RE: [cobalt-users] Virtual Sites-Moving and Preserving Paths

["Tony" <isplists@xxxxxxxxxxxx>]

+-----Original Message-----
+From: cobalt-users-admin@xxxxxxxxxxxxxxx
+[mailto:cobalt-users-admin@xxxxxxxxxxxxxxx]On Behalf Of Carrie
+Bartkowiak
+Sent: Sunday, January 14, 2001 7:11 PM
+To: cobalt-users@xxxxxxxxxxxxxxx
+Subject: RE: [cobalt-users] Virtual Sites-Moving and Preserving Paths
+
+
+||>There is no software
+||>available today that can correct hard-coded paths without potentially
+||>corrupting binaries or even regular code.
+<snip>
+||>So, THIS TIME, Cobalt is more or less innocent,
+
+Jan,
+I have to disagree.
+If the migration software moved the sites *as they are* on the
+original box,
+none of the hard-coded scripts would break. Site12 on the original machine
+would be Site12 on the new machine, etc. etc.

Yes, this could have been avoided if the migration tool moved the sites
along
with their site number. There was no reason it couldn't have. The target
machine
was virgin...no previous sites. I think it reinstalled the sites based on
the file
size of the archive with the largest becoming site1 even if it was site45 on
the old
box. In a perfect world the best CGI/Perl coders don't use full paths but
the majority of
the hosting clients did. There has always been a denial of cgi usage in the
Cobalt collective
mentality that leads to problems like this. I tried to be grateful to Cobalt
for such a neat
free resource like the migration script but that's hard when you have to
spend 12 unpaid
hours vi'ing through script after script that the free tool broke.
Sometimes it seems Cobalt goes out of their way to deliver stuff that ends
up making our
business lives a living hell. i.e.: patches that are pulled 5 hours after
public posting.