diff options
Diffstat (limited to 'indra/llcommon/tests/llunits_test.cpp')
-rw-r--r-- | indra/llcommon/tests/llunits_test.cpp | 388 |
1 files changed, 388 insertions, 0 deletions
diff --git a/indra/llcommon/tests/llunits_test.cpp b/indra/llcommon/tests/llunits_test.cpp new file mode 100644 index 0000000000..57cf9810af --- /dev/null +++ b/indra/llcommon/tests/llunits_test.cpp @@ -0,0 +1,388 @@ +/** + * @file llsingleton_test.cpp + * @date 2011-08-11 + * @brief Unit test for the LLSingleton class + * + * $LicenseInfo:firstyear=2011&license=viewerlgpl$ + * Second Life Viewer Source Code + * Copyright (C) 2011, Linden Research, Inc. + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; + * version 2.1 of the License only. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + * + * Linden Research, Inc., 945 Battery Street, San Francisco, CA 94111 USA + * $/LicenseInfo$ + */ + +#include "linden_common.h" + +#include "llunits.h" +#include "../test/lltut.h" + +namespace LLUnits +{ + // using powers of 2 to allow strict floating point equality + LL_DECLARE_BASE_UNIT(Quatloos, "Quat"); + LL_DECLARE_DERIVED_UNIT(Latinum, "Lat", Quatloos, / 4); + LL_DECLARE_DERIVED_UNIT(Solari, "Sol", Latinum, * 16); +} + +LL_DECLARE_UNIT_TYPEDEFS(LLUnits, Quatloos); +LL_DECLARE_UNIT_TYPEDEFS(LLUnits, Latinum); +LL_DECLARE_UNIT_TYPEDEFS(LLUnits, Solari); + +namespace LLUnits +{ + LL_DECLARE_BASE_UNIT(Celcius, "c"); + LL_DECLARE_DERIVED_UNIT(Fahrenheit, "f", Celcius, * 9 / 5 + 32); + LL_DECLARE_DERIVED_UNIT(Kelvin, "k", Celcius, + 273.15f); +} + +LL_DECLARE_UNIT_TYPEDEFS(LLUnits, Celcius); +LL_DECLARE_UNIT_TYPEDEFS(LLUnits, Fahrenheit); +LL_DECLARE_UNIT_TYPEDEFS(LLUnits, Kelvin); + + +namespace tut +{ + using namespace LLUnits; + struct units + { + }; + + typedef test_group<units> units_t; + typedef units_t::object units_object_t; + tut::units_t tut_singleton("LLUnit"); + + // storage type conversions + template<> template<> + void units_object_t::test<1>() + { + LLUnit<F32, Quatloos> float_quatloos; + ensure("default float unit is zero", float_quatloos == F32Quatloos(0.f)); + + LLUnit<F32, Quatloos> float_initialize_quatloos(1); + ensure("non-zero initialized unit", float_initialize_quatloos == F32Quatloos(1.f)); + + LLUnit<S32, Quatloos> int_quatloos; + ensure("default int unit is zero", int_quatloos == S32Quatloos(0)); + + int_quatloos = S32Quatloos(42); + ensure("int assignment is preserved", int_quatloos == S32Quatloos(42)); + float_quatloos = int_quatloos; + ensure("float assignment from int preserves value", float_quatloos == F32Quatloos(42.f)); + + int_quatloos = float_quatloos; + ensure("int assignment from float preserves value", int_quatloos == S32Quatloos(42)); + + float_quatloos = F32Quatloos(42.1f); + int_quatloos = float_quatloos; + ensure("int units truncate float units on assignment", int_quatloos == S32Quatloos(42)); + + LLUnit<U32, Quatloos> unsigned_int_quatloos(float_quatloos); + ensure("unsigned int can be initialized from signed int", unsigned_int_quatloos == S32Quatloos(42)); + + S32Solari int_solari(1); + + float_quatloos = int_solari; + ensure("fractional units are preserved in conversion from integer to float type", float_quatloos == F32Quatloos(0.25f)); + + int_quatloos = S32Quatloos(1); + F32Solari float_solari = int_quatloos; + ensure("can convert with fractional intermediates from integer to float type", float_solari == F32Solari(4.f)); + } + + // conversions to/from base unit + template<> template<> + void units_object_t::test<2>() + { + LLUnit<F32, Quatloos> quatloos(1.f); + LLUnit<F32, Latinum> latinum_bars(quatloos); + ensure("conversion between units is automatic via initialization", latinum_bars == F32Latinum(1.f / 4.f)); + + latinum_bars = S32Latinum(256); + quatloos = latinum_bars; + ensure("conversion between units is automatic via assignment, and bidirectional", quatloos == S32Quatloos(1024)); + + LLUnit<S32, Quatloos> single_quatloo(1); + LLUnit<F32, Latinum> quarter_latinum = single_quatloo; + ensure("division of integer unit preserves fractional values when converted to float unit", quarter_latinum == F32Latinum(0.25f)); + } + + // conversions across non-base units + template<> template<> + void units_object_t::test<3>() + { + LLUnit<F32, Quatloos> quatloos(1024); + LLUnit<F32, Solari> solari(quatloos); + ensure("conversions can work between indirectly related units: Quatloos -> Latinum -> Solari", solari == S32Solari(4096)); + + LLUnit<F32, Latinum> latinum_bars = solari; + ensure("Non base units can be converted between each other", latinum_bars == S32Latinum(256)); + } + + // math operations + template<> template<> + void units_object_t::test<4>() + { + // exercise math operations + LLUnit<F32, Quatloos> quatloos(1.f); + quatloos *= 4.f; + ensure(quatloos == S32Quatloos(4)); + quatloos = quatloos * 2; + ensure(quatloos == S32Quatloos(8)); + quatloos = 2.f * quatloos; + ensure(quatloos == S32Quatloos(16)); + + quatloos += F32Quatloos(4.f); + ensure(quatloos == S32Quatloos(20)); + quatloos += S32Quatloos(4); + ensure(quatloos == S32Quatloos(24)); + quatloos = quatloos + S32Quatloos(4); + ensure(quatloos == S32Quatloos(28)); + quatloos = S32Quatloos(4) + quatloos; + ensure(quatloos == S32Quatloos(32)); + quatloos += quatloos * 3; + ensure(quatloos == S32Quatloos(128)); + + quatloos -= quatloos / 4 * 3; + ensure(quatloos == S32Quatloos(32)); + quatloos = quatloos - S32Quatloos(8); + ensure(quatloos == S32Quatloos(24)); + quatloos -= S32Quatloos(4); + ensure(quatloos == S32Quatloos(20)); + quatloos -= F32Quatloos(4.f); + ensure(quatloos == S32Quatloos(16)); + + quatloos /= 2.f; + ensure(quatloos == S32Quatloos(8)); + quatloos = quatloos / 4; + ensure(quatloos == S32Quatloos(2)); + + F32 ratio = quatloos / LLUnit<F32, Quatloos>(2.f); + ensure(ratio == 1); + ratio = quatloos / LLUnit<F32, Solari>(8.f); + ensure(ratio == 1); + + quatloos += LLUnit<F32, Solari>(8.f); + ensure(quatloos == S32Quatloos(4)); + quatloos -= LLUnit<F32, Latinum>(1.f); + ensure(quatloos == S32Quatloos(0)); + } + + // comparison operators + template<> template<> + void units_object_t::test<5>() + { + LLUnit<S32, Quatloos> quatloos(1); + ensure("can perform less than comparison against same type", quatloos < S32Quatloos(2)); + ensure("can perform less than comparison against different storage type", quatloos < F32Quatloos(2.f)); + ensure("can perform less than comparison against different units", quatloos < S32Latinum(5)); + ensure("can perform less than comparison against different storage type and units", quatloos < F32Latinum(5.f)); + + ensure("can perform greater than comparison against same type", quatloos > S32Quatloos(0)); + ensure("can perform greater than comparison against different storage type", quatloos > F32Quatloos(0.f)); + ensure("can