I just read a PhD thesis where somebody wrote an LLVM back-end for the 
Haskell compiler. The "conclusions" section is entertaining. Consider this:


current complete uselessness is surprising. In closing, I believe that 
LLVM offers a compelling and viable back-end target for GHC today."

Obviously I'm taking this out of context. But seriously... it's useless, 
and that makes it a compelling and viable back-end? I LOL'd.

Post a reply to this message