That would require an understanding of how things work that even some electricians lack.
IDK, I generally use Legrand GFCI's. They are shipped in tripped position, you must have power to actually reset it. No power the button will stay in but does pop back out when power is applied, you must be pressing it while power is on to reset it. If it is reset and you then reverse line and load leads, it will stay set. It only fails to reset because of line load reversal after it has tripped. I don't see this as too big of a problem as that isn't too likely to happen in general use though. I don't know what it will do if the "self test" feature fails. I can see total failure of the logic components can possibly leave it still "set" but with no protection, though the features descriptions sort of say it shouldn't. Really don't know how this self test works though or if it can catch all failures or just catches some of the most common failures.Athough the Leviton approach leaves some portion of the electromechnical trip mechanism untested, I think there's a valid argument that it provides some extra margin of safety. With a traditional GFCI if the test button fails to trip it then this could be ignored because the receptacle would still be useable even though the ground fault protection is defective. In the Leviton design, after the test button mechanically trips open the contacts it will not reset and close these contacts if it detects that the GFCI protection circuitry is not functional. In this case there will be no power available from the receptacle and therefore some extra motivation to replace the defective GFCI.
There is no code section requiring the use of a SEPARATE tester unit for a GFCI, only that a GFCI must be tested. If you test it with its own Test button, code is met.
