[clang] Implement CWG2627 Bit-fields and narrowing conversions

clang/docs/ReleaseNotes.rst

@@ -111,6 +111,11 @@ Resolutions to C++ Defect Reports
   Otherwise, if there is no initializer list constructor, the copy will be elided as if it was ``T(e)``.
   (`CWG2311: Missed case for guaranteed copy elision <https://cplusplus.github.io/CWG/issues/2311.html>`)
 
+- Casts from a bit-field to an integral type is now not considered narrowing if the
+  width of the bit-field means that all potential values are in the range
+  of the target type, even if the type of the bit-field is larger.
+  (`CWG2627: Bit-fields and narrowing conversions <https://cplusplus.github.io/CWG/issues/2627.html>`_)
+
 C Language Changes
 ------------------
 

clang/lib/Sema/SemaOverload.cpp

@@ -489,61 +489,93 @@ NarrowingKind StandardConversionSequence::getNarrowingKind(
 
   // -- from an integer type or unscoped enumeration type to an integer type
   //    that cannot represent all the values of the original type, except where
-  //    the source is a constant expression and the actual value after
+  //    (CWG2627) -- the source is a bit-field whose width w is less than that
+  //    of its type (or, for an enumeration type, its underlying type) and the
+  //    target type can represent all the values of a hypothetical extended
+  //    integer type with width w and with the same signedness as the original
+  //    type or
+  //    -- the source is a constant expression and the actual value after
   //    conversion will fit into the target type and will produce the original
   //    value when converted back to the original type.
   case ICK_Integral_Conversion:
   IntegralConversion: {
     assert(FromType->isIntegralOrUnscopedEnumerationType());
     assert(ToType->isIntegralOrUnscopedEnumerationType());
     const bool FromSigned = FromType->isSignedIntegerOrEnumerationType();
-    const unsigned FromWidth = Ctx.getIntWidth(FromType);
+    unsigned FromWidth = Ctx.getIntWidth(FromType);
     const bool ToSigned = ToType->isSignedIntegerOrEnumerationType();
     const unsigned ToWidth = Ctx.getIntWidth(ToType);
 
-    if (FromWidth > ToWidth ||
-        (FromWidth == ToWidth && FromSigned != ToSigned) ||
-        (FromSigned && !ToSigned)) {
-      // Not all values of FromType can be represented in ToType.
-      const Expr *Initializer = IgnoreNarrowingConversion(Ctx, Converted);
+    constexpr auto CanRepresentAll = [](bool FromSigned, unsigned FromWidth,
+                                        bool ToSigned, unsigned ToWidth) {
+      return (FromWidth < ToWidth + (FromSigned == ToSigned)) &&
+             (FromSigned <= ToSigned);
+    };
 
-      // If it's value-dependent, we can't tell whether it's narrowing.
-      if (Initializer->isValueDependent())
-        return NK_Dependent_Narrowing;
+    if (CanRepresentAll(FromSigned, FromWidth, ToSigned, ToWidth))
+      return NK_Not_Narrowing;
 
