📄 udt_support_test.cpp
字号:
// (C) Copyright 2003, Fernando Luis Cacciola Carballal.//// Distributed under the Boost Software License, Version 1.0. (See// accompanying file LICENSE_1_0.txt or copy at// http://www.boost.org/LICENSE_1_0.txt)////#include<iostream>#include<iomanip>#include<string>#include<typeinfo>#include<vector>#include<algorithm>#include "boost/numeric/conversion/converter.hpp"#ifdef __BORLANDC__#pragma hdrstop#endif#include "test_helpers.cpp"#include "test_helpers2.cpp"#include "test_helpers3.cpp"using namespace std ;using namespace boost ;using namespace numeric ;using namespace MyUDT ;//-------------------------------------------------------------------------// These are the typical steps that are required to install support for// conversions from/to UDT which need special treatment.//-------------------------------------------------------------------------//// (1) Instantiate specific convesions traits.// This step is only for convenience.// These traits instances are required in order to define the specializations// that follow (and which *are required* to make the library work with MyInt and MyFloat)//namespace MyUDT {typedef conversion_traits<double , MyFloat> MyFloat_to_double_Traits;typedef conversion_traits<int , MyFloat> MyFloat_to_int_Traits;typedef conversion_traits<MyInt , MyFloat> MyFloat_to_MyInt_Traits;typedef conversion_traits<int , MyInt > MyInt_to_int_Traits;typedef conversion_traits<MyFloat, MyInt > MyInt_to_MyFloat_Traits;typedef conversion_traits<MyInt , double > double_to_MyInt_Traits;} // namespace MyUDT//// (2) Define suitable raw converters.//// Our sample UDTs don't support implicit conversions.// Therefore, the default raw_converter<> doesn't work,// and we need to define our own.//// There are two ways of doing this://// (a) One is to simply specialize boost::numeric::raw_converter<> directly.// This way, the default converter will work out of the box, which means, for instance,// that numeric_cast<> can be used with these UDTs.//// (b) Define a user class with the appropriate interface and supply it explicitely// as a policy to a converter instance.//// This test uses chice (a).//namespace boost {namespace numeric {template<>struct raw_converter<MyUDT::MyFloat_to_double_Traits>{ static double low_level_convert ( MyUDT::MyFloat const& s ) { return s.to_builtin() ; }} ;template<>struct raw_converter<MyUDT::MyFloat_to_int_Traits>{ static int low_level_convert ( MyUDT::MyFloat const& s ) { return static_cast<int>( s.to_builtin() ) ; }} ;template<>struct raw_converter<MyUDT::MyFloat_to_MyInt_Traits>{ static MyUDT::MyInt low_level_convert ( MyUDT::MyFloat const& s ) { return MyUDT::MyInt( static_cast<int>(s.to_builtin()) ) ; }} ;template<>struct raw_converter<MyUDT::MyInt_to_int_Traits>{ static int low_level_convert ( MyUDT::MyInt const& s ) { return s.to_builtin() ; }} ;template<>struct raw_converter<MyUDT::MyInt_to_MyFloat_Traits>{ static MyUDT::MyFloat low_level_convert ( MyUDT::MyInt const& s ) { return MyUDT::MyFloat( static_cast<double>(s.to_builtin()) ) ; }} ;template<>struct raw_converter<MyUDT::double_to_MyInt_Traits>{ static MyUDT::MyInt low_level_convert ( double s ) { return MyUDT::MyInt( static_cast<int>(s) ) ; }} ;} // namespace numeric} // namespace boost//// (3) Define suitable range checkers//// By default, if a UDT is involved in a conversion, internal range checking is disabled.// This is so because a UDT type can have any sort of range, even unbounded, thus// the library doesn't attempt to automatically figure out the appropriate range checking logic.// (as it does when builtin types are involved)// However, this situation is a bit unsufficient in practice, specially from doing narrowing (subranged)// conversions from UDTs.// The library provides a rudimentary hook to help this out: The user can plug in his own// range checker to the converter instance.//// This test shows how to define and use a custom range checker.//namespace MyUDT {//// The following are metaprogramming tools to allow us the implement the// MyCustomRangeChecker generically, for either builtin or UDT types.