[Mlir-commits] [mlir] [mlir][vector] extend `createReadOrMaskedRead`/`createWriteOrMaskedWrite` with permutation map support (PR #202766)

[Federico Bruzzone]

================
@@ -434,30 +457,45 @@ Value vector::createReadOrMaskedRead(OpBuilder &builder, Location loc,
   int64_t vecToReadRank = vecToReadTy.getRank();
   auto vecToReadShape = vecToReadTy.getShape();
 
-  assert(sourceShape.size() == static_cast<size_t>(vecToReadRank) &&
+  assert((permutationMap ||
+          sourceShape.size() == static_cast<size_t>(vecToReadRank)) &&
----------------
FedericoBruzzone wrote:

When permutationMap is provided, sourceShape must have the correct rank. 
I updated the assert message.

https://github.com/llvm/llvm-project/pull/202766