[PD-dev] local canvas-only pd_bind
katjavetter at gmail.com
Fri Mar 22 10:54:30 CET 2019
What will be provided in Pd (if any of this) is up to Miller of
course. The current discussion explores implementation options and
challenges. [mousestate] in mode 1 ('patch relative') comes close to
'GOP-relative' mouse tracking as proposed by Christof Ressi earlier in
Unfortunately cyclone GUI bypasses Pd own GUI framework to a large
extent and is still full of 'LATER rethink' and 'FIXME'. That makes it
hard to learn from [mousestate] how to design a mouse class to be
(hopefully) embedded in vanilla. Maybe a cyclone expert can port
[mousestate] to a version employing Pd's own GUI API / core functions
to the best possible extent. With that achieved it will be a small
step to upgrade 'patch-relative' to the more useful 'GOP-relative'.
Such a class would be a dream to have in vanilla.
To be clear: the [mousepad] trial, using Pd's widgetbehavior
interface, can not reasonably be extended to include [mousestate]'s
canvas/screen-wide functionality. These are two totally different
behaviors. A class should do one thing, and do it well.
On 3/21/19, Alexandre Torres Porres <porres at gmail.com> wrote:
> Yeah, I think you're right Katja.
> But i could still ask how about providing the same functionality from
> [cyclone/mousestate], but maybe in another mouse oriented object? Not sure
> how to call it.
> Em qui, 21 de mar de 2019 às 17:23, katja <katjavetter at gmail.com> escreveu:
>> Hi, the concepts of [mousestate] and [mousepad] are different, they
>> complement each other. [mousepad] is a widget, like slider & Co. If
>> they are stacked, only one at a time can have mouse focus.
>> [mousestate] doesn't steal mouse focus of widgets above it. Both
>> concepts have their own use cases.
>> On 3/21/19, Alexandre Torres Porres <porres at gmail.com> wrote:
>> > Em qui, 21 de mar de 2019 às 06:18, Lucas Cordiviola
>> > <lucarda27 at hotmail.com>
>> > escreveu:
>> >> [cyclone/mousestate] is very raw compared to [mousepad]. You can have
>> >> or more [cyclone/mousestate] and they all will just output the same xy
>> >> coords for the screen (not the patch).
>> > actually there are 3 modes of operation in [mousestate], but it's not
>> > I'm defendind the idea we should have the exact same object, design and
>> > functionalities
>> >> Probably is very easy for [mousepad] to incorporate an option like
>> >> "global
>> >> sender" to give the same functionality as [cyclone/mousestate].
>> > yup, I think so. I'd be happy with just being able to get mouse clicks
>> > anywhere in my patch window along with coordinates (which could be for
>> > whole screen or relative to the patch window, two of the options from
>> > [mousestate].
>> > Well, I don't know, what do you all think?
>> > cheers
More information about the Pd-dev