Good explanation Kevan.
Seems like it makes sense to be a sub-project. IIUC there isn't a very
tight coupling between Geronimo and GShell since the Geronimo specifics
are in the command modules. So there is value to be gained in GShell
being a subproject.
Joe
Kevan Miller wrote:
On Oct 26, 2007, at 10:35 AM, Prasad Kashyap wrote:
I don't see why we shouldn't. But can someone more informed please
list the pros and cons.
Here's my list:
Pro's
* Easier for other projects to reuse GShell
* Release cycle not tied to Geronimo server release cycle
Con's
* Small overhead for being a separately released project --
documentation, release voting, etc
* Separate source tree can complicate debugging (can make the
counterpoint that debugging GShell is easier...)
The Geronimo tx-manager components (transaction and connector) is
another example where we've done this. Note that prior to (or concurrent
with) voting on our last two releases, we've been voting on a tx-manager
release. Although it need not be that way, we're falling into a
lock-step release cycle...
I assume that Guillaume is interested in using GShell outside of
Geronimo. I assume that there will be others...
I'd support GShell as a subproject...
--kevan
