See http://jsfiddle.net/aEEUN/
why when I use backreference vs no backrefences, I get different results?
var str = "Hello World\nHello ";
document.write("Without backreference: <br />");
var match = str.match(/\S(?![\s\S]*\S)/);
document.write("- match.index: " + match.index + "<br />"); // index is 16
document.write("With backreference: <br />");
var match = str.match(/(\S)(?![\s\S]*\1)/);
document.write("- match.index: " + match.index); // index is 6
The difference is that you constrain the two character to be exactly the same.
In the first regular expression there must not be any non-whitespace character following:
/\S(?![\s\S]*\S)/
This basically means to match any non-whitespace character that is not followed by any other non-whitespace character. That’s why it matches the last o that is only followed by whitespace:
"Hello World\nHello "
^ no other non-whitespace character following
But in the second regular expression there must not be that specific character following that was matched before:
/(\S)(?![\s\S]*\1)/
This basically means to match any non-whitespace character that does not appear again in the rest of the string. That’s why the W is matched as it’s the first non-whitespace character that does not appear again after the first occurrence:
"Hello World\nHello "
^ no other “W” following
That’s why it’s called a backreference: You reference a previously matched string.
That is some weird regex. The character class [\s\S] matches everything doesn't it? A space or non-space? Not sure exactly, but the \1 in the latter example has to refer to the same character that it did at the start (not just any non-space again)... which would give it a different result.
