Material profiles need to serialise subprofiles that belong to different printers as well. Some of these materials may not be loaded in the ContainerTree structure. To prevent having to load that as well, we're just not going to use the container tree any more. It turns out that the only reason it was using the container tree was to get the hardware_type metadata from the node in the tree. So just get that from the container itself and we're fine.
Contributes to issue CURA-6600.
Just like MaterialManager used to be. There can be only one instance of the page then.
This prevents a crash when Qt deletes the QObject because it's no longer used in the page when you close the preferences screen. But when you open it again it doesn't construct a new one. Now there is always one instance so that's not a problem any more. Also it allows other pages to access this item.
Contributes to issue CURA-6600.
Just like the QualityManagementModel, this class is intended to be used as proxy for the material management page in the preferences. I'm intending to move all relevant pyqtSlots from the material manager into this one. The advantage of this switch is that the material manager had no well-bounded scope and so tended to become a big mess of all sorts of functions. This one has a clear scope: serve as a proxy for the buttons you can press in the preferences screen for materials.
Contributes to issue CURA-6600.
Get the quality changes group back from whichever one is actually active on the stack. This prevents the two from getting out of sync, which makes the code easier to maintain.
Contributes to issue CURA-6600.
Removed duplicate fallback mechanism if quality is empty. And removed dependency on shadow administration in _current_quality_group.
Contributes to issue CURA-6600.
The stack already makes it fall through properly, so there's no need to implement the fallback again here.
The only change is that it now displays 0.1mm as default layer height if there is no quality profile active. I don't think this makes a difference since we don't show the layer height then anyway. And technically it would be more correct too.
Contributes to issue CURA-6600.
A good default.
Possible bug: If there are multiple matching quality profiles but the preferred quality profiles matches none or multiple of them, a random one is chosen. The random profile for the global stack may not match the random one for the extruder?
Contributes to issue CURA-6600.
The speed improvement didn't work.
It still doesn't seem to have a lot of effect. Maybe it's not the creating of the tree that causes this slowdown?
Contributes to issue CURA-6600.
Otherwise the dictionary ends up being filled by container ID which then causes the quality type to not be present in the dictionary, ending up with an empty list of quality groups.
Contributes to issue CURA-6600.
