I have a vb.net based windows application, where when "GO" button is clicked a bunch of data is loaded into DB. So in my application as soon as "GO" button is clicked I want to just disable it and would like to enable it back when the uploading has completed.
Now in my specific method for btnGo_Click() I have:
btnGo.Enabled = False
as first line and
btnGo.Enabled = True
as last line in the same method.
But I fail to understand why the "GO" though appears as being disabled still allows click when processing is going on. Also if I remove the last line, it gets disabled permanently and doesn't allow the click event.
Kindly suggest what am I doing wrong?
Edit (Dated: 25th Jan 2012): I made changes as suggested by our collegues, but I am facing a new issue here. I am facing an issue where the textbox gets updated but not always. I have updated my textbox in "_ProgressChanged" event of the background worker thread. In my case if there is 10 records uploaded. Then there are 10 lines of updates that are expected in the texbox. But only few lines are shown in the textbox. Is it the repaint issue again? Kindly suggest...Because all other things are done as per your suggestion
You're not doing anything wrong. The problem is that the UI doesn't get updated until the code inside of your event handler method finishes executing. Then, the button is disabled and immediately enabled in rapid sequence.
That explains why if you forget to reenable the button control at the end of the event handler method, it is still disabled—because you told it to disable the button in the first line of the method.
This is a classic case of why you should never perform long-running computational tasks inside of an event handler method, because it blocks the UI from being updated. The computation actually needs to happen on a separate thread. But don't try to create the thread manually, and definitely don't try to update your UI from a separate thread. Instead, use the BackgroundWorker component to handle all of this for you automatically. The linked MSDN documentation has a great sample on how to use it.
Disable the button before starting the BackgroundWorker, and then re-enable it in its Completed event, signaling the completion of your database load.
Since you're trying to execute a function that can take some time, I'd advise you to make use of threading. In .NET there's a BackgroundWorker component which is excellent for performing tasks asynchronous.
On button click, invoke the BackgroundWorker like this:
if not bgwWorker.IsBusy then
btnGo.enabled = false
bgwWorker.RunWorkerAsync()
end if
And use the completed event to enable the button again:
Private Sub bgwWorker_DoWork(ByVal sender As System.Object, _
ByVal e As System.ComponentModel.DoWorkEventArgs) _
Handles bgwWorker.DoWork
' Do your things
End Sub
Private Sub bgwWorker_RunWorkerCompleted(ByVal sender As System.Object, _
ByVal e As System.ComponentModel.RunWorkerCompletedEventArgs) _
Handles bgwWorker.RunWorkerCompleted
' Called when the BackgroundWorker is completed.
btnGo.enabled = true
End Sub
In the example above, I've used bgwWorker as the instance of a BackgroundWorker.
The button click event is handled as soon as the UI thread has idle time.
After you disable your button, the UI thread is keept busy by your code. At the end of your method, you re-enable the button, and after that you exit the method and allow for idle time.
As a consequence, the button will already be enabled at the point in time where the click event is handled, so your click is "recognized".
The solution is, as others already suggested, to use a Backgroundworker.
Dont try to use doEvents() as a solution (never do), since this would be prone to introduce other subtle problems. That said, you can prove the above explanation with some experimental doEvents in your code. You will see that the click is discarded if a doEvents is performed before the button gets re-enabled. On the other hand, performing a doEvents directly after the button.disable (to "update the GUI") will not help if it is executed before the click.
If your btnGo_Click() is ran inside main thread, UI could not be updated correctly inside a time-consuming task.
The best way you can do what you need is running your method in a BackgroundWorker.
I just tried disabling a button, Updateing the form, Sleeping, and enabling it again. It still performed the click (A click that was done while it "slept" with the button disabled) after it was enabled.
I guess forms "remember" clicks.
(EDIT: I did this in C#.)
It's usually not a good idea to manage the state of a submit button. Instead, perform validation on submit.
I would like to add 2 enhancements to the answer generally described here which is to 'do the work in another thread'.
Ensure button.enable=true always gets called
1.a. You should use a try block in button_click . If there is an error in launching the thread, CATCH should re-enable the button.
1.b. The task complete call back should also ensure the button is enabled using try/catch/finally
1.c The task timeout should also re-enable the button
A common error based on exactly the situation described here is rapid-clicker-person clicks the button twice in rapid succession.
This is possible because its possible, even if unlikely, that 2 click messages get queued and processed before the button is disabled. You can not assume the events happen synchronously.
IMHO a best practice is to use a static variable. Initialize it to 0. Set it to one as the very first statement and set it to 0 following the guidelines in POINT 1.
The second statement in button click should simply RETURN/EXIT if the value > 0
If you are using a worker thread, the static variable may have to be located in that code. I would not advise making it a form level variable.
I had a slightly different issue not being able to call click.
I have a routine that turns the UI on/off based on a validation routine.
i will say that I disagree w/ the suggestion to do validation in the submit. The button should not be enabled if we are able to tell the form is invalid.
My issue was that I was calling the validation from several places. One of the calls was the CustomCellDraw event of a grid which was firing very frequently.
So while it appeared that I was simply disabling/enabling the button a few times, I really was doing this almost continually.
I was able to trouble shoot by placing a label on the form and kind of doing a console.log thing. I immediately realized button.Enabled was flickering, which led me down the correct trouble shooting path.
I realize this addresses a different root cause than op described. But it does address the problem the op describes.
