[PD] Behaviour of nested GoP abstractions

[Frank Barknecht]

Hallo,
Urs Liska hat gesagt: // Urs Liska wrote:

> 1)
> What's the behaviour of GUI objects that are not visible because they
> are in subpatches? Are they updated anyway and use computation time for
> this?

They seem to be updated as well, but invisible GUI objects suck much
less CPU than visible ones. Still you may want to disconnect or delete
unused GUI objects even in invisible subpatches. I'd recommend a
benchmark where you run, say, 50 instances of the GUI visible and then
50 instances of a version without GUI visible and maybe 50 instances
of a version without GUI at all. 

> The point is that I have a few GoP abstractions that I sometimes would
> like to use without needing the GoP feature (because they are located in
> invisible subpatches). I wouldn't like to duplicate the code to have two
> versions of the abstraction (one GoP and one without). If I then have an
> abstraction with a rather heavy GUI load (but not visible), would that
> have impact on the performance of the patch?

You can also make a wrapper abstraction: let [gop-abs] be a GOP
abstracion. A wrapper would be [gop-abs-nogui] which has GOP switched
*off* and just has an instance of [gop-abs] with connected in/outlets
inside. 

> 2)
> Is it impossible to nest GoP abstractions?

Yes, but it is severly broken in pd-0.39 so basically it currently is
impossible to use 0.39 to build nested GOPs. Older nested GOP patches
build with e.g. 0.38 however still work, check out for example my
rrad.fm2~.pd in CVS.

Ciao
-- 
 Frank Barknecht                 _ ______footils.org_ __goto10.org__