sqweek <[EMAIL PROTECTED]> writes: > Quoting from copy.c (just above the patch):
Sure, but that quote talks about another problem: cross-device renames that POSIX says should fail with errno == EXDEV but instead fail with errno having some other value. What we have here is a file system where POSIX says a within-device rename should succeed, but instead it fails with errno == EPERM. That is a different situation. That being said, it might be nice to work around the incompatibility if it doesn't adversely affect ordinary uses of 'mv'. However, the patch suggested in <http://lists.gnu.org/archive/html/bug-coreutils/2007-01/msg00018.html> is obviously too broad, since it will cause 'mv' to mishandle ordinary cases of EPERM. If there is a cheap way to figure out that this file system is non-conforming, and to work around the problem just for this file system and just for files on that file system where the EPERM is not conforming to POSIX, then we'd be on the right track. But I'm skeptical that a cheap workaround is possible (i.e., one that doesn't harm performance for all other file systems). This problem will affect lots of applications, so to some extent you're swimming upstream trying to patch coreutils to work around the problem. Perhaps you might try changing the 9p file system instead, so that it supports renames as POSIX requires. Another (far less satisfactory, but perhaps adequate) possibility would be to modify 9p so that it reports EXDEV for cross-directory renames. _______________________________________________ Bug-coreutils mailing list Bug-coreutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-coreutils
