In general, trying to parse XML (or HTML) with regex is not a good idea, and almost certainly doomed to failure. However, to match this specific case you might try: (  *)+" That's an extended POSIX regex, and seems to do the job. It matches any of the following:                    If you need to accept any white space you might try using ( [[:space:]]*) as the sub-pattern. If you may have line breaks in the span text, then you may need to tell your regex engine to not treat them as end-of-text markers. Keep Calm and Carry On