[jdk8u-dev] RFR: 8280963: Incorrect PrintFlags formatting on Windows

[Andrew John Hughes]

On Wed, 20 Apr 2022 21:01:34 GMT, Alex Kasko <akasko at openjdk.org> wrote:

> A fix to 8u-only Windows-specific problem with -XX:+PrintFlags* output.
> 
> Problem description on Stack Overflow by Andrei Pangin: https://stackoverflow.com/a/63309395
> 
> The problem was fixed in jdk9 as part of [JDK-8042893](https://bugs.openjdk.java.net/browse/JDK-8042893), but this [jdk9 change](https://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/115188e14c15) doesn't look like a good candidate for the 8u backport according to [8u Best Practices](https://mail.openjdk.java.net/pipermail/jdk8u-dev/2020-June/012002.html). Still the problem manifests itself as a customer-visible bug on Windows because people do use -XX:+PrintFlagsFinal to find out the actual flags picked up by JVM and can be confused by unexpected zeros in the output. Proposed patch cherry-picks the minimal change from [JDK-8042893 commit](https://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/115188e14c15#l53.1).
> 
> Mailing list review: https://mail.openjdk.java.net/pipermail/jdk8u-dev/2022-February/014532.html
> 
> Testing:
> 
>  - [x] regression test is included with the proposed patch
>  - [x] checked that it builds on Windows 64-bit and 32-bit on VS2017 and VS2010
>  - [x] ran jtreg:hotspot/test/runtime on Windows and Linux

This mostly matches the changes to this file in JDK-8042893, but it doesn't remove [the pragma](https://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/115188e14c15#l53.7)

Can we include that change as well to make sure the code is now compiling without warnings?

Also, the PR needs to be rebased anyway to fix the GHA run on Linux (bad dependency issue now fixed by [JDK-8284772](https://bugs.openjdk.java.net/browse/JDK-8284772). The Mac failure is more complicated (see #26) but is at least compiling the code before failing. Despite the title, this changes code shared by all platforms.

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

PR: https://git.openjdk.java.net/jdk8u-dev/pull/45