When a matrix is divided by a scalar it tries to take the reciprocal
of the scalar to convert the operation into a multiply. However it was
always doing this by making a 32-bit constant. If the scalar is a
double then this would end up making an FDiv instruction with
different types in the operands.

This patch adds a helper method called makeFpConstant which makes a
floating-point constant of the given type. The code to take the
reciprocal now uses it to make the same type as the result.

Fixes https://github.com/KhronosGroup/glslang/issues/1278

Update known_good to SPIR-V 1.3 support

HLSL: Correct some mistakes for min16 types

include/Common.h: reorder includes to avoid redefinition on some platforms

- Add missing constructor ops to support float16/int16/uint16 types
- Allow half float literals
- Correct two errors of double literal parse in HLSL: extension check and
  postfix

On some systems sstream defines snprintf, this leads to the a redefinition of sprintf_s if snprintf is defined before checking if it exists in sstream.

Fix memory leak upon ProcessDeferred failure.

Allow --hlsl-enable-16bit-types to be used without AMD_EXTENSIONS

Fixed GCC+Clang errors and warnings on Linux and OSX.

Merge branch 'fixup_sm_60_wording' of https://github.com/sheredom/glslang into sheredom-fixup_sm_60_wording

Now, version 5.* is all connected to making the uint type, which doesn't quite work.
Generator versions 4 and 6 do not do this.

SPV: Implement Vulkan 1.1 features and extensions.

This reverts commit 2c65069e, reversing
changes made to fa9b465b.

Legalization: Omit legalization warning when spirv-opt is linked.

Update spirv-tools known_good

Add more error checks for interpolateAtVertexAMD()

Still missing is automation and final decisions on mechanisms in general.