Update full screen API.

[avandecreme]

While fixing #347 I ended up updating the whole API as the methods names where not up to date.

I am wondering if we can remove the MIT license. The only code which is potentially under MIT is the definition of the fullScreenApi object. All other code has been rewritten.

[avandecreme]

I will need testers. Here is a test url: http://avandecreme.github.io/openseadragon-debug-site/

Things to test:

• Can you type in the input text while in fullscreen
• OpenSeadragon still working.

I already tested:

• Opera Linux (implement the standard API)
• Chromium Linux (webkit prefix)
• Firefox Linux (moz prefix + FullScreen instead of Fullscreen)
• IE 11 (ms prefix + events camel case)

Tests needed:

• IE < 11 (no support but activeX)
• Safari (webkit prefix but will probably complain about Element.ALLOW_KEYBOARD_INPUT)

[avandecreme]

Got access to IE11 and fixed a bug.
The doc here http://msdn.microsoft.com/en-us/library/ie/dn265028%28v=vs.85%29.aspx is telling that msFullscreenElement returns undefined when not in full screen but it actually returns null.

[msalsbery]

On Windows 7...

IE 11
Text box ok, no viewer after clicking "Toggle Fullscreen"

FF 27.0.1
Text box ok, no viewer after clicking "Toggle Fullscreen"

Chrome 33.0.1750.154 m
Text box ok, no viewer after clicking "Toggle Fullscreen"

Opera 20.0.1387.77
Text box ok, no viewer after clicking "Toggle Fullscreen"

[avandecreme]

Actually I have 2 tests here:

• toggle fullscreen on the input text
• toggle fullscreen on OpenSeadragon with the usual button

So it seems to work fine for you.
Thanks for testing.

[msalsbery]

Actually I have 2 tests here

Oh ok, sorry :) Then yes works fine.

[iangilman]

On Mac Mavericks, tested on Firefox, Chrome, Safari and Opera. Everything worked fine except you can't type in fullscreen in Safari.

[iangilman]

Code looks good.

On the license question, I'm certainly not an expert. It does seem like more or less a complete rewrite, so it seems like we could reasonably lose the MIT license. @bgilbert might have thoughts, as well as @johndyer, who wrote the code we started with.

[avandecreme]

On Mac Mavericks, tested on Firefox, Chrome, Safari and Opera. Everything worked fine except you can't type in fullscreen in Safari.

Good.
From what I read, Apple does not want to allow typing in fullscreen so that is the best we can do.
We still need a feedback from someone with an IE < 11.

[msalsbery]

I ran it through IE11 dev tools in IE8, 9, 10 modes - all gave the error "Automation server can't create object" at the "var wscript = new ActiveXObject("WScript.Shell");" line.

I could get it to work by trusting the site and relaxing the activex security settings as described in the last answers here

That may be a problem for regular users....just thought I'd mention it in case it helps. Or, it could be a problem with just the legacy version emulation...still need test on real IE versions. :)

[avandecreme]

Ok thanks for testing.
I am thinking to remove the activex part. OpenSeadragon would just go in fullpage.

[msalsbery]

That reminds me...The F11 just did a pseudo full page where the title and menu bars etc. are removed giving the page as much real estate as possible. Not a real full screen. That too could be just a missing feature of the emulation mode, but I think that may be what you'd get on the real versions if I recall...I could be wrong.

[avandecreme]

Ok, I removed the activeX part so that only real full screen mode is supported.

Also, it seems that Element.ALLOW_KEYBOARD_INPUT might crash old Safari. See code-lts/jquery-fullscreen-plugin#11
What versions of Safari are we supporting?
I am also surprised that they mention that Chrome doesn't need the parameter as discovering that I needed it, is what made me dig into this.

[avandecreme]

Alright, I figured out what was going on with Chrome.
Element.ALLOW_KEYBOARD_INPUT is needed with the old method webkitRequestFullScreen but not with the new one (webkitRequestFullscreen with a lower case S).

I am wondering if we are dropping full screen support on old Safari by using the lower case S though.

[iangilman]

How old of a Safari are you talking about? If you go by the browser usage stats:

http://caniuse.com/usage_table.php

...looks like there's a little bump at 5.1, but 5.0 and certainly 4.x don't seem to be getting much use.

I guess we still haven't defined an explicit browser support list, but there's some discussion in #106; perhaps it's worth revisiting.

[iangilman]

If we want to support webkitRequestFullScreen, it should be easy enough to test for the existence of element.requestFullScreen, right?

The code changes look good.

In the absence of response from @bgilbert or @johndyer, I'm happy to go either way on the license. It does seem like a full rewrite, FWIW.

[avandecreme]

If we want to support webkitRequestFullScreen, it should be easy enough to test for the existence of element.requestFullScreen, right?

Good idea. I added support for old Webkit just in case some old Safari supports it but not the new API.
And I added getFullScreenElement.

I also removed the license.

I updated the test page here: http://avandecreme.github.io/openseadragon-debug-site/

[msalsbery]

I updated the test page

Tested in IE 8, 9, 10 emulation on IE 11... toggle fullscreen button has no effect (no errors reported), viewer button does full page.

[avandecreme]

Tested in IE 8, 9, 10 emulation on IE 11... toggle fullscreen button has no effect (no errors reported), viewer button does full page.

Thanks! The comportment sounds good to me. I think it is up to the API user's to properly handle the case where the full-screen is not supported (like we do in OSD by switching to full-page instead).

[iangilman]

Tested again on Mac Firefox, Chrome and Safari; looking good.

Code looks good too. Is this ready to land?

[avandecreme]

Yes! Thanks for testing.

[iangilman]

Thank you for making it happen!