[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: AC_C_BIGENDIAN answers "universal" on powerpc-aix

From: Peter O'Gorman
Subject: Re: AC_C_BIGENDIAN answers "universal" on powerpc-aix
Date: Fri, 15 Aug 2008 00:02:02 -0500
User-agent: Mutt/1.5.17 (2007-11-01)

On Thu, Aug 14, 2008 at 10:34:06PM -0600, Eric Blake wrote:
> Hash: SHA1
> > 
> > I think this one addresses your nits.
> Two other nits:
> This name means that since AC_[A-Z]* is now expected to appear in
> configure, it should trip the m4_pattern_forbid of the autoconf m4
> namespace as a potentially unexpanded macro (I wonder why it didn't seem
> to trip in your testing?).  Why not just strip the AC_ prefix, and go with

It seems likely that someone is already using that define in a progect.
The macro is meant to be internal to autoconf, why not have it in
autoconf namespace?

> > +#  define WORDS_BIGENDIAN 1
> Not quite right, since this is for an autoheader template.  In the
> AH_VERBATIM, it should look like the instance four lines later:
> > +#  undef WORDS_BIGENDIAN
> and let config.status do the conversion to the desired state in both
> locations.

No, it is correct as is. The compiler, for one invocation will go over
the sources multiple times, once for each -arch setting. For x86 builds
it will do so with the __LITTLE_ENDIAN__ macro defined and set to 1, for
ppc builds it will have the __BIG_ENDIAN__ macro defined and set to 1,
regardless of the build machines endianness.

Peter O'Gorman

reply via email to

[Prev in Thread] Current Thread [Next in Thread]