It would be a monumental and incredibly expensive amount of work for a relatively small userbase.
People saying this are missing the point.
The only reason the userbase* is still smaller is that very few people dare to make a difference.
Do you really think consumers would willingly use Windows if their favourite software was available for an OS like Ubuntu or MacOSX (both way more user friendly than Windows)? An OS that seldom crashes and, when it does, prints concise messages rather than the Gobbledy-Gook Windows treats it's users to?
Heck. I think a company supporting a game development framework owes it to themselves and their userbase to publish their whole software to Linux. This way, users can grow and actually acquire some skills they might need later on (if they chose to pursue a professional, IT centric career).
I can't speak for MacOSX, but porting to Linux is blatantly easy and having a varying amount of distros isn't an issue, at all. Most users will choose a Debian based distro, which means you can use dpkg to pack Construct2 as a .deb. Debian also has an rpm packager available, so you can port your .deb to .rpm (used by RHEL, CentOS, etc) in a matter of minutes. Dependencies also won't be much of a problem. Especially if your download page lists certain distros that should be used.
It looks to me as if most of these issues come down to a quite error prone Direct3D implementation (AFAIK, the editor view uses D3D rather then OGL). By porting it over to Unix, you'd have to convert all your rendering code to OpenGL, which should also allow you to iron out all the little flaws within the rendering code.
*That's homeusers, btw. In enterprise applications (especially the server space), Windows is slowly but steadily going the way of the Dodo.