leighhunt

I've recently started looking into external object development. Looking through the pattr sdk pdf certainly made my head spin a little (I've no experience in C programming btw), but, working through the myobex example, I have got as far as successfully compiling a pattrified external (a hackedup myobex) with 6 attributes which installs fine and receives/sends attributes correctly. The attibutes are;

2 - @minval - to be a minumum argument to the value

3 - @maxval - to be the maximum argument to the value

4 - @encadj - to be an inc/dec to the value

5 - @modadj - to be a secondary inc/dec to the value. post encadj

6 - @state1 - to be a switch to enable/disable loopback of encadj

The processing I have already programmed in both max objects and .js. For my particular needs, moving this code into an external will most certainly remove 500+ pattr objects from my current patch.

Please excuse my ignorance if these questions are basic and obvious

1 - Is it possible, when receiving an attribute value from a bound pattr, to create a trigger to process/output changes to another attribute? 

eg - @encadj receives a 1, initiates a process to increment @floatval by 1 and output new value to secondary inc/dec process...

2 - If so, can the repeated receiving of the SAME attribute value from a bound pattr do the same, or must the value change?

Any pointers to these would be brill!!lol

3 - I shall refrain from asking 1037 questions regarding programming in C for now (I feel that's down to me to learn!), but if anyone can point me in the direction of the meaning of the syntax "->", mostly preceded with an x, it would stop me going completely crazy!

Adding to that, if anyone could give me any hints to C resources that might help my learning I would be more than grateful!

Macbookpro2.33,OSleopard.latest,maxmsp4.6.latest,xcode3.latest

Greetings All...
I've recently started looking into external object development. Looking through the pattr sdk pdf certainly made my head spin a little (I've no experience in C programming btw), but, working through the myobex example, I have got as far as successfully compiling a pattrified external (a hackedup myobex) with 6 attributes which installs fine and receives/sends attributes correctly. The attibutes are;
1 - @floatval - value of object
2 - @minval - to be a minumum argument to the value
3 - @maxval - to be the maximum argument to the value
4 - @encadj - to be an inc/dec to the value
5 - @modadj - to be a secondary inc/dec to the value. post encadj
6 - @state1 - to be a switch to enable/disable loopback of encadj

Please excuse my ignorance if these questions are basic and obvious
1 - Is it possible, when receiving an attribute value from a bound pattr, to create a trigger to process/output changes to another attribute? 
eg - @encadj receives a 1, initiates a process to increment @floatval by 1 and output new value to secondary inc/dec process...

System
Macbookpro2.33,OSleopard.latest,maxmsp4.6.latest,xcode3.latest


some-basic-questions-regarding-attributes

Quote: leighhunt wrote on Mon, 07 January 2008 14:12

----------------------------------------------------

> 3 - I shall refrain from asking 1037 questions regarding programming in C for now (I feel that's down to me to learn!), but if anyone can point me in the direction of the meaning of the syntax "->", mostly preceded with an x, it would stop me going completely crazy!

I have no ideas on the pattr stuff a it's not something I've done.

However, I can indeed point you in the direction of the -> operator:

This operator accesses a member of a struct from a pointer to the struct. So in practice:

accesses member bar from the struct which foo points to. For the code you're looking at x is almost certainly a pointer to struct representing an instance of a max-msp object and the operator is being used to read and write values to certain members of that instance of the object.

Not sure how clear this will be - but hopefully of some help - a good understanding of pointers and structures would be helpful - so I would read up on that if needs be. There are several good c tutorials and references online that you can find by googling....

Quote: leighhunt wrote on Mon, 07 January 2008 14:12
----------------------------------------------------

accesses member bar from the struct which foo points to. For the code you're looking at x is almost certainly a pointer to struct representing an instance of a max-msp object and the operator is  being used to read and write values to certain members of that instance of the object.

