Hi Gail,
the fix re-introduces the old behaviour (pre 5.0), the change in 5.0 and
5.1 was an oversight.
On Wed, 6 Feb 2019 at 23:45, Gail Badner <gbadner(a)redhat.com> wrote:
I was looking into backporting HHH-13076 for 5.3.8.
Looking at the fix [1], I see that starting a JDBC transaction that is
already started will fail, even
if hibernate.jpa.compliance.transaction=false.
AFAICT, this didn't happen in 5.1. Was there a reason in particular why
this was done?
Thanks,
Gail
[1]
https://github.com/hibernate/hibernate-orm/commit/a15dfe0e056c55664846b62...