1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
|
.TH fort 1 "April 1993" "Scilab Group" "Scilab Function"
.so ../sci.an
.SH NAME
fort - Fortran or C user routines call
.SH CALLING SEQUENCE
.nf
// long form 'out' is present
[y1,...,yk]=fort("ident",x1,px1,"tx1",...,xn,pxn,"txn",
"out",[ny1,my1],py1,"ty1",...,[nyl,myl],pyl,"tyl")
// short form : no 'out' parameter
[y1,....,yk]=fort("ident",x1,...,xn)
.fi
.SH PARAMETERS
.TP 10
"ident"
: string.
.TP
xi
: real matrix or string
.TP
pxi, pyi
: integers
.TP
txi, tyi
: character string \fV"d"\fR, \fV"r"\fR, \fV"i"\fR or \fV"c"\fR.
.SH DESCRIPTION
Interactive call of Fortran (or C) user program from Scilab. The
routine must be previously linked with Scilab. This link may be done:
.RS
.TP
-
with Scilab "\fVlink\fR" command (incremental "soft" linking) during the
Scilab session.(see \fVlink\fR)
.TP
-
by "hard" re-linking. Writing the routine call within Scilab routine
\fVdefault/Ex-fort.f\fR, adding the entry point in the file \fVdefault/Flist\fR
and then re_linking Scilab with the command \fVmake bin/scilex\fR
in main Scilab directory.
.RE
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) whitout modification of the code but the syntax is more
complex and the interpreted code slower.
.LP
The meaning of each parameter is described now:
.RS
.TP 15
"ident"
is the name of the called subroutine.
.TP
x1,...,xn
are input variables (real matrices or strings) sent to the routine,
.TP
px1,...,pxn
are the respective positions of these variables in the calling
sequence of the routine \fV"ident"\fR and
.TP
tx1,...,txn
are their types (\fV"r"\fR, \fV"i"\fR, \fV"d"\fR and \fV"c"\fR
for real (float) , integer, double precision and strings)
.TP
"out"
is a keyword used to separate input variables from output
variables. when this key word is present it is assumes that
the long form will be used and when it is not prsent,
the short form is used.
.TP
[ny1, my1]
are the size (# of rows and columns. For 'c' arguments,\fVm1*n1\fR is the number of charaters )
of output variables and
.TP
py1, ...
are the positions of output variables (possibly equal to \fVpxi\fR )
in the calling sequence of the routine.
The \fVpyi\fR's integers must be in increasing order.
.TP
"ty1", ...
are the Fortran types of output variables.
The \fVk\fR first output variables are put in \fVy1,..., yk\fR.
.RE
.LP
If an output variable coincides with an input variable
(i.e. \fVpyi=pxj\fR ) one can pass only its position \fVpyi\fR .
The size and type of \fVyi\fR are then the same as those of \fVxi\fR.
If an output variable coincides with an input variable and one specify
the dimensions of the output variable \fV[myl,nyl]\fR must follow the
compatibility condition \fVmxk*nxk >= myl*nyl\fR.
.LP
In the case of short syntax , \fV[y1,....,yk]=fort("ident",x1,...,xn)\fR,
the input parameters \fVxi\fR's and the name \fV"ident"\fR are sent to
the interface routine \fVEx-fort\fR. This interface routine is then
very similar to an interface (see the source code in the directory \fVSCIDIR/default/Ex-fort.f\fR).
.LP
For example the following program:
.nf
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
link("foof.o","foof")
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=fort("foof",a,2,"d",b,3,"d",n,4,"i",m,5,"i","out",[m,n],1,"d");
.fi
returns the matrix \fVc=2*a+b\fR.
If your machine is a DEC Alpha, SUN Solaris or SGI you may have to change the
previous command line \fVlink("foo.o","foo") \fR by one of the followings:
.nf
link('foof.o -lfor -lm -lc','foof').
link('foof.o -lftn -lm -lc','foof').
link('foof.o -L/opt/SUNWspro/SC3.0/lib/lib77 -lm -lc','foof').
.fi
The same example coded in C:
.nf
void fooc(c,a,b,m,n)
double a[],*b,c[];
int *m,*n;
{ double sin();
int i;
for ( i =0 ; i < (*m)*(*n) ; i++)
c[i] = sin(a[i]) + *b;
}
link("fooc.o","fooc","C") // note the third argument
a=[1,2,3;4,5,6];b= %pi;
[m,n]=size(a);
c=fort("fooc",a,2,"d",b,3,"d",m,4,"i",n,5,"i","out",[m,n],1,"d");
.fi
.SH SEE ALSO
link, c_link, intersci, addinter
|
