On Fri, 29 Jan 1999 "Tabor J. Wells" <twells@???> writes:
> On Fri, Jan 29, 1999 at 02:40:36PM -0500,
> Peter Radcliffe <pir@???> is thought to have said:
>
> > There is a comment on the web page:
> > http://www-th.phys.rug.nl/~schut/gnulist.html
> > stating that exim is not Y2k compliant:
> >
> > exim-2.02 1998-09-24 JJ.Schut NOT-OK Assumes year between
> > 1990 and 2090 in smartuser.c, lines 243+.
> > if (year < 90) year += 100;
> > else if (year > 1900) year -= 1900;
> >
> > This doesn't make sense to me. If exim is ok until 2090, why is it not y2k
> > safe ?
> > (besides the point that using 2 digit years in any context is a Bad Thing).
>
> It looks like his definition of "NOT-OK" is (from elsewhere on that page):
>
> "Not compliant code is code which handles dates wrong in any sense. This
> does not necessary mean that the package does not work anymore after
> 1999-12-31...Reason is that the package apparently expects the year to be
> 2-digit."
>
> Looks like he's really only checking for 2 digit years.
If this is the only "non-compliance" he can find in Exim, then we really
are well off! The code complained of is included only if ALLOW_EXPIRE
is set at compilation time, and it's part of an experimental scheme for
automatically timing out mail addresses in order to defeat spammers who
collect email addresses from Usenet postings. I think Philip has pretty
much abandoned it by now.
If you look at the code carefully, and assume that time_t and tm structs
have fields big enough not to worry about 1970-2038 range restrictions,
then the complaint boils down to "it's impossible to make an address expire
earlier than 1900 by using this feature". (Earlier than 1990 can be achieved
by using 4-digit years.)
Chris Thompson Cambridge University Computing Service,
Email: cet1@??? New Museums Site, Cambridge CB2 3QG,
Phone: +44 1223 334715 United Kingdom.
--
*** Exim information can be found at
http://www.exim.org/ ***
