Gavin King

Look, I found the docs for JBPM-managed conversations in Seam: https://docs.jboss.org/seam/2.3.1.Final/reference/html/jbpm.html

> My understanding was that the boundaries was not clear when it should happen + context propagation as consequences became excessively expensive. So, I mean _at a guess_ there's two...

Agreed. We really don't have a good way to know that the warning is completely harmless.

@FroMage > Pretty sure it's postgres. Not sure if you use `IF EXISTS`. We do (of course) use `IF EXISTS` on Postgres. But I suppose Postgres sends us a warning...

If someone *really* wants this fixed, they could write code to check with the JDBC metadata (or `information_schema`) that the table exists before sending the `drop` statement.

> If someone _really_ wants this fixed, they could write code to check with the JDBC metadata (or `information_schema`) that the table exists before sending the `drop` statement. Actually, this...

I mean: - 00 means "successful completion" and - 01 means warning. So I don't see how a warning can be 00, unless it's just .... wrong, and therefore untrustworthy....

> There may be another way to solve this Yeah, I hope so, since I don't like multiplying the `JdbcType`s like this! > Solution similar to `DB2Dialect`, which was accepted...

See https://hibernate.atlassian.net/browse/HHH-18314. We need to start by figuring out the elegant/simple way to fix that one.

@VladoKuruc If the only difference between `CHAR` and `NCHAR` on Informix is in the default collation order, perhaps the best option is to just define `InformixDialect` with `NationalizationSupport.IMPLICIT`. WDYT?