On Fri, Feb 24, 2006 at 08:37:36PM +0100, Jacob Sparre Andersen wrote: > > Isn't the real problem that Miroslav should have used '\b' > to identify the boundary between word and non-word text or > '[:^word:]' to identify all non-word characters.
Not really, because I wanted to match the whitespace before the "text", where "text" does not necessarily belong to the word class -- it can also start with a dot, which would be consumed by your suggestion. Of course there are many solutions to overcome this *after* you realize that not everything what looks like a whitespace is a whitespace. In my case I chaged the matches from \s* to (\s| )* -- Miroslav Kure -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
