RegEx matching for URLs in HTML

And in a one liner delimited by ' one can use \x27 for ', like /...("|\x27).../. There should be no need for parentheses around every element (they capture), except for the ones around the part that they are after, .*?, and one has to delimit the alternation, ("|\x27). That can also be done by non-capturing ones, (?:"|\x27). The ?: makes parens not capture (remember) what's between them, but only group those things.

(the comment above is meant to be constructive/helpful, etc -- not criticism!)

Yes, that looks good, except for the parenthesis. They "capture" so that what is matched inside is then in $1 (the first parens), $2 (the second), etc. So capture only what is needed later. Judged by the question that's only the URL, so I'd put parens only around .*? before > (which matches the URL). But, we also must have them around the alternation, ("|'), to separate it from the rest. Since this isn't needed later, use (?:) (with ?: they don't "capture" but only group). So I'd say /^<.*?href=(?:\x22|\x27)(.*?)(?:\x22|\x27)>$/gm and after this the URL is availble in $1.

[continued...] I've added the pattern for the closing quote as well. So what is captured (.*?) is really only the URL. Without that the (.*?) itself would capture the closing quote as well.

Right -- one thing is inefficiency, but also (for me) the trouble with unneeded captures is that later it's harder to keep track of which $N I need! If I actually need only one thing, and capture only that one, then I know it's $1 :)