Due to the inaccuracies introduced by the "months_worked" method, we have decided to calculate the "fraction_of_year" using "days_worked" instead. This avoids the need for rounding.
The new "edge cases" test added to the test file, is based on user feedback, where an example scenario producing incorrect results is given:Feedback
The change made to the calculator successfully passes the new "edge cases" test, as well as 175 other tests. However, there are 46 tests that are currently failing.
We are waiting for content designers to confirm with the department if the expected assertions, for the failing tests, are correct. We believe the data used for the assertions in the failing tests, are inaccurate and will need to be changed.
As we wait for this feedback from the content designers, we will leave these tests failing.
P.S: some parts of the tests are commented out because they are testing the implementation of the calculator as opposed to testing behaviour, which is not BDD-friendly and makes refactoring difficult.
Due to the inaccuracies introduced by the "months_worked" method, we have decided to calculate the "fraction_of_year" using "days_worked" instead. This avoids the need for rounding.
The new "edge cases" test added to the test file, is based on user feedback, where an example scenario producing incorrect results is given:Feedback
The change made to the calculator successfully passes the new "edge cases" test, as well as 175 other tests. However, there are 46 tests that are currently failing.
We are waiting for content designers to confirm with the department if the expected assertions, for the failing tests, are correct. We believe the data used for the assertions in the failing tests, are inaccurate and will need to be changed.
As we wait for this feedback from the content designers, we will leave these tests failing.
P.S: some parts of the tests are commented out because they are testing the implementation of the calculator as opposed to testing behaviour, which is not BDD-friendly and makes refactoring difficult.
Trello card