Attributions display twice on `mapbox`

[archmoj]

The image tests are failing on master after merging #7060: https://app.circleci.com/pipelines/github/plotly/plotly.js/10950/workflows/2c1cc796-b8cd-47b9-adcc-328d0ae4a9f2/jobs/242299/parallel-runs/0/steps/0-102

This is not related to #7060 itself since if I checkout v2.34.0

git checkout tags/v2.34.0 && npm ci && npm start

then

npm run baseline mapbox_0

Gives me the baseline with duplicate source attribution!

I suspect this was related to the order of rendering on CI but somehow wasn't noticed before.

[archmoj]

Re-running CircleCi tests on the same commit also confirms that something is changed on the mapbox side in the past two weeks that caused these tests to start failing.

[archmoj]

Strangely enough if I call

python3 test/image/make_baseline.py mapbox_0

It works just fine.
But

npm run baseline mapbox_0

fails although it calls the same python script using node.js!

🤯

[archmoj]

@birkskyum this issue doesn't seem related to your PR; although it start showing up after we merged the maplibre PR.

Anyway I was not able to figure out what causes this.
Please note that in addition to the image test, there is one jasmine test is failing which might be easier for debugging.

[archmoj]

Re-running CircleCi tests on the same commit also confirms that something is changed on the mapbox side in the past two weeks that caused these tests to start failing.

@birkskyum Yes the same commit before merging maplibre PR passed the tests.
But then few days later it start failing!!

[archmoj]

The issues are also present in the existing builds, like the examples: https://plotly.com/python/mapbox-layers/

Could you send a screenshot?
I don't see duplicates there.

[archmoj]

@birkskyum That looks like a different bug. No?

[birkskyum]

@archmoj , yes, maybe separate but can't say for sure. I made a ticket for that and removed the noise about it from this ticket.

• mapbox attribute control toggle, wrong styling #7082

[archmoj]

I skipped those mapbox tests on CI in #7083.
And I close this issue for now.