Treeview with custom tree

Helper

On 19/01/2014 at 06:28, xxxxxxxx wrote:

Originally posted by xxxxxxxx

It is quite possible (and easy) to use your own class structures in a tree view. InterPoser Pro uses a folder/file system to display items as a tree view. You need to have a root and items and associate levels with them (parents/children). If you would like to see my code, PM me.

Hi Robert, your inbox is full.
I like to see your code very much.

Regards, Pim

Helper

On 19/01/2014 at 10:52, xxxxxxxx wrote:

I'd like to see an example too Robert.
Because while I can also wrote the gizmo. It crashes when I try to add things to it due to using a custom Node.

And I still don't understand why we need to write a completely new Node for this?
If we don't want to use a baselist2D object. Can't we just substitute another type of object in the various TreeViewFunctions class methods?

Seeing a working example might help me answer those questions.

-ScottA

Helper

On 19/01/2014 at 15:47, xxxxxxxx wrote:

I think this example should set things clear now.

Best,
-Niklas

Helper

On 19/01/2014 at 17:04, xxxxxxxx wrote:

Thanks Niklas,

Only you're using a bunch of C++11 stuff.
And when I convert it to work in R13. The selections don't work. And the new items are never created as children.

I'm still trying to figure out why they don't work.

-ScottA

Helper

On 20/01/2014 at 02:51, xxxxxxxx wrote:

Great, thanks Niklas.
For me with R15 and VS Express 2010 it is working great.
Adding new nodes, drag&drop, rename, delete, everything works!
It looks very advanced C++ programming, so I will learn a lot from it.

Thanks, Pim

Helper

On 20/01/2014 at 08:23, xxxxxxxx wrote:

Originally posted by xxxxxxxx

@Scott: I want to download information from the Internet and present it in a treeview.
So, the information is not cinema 4d info, thus I need to define my own nodes and present it in the treeview.

Does the code Niklas provided answer your question though Pim?
Because as far as I can see. Even if you create a tree node system from scratch. Which I still don't think you have to do. You still have the problem of how to display text based information in the tree.

There's void *parameters in the tree's methods that I think are there to use for things like this. But I keep running into type errors.
A good example is the help system in C4D.
It's using a tree to display names of pages. And jumps you to them when clicked on in the tree gizmo.
How do we do that?

That's the kind of example I think Robert was going to show.
Are still will to share that code Robert?

-ScottA

Helper

On 20/01/2014 at 09:45, xxxxxxxx wrote:

Originally posted by xxxxxxxx

Because as far as I can see. Even if you create a tree node system from scratch. Which I still don't think you have to do. You still have the problem of how to display text based information in the tree.

Of course you could use the BaseList2D class, but it is not optimized for this purpose. It would come with
a huge overhead and you would make your life hard yourself if you want to display any data more than
just a name and folding and selection states.

Originally posted by xxxxxxxx

A good example is the help system in C4D.
It's using a tree to display names of pages. And jumps you to them when clicked on in the tree gizmo.

How do we do that?

It's a Treeview Gui and an HTML Gui. When one item in the tree is clicked, the HTML Gui changes its
contents. What is the problem you have here?

-Niklas

Helper

On 20/01/2014 at 10:38, xxxxxxxx wrote:

I know how to handle the HTML GUI part.
It's making the tree display data other than a BaseObject() or a BaseMaterial() that I can't figure out how to do.

For example. This code displays BaseObjects:

    virtual void *GetFirst(void *root,void *userdata)  
  {  
      //Casting from BaseDocument() to root works fine here  
      BaseDocument *doc = (BaseDocument* )root;   
  
      //Grabs the first object in the OM and fills the node data with a BaseObject type of data  
      return GetActiveDocument()->GetFirstObject();   
  }  
...  
  virtual String GetName(void *root, void *userdata, void *obj)  
  {  
      //The text in our tree branches comes from the string values in BaseObject  
      return ((BaseObject* )obj)->GetName();  
  }

Now what if we don't want to display a BaseObject() or BaseMaterial() in our tree?
What if for example. We want to use the objects stored in a BaseContainer? Or an array? Or maybe the file names of the HTML pages stored in a folder like the C4D help tree gui works?

This does not work:

  
  virtual void *GetFirst(void *root,void *userdata)  
  {  
      //Casting from BaseContainer() to root does not work like BaseDocument() does  
      BaseContainer *bc = (BaseContainer* )root;  //<---Ouch! void type error!!  
    
      return  bc->GetLong(1)  //<---Trying to return the first object in the container???  
  }

From what I'm guessing.
The tree gizmo is a linked list that takes a bunch of items that are stored somewhere (the document, an array, a container, etc...).
Using the SDK tree gizmo we can easily use the objects stored in a BaseDocument() to populate the tree. Because somehow...Somewhere. The node for the tree is set up to handle these type of data.

But how do we use data from an array, or a BaseContainer, or a folder containing a bunch of HTML files like the C4D help tree gizmo works?
Making a custom node only shows how to set up the way a tree looks and behaves. But it doesn't address how to use it with data stored someplace other than BaseDocument or BaseList2D.

I had thought that we could maybe use the void *params. to do that. But maybe not?

-ScottA

Helper

On 20/01/2014 at 11:27, xxxxxxxx wrote:

That is the very reason to use your own Node class, Scott. First, you build the tree, then you display it
in the TreeViewCustomGui. The TreeViewCustomGui is nowhere fixed to be used with objects from a
document, it is just a coincidence (intentionally of course) that it is easy to use with them.

Regarding your pseudo-example code, GetFirst(), just as many of the other methods in the
TreeViewFunctions class, returns a void* pointer. It is up to you to be sure about the actual type of
the memory address to return, all it wants is a memory address. From this memory address, it will
ask you for more data (GetNext(), GetName(), IsSelected() all get that void pointer passed as an argument).

In your example, you will probably want to do something like this instead?

virtual void* GetFirst(void* root, void* ud) {
    BaseDocument* doc = GetActiveDocument();
    if (doc)
        return doc->GetFirstObject();
    return nullptr;
}
  
// ... GetNext(), GetPred(), GetDown(), GetUp() ...
  
virtual String GetName(void* root, void* ud, void* obj) {
    BaseObject* op = static_cast<BaseObject*>(obj);
    BaseContainer* bc = op->GetDataInstance();
    if (bc)
        return bc->GetString(1);
    return "<???>";
}

Best,
-Niklas

Helper

On 20/01/2014 at 12:36, xxxxxxxx wrote:

Ok Thanks.
That answers the question I had whether or not we must create a new Node or not.

But I still don't know how to write the code that declares what data type the node can hold and use.
So that I can display a list of filenames in the tree rather than object names?

You've got three classes involved in your tree gizmo
-class BaseNode
-class Node : public BaseNode
-class TreeModel : public TreeViewFunctions

I assume we don't declare the data type in the TreeModel class.
So where (how) do we declare the data type we want to display in our tree. If I wanted to display the names of HTML files I have in a folder on my HD?

-ScottA

*Edit-- Oh wait..I think I can do that in the AddNode() function. And use Filename() there to set the node's text instead of "NewNode".