Use single module and macro for derived node attribute types

[sgoll]

Description

This adds the remaining subtypes of NodeAttributes, i.e.

• ua::MethodAttributes
• ua::ObjectTypeAttributes
• ua::VariableTypeAttributes
• ua::ReferenceTypeAttributes
• ua::DataTypeAttributes
• ua::ViewAttributes

[sgoll]

This is an alternative approach to the enum type in #126 that sticks closer to the original types from open62541. It is not clear yet which approach is better, so feel free to discuss 🙂

[sgoll]

But I don't want to loose the server.add_node() method which supports all attribute/node types. Could we use generics for that method and for the Node struct? Then we wouldn't require an enum, would we?

The idiomatic way to enable this use-case would be a trait. I have added this in 39e79fb30c91a352ba4964ca3a41b5a7ffe061c5. Please take a look.

[LeanderGlanda]

Is there any reason why you called the trait AsNodeAttributes?

There is also the display_name which requires a method on all attribute types. So a more generic name than AsNodeAttributes would be more use-describing in the future.

For fields which are only present in two Attribute types (e.g. value_rank), would you also create a trait or simply define the methods in the structs?

The changes here mostly replace how the attribute types are defined. In #126 I added a Node struct. It uses the enum Attributes, so a Node can contain any attribute type. This Node struct was used to enable the generic server.add_node() method.

Now one could use the AsNodeAttributes as trait for a generic Node struct. Like this:

pub struct Node<T: AsNodeAttributes> {
    pub node_id: ua::NodeId,
    pub parent_node_id: ua::NodeId,
    pub reference_type_id: ua::NodeId,
    pub browse_name: ua::QualifiedName,
    pub type_definition: Option<ua::NodeId>,
    pub node_context: Option<NodeContext>,
    pub attributes: T,
}

Again, a different name than AsNodeAttributes would fit better. But in general, what do you think of using this declaration for Node? Does that meet your expectation of how this should work?

[LeanderGlanda]

In #133 I added the changes which i described above. So maybe take a look at that to investigate whether my requested changes here are good.

[LeanderGlanda]

I wouldn't necessarily say an addition, but a better way to handle all the different attribute types. So it replaces code from #126.