[Lazarus] Comments on rev 36091 / issue #21459

[Paul Ishenin]

03.05.2012 1:05, David Jenkins wrote:
> We've run into a conflict between the fix applied for issue #21459 and
> our application that has lead to some examination of that patch and we
> believe that it is not correct. Since the issue is closed I chose to
> start out discussion here. If desired I can enter a new issue.
>
> We agree that OSX does not easily supply a method for doing non-modal
> over modal but we disagree that setting up the non-modal as a sheet on
> the modal is the correct response. Doing so allows the non-modal to
> remain active but it essentially makes the non-modal form a modal form
> (as it is attached to the modal form).
>
> Similar behavior on VCL/Windows/Qt is that the non-modal form can be
> activated independent of the modal form (example: a floating toolbar).
> In the issue comments Jaco Voorham indicates that StayOnTop does not
> work because it doesn't allow the window to move backwards when it is
> not focused. The current fix does not allow that either. When the form
> drops down as a sheet on the modal form, nothing can be done with the
> modal form until the sheet goes away (because sheets are modal).

Then the applied fix needs to be reimplemented. The need to show a 
non-modal form on top of modal is needed even for LCL components - for 
TDateEdit and TCalcEdit as they use forms for dropping down editors. 
Similar needs exists in many other component sets.

I hope you will be able to change the carbon implementation yourself soon.

Best regards,
Paul Ishenin