Re: regression failure on Solaris contrib/cube

Andrew Dunstan <andrew@dunslane.net> writes:
> Machine details:
> Solaris / 2.8

> regression diffs:

> *** ./expected/cube.out    Tue Sep 28 15:25:33 2004
> --- ./results/cube.out    Tue Sep 28 15:55:16 2004
> ***************
> *** 142,148 ****
>   SELECT '-1e-700'::cube AS cube;
>    cube 
>   ------
> !  (0)
>   (1 row) 
>   SELECT '1234567890123456'::cube AS cube;
> --- 142,148 ----
>   SELECT '-1e-700'::cube AS cube;
>    cube 
>   ------
> !  (-0)
>   (1 row) 
>   SELECT '1234567890123456'::cube AS cube;

Yeah, I've seen the same on Mac OS X.  I've committed a variant expected
file to make the contrib tests pass cleanly on these machines.
        regards, tom lane