RE: mktime()

["Howland Craig D (Craig)" <howland at LGSInnovations dot com>]

I've taken a look and agree with the logical analysis that an adjustment
is needed, and that the patch given does so appropriately.  (Perhaps
this saves Corinna a few minutes.)
Craig

-----Original Message-----
From: newlib-owner@sourceware.org [mailto:newlib-owner@sourceware.org]
On Behalf Of Steven Abner
Sent: Tuesday, September 13, 2011 12:29 PM
To: newlib@sourceware.org
Subject: Re: mktime()


Hi
   Thought you might like this? :0 Figured I'd keep within thread.
Test case not supplied, just logic:
Given Jan 1st in the British Isles, UTC0, and they decide to go DST on
Jan 1st at 2AM of 1 hour, the POSIX standard format for that is ",J1/2".
From tzset()
	__tzrule[i].ch = J;
	__tzrule[i].d = 1;
From __tzcalc_limits
	days = 0 + 1 + 0;
and
	change = 1 * 86400 + 7200 + 0;
instead of 7200.

  Then again I might have a loose light bulb?

Steve