RFR: 8293000: Review running times of jshell regression tests

[Joe Darcy]

On Mon, 26 Sep 2022 17:21:07 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> Some of the jshell tests have been identified as long-running.
> 
> This patch reduces tests runtimes with multiple tricks:
> 
> - local execution control is used in heavily parametrized and fragmented tests to eliminate JVM startups
> - some tests have been defragmented
> - unnecessary Thread::stop has been removed
> 
> Measured test run times before and after application of this patch:
> - ClassMembersTest 23.964ms ->  9.195ms
> - ToolBasicTest 24.281ms -> 14.955ms
> - FailOverExecutionControlHangingListenTest 31.221ms -> 5.780ms
> - FailOverExecutionControlDyingLaunchTest 31.221ms -> 6.050ms
> - FailOverExecutionControlHangingLaunchTest 31.303ms -> 5.832ms
> - CompletionSuggestionTest 10.523ms -> 4.940ms
> - SimpleRegressionTest 3.629ms -> 1.304ms
> 
> Please review the test changes.
> 
> Thanks,
> Adam

While not being familiar with the jshell test framework, the changes looks plausible. Thanks.

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

PR: https://git.openjdk.org/jdk/pull/10428