Show duplicate revisions when experiment finishes running in the workspace

[mattseddon]

2/2 main <- #3639 <- this

This PR replaces a small piece of functionality that was purged in #3624. Instead of overwriting the revision/data, we send duplicate revisions whenever an experiment finishes so that the experiment does not appear to vanish and reappear.

Demo

Screen.Recording.2023-04-05.at.3.38.34.pm.mov

Prior behaviour

Screen.Recording.2023-04-05.at.4.34.50.pm.mov

Note: this is a temporary patch as the new data will fix the problem properly.

[julieg18]

Great work!

[sroy3]

This rule really has some hits and misses. Sometimes we want to list things in a special order (sizes: large, larger, regular, small... number: one, two, three, four...), sometimes it just make sense to write have something come before ({ type: X, payload: Y}, to me it feels more natural to have type come before it limits what can come after). It does not have any say on enums though (because it's linked to a number underneath, I get it), but this make it feel inconsistent. I'd say we remove it / review it (can we enforce after 5 properties?) and use judgment when creating objects, but I don't mind if people want to keep it.

[mattseddon]

It can definitely be annoying. I'll put an item in the retro to discuss, I like the rule because it helps me avoid using smol 🧠 and gives me a few fewer decisions to make in the sea of decisions.

[codeclimate]

Code Climate has analyzed commit 8f80cce and detected 0 issues on this pull request.

The test coverage on the diff in this pull request is 97.0% (85% is the threshold).

This pull request will bring the total coverage in the repository to 94.9% (0.0% change).

View more on Code Climate.