Just to clarify first: we may be mixing male and female. A male connector is normally the one that has directly accessible pins; a female connector is usually not accessible but the spacing is very small (e.g. ~1mm).
My thinking was that the test house "fails" the male connector because the pins are accessible. However, to be consistent they should also "fail" the female connectors, as a spacing of ~1mm is also not enough.
Actually the 2nd edition is very messy on this point. Strictly, both male and female connectors do not comply with Clause Clause 17g (see the compliance statement, 2nd paragraph, which allows shorting of any cr/cl below limits).
The exclusion under 16a)5) still requires basic insulation, which most connectors do not have (male or female). And the exclusion in 16e)1) requires a cover, which again most connectors do not have.
However, the principle is that if contact with the patient is unlikely, then they can be excluded. Most test labs understand this and quietly consider signal connectors excluded under 16e)1). This quiet exclusion should apply to both male and female connectors.
This area has been fixed in the 3rd edition. Clause 8.4.2 c) sets up the basic exclusion (using risk management), and specifically refers to contacts of connectors, whether male or female.
The flow from 8.4 to requirements in the standard is:
8.4 --> 8.7 (leakage currents), and
8.4 --> 8.5 (MOP) --> 8.8 (dielectric strength, cr/cl)
Therefore, once the basic exclusion is set up under 8.4.2 c), the part is exempt from all tests for leakage, dielectric strength and cr/cl.
So to summarize:
Male and female connectors formally fail the 2nd edition, but it is industry practice to accept under 16e;
Male and female connectors pass the 3rd edition, provided location is such that direct or indirect contact with the patient is unlikely (e.g. rear of the device).
