Remove excessive filtering of non-ASCII charcters

Atto has a cleanHTML() function (run both when the text is saved, and when using Undo) that replaces certain non-ASCII characters with ASCII renderings, e.g.

1. "½" -> "1/2"
2. "©" -> "(c)"
3. "…" -> "..."

This was added in MDL-43857, from a third party filtering script from around 2006 - when presumably document encodings and UTF-8 handling was less robust.

As UTF-8 handling is better now than in 2006, and the ASCII rendeerings are quite ugly, filtering of these extended characters should be removed.

Additionally - given that TinyMCE doesn't do this filtering - this could be considered a regression in functionality resulting from making Atto the default editor.

Replication steps:

1. Create Page activity.
2. Enter some text including extended charcters "½©…" - e.g. this line.
3. Make some of the text bold.
4. Use the Undo function to remove the bold formatting.

Expected results: No change to the extended characters.

Actual results: The extended characters are replaced with ASCII renditions.

Forum discussion: https://moodle.org/mod/forum/discuss.php?d=265622