perform greater than comparison against different units", quatloos > S32Latinum(0)); + ensure("can perform greater than comparison against different storage type and units", quatloos > F32Latinum(0.f)); + + } + + bool accept_explicit_quatloos(S32Quatloos q) + { + return true; + } + + bool accept_implicit_quatloos(S32Quatloos q) + { + return true; + } + + // signature compatibility + template<> template<> + void units_object_t::test<6>() + { + S32Quatloos quatloos(1); + ensure("can pass unit values as argument", accept_explicit_quatloos(S32Quatloos(1))); + ensure("can pass unit values as argument", accept_explicit_quatloos(quatloos)); + } + + // implicit units + template<> template<> + void units_object_t::test<7>() + { + LLUnit<F32, Quatloos> quatloos; + LLUnitImplicit<F32, Quatloos> quatloos_implicit = quatloos + S32Quatloos(1); + ensure("can initialize implicit unit from explicit", quatloos_implicit == 1); + + quatloos = quatloos_implicit; + ensure("can assign implicit unit to explicit unit", quatloos == S32Quatloos(1)); + quatloos += quatloos_implicit; + ensure("can perform math operation using mixture of implicit and explicit units", quatloos == S32Quatloos(2)); + + // math operations on implicits + quatloos_implicit = 1; + ensure(quatloos_implicit == 1); + + quatloos_implicit += 2; + ensure(quatloos_implicit == 3); + + quatloos_implicit *= 2; + ensure(quatloos_implicit == 6); + + quatloos_implicit -= 1; + ensure(quatloos_implicit == 5); + + quatloos_implicit /= 5; + ensure(quatloos_implicit == 1); + + quatloos_implicit = quatloos_implicit + 3 + quatloos_implicit; + ensure(quatloos_implicit == 5); + + quatloos_implicit = 10 - quatloos_implicit - 1; + ensure(quatloos_implicit == 4); + + quatloos_implicit = 2 * quatloos_implicit * 2; + ensure(quatloos_implicit == 16); + + F32 one_half = quatloos_implicit / (quatloos_implicit * 2); + ensure(one_half == 0.5f); + + // implicit conversion to POD + F32 float_val = quatloos_implicit; + ensure("implicit units convert implicitly to regular values", float_val == 16); + + S32 int_val = quatloos_implicit; + ensure("implicit units convert implicitly to regular values", int_val == 16); + + // conversion of implicits + LLUnitImplicit<F32, Latinum> latinum_implicit(2); + ensure("implicit units of different types are comparable", latinum_implicit * 2 == quatloos_implicit); + + quatloos_implicit += F32Quatloos(10); + ensure("can add-assign explicit units", quatloos_implicit == 26); + + quatloos_implicit -= F32Quatloos(10); + ensure("can subtract-assign explicit units", quatloos_implicit == 16); + + // comparisons + ensure("can compare greater than implicit unit", quatloos_implicit > F32QuatloosImplicit(0.f)); + ensure("can compare greater than non-implicit unit", quatloos_implicit > F32Quatloos(0.f)); + ensure("can compare greater than or equal to implicit unit", quatloos_implicit >= F32QuatloosImplicit(0.f)); + ensure("can compare greater than or equal to non-implicit unit", quatloos_implicit >= F32Quatloos(0.f)); + ensure("can compare less than implicit unit", quatloos_implicit < F32QuatloosImplicit(20.f)); + ensure("can compare less than non-implicit unit", quatloos_implicit < F32Quatloos(20.f)); + ensure("can compare less than or equal to implicit unit", quatloos_implicit <= F32QuatloosImplicit(20.f)); + ensure("can compare less than or equal to non-implicit