|
|
|
|
|
| (no subject) [message #727354 is a reply to message #727265] |
Tue, 20 September 2011 23:25   |
Eike Stepper Messages: 5152 Registered: July 2009 |
Senior Member |
|
|
Am 20.09.2011 20:44, schrieb Steve Robenalt:
> Hi Eike,
>
> I've been looking at the definition models, which seem to have most of what would be needed, though I don't see an
> o.e.e.cdo.server.defs model with (for example) FailoverParticipantDef or RepositorySynchronizerDef in it, so that
> presumably needs to be added (or am I just not seeing it?).
I'm sure that synchronizable repositories have been developed way after André has stopped working on the definition
models. So yes, all kinds of new constructs wold have to be added, such as the different new repository types, the repo
synchronizer, session connfiguration factories, etc.
>
> I agree that it's best not to reinvent the wheel on something like this. My main concern with the defs model was
> simply that the configuration would end up being a fairly literal reflection of the underlying CDO infrastructure
That was the intend.
> when perhaps something simpler would suffice.
I doubt it. We'd likely end up in the same one way route as with RepositoryConfigurator.
> I'll experiment with it a bit to determine if my concerns are legitimate or not. If they are, I still wouldn't propose
> changing the defs model, but perhaps a simplified configuration model with a Model-To-Model transform into the
> underlying defs model would be appropriate.
Possible that you in particular could go with something simpler, but in general there will hardly be a consensus about
how simple simple may be or not.
Cheers
/Eike
----
http://www.esc-net.de
http://thegordian.blogspot.com
http://twitter.com/eikestepper
|
|
|
|
|
|
Powered by
FUDForum. Page generated in 0.02036 seconds