On Mon, 21 Aug 2023 19:25:25 GMT, Chris Plummer <cjplum...@openjdk.org> wrote:
> The test tries to match up various GC -XX:+PrintFlagsFinal values with the
> output of jhsdb -jmap --pid <pid> --heap. With ZGC, MaxNewSize set to
> (size_t)-1, but PrintFlagsFinal prints it as an unsigned long:
>
> size_t MaxNewSize = 18446744073709551615 {product} {default}
>
> jmap normally prints out both the raw size and the MB size. For example:
>
> MaxHeapSize = 805306368 (768.0MB)
>
> But as part of the fix for
> [JDK-6718125](https://bugs.openjdk.org/browse/JDK-6718125), it stopped
> printing the raw value for negative values an only printed the MB value. So
> for MaxNewSize we had:
>
> MaxNewSize = 17592186044415 MB
>
> Instead of:
>
> MaxNewSize = 18446744073709551615 (17592186044415 MB)
>
> So the test fails to find 18446744073709551615 in the output. I fixed jmap to
> include the raw value as an unsigned long, even if negative as a signed long
> (to be consistent with PrintFlagsFinal output), so now the test passes even
> when MaxNewSize is included in the list.
`(size_t)-1` is a weird value to pass, which requires this gymnastics in the
output code.
But I guess this is as good as it gets for the test-side fix.
Please test with all current GCs? We have seen regressions in this test with
Shenandoah previously.
-------------
Marked as reviewed by shade (Reviewer).
PR Review: https://git.openjdk.org/jdk/pull/15368#pullrequestreview-1593149095
