Fix/script textcontent parsing

[bourgeoa]

Migrates the fix from #693 (csarven's fork) into the main repository to resolve #692.

Change
Prevents parsing empty script tags in RDFa documents by checking content length before invoking the parser:

// reference/fetcher-classes.js, line 116

• if (Parsable[contentType]) {

• if (Parsable[contentType] && scripts[i].textContent.trim().length) {
  rdfParse(scripts[i].textContent, kb, xhr.original.uri, contentType)
  }
  Fixes errors when encountering <script type="text/turtle" src="..."></script> tags with no body content.