[buckminster-dev] Field report

[Thomas Spiessens <thomas.spiessens@xxxxxxxxx>]

Hi guys,
just wanted to give you an update on how we are working with 
Buckminster. We currently have one product (a distributed application) 
we are working on which consists in total of 24 Eclipse projects 
(mixture of regular Java, plugin and feature projects). All of these are 
stored in our Subversion repository.

The developers themselves use the cquery to populate their workspace and 
we have a number of Ant builders running that do some additional tasks 
during incremental and full builds (e.g. packaging and copying JAR files 
to certain locations).

We are using Buckminster to control our build and continuous integration 
process in a headless mode as well.  So far we have been using 
Buckminster to retrieve and build all of the projects, but we are using 
PDE to export the product. We have hooked this headless process into our 
CruiseControl server so we get continuous integration support.

As you see, Buckminster is being used in real life situations :)


We do have a number of things we want to be able to do with Buckminster 
(all of these in the context of a headless build):

1. Being able to export the product using Buckminster (perhaps by a new 
Buckminster command). It is a bit annoying to be using the PDE build 
because PDE expects a particular directory structure of the feature and 
plugin projects. This means that after the materialization we need to 
reorganize the directory structure through a script. If Buckminster 
would be able to export a product, we might not have to do this, making 
our script a bit simpler and easier to maintain in the future.

2. Some of the projects are also dependent on a number of JBoss 
libraries. These libraries become available after register a new JBoss 
server (available after installing the JBossIDe plugin). The way we do 
that now in headless mode is by copying a number of files in the 
metadata folder of the workspace that was created by Buckminster after 
the materialization. We were thinking of writing our own Buckminster 
command for registering that server.

3. Generating Javadoc from Buckminster

4. Running all of our unit test, but we are currently investigating that.

5. We are also in the progress of investigating the actor mechanism.

Keep up the good work,
Thomas.