[cfe-commits] r148701 - in /cfe/trunk: include/clang/Basic/DiagnosticSemaKinds.td lib/Sema/SemaInit.cpp test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-0x.cpp test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp
Douglas Gregor
dgregor at apple.com
Mon Jan 23 07:29:33 PST 2012
Author: dgregor
Date: Mon Jan 23 09:29:33 2012
New Revision: 148701
URL: http://llvm.org/viewvc/llvm-project?rev=148701&view=rev
Log:
Downgrade C++11 narrowing conversion errors to warnings default-mapped
to an error, so that users can turn them off if necessary. Note that
this does *not* change the behavior of in a SFINAE context, where we
still flag an error even if the warning is disabled. This matches
GCC's behavior.
Added:
cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp (with props)
Modified:
cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td
cfe/trunk/lib/Sema/SemaInit.cpp
cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-0x.cpp
Modified: cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td?rev=148701&r1=148700&r2=148701&view=diff
==============================================================================
--- cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td (original)
+++ cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td Mon Jan 23 09:29:33 2012
@@ -2942,13 +2942,22 @@
def err_illegal_initializer : Error<
"illegal initializer (only variables can be initialized)">;
def err_illegal_initializer_type : Error<"illegal initializer type %0">;
-def err_init_list_type_narrowing : Error<
+def err_init_list_type_narrowing_sfinae : Error<
"type %0 cannot be narrowed to %1 in initializer list">;
-def err_init_list_variable_narrowing : Error<
+def err_init_list_type_narrowing : Warning<
+ "type %0 cannot be narrowed to %1 in initializer list">,
+ InGroup<CXX11Narrowing>, DefaultError;
+def err_init_list_variable_narrowing_sfinae : Error<
"non-constant-expression cannot be narrowed from type %0 to %1 in "
"initializer list">;
-def err_init_list_constant_narrowing : Error<
+def err_init_list_variable_narrowing : Warning<
+ "non-constant-expression cannot be narrowed from type %0 to %1 in "
+ "initializer list">, InGroup<CXX11Narrowing>, DefaultError;
+def err_init_list_constant_narrowing_sfinae : Error<
"constant expression evaluates to %0 which cannot be narrowed to type %1">;
+def err_init_list_constant_narrowing : Warning<
+ "constant expression evaluates to %0 which cannot be narrowed to type %1">,
+ InGroup<CXX11Narrowing>, DefaultError;
def warn_init_list_type_narrowing : Warning<
"type %0 cannot be narrowed to %1 in initializer list in C++11">,
InGroup<CXX11Narrowing>, DefaultIgnore;
Modified: cfe/trunk/lib/Sema/SemaInit.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Sema/SemaInit.cpp?rev=148701&r1=148700&r2=148701&view=diff
==============================================================================
--- cfe/trunk/lib/Sema/SemaInit.cpp (original)
+++ cfe/trunk/lib/Sema/SemaInit.cpp Mon Jan 23 09:29:33 2012
@@ -5817,9 +5817,11 @@
// narrowing conversion even if the value is a constant and can be
// represented exactly as an integer.
S.Diag(PostInit->getLocStart(),
- S.getLangOptions().CPlusPlus0x && !S.getLangOptions().MicrosoftExt
- ? diag::err_init_list_type_narrowing
- : diag::warn_init_list_type_narrowing)
+ S.getLangOptions().MicrosoftExt || !S.getLangOptions().CPlusPlus0x?
+ diag::warn_init_list_type_narrowing
+ : S.isSFINAEContext()?
+ diag::err_init_list_type_narrowing_sfinae
+ : diag::err_init_list_type_narrowing)
<< PostInit->getSourceRange()
<< PreNarrowingType.getLocalUnqualifiedType()
<< EntityType.getLocalUnqualifiedType();
@@ -5828,9 +5830,11 @@
case NK_Constant_Narrowing:
// A constant value was narrowed.
S.Diag(PostInit->getLocStart(),
- S.getLangOptions().CPlusPlus0x && !S.getLangOptions().MicrosoftExt
- ? diag::err_init_list_constant_narrowing
- : diag::warn_init_list_constant_narrowing)
+ S.getLangOptions().MicrosoftExt || !S.getLangOptions().CPlusPlus0x?
