RFR: JDK-8306462: Remove NMT gtests that test dead-block-printing

[Thomas Stuefe]

On Thu, 20 Apr 2023 17:13:27 GMT, Daniel D. Daugherty <dcubed at openjdk.org> wrote:

> Thumbs up on the technical changes of removing the tests.
> 
> However, @dholmes-ora added comment to the bug report that should be addressed before this integrated: https://bugs.openjdk.org/browse/JDK-8306462?focusedCommentId=14575199&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14575199
> 
> We need to clarify that the fault lies with the new tests and not the new feature code that was added.

Hmm. Let's keep the perspective. 

The removed tests did test that, if we use NMT during hs-err file reporting to identify and unknown pointer, and that unknown pointer points into a malloced block that recently got freed, and its still around in memory, that printing will tell me "near dead block X". 

The test only exist because I am obsessive about writing tests. Others would probably not  have bothered in the first place. But the tests now failed repeatedly, and always on platforms I cannot access. Fixing them is very time consuming. Seeing the maintenance effort the tests costs, that the feature works (manually verified), I would vote for test removal.

If you guys disagree, that's fine - in that case I just disable the tests but leave them in. Not sure when I have the time to pick it up though, I am snowed in atm.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/13539#issuecomment-1516832451