unit", quatloos_implicit <= F32Quatloos(20.f)); + } + + // precision tests + template<> template<> + void units_object_t::test<8>() + { + U32Bytes max_bytes(U32_MAX); + S32Megabytes mega_bytes = max_bytes; + ensure("max available precision is used when converting units", mega_bytes == (S32Megabytes)4095); + + mega_bytes = (S32Megabytes)-5 + (U32Megabytes)1; + ensure("can mix signed and unsigned in units addition", mega_bytes == (S32Megabytes)-4); + + mega_bytes = (U32Megabytes)5 + (S32Megabytes)-1; + ensure("can mix unsigned and signed in units addition", mega_bytes == (S32Megabytes)4); + } + + // default units + template<> template<> + void units_object_t::test<9>() + { + U32Gigabytes GB(1); + U32Megabytes MB(GB); + U32Kilobytes KB(GB); + U32Bytes B(GB); + + ensure("GB -> MB conversion", MB.value() == 1024); + ensure("GB -> KB conversion", KB.value() == 1024 * 1024); + ensure("GB -> B conversion", B.value() == 1024 * 1024 * 1024); + + KB = U32Kilobytes(1); + U32Kilobits Kb(KB); + U32Bits b(KB); + ensure("KB -> Kb conversion", Kb.value() == 8); + ensure("KB -> b conversion", b.value() == 8 * 1024); + + U32Days days(1); + U32Hours hours(days); + U32Minutes minutes(days); + U32Seconds seconds(days); + U32Milliseconds ms(days); + + ensure("days -> hours conversion", hours.value() == 24); + ensure("days -> minutes conversion", minutes.value() == 24 * 60); + ensure("days -> seconds conversion", seconds.value() == 24 * 60 * 60); + ensure("days -> ms conversion", ms.value() == 24 * 60 * 60 * 1000); + + U32Kilometers km(1); + U32Meters m(km); + U32Centimeters cm(km); + U32Millimeters mm(km); + + ensure("km -> m conversion", m.value() == 1000); + ensure("km -> cm conversion", cm.value() == 1000 * 100); + ensure("km -> mm conversion", mm.value() == 1000 * 1000); + + U32Gigahertz GHz(1); + U32Megahertz MHz(GHz); + U32Kilohertz KHz(GHz); + U32Hertz Hz(GHz); + + ensure("GHz -> MHz conversion", MHz.value() == 1000); + ensure("GHz -> KHz conversion", KHz.value() == 1000 * 1000); + ensure("GHz -> Hz conversion", Hz.value() == 1000 * 1000 * 1000); + + F32Radians rad(6.2831853071795f); + S32Degrees deg(rad); + ensure("radians -> degrees conversion", deg.value() == 360); + + F32Percent percent(50); + F32Ratio ratio(percent); + ensure("percent -> ratio conversion", ratio.value() == 0.5f); + + U32Kilotriangles ktris(1); + U32Triangles tris(ktris); + ensure("kilotriangles -> triangles conversion", tris.value() == 1000); + } + + bool value_near(F32 value, F32 target, F32 threshold) + { + return fabsf(value - target) < threshold; + } + + // linear transforms + template<> template<> + void units_object_t::test<10>() + { + F32Celcius float_celcius(100); + F32Fahrenheit float_fahrenheit(float_celcius); + ensure("floating point celcius -> fahrenheit conversion using linear transform", value_near(float_fahrenheit.value(), 212, 0.1f) ); + + float_celcius = float_fahrenheit; + ensure("floating point fahrenheit -> celcius conversion using linear transform (round trip)", value_near(float_celcius.value(), 100.f, 0.1f) ); + + S32Celcius int_celcius(100); + S32Fahrenheit int_fahrenheit(int_celcius); + ensure("integer celcius -> fahrenheit conversion using linear transform", int_fahrenheit.value() == 212); + + int_celcius = int_fahrenheit; + ensure("integer fahrenheit -> celcius conversion using linear transform (round trip)", int_celcius.value() == 100); + } +} |