How to achieve 100% MCDC with 'after' transition in stateflow ?

Hi,

I have a fairly simple model with near 100% MCDC coverage. But I don't understand the meaning of the remaining issue.

Here we can see the actual condition `after(...)` is fully covered, but the condition `msec` is never False. But how is `msec` a condition on its own ? And what does it mean for `msec` to be False ?

This question seemed relevant :

https://fr.mathworks.com/matlabcentral/answers/332053-different-coverage-for-after-1-sec-and-after-1-msec-my-misunderstanding-or-just-a-bug

The answer suggests the issue might come from a time step too large. My model uses 10ms steps, and I did get a warning about using `msec` with simulation step larger than 1ms at some point. So I tried to use `sec` instead of `msec`, but it didin't solve it, it just says "sec" is never False now.

There is also this one :

https://fr.mathworks.com/matlabcentral/answers/109274-how-to-achieve-100-mc-dc-coverage-for-model-having-temporal-logic-tick

I believe the author had the same issue than me. If I understand the answer correctly, I need to have a test that ends before the timeout occurs. So I tried a 200ms total test with a timeout configured at 500ms. Still the same '"msec" is never False'.

Can anyone help me understanding the meaning of this please ?

edit: fixed second link

Answer 1

[Samar]

Hi Theotime,

• In after(n, msec) (or sec), msec is a time‑unit keyword, not a logical condition. Stateflow uses it only to interpret how time is measured. It has no True/False runtime meaning on its own. Please refer the MathWorks documentation: https://www.mathworks.com/help/stateflow/ug/using-temporal-logic-in-state-actions-and-transitions.html
• Simulink Coverage sometimes reports the unit keyword (msec, sec, tick) as a separate “condition”, which then appears as “never False”. This is a known coverage artifact, not a modeling or testing error.
• Changing msec → sec, changing the simulation step size, or shortening the test cannot fix this, because the unit keyword can never toggle to False by design.
• The actual temporal condition after(...) is being evaluated True/False correctly (as seen in your report); only the unit token is unachievable for MC/DC.

Practical actions (MATLAB‑accepted)

1. Use guard form: [after(..., msec)] (often removes the extra unit condition) [matlabsolutions.com], [mathworks.com]
2. Or justify/exclude the msec condition as unachievable by design in the coverage report (standard practice for certification). Please refer this documentation: https://www.mathworks.com/help/slcoverage/ug/coverage-for-stateflow-charts.html

Your MC/DC gap is due to a tool reporting limitation with temporal logic units, not missing tests or incorrect timing logic.

Hope this helps!

Regards,

Samar

[Théotime]

Hello,

Thank you for your answer.

So this is essentially a bug. That's how I would call an artifact that forces user to use workarounds to mask it.

I thought it might be the case but because this exact behaviour is shown here with no explanation, so I thought it was an actual lack of testing https://fr.mathworks.com/help/slcoverage/ug/coverage-for-stateflow-charts.html#:~:text=equivalent%20control%20flow.-,Transition%20Details%20Report%20Section,-Reports%20for%20transitions