util-linux-2.11n
Quoted 7 lines
Quoted 5 lines
Very off-topic: FSF promotes free software, but produces message translation files with header
# Copyright (C) 2001 Free Software Foundation, Inc.
and without any indication that these files can be freely modified and distributed. No GPL or anything. Very unfree.Ask RMS... and he is likely to fix that.
I did. You may well be right. We'll see.
Andries
Quoted 9 lines · On Thu, 20 Dec 2001 [email protected] wrote:
Quoted 7 lines
Quoted 5 lines
Very off-topic: FSF promotes free software, but produces message translation files with header
# Copyright (C) 2001 Free Software Foundation, Inc.
and without any indication that these files can be freely modified and distributed. No GPL or anything. Very unfree.Ask RMS... and he is likely to fix that.
I did. You may well be right. We'll see.
ObMount: now that MS_MOVE is in 2.5 (and submitted for merge in 2.4) mount --move <oldpath> <newpath> would be nice...
Effect of mount(old, new, NULL, MS_MOVE, NULL): subtree mounted at "old" is taken to "new". Move is atomic and takes all references to objects in a subtree with it - opened files, current directories, etc. Requires CAP_SYS_ADMIN.
Errors:
requires CAP_SYS_ADMIN (EPERM)
old and new must exist (ENOENT)
old must be a mountpoint (EINVAL)
new must not be a descendent of old (ELOOP)
old must not be the absolute root (EINVAL)
if old is directory new must be a directory and vice versa (EINVAL)
+ usual set of errors from lookups for old and new (EFAULT if either is not a valid address, ENOTDIR if component in the middle is not a directory, ELOOP if too many links, EACCES if no execute permissions on some component in the middle, ENOMEM if someone's out of memory, EIO if fs feels like that, etc. - the usual)
Quoted 2 lines · Alexander Viro
new must not be a descendent of old (ELOOP) old must not be the absolute root (EINVAL)
Isn't this second one redundant?
Signature · 1 line
"Love the dolphins," she advised him. "Write by W.A.S.T.E.."