But - there have been bikes made where the very likely crack was in a place that was impossible to see without either destructive disassembly or X-rays. I rode one of those bikes and paid dearly. (Being an engineering student when I purchased the bike, I wondered how the fork crown/steerer was executed but because I had some engineering at that point, I knew that no one in their right mind would cast and machine a plug extending up from the crown that inserted into the steerer. So I kept riding in ignorance.) Yes, the crack that ultimately failed was well along for a while. But the crack was completely obscured by the steerer that covered it, then the headset over that, then the head tube over that. No easy way to check.

As for creaking and groaning: yes you are right, but a lot of stem/handlebar joints and crank/axle interfaces creak and groan, sometimes for years. It is good to know as best we can where they are coming from, but it often means either not riding a bike or riding in paranoia. I do make it a point to look for cracks when I clean an area up.

I did have a steel fork nearly break (after cracking a full circumference between the two blades from hydrogen embrittlement. My first warning was shudder braking. No sound ever. Yes, there probably had been a crack to be seen for a while but who looks regularly at the tops of their fork blades on traditional steel forks built by skilled framebuilders? (The embrittlement was caused by nickle plating and not heat treating after; something the plater knew but neither did nor alert the builder.)

Ben