blueMarine, seeing the light at the end of the tunnel
After realizing that Ant and the software factory I had built on top of it had become unfit for managing my cluster of projects, about six months ago I decided to switch to Maven. I knew I was going to open a can of worms, as there are fundamental differences between the two systems, I was abandoning a tool that I was proficient with and moving to another that I didn't know at all; this was going to be applied to a dozen of projects, hundreds of modules and thousands of classes. In the meantime, I also had to substantially optimize my software factory because of test runs taking too long (thus breaking the meaning of continuous integration), to work on paid duties and to respect some constraints as a few of the projects involved are dependencies to some paid duties.
Of course the ultimate reason of my cluster of projects is blueMarine, the funniest of my projects, but I knew that, being it the top of the pyramid, I'd have been forced to stop working on its core until the conversion was completed. Yes, I have been to able work on other components in the meantime - the simpler one got mavenized in a short time - so in the end the development was never stopped, but they are not as fun as blueMarine itself.
Last but not least, in the latest two months I've had an incredible sequence of small "take-one-hour-for-a-five-minute-job" troubles, including two laptop disks broken and two hacker attacks to my sites.
So you can figure it out how I felt yesterday when, with the exception of a secondary module which triggers a compiler fault, I was able to compile and run the mavenized core of blueMarine! Now there are some broken test and missing resources, but getting back in having a yellow light on Hudson after so many months is refreshing.
The screenshot below is the first I've seen after being able to run for the first time the mavenized project.