Hello JUG leaders...

Java Champion Kirk Pepperdine blogs in the TSS Community about Sun's Mark Reinhold's blog entry about how the JDK source code has been kept in the past at Sun; and, the need for a globally distributable Network File System for potential committers in the Open Source community. Mercurial was a solution and a "lesson-learned" from the OpenSolaris Community.

http://www.theserverside.com/blogs/thread.tss?thread_id=43000

Kirk *Kirk Pepperdine* Java Champion

JDK source to be held in Mercurial?

Posted by: Kirk Pepperdine <http://www.theserverside.com/user/userthreads.tss?user_id=19973> on November 09, 2006 DIGG <javascript:location.href=_parseDiggURL();> If you are speculating about the details surrounding the Sun's <http://www.theserverside.com/news/thread.tss?thread_id=41760> efforts to open source their implementation of the Java specification than there are clues in various blogs. For example, take the important issue of Source Code Control <http://en.wikipedia.org/wiki/Source_Code_Control_System>. Once Sun Java is open sourced, how will all the bits get stored and versioned? To answer that question all we need to do is look at Mark Reinhold's <http://blogs.sun.com/mr/entry/openjdk_scm> blog.

The blog starts with a brief history of how the JDK source code has been kept in the past. The internal product developed and used by Sun engineering is known as Teamware. Though Mark starts by pointing out the advantages of using Teamware over a more tradition SCC system such as CVS <http://en.wikipedia.org/wiki/Concurrent_Versions_System>, he also recognizes the need for an alternative solution.

"...The main trouble with TeamWare is that it depends upon Network File System <http://en.wikipedia.org/wiki/Network_file_system> (NFS) for distributed operation. This works reasonably well so long as you're not too far from the repository..."

As you can imagine, NFS is not so well suited for a project that consists of contributors that will be strewn across the globe. Mark points out that this and others problems could be solved but in the end concluded that is was time to move on. It was at this point that the team decided to utilize the investigative work done by the OpenSolaris tools team <http://opensolaris.org/os/community/tools/scm/>. They didn't want to blindly follow the findings of that team but rather to use it as a starting point. In Marks words,

"...after doing some research and experiments of our own we are strongly leaning toward Mercurial..." <http://www.selenic.com/mercurial/wiki/index.cgi>

The team is interested in your feedback and to be sure there are already a number of very thoughtful opinions posted. Mark is open to your ideas so your comments on his blog will not go unnoticed. And, if you have experience with Mercurial than I'd like to hear about them right here.