Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!uakari.primate.wisc.edu!aplcen!haven!adm!smoke!gwyn
From: gwyn@smoke.BRL.MIL (Doug Gwyn)
Newsgroups: comp.sys.apple
Subject: Re: date converter
Message-ID: <11923@smoke.BRL.MIL>
Date: 10 Jan 90 09:06:00 GMT
References: <1990Jan7.165450.29589@mintaka.lcs.mit.edu> <1184@husc6.harvard.edu>
Reply-To: gwyn@brl.arpa (Doug Gwyn)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 16

In article <1184@husc6.harvard.edu> huang@husc4.UUCP (Howard Huang) writes:
>ANSI C is supposed to support the conversion of seconds, but I believe
>the actual details are machine-dependent.

This is wrong.  ANSI C time conversion involves an implementation-
dependent representation of time as a time_t data type, which need
not have any simple relationships to seconds since some epoch.  The
Standard C functions clock() and difftime() do produce some form of
time in seconds, but these do not involve conversion to or from
either a character representation of time or a struct tm (a "broken-
down time", i.e. a collection of seconds, minutes, etc. component
parts).

IEEE Std 1003.1 (POSIX) and UNIX do impose the additional requirement
that time_t be expressed in seconds since a particular epoch; however
this is not an ANSI C requirement.