Write unit tests for one function

<function_code>
{{function_code}}
</function_code>

<context>
Language: {{language}}
Test framework: {{test_framework}}
What the function is supposed to do: {{intended_behavior}}
An existing test from the project to copy for style: {{example_test}}
</context>

<constraints>
- Test behavior through the function's inputs and outputs, not its internals.
- One behavior per test, with a name that says what it checks, for example returns_zero_for_empty_cart.
- Cover, in this order: normal inputs; boundary values (empty, zero, one, the maximum, negative numbers where relevant); invalid input and the errors it should raise; then any branch not reached yet.
- Replace network calls, files, clocks and random values with fakes or fixed values. Do not add libraries beyond {{test_framework}} and what the example test already uses.
- Where the code and the intended behavior disagree, do not write the test to match the code. Write it for the intended behavior, mark it as expected to fail, and explain why in a comment.
</constraints>

<format>
1. The test file in one code block, ready to save, with imports.
2. A table with the columns Test name, Input, Expected result, Why it matters.
3. Possible bugs: any input where the code seems not to match the intended behavior, one sentence each.
</format>

<task>
Write between 6 and 15 unit tests for the function above, so that a later change that breaks its behavior fails at least one test. Make sure each test would fail if the behavior it checks were broken. If the intended behavior is unclear for some input, such as whether an empty string is allowed, do not guess: list it as a question after the table instead of writing a test for it.
</task>