Implement floating point numbers about llvm HOT 1 CLOSED

Update as of 2016-12-01.

The parser branch has gained support for most aspects of floating-point numbers (latest commit 778a42e).

The only remaining aspects are fast-math-flags. The 0.2 release is intended to focus on a simplified subset of LLVM IR, as covered by the data types defined in https://godoc.org/github.com/llir/llvm/ir.

A future release will include support for further details of the LLVM IR language, including fast-math-flags, and other instructions specific flags, such as nsw and nuw of binary operations.

The work to implement a the full subset of LLVM IR is covered by issue #7, and the specific issue to implement fast-math-flags has been move to that issue for future tracking.

Closing this issue for now, as the intended subset of floating-point instructions have been implemented for the release of v0.2.

@sangisos and @7i, great work! :)

from llvm.