why atlas for py-scipy?

[Jack Howarth]

On Sun, Jul 24, 2011 at 05:43:17AM +1000, Joshua Root wrote:
> On 2011-7-24 05:37 , Jack Howarth wrote:
> > Are there any plans to depreciate
> > some of the older gcc4x releases in MacPorts?
> 
> I can't say if anyone has concrete plans to do this in the near future,
> but it would probably be a good idea. It's certainly about time to move
> the default variant up from gcc44 in ports that use gfortran/gcj.
> 
> - Josh

Josh,
   We have made significant progress in keeping the testsuite results in decent
shape for darwin while adapting to changes in the OS...

http://gcc.gnu.org/gcc-4.6/changes.html
http://gcc.gnu.org/ml/gcc-testresults/2011-07/msg00877.html

Most of those failures are due to changes from Xcode 4.1. For Xcode 3.2.6/4.0.2, the
results are even better...

http://gcc.gnu.org/ml/gcc-testresults/2011-06/msg02515.html (really Xcode 3.0.2)
http://gcc.gnu.org/ml/gcc-testresults/2011-06/msg02555.html

In fact, except for a change discovered immediately before gcc 4.6's release that 
broke our model of handling LTO sections in Mach-O files, we would have functional
lto support in Xcode 4.x. In my fink gcc46 package, I have implemented the current
work-in-progress patch for containerization of the LTO sections so that feature
is restored.
       Jack