Re: [Bug-gsl] about gsl_complex_poly_complex_eval

[Liam Healy]

A good data point; I had a colleague try this test on a Mac with gcc
4.0.x, and there was no problem getting through the tests.  So it's
clear the problem was introduced after gcc 4.1.

Meanwhile, I have tried recompiling on the same computer that
originally gave me the problem with version
gcc (Debian 4.3.1-9) 4.3.1
but this time with -O0:

make clean
configure CFLAGS=-O0
make
make check

and I get through all the tests with no problems.  So it looks like
there's a problem with the optimizer in gcc 4.3.  Since this seems to
be a gcc problem, I've started going through the gcc bug reporting
procedure http://gcc.gnu.org/bugs.html.  So far I've recompiled with
-Wall to make sure there aren't any coding problems in the gsl
library.  There aren't.   The next step is to come up with a
simplified example.

Liam

On Sun, Sep 21, 2008 at 1:54 PM, L. Kliemann
<address@hidden> wrote:
> I saw this problem (see below) being discussed here already, so
> it's not new. Maybe I can provide some additional information,
> however: it happens for me when compiling with gcc 4.3.2, but not
> with gcc 4.1.2.
>
> FAIL: y.real, gsl_complex_poly_complex_eval ({-2.31 + 0.44i, 4.21 - 3.19i, 
> 0.93 + 1.04i, -0.42 + 0.68i}, 0.49 + 0.95i) (-1.68450788000000018 observed vs 
> 1.82462012000000007 expected) [9]
> FAIL: y.imag, gsl_complex_poly_complex_eval ({-2.31 + 0.44i, 4.21 - 3.19i, 
> 0.93 + 1.04i, -0.42 + 0.68i}, 0.49 + 0.95i) (-0.30943988 observed vs 
> 2.30389411999999982 expected) [10]
> FAIL: test
>