Steve, please let's not mix up stuff here. I maintain Spring Data JPA, I report builds fine on Beta 2, I report it doesn't on the snapshots. I take time to investigate intermediate snapshots. I single out for commits that cause the change in the otherwise unchanged reference project. If that's what you consider "lazy", so be it. I am sorry I reported an error that other people might run into, too.

My reference project has test cases that bootstrap multiple EntitiyManagerFactories in top of the very single database instance (by the very nature of the way Spring integration tests work). So a second bootstrap of the EMF basically "sees" the schema which the first run created. That's why I think the check whether a column exists during the second run doesn't get the quotation applied but applies it in the attempt to actually create the column. I am guessing here, but I guess I proved that the error I am reporting here is caused by changes in the quoting. Also, why would the exception mention an already existing artifact being present otherwise if it starts on a clean database.

I am not even saying this related to the other report you got, I just report what I stumble above and try to be as precise as possible. Calling stuff, that you don't like BS doesn't help by any means. Unfortunately removing Spring from a Spring project is sort of impossible.