1786171 - Exiting Fullscreen doesn't work on YouTube in GeckoViewExample (but not Fenix, Focus, or RB)

Status: REOPENED

(GeckoView :: Media, defect, P3)

Description

[sgberh243]

Attachment: This is my current Main Activity. This is the only code that I have for Geckoview besides the geckoview tag in my layout xml file. I'm currently on GeckoView 96

Steps to reproduce:

1. Used a basic Geckoview app(with no injected js/extensions), and visited https://m.youtube.com with it.
2. Selected a video(for example https://www.m.youtube.com/watch?v=dy8uTMKw_tc)
3. Entered Full Screen, then tried to exit

Actual results:

When trying to exit full screen, it stays there and won't exit at all(however, entering fullscreen works fine).

Expected results:

It should exit from full screen when the button is clicked.

Comment 1

[Chris Peterson [:cpeterson]]

Thanks for the bug report. One of our Android engineers will test this bug.

Are you trying to exit full screen using the YouTube video player's "Exit full screen" button or your Android phone's back button?

Does this bug happen with full screen web pages other than YouTube, such as https://www.w3schools.com/jsref/tryit.asp?filename=tryjsref_element_requestfullscreen?

Comment 2

[Jonathan Almeida [:jonalmeida]]

I tested this locally and I can indeed see this bug on GVE 106, but not on Fenix/Focus Nightly 106 (implements auto-rotation on fullscreen) or Reference Browser 106 (does NOT implement auto-rotation on fullscreen.

It's not clear to me what is causing this.

Comment 3

[sgberh243]

(In reply to Chris Peterson [:cpeterson] from comment #1)

Thanks for the bug report. One of our Android engineers will test this bug.

Are you trying to exit full screen using the YouTube video player's "Exit full screen" button or your Android phone's back button?

Does this bug happen with full screen web pages other than YouTube, such as https://www.w3schools.com/jsref/tryit.asp?filename=tryjsref_element_requestfullscreen?

I am trying to use the YouTube players exit full screen button, not the back button. As to trying this with the other fullscreen web page, entering fullscreen didn't even work(the emulator may have been slow though, I'm not sure what the cause was).

Comment 4

[Chris Peterson [:cpeterson]]

(In reply to Jonathan Almeida (:jonalmeida) from comment #2)

I tested this locally and I can indeed see this bug on GVE 106, but not on Fenix/Focus Nightly 106 (implements auto-rotation on fullscreen) or Reference Browser 106 (does NOT implement auto-rotation on fullscreen.

Jonathan, can we use mozregression with GeckoViewExample or Reference Browser to check if this bug is a regression?

Makoto, you fixed bug related to screen orientation and exiting full screen (bug 1757431) in Fenix 101. Could this GeckoViewExample bug be a regression from your fix for bug 1757431?

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

The bug has a release status flag that shows some version of Firefox is affected, thus it will be considered confirmed.

Comment 6

[Makoto Kato [:m_kato]]

Makoto, you fixed bug related to screen orientation and exiting full screen (bug 1757431) in Fenix 101. Could this GeckoViewExample bug be a regression from your fix for bug 1757431?

No, we don't turn on screen orientation API on GV as default. Although YouTube uses screen orientation API if available, we don't do anything since screen orientation API throws exception.

I guess that this depends on other delegate (MediaSession and ContentDelegate) implementation. We need a complete sample to reproduce this.

Comment 7

[Jonathan Almeida [:jonalmeida]]

(In reply to Makoto Kato [:m_kato] from comment #6)

I guess that this depends on other delegate (MediaSession and ContentDelegate) implementation. We need a complete sample to reproduce this.

That was my guess as well - we have the ContentDelegate in GVE, so it might be the MediaDelegate?

(In reply to Chris Peterson [:cpeterson] from comment #4)

Jonathan, can we use mozregression with GeckoViewExample or Reference Browser to check if this bug is a regression?

I'll try. Keeping ni for now.

Comment 8

[Chris Peterson [:cpeterson]]

Jonathan will try testing Makoto's suggestion to add a MediaDelegate.

Comment 9

[Jonathan Almeida [:jonalmeida]]

Ah, I see the problem now. There is an offset for the dynamic toolbar that moves the touch targets for the exit control higher up. Could you try to test this by setting GeckoView.setDynamicToolbarMaxHeight to the height of your url?

You can see how this is done in GeckoView Example here.

Comment 10

[sgberh243]

I have tried setting it to the height of my navbar, but this does not solve the issue. When you say url, do you mean navbar? To make sure I am doing this right though, When I log to see the value of the height param of the navbar, I get -2, is this a feasible value.

Comment 11

[sgberh243]

(In reply to Jonathan Almeida (:jonalmeida) from comment #9)

Ah, I see the problem now. There is an offset for the dynamic toolbar that moves the touch targets for the exit control higher up. Could you try to test this by setting GeckoView.setDynamicToolbarMaxHeight to the height of your url?

You can see how this is done in GeckoView Example here.

In my original example, where there is no toolbar for the Geckoview(just the GeckoView itself), would this still apply?

Comment 12

[Chris Peterson [:cpeterson]]

In my original example, where there is no toolbar for the Geckoview(just the GeckoView itself), would this still apply?

^ Jonathan?

Comment 13

[Jonathan Almeida [:jonalmeida]]

(In reply to sgberh243 from comment #11)

In my original example, where there is no toolbar for the Geckoview(just the GeckoView itself), would this still apply?

In this case, if you could set: GeckoView.setDynamicToolbarMaxHeight(0) that might ensure this is not a toolbar bug.

(In reply to sgberh243 from comment #10)

When I log to see the value of the height param of the navbar, I get -2, is this a feasible value.

No, that's probably not right, if you're getting -2. Sorry, by url, I meant the height of the toolbar where you enter in URL address. If you do not have one in your project for now, then ignore then and see the above comment.

Comment 14

[sgberh243]

(In reply to Jonathan Almeida (:jonalmeida) from comment #13)

(In reply to sgberh243 from comment #11)

In my original example, where there is no toolbar for the Geckoview(just the GeckoView itself), would this still apply?

In this case, if you could set: GeckoView.setDynamicToolbarMaxHeight(0) that might ensure this is not a toolbar bug.

(In reply to sgberh243 from comment #10)

When I log to see the value of the height param of the navbar, I get -2, is this a feasible value.

No, that's probably not right, if you're getting -2. Sorry, by url, I meant the height of the toolbar where you enter in URL address. If you do not have one in your project for now, then ignore then and see the above comment.

I have tried this as well, this still doesn't work.

Comment 15

[sgberh243]

(In reply to Jonathan Almeida [:jonalmeida] from comment #13)

(In reply to sgberh243 from comment #11)

In my original example, where there is no toolbar for the Geckoview(just the GeckoView itself), would this still apply?

In this case, if you could set: GeckoView.setDynamicToolbarMaxHeight(0) that might ensure this is not a toolbar bug.

(In reply to sgberh243 from comment #10)

When I log to see the value of the height param of the navbar, I get -2, is this a feasible value.

No, that's probably not right, if you're getting -2. Sorry, by url, I meant the height of the toolbar where you enter in URL address. If you do not have one in your project for now, then ignore then and see the above comment.

Do you have an activity layout file and corresponding Java/Kotlin file that has worked for you that I could try, because so far the dynamic toolbar max heught hasn'tb een working, even in a new boilerplate app. Thank You.

Comment 16

[sgberh243]

So this issue has been closed and won't be fixed is the final resolution?