[Mlir-commits] [mlir] a58ed7f - Fix ConstShapeOp::inferReturnTypes to be resilient to lack of properties

[Mehdi Amini]

Author: Mehdi Amini
Date: 2023-05-15T12:12:28-07:00
New Revision: a58ed7fcf6fcae7b622f8a1873459f28725ce7d6

URL: https://github.com/llvm/llvm-project/commit/a58ed7fcf6fcae7b622f8a1873459f28725ce7d6
DIFF: https://github.com/llvm/llvm-project/commit/a58ed7fcf6fcae7b622f8a1873459f28725ce7d6.diff

LOG: Fix ConstShapeOp::inferReturnTypes to be resilient to lack of properties

The Python bindings test aren't using properties yet, this is a bit
of a hack to support this here, but hopefully it'll be temporary.

Added: 
    

Modified: 
    mlir/lib/Dialect/Shape/IR/Shape.cpp

Removed: 
    


################################################################################
diff  --git a/mlir/lib/Dialect/Shape/IR/Shape.cpp b/mlir/lib/Dialect/Shape/IR/Shape.cpp
index 24302544ea06..58d0e6aa3964 100644
--- a/mlir/lib/Dialect/Shape/IR/Shape.cpp
+++ b/mlir/lib/Dialect/Shape/IR/Shape.cpp
@@ -921,7 +921,13 @@ LogicalResult mlir::shape::ConstShapeOp::inferReturnTypes(
     SmallVectorImpl<Type> &inferredReturnTypes) {
   Builder b(context);
   Properties *prop = properties.as<Properties *>();
-  DenseIntElementsAttr shape = prop->shape;
+  DenseIntElementsAttr shape;
+  // TODO: this is only exercised by the Python bindings codepath which does not
+  // support properties
+  if (prop)
+    shape = prop->shape;
+  else
+    shape = attributes.getAs<DenseIntElementsAttr>("shape");
   if (!shape)
     return emitOptionalError(location, "missing shape attribute");
   inferredReturnTypes.assign({RankedTensorType::get(