[Mlir-commits] [mlir] [mlirbc] Add missing encoding for float types (PR #191962)

[Mehdi Amini]

================
@@ -231,6 +231,12 @@ def BFloat16Type : DialectType<(type)>;
 
 def Float16Type : DialectType<(type)>;
 
+// Stage the addition of new floating point types so that readers can be updated
+// first.
----------------
joker-eph wrote:

I'm still confused about the process here, do we have precedent or are you innovating on this? How is this not just dead code here?

https://github.com/llvm/llvm-project/pull/191962