rick rineholt wrote:
A conversation got started on the IRC chat yesterday
that started with the directory structure of the samples and spilled over to the form of down loadable release. To summarize for all this is how it ended:

Samples in svn will be divided by technologies each
will have a "samples" directory and there is also a top level directory (just below tuscany/java) that will have "business samples". The name of the "business samples" was never really decided on. (I thought about it some and came to like "sampleapps" for sample applications, but that just my preference). In each of these samples/sampleapps directories the structure will be flat for now but we may need to change that as the number of samples grows

So roughly...
tuscany/java/sampleapps/bigbank
tuscany/java/sampleapps/petstore
...
tuscany/java/sca/samples/helloworld
tuscany/java/sca/samples/helloworldaxis
tuscany/java/sca/samples/helloworldaxis2
tuscany/java/sca/samples/helloworldaxissvc
tuscany/java/sca/samples/helloworldjsonrpc
tuscany/java/sca/samples/helloworldmc
tuscany/java/sca/samples/helloworldweb
tuscany/java/sca/samples/helloworldws
tuscany/java/sca/samples/helloworldwsclient
etc..
tuscany/java/sdo/samples/myfirstsdo
...
tuscany/java/das/samples/dasstarter
tuscany/java/das/samples/dasboot
+1, I am assuming that we have three groups of technology samples, sca, sdo, das.
We agreed that the package names need to be shortened.
I think we agreed that giving samples even short names
was preferable than a number scheme.
+1, we already have JIRA issues for that.
Now for the release artifacts:
There would be a deliverable download that is just
Tuscany core source, minus svn artifacts. There is a core binary download, and a download per technology samples (i.e. sca/samples, sdo/samples, das/samples). Also one for the sampleapps and once there was more than just BigBank we'll decide then if it needs to be broken up or not.

The core deliverable has Tomcat embedded ready to run
Tuscany with all Tuscany jars added including dependencies and configuration files updated. NO demos, NO samples, NO sample application etc. No unit testcase. I'm pretty sure this does not include Derby needed by just some technologies either. It *may* contain as of yet an undefined *simple* "smoke test". It has one prerequisite: a 1.5 level JRE.

The other down loadable deliverables should include
Tuscany source only. No svn, no maven, and no unit test artifacts. They can include Ant with shell and bat scripts that will build the samples and optionally deploy them to the Tuscany core embedded server. The only prerequisite they should have is a 1.5 level JDK and the Tuscany core.

So roughly the user experience is :
-Have/get JDK/JRE 1.5 level.
-Go to Tuscany download page ... download core
deliverable.
-Unpack core there should be a readme.html at the root
that tells you how to start your server.(go to bin startup and also probably mentions the demos and sample downloads etc)

(If you have your own SCA applications already to run
you just need to deploy them your done)

(Sample users)
-Download any technology samples or sampleapplication
that appeals to you.
-Unpack sample.  There should should be a readme.htm
at the root that mentions the prerequisites guides you through the rest:
-Possibly set some environment variables to point to
core, JDK (JAVA_HOME).
-Run bat or shell script that kicks off an Ant build
of the samples and optionally deploys to Tuscany core Tomcat embedded
server.
-The sample readme.htm  has directions to run any bat
shell script for J2SEE based samples and links to the running embedded core
WEB based samples.




__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com


--
Jean-Sebastien

Reply via email to