This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH]: Pass -D__HAVE_SHORT_DOUBLE__ when -fshort-double(HC11/HC12)
On Sat, May 05, 2001 at 08:50:04AM +0200, Stephane Carrez wrote:
> Hi!
>
> "Joseph S. Myers" a ?crit :
> >
> > On Fri, 4 May 2001, Stephane Carrez wrote:
> >
> > > To compile newlib correctly (_DOUBLE_IS_32BITS), pass -D__HAVE_SHORT_DOUBLE__
> > > when -fshort-double option is specified.
> >
> > I think this should be target-independent, since the -fshort-double option
> > is. (In that case, I'd also prefer __SHORT_DOUBLE__ to
> > __HAVE_SHORT_DOUBLE__ as the name of the macro.) It should also be
> > documented in cpp.texi (but so should machine-dependent defines).
> >
>
> Yes. It would be better. If nobody has a patch, I'll propose one.
>
> Concerning the documentation, it seems that machine-dependent defines are
> not documented for many architectures. Can you clarify where such defines
> should be documented?
>
> Would: `* Nonstandard Predefined' be the appropriate place?
> or, should we introduce a machine-dependent section?
In a new section, please. 'Nonstandard Predefined' has been brutally
trimmed in my cpp rewrite - please take a look:
http://www.stanford.edu/~zackw/cpp-doc/cpp_toc.html.
I'd prefer if people would hold off on any cpp manual patches at this
point until I check in the rewrite, which will be no later than the
end of next week.
zw