I was advocated to bump to JDK 22 last year, with FFM as a main reason to replace sun.misc.Unsafe [1], so of course I endorse this. The main rebuttal was that companies prefer to use LTS versions (although any distributor can declare any version as LTS), so I wonder if these considerations still take precedence or if FFM is too important to wait with.
- Nir [1] https://mail.openjdk.org/pipermail/openjfx-dev/2023-December/044081.html On Wed, Oct 2, 2024 at 5:45 PM Kevin Rushforth <kevin.rushfo...@oracle.com> wrote: > All, > > Even though we build JavaFX 24 binaries with JDK 22 (and soon will build > with JDK 23) as the boot JDK, the latest version of JavaFX still runs > with JDK 21, although it isn't tested with older JDK versions. In order > for JavaFX to be able to use newer JDK features, such as FFM (Panama), > we need to increase the minimum version of the JDK that can run the > latest JavaFX. Additionally, there is an ongoing cost to keeping JavaFX > buildable and runnable on older versions of Java, and very little reason > to continue to do so. > > To this end, I propose to bump the minimum version of the JDK needed to > run JavaFX 24 to JDK 22. I filed JDK-8340003 [1] to track this and > prepared Draft PR #1588 [2]. This will *not* affect update releases of > earlier versions of JavaFX (e.g., JavaFX 23.0.NN or JavaFX 21.0.NN), > which will continue to run with the same minimum JDK that they run on > today. > > The main driver for this is that we need to convert the memory > management methods used by Marlin from sun.misc.Unsafe to something > else, both for Java2D and JavaFX, and the natural choice is to use FFM > (Panama), which is what will be done for Java2D. We want to do the same > for JavaFX, which requires bumping the minimum to JDK 22. See > JDK-8334137 [3]. > > NOTE: this will not be an invitation to do wholesale refactoring of > existing classes or methods to use newer language features (e.g., a PR > that refactors existing switch statements and switch expressions into > pattern-matching switch expressions would not be welcome). Rather, this > can be seen as enabling judicious use of new features in new code, much > as we did when we started allowing the use of "var", records, and > pattern-matching instanceof. > > As a reminder, our stated position is that: A) we ensure that JavaFX N > runs on JDK N-1 or later; and B) we encourage developers to use JDK N to > run JavaFX N. It follows from this that if developers want to run their > application on an LTS of the JDK, they should also get a corresponding > LTS of JavaFX. > > Up until now we've been pretty conservative about bumping the minimum > JDK version, and we've chosen an LTS version. However, this has never > been a hard requirement nor guarantee; whether or not the minimum > happens to be an LTS should not be consideration. In the future, we > could consider bumping the minimum version more automatically to, say, > JDK N-2, which could be done fairly shortly after the fork for each new > feature release. This proposal doesn't do that, but we could have a > follow-on discussion as to whether to consider that. > > Comments are welcome. > > -- Kevin > > [1] https://bugs.openjdk.org/browse/JDK-8340003 > [2] https://github.com/openjdk/jfx/pull/1588 > [3] https://bugs.openjdk.org/browse/JDK-8334137 >
