Thread: pgwin32_backend_usleep seems overly cute/confusing

pgwin32_backend_usleep seems overly cute/confusing

From
Tom Lane
Date:
Instead of this rigmarole with pg_usleep() being a macro to
pgwin32_backend_usleep, why don't we just call the function in
backend/port/win32/signal.c pg_usleep(), and put something like
#if defined(FRONTEND) || !defined(WIN32)

around the version in src/port/pgsleep.c?  We don't really want
that version present in the backend at all on Windows, ISTM.
        regards, tom lane