Sanne Grinovero
Sanne Grinovero
Merged. Thanks again!
> Type mappings have changed between Hibernate 5 and 6, so we should just let it handle it upon install. And, then apply a .sql script with our other requirements...
P.S. I realize this is labour intensive and rather annoying to do, but Hibernate ORM upgrades that require such intervention should be extremely rare.
+1 to have "real" Read-Only connections, seems to be highly requested and we have now some real needs in the form of the Keycloack request. This would need to go...
> I was saying that _mimicking support while not enforcing anything_ for JDBC drivers that _don't support read-only transactions_ is not worth the effort. > > The work you'd be...
> What doesn't work well is that at the time we're starting the transaction we don't know which of the datasources will be the writable one. This concerns me, it...
>So essentially, flip the default in general, but override it in some integration tests that are known to not need it, just to be sure? Exactly: some integration tests or...
> > > Set --allow-incomplete-classpath by default in when Quarkus runs a native compilation > > > > > > You essentially need to avoid passing `--link-at-build-time`, which can be...
@mkouba ah I see what you mean, sorry - but why would you remove it, could you elaborate on the scenario?
Is anyone planning to work on this?