Re: Re: Sparse matrix problem

[Dmitri A. Sergatskov]

On 1/4/07, John W. Eaton <address@hidden> wrote:
> | BTW, if I compile it with optimization on it would give (inf, inf).
>
> Then that definitely seems like a bug, since I don't think we should
> see a difference like that just because of compiler optimizations.

I do not dispute that this is a bug. Perhaps this difference
even amplifies the fact that this is an issue with the compiler --
result of a conformant program should not change depending on
optimization level. But perhaps this test
program is not the representative for the issues with octave, since
octave is compiled with optimization on...

> jwe

Sincerely,

Dmitri.