RFR: 8260366: ExtendedSocketOptions <clinit> can deadlock in some circumstances [v3]
Jaikiran Pai
jpai at openjdk.java.net
Thu Feb 18 05:04:42 UTC 2021
On Thu, 18 Feb 2021 04:31:56 GMT, Vyom Mani Tewari <github.com+4410404+vyommani at openjdk.org> wrote:
> What about improving the test little bit, your test wants to load both classes at the same time. Please have a look on modified test.
Hello Vyom, I think that's a good suggestion to use a latch for deciding when to trigger the classloading. I've taken your input and have made some relatively minor change to the way that latch gets used and updated my PR with that change. The latch now waits for both the tasks to reach the point where they are going to do a `Class.forName` on their respectively class names. This should make the test trigger that classloading in separate threads as simultaneously as possible.
-------------
PR: https://git.openjdk.java.net/jdk/pull/2601
More information about the net-dev
mailing list