On Mon, Dec 26, 2005 at 04:22:14PM +0000, Christos Zoulas wrote:
> In article <8764pd5c21.fsf@snark.piermont.com>,
> Perry E. Metzger <perry@piermont.com> wrote:
> >
> >YAMAMOTO Takashi <yamt@mwd.biglobe.ne.jp> writes:
> >>> Module Name:	src
> >>> Committed By:	perry
> >>> Date:		Sat Dec 24 23:04:05 UTC 2005
> >>> 
> >>> Modified Files:
> >>> 	src/include: stdlib.h
> >>> 
> >>> Log Message:
> >>> __const__ -> const
> >>> This is in an __attribute__ so it isn't really necessary, but it gets
> >>> it out of the listing of __keyword__s...
> >>> 
> >>> 
> >>> To generate a diff of this commit:
> >>> cvs rdiff -r1.73 -r1.74 src/include/stdlib.h
> >>
> >> why?
> >
> >The "why" is so that sweeps looking for __const__ don't find it any
> >more. It isn't actually important -- it just makes the rest of the job
> >easier. As I said "isn't really necessary, but gets it out of the
> >listing".
> >
> 
> I am not sure if this is correct though. I.e. some programs do -Dconst=
> to avoid const, but in this case "const" means something different.
> 

Programs which do that are arguably just plain wrong. They fall into the
same bucket with me as ones which compile with > /dev/null 2>&1 on the compiler
line...

James