Floating point equality

[Crzyrndm]

GCC and Clang have -Wfloat-equal that warns on the use of float/double::operator==, and for good reason (MSVC doesn't appear to have a similar warning).

Some references on float equality, why it should be enabled, and how to make the comparison in a less error prone manner

• Lots of detail on the problem, includes multiple exceptional cases https://randomascii.wordpress.com/2012/02/25/comparing-floating-point-numbers-2012-edition/
• Has an example simple ULP comparer (that this doesn't make any attempt to deal with the exceptions noted in the first link) https://en.cppreference.com/w/cpp/types/numeric_limits/epsilon

[Crzyrndm]

Also, the Approx class in the Catch2 test framework

• https://github.com/catchorg/Catch2/blob/master/include/internal/catch_approx.cpp

[sukoi26]

i recommend reading on the UNUM ( universal number ) to understand the error with the use of IEEE Standard for Floating-Point Arithmetic (IEEE 754)

https://en.wikipedia.org/wiki/Unum_(number_format)

[paulharris]

This is excellent reading, and has code to match that you can copy paste if you don't want to suck in the whole boost library (copy from header)

https://www.boost.org/doc/libs/1_67_0/libs/math/doc/html/math_toolkit/float_comparison.html

On Sun., 22 Jul. 2018, 1:59 pm sukoi26, notifications@github.com wrote:

i recommend reading on the UNUM ( universal number ) to understand the error with the use of IEEE Standard for Floating-Point Arithmetic (IEEE 754)

https://en.wikipedia.org/wiki/Unum_(number_format)

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/tfussell/xlnt/issues/318#issuecomment-406843298, or mute the thread https://github.com/notifications/unsubscribe-auth/ABkgSyTQd2aqYkB7CxAQdda21JWRJ-VNks5uJBTHgaJpZM4VZZvp .