Boost logo

Boost-Commit :

Subject: [Boost-commit] svn:boost r74212 - sandbox/e_float/libs/e_float/test/e_float_numerical_test
From: pbristow_at_[hidden]
Date: 2011-09-03 12:32:18


Author: pbristow
Date: 2011-09-03 12:32:17 EDT (Sat, 03 Sep 2011)
New Revision: 74212
URL: http://svn.boost.org/trac/boost/changeset/74212

Log:
All numerical tests included using Boost.Test.
Added:
   sandbox/e_float/libs/e_float/test/e_float_numerical_test/ (props changed)
   sandbox/e_float/libs/e_float/test/e_float_numerical_test/e_float_test.cpp (contents, props changed)

Added: sandbox/e_float/libs/e_float/test/e_float_numerical_test/e_float_test.cpp
==============================================================================
--- (empty file)
+++ sandbox/e_float/libs/e_float/test/e_float_numerical_test/e_float_test.cpp 2011-09-03 12:32:17 EDT (Sat, 03 Sep 2011)
@@ -0,0 +1,309 @@
+//! \file
+// e_float_test.cpp
+
+// Copyright Paul A. Bristow 2011.
+// Copyright Christopher Kormanyos 2011.
+
+// Use, modification and distribution are subject to 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)
+
+// Test of using e_float and Boost.Check macro/functions.
+
+#define BOOST_TEST_MAIN
+#define BOOST_LIB_DIAGNOSTIC "on"// Show library file details.
+// Linking to lib file: libboost_unit_test_framework-vc100-mt-gd-1_48.lib (trunk at Jul 11)
+
+// #define E_FLOAT_DIGITS10 50 as command line using MSVC or in jamfile (but needs b2 -a option?)
+
+#include <iostream>
+using std::cout;
+using std::cin;
+using std::endl;
+using std::dec;
+using std::hex;
+using std::boolalpha;
+using std::scientific;
+using std::fixed;
+using std::defaultfloat;
+using std::showpos;
+using std::showpoint;
+
+#include <iomanip>
+using std::setprecision;
+using std::setw;
+#include <string>
+using std::string;
+#include <sstream>
+//using std::istringstream;
+//using std::ostringstream
+
+#include <boost/test/unit_test.hpp> // Enhanced for unit_test framework autolink.
+#include <boost/test/floating_point_comparison.hpp>
+
+#include <boost/e_float/e_float.hpp>
+#include <boost/e_float/e_float_constants.hpp>
+
+namespace test
+{
+ namespace real
+ {
+ // See \boost-sandbox\e_float\libs\e_float\test\real\test_real.cpp and /cases/ *.cpp
+ // \boost-sandbox\e_float\libs\e_float\test\real\cases\test_case_0000x_overflow_underflow.cpp
+ bool test_case_00001_overflow_mul_x(const bool);
+ bool test_case_00002_underflow_mul_x (const bool);
+ bool test_case_00003_overflow_x_mul_by_n(const bool);
+ bool test_case_00004_underflow_x_div_by_n(const bool);
+ // \boost-sandbox\e_float\libs\e_float\test\real\cases\test_case_0000y_write_to_ostream.cpp
+ bool test_case_00006_write_os_floatfield_fixed(const bool);
+ bool test_case_00007_write_os_floatfield_scientific(const bool);
+ bool test_case_00008_write_os_floatfield_not_set(const bool);
+ // \boost-sandbox\e_float\libs\e_float\test\real\cases\test_case_0000z_global_ops_pod.cpp
+ bool test_case_00009_global_ops_pod_equate (const bool);
+ bool test_case_00010_global_ops_pod_operations (const bool);
+ // \boost-sandbox\e_float\libs\e_float\test\real\cases\test_case_00011_various_elem_math.cpp
+ bool test_case_00011_various_elem_math (const bool);
+ bool test_case_00021_bernoulli (const bool);
+ bool test_case_00051_factorial (const bool);
+ bool test_case_00052_factorial2 (const bool);
+ bool test_case_00071_various_int_func (const bool);
+ bool test_case_00101_sin (const bool);
