RFR: 8269032: Stringdedup tests are failing if the ergonomically select GC does not support it

Kim Barrett kbarrett at openjdk.java.net
Wed Jun 30 06:31:02 UTC 2021


On Tue, 29 Jun 2021 11:09:31 GMT, Per Liden <pliden at openjdk.org> wrote:

>> test/hotspot/jtreg/gc/stringdedup/TestStringDeduplicationTools.java line 73:
>> 
>>> 71:         }
>>> 72:     }
>>> 73: 
>> 
>> I think it would be good to parse the gc string. Otherwise it might be easy to make a spelling mistake or add another argument (before) to the driver without realizing the error.
>
> I think we should just blindly pass this option (or options) to the spawned JVMs, without looking at them.

I don't think there's any point to doing any parsing here, and good reason not to.  If the value isn't sensible as a GC name the test will very likely fail because of an invalid CLA to the subprocess.  And validating that it's a valid GC name involves using vm.gc.GC, which brings in WhiteBox and all its attendant cruft.

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

PR: https://git.openjdk.java.net/jdk/pull/4603


More information about the hotspot-gc-dev mailing list