This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix Cloog configure check


On Thu, 4 Sep 2008, Matthias Klose wrote:

> Sebastian Pop writes:
> > On Wed, Sep 3, 2008 at 9:35 AM, Richard Guenther <rguenther@suse.de> wrote:
> > >
> > > If you do not specify --with-cloog then the cloog test fails even though
> > > everything is ok because -DCLOOG_PPL_BACKEND is not passed to the
> > > check causing a non-existant polylib header to be included from cloog.h.
> > >
> > 
> > I see: this happens only when cloog is installed in a standard
> > location, and then one does not have to specify --with-cloog to build
> > GCC with graphite.
> 
> I was a bit confused how to configure this. the toplevel configure
> checks for ppl and cloog independently regardless of any --with-*
> options. however the gcc configure requires the --with-*=/usr
> explicitely or else it won't be picked up. is this intended? please
> add documentation in gcc/doc/install.texi about this.

Are you sure?  (ok, I didn't check if the built gcc actually enabled
-fgraphite)  At least the toplevel claims to find both ppl and
cloog for me and the gcc/ configure doesnt' contain any cloog or
ppl checks.

Doh, you are right:

/* Define if cloog is in use. */
#ifndef USED_FOR_TARGET
/* #undef HAVE_cloog */
#endif

this sucks ;)

Richard.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]