Thomas Munro <thomas.munro@gmail.com> writes:
> ... but now I'm wondering if we should be more defensive and possibly
> even save/restore the mask.
+1, sounds like a more future-proof solution.
> Originally I discounted that because I
> thought I had to go through PG_SETMASK for portability reasons, but on
> closer inspection, I don't see any reason not to use sigprocmask
> directly in Unix-only code.
Seems like the thing to do is to add a suitable operation to the
pqsignal.h API. We could leave it unimplemented for now on Windows,
and then worry what to do if we ever need it in that context.
regards, tom lane