Quote: leighhunt wrote on Mon, 07 January 2008 22:12

> I've recently started looking into external object development. Looking through the pattr sdk pdf certainly made my head spin a little (I've no experience in C programming btw), but, working through the myobex example, I have got as far as successfully compiling a pattrified external (a hackedup myobex) with 6 attributes which installs fine and receives/sends attributes correctly. The attibutes are;

> 2 - @minval - to be a minumum argument to the value

> 3 - @maxval - to be the maximum argument to the value

> 4 - @encadj - to be an inc/dec to the value

> 5 - @modadj - to be a secondary inc/dec to the value. post encadj

> 6 - @state1 - to be a switch to enable/disable loopback of encadj

> The processing I have already programmed in both max objects and .js. For my particular needs, moving this code into an external will most certainly remove 500+ pattr objects from my current patch.

> Please excuse my ignorance if these questions are basic and obvious

> 1 - Is it possible, when receiving an attribute value from a bound pattr, to create a trigger to process/output changes to another attribute? 

> eg - @encadj receives a 1, initiates a process to increment @floatval by 1 and output new value to secondary inc/dec process...

Sure, you could use a custom setter function:

t_max_err myobject_set_encadj(t_myobject *x, void *attr, long ac, t_atom *av)

float f = object_attr_getfloat(x, gensym("floatval"));

x->encadj = 1; // maybe you don't even need this? do you really need to store the increment?

object_attr_setfloat(x, gensym("floatval"), f + enc);

Presumably, you have a custom setter for floatval as well, which takes your min and max into account.

> 2 - If so, can the repeated receiving of the SAME attribute value from a bound pattr do the same, or must the value change?

Can you rephrase the question? I don't quite understand what you're getting at.

Quote: leighhunt wrote on Mon, 07 January 2008 22:12
----------------------------------------------------
> Greetings All...
> I've recently started looking into external object development. Looking through the pattr sdk pdf certainly made my head spin a little (I've no experience in C programming btw), but, working through the myobex example, I have got as far as successfully compiling a pattrified external (a hackedup myobex) with 6 attributes which installs fine and receives/sends attributes correctly. The attibutes are;
> 1 - @floatval - value of object
> 2 - @minval - to be a minumum argument to the value
> 3 - @maxval - to be the maximum argument to the value
> 4 - @encadj - to be an inc/dec to the value
> 5 - @modadj - to be a secondary inc/dec to the value. post encadj
> 6 - @state1 - to be a switch to enable/disable loopback of encadj
>
> The processing I have already programmed in both max objects and .js. For my particular needs, moving this code into an external will most certainly remove 500+ pattr objects from my current patch.
>
> Please excuse my ignorance if these questions are basic and obvious
> 1 - Is it possible, when receiving an attribute value from a bound pattr, to create a trigger to process/output changes to another attribute? 
> eg - @encadj receives a 1, initiates a process to increment @floatval by 1 and output new value to secondary inc/dec process...

t_max_err myobject_set_encadj(t_myobject *x, void *attr, long ac, t_atom *av)
{
  if (ac && av) {
    long enc = atom_getlong(av);
    float f = object_attr_getfloat(x, gensym("floatval"));

    x->encadj = 1; // maybe you don't even need this? do you really need to store the increment?
    object_attr_setfloat(x, gensym("floatval"), f + enc);
  }
}

Quote: AlexHarker wrote on Tue, 08 January 2008 11:29

> However, I can indeed point you in the direction of the -> operator:

> This operator accesses a member of a struct from a pointer to the struct. So in practice:

> accesses member bar from the struct which foo points to. For the code you're looking at x is almost certainly a pointer to struct representing an instance of a max-msp object and the operator is being used to read and write values to certain members of that instance of the object.

> Not sure how clear this will be - but hopefully of some help - a good understanding of pointers and structures would be helpful - so I would read up on that if needs be. There are several good c tutorials and references online that you can find by googling....

I certainly am checking out some online reference right now. The -> is so obvious when you know eh?!

