Data structure for Unix time.

Please note that this uses GHC-derived Eq and Ord instances. Notably

>>> UnixTime 1 0 > UnixTime 0 999999999
True

You should instead use UnixDiffTime along with its helpers such as microSecondsToUnixDiffTime which will ensure that such unusual values are never created.

Getting UnixTime from OS.

Parsing ByteString to UnixTime interpreting as localtime. This is a wrapper for strptime_l(). Many implementations of strptime_l() do not support %Z and some implementations of strptime_l() do not support %z, either. utMicroSeconds is always set to 0.

Parsing ByteString to UnixTime interpreting as GMT. This is a wrapper for strptime_l(). utMicroSeconds is always set to 0.

>>> parseUnixTimeGMT webDateFormat "Thu, 01 Jan 1970 00:00:00 GMT"
UnixTime {utSeconds = 0, utMicroSeconds = 0}

Formatting UnixTime to ByteString in local time. This is a wrapper for strftime_l(). utMicroSeconds is ignored. The result depends on the TZ environment variable.

Formatting UnixTime to ByteString in GMT. This is a wrapper for strftime_l(). utMicroSeconds is ignored.

>>> formatUnixTimeGMT webDateFormat $ UnixTime 0 0
"Thu, 01 Jan 1970 00:00:00 GMT"
>>> let ut = UnixTime 100 200
>>> let str = formatUnixTimeGMT "%s" ut
>>> let ut' = parseUnixTimeGMT "%s" str
>>> ((==) `on` utSeconds) ut ut'
True
>>> ((==) `on` utMicroSeconds) ut ut'
False

Format of the strptime()/strftime() style.

Format for web (RFC 2616). The value is "%a, %d %b %Y %H:%M:%S GMT". This should be used with formatUnixTimeGMT and parseUnixTimeGMT.

Format for e-mail (RFC 5322). The value is "%a, %d %b %Y %H:%M:%S %z". This should be used with formatUnixTime and parseUnixTime.

Data structure for UnixTime diff.

It is up to the user to ensure that udtMicroSeconds < 1000000. Helpers such as microSecondsToUnixDiffTime can help you to create valid values. For example, it's a mistake to use addUnixDiffTime with a value UnixDiffTime 0 9999999 as it will produce an incorrect value back. You should instead use functions such as microSecondsToUnixDiffTime to create values that are in-range. This avoids any gotchas when then doing comparisons.

Calculating difference between two UnixTime.

>>> UnixTime 100 2000 `diffUnixTime` UnixTime 98 2100
UnixDiffTime {udtSeconds = 1, udtMicroSeconds = 999900}

Adding difference to UnixTime.

>>> UnixTime 100 2000 `addUnixDiffTime` microSecondsToUnixDiffTime (-1003000)
UnixTime {utSeconds = 98, utMicroSeconds = 999000}

Creating difference from seconds.

>>> secondsToUnixDiffTime 100
UnixDiffTime {udtSeconds = 100, udtMicroSeconds = 0}

Creating difference from micro seconds.

>>> microSecondsToUnixDiffTime 12345678
UnixDiffTime {udtSeconds = 12, udtMicroSeconds = 345678}

>>> microSecondsToUnixDiffTime (-12345678)
UnixDiffTime {udtSeconds = -12, udtMicroSeconds = -345678}

From EpochTime to UnixTime setting utMicroSeconds to 0.

From UnixTime to EpochTime ignoring utMicroSeconds.

From ClockTime to UnixTime.

From UnixTime to ClockTime.