Ryan VanderMeulen [:RyanVM] Reporter
	Description • 10 years ago

Per this bug's deps, the serviceworker tests are unreliable on every platform except Windows. Due to their ongoing issues and the upcoming merge, they are being disabled on affected platforms. This bug tracks fixing and re-enabling them.

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 1 • 10 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/69615044a960

	Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)
	Comment 2 • 10 years ago

https://hg.mozilla.org/mozilla-central/rev/69615044a960

	Chris Jones [:cjones] inactive; ni?/f?/r? if you need me
	Comment 3 • 10 years ago

Hey bent --- is the cause of this timeout known already?  It's not clear from the context in bug 1037739.

p.s. I'd like to use rr to debug this ;).

	Ben Turner (not reading bugmail, use the needinfo flag!)
	Comment 4 • 10 years ago

Chris, I think our best theory was bug 1042908. Maybe apply that first and then see if rr shows anything? Thanks!

	Andrea Marchesini [:baku] Assignee
	Comment 5 • 10 years ago

The dependence is fixed. Can we try to reenable the tests?

https://tbpl.mozilla.org/?tree=Try&rev=b0b3e94441af

	Andrea Marchesini [:baku] Assignee
	Comment 6 • 10 years ago

This is the latest result: https://tbpl.mozilla.org/?tree=Try&rev=5d676000e336

To me we can re-enable it and fix the issue when we find it.

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 7 • 10 years ago

Looks like Android 2.3 is still in pretty bad shape. I'd support re-enabling everywhere else, though. Changing the annotation to |skip-if = android_version == "10"| should do the trick. Would also be nice if bug 1042908 could be completed in nsm's absence.

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 8 • 10 years ago

Looks like there's a new B2G debug intermittent as well:
https://tbpl.mozilla.org/php/getParsedLog.php?id=45534564&tree=Try

	Ben Kelly [:bkelly, not reviewing]
	Comment 9 • 10 years ago

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #7)
> Looks like Android 2.3 is still in pretty bad shape. I'd support re-enabling
> everywhere else, though. Changing the annotation to |skip-if =
> android_version == "10"| should do the trick. Would also be nice if bug
> 1042908 could be completed in nsm's absence.

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #8)
> Looks like there's a new B2G debug intermittent as well:
> https://tbpl.mozilla.org/php/getParsedLog.php?id=45534564&tree=Try

Since the code is going to be changing a lot over the next few weeks I'm not sure its worth hunting these down right now.  There is a good chance they will get fixed as we finish implementing the rest of SW.

So lets just skip on android and b2g-emulator.  Please just file a follow-up bug to re-enable and hang it off bug 903441.

	Ben Kelly [:bkelly, not reviewing]
	Comment 10 • 10 years ago

Just to clarify, I don't think its worth hunting down given how slow it is to investigate android / b2g-emulator issues remotely via try.  We should definitely hunt down any desktop intermittents, though.

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 11 • 10 years ago

Trying to keep things as targeted as possible, let's go with:
skip-if = android_version == "10" || (toolkit == "gonk" && debug) # <new bug for re-enabling>

We can always get more restrictive later if we find more intermittents occurring in production. Andrea, mind filing said bug and updating the patch? :)

	Andrea Marchesini [:baku] Assignee
	Comment 12 • 10 years ago

Sure. Before doing this I want to see the 'try' result of bug 1056259.

	Andrea Marchesini [:baku] Assignee
	Comment 13 • 10 years ago

https://tbpl.mozilla.org/?tree=Try&rev=0389ba661302

Ok, the android implementation is still buggy.
And for b2g I have a separated patch.

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 14 • 10 years ago

How does the Try push in comment 13 relate to the Try pushes for the other checkin-needed patches of yours I pushed earlier today that had failures you said were OK to ignore?

	Andrea Marchesini [:baku] Assignee
	Comment 15 • 10 years ago

Yesterday we landed bug 1056259 and this fixed several failures of this try push: https://tbpl.mozilla.org/?tree=Try&rev=28d7624a47fe

Today the situation was much better, talking about ServiceWorkers.
There are still issues with b2g where window.open() fails, but my solution is still not enough:
https://tbpl.mozilla.org/?tree=Try&rev=ec72ac6fb344

Does it answer to your question?

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 16 • 10 years ago

Comment on attachment 8476597 [details] [diff] [review]
test1.patch

Review of attachment 8476597 [details] [diff] [review]:
-----------------------------------------------------------------

Worth a shot. We can always make it more restrictive later if need-be.

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 17 • 10 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/171355fe87e1

	Ed Morley [:emorley]
	Comment 18 • 10 years ago

Reverted for failures on B2G desktop (the try run didn't cover these and the skip-if syntax doesn't exclude B2G desktop):
https://tbpl.mozilla.org/php/getParsedLog.php?id=46560123&tree=Mozilla-Inbound
https://tbpl.mozilla.org/php/getParsedLog.php?id=46560372&tree=Mozilla-Inbound

remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/2ee31951b96f

	Andrea Marchesini [:baku] Assignee
	Comment 19 • 10 years ago

This excludes b2g as well.

	Ryan VanderMeulen [:RyanVM] Reporter
	Comment 20 • 10 years ago

Comment on attachment 8478336 [details] [diff] [review]
test1.patch

Review of attachment 8478336 [details] [diff] [review]:
-----------------------------------------------------------------

No need for the (toolkit == "gonk" && debug) part if you're skipping all of B2G.

	Andrea Marchesini [:baku] Assignee
	Comment 22 • 10 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/6ab548becb6f

	Carsten Book [:Tomcat]
	Comment 23 • 10 years ago

https://hg.mozilla.org/mozilla-central/rev/6ab548becb6f

	Chris Jones [:cjones] inactive; ni?/f?/r? if you need me
	Comment 24 • 9 years ago

(In reply to ben turner [:bent] (use the needinfo? flag!) from comment #4)
> Chris, I think our best theory was bug 1042908. Maybe apply that first and
> then see if rr shows anything? Thanks!

I was able to repro either this or another of the serviceworker intermittent-fails with the fix from bug 1042908, but I left this open to double-check this particular symptoms.  The odds of that being meaningful are pretty low at this point so clearing request.  Sorry bent.

	Nikhil Marathe [:nsm] (No longer reading bugmail, please needinfo?)
	Comment 25 • 9 years ago

We already have these enabled on central for everything except b2g. Bug 1053275 will fix intermittents.