On Fri, 17 Aug 2018 10:20:09 -0500, Scott Cheloha wrote: > I don't think we should encourage or even mention the possibility of the > use of poll(2) to effect millisecond timeouts. Even if the standard library > lacks such an interface. > > I'm pretty sure you can't use this hack portably, either. > > But mostly I just think it's a misuse of the interface and potentially > confusing to a reader of such code:
The way to effect a sleep using poll() is to set nfds to 0, not by setting fds to NULL so that text was incorrect anyway. Since POSIX doesn't explicitly say it is supported, and since there are better APIs for this anyway I think removing the text makes sense. OK millert@ - todd
