[tz] Fwd: [IANA #1283803] Europe/Amsterdam
Paul Eggert
eggert at cs.ucla.edu
Wed Oct 18 02:46:23 UTC 2023
On 10/17/23 04:08, Leo Bekhuis via tz wrote:
> *Is Europe/Amsterdam currently an IANA compliant timezone?*
"An IANA compliant timezone" is not a well-defined concept. As
<https://data.iana.org/time-zones/tz-link.html> says, the tz code and
data are by no means authoritative. (Brian discussed this in more detail
in his reply.)
Possibly you are asking about timestamps before the year 1970? Although
TZDB's scope is timestamps since 1970, because of its formats it also
covers pre-1970 timestamps in limited ways. In the latest TZDB release
(2023c) with the default settings, Europe/Amsterdam is a Link to
Europe/Brussels so their timestamps are identical. However, if you use
some non-default options when building from 2023c, Europe/Amsterdam
becomes a separate Zone with data that differ from Europe/Brussels for
some pre-1970 timestamps; although this optional Zone is also incorrect
for some pre-1970 timestamps, it's closer to being right. The
non-default options exist for compatibility with older verions of TZDB
that were more ambitious (arguably too ambitious) in their attempt to
cover pre-1970 timestamps.
One other thing: downstream distributions commonly tailor TZDB for their
own needs. Often there's no combination of options for the
IANA-distributed TZDB that will generate timestamps that exactly match
what you observe on your own system. So one must be careful about
phrases like "IANA compliant timezone". If you want 100% reproducible
timestamps, you'll need to give more detail about how you derived the
timestamps.
More information about the tz
mailing list