question

question

[Klauko Mota]

Can I use the gsl libraries with fortran90? What should I do to compiler
the code?

		******************************************
		* Klauko Mota                            *
		* Instituto de Física                    *
		* Campus da Praia Vermelha               *
		* Avenida Litorânea, s/n, Boa Viagem     *
		* Cep.: 24210-340, Niteroi - RJ - Brazil *
		* Voice: +55-21-2620-6735                *
		* Fax: +55-21-2620-3881                  *
		* e-mail: klauko@if.uff.br               *
		******************************************

Re: question

[Brian Gough]

> From: Klauko Mota <klauko@if.uff.br>
> Can I use the gsl libraries with fortran90?

Perhaps, but it's not recommended.

Re: Question

[Brian Gough]

Gerard Jungman writes:
 > Unfortunately, some compilers are truly brain-damaged and
 > will claim that the prototypes do not match if you remove
 > that 'const' from the declaration. As I recall, there was
 > a complaint to this effect several years ago; I think it
 > was an HP compiler.

That's right..  I think their man page claimed that functions with
const arguments could be compiled with a different calling convention,
and so the prototypes had to match.

-- 
Brian Gough

Re: Question

[Jordi Burguet Castell]

On Tuesday 01 February 2005 22:28, John Lamb wrote:
> Przemyslaw Sliwa wrote:
> [...]
> Nonetheless, it is often a good idea to declare an automatic function
> parameter const in the function definition. This forces the compiler to
> check that you don't change the parameter value in the body of the
> function. Similarly, it is a good idea to declare any variable const if
> you don't intend to change it.

IMHO, you lose more with the potential confusion for the programmer than what 
you may earn.

jordi

Re: Question

[Gerard Jungman]

On Tue, 2005-02-01 at 21:28 +0000, John Lamb wrote:
> Przemyslaw Sliwa wrote:

> > I was just wondering what means
> > 
> > double gsl_cdf_ugaussian_Pinv (const double P)


> I think you are right, const is not required and you don't need to worry 
> about it when calling the function. This is certainly true in C++.

Unfortunately, some compilers are truly brain-damaged and
will claim that the prototypes do not match if you remove
that 'const' from the declaration. As I recall, there was
a complaint to this effect several years ago; I think it
was an HP compiler.

Anyway, it may be unnecessary now. It certainly is silly.

-- 
Gerard Jungman <jungman@lanl.gov>
Los Alamos National Laboratory

Re: Question

[John Lamb]

Przemyslaw Sliwa wrote:
> All,
> 
> I was just wondering what means
> 
> double gsl_cdf_ugaussian_Pinv (const double P)
> 
> Correct me if I am wrong but it seems the const word is not required
> in this case and we cannot change the value of P.
> 

I think you are right, const is not required and you don't need to worry 
about it when calling the function. This is certainly true in C++.

Nonetheless, it is often a good idea to declare an automatic function 
parameter const in the function definition. This forces the compiler to 
check that you don't change the parameter value in the body of the 
function. Similarly, it is a good idea to declare any variable const if 
you don't intend to change it.

You can declare the function as

double gsl_cdf_ugaussian_Pinv (double);

and define it as

double gsl_cdf_ugaussian_Pinv (const double P){
...
}

but it may be better to make the two match precisely either because you 
may want to use some editing tools that rely on a match or because you 
want the code to work on a compiler that can't work oout that the 
declaration and definition do match.

-- 
JDL

Question

[Przemyslaw Sliwa]

All,

I was just wondering what means

double
gsl_cdf_ugaussian_Pinv (const double P)

Correct me if I am wrong but it seems the const word is not required in this case and we cannot change the value of P.

Thanks for help,

Pshemek


--

This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.