On Tue, 25 May 2010, Andrew Morton wrote:
>
> > I suspect fixing this entails moving all endian tests to #if __BYTE_ORDER ...
> > and then always define both __LITTLE_ENDIAN and __BIG_ENDIAN
>
> Yes we can't define __LITTLE_ENDIAN to anything on big-endian without
> breaking lots of other things.
Indeed. The whole glibc thing of doing
# if __BYTE_ORDER == __BIG_ENDIAN
(see "grep ENDIAN /usr/include/*") is a f*cking idiotic. It is way less
readable, and insane to do. I guess it's some traditional unix disease.
The way to do it in Linux is to just do
#ifdef __{BIG,LITTLE}_ENDIAN
which is a lot more readable.
If anybody wants to change that, they still shouldn't change it to the
idiotic glibc model. If you want it to be more readable, maybe somebody
can introduce a CONFIG_{BIG,LITTLE}_ENDIAN{_BITFIELDS}
Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-next" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
