Just looking around at the AddThis code, and it appears these ASCII control characters are intentional (although I’m sure not where they are ending up, lol). There are references to these control codes in at least one of the AddThis PHP files and at least one AddThis Javascript file.
Here’s a code excerpt from the “add-divs-to-coded-excerpts.js” file:
/**
* Queries window.document for a 3-letter non-printing code
* The order of the code identifies a type of excerpt (archive, category, etc).
* Inserts sharetoolbox and recommendedbox divs on either side of the excerpt.
*
* @alters window.document
*/
function addDivsToCodedExcerpts() {
var excerptNodes = getDocumentNodesWithCode();
for(var i=0; i<excerptNodes.length; i++) {
var excerptNode = excerptNodes[i];
var excerptCode = excerptNode.data.substring(0,3);
var suffix = "";
if(excerptCode === String.fromCharCode(5,6,7)) {
suffix = "-homepage";
} else if(excerptCode === String.fromCharCode(5,7,6)) {
suffix = "-page";
} else if(excerptCode === String.fromCharCode(6,7,5)) {
suffix = "";
} else if(excerptCode === String.fromCharCode(6,5,7)) {
suffix = "-cat-page";
} else if(excerptCode === String.fromCharCode(7,5,6)) {
suffix = "-arch-page";
}
That’s not the complete function – just showing some of it to show the ASCII codes 5, 6, and 7 are used to identify and alter with surrounding divs various types of WP page excerpts.
AddThis: I don’t know if this matters, but just poking around your add-divs-to-coded-excerpts.js code, I see that you’re doing a regular expression match search on line 15 where you explicitly list the repetition by using the {3} parameter.
Code: if(data && data.match && data.match(/[05-07]{3}/g))
However (and again, I don’t know that this actually matters or not) later you don’t specify a repetition when you’re replacing those character codes with empty strings on line 72:
Code: excerptNode.data = excerptNode.data.replace(/[05-07]/g, "");