> However, I can indeed point you in the direction of the -> operator:
>
> This operator accesses a member of a struct from a pointer to the struct. So in practice:
>
> given a pointer foo with a member bar
>
> foo->bar
>
> accesses member bar from the struct which foo points to. For the code you're looking at x is almost certainly a pointer to struct representing an instance of a max-msp object and the operator is  being used to read and write values to certain members of that instance of the object.
>
> Not sure how clear this will be - but hopefully of some help - a good understanding of pointers and structures would be helpful - so I would read up on that if needs be. There are several good c tutorials and references online that you can find by googling....
>
> Regards,
>
> Alex
----------------------------------------------------

Quote: Jeremy Bernstein wrote on Tue, 08 January 2008 11:41

> Sure, you could use a custom setter function:

> t_max_err myobject_set_encadj(t_myobject *x, void *attr, long ac, t_atom *av)

> float f = object_attr_getfloat(x, gensym("floatval"));

> x->encadj = 1; // maybe you don't even need this? do you really need to store the increment?

> object_attr_setfloat(x, gensym("floatval"), f + enc);

aha - Thanks Jeremy. That small piece of code just switched on a great big light bulb here!

> Presumably, you have a custom setter for floatval as well, which takes your min and max into account.

> > 2 - If so, can the repeated receiving of the SAME attribute value from a bound pattr do the same, or must the value change?

> Can you rephrase the question? I don't quite understand what you're getting at.

I think question 2 is now defunct, I shall return to it should I get stuck, but onwards with xcode for now! 

Quote: Jeremy Bernstein wrote on Tue, 08 January 2008 11:41
>
> Sure, you could use a custom setter function:
>
> t_max_err myobject_set_encadj(t_myobject *x, void *attr, long ac, t_atom *av)
> {
>   if (ac && av) {
>     long enc = atom_getlong(av);
>     float f = object_attr_getfloat(x, gensym("floatval"));
>
>     x->encadj = 1; // maybe you don't even need this? do you really need to store the increment?
>     object_attr_setfloat(x, gensym("floatval"), f + enc);
>   }
> }

> Presumably, you have a custom setter for floatval as well, which takes your min and max into account.
>
> > 2 - If so, can the repeated receiving of the SAME attribute value from a bound pattr do the same, or must the value change?
>
> Can you rephrase the question? I don't quite understand what you're getting at.
>
----------------------------------------------------
I think question 2 is now defunct, I shall return to it should I get stuck, but onwards with xcode for now! 
Thank you very much.

This is asking very much I know....but i am having a problem implementing your suggestion Jeremy.

If you could take a moment in your very busy schedule to have a quick peek at this......

This is the reworked version of the myobex example.

Please be aware I have never looked at C until 3 days ago, so my changes/additions might be very odd/incorrect....but the object does seem to function without error when instantiated in Max.

In Max, I can set all the attributes either by 

1 - a message 'attribute_name value' at the object's inlet.

2 - remotely - via a pattr object, bound to the relevent attribute.

3 - remotely - via a pattrforward object, when sent a 'attribute_name value' message to it's inlet.

The remote, "attribute bound" pattr object also correctly updates it's value correctly in all cases.

The code you gave me makes total sense, especially when comparing against the other custom setters, and compiles correctly, but the output of floatval doesn't update/change. I guessed I might still have to bang the object's input to output the changed floatval, but that doesn't seem to do anything either.

