[ARGUS] RE: problem with dates (fwd)

[Carter Bullard]

Hey James,
   Hmmm, I looked into this and there are some problems with
running this on sparc64 machines, as there are several
long/int casts that will cause problems.  I will do something
about the 64 bit thing after we get the current release out.

   Not to be toooooo reactive to John-Mark's reply, but if Sun
changes the size of struct time_t, hard to rationalize that
it's the programs that are broken.

Carter





-----Original Message-----
From: Gill, James [mailto:jamgill at mci.com]
Sent: Tuesday, March 30, 2004 2:45 PM
To: Carter Bullard
Subject: Re: problem with dates (fwd)


Hi Carter,

I haven't gotten the 2.0.6 version tested for my platform yet, but this
seems to detail why we're seeing the date wierdness.  Do you recognize
this and is it addressed in the 2.0.6 version?

Thank You,

--gill


        -----------------------------------------------------
        MCI/UUNET Network Security & Abuse * 1-800-900-0241,4
        -----------------------------------------------------
             v-net:  desk = 806-3834 ; group = 806-8805

---------- Forwarded message ----------
Date: Tue, 30 Mar 2004 10:05:20 -0800
From: John-Mark Gurney <gurney_j at efn.org>
To: RJ45 <rj45 at slacknet.com>
Cc: freebsd-sparc64 at freebsd.org
Subject: Re: problem with dates

RJ45 wrote this message on Tue, Mar 30, 2004 at 10:39 -0700:
> even if this si the current time:
>
> Tue Mar 30 19:39:04 CEST 2004
>
> anyone got this problem?
> thanks

Looks like the code is broken in that it assumes that time_t is an int,
and passes a pointer to a time_t to another function that casts it to
an integer...  This would turn up this problem since time_t's are now
64bits, and int's are 32bits, and since sparc64 is a big endian platform,
the first 32bits of a 64bit time_t are usually zero...

Tell the author of the program that his code is broken.

--
  John-Mark Gurney				Voice: +1 415 225 5579