The problem with that method is that any link between primary and secondary is likely to be capactive, so it would measure open on an ohmmeter. You would need to measure the voltage between the secondary and earth ground with an AC voltmeter.
Good point.
Tested this: it measures 96 V AC between mains earth and any of the DC output terminals (230V being the nominal mains voltage). The generator's DC negative is common with the output ground. Capacitive coupling, yeah.
Unfortunately, my DMM doesn't have a Low-Z mode, it would be interesting to measure that.
However, I have also tested 4 (whatever I could grab around) other PSUs, some with a 2-wire mains connection, some with 3-wire (with earth). One was a cheap(ish) noname, one was a genuine Japanese laptop PSU, other came with usual devices like any other.
*All of them* measured 93..115 V AC between any of the output terminals and the mains earth, just like the one in question. I measured them all without load.
I don't know what to make of it. It looks like any PSU will be the same in this regard.
Yes it can be a problem if you touch a small MOSFET's (such as 2N7000) gate with the generator's ground (maybe the signal wire as well) -- it may (or may not, or may partially) kill it, I did experience this in the past.
My oscilloscope measures about +/- 36 Vpp if I touch the generator's ground with a 10:1 probe (ground clip being disconnected), which may be just enough to damage small transistors.
But is there a PSU that won't be capacitor-coupled to mains? Sounds like this can be only solved by making the generator's output ground connected to mains earth, which will create its own set of well known problems.