Canvas with controls on top (z-order)

[Thumper]

Briefly, I've got a canvas that that fills the window. On top of the canvas, I am placing buttons (and later other controls, including canvas controls). I am drawing content on the canvas and the expectation is that the controls to remain visible and functioning.

If I create the background first then a button, I can see the button (looks perfect) but cannot get to respond to a click.

canvasGadget ( #g_background_logo, 0, 0, windowW, windowH )
drawLogo ( 0 )
then
buttonGadget ( #g_background_config, x1, y1, w1, h1, "Config", #PB_Button_MultiLine )

If I reverse the creation order

buttonGadget ( #g_background_config, x1, y1, w1, h1, "Config", #PB_Button_MultiLine )
then
canvasGadget ( #g_background_logo, 0, 0, windowW, windowH )
drawLogo ( 0 )

I cannot see the button but it responds to click events. Ah, but I can tab between them which changes the state invalidates and redraws.

(Of course, if I get rid of the background, the buttons work as expected.)



I just did another test. I removed the drawLogo from the OpenWindow and added a two second timer and THEN called drawLogo. The program created a white canvas with the buttons added after the canvasGadget then two seconds later, the canvas was updated and drew over the buttons that were previously visible. Tabbing through the controls restores the visibility regardless of whether they are above or under the canvas but only the ones on top of the canvas respond to clicks.

buttonGadget ( #g_background_config, x1, y1, w1, h1, "Config", #PB_Button_MultiLine )
canvasGadget ( #g_background_logo, 0, 0, windowW, windowH )
buttonGadget ( #g_background_showHDMI, x1, y1, w1, h1, "HDMI", #PB_Button_MultiLine )

Is there a trick that I'm missing to handling control z-order or perhaps using the wrong functions ?


And if I may ask another semi-related question... in the OpenWindow, I'm using #pb_window_invisible

if openWindow ( wind, windowx, windowy, windoww, windowh, "", #PB_Window_Borderless | #PB_Window_Invisible, parent )

and then

setWindowPos_ ( windowID ( wind ), #HWND_TopMost, 0, 0, 0, 0, #SWP_NoSize | #SWP_NoMove | #SWP_ShowWindow )

I'm doing this because the routine, I'm watching the canvas being created and updated, which is causing a bit of flashing.

Is this the correct procedure to ensure the screen just appears in an instant ?

[TI-994A]

Hi @Thumper. Just use the container flag for the canvas gadget. As for the flashing issue, you could simply hide the window until all the drawing functions complete.

Here's an example:

Code: Select all

win = OpenWindow(#PB_Any, 0, 0, 600, 150, "Canvas Gadget Container",
                 #PB_Window_SystemMenu | #PB_Window_ScreenCentered)
HideWindow(win, #True) ; to hide the window

CanvasGadget(0, 0, 0, 600, 150, #PB_Canvas_Container)

; drawing then gadget
StartDrawing(CanvasOutput(0))
Box(50, 50, 200, 50, #Red)
StopDrawing()
ButtonGadget (1, 60, 60, 180, 30, "Config 1", #PB_Button_MultiLine)

; gadget then drawing
ButtonGadget (2, 360, 60, 180, 30, "Config 2", #PB_Button_MultiLine)
StartDrawing(CanvasOutput(0))
Box(350, 50, 200, 50, #Red)
StopDrawing()

HideWindow(win, #False) ; to show the window again
While WaitWindowEvent() ! #PB_Event_CloseWindow : Wend

[mk-soft]

Missing CloseGadgetList() for CanvasGadget as container. I also forget sometimes

[TI-994A]

Missing CloseGadgetList() for CanvasGadget as container. I also forget sometimes

Hi @mk-soft. Thanks for pointing that out. But technically, it's not missing.

The CloseGadgetList() function closes the current gadget list and reopens the gadget list of the last container; in this case the window. It would be required if more gadgets were to be added to the window after a container gadget was instantiated. But in this example, all gadget creation has been wrapped up, so it's not required.

Even in the case where gadgets might be added in other parts of the code, for multi-window, multi-container applications, it would be better and safer to use the UseGadgetList() or OpenGadgetList() functions.

Nonetheless, it's a good practice to follow.

[Thumper]

Last night, I was trying things like invalidating the controls and drawing on the window.

Now I'm so confused as to why putting the controls into a container changes its behavior so much. Now many questions.

I was planning on using the containers in a few other ways. It appears that I can have containers within containers, so that should be fine, but I'm wondering what else to expect. (First idea - using containers like tabs that aren't tabs.)

Do we know what containers are doing under the hood in Windows ?

Is there a case for putting everything in a container ?

(Future thinking, like maybe next month) What if I end up replacing the canvas with a "screen" with controls on top of it ? That seem the best way to achieve some animations and maybe 3D effects.


(Not critical at this junction) Is this behavior shared if compiled to Mac and Linux ?



On the hiding, using #PB_Window_Invisible and the (un)hideWindow wouldn't cover up the task bar on some monitors. setWindowPos call appears to work on all, at least on this computer.


Thanks for that direction. I saw the capability but never associated it with my problem.

(And just a heads-up.... I'm diving deep into my first "real" (major) PB application during the next few weeks. I apologize as I'll probably have a few more questions...)

Cheers!