[2256] in Kerberos_V5_Development
Re: leap seconds
daemon@ATHENA.MIT.EDU (Theodore Y. Ts'o)
Thu Feb 20 22:41:06 1997
Date: Thu, 20 Feb 1997 22:38:15 -0500
From: "Theodore Y. Ts'o" <tytso@MIT.EDU>
To: Ken Raeburn <raeburn@cygnus.com>
Cc: krbdev@MIT.EDU
In-Reply-To: Ken Raeburn's message of Thu, 20 Feb 1997 21:38:44 -0500,
<9702210238.AA30905@cujo.cygnus.com>
Date: Thu, 20 Feb 1997 21:38:44 -0500
From: Ken Raeburn <raeburn@cygnus.com>
While investigating some ASN.1 test failures I ran into, and talking
to Mark and Marc, the question of leap seconds came up...
Is the ASN.1 representation intended to take leap seconds into
account? My freshly-updated Debian Linux box now appears to have a
gmtime that does so, and the tests fail.
Well, the important thing is that a time which is asn1_decoded, and then
asn1_encoded, is the same as it originally started. Currently, the
gmt_mktime() in doesn't take this into account, so a gmtime() which does
do the leap second thing will get us into trouble.
Jik once sent in a patch which iterated with gmtime() in order account
for leap seconds, and I think that if we do have a problem with gmtime,
we're going to have to use a patch like that. It sounds like we need to
get that patch into the tree.
- Ted
