1489224 - [remote-dbg-next] Tests: check that devices list shows "No device discovered" if empty

Status: RESOLVED FIXED

(DevTools :: about:debugging, enhancement, P1)

Description

[Julian Descottes [:jdescottes]]

Check that the devices list shows a message when no devices are available, and that the message disappears if devices are added.

Comment 1

[Julian Descottes [:jdescottes]]

Attachment: Bug 1489224 - Add mochitest to check runtimes section of aboudebugging sidebar;r=daisuke

Comment 2

[Julian Descottes [:jdescottes]]

Attachment: Bug 1489224 - Fix order of tests in aboutdebugging-new;r=daisuke

Comment 3

[Julian Descottes [:jdescottes]]

try at https://treeherder.mozilla.org/#/jobs?repo=try&revision=f7b501ba8645855f08afd10e84df8d7c80acd0df

Comment 4

[Daisuke Akatsuka (:daisuke)]

Comment on attachment 9010663 [details]
Bug 1489224 - Fix order of tests in aboutdebugging-new;r=daisuke

Daisuke Akatsuka (:daisuke) has approved the revision.

Comment 5

[Daisuke Akatsuka (:daisuke)]

Comment on attachment 9010662 [details]
Bug 1489224 - Add mochitest to check runtimes section of aboudebugging sidebar;r=daisuke

Daisuke Akatsuka (:daisuke) has approved the revision.

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/6300cdbc0e04
Fix order of tests in aboutdebugging-new;r=daisuke
https://hg.mozilla.org/integration/autoland/rev/d414a44d8119
Add mochitest to check runtimes section of aboudebugging sidebar;r=daisuke

Comment 7

[Arthur Iakab [arthur_iakab]]

Backed out 2 changesets (bug 1489224) 
Backout revision https://hg.mozilla.org/integration/autoland/rev/3ea2713d067e4753c275a62b576c1e9ec7fdd9ba
Failed push https://treeherder.mozilla.org/#/jobs?repo=autoland&resultStatus=testfailed%2Cbusted%2Cexception&classifiedState=unclassified&selectedJob=200770518&revision=d414a44d81195403aabdfb91994c5df9c9acbbc9
Failure log: https://treeherder.mozilla.org/logviewer.html#?job_id=200766932&repo=autoland

:jdescottes Could you please take a look?

Comment 8

[Cristian Brindusan [:cbrindusan]]

https://hg.mozilla.org/mozilla-central/rev/6300cdbc0e04
https://hg.mozilla.org/mozilla-central/rev/d414a44d8119

Comment 9

[Julian Descottes [:jdescottes]]

It seems this new test leaks on windows when running in test-verify only. Both the runs from autoland and central seem to confirm this.

When looking at the test-verify logs, the test passes the first phase successfully (tests running in the same browser), but fails at the first iteration of the second phase:

16:09:46     INFO -  TEST-INFO | checking window state
16:09:46     INFO -  Browser Chrome Test Summary
16:09:46     INFO -  	Passed: 33
16:09:46     INFO -  	Failed: 1
16:09:46     INFO -  	Todo: 0
16:09:46     INFO -  	Mode: e10s
[...]
16:09:46     INFO - ::: 1. Run each test 10 times in one browser. : Pass
16:09:46     INFO - ::: 2. Run each test 5 times in a new browser each time. : FAIL

Consistent results for the 4 TV jobs that failed, "Passed: 33, Failed: 1" means the leak check failed exactly after the first test of the second phase, since the test only contains 3 asserts.

Maybe what happens is that the browser is killed too fast and the destroy of the page doesn't have time to complete (since it is async)

Comment 10

[Julian Descottes [:jdescottes]]

I can't seem to reproduce this on try in isolation: https://treeherder.mozilla.org/#/jobs?repo=try&selectedJob=201132195&revision=e6618c69a91c3eb0cb55812140ff38e16408af44

The autoland TV job only had this test running. Will move on the autoland changeset to see if I can reproduce on try there.

Comment 11

[Julian Descottes [:jdescottes]]

> I can't seem to reproduce this on try in isolation

After rebasing the test on a more recent changeset, I managed to get the same failure on try.

This actually affects all the new about:debugging tests. I tried re-running our first initial test on try with test-verify and it fails in the same way. 

So something happened between https://hg.mozilla.org/try/rev/8265be8d02b0 and https://hg.mozilla.org/integration/autoland/rev/35c033207fa3 that makes our tests fail.

Comment 12

[Julian Descottes [:jdescottes]]

It looks like this started failing with the first changeset for USB devices:

https://hg.mozilla.org/try/rev/8ac3f47057c4

Comment 13

[Julian Descottes [:jdescottes]]

Attachment: Bug 1489224 - Enable adb scanner only if addon status is installed;r=daisuke

Using await check() to see if ADB is running is causing aboutdebugging to
leak on debug windows test-verify.

Ultimately we should be able to call ADB scanner without worrying if
the addon is installed or not but for now we should keep things simple
and only start the scanner if the extension is available.

Comment 14

[Daisuke Akatsuka (:daisuke)]

Comment on attachment 9011541 [details]
Bug 1489224 - Enable adb scanner only if addon status is installed;r=daisuke

Daisuke Akatsuka (:daisuke) has approved the revision.

Comment 15

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/26719c685fd3
Enable adb scanner only if addon status is installed;r=daisuke
https://hg.mozilla.org/integration/mozilla-inbound/rev/a581d5bc795d
Fix order of tests in aboutdebugging-new;r=daisuke
https://hg.mozilla.org/integration/mozilla-inbound/rev/573a32cebe76
Add mochitest to check runtimes section of aboudebugging sidebar;r=daisuke

Comment 16

[Andrei Ciure[:aciure]]

https://hg.mozilla.org/mozilla-central/rev/26719c685fd3
https://hg.mozilla.org/mozilla-central/rev/a581d5bc795d
https://hg.mozilla.org/mozilla-central/rev/573a32cebe76