The best outcome is if you can stick with the same size screw and NOT have to drill it out larger. I'm just saying it's often a good idea to run a tap down into the hole to clean out any loose threads that may be lieing in there and would hinder the smooth installation of the new bolt. In this case the tap would be the same size and thread as the original screw - no drilling required. Running a tap in there isn't an absolute requirement though, as long as the new screw will screw in without a lot effort. Also, any time you do use a tap, you want to be very careful not to snap it off in the hole. That can ruin your day. If it's too hard to turn, stop, wait a minute, and then carefully back it out.

There are a couple reasons you want to simulate the manifold thickness when test fitting the bolt. One is to keep you from accidentally driving the screw too deep into the hole and "bottoming-out" which might do serious damage like punching a crack into the water jacket. Another reason is so that you are engaging close to the same number of threads that will be engaged when the manifold is in place. This will give you more confidence that the threads can support the finish torque without stripping.

One thing you can sometimes do is to measure the depth of the hole with a sturdy piece of wire and see if you can get away with using a SLIGHTLY longer bolt in order to engage some more threads. But you have to be VERY CAREFUL in measuring the hole depth, adding the manifold thickness, and comparing this to the screw length, due to the possibility of bottoming out as mentioned above. If you decide to try this, make sure you can EASILY insert the screw at least a bit further than it will need to go when you are actually tightening the manifold.

At any rate, it's only if you can't get the screw to hold torque, that you would ever have to consider drilling and tapping oversize.