[Info-vax] EU will abandon daylight savings time in 2021

[Craig A. Berry]

On 4/4/19 12:53 PM, Bob Koehler wrote:
> In article <q84l2a$s2o$1 at dont-email.me>, =?UTF-8?Q?Jan-Erik_S=c3=b6derholm?= <jan-erik.soderholm at telia.com> writes:
>> Den 2019-04-04 kl. 03:27, skrev Arne Vajhøj:
>>> On 4/3/2019 1:37 PM, Jan-Erik Söderholm wrote:
>>>>                         We have some local timestamps that has to be
>>>> converted to GMT/UTC before sent to another system, and these are both
>>>> from "winter time" and "summer time". We can of course hardcode the
>>>> dates for the switches and compare, but it would be nice if the rule
>>>> could be used...
>>>
>>> What do you have and what do you need to convert it to?
>>
>> An INTEGER "date" with format YYYYMMDD, ag 20190404 for today.
>>
>> A SMALLINT "time" with format HMM or HHMM, range from 1 (one minute
>> after midnight) to 2359 (one minute before midnight). I guess
>> it could be zero also for the first minute after midnight, but
>> the system is having it's night-batch windows then.
>>
> 
>     Double precisionn floating point date/time in the format
>        YYYMMDD.hhmmss .... (to whatever fraction of seconds 64 bits holds)
> 
>     Been in use since the 1960s in a package that operates with lots of
>     different epochs, handles leap seconds, ...

Or you can use the 53 bits of the mantissa of an IEEE double as the
number of seconds since the epoch and have plenty left for that brief
period of time in which our sun is still burning.  After that it's Dr.
Who's problem.