Seeking clarifaction of tai64nlocal and leap seconds

Brian Reichert <[email protected]> Fri, 1 Apr 2011 14:47:41 -0400
Newsgroups gmane.comp.djb.syslog
Message-ID <[email protected]>
This page about tai64nlocal:

  http://cr.yp.to/daemontools/tai64nlocal.html

calls out

  Some localtime implementations use a broken time scale that does not
  account for leap seconds. On systems that use the Olson tz library (with
  an up-to-date leap-second table), you can fix this problem by setting your
  time zone to, e.g, right/US/Pacific instead of US/Pacific.

This sentence to me implies:

If you have both right/US/Pacific and US/Pacific timezones, then 

- right/US/Pacific has the leap seconds included

- and hence should be twentyish seconds 'ahead' of US/Pacific.

Om my CentOS box, though, it seems the other way around:

  # env TZ=right/UTC date; env TZ=UTC date
  Fri Apr  1 18:43:02 UTC 2011
  Fri Apr  1 18:43:26 UTC 2011

Have I misenterpreted djb's statement on that web page?

FWIW, my enviromment:

   # cat /etc/redhat-release 
   CentOS release 5.4 (Final)
   # rpm -qf /usr/share/zoneinfo/right/UTC 
   tzdata-2011b-1.el5

Thanks for any advice...

-- 
Brian Reichert				<[email protected]>
BSD admin/developer at large