> And the result would be correct wether x is removed, or y was replaced (I > see this similar to rm not doing the same thing as unlink to achieve removal > of a name, but resorting to a variety of syscalls).
This corner case is problematic in POSIX. The fact that GNU even removes x is counter to a strict reading of POSIX, which requires that mv defer to rename(2), and that rename is a no-op when the old and new name refer to the same file (even when it is two different names of a hard link). But this is so counter-intuitive that GNU has chosen to break with POSIX on this regards (except perhaps under POSIXLY_CORRECT; I'd have to re-read the source to be sure). So I'm not sure if Jim should spend much more effort improving the verbose output; the fact that it is different already goes to show that hard links are a special case. -- Eric Blake -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
