[jdk11u-dev] RFR: 8244508: JFR: FlightRecorderOptions reset date format
Christoph Langer
clanger at openjdk.org
Fri Oct 27 08:12:37 UTC 2023
On Tue, 5 Sep 2023 11:07:20 GMT, Pushkar N Kulkarni <duke at openjdk.org> wrote:
> Hello all,
>
> This backport request is related to JDK-8244508 [1]. When using JFR (`-XX:StartFlightRecording`), the configured default Locale is not honored, with FormatData or TimezoneNames(for example), because the relevant ResourceBundles do not get loaded when an attempt is made to locate and load them during VM startup (i.e when `-XX:StartFlightRecording` triggers JFR initialiazation). Though [1] quotes a mailing-list post [2] mentioning this issue was not seen with jdk11 then, the testcase in [1] does fail with 11.0.20 (as well as 11.0.22/master branch) as of today:
>
>
> $ java --show-version -Duser.country=DE -Duser.language=de -XX:FlightRecorderOptions=stackdepth=128 Main
> openjdk 11.0.22-internal 2024-01-16
> OpenJDK Runtime Environment (build 11.0.22-internal+0-adhoc.pushkarnk.jdk11u-dev)
> OpenJDK 64-Bit Server VM (build 11.0.22-internal+0-adhoc.pushkarnk.jdk11u-dev, mixed mode)
> expected: 01.01.2020, 00:00:00
> actual: 2020 Jan 1 00:00:00
>
> Testing: Linux x86-64 build, tier1 tests with `release` and `fastdebug` builds.
>
> [1] https://bugs.openjdk.org/browse/JDK-8244508
> [2] https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2020-March/001313.html
OK, looks good. I'll run the backport through our tests and will approve when no regressions spotted.
-------------
PR Comment: https://git.openjdk.org/jdk11u-dev/pull/2121#issuecomment-1782486843
More information about the jdk-updates-dev
mailing list