[Scilab-Dev] Primitives id limit on gateways

Sylvestre Ledru sylvestre.ledru at scilab.org
Mon Jul 13 17:41:34 CEST 2009


Le lundi 13 juillet 2009 à 14:19 +0200, Calixte Denizet a écrit :
> Hi,
> 
> I met this problem and I think it is due to the internal representation
> of functions coming from library, see the doc about funptr in Scilab.
> The number fin (funptr('your fun') modulo 100) is the entry in the array
> created in the file libtoolbox.c.
If it is the actual reason, it is a definitely a bug! Could you report
it ? (with the workaround... to help others).
Thanks!

Cheers,
Sylvestre
 
> You can pass this limitation in creating a function gateway(int n,
> list_args){(*(array[n]))(list_args);}, an array containing pointers of
> all your functions and generates Scilab's macros of the kind 
> function y=fun1(varargin), gateway(1,varargin), endfunction 
> 
> Calixte
> 
> Le lundi 13 juillet 2009 à 10:50 +0200, sylvestre.ledru at scilab.org a
> écrit :
> > On Mon, 13 Jul 2009 02:18:21 -0500, Jorge Cardona <jorgeecardona at gmail.com>
> > wrote:
> > > Hi,
> > > 
> > > I just get to more that 100 functions on the symbolic module and its
> > > appear that i can't define more than 100 functions, if i call one that
> > > has a primitiveId bigger than 100 scilab just crash, i even
> > > interchange a pair, and with a lower id it work fine the function, and
> > > the other one start to crash scilab, do i need to define a new
> > > gatewayid and start again from 1???
> > Hello Jorge,
> > 
> > I don't think it is the reason.
> > Try to increase
> >  #define MAXELEMENTFUNCTIONLIST 1536
> > in
> > modules/core/src/c/hashtable_core.h
> > This is something we have to work on.
> > 
> > Sylvestre
> > 
> > 
> > > Thanks.
> > > 
> > > --
> > > Jorge Eduardo Cardona
> > > jorgeecardona at gmail.com
> > > jorgeecardona.blogspot.com
> > > ------------------------------------------------
> > > Linux registered user  #391186
> > > Registered machine    #291871
> > > ------------------------------------------------
> > > 
> 




More information about the dev mailing list