Problem with pointer, array and function parameter

[fl]

Hi,

I have c function {bitrev_cplx} with having function parameters float
array. In the function bitrev_cplx internal, some array members will
be swap value. I find that when array index i1=1, xi1 gets the right
value, but when j1=16, xj1 gets the wrong value (very large 1.32xx e+8
for example). I suspect the pointer, function parameter usage are
wrong, but I cannot find what is wrong. Could you help check my code?

Thanks a lot.


.....................
void bitrev_cplx (float *, int *, int );
int main()
{
float x[32*2];
float fw[16*2], fwc[16], fws[16];

for (i=0; i<32/2; i++){
fwc = cos(2*PI*i/32);
fws = sin(2*PI*i/32);
}

bitrev_cplx (fwc, temp, nx);
}


void bitrev_cplx (float *x, int *index, int nx)
{
int i;
int i0, i1, i2, i3;
int j0, j1, j2, j3;
float xi0, xi1, xi3;
float xj0, xj1, xj3;

if (t){
x[i0] = xj0;
x[j0] = xi0;
}
i1 = i0 + 1;
j1 = j0 + halfn;
xi1 = x[i1];
xj1 = x[j1];
}

 

[Ike Naar]

I have c function {bitrev_cplx} with having function parameters float
array. In the function bitrev_cplx internal, some array members will
be swap value. I find that when array index i1=1, xi1 gets the right
value, but when j1=16, xj1 gets the wrong value (very large 1.32xx e+8
for example). I suspect the pointer, function parameter usage are
wrong, but I cannot find what is wrong. Could you help check my code?

void bitrev_cplx (float *, int *, int );
int main()
{
float x[32*2];
float fw[16*2], fwc[16], fws[16];

for (i=0; i<32/2; i++){
fwc = cos(2*PI*i/32);
fws = sin(2*PI*i/32);
}

bitrev_cplx (fwc, temp, nx);
}


void bitrev_cplx (float *x, int *index, int nx)
{
int i;
int i0, i1, i2, i3;
int j0, j1, j2, j3;
float xi0, xi1, xi3;
float xj0, xj1, xj3;

if (t){
x[i0] = xj0;
x[j0] = xi0;
}
i1 = i0 + 1;
j1 = j0 + halfn;
xi1 = x[i1];
xj1 = x[j1];
}

Sorry, but this code is so incomplete that it is almost
impossible to say anything about it. Look at the
bitrev_cplx function, for instance. What is t ?
Why are i0 and j0 used uninitialized? What is halfn?

 

[James Kuyper]

Hi,

I have c function {bitrev_cplx} with having function parameters float
array. In the function bitrev_cplx internal, some array members will
be swap value. I find that when array index i1=1, xi1 gets the right
value, but when j1=16, xj1 gets the wrong value (very large 1.32xx e+8
for example). I suspect the pointer, function parameter usage are
wrong, but I cannot find what is wrong. Could you help check my code?

Thanks a lot.


....................
void bitrev_cplx (float *, int *, int );
int main()
{
float x[32*2];
float fw[16*2], fwc[16], fws[16];

x and fw are not used anywhere. Why are they here? If this was trimmed
down from a larger example, that's a good idea, but you should trim
everything that's not actually used, to avoid confusion.

for (i=0; i<32/2; i++){

No variable named 'i' has been declared in the code you've given us.
Please provide a complete, compilable program demonstrating your
program. It's nearly useless looking at code that isn't the actual code
that compiled and failed.

fwc = cos(2*PI*i/32);
fws = sin(2*PI*i/32);
}

bitrev_cplx (fwc, temp, nx);

Neither temp nor nx are declared anywhere.

}


void bitrev_cplx (float *x, int *index, int nx)

Neither index nor nx are used anywhere in this function, so I suppose
the cancels out the fact that neither temp nor nx are declared anywhere
that's visible in main(). If they're relevant, show us how they're used.
If not, remove them.

{
int i;
int i0, i1, i2, i3;
int j0, j1, j2, j3;

Variables named that way are almost always a bad idea. If i0 through i3
are all closely related to each other, you're probably better off
declaring a single array named 'i', and using i[1] instead of i1. It may
seem wasteful typing the extra brackets, but if the values are
sufficiently closely related, you'll often find contexts in which
accessing them as elements of an array simplifies your code.

On the other hand, if they're not sufficiently closely related to
justify making them array, they should be given distinct names, not
closely related names as you have done.

float xi0, xi1, xi3;
float xj0, xj1, xj3;

i2, j2, i3, j3, xi3, and xj3 are not used anywhere in this function.
Not knowning the purpose of your code, I can't be sure, but are xi2 and
xj2 missing?

if (t){

t is not declared anywhere.

x[i0] = xj0;

Neither i0 nor xj0 have been initialized, at best they have unspecified
values; at worst, they could be trap values, in which case the behavior
is undefined. If i0 has a value greater than 15, this will result in
trying to access a part of fwc that doesn't exist, rendering the
behavior of your code undefined.

x[j0] = xi0;

Neither j0 nor xi0 have been initialized.

}
i1 = i0 + 1;
j1 = j0 + halfn;

halfn is not declared anywhere. j0 is still uninitialized.

xi1 = x[i1];
xj1 = x[j1];

The values of i1 and j1 are, at best, unspecified at this point. If
either one has a value greater than 15, this code results in an attempt
to access a part of fwc that doesn't exist, rendering the behavior of
your code undefined.

I reiterate: give us a simple complete compilable example program that
demonstrates your problem, and it will be a lot easier for us to help
you. "simple" means, among other things, removing every feature not
needed to demonstrate the problem. "complete" means that it must define
or declare everything needed to demonstrate the problem. "compilable" is
something that you should confirm by compiling it yourself, with warning
levels on the compiler set as high as possible, before you post it here.
If it produces any warning messages, please give us the complete text of
those messages. Telling us precisely which compiler you're using, and
what options you selected for the compiler, would also help.

Finally, make sure that you present us with the complete actual code
that actually compiled into a program that demonstrates your problem.
Don't try to copy it by hand, any errors you make while copying it (and
you'll almost certainly make some) will only confuse the issue. Cut and
paste the actual code directly from your source code files.