+ bool test_case_00102_cos (const bool);
+ bool test_case_00103_exp (const bool);
+ bool test_case_00104_log (const bool);
+ bool test_case_00105_sqrt (const bool);
+ bool test_case_00106_rootn (const bool);
+ bool test_case_00111_sin_small_x (const bool);
+ bool test_case_00112_cos_x_near_pi_half (const bool);
+ bool test_case_00113_atan_x_small_to_large (const bool);
+ bool test_case_00114_various_trig (const bool);
+ bool test_case_00115_various_elem_trans (const bool);
+ bool test_case_00121_sinh (const bool);
+ bool test_case_00122_cosh (const bool);
+ bool test_case_00123_tanh (const bool);
+ bool test_case_00124_asinh (const bool);
+ bool test_case_00125_acosh (const bool);
+ bool test_case_00126_atanh (const bool);
+ bool test_case_00201_gamma (const bool);
+ bool test_case_00202_gamma_medium_x (const bool);
+ bool test_case_00203_gamma_small_x (const bool);
+ bool test_case_00204_gamma_tiny_x (const bool);
+ bool test_case_00205_gamma_near_neg_n (const bool);
+ bool test_case_00221_various_gamma_func (const bool);
+ bool test_case_00901_zeta_small_x (const bool);
+ bool test_case_00902_zeta_all_x (const bool);
+ bool test_case_00903_zeta_neg_x (const bool);
+ } // namespace real
+} // namespace test
+
+
+//Macros to Check using manips output expected string result, for example:
+// CHECK_OUT(hex << showbase << setw(10) << i, " 0xf")
+// CHECK_OUT(scientific << setw(20) << d, " 1.234568e+001");
+
+// Compare a output with expected result.
+#define CHECK_OUT(manips, result)\
+{ \
+ std::ostringstream oss;\
+ oss << manips;\
+ BOOST_CHECK_EQUAL(oss.str(), result);\
+}
+
+#define CHECK_LENGTH(manips, result)\
+{\
+ BOOST_CHECK_EQUAL(oss.str().length(), strlen(result));\
+}\
+
+// Compare results of reading string in,
+#define CHECK_IN(in, value)\
+{\
+ e_float r;\
+ std::istringstream iss(in);\
+ iss >> r;\
+ BOOST_CHECK_CLOSE_FRACTION(r, value, std::numeric_limits<e_float>::epsilon());\
+} // #define CHECK_IN(in, value, sd, df, types)
+
+// CHECK_OUT_IN Output via manips, and read back in, check is same. 'Loopback'.
+#define CHECK_OUT_IN(manips, result, value)\
+{\
+ std::stringstream ss;\
+ ss << manips;\
+ BOOST_CHECK_EQUAL(ss.str(), result);\
+ e_float r;\
+ ss >> r;\
+ BOOST_CHECK_CLOSE_FRACTION(r, value, std::numeric_limits<e_float>::epsilon());\
+}// #define CHECK_OUT_IN(manips, result)
+
+
+// Must #define E_FLOAT_TYPE_EFX; in project properties.
+
+// Note: exact double means exactly representable as double, for example: 0.5,
+// but NOT 0.1.
+// Integral may be an integer or a double having an integral value.
+
+BOOST_AUTO_TEST_CASE(e_float_test_template)
+{ // These are just examples of using Boost.Test.
+ // Need to be removed when no longer helpful.
+ BOOST_TEST_MESSAGE("Test Boost.e_float"); // Only appears if command line has --log_level="message"
+
+ string m = "Test with ";
+ m+= __FILE__;
+ m+= " edited ";
+ m += __TIMESTAMP__ ".\n";
+ BOOST_TEST_MESSAGE(m);
+ BOOST_CHECK(true);
+ BOOST_CHECK_EQUAL(1, 1);
+ BOOST_CHECK_NE(1, -1);
+ BOOST_CHECK_CLOSE_FRACTION(1.0, 1.0+std::numeric_limits<double>::epsilon(), std::numeric_limits<double>::epsilon());
+ double d = 123.456789;
+ // cout << scientific << d << endl; // Outputs: "1.234568e+003"
+ CHECK_OUT(d, "123.457"); // Default. == << std::defaultfloat
+ CHECK_OUT(defaultfloat << d, "123.457"); // Default. == << std::defaultfloat
+ string ddef = "123.457"; // default float output.
