Where-ever an app API request makes an assumption about content (e.g. this class will be present and mean x, this attribute will be preset).
Examples:
Outcome:
Please fold assumptions into a task so reading web team can capture these things in test (see T139003)
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Invalid | None | T139003 Write tests to protect app based API queries based on assumptions made on content. | |||
| Declined | None | T139004 Spike: Audit mobileview API assumptions |
Hello! We'd like to work T139003 but we need this to happen first. We've dropped priority to normal for that task. Please let us know if this is no longer relevant.
This seems like a pretty good supplement to Android's flaky WebView tests.
FWIW, I think a lot of the iOS and Android test coverage will now appear as part of T152991. That is, T152991 will consolidate both app's JavaScript code, especially transforms, into a single, platform independent NPM module that has simple Mocha tests. We will still need the service to unit test its responses but we might be able to add integration level tests to the library where practical. We haven't gotten far enough yet to add more detail on what assumptions exist :/