I want to convert the second element of a list of atoms, to a float
and it’s driving me crazy. I can only get the first one to work,
everything else i try gives me the address of the pointer instead.

Please help

My list is Atom *av

float target = 0;
target = av->a_w.w.float;
outlet_float(x->t_floatout, target);

gives me the first member of the list. How to get the second one ?

f.e

float target;
target = av[1].a_w.w_float;
outlet_float(x->t_floatout, target);

On Nov 16, 2007 7:34 AM, f.e wrote:
> Dear list,
>
> I want to convert the second element of a list of atoms, to a float
> and it’s driving me crazy. I can only get the first one to work,
> everything else i try gives me the address of the pointer instead.
>
> Please help
>
> My list is Atom *av
>
> float target = 0;
> target = av->a_w.w.float;
> outlet_float(x->t_floatout, target);
>
> gives me the first member of the list. How to get the second one ?
>
> f.e
>

f.e

On Nov 16, 2007 1:48 PM, P C wrote:
> Try this:
>
> float target;
> target = av[1].a_w.w_float;
> outlet_float(x->t_floatout, target);
>
>
>
>
> On Nov 16, 2007 7:34 AM, f.e wrote:
> > Dear list,
> >
> > I want to convert the second element of a list of atoms, to a float
> > and it’s driving me crazy. I can only get the first one to work,
> > everything else i try gives me the address of the pointer instead.
> >
> > Please help
> >
> > My list is Atom *av
> >
> > float target = 0;
> > target = av->a_w.w.float;
> > outlet_float(x->t_floatout, target);
> >
> > gives me the first member of the list. How to get the second one ?
> >
> > f.e
> >
>

double target = av[1].a_w.w_float;

On Nov 16, 2007 7:55 AM, f.e wrote:
> It’s the first thing i’ve done but it doesn’t work. It should but it
> doesn’t. It gives me something like-15048510125 instead of 1.0
>
> f.e
>
>
>
> On Nov 16, 2007 1:48 PM, P C wrote:
> > Try this:
> >
> > float target;
> > target = av[1].a_w.w_float;
> > outlet_float(x->t_floatout, target);
> >
> >
> >
> >
> > On Nov 16, 2007 7:34 AM, f.e wrote:
> > > Dear list,
> > >
> > > I want to convert the second element of a list of atoms, to a float
> > > and it’s driving me crazy. I can only get the first one to work,
> > > everything else i try gives me the address of the pointer instead.
> > >
> > > Please help
> > >
> > > My list is Atom *av
> > >
> > > float target = 0;
> > > target = av->a_w.w.float;
> > > outlet_float(x->t_floatout, target);
> > >
> > > gives me the first member of the list. How to get the second one ?
> > >
> > > f.e
> > >
> >
>

> Replace float by double.
> as this:
>
> double target = av[1].a_w.w_float;

Something like that should work, unless I miss typed something.

void myobject_toto(t_myobject *x, t_symbol *s, long argc, t_atom *argv)
{
double f;

if (argv && argv) {
f = atom_getfloat(&argv[0]);
outlet(x->floatout, f);
}
}

ej

Something else : this few lines of codes were placed AFTER a for to
loop (reading all the atoms), and, for an unknown reason, av[1] was
giving me av[7] instead. I move these lines BEFORE the for to loop,
and it works again… Weird, isn’t it ?

Thanks a lot, anyway

f.e

On Nov 16, 2007 2:32 PM, Emmanuel Jourdan wrote:
> On 16 nov. 07, at 14:18, P C wrote:
>
> > Replace float by double.
> > as this:
> >
> > double target = av[1].a_w.w_float;
>
> Something like that should work, unless I miss typed something.
>
> void myobject_toto(t_myobject *x, t_symbol *s, long argc, t_atom *argv)
> {
> double f;
>
> if (argv && argv) {
> f = atom_getfloat(&argv[0]);
> outlet(x->floatout, f);
> }
> }
>
>
> ej
>
>

> Incredible. It works with double !
>
> Something else : this few lines of codes were placed AFTER a for to
> loop (reading all the atoms), and, for an unknown reason, av[1] was
> giving me av[7] instead. I move these lines BEFORE the for to loop,
> and it works again… Weird, isn’t it ?

