On Fri, Aug 13, 2010 at 7:59 AM, Axel Dörfler <axeld@xxxxxxxxxxxxxxxx> wrote: > Interesting, but I don't get it :-) I believe it has to do with the rather interesting definition of Length(), which is also being inlined into that statement. > Anyway, the incorrect check did not have any consequences, as the > trailing null byte is not contained in Length(). That's true for now, yes, but there's no guarantee (or promise by specification) that BString's internal representation will always contain that, especially seeing as ours has already been completely rewritten at least once. Regards, Rene