From: Jeff Garland (jeff_at_[hidden])
Date: 2003-06-24 07:50:07
> On Mon, Jun 23, 2003 at 06:53:28AM -0700, Jeff Garland wrote:
> > [snip]
> > Ok will do. I'll add some protections or rounding for when durations
> > support higher than microsecond resolution. I'll let you know when
> > this gets added.
> It seems that since 1) timeval is non-(C++) standard, and 2) it is
> used only by select, that this conversion is more appropriate in a
> Boost socket library, and even there only as an internal component.
I think that it is broader than sockets. The thread library, for
example, uses something that basically amounts to timeval in
some cases. Finally, if someone wanted to use a different socket
library (especially since boost.socket isn't in boost) why make
them rewrite this code?
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk