[jdk13u-dev] RFR: 8198540: Dynalink leaks memory when generating type converters

[Attila Szegedi]

On Thu, 4 Mar 2021 09:06:52 GMT, Yuri Nesterenko <yan at openjdk.org> wrote:

>> I would like to backport JDK-8198540 to 13u.
>> 
>> The patch doesn't apply cleanly due to minor context differences, namely that [JDK-8251538](https://github.com/openjdk/jdk/commit/4b1b547020effb72ba9ef92807fec39a0db99973) is not in 13u. This was a purely code refactoring task, which among other things refactored some anonymous inner classes in `TypeConverterFactory.java` and `ClassMap.java` to lambdas. All the lines that failed to apply cleanly were deletes, so after this change these differences no longer even exist.
>> 
>> Additionally, 13u did not have running of tests in `test/jdk/jdk/dynalink` enabled as part of the `jdk_other` test group, so a relevant single line of [JDK-8252124](https://github.com/openjdk/jdk/commit/97f8261e4190b8edf83c1d8f11ea57f6c8338284#diff-687d4f32a29eeafbc88c0807e48388ea3e3b16dee622db0fef1473cb387f22ac) was also added to `test/jdk/TEST.groups` to enable the two tests that are part of this change to run.
>
> There should be a better reviewer but time is nigh: ship it!

Thank you @yan-too ! I'll shortly follow it up with a backport of the change that makes the tests in this PR more reliable.

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

PR: https://git.openjdk.java.net/jdk13u-dev/pull/127