> May bug n. 21427 be caused by this? possibly, if the client app also has a bug so it ignores truncate failures. I was quite surprised by this vfat/extend behaviour actually. It seems to be quite deliberate in the fat code, but I don't really understand why.