[Clang][NFC] Mark P2552 as implemented.

[cor3ntin]

wg21.link/P2552 suggest that __has_cpp_attribute
should return a non-zero value for all attributes that the implementation does something interesting with.

Clang does something meaninful with all attributes except for:

• no_unique_address which we do not support for msvc target
• carries_dependency which arguably does nothing interesting. P2552 shies away from specifying a behavior for that attribute (despite being the only one for which a recommandation would have been interesting, arguably)

As such, we have nothing to change for this paper. This paper is a DR and clang always behaved reasonably.

[llvmbot]

@llvm/pr-subscribers-clang

Author: cor3ntin (cor3ntin)

Changes

wg21.link/P2552 suggest that __has_cpp_attribute
should return a non-zero value for all attributes that the implementation does something interesting with.

Clang does something meaninful with all attributes except for:

• no_unique_address which we do not support for msvc target
• carries_dependency which arguably does nothing interesting. P2552 shies away from specifying a behavior for that attribute (despite being the only one for which a recommandation would have been interesting, arguably)

As such, we have nothing to change for this paper. This paper is a DR and clang always behaved reasonably.

---

Full diff: https://github.com/llvm/llvm-project/pull/92007.diff

2 Files Affected:

• (added) clang/test/SemaCXX/cxx2c-attributes.cpp (+23)
• (modified) clang/www/cxx_status.html (+1-1)

diff --git a/clang/test/SemaCXX/cxx2c-attributes.cpp b/clang/test/SemaCXX/cxx2c-attributes.cpp
new file mode 100644
index 0000000000000..f9f76efd01be0
--- /dev/null
+++ b/clang/test/SemaCXX/cxx2c-attributes.cpp
@@ -0,0 +1,23 @@
+// RUN: %clang_cc1 -x c++ -std=c++11 -triple x86_64-pc-linux -fsyntax-only
+// RUN: %clang_cc1 -x c++ -std=c++11 -triple x86_64-windows-msvc -fsyntax-only
+
+// Check we return non-zero values for supported attributes as per
+// wg21.link/p2552r3.pdf
+static_assert(__has_cpp_attribute(assume));
+
+// The standard does not prescribe a behavior for [[carries_dependency]]
+
+static_assert(__has_cpp_attribute(deprecated));
+static_assert(__has_cpp_attribute(fallthrough));
+static_assert(__has_cpp_attribute(likely));
+static_assert(__has_cpp_attribute(unlikely));
+static_assert(__has_cpp_attribute(maybe_unused));
+static_assert(__has_cpp_attribute(nodiscard));
+static_assert(__has_cpp_attribute(noreturn));
+
+#ifdef _MSC_VER
+// We do not support [[no_unique_address]] in MSVC emulation mode
+static_assert(__has_cpp_attribute(no_unique_address) == 0);
+#else
+static_assert(__has_cpp_attribute(no_unique_address));
+#endif
diff --git a/clang/www/cxx_status.html b/clang/www/cxx_status.html
index 1338f544ffcb5..06777eaa6df69 100755
--- a/clang/www/cxx_status.html
+++ b/clang/www/cxx_status.html
@@ -130,7 +130,7 @@ <h2 id="cxx26">C++2c implementation status</h2>
  <tr>
   <td>On the ignorability of standard attributes</td>
   <td><a href="https://wg21.link/P2552R3">P2552R3</a> (<a href="#dr">DR</a>)</td>
-  <td class="none" align="center">No</td>
+  <td class="full" align="center">Yes</td>
  </tr>
  <tr>
   <td>Static storage for braced initializers</td>

[AaronBallman]

LGTM aside from a potential improvement to the test for clarity.

[AaronBallman]

I think it's more clear to use -verify= instead of using #ifdef so it's clear that we expect the static assertion to fail in MSVC mode, but I don't insist.