RFR: 8295239: Refactor java/util/Formatter/Basic script into a Java native test launcher

[Justin Lu]

On Fri, 14 Oct 2022 21:48:06 GMT, Naoto Sato <naoto at openjdk.org> wrote:

>> Issue: Formatter unit tests are launched via basic.sh
>> 
>> Fix: Replace basic.sh with a Java test launcher
>> 
>> Note: Java.internal.math was included in the original configuration of Basic, but I removed it as it was not used within the Basic unit tests 
>> 
>> 
>> Original output on success
>> <img src="https://user-images.githubusercontent.com/67398801/195936541-bc90db50-8d03-47be-9c4f-95176b19a6a7.png" width="350" height="350">
>> 
>> 
>> New output on success
>> <img src="https://user-images.githubusercontent.com/67398801/195936558-f85f4d48-dae2-4c38-aa50-46ef47db3d89.png" width="350" height="450">
>
> test/jdk/java/util/Formatter/Basic.java line 24:
> 
>> 22:  */
>> 23: 
>> 24: import static java.lang.System.out;
> 
> It's not your change, but probably this does not seem necessary, as `out` appears only once.

Got rid of it!

> test/jdk/java/util/Formatter/BasicTestLauncher.java line 48:
> 
>> 46:         runFormatterTests(TZ_UP);
>> 47:         runFormatterTests(TZ_AN);
>> 48:     }
> 
> Could use testng/junit, instead of normal main. This way both timezones are guaranteed to be tested. Currently, if an error occurs with `US/Pacific`, `Asia/Novosibirsk` will not run.

Thank you for the catch, replaced with JUnit

> test/jdk/java/util/Formatter/BasicTestLauncher.java line 74:
> 
>> 72:             throw new RuntimeException(String.format("$$$ Error(s) found within %s subprocess: " +
>> 73:                     "%s%n", timeZone, err.getMessage()));
>> 74:         }
> 
> I'd prefer not to catch an exception (and turn it into `RuntimeException`) here, as the exception type will be lost.

Thank you, separated to distinguish the IO and runtime exceptions

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

PR: https://git.openjdk.org/jdk/pull/10715