[clang] [llvm] [HLSL] Implement elementwise firstbitlow builtin (PR #116858)

Ashley Coleman via cfe-commits cfe-commits at lists.llvm.org
Tue Nov 19 10:58:08 PST 2024


================
@@ -17,12 +17,10 @@ double test_int_builtin(double p0) {
 
 double2 test_int_builtin_2(double2 p0) {
   return __builtin_hlsl_elementwise_firstbithigh(p0);
-  // expected-error at -1 {{1st argument must be a vector of integers
-  // (was 'double2' (aka 'vector<double, 2>'))}}
+  // expected-error at -1 {{1st argument must be a vector of integers (was 'double2' (aka 'vector<double, 2>'))}}
 }
 
 float test_int_builtin_3(float p0) {
   return __builtin_hlsl_elementwise_firstbithigh(p0);
-  // expected-error at -1 {{1st argument must be a vector of integers
-  // (was 'float')}}
+  // expected-error at -1 {{1st argument must be a vector of integers (was 'double')}}
----------------
V-FEXrt wrote:

The test framework doesn't actually assert these lines when split. I had to unsplit them which raised a test failure that was then fixed

https://github.com/llvm/llvm-project/pull/116858


More information about the cfe-commits mailing list