Could you post this extract of the code?

ej

About your loop: if it’s incrementing the base address of av, well av[1] is going to mean something different after the loop than it did before. Incrementing the base address of an array is a vile and error-prone programming habit, but unfortunately most C programmers are addicted to it.

// The following is code to crash & die for
void foo(Object *x, short ac, Atom *av)
{
long x, y;

// av points to a vector of Atoms
x = av[0].a_w.w_long;

while (i– > 0) {
av++;
}
// If (i>0) av now points to a *different* address
y = av[0].a_w.w_long;

ASSERT (probably_not_equal(x, y));
}

> Correct me if i’m wrong, this is not what i do there ? (because
> here it’s called l_templistout) :
>
> typedef struct _vcparse
> {
> t_object m_ob;
> void *t_intout;
> void *t_listout;
> t_atom l_templistout[8];
> } t_vcparse;

Yes, here you are allocating an array of 8 t_atoms.

> And, like a fool, i declare
>
> t_atom *templistout;

Here, you are declaring a (global) t_atom * (4 bytes on a 32-bit OS),
which you can later use to store the value of an allocated pointer.
For instance:

templistout = (t_atom *)sysmem_newptr(sizeof(t_atom) * 8);

But until you’ve assigned templistout to a block of allocated memory,
you should’t try to read or write from it (as it isn’t pointing to
anything valid).

I wouldn’t use a global variable for this, anyway, but I think you
understand that.

Changing your code (just email typing — not verified that it works):

void vcparse_list(t_vcparse *x, Symbol *s, int ac, Atom *av)
{
int i;
int j = 3;
short target = 0;
t_atom templistout[8]; // in this case is the easier solution
// OR
// t_atom *templistout = (t_atom *)sysmem_newptr(sizeof(t_atom)
* 8); // better for dynamic sized templist — don’t forget to free
with sysmem_freeptr();

if ((ac==34) & (av[1].a_w.w_float!=0)) {
for (i=0;i<8;i++)
{
SETFLOAT(&templistout[i], av[j].a_w.w_float);
j = j+4;
}

target = (short) av[1].a_w.w_float;
outlet_int(x->t_intout, target);
outlet_list(x->t_listout, 0L, 8, templistout);
}
else if ((ac==34) & (av[1].a_w.w_float==0))
{
post(“[sd.vc.parse] ERROR ! Packet dropped because target 0
received”);
}
else if (ac==6)
{
post(“[sd.vc.parse] ERROR ! Packet dropped because source
coordinates received”);
}
else post(“[sd.vc.parse] ERROR ! Packet dropped because list too
short (%d)”, ac);
}

However, to pick up on your question about my example (which is quite likely to happen to a beginner copying and modifying existing C code), there are several ways to avoid the problem I focussed on. One would be as follows:

========================

void foo(Object *x, short ac, Atom *av)
{
long x, y, j;

// av points to a vector of Atoms
x = av[0].a_w.w_long;

j = 0;
while (j < ac) { // this was a typo in prev. example
long z = av[j++].a_w.w_long;
// Real code would do something with the data here
}

// av hasn’t changed, so the following works as expected
y = av[0].a_w.w_long;

ASSERT (absolutely_definitely_equal(x, y));
}

———————————-

As a matter of principal I always declare my argc/argv functions something like the following:

void foo(Object* x, short ac, Atom av[])

Declaring av as an array of Atoms (rather than a pointer to an Atom) serves as a reminder to me about what’s being passed. And I never move the base address. That comes from training in Pascal. C programmers liked to treat arrays as movable pointers because they could get marginally better performance that way with early C compilers. Modern optimizing C compilers will, as a rule, generate equally efficient machine code with either this example or the code I sent yesterday.

Hope this little tangent is helpful.

– P.

> In the case i use :
>
> t_atom *templistout = (t_atom *)sysmem_newptr(sizeof(t_atom) * 8);
>
> Only things i have to do is :
>
> sysmem_freeptr(templistout);

Exactly.

ej