[clang] Add test for CWG2149 "Brace elision and array length deduction"

[Endilll]

This patch adds test for CWG2149, following P3106R1 "Clarifying rules for brace elision in aggregate initialization" and a clarification note on top of it added on April 2024.

I haven't found a better way to check for equality of values inside array in 98 mode than to dump AST. I'm open to suggestions there.

[llvmbot]

@llvm/pr-subscribers-clang

Author: Vlad Serebrennikov (Endilll)

Changes

This patch adds test for CWG2149, following P3106R1 "Clarifying rules for brace elision in aggregate initialization" and a clarification note on top of it added on April 2024.

I haven't found a better way to check for equality of values inside array in 98 mode than to dump AST. I'm open to suggestions there.

---

Full diff: https://github.com/llvm/llvm-project/pull/90079.diff

2 Files Affected:

• (added) clang/test/CXX/drs/cwg2149.cpp (+77)
• (modified) clang/www/cxx_dr_status.html (+1-1)

diff --git a/clang/test/CXX/drs/cwg2149.cpp b/clang/test/CXX/drs/cwg2149.cpp
new file mode 100644
index 00000000000000..d0f8cb2dfc0a93
--- /dev/null
+++ b/clang/test/CXX/drs/cwg2149.cpp
@@ -0,0 +1,77 @@
+// RUN: %clang_cc1 -std=c++98 -triple x86_64-unknown-unknown %s -verify=expected,cxx98 -fexceptions -fcxx-exceptions -pedantic-errors -ast-dump | FileCheck %s --check-prefixes CXX98
+// RUN: %clang_cc1 -std=c++11 -triple x86_64-unknown-unknown %s -verify=expected,since-cxx11 -fexceptions -fcxx-exceptions -pedantic-errors
+// RUN: %clang_cc1 -std=c++14 -triple x86_64-unknown-unknown %s -verify=expected,since-cxx11 -fexceptions -fcxx-exceptions -pedantic-errors
+// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-unknown %s -verify=expected,since-cxx11 -fexceptions -fcxx-exceptions -pedantic-errors
+// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-unknown %s -verify=expected,since-cxx11 -fexceptions -fcxx-exceptions -pedantic-errors
+// RUN: %clang_cc1 -std=c++23 -triple x86_64-unknown-unknown %s -verify=expected,since-cxx11 -fexceptions -fcxx-exceptions -pedantic-errors
+// RUN: %clang_cc1 -std=c++2c -triple x86_64-unknown-unknown %s -verify=expected,since-cxx11 -fexceptions -fcxx-exceptions -pedantic-errors
+
+#if __cplusplus == 199711L
+#define static_assert(...) __extension__ _Static_assert(__VA_ARGS__)
+// cxx98-error@-1 {{variadic macros are a C99 feature}}
+#endif
+
+namespace cwg2149 { // cwg2149: 3.1 drafting 2024-04
+#if __cplusplus <= 201103L
+struct X { int i, j, k; };
+#else
+struct X { int i, j, k = 42; };
+#endif
+
+template<int N> 
+void f1(const X(&)[N]); // #cwg2149-f1
+
+template<int N>
+void f2(const X(&)[N][2]); // #cwg2149-f2
+
+void f() {
+  X a[] = { 1, 2, 3, 4, 5, 6 };
+  static_assert(sizeof(a) / sizeof(X) == 2, "");
+  X b[2] = { { 1, 2, 3 }, { 4, 5, 6 } };
+  X c[][2] = { 1, 2, 3, 4, 5, 6 };
+  static_assert(sizeof(c) / sizeof(X[2]) == 1, "");
+  
+  #if __cplusplus >= 201103L
+  constexpr X ca[] = { 1, 2, 3, 4, 5, 6 };
+  constexpr X cb[2] = { { 1, 2, 3 }, { 4, 5, 6 } };
+  static_assert(ca[0].i == cb[0].i, "");
+  static_assert(ca[0].j == cb[0].j, "");
+  static_assert(ca[0].k == cb[0].k, "");
+  static_assert(ca[1].i == cb[1].i, "");
+  static_assert(ca[1].j == cb[1].j, "");
+  static_assert(ca[1].k == cb[1].k, "");
+
+  f1({ 1, 2, 3, 4, 5, 6 });
+  // since-cxx11-error@-1 {{no matching function for call to 'f1'}}
+  //   since-cxx11-note@#cwg2149-f1 {{candidate function [with N = 6] not viable: no known conversion from 'int' to 'const X' for 1st argument}}
+  f2({ 1, 2, 3, 4, 5, 6 });
+  // since-cxx11-error@-1 {{no matching function for call to 'f2'}}
+  //   since-cxx11-note@#cwg2149-f2 {{candidate function [with N = 6] not viable: no known conversion from 'int' to 'const X[2]' for 1st argument}}
+  #endif
+}
+} // namespace cwg2149
+
+// Constant evaluation is not powerful enough in 98 mode to check for equality
+// via static_assert, even with constant folding enabled.
+
+// CXX98:       VarDecl {{.+}} a 'X[2]'
+// CXX98-NEXT:  `-InitListExpr {{.+}} 'X[2]'
+// CXX98-NEXT:    |-InitListExpr {{.+}} 'X':'cwg2149::X'
+// CXX98-NEXT:    | |-IntegerLiteral {{.+}} 'int' 1
+// CXX98-NEXT:    | |-IntegerLiteral {{.+}} 'int' 2
+// CXX98-NEXT:    | `-IntegerLiteral {{.+}} 'int' 3
+// CXX98-NEXT:    `-InitListExpr {{.+}} 'X':'cwg2149::X'
+// CXX98-NEXT:      |-IntegerLiteral {{.+}} 'int' 4
+// CXX98-NEXT:      |-IntegerLiteral {{.+}} 'int' 5
+// CXX98-NEXT:      `-IntegerLiteral {{.+}} 'int' 6
+
+// CXX98:       VarDecl {{.+}} b 'X[2]'
+// CXX98-NEXT:  `-InitListExpr {{.+}} 'X[2]'
+// CXX98-NEXT:    |-InitListExpr {{.+}} 'X':'cwg2149::X'
+// CXX98-NEXT:    | |-IntegerLiteral {{.+}} 'int' 1
+// CXX98-NEXT:    | |-IntegerLiteral {{.+}} 'int' 2
+// CXX98-NEXT:    | `-IntegerLiteral {{.+}} 'int' 3
+// CXX98-NEXT:    `-InitListExpr {{.+}} 'X':'cwg2149::X'
+// CXX98-NEXT:      |-IntegerLiteral {{.+}} 'int' 4
+// CXX98-NEXT:      |-IntegerLiteral {{.+}} 'int' 5
+// CXX98-NEXT:      `-IntegerLiteral {{.+}} 'int' 6
diff --git a/clang/www/cxx_dr_status.html b/clang/www/cxx_dr_status.html
index 83b71e7c122d1e..ea8872c91be604 100755
--- a/clang/www/cxx_dr_status.html
+++ b/clang/www/cxx_dr_status.html
@@ -12702,7 +12702,7 @@ <h2 id="cxxdr">C++ defect report implementation status</h2>
     <td><a href="https://cplusplus.github.io/CWG/issues/2149.html">2149</a></td>
     <td>drafting</td>
     <td>Brace elision and array length deduction</td>
-    <td align="center">Not resolved</td>
+    <td title="Clang 3.1 implements 2024-04 resolution" align="center">Not Resolved*</td>
   </tr>
   <tr id="2150">
     <td><a href="https://cplusplus.github.io/CWG/issues/2150.html">2150</a></td>

