MathJax widget improvements

[Piotr Jasiun]

Umbrella ticket for MathJax widget. Because most of these ticket are related to each other they will be done in one branch.

Ticket	Summary	Owner	Type	Status	Priority	Component
#10840	MathJax widget - same origin policy issue	Piotr Jasiun	Bug	closed	Normal	General
#10841	MathJax widget - support subtree changes	Piotr Jasiun	Bug	closed	Must have (possibly next milestone)	General
#10842	MathJax widget - undo change done through a dialog		Bug	closed	Normal	General
#10843	MathJax widget tests	Piotr Jasiun	Bug	closed	Normal	General
#10844	MathJax widget - IE8 support	Piotr Jasiun	Bug	closed	Normal	General
#10857	MathJax widget [FF,IE9] - clipboard totally broken	Piotr Jasiun	Bug	closed	Must have (possibly next milestone)	Core : Pasting
#10930	IE9 MathJax crash on undo	Piotr Jasiun	Bug	closed	Normal	General
#10948	MathJax widget - loading indicators	Piotr Jasiun	New Feature	closed	Normal	General
#11078	MathJax sample should mention that MathJax widget does not support IE8	Piotr Jasiun	Task	closed	Normal	Documentation & Samples

[Piotr Jasiun]

All of changes and tests are in ​t/10937 branch and corresponding branch in tests project.

Both branches have been rebased on the top of the major and tested with IE11. Everything there is fine.

[Piotr Jasiun]

MathJax is stable now. Only IE8 makes some problems but it seems to be pointless to spend more time in this. All of the tests are green in all supported browsers in both dev and build versions.

All of the changes are in: ​​t/10937 and corresponding test branch.

[Frederico Caldeira Knabben]

I think we reached a good situation with this feature, therefore I'm not going so hard on the review.

There are some minor things though.

---

Just for simplification, I would do this:

node.hasAttributes[ 'data-cke-survive' ]

... instead of this:

node.attributes[ 'data-cke-survive' ] === 'true'

I would then simply use "1" instead of "true".

In this way this attribute would behave just like a HTML "boolean". In fact, the current "false" behavior is misleading, because it is not a guarantee that the element will get removed if empty.

---

The loader image is in the "icons" folder and this is wrong. That folder is reserved to toolbar icons. It should be into a new "images" folder.

---

Other than this, we should be ready to close this ticket and all its siblings.

[Piotrek Koszuliński]

Wrong quotes used:

document.domain = "ckeditor.dev";

---

iframe.getName() != 'iframe'

Could be changed to:

!iframe.is( 'iframe' )

---

Missing spaces:

if( CKEDITOR.env.ie ) // Twice!

attrs['data-cke-survive'] = 'true';

return  match.replace(/(<iframe.*?\/iframe>)/i, ""); // Plus wrong quotes.

preview.setHtml( '<img src=' + CKEDITOR.plugins.mathjax.loadingIcon + ' alt=' + editor.lang.mathjax.loading +'>' ); // After '+'.

if( doc.getById( 'preview' ) )

---

There's a space in indentation after this line:

CKEDITOR.plugins.mathjax.fixSrc =

---

We don't use curly braces in this case:

+                                       if ( val ) {
+                                               preview.setStyle( key, val );
+                                       }

[Piotr Jasiun]

There is no hasAttributes method in CKEDITOR.htmlParser.element (it exists only in CKEDITOR.dom.element) so changed to:

if( node.attributes[ 'data-cke-survive' ] )

I agree that data-cke-survive=false does not make much sense.

I also moved loader.gif form icons to images.

[Piotrek Koszuliński]

Replying to pjasiun:

There is no hasAttributes method in CKEDITOR.htmlParser.element (it exists only in CKEDITOR.dom.element) so changed to:

if( node.attributes[ 'data-cke-survive' ] )

Space!

[Piotr Jasiun]

Code style fixes applied and everything seems to work.

[Piotr Jasiun]

• ​git:95bd2fe
• tests:b802386