On 20 Jun 2002 20:08:38 -0500
Ron Johnson <[EMAIL PROTECTED]> wrote:
> On Thu, 2002-06-20 at 17:21, csj wrote:
> > On 20 Jun 2002 04:56:41 -0500
> > Ron Johnson <[EMAIL PROTECTED]> wrote:
> >
> > > If I copy /boot to /new/boot on /dev/hdc1, and / to /new/treeroot
> > > on /dev/hdc2, then symlinks like /vmlinuz which are now
> > > /new/treeroot/vmlinuz still point back to /dev/hda1 instead of
> > > automagically pointing to /dev/hdb1.
> >
> > Why should vmlinuz point back to hda1? It should point back to a file,
> > not a device.
>
> But files don't exist out in the ether, they exist on devices. I
> presumed you would understand the general thrust of the paragraph.
> Of course, I meant that it still points back to /boot, but /boot is
> /dev/hda1... Sometimes being algorithmically specific clutters up
> the meaning.
I believe I understood your original meaning. The example link vmlinuz
won't point to /boot, but to boot (no slash). /boot is absolute, boot
(without the slash) isn't.
> cp may break the link, but tar does not.
The link will break if the filesystem is not mounted properly as /. You
snipped out my example. For soft links cp -auv should yield the same
results as tar.
>From cp --help:
-a, --archive same as -dpR
-d, --no-dereference never follow symbolic links
-p, --preserve preserve file attributes if possible
-R, --recursive copy directories recursively
Add this special caveat:
-r copy recursively, non-directories as files
WARNING: use -R instead when you might copy
special files like FIFOs or /dev/zero
>From the above I get the impression that cp -a can be used to copy
devices (/dev/zero).
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
