coverage.py always reports one missing statement but the html report doesen't show what the missing statement is

Question

I am using coverage.py to measure the statement/code coverage of my test.

There are two files that are being tested:

• example_solution.py
• created_unit_test.py

"example_solution" contains a function "contains_negative()" that should return a boolean if any negative number is found within a list. "created_unit_test.py" is an automatically generated unit-test for "example_solution".

I run that code in my linux cli: coverage run --branch created_scripts/created_unit_test.py

And receive the report in ASCII form:

----------------------------------------------------------------------
Ran 4 tests in 0.001s
OK
Name Stmts Miss Branch BrPart Cover
------------------------------------------------------------------------
created_scripts/created_unit_test.py 12 1 0 0 92%
created_scripts/example_solution.py 3 1 2 0 80%
------------------------------------------------------------------------
TOTAL 15 2 2 0 88%

But the HTML reports, as shown in the images below, don't show where that one missed statement is.

My questions are:

1. What is the one missing statement that doesn't get highlighted and how do I get rid of it?
2. How does coverage.py calculate the score? As an example, "example_solution.py" has 3 statements, 1 seemingly is not covered. So one would expect a statement coverage score of 66%, but why is it 80%?

This is the HTML report of the unit-test not highlighting the missed statement

Other HTML report but for the code with the same issue

Answer 1

The reports look very odd, but I also notice they are using coverage 5.5, which is quite old. Can you try with an updated version of coverage?

As to how the total is calculated: https://coverage.readthedocs.io/en/7.6.0/faq.html#q-how-is-the-total-percentage-calculated

• rpommes: Thank you very much. I used coverage.py in a venv and didn't notice that somehow an old version was installed. After updating everything worked as intendet.