Re: Pdns-dev Digest, Vol 121, Issue 4

Hallo,

On Wed, Feb 19, 2014 at 12:00:01PM +0100, [email protected] wrote:
> ----------------------------------------------------------------------
> Message: 1
> Date: Wed, 19 Feb 2014 04:55:18 +0100
> From: a b <[email protected]>
> To: "[email protected]" <[email protected]>
> Subject: [Pdns-dev] pdns/dnspacket.cc patch
> Message-ID: <[email protected]>
> Content-Type: text/plain; charset="iso-8859-1"
> 
> This patch fixes the compile error reported earlier with GCC 4.7.1. If
> someone with more insight could review it and provide feedback, it wouldi
> be much appreciated. In the patch I replaced uint64_t with long because
> that is labs(3C) calls for in the prototype. However, I am not sure that
> this is correct. Presumably, with uint64_t the intent was much finer time
> granularity?

I think both variants (the original as well as your "correction") are not
really correct. According to the man-page for the libraryfunction time()
returns something from type time_t, so the variable to which this return-
value is assigned, should have type time_t. And if you want to build the
difference from 2 time-stamps returned by different calls to time(), you
should use the function difftime, which returns the difference in seconds
as a double value. I think, all other possibilities would be system-dependent
and *not* really portable.

regards,
	Juergen Ilse
-- 
Ein Domainname (auch wenn er Teil einer Mailadresse ist) ist nur ein Name,
nicht mehr und nicht weniger ...