Ok,
This is a really good start, but what is in CVS isn't really what I was
thinking of entirely.
The issue is that there really should not be any .jar files in the TDK's
CVS. The whole process should work more like this:
cvs co turbine
cvs co tdk
cd tdk/build
./build-tdk.sh package-all
What would then happen is that the TDK would build itself by first trying to
build Turbine and then copying all the necessary .jar files from Turbine
into a distribution for itself (ie: tdk/tdk-1.0.tar.gz).
Of course you would leave the files like what is in tdk/bin and tdk/share/
since that is specific to the TDK only.
The point is that the TDK can use whatever version of Turbine you want and
the TDK itself would be version controlled with specific versions of
Turbine. Not only that, but since the .jar files are all in Turbine, that is
only one single place that we need to do updates. In the future when Turbine
is tagged in CVS, if you want to build the TDK with a specific version of
Turbine that is easy by simply checking that version out of CVS.
I would go so far as even requiring that Tomcat/ServletAPI be checked out of
CVS as well and simply building it and grabbing the servlet.jar and
webserver.jar files.
Does this make sense?
-jon
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
Problems?: [EMAIL PROTECTED]
