[Mlir-commits] [mlir] [Mlir] decompose generic by unfolding projected permutation crash fix (PR #122449)

[Andrzej Warzyński]

================
@@ -159,6 +160,16 @@ LogicalResult DecomposeProjectedPermutation::matchAndRewrite(
     auto map = op.getMatchingIndexingMap(&opOperand);
     if (!map.isProjectedPermutation(false))
       return failure();
+
+    // If we have any inputs that aren't ranked tensor types, reject
+    // the pattern.
+    if (!dyn_cast<RankedTensorType>(opOperand.get().getType()))
+      return rewriter.notifyMatchFailure(
+          opOperand.get().getLoc(),
+          llvm::formatv("Expected operand #{0} to be "
+                        "ranked tensor of any type values, but got {1}",
+                        opOperand.getOperandNumber(),
+                        opOperand.get().getType()));
----------------
banach-space wrote:

[nit] `notifyMatchFailure` messages tend to be fairly simple. My preference would be to keep this one simple as well and avoid this formatting.

https://github.com/llvm/llvm-project/pull/122449