Hi,

I just finished working out some final issues with the installer I was
working on for Tambora so that completes the first iteration over the TDK
that allows development in a cvs layout and supports the development process
all the way through to deployment by the end user.

In the case of this particular install postgres was the target database and
the torque schemas were used on the client machine in conjunction with a
client profile to produce a fully configured turbine application. The
databases are created and populated using some new torque code, and the TRP
is fully configured on the client end by using values from the client
profile. Hopefully in the future maybe a nice installer like zeroG or one of
the OSS Swing installers could be used. But the configuration for the client
is isolated to a single client profile and those values are pushed through
the install procedure and used where necessary.

There is a separation between processes which can be shared by all turbine
applications and processes which are project specific so any project should
be able to extend what's provided in the TDK. Tambora uses multiple
databases and generates self-signed certificates and shares the public
certificates, in addition we have a test machine setup so that client
machines with new installs can bounce transactions off the test machine to
verify the install is working correctly.

So I feel now that what's in the TDK is viable for development all the way
through to deployment and I will start checking in changes slowly over the
next few days and on the weekend. There's a lot of them but I wanted to make
sure that everything worked with the test client I was working with before
checking everything in.

As a side note I did get burned by having the create database scripts. The
installer provides everything so that no additional downloads are required
but the <exec> ant tasks requires the "antRun" scripts in order to run
native scripts. I was thinking a little ANT_HOME could be setup by the
install process, but I would rather go the route of getting rid of the
native scripts and use SQL to produce the databases. That was pretty much
the only snag in the install.

Tambora probably has a more difficult install because of the multiple
databases, use of certificates and SSL, and the requirement that a test
machine be know to the install process so I think that the process in the
TDK will work quite well for any existing Turbine project.

I will get the TDK updated over the next couple of days, and than I plan to
use what is in the TDK to make a big sandbox that will generate the sample
TDK app for all known databases, build them, and test them. Tambora has
normally run with Postgres but we now have the requirement for MSSQL and
HSQL (for really easy installs) so I want to get into place tests for
turbine for these combinations first and than move on to MySQL and Oracle.
The TDK currently doesn't test an installed version but this could be added
so that the TDK could be used to test functionality of the different
databases but also make sure the installer process works with all the
different databases.

I will also get Santiago's patch in to try and get Jetspeed closer to
Turbine.

-- 

jvz.

Jason van Zyl

http://tambora.zenplex.org
http://jakarta.apache.org/turbine
http://jakarta.apache.org/velocity
http://jakarta.apache.org/alexandria
http://jakarta.apache.org/commons



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to