httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jim Jagielski <...@jaguNET.com>
Subject Re: binary backwards compatability.
Date Thu, 30 Mar 2000 03:15:33 GMT
by "box" I'm assuming possibly totally different boxes. The below
is still valid, in many cases, even on the _same_ box (as far
as alignments and sizes, though I don't know of any that change
endian).

Jim Jagielski wrote:
> 
> Sure, that'll happen. Heck, depending on the box, the endian-ness
> can change too. And if one is a 32bit int, and the other is a 64bit
> int, then the overall size will change as well. But that's out
> of ANSI's area of control, however.
> 
> Manoj Kasichainula wrote:
> > 
> > On Wed, Mar 29, 2000 at 09:51:34PM -0500, Jim Jagielski wrote:
> > > Hmmm. I don't think that ANSI can re_order_ structs. I think it
> > > says stuff about not expecting anything non-portable about
> > > alignment between struct entries. But it can't shuffle those
> > > entries around.
> > 
> > Well, lets say that I compile Apache with gcc -march=i386 on one box
> > and gcc -march=pentiumpro on another. I'm guessing (but haven't
> > checked) that alignment of the structures will change. Maybe different
> > cc's on the same platform will produce different alignments as well.
> > This is just as bad as field reordering.
> > 
> 
> 
> -- 
> ===========================================================================
>    Jim Jagielski   [|]   jim@jaguNET.com   [|]   http://www.jaguNET.com/
>                 "Are you suggesting coconuts migrate??"
> 


-- 
===========================================================================
   Jim Jagielski   [|]   jim@jaguNET.com   [|]   http://www.jaguNET.com/
                "Are you suggesting coconuts migrate??"

Mime
View raw message