1161109 - Don't indiscriminately sleep in Robocop tests

Status: RESOLVED WONTFIX

(Firefox for Android Graveyard :: Testing, defect)

Description

[Nick Alexander :nalexander [he/him]]

Right now we have two *long* indiscriminate sleep commands in our Robocop tests.  By indiscriminate, I mean waits that do not periodically check a condition.

This is perhaps expedient in automation but hostile to local testers.  Local testers expect a command to complete essentially immediately or never.  We shouldn't be blocking for *two minutes* just in case.

Let's find a way to check whatever condition we're looking for -- is that GeckoThread switching to killed status?

The two places are:

	mobile/android/tests/browser/robocop/BaseTest.java
153 	mSolo.sleep(120000);
	mobile/android/tests/browser/robocop/UITest.java
67 	mSolo.sleep(120000);

Comment 1

[Geoff Brown [:gbrown]]

The sleep() follows a request to quit. The quit request should result in a process exit well before the sleep completes. In automation, currently, the process typically goes down within 10 seconds of the quit request. Are you seeing longer waits locally?

Historically, we have had problems in automation with reliably quitting. Intermittently, the quit request would not result in process exit for 90+ seconds. (Someone (jchen?) changed fennec shutdown recently (a couple of weeks ago?) and I have not noticed any long quit delays since.) If the quit request does not bring the process down, the robocop framework needs to take action eventually to force it down, but that usually results in some form of crash report - we wait for 120 seconds to avoid the majority of those cases.

Comment 2

[Geoff Brown [:gbrown]]

Mass closing Firefox for Android :: Testing bugs with no progress in 2017.

If this bug is important to you, please re-open.