Proposed change to "systemv" file

[Paul Eggert]

Olson, Arthur David (NIH/NCI) wrote:

>The attached change avoids a "link-to-a-link" situation in the "systemv"
>data file.

Thanks for catching that; I had missed systemv somehow.  However,
Indiana will observe DST next year, so we probably should stop
equating Indianapolis with "no DST".  Also, in reviewing the comments,
they need updating in the light of the current US plans for DST, and I
found the first line a bit confusing.

How about the following further patch?  It attempts to address these issues.

Does anyone know when the SystemV rules were written?  If so, that
"1976" below should be changed to that date.  It had to be some time
in the 1976/1987 time frame.

--- systemv	2005/10/03 13:56:37	2001.1.0.1
+++ systemv	2005/10/04 04:31:57	2001.1.0.2
@@ -1,10 +1,10 @@
 # @(#)systemv	7.4
 
-# Old rules, should the need arise.
+# The commented-out entries below contain old rules, should the need arise.
 # No attempt is made to handle Newfoundland, since it cannot be expressed
 # using the System V "TZ" scheme (half-hour offset), or anything outside
 # North America (no support for non-standard DST start/end dates), nor
-# the change in the DST rules in the US in 1987 (which occurred before
+# the changes in the DST rules in the US after 1976 (which occurred after
 # the old rules were written).
 #
 # If you need the old rules, uncomment ## lines and comment-out Link lines.
@@ -42,7 +42,7 @@ Link	America/Denver		SystemV/MST7MDT
 Link	America/Los_Angeles	SystemV/PST8PDT
 Link	America/Anchorage	SystemV/YST9YDT
 Link	America/Puerto_Rico	SystemV/AST4
-Link	America/Indiana/Indianapolis	SystemV/EST5
+Link	America/Panama		SystemV/EST5
 Link	America/Regina		SystemV/CST6
 Link	America/Phoenix		SystemV/MST7
 Link	Pacific/Pitcairn	SystemV/PST8