+ CHECK_OUT(d, ddef); // Default. == << std::defaultfloat
+ CHECK_OUT(defaultfloat << d, ddef); // Default. == << std::defaultfloat
+ CHECK_OUT(scientific << d, "1.234568e+002");
+ CHECK_OUT(fixed << d, "123.456789");
+ int m1 = -1; // negative variable.
+ CHECK_OUT(m1, "-1"); // negative constant.
+ CHECK_OUT(hex << m1, "ffffffff"); // with hex manipulator.
+ // Checking input with inline.
+ e_float r;
+ std::istringstream iss("123.456");
+ iss >> r;
+ BOOST_CHECK_CLOSE_FRACTION(r, e_float("123.456"), std::numeric_limits<e_float>::epsilon());
+ BOOST_CHECK_EQUAL(r, e_float("123.456")); // Also works
+ // Repeat same test using CHECK_IN macro defined above.
+ CHECK_IN("123.456", e_float("123.456"));
+
+ //CHECK_IN("123.456", e_float(123.456)); // Mistaken conversion from less accurate double.
+ // Fails r{123.456} and e_float(123.456){123.4560000000000030695446184836328029632568359375}
+ // differ by more than 1e-49.
+
+ // CHECK_OUT_IN Output via manips, and read back in, check is same. 'Loopback'.
+ // #define CHECK_OUT_IN(manips, result, value)
+ { // Integer example.
+ int i = 255;
+ std::string result = "ff";
+ int value = 255;
+ std::stringstream ss;
+ ss << hex << i;
+ BOOST_CHECK_EQUAL(ss.str(), result);
+ int read;
+ ss >> read;
+ BOOST_CHECK_EQUAL(read, value);
+ }
+ { // double example
+ double w = 123.456;
+ std::string result = "123.456000";
+ double value = 123.456;
+ std::stringstream ss;
+ ss << std::fixed << w;
+ //cout << ss.str() << endl;
+ BOOST_CHECK_EQUAL(ss.str(), result);
+ double read;
+ ss >> read;
+ BOOST_CHECK_EQUAL(read, value);
+ BOOST_CHECK_CLOSE_FRACTION(read, value, std::numeric_limits<double>::epsilon());
+ }
+ { // e_float example
+ e_float w("123.456");
+ std::string result = "123.456000"; // double result was "123.456000";
+ e_float value("123.456");
+ std::stringstream ss;
+ ss << std::fixed << w;
+ //cout << ss.str() << endl;
+ BOOST_CHECK_EQUAL(ss.str(), result);
+ e_float read;
+ ss >> read;
+ BOOST_CHECK_EQUAL(read, value);
+ BOOST_CHECK_CLOSE_FRACTION(read, value, std::numeric_limits<e_float>::epsilon());
+
+ CHECK_OUT_IN(std::fixed << w, result, value);
+ }
+
+} // BOOST_AUTO_TEST_CASE(e_float_template)
+
+BOOST_AUTO_TEST_CASE(e_float_test_macros)
+{ // Check some macro values.
+ //
+ BOOST_CHECK_EQUAL(E_FLOAT_DIGITS10, 50); // Assumes we are testing at 50 digits, NOT the default.
+} // BOOST_AUTO_TEST_CASE(e_float_test_macros)
+
+
+BOOST_AUTO_TEST_CASE(e_float_test_ios)
+{ // Check some IOS defaults.