(and also to anyone else who might be able to chime in on this)

    jeremy bernstein - jeremy@bootsquad.com    

    t_object                    ob;

    void                        *obex;

    double                        floatval;

    t_atom                        atomarray[32];

    long                        atomarraycount;

    char                        mode;

    void                         *out;

    double                        minval;

    double                        maxval;

    double                        encadj;

    double                        modadj;

    long                        state1;

void *lhpattr_new(t_symbol *s, long argc, t_atom *argv);

void lhpattr_assist(t_lhpattr *x, void *b, long m, long a, char *s);

void lhpattr_anything(t_lhpattr *x, t_symbol *s, long argc, t_atom *argv);

void lhpattr_float(t_lhpattr *x, double f);

void lhpattr_min(t_lhpattr *x, double f);

void lhpattr_max(t_lhpattr *x, double f);

void lhpattr_enc(t_lhpattr *x, double f);

void lhpattr_mod(t_lhpattr *x, double f);

void lhpattr_addition1(t_lhpattr *x, long m, double f);

// attribute method prototypes, for attributes which override their defaults

t_max_err lhpattr_floatval(t_lhpattr *x, void *attr, long *ac, t_atom **av);

t_max_err lhpattr_minval(t_lhpattr *x, void *attr, long *ac, t_atom **av);

t_max_err lhpattr_maxval(t_lhpattr *x, void *attr, long *ac, t_atom **av);

t_max_err lhpattr_encadj(t_lhpattr *x, void *attr, long ac, t_atom *av);

t_max_err lhpattr_modadj(t_lhpattr *x, void *attr, long *ac, t_atom **av);

t_max_err lhpattr_state1(t_lhpattr *x, void *attr, long *ac, t_atom **av);

t_max_err lhpattr_pattrmode_set(t_lhpattr *x, void *attr, long ac, t_atom *av);

t_max_err lhpattr_getvalueof(t_lhpattr *x, long *ac, t_atom **av);

t_max_err lhpattr_setvalueof(t_lhpattr *x, long ac, t_atom *av);

    c = class_new("lhpattr", (method)lhpattr_new, (method)lhpattr_free, (short)sizeof(t_lhpattr), 

    // initialize the common symbols, since we want to use them

    // register the byte offset of obex with the class

    class_obexoffset_set(c, calcoffset(t_lhpattr, obex));

    attr = attr_offset_new("floatval", _sym_float64, attrflags, 

        (method)0L, (method)0L, calcoffset(t_lhpattr, floatval));

    attr = attr_offset_new("minval", _sym_float64, attrflags, 

        (method)0L, (method)0L, calcoffset(t_lhpattr, minval));        

    attr = attr_offset_new("maxval", _sym_float64, attrflags, 

        (method)0L, (method)0L, calcoffset(t_lhpattr, maxval));

    class_addattr(c, attr);        

    attr = attr_offset_new("encadj", _sym_float64, attrflags, 

        (method)0L, (method)0L, calcoffset(t_lhpattr, encadj));        

    attr = attr_offset_new("modadj", _sym_float64, attrflags, 

        (method)0L, (method)0L, calcoffset(t_lhpattr, modadj));        

    attr = attr_offset_new("state1", _sym_long, attrflags, 

        (method)0L, (method)0L, calcoffset(t_lhpattr, state1));

    attr = attr_offset_array_new("array", _sym_atom, 32, attrflags, 

        calcoffset(t_lhpattr, atomarraycount), calcoffset(t_lhpattr, atomarray));

    // this attribute will determine which of the previous attributes will

    // be visible to pattr, for the getvalueof and setvalueof methods. since

    // setting this value should cause a few things to happen, we're using a

    attr = attr_offset_new("pattrmode", _sym_char, attrflags, 

        (method)0L, (method)lhpattr_pattrmode_set, calcoffset(t_lhpattr, mode));

    class_addmethod(c, (method)lhpattr_bang,        "bang",        0);

    class_addmethod(c, (method)lhpattr_float,    "float",    A_FLOAT, 0);

    class_addmethod(c, (method)lhpattr_min,    "float",    A_FLOAT, 0);

    class_addmethod(c, (method)lhpattr_max,    "float",    A_FLOAT, 0);

    class_addmethod(c, (method)lhpattr_enc,    "float",    A_FLOAT, 0);

    class_addmethod(c, (method)lhpattr_mod,    "float",    A_FLOAT, 0);

    class_addmethod(c, (method)lhpattr_sta,    "int",    A_LONG, 0);

