Hi Timur,

I worked with a client where we moved the library functionality from Maven to RAM. This is not simply 'using RAM as a Maven repository' - it was a systematic effort to move away from Maven to RTC and RAM over time, which ended up very successful.

The first step we took was to create build definitions in RTC for the Maven projects so that a developer could trigger a Maven based build from RTC. This was really very painless, the only problem or two we had being when one of the teams had tried to make Maven do something tricky and non standard. We found we were always able to rewrite the project to be simpler and easier without the trick.

We then changed the continuous integration steps in Maven so that they also added the built components to RAM. This was done with the Ant library that comes with RAM that we also use for BuildForge and RTC. It was pretty straightforward.

The final part was modifying the Maven steps to retrieve the correct versions of build components from RAM. Again we used the Ant library that comes with RAM to do this.

At this point in time, while Maven was still adding build components to its own library, all the actual builds were being done with components from RAM, and were putting the finished build components into RAM.

I don't know if they ever turned off the actual Maven repository or not but it made no difference - RAM had become the definitive library for build and deploy.

See also https://jazz.net/downloads/rational-asset-manager/milestones/7.5.1.2M2?p=news

Can you shed some light on what was the tricky things that Maven had hard time to do, and what was it that RAM provides to them on top of Maven?