rave-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chris Geer <ch...@cxtsoftware.com>
Subject Re: Reintegration of MongoDB Branch
Date Thu, 03 Jan 2013 19:27:50 GMT
On Thu, Jan 3, 2013 at 11:49 AM, Matt Franklin <m.ben.franklin@gmail.com>wrote:

> On Fri, Dec 28, 2012 at 10:54 AM, Matt Franklin
> <m.ben.franklin@gmail.com> wrote:
> > The MongoDb branch is about ready to be re-integrated into trunk.  I
> > am currently going through and fixing merge conflicts caused by the
> > model-interface split.  Currently, the only changes to the current
> > working code are some application context improvements and additional
> > configuration attributes.  In order to ensure that the jpa & mongo
> > modules were swappable, I tried to isolate the changes as much as
> > possible.  Because of this, I think the risk to trunk is pretty low.
> >
> > Assuming no objections, I will merge it back to trunk before the end of
> Monday.
> Of course, I forgot Monday was New Year's eve.  I have completed all
> of the work now and committed the branch back to trunk.
> The vast majority of changes are isolated into the rave-mongodb
> project.  I did add new build profiles for rave-portal and rave-portal
> resources so that the entire rave-project can be built with MongoDB as
> the default persistence provider.  However, the default profile is JPA
> and the rave-portal & rave-portal-resources releases will still be set
> for JPA persistence.
> I tested a clean build of the JPA version of rave and saw no issues
> with the merged code.

Matt, quick question for you. I noticed that the group table (collection)
doesn't just contain a list of user IDs but instead contains complete user
objects. What happens if the user object stored in the group and the user
object stored in the users table get out of sync? Can that be simplified to
just store the IDs?

Have you done any tests using multiple providers for different entities?
For example using Mongo for Users but JPA for Pages? The model split should
allow that but I don't know if the providers will allow for that.

Can't wait to try this out!


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message