[PD] datastruct newbie mistake

[Mathieu Bouchard]

On Tue, 26 Sep 2006, Chuckk Hubbard wrote:

> I was about to say the coolest thing would be to be able to delete them. 
> But as I think about it, it wouldn't be hard to make a patch that would 
> set a certain flag to -1 in place of deleting a scalar, and then before 
> appending new scalars, always search for any extant scalars with -1 in 
> that value and alter their values instead of creating new scalars.

I have no idea what you're talking about because I've never used DS, but I 
can tell you that scalars are part of a canvas in exactly the same way 
that all patchable objects are: they're all a big brochette - a kind of 
linked list in which each element points to the next one. In theory, 
scalars and patchables can be mixed in that list (though I don't know 
how much it can happen).

The common "class" of scalars (t_scalar) and patchables (t_text) is 
gobjects (t_gobj), which is obviously a lot generic than what we usually 
mean by GUI objects (which means t_text that use a custom 
t_widgetbehavior). The "brochette feature" is part of t_gobj.

There is already code for deleting any t_gobj you want from the brochette: 
that's what is used when you delete a patchable. There's just something at 
some point that prevents it.

To me, the "-1" flag is really not needed, but then I might be missing 
something... something like: a t_gpointer is not a weakref to a t_scalar, 
it's a weakref to a t_canvas with an extra unsafe pointer to the t_scalar, 
so the only way you can delete the t_scalar safely is by first deleting 
all t_gpointers that refer to it, but that's not possible... anyone 
correct me if I'm wrong. (I'll touch that code again eventually, but maybe 
not this year)

  _ _ __ ___ _____ ________ _____________ _____________________ ...
| Mathieu Bouchard - tél:+1.514.383.3801 - http://artengine.ca/matju
| Freelance Digital Arts Engineer, Montréal QC Canada