Custom style with anchor element fails

[Olek Nowodziński]

The following anchor style is invisible on the list of available styles:

CKEDITOR.stylesSet.add( 'default', [
	{
		name: 'Custom anchor',
		element: 'a',
		styles: {
			'line-height': '18px',
			color: 'red'
		}
	}
]);

The following code also fails:

var range = CKEDITOR.instances.editor1.getSelection().getRanges()[ 0 ];
new CKEDITOR.style( { element: 'a', ... } ).applyToRange( range );

It is visible when accessing via getStylesSet though:

CKEDITOR.instances.editor1.getStylesSet( function( stylesDefinitions ) { 
	console.log( stylesDefinitions );
});
>>> [ Object ]

This problem is only for element: 'a'. It works with p and similar. Also in v3.

[Garry Yao]

The DTD list of object elements need to be reviewed - style.js#L63, actually with the presence of link dialog, it makes little sense of having style with element <a> to be applied as "object", which manipulate a ancestor in the element path.

While we can definitely avoid this bug, since the style system support an explicit "type" property on style def, which will yield the proper style application result.

{
	element: 'a',
	type: CKEDITOR.STYLE_INLINE,
	attributes: { href: '#' },
	styles : { 'line-height': '18px' }
}

Added tc dt/core/style/style.html#test-apply-anchor-style for such usage.

Meanwhile we're working on the html5 DTD at #9301, the DTD is to be unified there, so I would leave this ticket open, to be checked after then.

[Jakub Ś]

1. Actually the same thing is happening with tables and images as described in #9405 comment5.
2. Problem can be reproduced from CKEditor 3.2.1 (when styles combo has started reflecting selection)
3. Styles are only visible when you click on element and then open styles dropdown.

Please also see the attached sample.

This may be a duplicate of #9405.

[Bhadresh Shiroya1]

same problem occur plz ans

[SebCorbin]

May be related ​https://groups.google.com/forum/#!topic/hippo-community/RS7gf_s8ai4

Classes have to be alphabetically ordered to work...

[Jakub Ś]

@SebCorbin if that is really the case then this issue might be related to #13206.