Earlier quoted context omitted.
Except now your fonts are rendered like trash because the subpixel rendering and hinting assumes each RGB pixel triplet is in a horizontal line.
I know nothing about the technical details of font rendering algorithms, but isn't that just an implementation detail? How hard would it be to write a font rendering algorithm that works when each RGB triplet is a vertical line? And having done that how hard would it be to implement a font renderer that switches between the two algorithms depending on screen orientation?
Edit: natesm posted a response to the grandparent of this comment with a screenshot of the subpixel orientation selection in Ubuntu.