This is a good question. JDK-8311806 is a P4 bug, so my (non-formal) algorithm would say not to backport this. But then, this clearly can create crashes and it is a simple fix that is not affecting other platforms, so I am very much in favour of backporting it. Note that I raised this point on the jdk-dev list [1] and it is discussed there as well [2]. >From those discussions, it becomes clear that a formal algorithm is most likely not enough, and there should always be room for manual judgement.
In this particular case, I would vote +1 for backporting it. - Johan [1] https://mail.openjdk.org/pipermail/jdk-dev/2024-November/009625.html [2] https://mail.openjdk.org/pipermail/jdk-dev/2024-November/009629.html On Thu, Nov 14, 2024 at 9:10 PM Kevin Rushforth <kevin.rushfo...@oracle.com> wrote: > This does look like JDK-8311806. > > Gluon maintains jfx17u, so Johan can comment on your request. > > -- Kevin > > > On 11/14/2024 2:29 AM, Glavo wrote: > > Hi, > > We received a JVM crash report[1] for our JavaFX application and I noticed > it contained these lines: > > objc[71933]: Class ButtonAccessibility is implemented in both >> /Library/Java/JavaVirtualMachines/jdk-22.jdk/Contents/Home/lib/libawt_lwawt.dylib >> (0x100d6caa0) and /Users/lolimaster/.openjfx/cache/17.0.13/libglass.dylib >> (0x14c4fe218). One of the two will be used. Which one is undefined. > > > *** Terminating app due to uncaught exception >> 'NSInvalidArgumentException', reason: '-[ButtonAccessibility >> initWithEnv:accessible:]: unrecognized selector sent to instance >> 0x6000024fb540' > > > This crash occurred after we bumped the JavaFX from 19.0.2.1 to 17.0.13. > It looks like it's related to JDK-8311806, is it? > If yes, I'd like to request a backport of the patch that fixes it to > JavaFX 17u. > > Glavo > > [1]: https://github.com/HMCL-dev/HMCL/issues/3409 > > > >
