================
@@ -0,0 +1,348 @@
+// RUN: %clang_cc1 --std=c++20 -fsyntax-only -Wdangling -Wdangling-field
-Wreturn-stack-address -verify %s
+
+#include "Inputs/lifetime-analysis.h"
+
+struct X {
+ const int *x;
+};
+X x;
+
+// ****************************************************************************
+// Capture an integer
+// ****************************************************************************
+namespace capture_int {
+void captureInt(const int &i [[clang::lifetime_capture_by(x)]], X &x);
+void captureRValInt(int &&i [[clang::lifetime_capture_by(x)]], X &x);
+void noCaptureInt(int i [[clang::lifetime_capture_by(x)]], X &x);
+
+void use() {
+ int local;
+ captureInt(1, // expected-warning {{object whose reference is captured by
'x' will be destroyed at the end of the full-expression}}
+ x);
+ captureRValInt(1, x); // expected-warning {{object whose reference is
captured by 'x'}}
----------------
bricknerb wrote:
I think it's very helpful for multiple reasons:
1. Documentation perspective: it clarifies what warning is expected and makes
use cases.
2. Makes tests easier to search for when copying the warning text.
3. While it doesn't add test coverage, the single case that does check the
entire text is also tests for a specific use case, and if we happen to remove
it (perhaps it's covered by somewhere else), we might forget to add the full
text to a different case and we will lose coverage.
https://github.com/llvm/llvm-project/pull/115921
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
