Bad code generation on HPPA platform

[Hans-Peter Nilsson]

On Wed, 7 May 2008, Peter Bergner wrote:
> Note that the rtlanal.c:ÿÿcommutative_operand_precedence() hunk was reverted
> because it caused some problems on CRIS

For the record, that's a simplification.  It breaking CRIS
wasn't the main reason for the revert.  I applied a fix, but
raised a general concern.  See original reference.  I'm just
objecting to CRIS and its maintainer being portrayed as a
squeaking wheel.  (Though having an autotester, I tend to see
breakage early, but it's usually quite general.)

> and was replaced by the following
> safer change:
>
>     http://gcc.gnu.org/ml/gcc-patches/2008-04/msg00693.html

brgds, H-P