Bug Report? - How to select which right-click linearMarkingMenu context menus to add my Add-in command into ?

[OceanHydroAU]

The "Bodies" and "Sketches" control-containers which appear inside components (other than the root one) inside the Browser are not populating the "args" in the MarkingMenuEventHandler

I've written an add-in with a "nesting" command (for laying out parts to CNC/laser cut) which makes sense to invoke when the user right-clicks on assorted different things, including:

• The root component in the browser ( adsk::fusion::Component )
• The "bodies" root entry inside the browser root ( adsk::fusion::BRepBodies )
• The "sketches" root entry inside a component ( adsk::fusion::Sketches )
• A sub component in the browser ( adsk::fusion::Occurrence )
• The "bodies" root entry inside a component ( NoneType )
• The "sketches" root entry inside a component ( NoneType )
• An actual body or face in the browser [or the body/face  itself] ( adsk::fusion::BRepBody / adsk::fusion::BRepFace )
• The root in a flatpattern ( adsk::fusion::FlatPatternComponent ) or the "bodies" entry inside it ( BRepBodies - adsk::fusion::BRepBodies ) or a body in that
• Any sketch profile ( adsk::fusion::Profile )

The point of the right-click is to give them rapid access (preselect + preview) to lay out copies of the shape(s) they've indicated by the right-click.

Is there a workaround to identify the "NoneType" browser elements I've highlighted above which they've clicked on and what bodies/profiles they're wanting to nest from there?  Can I "work backwards" from the "NoneType" passed in to me to find the component they clicked in?

My code so far (simply prints out what they right-clicked on, to help me work out how to identify the context):-

        class MyMarkingMenuHandler(adsk.core.MarkingMenuEventHandler):
            global uiel, handlers, userdata
            def __init__(self, pnlid):
                super().__init__()
                self.pnlid=pnlid 
            def notify(self, args):
                try:

                    if 1:
                        if args.selectedEntities.size:
                            myt=args.selectedEntities.front().__class__.__name__
                        if myt != 'NoneType':
                            try:
                                t=args.selectedEntities.front()
                                myt = f"{myt} - {t.classType()}"
                                myt = f"{myt} '{getattr(t,'name')}'"
                                myt = f"{myt} 'id={getattr(t,'id')}'"
                            except: pass
                            try:
                                myt = f"{myt}: {args.front()._get_name()}"
                            except: pass
                        elif getattr(args,'linearMarkingMenu'):
                            try:
                                mnu=args.linearMarkingMenu
                                ctrl=mnu.controls
                                for m in ctrl:
                                    if m.isVisible:
                                        eprint(f"\t{m.id}")
                            except: pass

                        eprint(myt)

[OceanHydroAU]

See also my bug report in the correct forum which has more details (same bug affects commands as well as menu creation):  https://forums.autodesk.com/t5/fusion-support/bug-report-arguments-in-markingmenueventhandler-are-mi...

[BrianEkins]

When NoneType is returned, in this case, something is selected for which the API doesn't have an object that represents the selection. What's selected only exists in the browser as a convenience to group a set of entities. The API will need to expose some kind of browser-related functionality before something useful can be returned. Right now, I can't think of a feasible workaround.

[OceanHydroAU]

The closest I could get to a "workaround" was to look through the entries within the menu they've clicked on, to guess what kind of thing it was.  Kindof works, but if there's 2+ components in the design, you still can't find *which* of those components the click happened on.

I also experimented with SelectionCommandInput - this works fine for individual things (Bodies, sketch profiles, etc):-

and happily lets me know what they clicked on

but frustratingly, even with every possible criteria enabled (*all* of these), while it lets me click on what I want:

and tricks me into thinking that worked:-

in reality, it only gives me the design root component (if "RootComponents" was in my selection criteria, otherwise I get the component's Occurrence with "Occurrences" my criteria) every time:

 

Neither of which was the group of bodies I clicked on and wanted to select all-at-once (and without  'RootComponents' /'Occurrences' selectable, nothing in the browser can be selected)

Oh well... could be worse... at least users can shift-click on things to manually select them all I suppose...