Hello all -- I would like to build, from source, on an Xserve running Leopard Server, a "standard" MacPorts system that could be deployed, in an organization, on numerous other Apple servers running the same version of the OS (e.g. Leopard Server). In doing so I would use Apple's command-line tools (such as hdiutil and asr) create a restorable disk image (.dmg) file of a disc volume that hangs off an external disk on my original Xserve which houses the MacPorts prefix (e.g., /opt/local). I would then restore from this disc image to the additional machines slated to be the recipients of the original standard. I don't see any immediate problems in doing so, however I have noticed when installing some ports (when I have occasionally enabled verbose mode "-v" during the install), depending on the port, some ports seem to take into consideration the Hostname of the Xserve I'm running it on. On my Xserve running Leopard Server, the Hostname is very importantly tied to the DNS name because I am running Apple's DNS server that is buncled with Leopard Server since I need it for other services on the machine which are not directly related to MacPorts such as running an OpenDirectory Master. The problem is, with hundreds of MacPorts tied together and with complex dependency chains, how can anyone determine when a specific MacPort reads the hostname and uses the hostname of the machine during the configuration process (unless a sys admin such as myself builds each and every port with verbose mode and analysis the output grepping for the hostname)? It might be problematic if the hostname of any given MacPort somehow becomes "hard wired" into the port itself during the configuration / install processing. Then again, I might be making a mountain out of a mole hill. Has anyone else built a Standardized MacPorts system that is cloned to multiple Macs? Thanks, T.M.