[PD] Self-construction &

[guenter geiger]

On Sun, 15 Jul 2001, HENRY Damien wrote:
<snip>
> To acheive this work i'll need some more informations about how we can use pd 
> via a socket.
> 
> I curently don't kown : 
> * how removing an (obj/msg/etc..) in a patch.
> * how unconnecting (obj/msg/etc..)
> * how showing or hiding a sub patch
> * what it's possible to do and what's is not...
> 
> where can i find more information about it ?
> 
> by the way, i'm very interested about how pd work.
> where can i find documentaion about it ?
> what are the most importants files in the code ? 
> 

I don't think there is too much documentation about the 
inner workings of the pd, it's all in the code though.

The messages pdscript uses are either in m_glob.c or in 
g_canvas.c.

AFAIK there is no delete message, although there has been some
talk about it some months ago.

You can fake it by using different messages to the canvas, 
like "motion" to position the mouse, select, and cut.

If you are able to remember the positions of your objects that 
might by ok. 
There is no way to store references to the objects, but coding
references in position information is ok as long as there are
no overlaps. 

A raw list of important source files:
s_inter.c, s_main.c (s_* could stand for system ..)

the m_* files implement the object structures,
and messageing. (m_sched.c where m_scheduler is the main loop of pd,
m_glob, .. well, almost everything is important because pd builds up
on these objects)
 
d_* for "DSP", I think, where d_ugen.c is the important 
part, 

g_ might stand for graph, everything that has to do with
graphical objects ..  

x_ is, well, other objects that do something, not dsp related


Hope that helps,
(and I hope it is correct :) ).

Guenter