-      std::optional<llvm::APSInt> OptInitializerValue;
-      if (!(OptInitializerValue = Initializer->getIntegerConstantExpr(Ctx))) {
-        // Such conversions on variables are always narrowing.
-        return NK_Variable_Narrowing;
-      }
-      llvm::APSInt &InitializerValue = *OptInitializerValue;
-      bool Narrowing = false;
-      if (FromWidth < ToWidth) {
-        // Negative -> unsigned is narrowing. Otherwise, more bits is never
-        // narrowing.
-        if (InitializerValue.isSigned() && InitializerValue.isNegative())
-          Narrowing = true;
-      } else {
-        // Add a bit to the InitializerValue so we don't have to worry about
-        // signed vs. unsigned comparisons.
-        InitializerValue = InitializerValue.extend(
-          InitializerValue.getBitWidth() + 1);
-        // Convert the initializer to and from the target width and signed-ness.
-        llvm::APSInt ConvertedValue = InitializerValue;
-        ConvertedValue = ConvertedValue.trunc(ToWidth);
-        ConvertedValue.setIsSigned(ToSigned);
-        ConvertedValue = ConvertedValue.extend(InitializerValue.getBitWidth());
-        ConvertedValue.setIsSigned(InitializerValue.isSigned());
-        // If the result is different, this was a narrowing conversion.
-        if (ConvertedValue != InitializerValue)
-          Narrowing = true;
-      }
-      if (Narrowing) {
-        ConstantType = Initializer->getType();
-        ConstantValue = APValue(InitializerValue);
-        return NK_Constant_Narrowing;
+    // Not all values of FromType can be represented in ToType.
+    const Expr *Initializer = IgnoreNarrowingConversion(Ctx, Converted);
+
+    bool DependentBitField = false;
+    if (const FieldDecl *BitField = Initializer->getSourceBitField()) {
+      if (BitField->getBitWidth()->isValueDependent())
+        DependentBitField = true;
+      else if (unsigned BitFieldWidth = BitField->getBitWidthValue(Ctx);
+               BitFieldWidth < FromWidth) {
+        if (CanRepresentAll(FromSigned, BitFieldWidth, ToSigned, ToWidth))
+          return NK_Not_Narrowing;
+
+        // The initializer will be truncated to the bit-field width
+        FromWidth = BitFieldWidth;
       }
     }
+
+    // If it's value-dependent, we can't tell whether it's narrowing.
+    if (Initializer->isValueDependent())
+      return NK_Dependent_Narrowing;
+
+    std::optional<llvm::APSInt> OptInitializerValue =
+        Initializer->getIntegerConstantExpr(Ctx);
+    if (!OptInitializerValue) {
+      // If the bit-field width was dependent, it might end up being small
+      // enough to fit in the target type (unless the target type is unsigned
+      // and the source type is signed, in which case it will never fit)
+      if (DependentBitField && (FromSigned <= ToSigned))
+        return NK_Dependent_Narrowing;
+
+      // Otherwise, such a conversion is always narrowing
+      return NK_Variable_Narrowing;
+    }
+    llvm::APSInt &InitializerValue = *OptInitializerValue;
+    bool Narrowing = false;
+    if (FromWidth < ToWidth) {
+      // Negative -> unsigned is narrowing. Otherwise, more bits is never
+      // narrowing.
+      if (InitializerValue.isSigned() && InitializerValue.isNegative())
+        Narrowing = true;
+    } else {
+      // Add a bit to the InitializerValue so we don't have to worry about
+      // signed vs. unsigned comparisons.
+      InitializerValue =
+          InitializerValue.extend(InitializerValue.getBitWidth() + 1);
+      // Convert the initializer to and from the target width and signed-ness.
+      llvm::APSInt ConvertedValue = InitializerValue;
+      ConvertedValue = ConvertedValue.trunc(ToWidth);
+      ConvertedValue.setIsSigned(ToSigned);
+      ConvertedValue = ConvertedValue.extend(InitializerValue.getBitWidth());
+      ConvertedValue.setIsSigned(InitializerValue.isSigned());
+      // If the result is different, this was a narrowing conversion.
+      if (ConvertedValue != InitializerValue)
+        Narrowing = true;
+    }
+    if (Narrowing) {
+      ConstantType = Initializer->getType();
+      ConstantValue = APValue(InitializerValue);
+      return NK_Constant_Narrowing;
+    }
+
     return NK_Not_Narrowing;
   }
   case ICK_Complex_Real:

clang/test/CXX/drs/cwg26xx.cpp

@@ -1,11 +1,39 @@
-// RUN: %clang_cc1 -std=c++98 -triple x86_64-unknown-unknown -pedantic-errors %s -verify=expected
-// RUN: %clang_cc1 -std=c++11 -triple x86_64-unknown-unknown -pedantic-errors %s -verify=expected,since-cxx11,cxx11
-// RUN: %clang_cc1 -std=c++14 -triple x86_64-unknown-unknown -pedantic-errors %s -verify=expected,since-cxx11
-// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-unknown -pedantic-errors %s -verify=expected,since-cxx11
-// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-unknown -pedantic-errors %s -verify=expected,since-cxx11,since-cxx20
-// RUN: %clang_cc1 -std=c++23 -triple x86_64-unknown-unknown -pedantic-errors %s -verify=expected,since-cxx11,since-cxx20,since-cxx23
-// RUN: %clang_cc1 -std=c++2c -triple x86_64-unknown-unknown -pedantic-errors %s -verify=expected,since-cxx11,since-cxx20,since-cxx23
+// RUN: %clang_cc1 -std=c++98 -pedantic-errors %s -verify=expected,cxx98
+// RUN: %clang_cc1 -std=c++11 -pedantic-errors %s -verify=expected,since-cxx11,cxx11
+// RUN: %clang_cc1 -std=c++14 -pedantic-errors %s -verify=expected,since-cxx11
+// RUN: %clang_cc1 -std=c++17 -pedantic-errors %s -verify=expected,since-cxx11
+// RUN: %clang_cc1 -std=c++20 -pedantic-errors %s -verify=expected,since-cxx11,since-cxx20
+// RUN: %clang_cc1 -std=c++23 -pedantic-errors %s -verify=expected,since-cxx11,since-cxx20,since-cxx23
+// RUN: %clang_cc1 -std=c++2c -pedantic-errors %s -verify=expected,since-cxx11,since-cxx20,since-cxx23
+
+#if __cplusplus == 199711L
+#define static_assert(...) __extension__ _Static_assert(__VA_ARGS__)
+// cxx98-error@-1 {{variadic macros are a C99 feature}}
+#endif
 
+namespace std {
+#if __cplusplus >= 202002L
+  struct strong_ordering {
+    int n;
+    constexpr operator int() const { return n; }
+    static const strong_ordering less, equal, greater;
+  };
+  constexpr strong_ordering strong_ordering::less{-1},
+      strong_ordering::equal{0}, strong_ordering::greater{1};
+#endif
+
+  typedef short int16_t;
+  typedef unsigned short uint16_t;
+  typedef int int32_t;
+  typedef unsigned uint32_t;
+  typedef long long int64_t;
+  // cxx98-error@-1 {{'long long' is a C++11 extension}}
+  typedef unsigned long long uint64_t;
+  // cxx98-error@-1 {{'long long' is a C++11 extension}}
+  static_assert(sizeof(int16_t) == 2 && sizeof(int32_t) == 4 && sizeof(int64_t) == 8, "Some tests rely on these sizes");
+
+  template<typename T> T declval();
+}
 
 namespace cwg2621 { // cwg2621: sup 2877
 #if __cplusplus >= 202002L
@@ -23,6 +51,87 @@ using enum E;
 #endif
 }
 
+namespace cwg2627 { // cwg2627: 20
+#if __cplusplus >= 202002L
+struct C {
+  long long i : 8;
+  friend auto operator<=>(C, C) = default;
+};
+
+void f() {
+  C x{1}, y{2};
+  static_cast<void>(x <=> y);
+  static_cast<void>(x.i <=> y.i);
+}
+
+template<typename T>
+struct CDependent {
+  T i : 8;
+  friend auto operator<=>(CDependent, CDependent) = default;
+};
+
+template<typename T>
+concept three_way_comparable = requires(T t) { { t <=> t }; };
+template<typename T>
+concept bf_three_way_comparable = requires(T t) { { t.i <=> t.i }; };
+static_assert(three_way_comparable<CDependent<long long>>);
+static_assert(bf_three_way_comparable<CDependent<long long>>);
+#endif
+
+#if __cplusplus >= 201103L
+template<typename T, int N>
+struct D {
+  T i : N;
+};
+
+template<typename T, int N>
+D<T, N> d();
+
+std::int32_t d1{ d<std::int64_t, 31>().i };
+std::int32_t d2{ d<std::int64_t, 32>().i };
+std::int32_t d3{ d<std::int64_t, 33>().i };
+// since-cxx11-error@-1 {{non-constant-expression cannot be narrowed from type 'long long' to 'std::int32_t' (aka 'int') in initializer list}}
+//   since-cxx11-note@-2 {{insert an explicit cast to silence this issue}}
+
+std::int16_t d6{ d<int, 16>().i };
+std::int16_t d7{ d<unsigned, 15>().i };
+std::int16_t d8{ d<unsigned, 16>().i };
+// since-cxx11-error@-1 {{non-constant-expression cannot be narrowed from type 'unsigned int' to 'std::int16_t' (aka 'short') in initializer list}}
+//   since-cxx11-note@-2 {{insert an explicit cast to silence this issue}}
+std::uint16_t d9{ d<unsigned, 16>().i };
+std::uint16_t da{ d<int, 1>().i };
+// since-cxx11-error@-1 {{non-constant-expression cannot be narrowed from type 'int' to 'std::uint16_t' (aka 'unsigned short') in initializer list}}
+//   since-cxx11-note@-2 {{insert an explicit cast to silence this issue}}
+
+bool db{ d<unsigned, 1>().i };
+bool dc{ d<int, 1>().i };
+// since-cxx11-error@-1 {{non-constant-expression cannot be narrowed from type 'int' to 'bool' in initializer list}}
+//   since-cxx11-note@-2 {{insert an explicit cast to silence this issue}}
+
+template<typename Target, typename Source>
+constexpr decltype(Target{ std::declval<Source>().i }, false) is_narrowing(int) { return false; }
+template<typename Target, typename Source>
+constexpr bool is_narrowing(long) { return true; }
+
+static_assert(!is_narrowing<std::int16_t, D<int, 16>>(0), "");
+static_assert(!is_narrowing<std::int16_t, D<unsigned, 15>>(0), "");
+static_assert(is_narrowing<std::int16_t, D<unsigned, 16>>(0), "");
+static_assert(!is_narrowing<std::uint16_t, D<unsigned, 16>>(0), "");
+static_assert(is_narrowing<std::uint16_t, D<int, 1>>(0), "");
+static_assert(!is_narrowing<bool, D<unsigned, 1>>(0), "");
+static_assert(is_narrowing<bool, D<int, 1>>(0), "");
+
+template<int N>
+struct E {
+  signed int x : N;
+  decltype(std::int16_t{ x }) dependent_narrowing;
+  decltype(unsigned{ x }) always_narrowing;
+  // since-cxx11-error@-1 {{non-constant-expression cannot be narrowed from type 'int' to 'unsigned int' in initializer list}}
+  //   since-cxx11-note@-2 {{insert an explicit cast to silence this issue}}
+};
+#endif
+} // namespace cwg2627
+
 namespace cwg2628 { // cwg2628: no
                    // this was reverted for the 16.x release
                    // due to regressions, see the issue for more details:

clang/test/Sema/constexpr.c

@@ -360,3 +360,10 @@ void infsNaNs() {
 
 constexpr struct S9 s9 = {  }; // expected-error {{variable has incomplete type 'const struct S9'}} \
                                // expected-note {{forward declaration of 'struct S9'}}
+
+struct S10 {
+  signed long long i : 8;
+};
+constexpr struct S10 c = { 255 };
+// FIXME-expected-error@-1 {{constexpr initializer evaluates to 255 which is not exactly representable in 'long long' bit-field with width 8}}
+// See: GH#101299

clang/test/SemaCXX/bitint-narrowing.cpp

@@ -0,0 +1,36 @@
+// RUN: %clang_cc1 -fsyntax-only -verify -std=c++11 %s
+// RUN: %clang_cc1 -triple x86_64 -fsyntax-only -verify -std=c++11 %s
+// RUN: %clang_cc1 -triple i386 -fsyntax-only -verify -std=c++11 %s
+
+struct {
+  _BitInt(35) i : 33;
+} x;
+struct {
+  _BitInt(35) i : 34;
+} y;
+_BitInt(33) xx{ x.i };
+_BitInt(33) yy{ y.i };
+// expected-error@-1 {{non-constant-expression cannot be narrowed from type '_BitInt(35)' to '_BitInt(33)' in initializer list}}
+//   FIXME-expected-note@-2 {{insert an explicit cast to silence this issue}}
+
+         _BitInt(2) S2 = 0;
+unsigned _BitInt(2) U2 = 0;
+         _BitInt(3) S3 = 0;
+unsigned _BitInt(3) U3 = 0;
+
+         _BitInt(2) bi0{ S2 };
+         _BitInt(2) bi1{ U2 }; // expected-error {{non-constant-expression cannot be narrowed from type 'unsigned _BitInt(2)' to '_BitInt(2)' in initializer list}}
+         _BitInt(2) bi2{ S3 }; // expected-error {{non-constant-expression cannot be narrowed from type '_BitInt(3)' to '_BitInt(2)' in initializer list}}
+         _BitInt(2) bi3{ U3 }; // expected-error {{non-constant-expression cannot be narrowed from type 'unsigned _BitInt(3)' to '_BitInt(2)' in initializer list}}
+unsigned _BitInt(2) bi4{ S2 }; // expected-error {{non-constant-expression cannot be narrowed from type '_BitInt(2)' to 'unsigned _BitInt(2)' in initializer list}}
+unsigned _BitInt(2) bi5{ U2 };
+unsigned _BitInt(2) bi6{ S3 }; // expected-error {{non-constant-expression cannot be narrowed from type '_BitInt(3)' to 'unsigned _BitInt(2)' in initializer list}}
+unsigned _BitInt(2) bi7{ U3 }; // expected-error {{non-constant-expression cannot be narrowed from type 'unsigned _BitInt(3)' to 'unsigned _BitInt(2)' in initializer list}}
+         _BitInt(3) bi8{ S2 };
+         _BitInt(3) bi9{ U2 };
+         _BitInt(3) bia{ S3 };
+         _BitInt(3) bib{ U3 }; // expected-error {{non-constant-expression cannot be narrowed from type 'unsigned _BitInt(3)' to '_BitInt(3)' in initializer list}}
+unsigned _BitInt(3) bic{ S2 }; // expected-error {{non-constant-expression cannot be narrowed from type '_BitInt(2)' to 'unsigned _BitInt(3)' in initializer list}}
+unsigned _BitInt(3) bid{ U2 };
+unsigned _BitInt(3) bie{ S3 }; // expected-error {{non-constant-expression cannot be narrowed from type '_BitInt(3)' to 'unsigned _BitInt(3)' in initializer list}}
+unsigned _BitInt(3) bif{ U3 };

clang/www/cxx_dr_status.html

@@ -15577,7 +15577,7 @@ <h2 id="cxxdr">C++ defect report implementation status</h2>
     <td><a href="https://cplusplus.github.io/CWG/issues/2627.html">2627</a></td>
     <td>C++23</td>
     <td>Bit-fields and narrowing conversions</td>
-    <td class="unknown" align="center">Unknown</td>
+    <td class="unreleased" align="center">Clang 20</td>
   </tr>
   <tr id="2628">
     <td><a href="https://cplusplus.github.io/CWG/issues/2628.html">2628</a></td>