[Endilll]

Status of 2149 changed just yesterday to say DR per 2024 Tokyo straw poll. That's where the discrepancy between official cwg_index.html and CWG GitHub repo comes (that I link in the description).

[cor3ntin]

Can you add examples 14, 16, 17, 18 of the paper?
I'd be if they are only tested in C++11+ modes
Then we can mark the paper as implemented (bonus point for finding the oldest conforming version)

[Endilll]

Can you add examples 14, 16, 17, 18 of the paper?

I'll add them to conformance test suite in a subsequent PR. They don't belong here, as the issue is about deducing array length from brace initializer, but the examples you're listing are focusing on various aspects of how initializer-clauses appertain to aggregate elements.

I can, though, derive more tests from them that check for deduced array length in various appertainment cases, if you feel current coverage of the issue is not sufficient.

[Sirraide]

Two questions, and more tests from the paper would probably be nice, as Corentin already pointed out, but what’s already here lgtm.

[Sirraide]

and more tests from the paper would probably be nice

Actually, yeah, now that I think about it the main question that the DR addresses is already being tested, so it’s probably fine the way it is atm.

[cor3ntin]

I'm happy to approve that as is, I expect additional tests to materialize for the rest of the paper.
I'm not concerned about the status page, a new core issue list should materialize any day now

Thanks