On Sun, Oct 4, 2009 at 5:20 PM, Abigail <abigail@booking.com> wrote: > I would personally favour if \d becomes just a shortcut for [0-9], and \w a > shortcut for [a-zA-Z_0-9]. All the time. Regardless of internal encoding > of the subject string, the locale, or how the regexp itself is encoded. > I noticed you didn't touch \s, which is the one that troubles me (too?). I often use \d and \w in patterns that are captured. It's good to match tightly, so I agree with you. \s, on the other hand, matches parts of the input I usually wish to discard. Having it behave laxly (i.e. match characters such as NBSP) would benefit me. - ELBThread Previous | Thread Next