The seat tube is loaded such that it is 'rotating' away from the top tube and the headtube in the mirro of this, too, both placing the top of the top tube in tension when the bicycle is being ridden, or at least to a first approximation. The true stress states are way more complex, but I'm trying to avoid worrying too much about the loading regime here, because it's not the major issue.
In short, don't worry about holing the tubes from a stress point of view. If you're going to worry, do it from a corrosion POV. Peter's solution is about the only way I'd like to see it.
/Falanx out :-D