due_friday

I would like to ask, what the exact purpose of object_parameter_free is and why it is needed?

The thing is, for one object of mine, it crashes Max when removing the object while patching. When closing Max with the patch open, it does not crash :/ Also, using my object in a Max for Live device does not crash Live when removing the device. I narrowed it down to the usage of "object_parameter_free" in ext_parameter.h. When including "object_parameter_free" in my free method my object crashes as described above, when commenting the line out, it does not crash. All the header file tells me is to "call in free function", but if I do not call it "in free function" I seem to have no problems. Also, this only occurs in one of my objects, others that use it are fine. Nevertheless, I cannot seem to figure out, what it causing object_parameter_free to crash when removing the object...

So, I guess my question really boils down to: is it bad if I do *not* call object_parameter_free? What will happen if I remove it?

So, I guess my question really boils down to: is it bad if I do *not* call object_parameter_free? What will happen if I remove it?
Thanks!


what-is-the-purpose-of-object_parameter_free

Well it shouldn't crash (maybe you can send us some source code so we can see the crash). Does your object have a parameter? :-)

Sounds like there's something else at work, since it only happens in one object. Again, source code would help.

I tried condensing my object down to something I can share with you and still reproduce the crash, but now I cannot reproduce it anymore with the original object either... kind of embarrassing, but I cannot get it to break anymore.

This is a bit unsettling, because who knows if this crash will happen again. Thanks for offering the help, I will write into this thread should something come up again.

BTW I am using object_parameter_init((t_object *)x, PARAM_TYPE_FLOAT); in my new method, but I am *not* calling class_parameter_init in the main method. I assume this is correct? Since I can only work off examples, I am never 100% sure what I am exactly doing :/ Other than that saving the value with the Ableton set and automating it works great!

Hey Jeremy,
I tried condensing my object down to something I can share with you and still reproduce the crash, but now I cannot reproduce it anymore with the original object either... kind of embarrassing, but I cannot get it to break anymore.
This is a bit unsettling, because who knows if this crash will happen again. Thanks for offering the help, I will write into this thread should something come up again.

BTW I am using object_parameter_init((t_object *)x, PARAM_TYPE_FLOAT); in my new method, but I am *not* calling class_parameter_init in the main method. I assume this is correct? Since I can only work off examples, I am never 100% sure what I am exactly doing :/ Other than that saving the value with the Ableton set and automating it works great!


In fact, you should be calling class_parameter_init() with your object's class as the argument (the pointer returned from class_new()). In addition to object_parameter_init().

EDIT: I WAS RIGHT! THIS IS AWESOME! (I'm sorry I'm spamming the board but it's been years of me trying to figure this stuff out)

Where are people learning this stuff?! I'm flying totally blind here. Could I trouble you for an example object? I got the object to hilite when map mode goes on but it crashes. I'm assuming it's looking for a "value" somewhere but I don't know where to put it. do I have to implement object_getvalueof somehow?

EDIT:  I WAS RIGHT!   THIS IS AWESOME!   (I'm sorry I'm spamming the board but it's been years of me trying to figure this stuff out) 

Where are people learning this stuff?!  I'm flying totally blind here.   Could I trouble you for an example object?  I got the object to hilite when map mode goes on but it crashes.   I'm assuming it's looking for a "value" somewhere but I don't know where to put it.  do I have to implement object_getvalueof somehow? 

@audiomatt without source code or a very clear description of what you're trying to do, your question is too vague. This thread has been entirely source-code-free thus far, which has meant that we can't help. Post some source, or send it privately to my handle @cycling74.com if it's not something you want to share. But the former would be ideal so that we can spread the knowledge around.

Apologies for not posting source code. I'm still in the stage of mix gibrish from the undocumented header file together and see what doesn't crash. After posting I did guess I need to implement getvalueof/setvalueof. I had no clue.

Right now I'm trying to figure out how to deal with different parameter data types. The included snippet crashes my machine with a vague bad memory error on the object_parameter_getinfo line. I also tried allocating the av atom with t_atom av[1]; and t_atom av[10];

t_max_err matjsuiplus_getValType(t_matjsuiplus *x,int *type){
    
    
    long ac;
    t_atom *av;
    
    object_parameter_getinfo(
                             (t_object *)x,
                             PARAM_DATA_TYPE_TYPE,
                             &ac,
                             &av);  //CrBrowserMain (1): EXC_BAD_ACCESS (code=2, address=0x1000c6e94)
    
    
    if(ac&&av){
        if(av->a_type==A_LONG){
            *type = atom_getlong(av);
        }else{
            return MAX_ERR_GENERIC;
        }
    }else{
        return MAX_ERR_GENERIC;
    }
    return MAX_ERR_NONE;
}

Apologies for not posting source code.  I'm still in the stage of mix gibrish from the undocumented header file together and see what doesn't crash.  After posting I did guess I need to implement getvalueof/setvalueof.  I had no clue.

Right now I'm trying to figure out how to deal with different parameter data types.  The included snippet crashes my machine with a vague bad memory error on the object_parameter_getinfo line.    I also tried allocating the av atom with t_atom av[1]; and t_atom av[10];

Try initializing ac = 0 and av = NULL, otherwise object_parameter_getinfo is going to try to use what it thinks is pre-allocated memory.

I have to admit it's obvious in retrospect but I wasn't skilled enough to guess that's how it worked ahead of time.

Don't forget to free av with sysmem_freeptr() when you're done. Even better would be:

object_parameter_getinfo((t_object *)x, PARAM_DATA_TYPE_TYPE, &ac, &ap);

then you don't need to deal with the memory allocation/freeing at all.

long ac = 1;
t_atom av;
t_atom *ap = &av;
object_parameter_getinfo((t_object *)x, PARAM_DATA_TYPE_TYPE, &ac, &ap);

// do something with the value in av

@jeremy , y'know I thought about that before and thought "too ugly!" but your reasoning makes sense.

I've heard through the grapevine forums like this are grunt work for devs.

From a user, It's 75% of the value. Thank you.

@jeremy , y'know I thought about that before and thought "too ugly!"  but your reasoning makes sense.

I've heard through the grapevine forums like this are grunt work for devs.  

What is the purpose of object_parameter_free?