Eric Pouech wrote: > however the last trace should be protected (and debugstr_a is a better > choice than testing for a null string) > A+ > Thanks, Eric. I've sent a replacement patch (with a revised title). -- Andy.