ALS-1927 Optimize all E2E test case #102
Conversation
…lTest function ALS-1927
ALS-1927 Optimize all E2E test case
| onView(withId(R.id.et_search_country)).perform(replaceText(TEST_WORD_ARG)) | ||
|
|
||
| Thread.sleep(DELAY_1000) | ||
| uiDevice.wait(Until.hasObject(By.text(mActivityRule.activity.getString(R.string.description_arg))), DELAY_5000) |
There was a problem hiding this comment.
For my own curiosity: What's the difference between
uDevice.waitwaitForViewmActivityRule.activity.findViewById
and when do you use each? I'm wondering why our method of testing if an element is visible or not isn't standardized across tests?
There was a problem hiding this comment.
I also see some usages of uDevice.wait replaced with waitForView. Is it possible to move entirely to one convention? Or is it use case dependent?
There was a problem hiding this comment.
uDevice.wait: This comes with UI Automator and waits for a UI element to appear within a given period.
waitForView: Ensures a view is present before interacting with it in UI tests.
mActivityRule.activity.findViewById: Directly accesses a view from the activity in tests.
I am standardizing most instances with waitForView and removing unnecessary uiDevice.wait calls. However, in some cases, findViewById is required.
For Example, findViewById is needed to check the value inside an element.
| waitForView(allOf(withId(R.id.rv_tracking), isDisplayed())) | ||
| val rvTracking = | ||
| mActivityRule.activity.findViewById<RecyclerView>(R.id.rv_tracking) |
There was a problem hiding this comment.
minor: A lot of such code can be simplified if the waitForView utility were to just return the view when it's available.
val rvTracking = waitForView(allOf(withId(R.id.rv_tracking), isDisplayed()))
| uiDevice.wait( | ||
| Until.hasObject(By.text(mActivityRule.activity.getString(R.string.label_start_tracking))), | ||
| DELAY_1000 | ||
| DELAY_5000 |
There was a problem hiding this comment.
This isn't really an optimization right? We're avoiding the Thread.Sleep and moving the wait. I'm not suggesting this is a problem or that it needs to change, but just something to consider if this can be reduced further.
| onView(withId(R.id.et_search_country)).perform(replaceText(TEST_WORD_RUS)) | ||
|
|
||
| Thread.sleep(DELAY_1000) | ||
| uiDevice.wait(Until.hasObject(By.text(mActivityRule.activity.getString(R.string.description_rus))), DELAY_5000) |
There was a problem hiding this comment.
Is 5000 the delay needed for this element to actually show up? Seems like we went from 1000 to 5000 instead.
| waitForView( | ||
| CoreMatchers.allOf( | ||
| withId(R.id.rv_geofence), | ||
| isDisplayed(), | ||
| hasMinimumChildCount(1), | ||
| ), | ||
| ) | ||
| onView(withId(R.id.rv_geofence)).perform( | ||
| RecyclerViewActions.scrollTo<RecyclerView.ViewHolder>( | ||
| hasDescendant(withText(geofenceName)), | ||
| ), | ||
| ) |
There was a problem hiding this comment.
minor: Thanks for the all the indentation/styling fixes! It would be nice to see if we can add static analysis (like FindBugs) on this package to flag these as potential warnings. That way these are fixed when the code is initially committed.
There was a problem hiding this comment.
Noted.
Ideally we should create a ticket and when bandwidth allow we can add it.
ALS-1927
ALS-1927 Optimize E2E test cases
Issue #, if available:
Description of changes:
As per the ask E2E test cases optimized and removed all unnecessary
Thread.sleepcalls. When there is a system-generated dialogue, we don't have access to it. In that case,Thread.sleepis needed. ex: Location permission.By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.