Re: b32ff1a86 Breaks OSX and FreeBSD

[Greg Troxel]

Fred Wright <address@hidden> writes:

> Thanks.  The usual build now works again on all the platforms I
> normally test on (see some of my recent commit messages), except that
> I can't test on OSX 10.14 or 10.15 without rebooting into them, which
> I don't plan to do until final testing after we stop breaking things.

That's not how gpsd works!

> It's not clear to me why that code even worked on Linux, since
> system-related definitions aren't supposed to be available without the
> proper includes.

I am unclear on whether POSIX requires that the definitions be
unavailable when the header wasn't included.  I also wonder if on Linux
stdlib of stio ends up including sys/time.h, or whether time_t is really
in sys/types, which is pulled in, or something else.