I agree with FrostWorks and Dark Dimension. I only published a Linux version of my bowling game because one person asked for it, but I didn't have a linux machine and had no idea if it ran or ran well (in fact, I only tested it for the first time yesterday after installing linux on a 2009 MacBook Pro).
Which brings up the other issue, if only one or a handful of linux users are going to download the game, then is it really worth the time to not only build, but get it to build (maybe there are scripting differences/limitations for linux, maybe some third party libraries or content aren't compatible), maybe you have to adjust for performance or controls (I had to separately tweak controls on Windows and Mac), and if you end up having to do builds and uploads and updates and support just for that small audience, it may not be the best use of a dev's time, and for indie devs, time is the scarcest resource.