Question

I have a JS function which is passed a string that a RegEx is run against, and returns any matches:

searchText= // some string which may or may not contain URLs
Rxp= new RegExp("([a-zA-Z\d]+://)?(\w+:\w+@)?([a-zA-Z\d.-]+\.[A-Za-z]{2,4})(:\d+)?(/.*)?/ig")
return searchText.match(Rxp);

The RegExp should return matches for any of the following (and similar derivations):

• google.com
• http://www.google.com
• http://www.google.com
• http://google.com
• google.com?querystring=value
• http://www.google.com?querystring=value
• http://www.google.com?querystring=value
• http://google.com?querystring=value

However, no such luck. Any suggestions?

Answer 1

In a string, \ has to be escaped: \\.

First, the string is interpreted. \w turns in w, because it has no significant meaning.
Then, the parsed string is turned in a RegEx. But \ is lost during the string parsing, so your RegEx breaks.

Instead of using the RegExp constructor, use RegEx literals:

Rxp = /([a-zA-Z\d]+:\/\/)?(\w+:\w+@)?([a-zA-Z\d.-]+\.[A-Za-z]{2,4})(:\d+)?(\/.*)?/ig;
// Note: I recommend to use a different variable name. Variables starting with a
//  capital usually indicate a constructor, by convention.

If you’re not 100% sure that the input is a string, it’s better to use the exec method, which coerces the argument to a string:

return Rxp.exec(searchText);

Here’s a pattern which includes the query string and URL fragment:

/([a-zA-Z\d]+:\/\/)?(\w+:\w+@)?([a-zA-Z\d.-]+\.[A-Za-z]{2,4})(:\d+)?(\/[^?#\s]*‌)?(\?[^#\s]*)?(#\S*)?/ig