[CLUE-Tech] Dumb tar question

Matt Gushee mgushee at havenrock.com
Tue Sep 17 14:38:02 MDT 2002


Hi, all--

Well, I thought I understood how tar worked. But now I'm confused. I
wanted to update my Web site. So (as I have done in the past) I
performed the following steps:

  Copied the directory tree rooted at 'public_html' from the
  development directory to ~/export.

  $ cd ~/export
  $ tar czf 2002091701.tar.gz public_html/

  FTP'ed the tarball to my home directory on the web server (not my
  own machine, unfortunately, or this would be easy to fix).

  telnet'ed in to the web host (running FreeBSD 4.6-STABLE)

  % tar zxvf 2002091701.tar.gz

The site isn't updated. Going back to my Telnet session, I tried

  % file public_html
  public_html/: directory

Wait a minute. It's supposed to be a symlink to the web site tree, 
which is somewhere else.

It appears that tar replaced the symlink with a new directory (in fact,
I tried unpacking the same tarball on my own machine, and it did replace
a symlink). But I thought it was supposed to unpack the contents into
the existing (symlinked directory). And I'm 99% sure I followed exactly
the same procedure to update my site on August 29; whatever I did worked
fine, at any rate.

Am I losing my mind?

-- 
Matt Gushee
Englewood, Colorado, USA
mgushee at havenrock.com
http://www.havenrock.com/



More information about the clue-tech mailing list