[Scilab-Dev] Oh scilab!what a beaut you are you are if only you'd build without the war

[Sylvestre Ledru]

Yep, this is an expected behavior.
The configure is just trying to link against the symbol of the library
(with the C library, it has no way to check the consistency of the input
argument of the C function at link time).
It is not called.

Sylvestre


Le samedi 06 novembre 2010 à 18:25 +0100, Calixte Denizet a écrit :
> Ceci dit le amd_info prend un double* en param et qd tu appelles
> amd_info() ca fait une erreur de segmentation (ca va prendre n'importe
> quoi dans la pile)...
> 
> Le samedi 06 novembre 2010 à 19:18 +0100, Sylvestre Ledru a écrit :
> > Le jeudi 04 novembre 2010 à 15:10 +0000, bv a écrit :
> > > On Monday 01 November 2010 15:23:12 bv wrote:
> > > > On Monday 01 November 2010 12:42:42 Calixte Denizet wrote:
> > > > > Hello,
> > > > > 
> > > > > what's the output of :
> > > > > # ldconfig -v | grep libamd
> > > > > 
> > > > > and the output of
> > > > > $ readelf -a /path/to/lib/libamd.so | grep amd_info
> > > > 
> > > > root [ ~ ]# ldconfig -v | grep libamd
> > > >         libamd.so.2.2.1 -> libamd.so.2.2.1
> > > > root [ ~ ]#
> > > > 
> > > > 
> > > > root [ ~ ]# readelf -a /usr/local/numerics/lib/libamd.so | grep amd_info
> > > >     23: 0000000000003800  1493 FUNC    GLOBAL DEFAULT   10 amd_info
> > > >     48: 0000000000000000     0 FILE    LOCAL  DEFAULT  ABS amd_info.c
> > > >     60: 0000000000000000     0 FILE    LOCAL  DEFAULT  ABS amd_info.c
> > > >     88: 0000000000003800  1493 FUNC    GLOBAL DEFAULT   10 amd_info
> > > > root [ ~ ]#
> > > 
> > > 
> > > I supplied the above answers on Monday  and I would be grateful to know if 
> > > there is a problem with them or otherwise.
> > --without-umfpack will be way easier here...
> > 
> > Sylvestre
> > 
> 
> 

-- 
-------------------------
Sylvestre Ledru
-------------------------
The Scilab Consortium
Digiteo
Domaine de Voluceau
Rocquencourt - B.P. 105
78153 Le Chesnay Cedex
France