Parse HTML data on input
|Reported by:||fredck||Owned by:||fredck|
The editor is currently taking the input HTML as is, doing lots of string processing (regexes mostly) and them loading it into the editing area.
There are occasions when we have buggy HTML being loaded though causing use problem. For example, it's not possible to remove the bold style for the "Test" word in the following case:
<strong> <table> <tr> <td>Test</td> </tr> </table> </strong>
We should use the HTML parser to "fix" the HTML before loading. The parser should also be used on the paste dialogs, insertHtml (with optional parameter to avoid it), or any function that accepts data from the outside.
Also, by using the "parsing filters", proposed with #3002, we could avoid all the regexes and special manipulation we have today, making it also easily customizable by plugins or external code.