| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |
From: John T Kohl <jtkohl@ATHENA.MIT.EDU> To: kerberos@ATHENA.MIT.EDU Cc: Jeffrey I. Schiller <jis@ATHENA.MIT.EDU> In-Reply-To: Jeffrey I. Schiller's message of Mon, 19 Dec 88 23:49:06 EST, I thought about this a bit last night, and after reading Jeff Schiller's concerns, I changed my mind and only made a slight change in format, extending the year field to 4 digits, and using GMT. The input side is capable of detecting old-style dumps and assuming 19xx and local time. It really isn't hard to parse the field into a (struct tm), and the routines I picked up from RCS convert that into a 32-bit time [That was the real hair in the code that came from /bin/date]. The problem that might remain is reading a new dump into an old kdb_util, which would get horribly confused by the date extension. Further comments? John
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |