Editing Unix time (section)

=== Representing the number ===
A Unix time number can be represented in any form capable of representing numbers. In some applications the number is simply represented textually as a string of decimal digits, raising only trivial additional problems. However, certain binary representations of Unix times are particularly significant.

The Unix <!-- time_t meaning time type is a C type name. DO NOT REPLACE «time_t» WITH «time t»!!! --><code>[[time_t]]</code><!-- DO NOT REPLACE «time_t» WITH «time t»!!! --> data type that represents a point in time is, on many platforms, a [[Integer (computer science)|signed integer]], traditionally of 32{{nbsp}}[[bit]]s (but [[#Range of representable times|see below]]), directly encoding the Unix time number as described in the preceding section. A signed 32-bit value covers about 68 years before and after the 1970-01-01 epoch. The minimum representable date is Friday 1901-12-13, and the maximum representable date is Tuesday 2038-01-19. One second after 2038-01-19T03:14:07Z this representation will [[arithmetic overflow|overflow]] in what is known as the [[year 2038 problem]].

[[UUID]]v7 encodes the Unix epoch timestamp (in milliseconds) in an unsigned 48-bit field. This representation is valid until the year 10889 AD.<ref>
{{IETF RFC|9562}}
</ref>

In some newer operating systems, <code>time_t</code> has been widened to 64 bits. This expands the times representable to about {{tooltip|1={{round|2^63/365.2422/24/3600/10^9|1}} billion years|2=More precisely, 292277264695 years}}<!-- note:including {{round}} inside {{tooltip}} doesn't work when the resulting number results has an exponent --> in both directions, which is over twenty times the present [[age of the universe]].

There was originally some controversy over whether the Unix <code>time_t</code> should be signed or unsigned. If unsigned, its range in the future would be doubled, postponing the 32-bit overflow (by 68 years). However, it would then be incapable of representing times prior to the epoch. The consensus is for <code>time_t</code> to be signed, and this is the usual practice. The software development platform for version 6 of the [[QNX]] operating system has an unsigned 32-bit <code>time_t</code>, though older releases used a signed type.

The [[POSIX]] and [[Open Group]] Unix specifications include the [[C standard library]], which includes the time types and functions defined in the [[time.h|<code>&lt;time.h&gt;</code>]] header file. The ISO C standard states that <code>time_t</code> must be an arithmetic type, but does not mandate any specific type or encoding for it. POSIX requires <code>time_t</code> to be an integer type, but does not mandate that it be signed or unsigned.

Unix has no tradition of directly representing non-integer Unix time numbers as binary fractions. Instead, times with sub-second precision are represented using [[composite data type]]s that consist of two integers, the first being a <code>time_t</code> (the integral part of the Unix time), and the second being the fractional part of the time number in millionths (in <code>struct timeval</code>) or billionths (in <code>struct timespec</code>).<ref>{{cite web |title=timespec |url=https://netbsd.gw.com/cgi-bin/man-cgi?timespec |website=NetBSD Manual Pages |access-date=5 July 2019 |date=12 April 2011 |archive-date=10 August 2019 |archive-url=https://web.archive.org/web/20190810060329/https://netbsd.gw.com/cgi-bin/man-cgi?timespec |url-status=live }}</ref><ref>{{cite web |title=time.h(0P) |url=http://man7.org/linux/man-pages/man0/time.h.0p.html |website=Linux manual page |access-date=5 July 2019 |archive-date=27 June 2019 |archive-url=https://web.archive.org/web/20190627152007/http://man7.org/linux/man-pages/man0/time.h.0p.html |url-status=live }}</ref> These structures provide a [[decimal]]-based [[Fixed-point arithmetic|fixed-point]] data format, which is useful for some applications, and trivial to convert for others.