+ diag::warn_init_list_constant_narrowing
+ : S.isSFINAEContext()?
+ diag::err_init_list_constant_narrowing_sfinae
+ : diag::err_init_list_constant_narrowing)
<< PostInit->getSourceRange()
<< ConstantValue.getAsString(S.getASTContext(), EntityType)
<< EntityType.getLocalUnqualifiedType();
@@ -5839,9 +5843,11 @@
case NK_Variable_Narrowing:
// A variable's value may have been narrowed.
S.Diag(PostInit->getLocStart(),
- S.getLangOptions().CPlusPlus0x && !S.getLangOptions().MicrosoftExt
- ? diag::err_init_list_variable_narrowing
- : diag::warn_init_list_variable_narrowing)
+ S.getLangOptions().MicrosoftExt || !S.getLangOptions().CPlusPlus0x?
+ diag::warn_init_list_variable_narrowing
+ : S.isSFINAEContext()?
+ diag::err_init_list_variable_narrowing_sfinae
+ : diag::err_init_list_variable_narrowing)
<< PostInit->getSourceRange()
<< PreNarrowingType.getLocalUnqualifiedType()
<< EntityType.getLocalUnqualifiedType();
Modified: cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-0x.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-0x.cpp?rev=148701&r1=148700&r2=148701&view=diff
==============================================================================
--- cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-0x.cpp (original)
+++ cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-0x.cpp Mon Jan 23 09:29:33 2012
@@ -193,3 +193,17 @@
// Template arguments make it harder to avoid printing qualifiers:
Agg<const unsigned char> c2 = {j}; // expected-error {{from type 'int' to 'const unsigned char' in}} expected-note {{override}}
}
+
+// Test SFINAE checks.
+template<unsigned> struct Value { };
+
+template<typename T>
+int &check_narrowed(Value<sizeof((T){1.1})>);
+
+template<typename T>
+float &check_narrowed(...);
+
+void test_narrowed(Value<sizeof(int)> vi, Value<sizeof(double)> vd) {
+ int &ir = check_narrowed<double>(vd);
+ float &fr = check_narrowed<int>(vi);
+}
Added: cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp?rev=148701&view=auto
==============================================================================
--- cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp (added)
+++ cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp Mon Jan 23 09:29:33 2012
@@ -0,0 +1,209 @@
+// RUN: %clang_cc1 -fsyntax-only -std=c++11 -Wno-error=c++11-narrowing -triple x86_64-apple-macosx10.6.7 -verify %s
+
+// Verify that narrowing conversions in initializer lists cause errors in C++0x
+// mode.
+
+void std_example() {
+ int x = 999; // x is not a constant expression
+ const int y = 999;
+ const int z = 99;
+ char c1 = x; // OK, though it might narrow (in this case, it does narrow)
+ char c2{x}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ char c3{y}; // expected-warning {{ cannot be narrowed }} expected-note {{override}} expected-warning {{changes value}}
+ char c4{z}; // OK: no narrowing needed
+ unsigned char uc1 = {5}; // OK: no narrowing needed
+ unsigned char uc2 = {-1}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ unsigned int ui1 = {-1}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ signed int si1 =
+ { (unsigned int)-1 }; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ int ii = {2.0}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ float f1 { x }; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ float f2 { 7 }; // OK: 7 can be exactly represented as a float
+ int f(int);
+ int a[] =
+ { 2, f(2), f(2.0) }; // OK: the double-to-int conversion is not at the top level
+}
+
+// Test each rule individually.
