RFR: 8311990: Two JDI tests may interfere with each other [v2]

Chris Plummer cjplummer at openjdk.org
Wed Jul 31 22:47:31 UTC 2024


On Wed, 31 Jul 2024 19:55:54 GMT, Alex Menkov <amenkov at openjdk.org> wrote:

>> test/jdk/com/sun/jdi/JdwpListenTest.java line 54:
>> 
>>> 52:     // It's off by default as it causes test time increase and test interference (see JDK-8231915).
>>> 53:     private static boolean allowNegativeTesting =
>>> 54:         "true".equalsIgnoreCase(System.getProperty("JDI_ALLOW_NEGATIVE_TESTING"));
>> 
>> That seems a strange name for a system property - it would normally be something like jdk.jdi.allowNegativeTesting. Is this style of naming being used elsewhere?
>
> Renamed as suggested.
> IIRC we had `-Duse.JTREG_TEST_THREAD_FACTORY=Virtual` some time ago. Other tests use "standard" naming

> use.JTREG_TEST_THREAD_FACTORY is not actually a property used by the source. It is just something we add to the mach5 JVM args to show that the test task was done with JTREG_TEST_THREAD_FACTORY=Virtual. It's basically a comment.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/20362#discussion_r1699179887


More information about the serviceability-dev mailing list