> <http://cr.openjdk.java.net/%7Estuefe/webrevs/8199518-SelectionResolution-tests-time-out-after-8198423/webrev.00/webrev/>
> JDK-8198423 made fastdebug too slow for some tests which heavily 
> stress metaspace allocation. One of these tests is 
>  runtime/SelectionResolution/*, where a lot of dynamic class 
> generation happens.
> The fix tones down a number of the more aggressive asserts. Similar 
> things were done before already, so there is already a static switch 
> to switch the expensive asserts off, which I reused.
> In a later change I'd like to make this a diagnostic switch. For now I 
> like to keep this patch simple.
> Thanks and sorry for the trouble,
> Kind Regards, Thomas

More information about the hotspot-runtime-dev mailing list