+
+template<typename T>
+struct Agg {
+ T t;
+};
+
+template<typename T>
+struct Convert {
+ constexpr Convert(T v) : v(v) {}
+ constexpr operator T() const { return v; }
+ T v;
+};
+template<typename T> Convert<T> ConvertVar();
+
+// C++0x [dcl.init.list]p7: A narrowing conversion is an implicit conversion
+//
+// * from a floating-point type to an integer type, or
+
+void float_to_int() {
+ Agg<char> a1 = {1.0F}; // expected-warning {{type 'float' cannot be narrowed to 'char'}} expected-note {{override}}
+ Agg<char> a2 = {1.0}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<char> a3 = {1.0L}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+
+ float f = 1.0;
+ double d = 1.0;
+ long double ld = 1.0;
+ Agg<char> a4 = {f}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<char> a5 = {d}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<char> a6 = {ld}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+
+ Agg<char> ce1 = { Convert<float>(1.0) }; // expected-warning {{type 'float' cannot be narrowed to 'char'}} expected-note {{override}}
+ Agg<char> ce2 = { ConvertVar<double>() }; // expected-warning {{type 'double' cannot be narrowed to 'char'}} expected-note {{override}}
+}
+
+// * from long double to double or float, or from double to float, except where
+// the source is a constant expression and the actual value after conversion
+// is within the range of values that can be represented (even if it cannot be
+// represented exactly), or
+
+void shrink_float() {
+ // These aren't constant expressions.
+ float f = 1.0;
+ double d = 1.0;
+ long double ld = 1.0;
+
+ // Variables.
+ Agg<float> f1 = {f}; // OK (no-op)
+ Agg<float> f2 = {d}; // expected-warning {{non-constant-expression cannot be narrowed from type 'double' to 'float'}} expected-note {{override}}
+ Agg<float> f3 = {ld}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ // Exact constants.
+ Agg<float> f4 = {1.0}; // OK (double constant represented exactly)
+ Agg<float> f5 = {1.0L}; // OK (long double constant represented exactly)
+ // Inexact but in-range constants.
+ Agg<float> f6 = {0.1}; // OK (double constant in range but rounded)
+ Agg<float> f7 = {0.1L}; // OK (long double constant in range but rounded)
+ // Out of range constants.
+ Agg<float> f8 = {1E50}; // expected-warning {{constant expression evaluates to 1.000000e+50 which cannot be narrowed to type 'float'}} expected-note {{override}}
+ Agg<float> f9 = {1E50L}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ // More complex constant expression.
+ constexpr long double e40 = 1E40L, e30 = 1E30L, e39 = 1E39L;
+ Agg<float> f10 = {e40 - 5 * e39 + e30 - 5 * e39}; // OK
+
+ // Variables.
+ Agg<double> d1 = {f}; // OK (widening)
+ Agg<double> d2 = {d}; // OK (no-op)
+ Agg<double> d3 = {ld}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ // Exact constant.
+ Agg<double> d4 = {1.0L}; // OK (long double constant represented exactly)
+ // Inexact but in-range constant.
+ Agg<double> d5 = {0.1L}; // OK (long double constant in range but rounded)
+ // Out of range constant.
+ Agg<double> d6 = {1E315L}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ // More complex constant expression.
+ constexpr long double e315 = 1E315L, e305 = 1E305L, e314 = 1E314L;
+ Agg<double> d7 = {e315 - 5 * e314 + e305 - 5 * e314}; // OK
+
+ Agg<float> ce1 = { Convert<double>(1e300) }; // expected-warning {{constant expression evaluates to 1.000000e+300 which cannot be narrowed to type 'float'}} expected-note {{override}}
+ Agg<double> ce2 = { ConvertVar<long double>() }; // expected-warning {{non-constant-expression cannot be narrowed from type 'long double' to 'double'}} expected-note {{override}}
+}
+
+// * from an integer type or unscoped enumeration type to a floating-point type,
+// except where 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, or
+void int_to_float() {
+ // Not a constant expression.
+ char c = 1;
+
+ // Variables. Yes, even though all char's will fit into any floating type.
+ Agg<float> f1 = {c}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<double> f2 = {c}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<long double> f3 = {c}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+
+ // Constants.
+ Agg<float> f4 = {12345678}; // OK (exactly fits in a float)
+ Agg<float> f5 = {123456789}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+
+ Agg<float> ce1 = { Convert<int>(123456789) }; // expected-warning {{constant expression evaluates to 123456789 which cannot be narrowed to type 'float'}} expected-note {{override}}
+ Agg<double> ce2 = { ConvertVar<long long>() }; // expected-warning {{non-constant-expression cannot be narrowed from type 'long long' to 'double'}} expected-note {{override}}
+}
+
+// * 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 conversion will
+// fit into the target type and will produce the original value when converted
+// back to the original type.
