Escape whitespace characters that XML parsers normalize

[Shaisolaris]

OutputXML passes text and attribute values through html.EscapeString, which leaves tabs, newlines and carriage returns as they are. Parsers rewrite some of those raw characters, so the output no longer holds the same values when read back:

doc, _ := xmlquery.Parse(strings.NewReader(`<a x="tab&#9;nl&#10;cr&#13;end">one&#13;two</a>`))
fmt.Print(doc.OutputXML(true))
// <?xml version="1.0"?><a x="tab<TAB>nl<LF>cr<CR>end">one<CR>two</a>

• Reading that output with xmlquery again, x comes back as "tab\tnl\ncr\nend" and the text as "one\ntwo", because the carriage returns are read as newlines.
• lxml (libxml2) also applies attribute-value normalization, so it reads x as "tab nl cr end".

This change writes \r in text as 
, and \t, \n and \r in attribute values as 	, 
 and 
. Everything html.EscapeString already escaped is unchanged, and newlines and tabs in text are still written raw, since they survive parsing.

Testing: TestOutputXMLKeepsWhitespaceCharacters parses the document above, writes it with OutputXML, parses it again, and compares the attribute and the text. It fails on master (both values differ) and passes with the change. go test . passes, and so does go vet .. I also checked the before and after output with lxml, which reads back the original values only after the change. I only ran this on Go 1.25.