Cray pointer paperwork is done.

[Andrew Pinski]

> 
>    So that basically nixes any chance of getting this code in time for
> gcc 4.1's branch. FSF really needs to simplify their license process
> as currently it sounds like they set an insanely high barrier to
> participation.

Actually the FSF's process is simple if followed correctly.  In this case
it was not followed so please don't blame it on the FSF.

-- Pinski