-
Notifications
You must be signed in to change notification settings - Fork 79
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Build] Disallow Java security-manager in all I-/Y-/Smoke-tests #2659
[Build] Disallow Java security-manager in all I-/Y-/Smoke-tests #2659
Conversation
IMO drop the argument. For #2654 builds will be moved to Java 21 soon as this is a prereq to actually do anything. |
Setting it to disallow now will let us detected problems sooner rather than later... |
c0a15b4
to
c63acd4
Compare
Since Java-18 the default is already
This resulted in the following additional failures. If we not disable it for all builds we will slightly reduce the flood of new failures.
|
The failing tests in Equinox are fixed respectively disabled if setting a security-manager is not permitted with Another reason to not explicitly disallow setting a I have created a PR to refine ANT in this regard but that's of course something we can't use immediately: Therefore we should just stick to the defaults or we'll end up with failing ANT tests for all configurations running on Java-17. There are still a few more ant.tests failing with a disabled security-manger and I'm currently looking into them. |
This should now be fixed with: and together with disabling the security-manager in all tests shouldn't cause regressions anymore. |
Yes, this looks good and is a good step forward. |
Great. Then lets submit this and see if really all I-build tests can handle a disallowed security-manager. |
Part of #2623
@merks do you think we should explicitly disallow the security-manager or just remove the argument to explicitly enable it?
Since Java-18 the default value of the
java.security.manager
property isdisallow
as part of JEP-411:In order to make the builds a bit simpler I would be in favor to just remove the
java.security.manager
argument. With that only the Java-17 tests would run with enabled security-manager and all others would have it disabled. And Java-17 will probably not last long any-ways.Btw. I have replayed a I-build test with explicit disallowed security-manager in order to test in advance if everything falls apart or not (you can see it in the diff):