help-gsl
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-gsl] GNU build configure.ac script


From: Rutger van Haasteren
Subject: Re: [Help-gsl] GNU build configure.ac script
Date: Thu, 16 Aug 2007 14:42:40 +0200

Thanks for you reply Brian.

Even with this command I still do not know why it was failing. I have found
a workaround for my case though, but the above mentioned case is still
valid. It is not that I have to check for a different function instead of
main, although I can do that.

My question boils down to this:
If I know the gslcblas library does exist on my system (it does), and I do
not check for it, I should be able to use:

AC_CHECK_LIB(m,main)
AC_SEARCH_LIBS(main, gsl, [], [], -lgslcblas)

instead of

AC_CHECK_LIB(m,main)
AC_SEARCH_LIBS(cblas_dgemm, gslcblas)
AC_SEARCH_LIBS(main, gsl, [], [], -lgslcblas)

The former doesn't work, whereas the latter does. I know the compile will
fail due to variables like DEPS_LIBS, but that is not the point. The point
is that the tests themselves fail. How can this be? Any ideas? TIA. Cheers,


Rutger


(please cc me, as I am not on the list)

On 8/15/07, Brian Gough <address@hidden> wrote:
>
> At Fri, 10 Aug 2007 08:52:38 +0200,
> Rutger van Haasteren wrote:
> > Now, I can also use the ATLAS version of CBLAS in step 2. But I fail to
> be
> > able to do this for the intel MKL library. What are all the requirements
> > that must be met for AC_CHECK_LIB(gsl,main) to succeed?
>
> Sometimes it's necessary to check for a different function instead of
> main, e.g. cblas_dgemm
>
> If you run a configure script in debugging mode it's possible to see
> why it is failing.
>
>   sh -x ./configure
>
> --
> Brian Gough
> (GSL Maintainer)
>
> Network Theory Ltd,
> Publishing the GSL Manual - http://www.network-theory.co.uk/gsl/manual/
>


reply via email to

[Prev in Thread] Current Thread [Next in Thread]