Recover all of usage section

[felixSchl]

Currently, the scanner changes the section so that it replaces the section header with whitespace. Since the usage section token parser is indent based, the text can then be parsed properly. However, this has the disadvantage that upon failure, we have to stitch together the title again which is less than ideal.

To mitigate this, the "fixSection" code could be done internal to the usage parser module, potentially here.