[clang] [HLSL] Add Append/ConsumeStructuredBuffer definitions to HLSLExternalSemaSource (PR #113643)

Greg Roth via cfe-commits cfe-commits at lists.llvm.org
Thu Nov 7 13:20:43 PST 2024


================
@@ -0,0 +1,50 @@
+// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-library -x hlsl -ast-dump -DEMPTY %s | FileCheck -check-prefix=EMPTY %s
+// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-library -x hlsl -ast-dump %s | FileCheck %s
+
+
+// This test tests two different AST generations. The "EMPTY" test mode verifies
+// the AST generated by forward declaration of the HLSL types which happens on
+// initializing the HLSL external AST with an AST Context.
+
+// The non-empty mode has a use that requires the AppendStructuredBuffer type be complete,
+// which results in the AST being populated by the external AST source. That
+// case covers the full implementation of the template declaration and the
+// instantiated specialization.
+
+// EMPTY: ClassTemplateDecl 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> <invalid sloc> implicit AppendStructuredBuffer
+// EMPTY-NEXT: TemplateTypeParmDecl 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> <invalid sloc> class depth 0 index 0 element_type
+// EMPTY-NEXT: CXXRecordDecl 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> <invalid sloc> implicit <undeserialized declarations> class AppendStructuredBuffer
+// EMPTY-NEXT: FinalAttr 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> Implicit final
+
+// There should be no more occurrences of AppendStructuredBuffer
+// EMPTY-NOT: {{[^[:alnum:]]}}AppendStructuredBuffer
+
+#ifndef EMPTY
+
+AppendStructuredBuffer<int> Buffer;
+
+#endif
+
+// CHECK: ClassTemplateDecl 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> <invalid sloc> implicit AppendStructuredBuffer
+// CHECK-NEXT: TemplateTypeParmDecl 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> <invalid sloc> class depth 0 index 0 element_type
+// CHECK-NEXT: CXXRecordDecl 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> <invalid sloc> implicit class AppendStructuredBuffer definition
+
+// CHECK: FinalAttr 0x{{[0-9A-Fa-f]+}} <<invalid sloc>> Implicit final
----------------
pow2clk wrote:

I guess this one isn't on the next line in this form unlike the EMPTY case above? 

https://github.com/llvm/llvm-project/pull/113643


More information about the cfe-commits mailing list