Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!ames!oliveb!sun!gorodish!guy
From: guy@gorodish.UUCP
Newsgroups: comp.unix.wizards
Subject: Re: More DST tinkering
Message-ID: <15590@sun.uucp>
Date: Wed, 25-Mar-87 02:57:03 EST
Article-I.D.: sun.15590
Posted: Wed Mar 25 02:57:03 1987
Date-Received: Fri, 27-Mar-87 00:46:19 EST
References: <8784@decwrl.DEC.COM>
Sender: news@sun.uucp
Lines: 40
Keywords: daylight time, elections
Summary: Yes, Olson's does

> That's right, in presidential election years, daylight time would be
> extended by two weeks in the Pacific time zone only. (No word on what happens
> in Hawaii.)
> 
> Do any of the recent proposals for parameter driven time conversion deal
> with this class of tinkering?

From the "pacificnew" file in the Arthur Olson timezone code, as
posted to net.sources:

# @(#)pacificnew	3.1

# "Pacific Presidential Election Time" has been contemplated by the US Congress

# Rule	NAME	FROM	TO	TYPE	IN	ON	AT	SAVE	LETTER/S
Rule	Twilite	1987	max	-	Apr	Sun>=1	2:00	1:00	D
Rule	Twilite	1987	max	uspres	Oct	lastSun	2:00	1:00	PE
Rule	Twilite	1987	max	uspres	Nov	Sun>=7	2:00	0	S
Rule	Twilite	1987	max	nonpres	Oct	lastSun	2:00	0	S

# Zone	NAME		GMTOFF	RULES/SAVE	FORMAT	[UNTIL]
Zone	US/Pacific-New	-8:00	US		P%sT	1987
			-8:00	Twilite		P%sT

The types "uspres" and "nonpres" are hardcoded into the time zone
file compiler for efficiency, but other types are supported; the
compiler calls "system" with the command

	yearistype <year> <type>

where <year> is the year number and <type> is the type string.  If
"yearistype" exits with an exit code of 0 (i.e., normal UNIX "true"),
the year is considered to be of that type; otherwise, if it exits
with an exit code of 1, the year is considered not to be of that
type; otherwise, "yearistype" is assumed to be badly confused and the
compiler quits after printing an error.

This is done when the tables are built; the tables are just lists of
transition times, so the only overhead this feature causes is
whatever overhead is caused by the increase in table size.