Re: [hibernate-dev] Supporting timezone in Timestamp Type

2016-08-31 Thread Gunnar Morling
> it is a huge mistake to save the specific TZ differences to the DB I once was working on a globally deployed shipping management application where that was a requirement. Say there is a shipment from Austin (one TZ) to London (another TZ). In that application an operator based in Hamburg (yet a

Re: [hibernate-dev] Centralized access to "bootstrap only" resources

2016-08-31 Thread Gunnar Morling
> Do we want to change JpaIntegrator#integrate signature to pass its context > as a parameter object If you think that's the better approach, I'd say 6 is the right time to do it. Users (or integrators) will be prepared for this sort of breakage in a major. > do other projects supply custom > org

Re: [hibernate-dev] Centralized access to "bootstrap only" resources

2016-08-31 Thread Steve Ebersole
One contract I would need to adjust for this is org.hibernate.service.spi.SessionFactoryServiceInitiator#initiateService. I can find all the implementations of that in ORM, but do other projects supply custom org.hibernate.service.spi.SessionFactoryServiceInitiator impls? On Wed, Aug 31, 2016 at

Re: [hibernate-dev] Supporting timezone in Timestamp Type

2016-08-31 Thread Steve Ebersole
We discussed this on HipChat, but for the benefit of all on this discussion... Part of this (the original report) speaks to a difference in how we map (org.hibernate.type.Type) java.time temporal versus a java.util temporal - specifically java.util.Calendar. When we are passed a Calendar (the mod

Re: [hibernate-dev] Supporting timezone in Timestamp Type

2016-08-31 Thread Emmanuel Bernard
On Wed 2016-08-31 17:07, Vlad Mihalcea wrote: > Hi, > > While reviewing the Pull Request for this issue: > https://github.com/hibernate/hibernate-orm/pull/1536 > > I realized that we can improve the default TimestampType as follows: > > 1. We could make it parameterizable so that it can also tak

[hibernate-dev] Supporting timezone in Timestamp Type

2016-08-31 Thread Vlad Mihalcea
Hi, While reviewing the Pull Request for this issue: https://github.com/hibernate/hibernate-orm/pull/1536 I realized that we can improve the default TimestampType as follows: 1. We could make it parameterizable so that it can also take a custom timezone (UTC) during mapping. 2. We could also def

Re: [hibernate-dev] Centralized access to "bootstrap only" resources

2016-08-31 Thread andrea boriero
I'm fine with combining native and JPA events handling, about the second point, ideally I would change the signature but due to the problems you listed I vote for the in-line solution. On 30 August 2016 at 19:20, Steve Ebersole wrote: > Any thoughts on the JpaIntegrator parts of the discussion?