RFR(T): 8214857: "bad trailing membar" assert failure at memnode.cpp:3220
tobias.hartmann at oracle.com
Thu Dec 6 09:16:45 UTC 2018
Thanks Igor, I understand that we are limited by time/resources.
On 05.12.18 20:09, Igor Ignatyev wrote:
> of course we can do it for other jar files, and we used to that, however benefit-cost ratio wasn't high enough, and due to this and other reasons it was decided to stop this kind testing. if someone knows good candidates (where good means one which have greater chances to crash during ctw) for regular testing, we can work on adding them into regular testing.
Yes, I remember that and I also thought that this testing caught quite a number of bugs back then
but my memory might be distorted (I used it a lot for stress testing of local changes). A big
advantage of CTW is also the reproducibility because there is no dynamically generated code.
More information about the hotspot-compiler-dev