On my current project we are constantly improving the way we deliver new patches to our customer’s environment.
A lot of the changes and fixes we apply are done via SQL Scripts or XMLA Scripts, and these are passed on to the client to run in their environment.
We have encountered deployment issues in the past and have added in some error handling to our SQL Scripts.
The thing is, it takes almost as long to write the custom error handling as it does to write the actual functional script itself!
At what point do you draw the line?