The International Earth Rotation and Reference Systems Service (IERS),
which has jusidiction over such things, has announced the introduction
of a leap-second for Universal Coordinated Time (UTC) at the end of
December 2005 (Bulletin C announcement:
<http://hpiers.obspm.fr/eoppc/bul/bulc/bulletinc.dat>).

Users of SRFI-19's UTC's functions will need to modify the internal
leap second table to take this additional leap second into account. If
you use the SRFI implementation, or some close variant, the correct
new leap second table definition is:

(define tm:leap-second-table
 '((1136073600 . 33)
  (915148800 . 32)
  (867715200 . 31)
  (820454400 . 30)
  (773020800 . 29)
  (741484800 . 28)
  (709948800 . 27)
  (662688000 . 26)
  (631152000 . 25)
  (567993600 . 24)
  (489024000 . 23)
  (425865600 . 22)
  (394329600 . 21)
  (362793600 . 20)
  (315532800 . 19)
  (283996800 . 18)
  (252460800 . 17)
  (220924800 . 16)
  (189302400 . 15)
  (157766400 . 14)
  (126230400 . 13)
  (94694400 . 12)
  (78796800 . 11)
  (63072000 . 10)))

Could the current SRFI editor for SRFI-19 make the above change in the
reference implementation?

--
Will Fitzgerald
weblog: <http://www.entish.org/willwhim>

"At the tone, the time will be 2005 December 31, 23h 59m 60s."