    // these methods support the pattr set of objects

    class_addmethod(c, (method)lhpattr_getvalueof,         "getvalueof",     A_CANT,0); 

    class_addmethod(c, (method)lhpattr_setvalueof,         "setvalueof",     A_CANT,0);

    // add methods for dumpout and quickref    

    class_addmethod(c, (method)object_obex_dumpout,        "dumpout",        A_CANT, 0); 

    class_addmethod(c, (method)object_obex_quickref,    "quickref",        A_CANT, 0);

    // we want this class to instantiate inside of the Max UI; ergo CLASS_BOX

// custom attribute get methods all need to be implemented using this basic form

t_max_err lhpattr_floatval(t_lhpattr *x, void *attr, long *ac, t_atom **av)

            // memory passed in; use it

            if (!(*av = (t_atom *)getbytes(sizeof(t_atom)))) {

                return MAX_ERR_OUT_OF_MEM;

        // query the attribute's value. we could do this directly, too (x->floatval)

        // but then that wouldn't be so pedagogically useful, would it?

        floatval = object_attr_getfloat(x, gensym("floatval"));

        atom_setfloat(*av, (float)((long)(floatval)));

t_max_err lhpattr_minval(t_lhpattr *x, void *attr, long *ac, t_atom **av)

        minval = object_attr_getfloat(x, gensym("minval"));

        atom_setfloat(*av, (float)((long)(minval)));

t_max_err lhpattr_maxval(t_lhpattr *x, void *attr, long *ac, t_atom **av)

        maxval = object_attr_getfloat(x, gensym("maxval"));

        atom_setfloat(*av, (float)((long)(maxval)));

t_max_err lhpattr_encadj(t_lhpattr *x, void *attr, long ac, t_atom *av)

t_max_err lhpattr_modadj(t_lhpattr *x, void *attr, long *ac, t_atom **av)

        modadj = object_attr_getfloat(x, gensym("modadj"));

        atom_setfloat(*av, (float)((long)(modadj)));

t_max_err lhpattr_state1(t_lhpattr *x, void *attr, long *ac, t_atom **av)

        state1 = object_attr_getlong(x, gensym("state1"));

        atom_setlong(*av, (int)((long)(state1)));

// custom attribute set methods all need to be implemented using this basic form

t_max_err lhpattr_pattrmode_set(t_lhpattr *x, void *attr, long ac, t_atom *av)

    /*    note that we're using this attribute to perform the notification call, too.

        as such, we call it whenever data changes, not just when we're changing the mode */

        // now, notify pattr that something has changed

        // get the box pointer from the obex

        object_obex_lookup(x, gensym("

void lhpattr_float(t_lhpattr *x, double f)

    object_attr_setfloat(x, gensym("floatval"), f);    

    object_attr_setlong(x, gensym("pattrmode"), 0);

    object_attr_setfloat(x, gensym("minval"), f);    

    object_attr_setlong(x, gensym("pattrmode"), 2);

    object_attr_setfloat(x, gensym("maxval"), f);    

    object_attr_setlong(x, gensym("pattrmode"), 3);

    object_attr_setfloat(x, gensym("encadj"), f);

    object_attr_setlong(x, gensym("pattrmode"), 4);

    object_attr_setfloat(x, gensym("modadj"), f);    

    object_attr_setlong(x, gensym("pattrmode"), 5);

    object_attr_setlong(x, gensym("state1"), f);    

    object_attr_setlong(x, gensym("pattrmode"), 6);

void lhpattr_anything(t_lhpattr *x, t_symbol *s, long ac, t_atom *av)

            object_attr_setvalueof(x, gensym("array"), ac, av);

            x->atomarraycount = ac;

            if (a = (t_atom *)getbytes(sizeof(t_atom) * (ac + 1))) {

                atom_setsym(a, s);

                sysmem_copyptr(av, a + 1, sizeof(t_atom) * ac);

some basic questions regarding attributes