Re: HEADS UP! MAJOR change to FreeBSD/sparc64

From: Garance A Drosihn <drosih_at_rpi.edu>
Date: Sun, 14 Mar 2004 21:12:32 -0500
At 5:16 PM +1000 3/14/04, Stephen McKay wrote:
>On Wednesday, 10th March 2004, Garance A Drosihn wrote:
>
>>The long-threatened change for FreeBSD/sparc64 has now been
>>committed.  This changes time_t to be a 64-bit quantity, the
>>same as it is for the AMD64 and IA64 architectures.  People
>>running FreeBSD/sparc64 *will* have to read the UPDATING.64BTT
>>file for instructions on how to safely build and install this
>>change.
>
>The change to 64-bit time is essential, of course, but I don't
>understand why it has to break backward compatibility.

When the people who were running sparc64 were asked about it, the
feeling was that there wasn't much point in backwards compatibility
for a platform that presently has so few users.  The 5.x branch is
the only branch which has supported sparc64, and that is still a
"cutting edge" branch where major changes like this are acceptable.
The sparc64 users (including me) felt that if we "just did it" now,
we would be much happier than waiting for 6.0 to roll around.

I'm sure there will still be a few bumps in making this transition,
but on the whole I think we're better off trying to make it now.

>What else is going on?  (I don't have a Sparc or I'd join
>your experiment.)

What?  This wasn't dramatic enough?   :-)

>  > This only affects freebsd-current, of course.  Later we'll
>  > have to decide the best upgrade method for people who make
>  > the jump from RELENG_5_2 to the upcoming RELENG_5_3.
>
>I'm thinking of the "later" where i386 is changed to 64-bit
>time_t. If that's not backward compatible, you'll get no takers.

The users of each platform will make the decision for that platform.

In the case of i386, there is a 10-year history of servers and
programs running with 32-bTT, in production.  I do also run
freebsd/i386, and for that platform I really can not imagine making
this major a change without providing backward-compatibility.  There
is just too much written which assumes 32-bTT, including programs
which perhaps can not be recompiled.  I do not see that happening
before the 6.0-branch.

There is some chance that the PowerPC port may make this same
transition in about the same way that sparc64 did.  But again, it
will be based on how the users of that platform wish to handle it.

-- 
Garance Alistair Drosehn            =   gad_at_gilead.netel.rpi.edu
Senior Systems Programmer           or  gad_at_freebsd.org
Rensselaer Polytechnic Institute    or  drosih_at_rpi.edu
Received on Sun Mar 14 2004 - 17:12:35 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:47 UTC