472: Adding autofill route if fingerprint onboarding is skipped

[eliserichards]

Fixes #472

Testing and Review Notes

Screenshots or Videos

To Do

• double check the original issue to confirm it is fully satisfied
• add testing notes and screenshots in PR description to help guide reviewers
• add unit tests
  • optional: consider adding instrumentation (integration/UI) tests
• consider running this branch in a debug simulator and check for memory leak notifications or any warnings
• request the "UX" team perform a design review (if/when applicable)
• make sure CI builds are passing (e.g.: fix lint and other errors)
• check on the accessibility of any added UI

[jhugman]

I think the logic here needs some small amount of rework, but I may be misunderstanding.

[jhugman]

What happens in Android versions at or above O, but where autofillAvailable is false?

Should it just return without dispatching the confirm route?

[eliserichards]

I believe that's what it should do. If autofill isn't available then we shouldn't tell them to use autofill
I see what you're saying now :)

Addressed comment

[sashei]

would like to see complete tests!

addressed comments :)

[jhugman]

🚜 🚜 🚜

This is looking pretty good. The tests need a bit of attention, but the main code is 👍 .

[jhugman]

👾

This is making AutofillManager part of the public interface for this store. If I understand correctly, this is to make it easier to test.

I think probably the best way of solving this problem is to mock out the autofill manager and context, then inject it using injectContext. The SettingStoreTest shows this quite nicely.

[eliserichards]

Ahh I think I tried every way of mocking the AutofillManager and SettingStore except whenCalled(context.getSystemService(AutofillManager::class.java)).thenReturn(autofillManager). Thanks!

[sashei]

nice work @eliserichards !! thanks for putting those tests together :))

Addressed comments and made the autofillManager private to the SettingStore again :)