[tz] DB incoherence

Guy Harris guy at alum.mit.edu
Tue Apr 7 21:22:56 UTC 2020


On Apr 7, 2020, at 7:55 AM, Américo Mendes <sednem69 at gmail.com> wrote:

> Was trying to use your DB and I found what I think it's a data incoherence...
> When you list the rules, i suppose that they must be in chronological order, I'm assuming that (X) rule start date always smaller than (X+1) rule start date,

That's not a requirement.  Software that reads tzdb source files needs to be able to handle that.

> I fund in the Haiti rules the following section:
> R HT 1983 o - May 8 0 1 D
> R HT 1984 1987 - Ap lastSu 0 1 D
> R HT 1983 1987 - O lastSu 0 0 S
> According to my assumption, to keep DB coherence, we should have, in 2nd position something like:
> R HT 1983 o - O lastSu 0 0 S
> 
> Am I right or is it some particularity that I didn't understand?

The rule

	Rule Haiti 1983 only - May 8 0:00 1:00 D

says "in 1983, and only in 1983, on 8 May, at 00:00, the clocks are set to be 1 hour later than the normal time".

The rule

	Rule Haiti 1984 1987 - Apr lastSun 0:00 1:00 D

says "from 1984 to 1987, on the last Sunday in April, at 00:00, the clocks are set to be 1 hour later than the normal time".

The rule

	Rule Haiti 1983 1987 - Oct lastSun 0:00 0 S

says "from 1983 to 1987, on the last Sunday in October, at 00:00, the clocks are set to be the normal time".

I assume they're grouped that way in the source file so that you have two rules describing the "when the clocks are set forward" between 1983 and 1987, followed by one rule describing "when the clocks are set back" between 1983 and 1987.


More information about the tz mailing list