[ntp:questions] Re: no more leap seconds?

Danny Mayer mayer at gis.net
Fri Aug 12 17:07:42 UTC 2005

Casper H.S.Dik wrote:
> "David Schwartz" <davids at webmaster.com> writes:
>>   What I advocated in 1998 when people were talking about the Y2K problem 
>>was a new time standard for computers that wouldn't break in 2038 and could 
>>properly handle leap seconds. However, on most current operating systems, 
>>it's impossible even today to build code that will work in 2070.
> Solaris 64 bit applications should withstand that test of time.
> (The current OS does not because it comes with many 32 bit
> applications, but applications build now in 64 bit mode have no
> 32 bit time issues)
> Casper

Don't be so sure about that. The word length doesn't mean anything by 
itself. It's what the clocks, functions, timers, etc. use that's 
important. Have you examined all of the 64-bit Sun code for the clock, 
adjtime. gettimeofday, etc.?


More information about the questions mailing list