Improve encoding recognition for html pages and js resources

Whenever html files are delivered using the single page course element, SCORM or CP modules etc. the page ocntent encoding is guessed to properly deliver the page. Following resources such as linked/embedded javascript files are delivered using the same encoding as the content page by default.

Current situation
There is no standard way to do this. OpenOLAT uses the HTML 4 meta tag

<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">

to detect the encoding. If not detected, the system wide default encoding is applied. This automatic detection can also be overridden using the configuration for that page.

Changes
Many pages nowadays use the HTML 5 meta tag:

<meta charset="utf-8">

The goal of this issue is to search for both the HTML4 and the HTML5 meta tag for encoding description.