Drop Java 17 support

[cdietrich]

we are considering dropping Java 17 support. Is there anyone still using Xtext on Java 17?

cc @iloveeclipse

[iloveeclipse]

Not Advantest. We are at 21 minimal now with latest platform.

[mmosconi1]

We just went to Java 17 in a DSL-based application runtime (not an IDE, no UI) ...
It has a minimal Xtext footprint of

• org.eclipse.xtext
• org.eclipse.xtext.common.types
• org.eclipse.xtext.util
• org.eclipse.xtext.xbase.lib
• org.eclipse.xtend.lib

Do you think, a subset like this or "org.eclipse.xtext.runtime.feature.group" could keep Java 17 compatibility?

[szarnekow]

What timeline would you have in mind @mmosconi1
How often do you update the Xtext version?

[mmosconi1]

Given that Java 17 is supported until 2027/2029, it may be that we need to use/support it at least until 2027, I guess.
(for the mentioned non-UI runtime parts, as mentioned above)

We just updated to recent Xtext 2.40, so we are in a good state currently.
But of course we we would like to catch up with newer versions, if possible.

[iloveeclipse]

Given that Java 17 is supported until 2027/2029, it may be that we need to use/support it at least until 2027, I guess

Java 8 is supported even longer, so what?

Oracle support of some Java version not argument.

The argument would be: "my organization uses Java 17 till 2027 and we are willing to provide all resources needed (manpower) for Xtext project maintenance in this time".

[cdietrich]

I assume the eclipse jdt version that supports 25 will require 21 to run
This will makes supporting 17-21 complicated

[iloveeclipse]

I assume the eclipse jdt version that supports 25 will require 21 to run This will makes supporting 17-21 complicated

Sure.

Honestly, if some organization needs Java 17 support and Xtext will move to Java 21, there are few alternatives available.

• Use older Xtext and backport whatever patches needed
• Provide engineers working on Xtext maintenance
• Spend enough money for release engineering / maintenance
• TBC

With given project setup Xtext must move forward and reduce maintenance costs as much as possible at same time.
The only solution is consequently raise minimal required Java / Eclipse versions and throw away as much ballast as possible.

[szarnekow]

Thank you for sharing your thoughts on that, @iloveeclipse - there's undoubtedly truth to that.

We've to weigh the options at hand, that's for sure. I'm still curious whether other people share their views on things here.

[mmosconi1]

I actually just answered the original question:

Is there anyone still using Xtext on Java 17?

We don't expect anything here, just wanted to give feedback.

[szarnekow]

We concluded that Xtext 2.43 will be compiled with Java21 bytecode and no longer support IDEs running with Java17