Might I suggest, zydubion, that you do some studying on Hyperbolic geometry? And make sure that relative consistency is on your cirriculum... one of the pertenint points is that any mathematical proof that Euclidean geometry is the only geometry works as a proof that Euclidean geometry itself is a logical contradiction.
There's plenty of evidence that space-time can be described by a non-Euclidean geometry. For example, according to General Relativity, the path that the Earth takes around the sun IS a straight line through space-time.
In general, lines don't have to bend to return to their starting points. A common example (at least it was a couple decades ago) is a video game like Asteroids; when you go off of one edge of the screen you appear on the other side. Believe it or not, but the "universe" in which those video game objects live is a perfectly valid geometry. We see edges since we've embedded the geometry into our 3-dimensional universe in a non-faithful way, but there are no edges to those living in this geometry, and it's everywhere flat to boot... yet you can go in a straight line and end up back where you started.
(Incidentally, this geometry is called the flat two-dimensional torus)
Mathematics (and the laws of physics) certainly permit this type of geometry for the universe, and even much stranger ones. Currently we only consider geometries of space that look Euclidean on small scales (actually, geometries of space-time that look Minowski on small scales), but theoretical physical theories consider more exotic things.