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
Use oracle-actions/setup-java
for cross-version workflow
#2658
Conversation
6af3093
to
a566f73
Compare
Just wondering: what are the benefits? |
The benefits are mainly going to the source and using early access builds earlier than when using |
This is actually incorrect. You can get easy access builds using |
I don't see any incorrect parts in my statement 😉 I didn't say it cannot be done with
If you check the code changes, there is nothing saying that 19 and 20 are EA builds, there is just a Also mentioned in their README:
Considering the target of our @sormuras did I get the idea behind it correctly? |
Testing with EA is cool, however, I believe it would be beneficial to add tests with JVMs from different vendors or even different JIT. I've explained the suggestion in #2660 |
In addition to 20-EA+1, 19-EA+26 builds are also missing from Azul. Both set of builds were released last Thursday on jdk.java.net. But more importantly than the lag and/or the missing EA builds is the fact that the jdk.java.net builds are aligned with the upstream OpenJDK development, ex. an issue should be logged against a specific jdk.java.net build, not against another vendor's build. |
@vlsi What about actively participating in the OpenJDK Quality Outreach program? |
Hey @delabassee, I was looking into it a long time ago but got distracted and lazy so never completed it 🙂 Happy to get it done for AssertJ, I'll get in touch with you separately. |
I read it as they aren't available, that is my bad.
I went through your workflows and I realized that you are not testing with Java 8 nor with Java 11. Am I seeing this wrong, or is the goal of the AssertJ project not to validate against those Java versions?
I completely, understand your position @delabassee. This is not the right place to discuss this, but with you not providing the Oracle JDK through |
Dropping the older LTS version has been a conscious decision we took, mostly related to Javadoc stylesheet compatibility issues between different LTS versions. That's why now AssertJ requires Java 17 for the build (daef040). In general, our experience is that the problems are never related to older LTS releases, so we accepted this as a reasonable compromise. |
Here's a set of issues where 1.8 produced invalid classfiles when code used type annotations: |
I don't get your point: is this about the Java version AssertJ users would use, or about the version used to build the AssertJ artifact? Maybe I should have said that AssertJ problems are usually not related to older LTS releases. |
I missed that and I didn't see it in the readme. I completely agree about building, but I would expect for AssertJ to still test that the AssertJ build artifact (with JDK 17) is still working properly when testing with JDK 11 and JDK 8. I haven't used the latest AssertJ versions, but it is something that might be useful for the project to validate that it works for your users on older JDKs. |
I fully agree, however that goes more in the direction of integration tests, something that it's currently not very flexible with the current Maven setup and would collide with the limitations I mentioned before. We plan to revise the entire module config in #2424 with a special focus on integration tests. |
@delabassee , what is the way to file a ticket for a bug in javac? The issue is reproducible in Java 17, 18, 19, 20. The test case is as follows:
Expected: AssertJ should compile
|
@vlsi To fill a ticket on https://bugs.openjdk.org, one of you needs to have the OpenJDK Author role. |
I've filed https://bugs.openjdk.org/browse/JDK-8288590 for tracking |
No description provided.