Re: 2.0.33: "ln -sf" reacts strange on directories

Harald Koenig (koenig@tat.physik.uni-tuebingen.de)
Wed, 15 Apr 1998 14:01:08 +0200


On Apr 15, Ulrich Windl wrote:

> I have found one odd thing about 2.0.33:
>
> If I do "ln -sf" for an existing symbolic link, it depends where the
> link points to.
>
> For a file it works as expected (by me).
> For a directory, the modification time of the link's target (the
> directory) changes. This is what I do not expect.
>
> Sorry, if it's not a kernel problem, but for me it seems to be one.
>
> (GNU fileutils 3.16, libc-5.4.44, kernel 2.0.33, (bash 2.01))

the same question came up in June 1995 (so almost 2 years ago)
and that time I answered:

use `ln -sfn linux-1.2.8 linux' for GNU-ln.

I've mailed with the GNU guys about this a long time ago.
they think it's a feature, I still call it a BUG.

unfortuneately I can't find my original mails to the GNU guys anymore
(maybe I'd have to dig even deeper in the attic;)

I tried to convice them for quite a while that this was a _bad_ change
to make `ln' behave incompatiblely to other non-GNU UN*X systems (where
ln behaves as `-n' is set by default) but they still were/are happy
about that move:(

On Apr 15, Andreas Schwab wrote:

> No, it isn't. Use "ln -sfd" (-d -> no dereference). If the target is a
> directory then ln creates the link there.

RTFM: it's the `-n' option. from `ln --help' :

-d, -F, --directory hard link directories (super-user only)
-n, --no-dereference treat destination that is a symlink to a
directory as if it were a normal file

Harald

--
All SCSI disks will from now on                     ___       _____
be required to send an email notice                0--,|    /OOOOOOO\
24 hours prior to complete hardware failure!      <_/  /  /OOOOOOOOOOO\
                                                    \  \/OOOOOOOOOOOOOOO\
                                                      \ OOOOOOOOOOOOOOOOO|//
Harald Koenig,                                         \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik                              //  /     \\  \
koenig@tat.physik.uni-tuebingen.de                     ^^^^^       ^^^^^

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu