uint8, uint16, uint32, and uint64 in kernels (#413)

A forthcoming PR will update the RNG to use these types.

Also:
- Add tests for the `//`, `<<`, and `>>` operators.
- Change `TensorWrapper` to unwrap objects when the resulting object would be simpler.
- Clean up `throw_unreachable`, since it was triggering compiler warnings.

lib/ir/context.cc

@@ -19,12 +19,16 @@ context_impl::context_impl(context &ctx)
       fp32_ty(ctx, type::FP32TyID),
       fp64_ty(ctx, type::FP64TyID),
       // integers
-      int1_ty(ctx, 1),
-      int8_ty(ctx, 8),
-      int16_ty(ctx, 16),
-      int32_ty(ctx, 32),
-      int64_ty(ctx, 64),
-      int128_ty(ctx, 128){
+      int1_ty(ctx, 1, signedness::SIGNED),
+      int8_ty(ctx, 8, signedness::SIGNED),
+      int16_ty(ctx, 16, signedness::SIGNED),
+      int32_ty(ctx, 32, signedness::SIGNED),
+      int64_ty(ctx, 64, signedness::SIGNED),
+      int128_ty(ctx, 128, signedness::SIGNED),
+      uint8_ty(ctx, 8, signedness::UNSIGNED),
+      uint16_ty(ctx, 16, signedness::UNSIGNED),
+      uint32_ty(ctx, 32, signedness::UNSIGNED),
+      uint64_ty(ctx, 64, signedness::UNSIGNED){
 
 }