[mlir][sparse] Improving error detection/messages for `get{RankedTensor,MemRef}Type` wrappers

[wrengr]

Summary

This helps catch some otherwise hard to track down segfaults. N.B., even though getSparseTensorType is not touched in this patch, it also gains the new error checking (via getRankedTensorType).

Depends On D149805

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensor.h

 
 //===----------------------------------------------------------------------===//
 // Additional convenience methods.
 //===----------------------------------------------------------------------===//
 
 namespace mlir {
 namespace sparse_tensor {
 
+// NOTE: `Value::getType` doesn't check for null before trying to
+// dereference things.  Therefore we check, because an assertion-failure
+// is easier to debug than a segfault.  Presumably other `T::getType`
+// methods are similarly susceptible.
+
 /// Convenience method to abbreviate casting `getType()`.
 template <typename T>
-inline RankedTensorType getRankedTensorType(T t) {
-  return t.getType().template cast<RankedTensorType>();
+inline RankedTensorType getRankedTensorType(T &&t) {
+  assert(static_cast<bool>(std::forward<T>(t)) &&
+         "getRankedTensorType got null argument");
+  return std::forward<T>(t).getType().template cast<RankedTensorType>();
 }
 
 /// Convenience method to abbreviate casting `getType()`.
 template <typename T>
-inline MemRefType getMemRefType(T t) {
-  return t.getType().template cast<MemRefType>();
+inline MemRefType getMemRefType(T &&t) {
+  assert(static_cast<bool>(std::forward<T>(t)) &&
+         "getMemRefType got null argument");
+  return std::forward<T>(t).getType().template cast<MemRefType>();
 }
 
 /// Convenience method to get a sparse encoding attribute from a type.
 /// Returns null-attribute for any type without an encoding.
 SparseTensorEncodingAttr getSparseTensorEncoding(Type type);
 
 /// Returns true iff the given sparse tensor encoding attribute has a trailing
 /// COO region starting at the given level.

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorType.h

 class SparseTensorType {
 public:
   // We memoize `lvlRank` and `dim2lvl` to avoid repeating the
   // conditionals throughout the rest of the class.
   SparseTensorType(RankedTensorType rtp)
       : rtp(rtp), enc(getSparseTensorEncoding(rtp)),
         lvlRank(enc ? enc.getLvlRank() : getDimRank()),
         dim2lvl(enc.hasIdDimOrdering() ? AffineMap() : enc.getDimOrdering()) {
+    assert(rtp && "got null RankedTensorType");
     assert((!isIdentity() || getDimRank() == lvlRank) && "Rank mismatch");
   }
 
   SparseTensorType(ShapedType stp, SparseTensorEncodingAttr enc)
       : SparseTensorType(
             RankedTensorType::get(stp.getShape(), stp.getElementType(), enc)) {}
 
   // Copy-assignment would be implicitly deleted (because our fields