Scilab Website | Contribute with GitLab | Mailing list archives | ATOMS toolboxes
Scilab Online Help
6.1.1 - Русский

Change language to:
English - Français - 日本語 - Português -

Please note that the recommended version of Scilab is 2025.0.0. This page might be outdated.
See the recommended documentation of this function

Справка Scilab >> Dynamic/incremental Link > call

call

Fortran or C user routines call

Syntax

// long form 'out' is present
[y1,...,yk] = call("ident",x1,px1,"tx1",...,xn,pxn,"txn","out",[ny1,my1],py1,"ty1",...,[nyl,myl],pyl,"tyl")
// short form : no 'out' parameter
[y1,....,yk] = call("ident",x1,...,xn)

Arguments

"ident"

string.

xi

real matrix or string

pxi, pyi

integers

txi, tyi

character string "d", "r", "i" or "c".

Description

Interactive call of Fortran (or C) user program from Scilab. The routine must be previously linked with Scilab. This link may be done:

  • with Scilab "link" command (dynamical linking) during the Scilab session.(see link)

    On Windows, C functions must use cdecl calling convention name (see options in your C compiler(default calling convention for x86 C compilers)).

There are two forms of calling syntax, a short one and a long one. The short one will give faster code and an easier calling syntax but one has to write a small (C or Fortran) interface in order to make the short form possible. The long one make it possible to call a Fortran routine (or a C one) without modification of the code but the syntax is more complex and the interpreted code slower.

The meaning of each parameter is described now:

"ident"

is the name of the called subroutine.

x1,...,xn

are input variables (real matrices or strings) sent to the routine,

px1,...,pxn

are the respective positions of these variables in the calling sequence of the routine "ident" and

tx1,...,txn

are their types ("r", "i", "d" and "c" for real (float) , integer, double precision and strings)

"out"

is a keyword used to separate input variables from output variables. when this key word is present it is assumed that the long form will be used and when it is not present, the short form is used.

[ny1, my1]

are the size (# of rows and columns. For 'c' arguments,m1*n1 is the number of characters ) of output variables and

py1, ...

are the positions of output variables (possibly equal to pxi ) in the syntax of the routine. The pyi's integers must be in increasing order.

"ty1", ...

are the Fortran types of output variables. The k first output variables are put in y1,..., yk.

If an output variable coincides with an input variable (i.e. pyi=pxj ) one can pass only its position pyi . The size and type of yi are then the same as those of xi. If an output variable coincides with an input variable and one specify the dimensions of the output variable [myl,nyl] must follow the compatibility condition mxk*nxk >= myl*nyl.

Examples

//Example 1 with  a simple C code
            f1=['#include <math.h>'
            'void fooc(double c[],double a[],double *b,int *m,int *n)'
            '{'
            '   int i;'
            '   for ( i =0 ; i < (*m)*(*n) ; i++) '
            '     c[i] = sin(a[i]) + *b; '
            '}'];

            mputl(f1,'fooc.c')

            //creating the shared library (a gateway, a Makefile and a loader are
            //generated.

            ilib_for_link('fooc','fooc.c',[],"c")

            // load the shared library

            exec loader.sce

            //using the new primitive
            a=[1,2,3;4,5,6];b= %pi;
            [m,n]=size(a);

            // Inputs:
            // a is in position 2 and double
            // b                3     double
            // n                4     integer
            // m                5     integer
            // Outputs:
            // c is in position 1 and double with size [m,n]
            c=call("fooc",a,2,"d",b,3,"d",m,4,"i",n,5,"i","out",[m,n],1,"d");

            //Example 2 with  a simple Fortran code
            f1=['      subroutine foof(c,a,b,n,m)'
            '      integer n,m'
            '      double precision a(*),b,c(*)'
            '      do 10 i=1,m*n '
            '        c(i) = sin(a(i))+b'
            '   10 continue'
            '      end'];
            mputl(f1,'foof.f')

            //creating the shared library (a gateway, a Makefile and a loader are
            //generated.

            ilib_for_link('foof','foof.f',[],"f")

            // load the shared library

            exec loader.sce

            //using the new primitive
            a=[1,2,3;4,5,6];b= %pi;
            [m,n]=size(a);
            c=call("foof",a,2,"d",b,3,"d",m,4,"i",n,5,"i","out",[m,n],1,"d");

Limitations to Shared Library Support

Scilab shared library interface (call function) supports only C and fortran libraries routines.

In others cases, we recommend to use api_scilab interfaces.

Troubleshooting Shared Library Applications (Windows)

Some shared libraries, compiled as Microsoft Windows 32 or 64 bit libraries, use a calling convention that is incompatible with the default Scilab calling convention.

The default calling convention for Scilab and for Microsoft C and C++ compilers is cdecl.

For more information, see the MSDN® Calling Conventions and Wikipedia articles.

See also

  • link — dynamic linker
  • c_link — check if symbol loaded with dynamic link
  • addinter — new functions interface dynamic link at run time.
Report an issue
<< c_link Dynamic/incremental Link chooselcccompiler >>

Copyright (c) 2022-2024 (Dassault Systèmes)
Copyright (c) 2017-2022 (ESI Group)
Copyright (c) 2011-2017 (Scilab Enterprises)
Copyright (c) 1989-2012 (INRIA)
Copyright (c) 1989-2007 (ENPC)
with contributors
Last updated:
Mon Jan 03 14:40:12 CET 2022