Philippe Marschall

Thanks. I'll try to have a look later this week.

I have a hard time reproducing this. I added two tests for the behavior you describe in 16e9d035e4d20404b08de9402c50e8a04770b85f. Both with a `FileChannel` open for `WRITE`, one with an additional lock...

You're correct. It looks like we are running into [HHH-17878](https://hibernate.atlassian.net/browse/HHH-17878) which is only [fixed](https://github.com/hibernate/hibernate-orm/commit/ddb23247368599e0d8da5e90e8c1b2700889984d) in 6.5. The 6.3 tests only pass because they run with 6.5, they would fail with...

I think I found a solution. I ran all three `H2BatchSequenceGeneratorTest` in my IDE and they now pass. I also updated the 6.3 submodule as I assume it is intended...

> Is there a typo or two in https://github.com/marschall/Seaside/blob/master/repository/Seaside-WebComponents-Core.package/WAHeadlessComponent.class/README.md ? Probably a rendering issue by GitHub. Check out the raw content https://raw.githubusercontent.com/marschall/Seaside/refs/heads/master/repository/Seaside-WebComponents-Core.package/WAHeadlessComponent.class/README.md > Also, I wouldn't mind if the classes...

How would you switch them off for production? Different canvas class? `#ifTrue:` in component code? Configuration attribute?