[NOT] Hopefully fixing timing issue in wokwi-ili9341/lcd-uno

[JakubAndrysek]

Oh, It sometimes fails and sometimes runs - weird.
But the same code once more again it works.

Added cache for quicker response.

https://github.com/JakubAndrysek/wokwi-part-tests/actions/runs/19279226502/job/55126435290

Run wokwi/wokwi-ci-action@v1
Run wget -q -O /usr/local/bin/wokwi-cli https://github.com/wokwi/wokwi-cli/releases/latest/download/wokwi-cli-linuxstatic-x64
Run wokwi-cli --timeout "20000" --expect-text "$CI_EXPECT_TEXT" \
Wokwi CLI v0.18.3 (786fa8e49d9c)
Connected to Wokwi Simulation API 1.0.0-20251028-g60747fe2
Starting simulation...
ILI9341 Test!
[Arduino Uno ILI9341 Test] Expected text matched: "ILI9341 Test!"
Display initialized
[Arduino Uno ILI9341 Test] Expected text matched: "Display initialized"
Counter: 0
Counter: 1
Counter: 2
Counter: 3
Counter: 4
Counter: 5
Counter: 6
Counter: 7
Counter: 8
Counter: 9
Counter: 10
Counter: 11
Counter: 12
Counter: 13
Counter: 14
Counter: 15
Counter: 16
Counter: 17
Counter: 18

[urish]

Thanks!

Can you please split into two PRs (one for cache, one for fixing)?

If I understand correctly, adding the delay helps achieving consistent behavior with this specific test?

[JakubAndrysek]

It seems that it helps, but exactly not. The issue persists, although it is sometimes still present, even with a delay; however, it appears to be much less frequent.

[JakubAndrysek]

I have added a new PR - add-chache and now it runs successfully: https://github.com/JakubAndrysek/wokwi-part-tests/actions/runs/19302626572

[urish]

It seems that it helps, but exactly not. The issue persists, although it is sometimes still present, even with a delay; however, it appears to be much less frequent.

Thanks! The proper solution would be to understand why we get inconsitent behavior and make sure the behavior is consistent. It's on my list for next week.

[urish]

Update: I have found and fixed the underlying issue in the CI server. The problem was the simulation would continue to run for several thousand clock cycles after pause points on AVR cores (so it only affected Arduino Uno / Mega / Nano, not the other MCUs).

I've just ran the CI test suite for 10 times to verify the fix, and we got 10 out of 10 green:

Thanks again for looking into this!