I use the code in Bruce McGee's reply in this post, to drag a borderless window, and it works great.
The problem was that I could not find a way to change the mouse cursor at the start of the drag (and return it at the end of the drag). OnMouseDown and OnMouseUp events seem to go wrong after reading this code in the OnMouseDown event.
How can I do this?
Related
In my application I have used the OnMouseDown and OnMouseUp to create a windows desktop style selection box. OnMouseDown I create a semi transparent blue window and get rid of it in the OnMouseUp.
The result is a nice transparent box to indicate the selection area in a grid control.
The problem is that on a touch screen the mouse down and up events do not get called when you press and drag like you would with a mouse. The up/down events are called if you simply press in one spot but not if you touch and drag.
My current thinking is to use WM_TOUCH messages and capture the touch and drag myself but I am hoping there is a better way or something I have missed. Any suggestions?
(This is a VCL app under XE5)
I have two 100% overlapping panels with different contents on a form.
The first (static display of information) should be visible by default, but the other (user interaction) should replace it if the user moves the mouse near the two - and if the mouse moves away, it should switch back.
Something like this:
if (*the mouse is near*) then
begin
Panel1.Hide;
Panel2.Show;
end
else
begin
Panel2.Hide;
Panel1.Show;
end;
My problem is: where to capture mouse movement?
Each component has its own OnMouseMove handler - of course I could forward each of them to the forms handler, but I'd rather have something a bit more elegant (and easier to maintain).
The perfect solution would be something like Form1.MousePreview := true;.
Another solution would be assigning a generic handler that translates coordinates and calls the forms handler; The assigning could be done in FormCreate.
But that's not as easy as it seems, because one TImage already has its own mouse event handlers.
I have tried OnMouseEnter and OnMouseLeave of the two panels but it didn't work; #1 disappeared, but #2 did not appear. I guess that's because if the mouse leaves Panel2, it should disappear - but it also leaves it if it enters a button on it.
That's why I'd like to use a coordinate based approach to make the check more reliable.
Maybe the two panel method is completely wrong?
Update: Yes, it is, as Remy said.
I am now using a TJVPageList because a TPageControl has visual tabs.
The OnMouseEnter handler of the PageList sets one page, the OnMouseLeave sets the other; But once the mouse moves over the PageList, both pages start flickering.
I have tried adding the event handlers to each page, too, but that made no difference.
Should I check all OnMouseEnter / Leave events to filter out the PageList, the two pages and all components sitting on the pages?
Another update: I am using a TPageControl now, but the behaviour is similar.
It doesn't flicker, but if I move the mouse ontop the TPageControl, no TTabSheet is displayed at all.
Only if I press down the left mousebutton, the UI sheet is displayed. The other sheet is displayed normally if I move away the mouse. (The TJVPageList displays the UI page if I press the left mousebutton, too.)
I have used the mouse event handlers (enter/leave) of the TPageControl and both TTabSheets.
Update 3:
Done it.
The static Panel / TabSheet / JvStandardPage (#1) must not trigger the OnMouseLeave handler.
What you describe might be better served using a single TPageControl instead of two TPanel controls. Use the TPageControl's own OnMouseEnter/Leave events (or intercept the CM_MOUSEENTER/LEAVE messages) to switch the TPageControl.ActivePage as needed.
I have a TGridLayout filled with some images in a Firemonkey HD Application. In the gridLayout's onMouseDown event I want to get the image object on which the user has clicked, but I have only mouse coordinates.
Implementing onMouseDown event for every image is not an option, because if an image is removed from the gridlayout an empty space remains. In this case I also want to know if the user has clicked in this empty area.
Is there a some kind of "getChildAtPos" or "FindVCLWindow" analog in Firemonkey?
Thanks!
TForm.ObjectAtPoint should do the trick.
FindVCLWindow does what you need. You need to specify the point at which the control is, in your case it's mouse position.
I have a heavily modified control based on TStringGrid. I want to allow the user to make selections in this grid but restrict the selection one single row (the current row).
Implementation:
When the user presses the left mouse button I am using a dirty hack: On OnMouseDown event I capture the mouse cursor and keep it on the current row. The user can move the mouse device up and down on its pad but the cursor will not go up or down. It will stay on the current row. When the user releases the button (OnMouseUp event), I release the capture.
However, this hack is as I said very dirty. There are several problems. For example, if the user presses the left mouse button (LMB) and then without releasing that button it presses the right button, the associated pop-up menu will pop but the mouse capture will never be released. So, the mouse will be locked in a screen region until the user will has the brilliant idea to click the LMB one more time to unlock the mouse. This may be a bug in D7. There is a separate post about it here: TStringGrid - OnMouseUp is not called!
There is a elegant way to do this?
Edit:
OnSelectCell is not working. OnSelectCell event is called only once when you click the cell. If you keep the button pressed and move the mouse to expand the selection, OnSelectCell will not be called again.
Indeed MoveCurrent appears only in TCustomGrid.MouseDown.
Use the OnSelectCell event and set the CanSelect var parameter depending on whether the ARow parameter is what you want.
A nice solution
Check the options property on TStringGrid
There is an Option called 'goRangeSelect'
Set this to false using the object inspector
Or programatically this can be done by
StringGrid1.Options - [goRangeSelect];
This answer is not elegant at all, but it works.
My solution is NOT to use the PopupMenu property of the StringGrid. Instead I implemented my own PopUpMenu property. Works almost perfect. There is on small problem, the bottom of the pop-up menu appears next to the cursor and not its top.
In Delphi, I've added a Scrollbar component (oriented vertical) to the right side of my form.
I've added a Scrollbar OnChange event so I can change the view of the form and the position of the scrollbar thumb when the user clicks on the UpArrow or DownArrow button with his mouse, and this works fine.
But the OnChange event only seems to get triggered when the mouse button is initially pressed on the arrow.
I notice all scrollbar controls repeat the command and continue scrolling while the mouse remains pressed on the arrow, and I'd like to implement this behavior.
So how can I easily detect if the user has not moved the mouse and continues to press the mouse button while the mouse remains over the arrow?
Conclusion. Somehow something in the scrollbar in my project got corrupted. After I deleted the ScrollBar, and added it again, the problem vanished.
This is one of those tricky ones that took me a lot of time to solve. Thanks for your help. I'm closing this question.
Use the OnScroll event.
The following code adds 'xxx' to a memo as long as the mouse is held down on the scrollbar arrow button. Tested with Delphi 6.
procedure TForm1.ScrollBar1Scroll(Sender: TObject; ScrollCode: TScrollCode;
var ScrollPos: Integer);
begin
Memo1.Lines.Add( 'xxx' );
end;
The usual way to handle auto-repeating is to enable a TTimer and check in the OnTimer() event handler whether the action needs to be performed again, and to deactivate the timer if not. If you need sample code, I seem to remember that the SynEdit control used a similar technique for autoscrolling in drag and drop operations.
If a component does not encapsulate the behaviour you are looking for and you can't easily simulate the behaviour with the methods available you should really subclass the closest component that does most of what you need and add the behaviours that are missing.
I know that some extra work is involved but it really is the better way to go. Now with Delphi, I seem to recall that subclassed components needed a bit of extra work as well to be able to be used from the IDE for form design, maybe this has changed since version 7.