+ BOOST_CHECK_EQUAL(cout.precision(), 6);
+ std::ostringstream oss;
+ BOOST_CHECK_EQUAL(oss.precision(), 6);
+ }
+
+BOOST_AUTO_TEST_CASE(e_float_test_input)
+{ //
+ BOOST_TEST_MESSAGE("Test Boost.e_float input.");
+
+ CHECK_IN("2", e_float("2"));
+ CHECK_IN("-2", e_float("-2"));
+ CHECK_IN("+2", e_float("+2"));
+ CHECK_IN("-2.", e_float("-2."));
+
+ CHECK_IN("123.456", e_float("123.456"));
+ CHECK_IN("0.0123456", e_float("0.0123456"));
+ CHECK_IN("1e-6", e_float("1e-6"));
+ CHECK_IN("-1e-6", e_float("-1e-6"));
+} // BOOST_AUTO_TEST_CASE(e_float_test_input)
+
+BOOST_AUTO_TEST_CASE(e_float_tests)
+{ // Original real tests from \e_float\libs\e_float\test\real\cases
+ BOOST_TEST_MESSAGE("Test Boost.e_float numerical.");
+
+ BOOST_CHECK(test::real::test_case_00001_overflow_mul_x (false));
+ BOOST_CHECK(test::real::test_case_00002_underflow_mul_x (false));
+ BOOST_CHECK(test::real::test_case_00003_overflow_x_mul_by_n (false));
+ BOOST_CHECK(test::real::test_case_00004_underflow_x_div_by_n (false));
+ BOOST_CHECK(test::real::test_case_00006_write_os_floatfield_fixed (false));
+ BOOST_CHECK(test::real::test_case_00007_write_os_floatfield_scientific (false));
+ BOOST_CHECK(test::real::test_case_00008_write_os_floatfield_not_set (false));
+ BOOST_CHECK( test::real::test_case_00001_overflow_mul_x (false));
+ BOOST_CHECK( test::real::test_case_00002_underflow_mul_x (false));
+ BOOST_CHECK( test::real::test_case_00003_overflow_x_mul_by_n (false));
+ BOOST_CHECK( test::real::test_case_00004_underflow_x_div_by_n (false));
+ BOOST_CHECK( test::real::test_case_00006_write_os_floatfield_fixed (false));
+ BOOST_CHECK( test::real::test_case_00007_write_os_floatfield_scientific(false));
+ BOOST_CHECK( test::real::test_case_00008_write_os_floatfield_not_set (false));
+ BOOST_CHECK( test::real::test_case_00009_global_ops_pod_equate (false));
+ BOOST_CHECK( test::real::test_case_00010_global_ops_pod_operations (false));
+ BOOST_CHECK( test::real::test_case_00011_various_elem_math (false));
+ BOOST_CHECK( test::real::test_case_00021_bernoulli (false));
+ BOOST_CHECK( test::real::test_case_00051_factorial (false));
+ BOOST_CHECK( test::real::test_case_00052_factorial2 (false));
+ BOOST_CHECK( test::real::test_case_00071_various_int_func (false));
+ BOOST_CHECK( test::real::test_case_00101_sin (false));
+ BOOST_CHECK( test::real::test_case_00102_cos (false));
+ BOOST_CHECK( test::real::test_case_00103_exp (false));
+ BOOST_CHECK( test::real::test_case_00104_log (false));
+ BOOST_CHECK( test::real::test_case_00105_sqrt (false));
+ BOOST_CHECK( test::real::test_case_00106_rootn (false));
+ BOOST_CHECK( test::real::test_case_00111_sin_small_x (false));
+ BOOST_CHECK( test::real::test_case_00112_cos_x_near_pi_half (false));
+ BOOST_CHECK( test::real::test_case_00113_atan_x_small_to_large (false));
+ BOOST_CHECK( test::real::test_case_00114_various_trig (false));
+ BOOST_CHECK( test::real::test_case_00115_various_elem_trans (false));
+ BOOST_CHECK( test::real::test_case_00121_sinh (false));
+ BOOST_CHECK( test::real::test_case_00122_cosh (false));
+ BOOST_CHECK( test::real::test_case_00123_tanh (false));
+ BOOST_CHECK( test::real::test_case_00124_asinh (false));
+ BOOST_CHECK( test::real::test_case_00125_acosh (false));
+ BOOST_CHECK( test::real::test_case_00126_atanh (false));
+ BOOST_CHECK( test::real::test_case_00201_gamma (false));
+ BOOST_CHECK( test::real::test_case_00202_gamma_medium_x (false));
+ BOOST_CHECK( test::real::test_case_00203_gamma_small_x (false));
+ BOOST_CHECK( test::real::test_case_00204_gamma_tiny_x (false));
+ BOOST_CHECK( test::real::test_case_00205_gamma_near_neg_n (false));
+ BOOST_CHECK( test::real::test_case_00221_various_gamma_func (false));
+ BOOST_CHECK( test::real::test_case_00901_zeta_small_x (false));
+ BOOST_CHECK( test::real::test_case_00902_zeta_all_x (false));
+ BOOST_CHECK( test::real::test_case_00903_zeta_neg_x (false));
+
+} // BOOST_AUTO_TEST_CASE(e_float_tests)


Boost-Commit list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk