LLVM C++ API How to create a constant f128

Question

how can I create a constant f128 (quad) using the LLVM C*+ API. Usually you do it like this (as per the tutorial):

ConstantFP::get(*TheContext, APFloat(Val))

But APFloat has no constructor that can take a long double or a StringRef where I can pass in the string repr of the quad. So how do I create a long double with the LLVM C++ API? As per the documentation of the IR language I know that a f128 type exists.

Answer 1

Try using this method instead with an f128 type gotten from this llvm::Type static member

Answer 2

Looks like APFloat has a constructor that takes a String and another that takes an abitrary precision integer (presumably containing the bits of the IEEE754 representation). The fp128 Type object has getFltSemantics() which both constructors need.