That's fine. And I'm fine with you using geotools packages on the
geoserver branch. It should just stay on the branch, for integration
we should put them in gt2.
Ah yes, sure, I was always, implicitly, talking about our branch only!!!
I think David's plan is to for GeoServer to truly be web only, do
everything else in GeoTools. I support this as well. If
there is some
super unique functionality, or some untested stuff that geotools
doesn't want, then it can live in geoserver. But we've always strived
to get as much as possible into gt2. At the time we made our global
classes, like Repository, it wasn't clear how much they'd be needed in
gt2. But with udig we now have diverse requirements, and the time is
definitely due.
Yes, I always thought that way too, I just wanted to be sure.
The big movement I've been trying to promote for awhile is to
have trunk
(which will start as 2.2.x) _always_ be stable, and to do RnD stuff
only on branches. When too much RnD stuff happens all at once we get
nasty situations. It's not totally sure this will work better, there
are other risks. But we'd like to have trunk be nice and stable, any
RnD work must get their branch totally stable, and when it rolls in
there should just be a day or less of instability. Other RnD branches
may experience instability longer, since it may take them longer to
align. But that's our current plan.
For what it counts I like this plan. To be honest I always thought that
trunk has always been stable, I was surprised when I discovered it was not..
Yeah, that was a mistake. We're hoping to correct it soon. But it's
tough given all the constraints.
...don't worry at all about that!!!
Now the situation seems clear to me, so we can go on working.
You'll have to ask David or James. I'm not up on svn, I still haven't
figured out checking out.
I'll write directly to David about this. Thank you anyway!!!
Chris
Bye Paolo