RFR: 8342376: More reliable OOM handling in ExceptionDuringDumpAtObjectsInitPhase test

[Aleksey Shipilev]

On Wed, 16 Oct 2024 18:51:43 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

> I'll look again tomorrow.

Yes, I think that's what happens. The test catches the secondary OOM from later code. Shenandoah is different a bit in that it tries to power through even with back-to-back GCs, while other GCs are OOM-ing with GC overhead limit. The easier patch is to drop the size of the waste array we have, so that there is less headroom after we get the first (omitted) OOM.

See new commit.

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

PR Comment: https://git.openjdk.org/jdk/pull/21536#issuecomment-2419027783