Oliver, you run a project. I think you understand the time crunch that goes into a major release. The nature of our respective projects is different in this next particular point so maybe you have not felt it, but I can assure you it gets extremely frustrating debugging bug reports that do not simplify away extraneous BS. To be frank, that extraneous BS exists generally speaking because the reporter is too lazy to remove it. Or they do not have time. Time... there is that word again...

So please understand that when I get 2 very similar reports, each with "tests that reproduce the issue" where each test uses Spring and I spend time looking into one of them and actually cannot reproduce it when Spring is not involved.. understand how reluctant I might be to "spend" more time on that second test.

Also, JPA does not define support for schema update. So what make you think that Hibernate's JPA impl is somehow causing schema update to be run twice?

And again, if the Spring usage here is so minuscule why have it at all? Here is the original reporter's test: https://github.com/paranoiabla/hibernate-hsql-issue. He has merged back the "Hibernate only" test I sent him as a PR. I have yet to see how "Hibernate only" causes this. And ultimately that is what I care about. This is the Hibernate issue tracker after all. If one of y'all can take that "Hibernate only" test and get it to reproduce the error, to me that is a proper Hibernate bug report and I will work on it for 5.0. If not I will work on it and decoding Spring's involvement etc when I have time.