-
Notifications
You must be signed in to change notification settings - Fork 368
Update versions of dependencies in pom.xml #5508
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
Conversation
Signed-off-by: Arjan Tijms <arjan.tijms@omnifish.ee>
Signed-off-by: Arjan Tijms <arjan.tijms@omnifish.ee>
|
Hi Arjan, |
The OSGi tests were already failing before I made any change. In case it matters, they were failing on my M1 Mac locally. |
Odd. The tests were run on my local env and on the CI env for the PRs. Any chance you run the tests in the IDE (so that they were not excluded by maven)? |
Nope, just "mvn clean install" on the command line. I could try to go back to the rev before I made changes and try again though. |
|
Regardless btw, I think for a new major version we should surely not withhold any version updates because something breaks. Instead, whatever breaks should be fixed then (at a later time perhaps, but should be fixed). |
|
@jansupol I reran the build from current 4.0 branch (without any changes), and got the following results: At a glance it seems to be related to Moxy, which I had to remove from GlassFish a long time ago. Lukash added something to EclipseLink, which caused two of the same providers on the class path, and I had to remove one. See eclipse-ee4j/eclipselink#1380 If the test is dependent on Moxy, and uses GlassFish as a test target, it should have failed on your side too I guess. |
|
Details: |
Signed-off-by: Maxim Nesen <maxim.nesen@oracle.com>
|
evidently |
|
@arjantijms I tested the branch on my M1 mac (JDK 17, maven 3.8.7, clean clone and build), too. All tests pass fine. That's 3 different environments that are passing for me. Given
I agree for Jakarta EE dependencies, but not for OSGi support dependencies that can prove to be broken.
No, the tests do not use Glassfish, just Felix to check the OSGi headers in runtime, so that we do not have issues with integration to any OSGi env (there are customers using OSGi servers without Glassfish). To proceed further, I'd go just with those Grizzly & Servlet changes, without the OSGi changes, which should be in a separate commit, especially given how much trouble the OSGi causes. I can volunteer to do so if your test issues persist. |
jansupol
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I can see Maxim reverted the OSGi test module exclusion, all good now!
senivam
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
OK, so we can merge it then :)
That's odd indeed, especially when your test environment is quite similar to mine (Mac and M1). I keep having these: When time allows I'll try do dig into these.
Okay thanks! :) |
Temporary disabled the osgi test module. It already failed on the current state of the 4.0 branch.
We should of course fix and activate it later again.