+void shrink_int() {
+ // Not a constant expression.
+ short s = 1;
+ unsigned short us = 1;
+ Agg<char> c1 = {s}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<unsigned short> s1 = {s}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<short> s2 = {us}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+
+ // "that cannot represent all the values of the original type" means that the
+ // validity of the program depends on the relative sizes of integral types.
+ // This test compiles with -m64, so sizeof(int)<sizeof(long)==sizeof(long
+ // long).
+ long l1 = 1;
+ Agg<int> i1 = {l1}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ long long ll = 1;
+ Agg<long> l2 = {ll}; // OK
+
+ // Constants.
+ Agg<char> c2 = {127}; // OK
+ Agg<char> c3 = {300}; // expected-warning {{ cannot be narrowed }} expected-note {{override}} expected-warning {{changes value}}
+
+ Agg<int> i2 = {0x7FFFFFFFU}; // OK
+ Agg<int> i3 = {0x80000000U}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<unsigned int> i4 = {-0x80000000L}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+
+ // Bool is also an integer type, but conversions to it are a different AST
+ // node.
+ Agg<bool> b1 = {0}; // OK
+ Agg<bool> b2 = {1}; // OK
+ Agg<bool> b3 = {-1}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+
+ // Conversions from pointers to booleans aren't narrowing conversions.
+ Agg<bool> b = {&b1}; // OK
+
+ Agg<short> ce1 = { Convert<int>(100000) }; // expected-warning {{constant expression evaluates to 100000 which cannot be narrowed to type 'short'}} expected-note {{override}} expected-warning {{changes value from 100000 to -31072}}
+ Agg<char> ce2 = { ConvertVar<short>() }; // expected-warning {{non-constant-expression cannot be narrowed from type 'short' to 'char'}} expected-note {{override}}
+}
+
+// Be sure that type- and value-dependent expressions in templates get the warning
+// too.
+
+template<int I, typename T>
+void maybe_shrink_int(T t) {
+ Agg<short> s1 = {t}; // expected-warning {{ cannot be narrowed }} expected-note {{override}}
+ Agg<short> s2 = {I}; // expected-warning {{ cannot be narrowed }} expected-note {{override}} expected-warning {{changes value}}
+ Agg<T> t2 = {700}; // expected-warning {{ cannot be narrowed }} expected-note {{override}} expected-warning {{changes value}}
+}
+
+void test_template() {
+ maybe_shrink_int<15>((int)3); // expected-note {{in instantiation}}
+ maybe_shrink_int<70000>((char)3); // expected-note {{in instantiation}}
+}
+
+
+// We don't want qualifiers on the types in the diagnostic.
+
+void test_qualifiers(int i) {
+ const int j = i;
+ struct {const unsigned char c;} c1 = {j}; // expected-warning {{from type 'int' to 'unsigned char' in}} expected-note {{override}}
+ // Template arguments make it harder to avoid printing qualifiers:
+ Agg<const unsigned char> c2 = {j}; // expected-warning {{from type 'int' to 'const unsigned char' in}} expected-note {{override}}
+}
+
+// Make sure we still get the right SFINAE behavior.
+template<unsigned> struct Value { };
+
+template<typename T>
+int &check_narrowed(Value<sizeof((T){1.1})>);
+
+template<typename T>
+float &check_narrowed(...);
+
+void test_narrowed(Value<sizeof(int)> vi, Value<sizeof(double)> vd) {
+ int &ir = check_narrowed<double>(vd);
+ float &fr = check_narrowed<int>(vi);
+}
Propchange: cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp
------------------------------------------------------------------------------
svn:eol-style = native
Propchange: cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp
------------------------------------------------------------------------------
svn:keywords = Id
Propchange: cfe/trunk/test/CXX/dcl.decl/dcl.init/dcl.init.list/p7-cxx11-nowarn.cpp
------------------------------------------------------------------------------
svn:mime-type = text/plain
More information about the cfe-commits
mailing list