For manually installed services, we -should- keep the behavior. What good is it to have a user install a reference to an Apache service which won't work?
Installing a service != starting a service. Users will figure out httpd.conf has problems as soon as they attempt to start the service they just installed. I guess I don't fully appreciate the value of the current behaviour :-)
For automated installs, e.g. the package installer, I see the benefit of offering more than one behavior. E.g. the win32 installer, or someone's external deployment scripts, should be able to override the tests.
Would that satisfy your requirements?
It would, but is it worth the extra code to enable switching between both behaviours? Got some code to review?
Bill
