I have set ckeditor 4 with the following overrides set in config.js for entity handling
config.entities_greek = false;
config.entities_latin = false;
config.entities_additional = '#39,#160';
The default for config.entities_additional
is '#39'
.
When I use the above configuration, the output encoding breaks - I get several &undefined;
entities in the output.
Is this a defect in ckeditor? There appears to be a workaround:
config.entities_processNumerical = 'force';
but this has the undesirable effect of encoding all non-ascii characters as numerical entities.
I know you wanted this to be done with config.js, but if you haven't found a solution, it could be done like this (assuming a CKEDITOR instance named editor1
):
CKEDITOR.instances.editor1.on('instanceReady', function(evt) {
evt.editor.dataProcessor.htmlFilter.addRules({
text: function(value) {
return value.replace(/ /g, ' ');
}
});
});