patch for merging graphite branch (before tuplification)

[Sebastian Pop]

Hi,

On Sun, Aug 3, 2008 at 10:25 AM, Jack Howarth <howarth@bromo.msbb.uc.edu> wrote:
>   I am a bit confused about the comment that building the cloog
> and ppl libs in the gcc directory is broken. Do you intend to
> actually merge their source code into gcc?

No.  I do not intend to have cloog and ppl included in gcc sources.
They should be separate libs.

> I thought they would
> be treated like mpfr (as an external development library provided
> by the the OS distribution).

Exactly.  Currently gcc supports the build of the mpfr and gmp libs
in the sources of gcc: that means that you can just drop in the toplevel
directory of gcc the mpfr and gmp libs, and just configure gcc that
will recursively configure and build these libs.  I had such support
for the previous cloog/polylib, but after ppl changes I was not able to
build ppl within the same directory.  If I cannot figure out how to fix
this, I will remove this building feature.

> Also, can the tuplified graphite
> changes be posted as a patch that applies to current gcc trunk?

I'm still addressing some of the comments that Joseph asked me
for, and then I was planing to send the patches split as Richi asked
for review.

> ps I am the fink maintainer of the gcc packages on Mac OS X
> and had intended to just create ppl/polylib and cloog packages
> that would be required to build/use our gcc44 package.

Cool.  Thanks for testing on this MacOS.  Let me know if something
is broken.  I was using a mac last year for developing graphite,
and there seemed to be no problems, but I don't have access to
such a machine anymore.

Sebastian