Treeview Get Image

[Thomas Noelker]

Hi there,
does anyone know how to get the index of an image from the selected item
in a treeview control. It's propably simple. I've beeing searching
up and down in all forums but couldn't find any answer yet.

Thanks ahead.
Thomas



------------------
[email protected]

[Lance Edmonds]

It's been a while, but I believe you use send a %TVM_GETITEM message to the control and the image index number is found in the TV_ITEM structure that is filled by the control.


------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>

[Paul Noble]

Thomas,

If you download the free TinyHelp project from my web site, you'll find it includes
a function TVW_GetImage() that does does exactly this. The project also includes lots
more useful code, including code for the treeview control.

Here's the function -

  function TVW_GetImage( byval hTree as long, byval hItem as long ) as long

    'Purpose : Returns the ordinal of the item's bitmap ; if the item is not
    '          found, returns -1
    '
    local ti as TV_ITEM

    ti.hItem = hItem                                              'set up the item structure
    ti.mask  = %TVIF_HANDLE or %TVIF_IMAGE                        'these are the valid fields

    if TreeView_GetItem( hTree, ti ) = 0 then                     'fetch info about the new item
      function = -1
    else
      function = ti.iImage
    end if

  end function

Regards,

Paul


------------------
http://www.zippety.net
mailto[email protected][email protected]</A>

[Thomas Noelker]

Thanks guys,

last question about this.
Is there anything wrong with this snippet?
Because, if I click onto a node. Not this image is changed but the
image of the node I'm leaving I mean from the previous node. I don't get it.


CASE %WM_NOTIFY
tagNMHDR = CBLPARAM
IF (@tagNMHDR.Code = %TVN_SELCHANGED) AND _
(@tagNMHDR.IdFrom = %ID_TREE) THEN
hTV = CBLPARAM
i = @hTV.ItemNew.hItem
pItem.hItem = i
pItem.mask = %TVIF_TEXT OR %TV_IMAGE
pItem.pszText = VARPTR(zt) 'get the text
pItem.cchTextMax = %MAX_PATH
TreeView_GetItem @tagNMHDR.hWndFrom, pItem
'// here I try to change the image
pItem.hItem = i 'this one and the next one should be still in struct but just in case
pItem.mask = %TVIF_TEXT OR %TV_IMAGE
pItem.iImage = newimage_index
TreeView_SetItem @tagNMHDR.hWndFrom, pItem
END IF

Is there a refresh needed for the control? If yes please how?

Thanks ahead
Thomas


------------------
[email protected]

[Thomas Noelker]

Paul,

I should better take a look before the other reply!
I think the Tiny Help answers all my questions.

Good job.
Many Thanks.
Thomas


------------------
[email protected]

[Lance Edmonds]

You are only setting the non-selected image (TI.iImage). Why not specify the TI.iSelectedImage when you add the item initially to the treeview?

If necessary, you can refresh the control with CALL UpdateWindow(hTree), but it seems that you just need to adjust your approach slightly.

BTW, anyone written a virtual-treeview (something faster than the native Windows one?). (hint hint!)


------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>

[Thomas Noelker]

Lance,

Yes Sir, absolutely. That was the mistake iSelectedImage - Stupid.

Thanks.
Thomas


------------------
[email protected]

[Paul Noble]

Thomas,

I believe the correct way to set up item images to reflect their state is to use a state
mask, which is a feature offered by the treeview control. In TinyHelp, I manage state
images in code, simply because I had got it working before I discovered state masks. If
you're interested in pursuing this, Rector & Newcomer cover it in detail. HTH -

Lance,

A virtual treeview eh, hmm. The Softel treeview control ( http://www.softelvdm.com ) has a
virtual mode, but I think it's only when it's being used as a listview (ie: with
only a single level of indentation). Also Sheridan (or whatever they're called these days)
sell an ActiveX tree with two virtual modes - one mode retains data from 'ground covered',
and the other always refreshes data as new ground is covered.

Quite a job I'd say. To date I've been able to avoid the need for a virtual mode by, if
necessary, delaying node population until the user demands a node expansion, for example
when drilling down through client-server data. Also, I usually use WM_SETREDRAW to prevent
the tree redrawing itself during loading and expanding, which provides an enormous speed
increase. For example in Lynx, all nodes are populated when a project loads, and when
loaded with the Lynx project itself, currently that's 4600 nodes, and they load seemingly
instantly (shameless plug )

But to return to your question, I imagine you'd implement such a scheme by populating
tree nodes purely from a callback fired by the tree, or at least, from a worker thread
acting on the tree's behalf. What's not clear is the algorithm you might use to determine
which nodes need to be added to, for instance if you have some level of child-node
expansion and the user then decides to scroll down some. Or even collapses a child node,
at some level of indentation. 'Gives me a headache just thinking about it

Regards,

Paul


------------------
http://www.zippety.net
mailto[email protected][email protected]</A>

[Lance Edmonds]

Of course, there are many ways to achieve the goal, but if one of our resident Custom Control guru's has created one already...

BTW, 4600 items? Thats nothing! Try loading a treeview with 1,000,000 non-expanded sub-items, then press * at the root level. You can go away and comeback another day and it will still be running...

While a million items seems excessive (and is probably bad app design to allow it), the TV control does allow it (although virtual memory gets a hammering at that level!)


------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>