================
@@ -0,0 +1,42 @@
+// RUN: %clang_cc1 %s -DTEST_EMPTY_COPYRIGHT -triple powerpc-ibm-aix -verify
+// RUN: %clang_cc1 %s -DTEST_OTHER_COMMENT_KINDS -triple powerpc-ibm-aix
-verify
+// RUN: %clang_cc1 %s -triple powerpc-ibm-aix -emit-llvm -disable-llvm-passes
-o - | FileCheck %s
+// RUN: %clang_cc1 %s -triple powerpc-ibm-aix -verify
+
+// RUN: %clang_cc1 %s -DTEST_EMPTY_COPYRIGHT -triple powerpc64-ibm-aix -verify
+// RUN: %clang_cc1 %s -DTEST_OTHER_COMMENT_KINDS -triple powerpc64-ibm-aix
-verify
+// RUN: %clang_cc1 %s -triple powerpc64-ibm-aix -emit-llvm
-disable-llvm-passes -o - | FileCheck %s
+// RUN: %clang_cc1 %s -triple powerpc64-ibm-aix -verify
+
+#ifdef TEST_EMPTY_COPYRIGHT
+
+// 1. Verify no diagnostics for empty copyright string
+#pragma comment(copyright, "") // expected-no-diagnostics
+int main() {return 0; }
+
+#elif defined(TEST_OTHER_COMMENT_KINDS)
+
+// 2. Verify warnings for lib/linker and silent ignore for others
+#pragma comment(lib, "m") // expected-warning {{'#pragma comment lib' ignored}}
+#pragma comment(linker, "foo") // expected-warning {{'#pragma comment linker'
ignored}}
+
+// These are recognized but ignored in CodeGen
----------------
w2yehia wrote:
why do we ignore them in Codegen? sorry if this was answered before.
why not accept only what's supported during Preprocessor or Sema, and have
Codegen deal with supported types of comments? Or is this the normal
implementation for pragma comment on other targets?
https://github.com/llvm/llvm-project/pull/178184
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
