Fix flaky iPad UI Test - testWPcomLoginLogout

[jostnes]

Description

The testWPcomLoginLogout test is flaky on iPad test runs. On iPhone test runs, it has a 100% passing rate.

When looking at the difference between iPad and iPhone, I found that on iPad, the MeTabScreen is a modal on top of the MySiteScreen. Because there is often a bit of a timing issue in CI, when the test is verifying that MeTabScreen is loaded, it's still testing against MySiteScreen element tree causing the test to hang and fail.

I'm not 100% sure this will be the right solution because, in theory, the test without any changes should work, but that isn't the reality. I'd like to try this out and monitor for a bit if it works. There are also other iPad-only flakiness that could be because of the same issue. If this works, we can apply the same to those tests too.

Testing

testWPcomLoginLogout is not flaky on iPad (with error: Failed to determine hittability of "appSettings" Cell: Activation point invalid and no suggested hit points based on element frame) and CI is 🟢

[wpmobilebot]

📲 You can test the changes from this Pull Request in WordPress Alpha by scanning the QR code below to install the corresponding build.

	App Name	WordPress Alpha
	Configuration	Release-Alpha
	Build Number	pr21200-7bc3089
	Version	22.9
	Bundle ID	org.wordpress.alpha
	Commit	7bc3089
	App Center Build	WPiOS - One-Offs #6524

Automatticians: You can use our internal self-serve MC tool to give yourself access to App Center if needed.

[wpmobilebot]

📲 You can test the changes from this Pull Request in Jetpack Alpha by scanning the QR code below to install the corresponding build.

	App Name	Jetpack Alpha
	Configuration	Release-Alpha
	Build Number	pr21200-7bc3089
	Version	22.9
	Bundle ID	com.jetpack.alpha
	Commit	7bc3089
	App Center Build	jetpack-installable-builds #5555

Automatticians: You can use our internal self-serve MC tool to give yourself access to App Center if needed.

[pachlava]

@jostnes 👋

Thanks for working on this! While in theory the change should not reduce the flakiness, because the change is about making MeTabScreen be instantiated separately, instead of having it as a return value of goToMeScreen 🤔 (correct me please, if I'm missing something), let give it a go and see how it runs. The CI is green, so

P.S. While launching the Simulator locally for the first time, this was blocking the execution:

It looks like this does not happen on CI, though.

[jostnes]

@pachlava thank you for reviewing!

While in theory the change should not reduce the flakiness, because the change is about making MeTabScreen be instantiated separately, instead of having it as a return value of goToMeScreen 🤔 (correct me please, if I'm missing something), let give it a go and see how it runs. The CI is green, so

yes that is correct, doesn't really make sense in theory (because it should already work as is! 😅) so a bit of experimenting to see if separating it would help with the flakiness on iPad

P.S. While launching the Simulator locally for the first time, this was blocking the execution:

It looks like this does not happen on CI, though.

that looks new, never seen it. will monitor and see if we need to handle that soon...