[basic-improvements] bootstrap failure on x86

[Michael Matz]

Hi,

On Thu, 5 Dec 2002, Jan Hubicka wrote:

> > -   && !x86_64_immediate_operand (operands[1], DImode)"
> > +   && !x86_64_immediate_operand (operands[1], DImode) && 1"
>
> Hmm, now I remember the reason for && 1.  The patterns looks same but
> when one fails the other is used.  OK I will re-add the && 1 as obvious
> fix.

And add a comment for that effect.  Otherwise later people will wonder
again about the useless "&& 1" and remove it.  It's highly magic ;-)


Ciao,
Michael.