new trunk dependence on nanosleep()

[Jonathan Wakely]

On 12 November 2015 at 16:38, Tim Prince wrote:
> I see that the current trunk chooses to configure so as to depend on a
> compatible nanosleep() regardless of the ability of the target system to
> support that (e.g. cygwin64).  I haven't been able to see whether there
> is a configure option or a prerequisites download installation to take
> care of it.

We've been using nanosleep there pretty much unchanged for about three years.

The only recent change is to check its return value and inspect errno.

> The reported compile error has to do with undefined errno and associated
> macros in the nanosleep() call.  Is there an include missing from the
> libstdc++ source where it calls nanosleep() ?

The missing include is fixed at r230266.