RFR: 7903671: jcstress: Update buffer tests for JDK-8318966
Jiří Vaněk
jvanek at openjdk.org
Mon Mar 3 10:51:11 UTC 2025
On Wed, 26 Feb 2025 19:27:27 GMT, Jiří Vaněk <jvanek at openjdk.org> wrote:
> This is another possible fix for https://bugs.openjdk.org/browse/CODETOOLS-7903671 - : jcstress: Update buffer tests for JDK-8318966 (Bug - P4)
> This one is attempting Experimental reduction of (0, 8) -> (0, 1) in all affected calls. Brief experiments seesm the change works. But whether it is still testing what it should.. I try to decrypt it, but up to now now proof.
>
> I have not yet tested all affected tests. That is in progress.
Ok. I made a much more deeper look, and consulted with Aph and Thomas, and this fix indeed looks wrong from all points of view.
To reduce it to (1,1),except obvious failures, seesm to no longer check the issue. Also it seems that for jdk 23 and up, the correct replacement are direct buffers - which - feel free to correct me - are well covered in jcstress,
Thus the correct fix seesm to be https://github.com/openjdk/jcstress/pull/159 or to remove the.
Closing this in favour of https://github.com/openjdk/jcstress/pull/159
-------------
PR Comment: https://git.openjdk.org/jcstress/pull/160#issuecomment-2693968932
More information about the jcstress-dev
mailing list