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

[Stephen Hoffman]

On 2019-04-01 20:08:57 +0000, pcoviello at gmail.com said:


> On 4/1/2019 10:00 AM, Stephen Hoffman wrote:
>> 
>> The Heisei Era is ending, and the Reiwa era commences on 1-May-2019.  
>> (Whether that's 1-May local time or 1-May UTC, I've not confirmed.)
>> 
>> There are changes happening several times a year, and variously more often.
>> 
>> It's rare to hear about TZ changes else-planet.  Including among 
>> customers, when folks are (knowingly or otherwise) running 
>> planetary-scale apps.
> 
> I have to ask why one would not be using NST for the time anyways?  
> would this not solve this issue?  or did I miss something?

If you meant NTP there and not NST, then NTP gets you the ~current UTC 
time; a value for ~now.

NTP doesn't get you the Timezone (TZ) data and the current Daylight 
Saving Time (DST) thresholds, and won't help with all but the simplest 
and most short-term of time-related math.

NTP won't get you the shift from the Heisei to Reiwa era, for instance.

Which means you'll have a UTC time value, but you'll not correctly 
display the local time in the Reiwa era that starts next month.

What time it is locally depends not only on UTC, but on where you are 
located, and when you are on the local calendar.

Those details and display-related details are acquired from the TZ 
file.  And the TZ file contents changes.  Such as with the Heisei to 
Reiwa change.

Which means downloading the TZ definitions

Either from VSI via a TZ-related patch, or ad-hoc across the installed 
base.  With the ad-hoc approach common now and the infrequently-issued 
TZ patches, most sites will ignore the TZ changes until after somebody 
runs into a problem.  Then various of the folks will either call for 
support and get a patch or get instructions on a manual change, or both.

Reading:  https://en.wikipedia.org/wiki/tz_database




-- 
Pure Personal Opinion | HoffmanLabs LLC