AB testing: implements an A/A experiment

[leandroalonso]

Ready for review after #15770 is merged

Automattic Tracks: Automattic/Automattic-Tracks-iOS#172

This PR implements the A/A experiment to validate that our ExPlat implementation is not biased.

To test

1. Add a breakpoint in ExPlatService.swift:70
2. Add a breakpoint in WPTabBarController+Swift.swift:89
3. Run the app
4. The first breakpoint in ExPlatService should be hit
5. Tap the Reader tab
6. The second breakpoint should be hit and the property variation should be control

Closing and reopening the app

1. Stop the app
2. Run it again
3. The first breakpoint in ExPlatService should be hit

Coming back from the background

1. Put the app in the background
2. Put the app in the foreground
3. The first breakpoint in ExPlatService should not be hit

PR submission checklist:

• I have considered adding unit tests where possible.
• I have considered adding accessibility improvements for my changes.
• I have considered if this change warrants user-facing release notes and have added them to RELEASE-NOTES.txt if necessary.

[peril-wordpress-mobile]

You can trigger an installable build for these changes by visiting CircleCI here.

[peril-wordpress-mobile]

You can trigger optional UI/connected tests for these changes by visiting CircleCI here.

[leandroalonso]

@emilylaguna can I bother you to review this one?

[leandroalonso]

@renanferrari JFYI, here's the iOS PR for the A/A experiment.

[aaronyan]

Does everyone that open an app get assigned an experiment variation? If so, that's okay, I just want to understand how this works. :)

[leandroalonso]

@aaronyan It works like this:

• If there's any active experiment (aka an enum added here), the request is made
• The returned experiments values are saved
• When the variation for an experiment is consulted, we either return "treatment" (if the API returned treatment) or "control" (in any other case: missing value, null, etc - as per suggested here)

[jkmassel]

👋 We're freezing 16.7 today, so this PR is being bumped to 16.8. If you need this to be part of the 16.7 release, please merge it into the release/16.7 branch and DM me – I'll be happy to cut a new beta release!

[aaronyan]

• If there's any active experiment (aka an enum added here), the request is made
• The returned experiments values are saved
• When the variation for an experiment is consulted, we either return "treatment" (if the API returned treatment) or "control" (in any other case: missing value, null, etc - as per suggested here)

This looks great!

[emilylaguna]

LGTM. Works as described.

[leandroalonso]

I tested the scenarios for the A/A testing as described in pbxNRc-Fx-p2#comment-1656