Re: Debian was hacked: The Canterbury Distribution (howto write the date)
On 04/02/2011 06:31 PM, Boyd Stephen Smith Jr. wrote:
In<[🔎] 4D96A8C3.9080208@cox.net>, Ron Johnson wrote:
I've always thought that Unix Time is *incredibly stupid* (who the heck
says "Fri Apr 1 23:27:41 CDT 2011"?)
and *monumentally shortsighted*
(did nothing happen before 01-Jan-1970?).
What makes you say this is UNIX time? The UNIX standard provides many ways of
displaying a time, and AFAIK, doesn't really prefer any particular string
format.
Because that's traditionally how ls presents the file date.
For me, UNIX time is nanoseconds from Epoch. The time_t and clock_t types are
The Epoch is still 01-Jan-1970.
allowed to be signed and any reasonable size.
Which is still 32 bits on 32-bit Linux.
$ cat time_t.c
#include <stdio.h>
#include <stdlib.h>
#include <time.h>
int main (int argc, char **argv )
{
printf("bits in time_t = %d\n", sizeof(time_t) * 8);
printf("bits in clock_t = %d\n", sizeof(clock_t) * 8);
}
$ gcc time_t.c
$ ./a.out
bits in time_t = 32
bits in clock_t = 32
The timespec structure
specifically records nanosecond and interprets a time_t as seconds. The
clock_t type was "always" in microseconds in SUSv2, but even then there was
warning that it might change.
Which breaks binary compatibility.
Traditionally, UNIX-like systems have used a 32-bit signed time_t, but I'm
pretty sure all the *BSDs (including Mac OS X) and the Linux kernel have moved
beyond that. I'm not sure about AIX and Solaris. I'm pretty sure there won't
be anymore HP-UX.
--
"Neither the wisest constitution nor the wisest laws will secure
the liberty and happiness of a people whose manners are universally
corrupt."
Samuel Adams, essay in The Public Advertiser, 1749
Reply to: