784324 - Turn on full-screen-api.enabled (Fullscreen and Mouse Pointer Lock don't work in SeaMonkey on BananaBread demo)

Status: RESOLVED FIXED

(SeaMonkey :: General, defect)

Description

[Robert Kaiser]

I just tried the BananaBread demos at https://developer.mozilla.org/media/uploads/demos/a/z/azakai/3baf4ad7e600cbda06ec46efec5ec3b8/bananabread_1345137842_demo_package/index.html in SeaMonkey trunk, and while the Fullscreen mode and the Mouse Pointer Lock API work fine in Firefox Nightly, they don't in SeaMonkey - we don't go fullscreen, and we still show the mouse pointer.

I guess there's some things we need to port to enable those things properly in SeaMonkey.

This demo will be pushed to the public with the Firefox 15 release next week, so it would be nice if we'd get this to work as well.

Comment 1

[Karsten Düsterloh]

Bug 610509, I assume, for fullscreen. I'm currently investigating the state of things there (it's been lingering for a while, unfortunately :-( ).

Comment 2

[Karsten Düsterloh]

Errors I get on the Error Console under Linux x86_64:

Warning: Error: WebGL: Can't get a usable WebGL context
Source File: https://developer.mozilla.org/media/uploads/demos/a/z/azakai/3baf4ad7e600cbda06ec46efec5ec3b8/bananabread_1345137842_demo_package/bb.js
Line: 1
(twice)
Error: TypeError: Module.a is undefined
Source File: https://developer.mozilla.org/media/uploads/demos/a/z/azakai/3baf4ad7e600cbda06ec46efec5ec3b8/bananabread_1345137842_demo_package/bb.js
Line: 1
(once)
Error: Couldn't get mousewheel.withnokey.action pref: [Exception... "Component returned failure code: 0x8000ffff (NS_ERROR_UNEXPECTED) [nsIPrefBranch.getIntPref]"  nsresult: "0x8000ffff (NS_ERROR_UNEXPECTED)"  location: "JS frame :: chrome://communicator/content/utilityOverlay.js :: GetIntPref :: line 1271"  data: no]
Source File: chrome://communicator/content/utilityOverlay.js
Line: 1273
(dozens)

Comment 3

[[e-mail]]

>Warning: Request for full-screen was denied because full-screen API is disabled by user preference.
>Source File: https://developer.mozilla.org/media/uploads/demos/a/z/azakai/3baf4ad7e600cbda06ec46efec5ec3b8/bananabread_1345137842_demo_package/main.html?low,low

Anyone know what that user preference is?

Comment 4

[[e-mail]]

(In reply to Karsten Düsterloh from comment #2)
> Error: Couldn't get mousewheel.withnokey.action pref: [Exception...
> "Component returned failure code: 0x8000ffff (NS_ERROR_UNEXPECTED)
> [nsIPrefBranch.getIntPref]"  nsresult: "0x8000ffff (NS_ERROR_UNEXPECTED)" 
> location: "JS frame :: chrome://communicator/content/utilityOverlay.js ::
> GetIntPref :: line 1271"  data: no]
> Source File: chrome://communicator/content/utilityOverlay.js
> Line: 1273
> (dozens)
I'm already looking into that; it happens whenever you use the mouse wheel (obviously).

Comment 5

[Philip Chee]

> I'm already looking into that; it happens whenever you use the mouse wheel (obviously).
That's probably _handleMouseScrolled() in viewZoomOverlay.js

Related perma-oranges:

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1345568671.1345571870.16167.gz

3368 ERROR TEST-UNEXPECTED-FAIL | chrome://mochitests/content/chrome/dom/tests/mochitest/chrome/test_MozEnteredDomFullscreen_event.xul | Test timed out.
8314 ERROR TEST-UNEXPECTED-FAIL | chrome://mochitests/content/chrome/dom/tests/mochitest/chrome/test_fullscreen.xul | Test timed out.
8317 ERROR TEST-UNEXPECTED-FAIL | chrome://mochitests/content/chrome/dom/tests/mochitest/chrome/test_fullscreen_preventdefault.xul | Test timed out.


http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1345569925.1345571217.14221.gz

14070 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/pointerlock/test_pointerlock-api.html | file_approval.html: Should only receive mozpointerlockchange when we've been approved for fullscreen. - got false, expected true
14071 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/pointerlock/test_pointerlock-api.html | file_approval.html: Should only receive mozpointerlockchange when we've been approved for fullscreen. - got false, expected true

Comment 6

[Robert Kaiser]

(In reply to [e-mail] from comment #3)
> >Warning: Request for full-screen was denied because full-screen API is disabled by user preference.

I'm pretty sure this is just because we miss the front end for the fullscreen implementation. I'm, seeing this as well.

(In reply to Karsten Düsterloh from comment #2)
> Warning: Error: WebGL: Can't get a usable WebGL context

Hmm, do you have WebGL support in about:support? I do (also Linux), and I don't see this one.

> Error: Couldn't get mousewheel.withnokey.action pref:

Hmm, I'm also not seeing that.

What I get is some of those:

Zeitstempel: 21.08.2012 21:24:06
Warnung: Error: WebGL: renderbufferStorage: internalformat: invalid enum value 0x81a6
Quelldatei: https://developer.mozilla.org/media/uploads/demos/a/z/azakai/3baf4ad7e600cbda06ec46efec5ec3b8/bananabread_1345137842_demo_package/bb.js
Zeile: 1
Zeitstempel: 21.08.2012 21:24:29
Warnung: Error: WebGL: bufferSubData: no buffer bound!
Quelldatei: https://developer.mozilla.org/media/uploads/demos/a/z/azakai/3baf4ad7e600cbda06ec46efec5ec3b8/bananabread_1345137842_demo_package/bb.js
Zeile: 1

but I'm seeing this in Firefox as well and the game/demo is still working fine (sorry for the German in the output, need to eat my own dogfood as a localizer).

Comment 7

[[e-mail]]

OK, so flip the full-screen-api.enabled pref and you're good to go.

Except that its mouse sensitivity is horribly high which makes it unplayable :-(

Comment 8

[[e-mail]]

I guess you might want to file bugs to port some of the front-end controls:

https://bugzilla.mozilla.org/showdependencytree.cgi?id=545812&maxdepth=1&hide_resolved=0

Comment 9

[Philip Chee]

> OK, so flip the full-screen-api.enabled pref and you're good to go.
Based on what I know of other tests, this test should flip the pref itself as part of the test.

Comment 10

[Frank-Rainer Grahl (:frg)]

The original web page is gone. 

This works for me in 2.48:
http://firefoxmania.uci.cu/common/dev_files/demos/bananabread/game.html

Can we close it?

Comment 11

[Robert Kaiser]

I see that full-screen-api.enabled defaults to true in recent SeaMonkey nightly builds, so yes, it was fixed somewhere along the way.