This test is about to be disabled, so this is a reminder bug to reenable it once it can be run again successfully on TBPL.
So what I discovered here is that this test is failing due to a memory spike launching the apps rapidly. This rapid launching of apps is a bit of an unrealistic scenario anyway. Although it's agreed in bug 915138 that there is some kind of memory regression, this test is a cards view functional test not a memory test. I don't like the sleeps but I'd rather have the functionality of this test and B2G team can chase memory regressions using other techniques. I did experiment with the length of sleep, apps in use and order of launching them, too.
I refactored it to use the app object and thus take advantage of any waits inside the app object's launch method. It's a little better but didn't entirely remove the need for the sleep.
Attachment #8333897 - Flags: review?(bob.silverberg) → review+
Attachment #8333897 - Flags: review?(florin.strugariu) → review+
Comment on attachment 8333897 [details] [review] github pr My bad. This doesn't seem to be a permanent fix.
:( What have you found?
Sorry, Zac. More comments in the PR. Basically, there is still an intermittent failure with these changes.
OK! I've closed the pull. Not sure what to do for the time being. I'll unassign myself from this incase somebody else wants to have a shot.
Assignee: zcampbell → nobody
Comment on attachment 8333897 [details] [review] github pr >https://github.com/mozilla-b2g/gaia/pull/13771
Attachment #8333897 - Attachment is obsolete: true
Made this obsolete as the pull is closed
Test no loner exists, marking invalid.
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.