//// get_builtin_type<N>::type extracts the built-in type of our UDT's//template<class N> struct get_builtin_type { typedef N type ; } ;template<> struct get_builtin_type<MyInt> { typedef int type ; } ;template<> struct get_builtin_type<MyFloat> { typedef double type ; } ;// U extract_builtin ( T s ) returns 's' converted to the corresponding built-in type U.// template<class N>struct extract_builtin{ static N apply ( N n ) { return n ; }} ;template<>struct extract_builtin<MyInt>{ static int apply ( MyInt const& n ) { return n.to_builtin() ; }} ;template<>struct extract_builtin<MyFloat>{ static double apply ( MyFloat const& n ) { return n.to_builtin() ; }} ;template<class Traits>struct MyCustomRangeChecker{ typedef typename Traits::argument_type argument_type ; // This custom range checker uses the fact that our 'fake' UDT are merely wrappers // around builtin types; so it just forward the logic to the correspoding range // checkers for the wrapped builtin types. // typedef typename Traits::source_type S ; typedef typename Traits::target_type T ; // NOTE: S and/or T can be either UDT or builtin types. typedef typename get_builtin_type<S>::type builtinS ; typedef typename get_builtin_type<T>::type builtinT ; // NOTE: The internal range checker used by default is *built* when you instantiate // a converter<> with a given Traits according to the properties of the involved types. // Currently, there is no way to instantiate this range checker as a separate class. // However, you can see it as part of the interface of the converter // (since the converter inherits from it) // Therefore, here we instantiate a converter corresponding to the builtin types to access // their associated builtin range checker. // typedef boost::numeric::converter<builtinT,builtinS> InternalConverter ; static range_check_result out_of_range ( argument_type s ) { return InternalConverter::out_of_range( extract_builtin<S>::apply(s) ); } static void validate_range ( argument_type s ) { return InternalConverter::validate_range( extract_builtin<S>::apply(s) ); }} ;} // namespace MyUDT//// Test here//void test_udt_conversions_with_defaults(){ cout << "Testing UDT conversion with default policies\n" ; // MyInt <--> int int mibv = rand(); MyInt miv(mibv); TEST_SUCCEEDING_CONVERSION_DEF(MyInt,int,miv,mibv); TEST_SUCCEEDING_CONVERSION_DEF(int,MyInt,mibv,miv); // MyFloat <--> double double mfbv = static_cast<double>(rand()) / 3.0 ; MyFloat mfv (mfbv); TEST_SUCCEEDING_CONVERSION_DEF(MyFloat,double,mfv,mfbv); TEST_SUCCEEDING_CONVERSION_DEF(double,MyFloat,mfbv,mfv); // MyInt <--> MyFloat MyInt miv2 ( static_cast<int>(mfbv) ); MyFloat miv2F ( static_cast<int>(mfbv) ); MyFloat mfv2 ( static_cast<double>(mibv) ); MyInt mfv2I ( static_cast<double>(mibv) ); TEST_SUCCEEDING_CONVERSION_DEF(MyFloat,MyInt,miv2F,miv2); TEST_SUCCEEDING_CONVERSION_DEF(MyInt,MyFloat,mfv2I,mfv2);}template<class T, class S>struct GenerateCustomConverter{ typedef conversion_traits<T,S> Traits; typedef def_overflow_handler OverflowHandler ; typedef Trunc<S> Float2IntRounder ; typedef raw_converter<Traits> RawConverter ; typedef MyCustomRangeChecker<Traits> RangeChecker ; typedef converter<T,S,Traits,OverflowHandler,Float2IntRounder,RawConverter,RangeChecker> type ;} ;void test_udt_conversions_with_custom_range_checking(){ cout << "Testing UDT conversions with custom range checker\n" ; int mibv = rand(); MyFloat mfv ( static_cast<double>(mibv) ); typedef GenerateCustomConverter<MyFloat,int>::type int_to_MyFloat_Conv ; TEST_SUCCEEDING_CONVERSION( int_to_MyFloat_Conv, MyFloat, int, mfv, mibv ); int mibv2 = rand(); MyInt miv (mibv2); MyFloat mfv2 ( static_cast<double>(mibv2) ); typedef GenerateCustomConverter<MyFloat,MyInt>::type MyInt_to_MyFloat_Conv ; TEST_SUCCEEDING_CONVERSION( MyInt_to_MyFloat_Conv, MyFloat, MyInt, mfv2, miv ); double mfbv = bounds<double>::highest(); typedef GenerateCustomConverter<MyInt,double>::type double_to_MyInt_Conv ; TEST_POS_OVERFLOW_CONVERSION( double_to_MyInt_Conv, MyInt, double, mfbv ); MyFloat mfv3 ( bounds<double>::lowest() ) ; typedef GenerateCustomConverter<int,MyFloat>::type MyFloat_to_int_Conv ; TEST_NEG_OVERFLOW_CONVERSION( MyFloat_to_int_Conv, int, MyFloat, mfv3 );}int test_main( int, char* [] ){ cout << setprecision( numeric_limits<long double>::digits10 ) ; test_udt_conversions_with_defaults(); test_udt_conversions_with_custom_range_checking(); return 0;}
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -