Affected versions

• Firefox 65.0.2 (BuildId:20190225143501)
• Firefox 66.0.1 (BuildId:20190322013140)
• Firefox 67.0b5 (BuildId:20190325125126)
• Firefox 68.0a1 (BuildId:20190326002640)

Unaffected versions

• Firefox 60.6.1esr (BuildId:20190322020346)

Affected platforms

• macOS 10.12.6

Unaffected platforms

• Ubuntu 18.04 64bit
• Windows 10 64bit

Steps to reproduce

1. Launch Firefox.
2. Access the following link.
3. Click the Full Screen button.

Expected result

• The video and audio playback is not interrupted.

Actual result

• Every time the user enters Full Screen mode, the audio is briefly interrupted. Also I noticed that if the user enters Full Screen mode while the video was playing for less than 5 seconds, the video starts over.

Regression range

• First bad revision: 51fa00bbe97e6553e67d8d85ae12e5cf98b4ba8b
• Last good revision: 27730939a1386619acfe4feb31e205977bde878a
• Pushlog: https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=27730939a1386619acfe4feb31e205977bde878a&tochange=51fa00bbe97e6553e67d8d85ae12e5cf98b4ba8b
• Potential regressor: Bug 1508277

Hi Haik,

It seems that mozregression pointed out Bug 1508277 for causing this regression.

Can you please have a look?

Thanks!

(In reply to Emil Ghitta, QA [:emilghitta] from comment #1)

Hi Haik,

It seems that mozregression pointed out Bug 1508277 for causing this regression.

Can you please have a look?

I'll take a look. Thanks for the report.

I have a couple of questions:

• Were other macOS versions also tested and you found that it only occurs on 10.12? Or we don't know if it also reproduces on other macOS versions?

• What Apple hardware was the bug reproduced on?

We initially encountered and reported this bug while we performed some tests on macOS 10.12.6. Now, after further investigations on this issue we noted the following:

This issue can be reproduce on iMac (21.5 inch, late 2013) 10.13.6, 10.12.6 and 10.9.5 with Intel Iris pro 1536 graphics.
This issue is not reproducible on macBook Pro (Retina 13 inch, early 2015) 10.14.2 and 10.13.6 with Intel Iris Pro 6100 graphics.

Please let me know if you need more info.

Haik, is this going to get a fix for 67 beta?

(In reply to Neha Kochar [:neha] from comment #4)

Haik, is this going to get a fix for 67 beta?

Too early to say. I will try to reproduce/debug this today and then we'll go from there.

(In reply to Emil Ghitta, QA [:emilghitta] from comment #3)

We initially encountered and reported this bug while we performed some tests on macOS 10.12.6. Now, after further investigations on this issue we noted the following:

This issue can be reproduce on iMac (21.5 inch, late 2013) 10.13.6, 10.12.6 and 10.9.5 with Intel Iris pro 1536 graphics.

Please let me know if you need more info.

Thanks.

I haven't been able to reproduce this so far, but I don't have access to a machine with an Intel Iris Pro graphics card. I will work on getting the same iMac (21.5" late 2013).

If you are able to reproduce it on any other hardware, please let us know.

Emil, I haven't been able to get access to a machine that reproduces this. Could you try these debugging steps on one of the affected machines?

• First, set the pref security.sandbox.logging.enabled to true and restart the browser.

• Run the Mac Console app (/Applications/Utilities/Console.app).

• In the search box, enter plugin-container so that log entries from plugin-container will only be shown.

• Launch the browser and go to Youtube, but don't start the video yet.

• Click the "Clear" button in the Console app.

• Reproduce the problem and then quit Firefox.

• Collect all the logged entries in the Console app using Edit->Select All, then copy, then paste into a text file and attach the text file.

Let me know if it would help to get on a Vidyo call and walk through this.

(In reply to Haik Aftandilian [:haik] from comment #6)

I haven't been able to reproduce this so far, but I don't have access to a machine with an Intel Iris Pro graphics card. I will work on getting the same iMac (21.5" late 2013).

For what it's worth, I can reproduce on a fairly recent MB13 and a clean profile

MacBook Pro (13-inch, 2017, Four Thunderbolt 3 Ports)
CPU 3,5 GHz Intel Core i7
Intel Iris Plus Graphics 650 1536 MB
10.14.4 (18E226)

default	07:02:30.468954 +0200	plugin-container	HALC_ProxyIOContext.cpp:1068:IOWorkLoop:  HALC_ProxyIOContext::IOWorkLoop: skipping cycle due to overload
default	07:02:30.491440 +0200	plugin-container	HALC_ProxyIOContext.cpp:1068:IOWorkLoop:  HALC_ProxyIOContext::IOWorkLoop: skipping cycle due to overload
default	07:02:30.555201 +0200	plugin-container	HALC_ProxyIOContext.cpp:1068:IOWorkLoop:  HALC_ProxyIOContext::IOWorkLoop: skipping cycle due to overload

And now I think I know why you can't reproduce it…

It doesn't happen on the main screen for me (hi-DPI), it happens when I move the window to an external non hi-DPI screen. That's the second issue I have with Firefox on this screen (the other is bug 1541129), I wonder if they are somehow related.

I found a machine I could reproduce this on reliably and running mozregression led me to bug 1480190.
The mozregression appears to be accurate for this machine because after setting media.media-capabilities.enabled=false and restarting the browser, I could not reproduce the problem.

The machine is a MacBook Air, 11-inch, early 2015 running OS X 10.11.6. I tested Release and Nightly.

$ mozregression --good 2018-01-01 
...
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=56aaeff290b7cbe2b9efc43ef8a651049c3fe6b0&tochange=47091a68bf3a7f0706cab19f00c414134a63a787

I used video https://www.youtube.com/watch?v=-vOcJtlTQAY. When switching to full screen, the sound briefly pauses and sometimes jumps backwards a few seconds and repeats along with the video.

For others hitting this bug, try setting media.media-capabilities.enabled=false and restarting the browser.

@jya, could you take a look?

(In reply to Francesco Lodolo [:flod] from comment #9)

And now I think I know why you can't reproduce it…

It doesn't happen on the main screen for me (hi-DPI), it happens when I move the window to an external non hi-DPI screen. That's the second issue I have with Firefox on this screen (the other is bug 1541129), I wonder if they are somehow related.

Thanks for the info. Could you confirm if setting media.media-capabilities.enabled=false (and restarting the browser) works around the problem for you?

If not, let's consider it a different bug. In that case, could you try running mozregression?

(In reply to Haik Aftandilian [:haik] from comment #10)

I used video https://www.youtube.com/watch?v=-vOcJtlTQAY. When switching to full screen, the sound briefly pauses and sometimes jumps backwards a few seconds and repeats along with the video.

For others hitting this bug, try setting media.media-capabilities.enabled=false and restarting the browser.

@jya, could you take a look?

the issue is YT resetting the source buffers and having to reload the content.

Nothing we can do.

Before enabling Media Capabilities, on mac we would have fully disallowed VP9 while now we allow some.

This is an issue that needs to be reported to YT.

(In reply to Haik Aftandilian [:haik] from comment #11)

Thanks for the info. Could you confirm if setting media.media-capabilities.enabled=false (and restarting the browser) works around the problem for you?

That removes the issue for me too.

(In reply to Jean-Yves Avenard [:jya] from comment #12)

This is an issue that needs to be reported to YT.

Has someone done it?

Hello Haik, any news regarding this issue? It is still reproducing on macOS.

(In reply to Catalin Sasca, QA [:csasca] from comment #14)

Hello Haik, any news regarding this issue? It is still reproducing on macOS.

:jya explained in comment 12 that this is a YouTube issue and we need to report it to YouTube.

@jya, do you have a contact we can use or know who to ask at Mozilla on that?

Hey there, YT engineer here. Please right-click on a playback after you've just reproduced the issue, choose Copy Debug Info, and paste the result here.

Shouldn't we better leave this bug open and move it instead over to webcompat?

(In reply to Henrik Skupin (:whimboo) [⌚️UTC+1] from comment #17)

Shouldn't we better leave this bug open and move it instead over to webcompat?

Yes, that would be better, thanks. I'll move the bug to Web Compatibility :: Desktop.

Adding a needinfo for myself to reproduce this bug and post the debug information requested by strobe in comment 16.

The debug info was collected after going full screen and experiencing the audio and video jump-back.

Thanks for taking a look, strobe.

MacBook Air (11-inch, Early 2015)
macOS 10.11.6 (15G22010)
Firefox 66.0.2

After retesting the issue I was unable to reproduce it.

The video and audio continuously played after accessing the link and enabling full-screen mode.

Tested with:
Browser / Version: Firefox Nightly 87.0a1 (2021-01-27)
Operating System: macOS 10.15.6

Based on the above, I'm closing the issue.

Emil can you still reproduce it on your side?

No need to keep this ni? around. I wasn't able to confirm myself (due to the lack of necessary hardware - macOS with Intel Iris pro 1536 graphics)

We do have the necessary hardware now and I asked acristea (CC'ed) today just to double check on her end and, as expected, this isn't reproducible on the latest Nightly build.

Sorry for how long it took me to get to this :)