For a time, the best FLOSS issue tracking system around was Redmine. Then, a couple of years ago, Redmine was forked by some of its core developers to create Chiliproject. They felt that they weren't able to get new features committed in a timely manner and that the existing system wasn't sufficiently open for product growth. The issues are described on Chiliproject's Why Fork? page.
So what happened? Did the community embrace Chiliproject, and leave Redmine behind?
It doesn't look like it. The biggest loss for the fork appears to be Eric Davis' departure. He was the project lead, and did a lot of work on the project. He detached himself from the community, and even shut down his paid support. Although there is a new project lead, there hasn't been much project activity over the past year. Chiliproject development has remained relatively flat, while Redmine activity has been steadily increasing. For the statistics, take a look at the Ohloh data and graphs.
It's unclear, but it could be that Redmine's governance structure has changed to allow for more development at the same time that development on Chiliproject has stagnated.
As far as feature comparisons go, Chiliproject is maintaining a table over at Differences Between Chiliproject and Redmine, but if you look at the wiki page revision history, you'll notice that it hasn't been updated in close to a year. Perhaps the current differences are best summed up in a blog post by deepHill Communications, where they provide a point-by-point explanation of why they're switching back to Redmine.
I'll keep a close eye on the situation, but for now, I think I'll recommend Redmine for all of my clients.