home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.protocols.kerberos
- Path: sparky!uunet!gatech!udel!wupost!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!agate!stanford.edu!EDDIE.MIT.EDU!harvard!scubed!riipsdev.waterloo.ncr.com!ajlill
- From: harvard!scubed!riipsdev.waterloo.ncr.com!ajlill@EDDIE.MIT.EDU (Tony Lill)
- Subject: Re: Porting problems in V5
- Message-ID: <9301041504.AA03631@riipsdev.waterloo.ncr.com>
- Sender: news@shelby.stanford.edu (USENET News System)
- Organization: Internet-USENET Gateway at Stanford University
- References: <1gll1cINNs0o@usenet.INS.CWRU.Edu>
- Date: Mon, 4 Jan 1993 15:04:04 GMT
- Lines: 44
-
- >> In article <9212082146.AA07100@riipsdev.waterloo.ncr.com> harvard!scubed!riipsdev.waterloo.ncr.com!ajlill@eddie.mit.edu writes:
- >> >For those of you lucky enough not to have dealt extensively with DOS don't
- >> >know, the DOS time() call returns the number of seconds since midnight, Dec
- >> >31, 1899, rather than Jan 1, 1970.
- >>
- >> That's strange. Every DOS C compiler I've used returns seconds since
- >> Jan. 1, 1970. The only time I've seen something based on 1899/1900
- >> was in code that queried a time server for the time. time()'s very
- >> definition depends on the Unix epoch of 1/1/70.
-
- Surprise! In C 7.0, they changed the definition of time.
-
- >> No matter what, you have the right solution, I think. A better solution,
- >> if possible, would be to fix the offset used by your time() call.
-
- Yes, after trying to make Kerberos work with this weird definition of
- time, I've found that for the decrypt integrity checks (as written) to
- work depends on both machines using the same epoch to encode time 0,
- so I'll be changing krb5_timestamp to offset it. I've already provided
- my own gmtime since the one in the Windows 3.1 SDK libraries is sooooo
- broken, it's not funny (they don't even have a large model object in
- the large model library!).
-
- Windows/NT the death of UNIX? At least UNIX knows what time it is!
-
- >> >Pepsy (from isode 8.0) turns that into an int when it creates it's
- >> >tables.
- >>
- >> That's a bug in Pepsy-on-DOS, or Pepsy-on-PDP-11 for that matter. :-)
- >> Someone did a port of ISODE to DOS a while back. You might see if you
- >> can find that port, since the Pepsy there probably has this bug fixed.
-
- Actually, that problem turned out to be a bug in int2prim() or
- something not getting what was expected in 0x1ff<<24. (Hint: this is 0
- on a 16 bit machine).
-
- - ---
- Tony Lill, Tony.Lill@Waterloo.NCR.COM
- President, A. J. Lill Consultants (519) 650 0660
- 539 Grand Valley Dr., Cambridge, Ont. (519) 653 9732
- presently at E&M Waterloo, NCR Canada Ltd. (519) 884 1710 x624
- voice plus 643 1624
-
- "Welcome to All Things UNIX, where if it's not UNIX, it's CRAP!"
-