Hi folks, I opened an issue about this on github, but since there was no answer I'm hoping to get some attention and a possible answer here on the dev list.
In Pd, if you have graphical objects in a GOP subpatch that are out of the visible range, it's possible they get drawn, and I opened the issue to mention that [vradio], for instance, does that (here it is: https://github.com/pure-data/pure-data/issues/1053 ). I also mentioned n there how graphs inside a GOP subpatch can also do that. But I also have this issue with some GUI externals I take care of.
So, is there a way for an object to know if it is visible in the parent pacth so we do not tell it to get drawn?
What I've now just stumbled upon is our code for cyclone/scope~ - unlike [vradio], if the object is out of bounds in a GOP and I tell it to change dimensions, it won't get drawn, but trigger a tcl/tk error instead. Note that no error happens if the subpatch is opened or if it is inside bounds (or both of course). Anyway, this line is what triggers the error => https://github.com/porres/pd-cyclone/blob/cyclone_0.5-1/cyclone_objects/bina...
The error gives me hope there's a way to check if it is possible to do it, and then not tell it to do it if it's not possible...
I used the "(glist_isvisible())" function to check if it's visible. It doesn't work as a check because the object is in a GOP, so it'll say it is visible, even though the object itself is actually out of bounds. Now, in the case of scope~, it tries to delete the GUI by referring to the subpatch canvas, and I assume it doesn't work because that window isn't active/opened.
So if I change the lines below
sys_vgui(".x%lx.c delete all%lx\n", x->x_cv, x); scope_draw(x, x->x_cv);
To refer to the parent canvas (and not "x->x_cv"), I don't get an error, but I get the object getting redrawn in the parent patch outside the graph, just like the issue I reported with [vradio].
Again, when bumping into this, I thought, since the object knows it can't delete and redraw (that is, since it gives an error), there could be a way to check it.
What do you say?
Thanks
I'm not sure I understand the issue. Pd obtains the bounding rect of a GUI object (by calling w_getrectfn in gobj_getrect()) to determine whether it should be visible or not and then calls the w_visfn as needed, see gobj_shouldvis(). This means you shouldn't have to perform the bound check against the GOP rect yourself, Pd will do it for you.
so it'll say it is visible, even though the object itself is actually out of bounds
Make sure that you report the actual bounding rect of your object.
Christof
On 27.06.2020 02:35, Alexandre Torres Porres wrote:
Hi folks, I opened an issue about this on github, but since there was no answer I'm hoping to get some attention and a possible answer here on the dev list.
In Pd, if you have graphical objects in a GOP subpatch that are out of the visible range, it's possible they get drawn, and I opened the issue to mention that [vradio], for instance, does that (here it is: https://github.com/pure-data/pure-data/issues/1053%C2%A0). I also mentioned n there how graphs inside a GOP subpatch can also do that. But I also have this issue with some GUI externals I take care of.
So, is there a way for an object to know if it is visible in the parent pacth so we do not tell it to get drawn?
What I've now just stumbled upon is our code for cyclone/scope~ - unlike [vradio], if the object is out of bounds in a GOP and I tell it to change dimensions, it won't get drawn, but trigger a tcl/tk error instead. Note that no error happens if the subpatch is opened or if it is inside bounds (or both of course). Anyway, this line is what triggers the error => https://github.com/porres/pd-cyclone/blob/cyclone_0.5-1/cyclone_objects/bina...
The error gives me hope there's a way to check if it is possible to do it, and then not tell it to do it if it's not possible...
I used the "(glist_isvisible())" function to check if it's visible. It doesn't work as a check because the object is in a GOP, so it'll say it is visible, even though the object itself is actually out of bounds. Now, in the case of scope~, it tries to delete the GUI by referring to the subpatch canvas, and I assume it doesn't work because that window isn't active/opened.
So if I change the lines below
sys_vgui(".x%lx.c delete all%lx\n", x->x_cv, x); scope_draw(x, x->x_cv);
To refer to the parent canvas (and not "x->x_cv"), I don't get an error, but I get the object getting redrawn in the parent patch outside the graph, just like the issue I reported with [vradio].
Again, when bumping into this, I thought, since the object knows it can't delete and redraw (that is, since it gives an error), there could be a way to check it.
What do you say?
Thanks
Pd-dev mailing list Pd-dev@lists.iem.at https://lists.puredata.info/listinfo/pd-dev
Em sex., 26 de jun. de 2020 às 22:07, Christof Ressi info@christofressi.com escreveu:
see gobj_shouldvis().
yeah, in my first test, that seems to do the trick :) THANKS!
will try to send a fix for https://github.com/pure-data/pure-data/issues/1053 soon
cheers
-
Alexandre Torres Porres -
Christof Ressi