Document ctypes support policy (or rather, the lack thereof)

[ncoghlan]

In writing https://bugs.python.org/issue30478#msg294528, I went looking in the ctypes docs to see if there was a note saying that it's relatively trivial to trigger segfaults with that module, so if you get a crash when using it, our default assumption is going to be that the bug is in how ctypes is being used, rather than being in ctypes itself.

However, I didn't see any such note, and I don't think we have anything explicit in the developer guide advising triagers to use "closed/third party" as the default resolution for such reports, with a pointer to the faulthandler module as a tool for investigating further (either the stdlib one for 3.x users, or the PyPI one for 2.7).

[willingc]

Next action: add a note to the triage page in closed/third party description.