[PD] boxless graphs?

Krzysztof Czaja czaja at chopin.edu.pl
Wed Feb 20 12:11:04 CET 2002

hi Miller,

nothing deeper really, maybe a little more general.  I mean being
able to write an external, which might create a graph (aka canvas),
and show it in a window, but to use a specialized handling of the
way it is displayed on parent -- both when child window is open,
and when it is closed.

The standard handling is to either show a text object box (if
on-parent flag is off), or a rectangular area.  There are several
reasons one may want to change that (and no, hiding of a border is
not that important):

1. In a poly~-alike external I first create a specified number of
graphs, each acting as a single voice.  Then user may want to look
into (and edit!) any of the voices, i.e. open any of the graphs in
a window.  But the voice-graph cannot be shown on parent, neither
as a graph, nor as an object box -- what is shown on parent is an
object box of an external holding all the graphs.  In this case the
preferred solution would be to inhibit on-parent flag setting, and
replace text_widgetbehavior (hard coded in graph_vis() routine) with
a `hidden_widgetbehavior'.

2. Also in another external I create a specified number of graphs,
each holding one array, but this time their _parent_ canvas is to
be opened by clicking on an object box of an external.  Now this
parent canvas is causing trouble -- it is itself a subgraph of
a graph in which an external object lives.  It has to be invisible
if on-parent flag is off, or show all the arrays otherwise.  (And
no, my external cannot simply mutate itself into a canvas_class.)
One example of something slightly similar to what I try to make as
vex_multi, is class-array of OpenMusic.

Well, in any case an ideal solution would be to subclass
a canvas_class -- just a dream...


(And sorry, when I want to be concise, I speak like a robofreak --
this is because of a fear of generating too much noise and sending
it to the list.)

Miller Puckette wrote:
> Do you just mean graphs without the drawn rectangular border??? Or
> something deeper that I'm not understanding yet...?

More information about the Pd-list mailing list