>From: "Craig McClanahan" <[EMAIL PROTECTED]> 
>
> On 9/18/06, Gary VanMatre wrote: 
> > 
> > > 
> > >> BTW, I'm going to take a try at moving commons valiator out of the 
> > core. 
> > > > I might need some help with the clean up. Does 
> > "shale-commons-validator" 
> > > > sound ok for the project under "framework"? 
> > > 
> > > 
> > > I'd think shale-validator would be better, unless we contemplate 
> > alternative 
> > > validation strategies. 
> > > 
> > 
> > That name would also work if we wanted to add other types of validators to 
> > the 
> > project. 
> > 
> > I almost had this done and realized that we really need to change the 
> > package names. 
> > I started moving the validator stuff out into corresponding packages but 
> > realized that 
> > that wouldn't work for "resources/Bundle.properties". 
> > 
> > What about moving these packages under "validator". 
> > 
> > shale/component 
> > shale/resource 
> > shale/taglib 
> > shale/renderer 
> > shale/faces 
> > shale/view 
> > 
> > shale/validator/component 
> > shale/validator/resource 
> > shale/validator/taglib 
> > shale/validator/renderer 
> > shale/validator/faces 
> > shale/validator/view 
> > 
> > Any thoughts? 
> 
> 
> I presume you're thinking only of moving the "validator" related classes, 
> right? Among other things, the token component would need to stay. 
> 

Ya, I was only going to move the validator related artifacts.  But, that
alone amounts to about 1/4 of the core jar.  Some folder can be moved
(svn mv) into another jar project.  The folder that only contain validator 
code.  
The other would have to be moved at a source file level.  In addition, the
property files and config files would need to be copied and content removed
from both core and the new project.

 
> Doing this means a new tag library for the validator stuff (as well as its 
> own faces-config.xml file). 
> 
> What do anticipate is the endgame? Would we pull the validator stuff out of 
> the release, or try to label it as something not release quality? 
> 

This is in reference to your request to simplify dependencies in the maven
builds.  That's my motivation.  It was a bullet point in your plan.



> I'm not sure I am game for all of this ... if we're willing to label this 
> API as not being stable, I don't see the point in doing the separation. If 
> we are not going to ship it at all, this might be OK. But any existing 
> users would still want the functionality available from the sandbox -- and 
> they would probably prefer it to not require any changes. 
> 

Well, it will require a lot of effort and I kind of agree.  If we have the 
maven dependencies under control, I think our efforts are better spent
elsewhere.  After looking into what it would take, I'm not sure I have
the courage right now anyway.



> Hmm.... 
> 
> Craig 
> 
> 
> > 
> > >> Gary 
> > > 
> > > 
> > > Craig 
> > > 
> > 
> > Gary 
> > 

Reply via email to