On Tue, 5 Oct 2004, Alex Kiernan wrote:
> Is it worth adding a check in milliwait() that setitimer() succeeds
> before going into sigsuspend()?
And what do you do if it fails? You could perhaps obey sleep(n) as an
emergency fixup, where n is the time rounded up to the nearest second.
Advantage: apart from a too-long delay, everything works.
Disadvantage: nobody ever notices the problem, and performance suffers.
I would certainly do it if failure to continue was sufficiently serious
(e.g. life-threatening), but I think perhaps in this case it would be
better in the long run to discover that there is a problem. Does that
make sense?
I am about to freeze the code for 4.43 (it includes this stuff) and
intend to release it later this week, maybe even tomorrow.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book
