[Mlir-commits] [mlir] [mlir][arith] Add more canonicalization and integration tests coverage (PR #92272)

[Jakub Kuderski]

================
@@ -3022,6 +3022,36 @@ func.func @minsi_i0() -> i0 {
   return %minsi : i0
 }
 
+// CHECK-LABEL: @extsiOnI1
+//       CHECK: %[[TRUE:.*]] = arith.constant true
+//       CHECK: %[[CST:.*]] = arith.constant -1 : i16
+//       CHECK: return %[[TRUE]], %[[CST]]
+func.func @extsiOnI1() -> (i1, i16) {
+  %true = arith.constant -1 : i1
----------------
kuhar wrote:

is there a particular reason for using -1 here instead of `arith.constant true` like in the test cases below?

https://github.com/llvm/llvm-project/pull/92272