Replies: 3 comments 8 replies
-
I think it's reasonable to upgrade to JDK 11, as most developers are using that (or newer) for new Java projects. Existing projects should have no problem continuing to use older versions of this plugin. Even users targeting older JREs can still build their projects in 11 or newer JDK, with the FWIW, I checked all the dependencies for JSDT (only checked one class file per distinct package/directory, so it was faster to check the whole 30MB) and there was nothing using anything newer than Java 8. However, that could change. In fact, it's likely to change over time. I think it's reasonable to move forward now, when it's a casual choice, rather than later, when it becomes an urgent one to avoid a breakage. So, I'd be strongly in favor of bumping to 11. We can communicate the change with a major version bump in the plugin, if it helps. If we find a critical bug in a previous version (to fix an actual bug, not just support a newer Eclipse release) and there's demand for it, we can always patch and release a bugfix for the last release that supported Java 8. If users don't want to use Java 11, they can use the older plugin version. If they do, then they can use the newer plugin version. No big deal. |
Beta Was this translation helpful? Give feedback.
-
Well looks like we are at jdk 11 already... Error: Failed to execute goal net.revelc.code.formatter:formatter-maven-plugin:2.17.0:format (default) on project waffle-jna: Execution default of goal net.revelc.code.formatter:formatter-maven-plugin:2.17.0:format failed: An API incompatibility was encountered while executing net.revelc.code.formatter:formatter-maven-plugin:2.17.0:format: java.lang.UnsupportedClassVersionError: org/eclipse/jdt/core/ToolFactory has been compiled by a more recent version of the Java Runtime (class file version 55.0), this version of the Java Runtime only recognizes class file versions up to 52.0 |
Beta Was this translation helpful? Give feedback.
-
we are on jdk 11 now due to underlying Eclipse forcing us up. |
Beta Was this translation helpful? Give feedback.
-
This discussion is intended to start discussing if we should move to jdk 11. I'm actually against this one given how big our community is as it could result in inabiilty for users to upgrade.
The situation currently is that Eclipse moved to jdk 11 with 2020-12. Not all libraries actually use java 11 as a requirement yet but this is a problem with how we do javascript (jsdt-core). We have to pull and rebundle that from p2. As a result, any update post our current point requires osgi with java 11. The size of that jar we rebundle from class files not source files is around 30mb making it rather hard for us to currently know if there was code tied to java 11 bytecode. I'm sure there are some tricks we could explore but not sure if its worth it.
Now, from a dev ops standpoint, in my daily work, I'm responsible for well over 500 java builds. At this time, we are currently north of 85% on jdk 11 or higher builds. I don't know how typical that is in the community overall and clearly even in my own angle, there is 15% that woudl not benefit if we forced up to jdk 11.
Some plugins have already moved to jdk 11 requirement, not a lot but some have. We could keep both active as well. That said, lets discuss this...
Beta Was this translation helpful? Give feedback.
All reactions