If you fit an extension I believe it gives a lower reading ?
How would that work?
You measure the peak pressure when the volume of the cylinder has been compressed into the volume left at TDC. If the extension adds volume then it will have the effect of reducing the peak pressure measured.
It doesn't really matter though because the compression test is most use at showing when one or two cylinders are significantly below par (teaching NickW how to suck eggs here).
Finally, if the extension has its own non-return valve then it won't have any effect on the measured pressure.
If you were only measuring one rotation, then I would agree with your first point. But you crank the engine on the starter for a couple of seconds, so there are several(30? 40?) cycles to 'fill up' the gauge. Which is how everyone I've ever used works: it takes a few turns for the reading to reach its maximum number, which is held until you release that pressure in the gauge.
Your second point is the most important here: it's only a comparison for basic diagnostics, not a search for a rigidly defined number. That would require a properly calibrated gauge(which my £20 one certainly isn't, especially after knocking about in a toolbox for 20 years) and a totally repeatable procedure for ALL of the cylinders. A series of leak-down tests over the life of the engine is a better thing to do, as each one gives a better idea of the health of each cylinder and how it's changing. But they require more kit and time to carry out, and are largely unnecessary on anything other than race engines. Or aeroplanes.