Now
- for the "close enough" method, shows the min of the fractions if they are both failing
- for the "very close" method, shows the min in all cases.
- the negation is included in the close_at_tolerance, otherwise the comparison is not storing the failing fraction (if the negation is outside)
* develop: (35 commits)
Fixing tests on GCCs
Fixing test on GCCs
Fixing the tests of this file for GCCs
Fix collection test on Gccs
Adding support for C++11 to old gcc compilers as well Fixing issues with arrays
Fixing counter for r-value ref supported but std::vector still lacking the proper constructor
Once for all the "Mismatch at" instead of "Mismatch in" for every type of collection
silence a warning in gcc 4.5
Revert "introduce fpc::tolerance_based to be used instead of if_floating_point. This should allow us to extend tolerance based comparisons t omore types"
Revert "introduce fpc::tolerance_based to be used instead of is_floating_point. This should allow us to extend tolerance based comparisons t omore types"
introduce fpc::tolerance_based to be used instead of is_floating_point. This should allow us to extend tolerance based comparisons t omore types
introduce fpc::tolerance_based to be used instead of if_floating_point. This should allow us to extend tolerance based comparisons t omore types
remove executable from header
remove executable form header
fix MSVC-12 tests
ENABLE new test tools fix few leftovers from implementation of new FPV comparisons restore test case in test_fp_comparisons.cpp
test to clear the error on CrystaX runners
some doc
const of the elements of the table
Trying to fix the array problem on Clang 3.4
...
Conflicts:
include/boost/test/tools/floating_point_comparison.hpp
include/boost/test/tools/old/impl.hpp
* local/fix_msvc_unit_tests:
Adding support for C++11 to old gcc compilers
Fixing tests on GCCs && Clang (Linux / OSX), VS 2013
Fixing issues with arrays
Fixing tests due to std::vector constructor
Once for all the "Mismatch at" instead of "Mismatch in" for every type of collection
- replacing all boost::is_floating_point by a trait is_floating_point_comparable
- reporting the errors messages in the static_asserts
- compiling now the boost.mulitprecision test with C++03 and C++11
- adding the new API macros in the boost.mulitprecision test
* tractickets/11054-multiprecision_test_fixes-doc:
Adding a test case for the floating points in multiprecision
changelog
close_at_tolerance : getting rid of redundant lines
casting for compliance with multiprecision expression templates
messages with the assertions
Remove the macro BOOST_TEST_USE_DEPRECATED_TIMER and replace it with a
runtime cla --deprecated_timer_format. Additionally change <TestingTime>
to write the testing time in microseconds and add <CpuTime> giving the
time in nanoseconds. Existing scripts and parsers that handle the XML
output should not be affected by this change.