Related
I am currently working on a LabVIEW project and have found myself stuck on how to make a while loop exit when I press the abort (stop) button. For a simple while loop I understand how to do this - but the problem is that this while loop is nested inside of an event structure and I'm guessing that the button cannot be pressed while the loop is executing. Attached here is a picture of part of my code (that contains this specific event case which is causing me problems): To spend a little more time explaining what the problem is - I think the code is doing what I want it to do (namely output a set of commands in a repeated cycle with a wait timer) but I cannot stop the code mid cycle (pressing the abort button with my mouse does nothing - as in the button doesn't show it being pressed and the indicator shows no change, I also can't use any other functionality of my program during the cycle which I'm assuming is related). And I do not want to stop the LabVIEW program from running entirely - just the code inside the while loop pictured above. This is what the front panel is configured too for completeness:
Essentially what I want to happen is the while loop to execute when I press DWG and in the middle of the cycle be able to abort it. Sorry if my code seems a little messy. Additionally, I've tried the same code with a for loop originally (via a conditional terminal so it could stop early) and that didn't work either. Thanks for any help I appreciate it!
Your problem is that inside the event structure, by default the UI is frozen so no UI actions (keyboard/mouse/etc) are processed until you exit that frame.
Option 1. You can right click the Event Structure and select "Edit events handled by this case" dialog and then uncheck the "Lock panel" checkbox -- that will allow the UI to be live while you are in that frame. I do not recommend this solution generally unless you have an extremely simple user interface because it leads to the user being able to change controls without the events behind those controls being processed (not a good UI experience for users). But if the UI is simple enough, that works.
Option 2. You can create a user event that is the code you want inside your While Loop. When the Deg Wait Go button is pressed, use the "Generate User Event" node to trigger that event. Do the same thing in the user event case so that the event re-triggers itself if and only if the Abort button has not been pressed.
Option 3. Create a separate loop OUTSIDE your UI loop that does your processing with some sort of command queue running between the UI loop and that other loop. The other loop moves into various states at the request of the UI loop... it's the one that does nothing until it receives a "go" message and then keeps looping until it receives a "stop" message. You can Google "queued message handler" for extensive details of this solution. This is the most common solution for complex UI requirements, especially useful for separating concerns of the UI code from the execution code.
I have a usercontrol placed on a form at runtime.
I'm experiencing a UserControl_MouseLeave after FormClosing.
I'm wondering if this should happen under normal circumstances.
To test if this would occur normally as well (and not just in my huge project) I set up a test project with a tiny usercontrol (just having a certain backcolor), and I haven't been able to get the UserControl_MouseLeave event
after receiving the Form_Closing event.
This makes me wonder if my other usercontrol just takes really long to unload therefore doesn't go away right away and thus even stays alive after the FormClosing event.
Unfortunately I did not find any information about when the events can occur.
Does anybody know if this behaviour is normal?
Nothing is destroyed when the FormClosing event is fired so your observation about the disposing taking a long time might actually be correct.
When the form is closed, the WM_CLOSE method is called internally. Here's a flowchart of what happens:
Taken from MSDN article Closing the Window
In WM_CLOSE, the FormClosing event is raised in MDI children and all owned forms. Even after that, there's still a possibility to cancel the operation.
If we continue with the closing, the window is removed from the screen (but it still exists) Only after that, the WM_DESTROY message is sent to both the form itself and the child windows. MSDN states that:
During the processing of the message, it can be assumed that all child windows still exist.
If you have a very complex application that takes longer to dispose of, it's certainly possible that events are still fired after the FormClosing event and even if the form is not visible at that time.
It's also a possibility that there's something wrong with your code but without seeing it, this is what I think is the reason for the behaviour.
Edit: You can take a look at the insides of how the Form handles everything by checking out the Reference Source.
I am using simple event handler in the while loop.
I have value change event for the boolean button. There is some code that takes 3-4 seconds to execute.
The problem is I am not able to click anything on my Front panel during this period. Is it possible to allow the user to click on other controls when event handler is working on some case (as I understand the event handler is able to collect all events and process them ASAP)?
I fully agree with Mikhail N Zakharov's answer, but anyway your problem can be easily solved by just unchecking the checkbox named Lock panel until the case for this event complates
Please see screenshot below.
PS. Once again it is not the best practise to make event structure to work for 3-4 seconds.
I think you need to restructure your application to make it more responsive. LabVIEW best development practices suggest keeping event handler code as fast as possible. One of the ways to handle this would be to send a message into the queue on the change of this Boolean control and process the queue in a separate loop.
I have a program where I need to do some initial work before calling the form, so I disabled the Application Framework setting and wrote my own Main function with a call to Application.Run(myForm) when it's time to run the form.
Everything was working fine, no problems, but now I have need of some other service before opening the form. Rather than add all that code to this program, it has all been moved into its own executable. This second program can edit files that the first program will use, so I need the first program to wait so that it will read up those changes (should they be made). I suppose could just as easily use the Shell function, but for various reasons I'm creating my own Process object and calling it/waiting on it through that.
Anyway, I make this call to the second program some time before the Application.Run call. The first program waits its turn, and I can interact with the second program successfully, no trouble at all. But when it's done, the window for the first program is hidden behind any other windows that are on the screen. This doesn't happen in XP, only in Vista (and maybe 7, but I haven't confirmed yet). I've already tried manually forcing the form to appear in front, minimize then maximize, get focus, etc, but nothing brings it to the front unless the user manually clicks on it with the mouse.
What am I doing wrong? Why does this behavior occur? I know it has something to do with waiting for the executable to finish, because if I don't force the first program to wait everything is fine (other than it not waiting). I can circumvent the issue by calling the second program in the Load event of the form, but then I have to read the file a second time to catch the changes instead of reading it once, and it also looks bad because the form is being drawn really slowly while the second program is sitting there.
If anyone has any input, I'd appreciate it.
This isn't really an answer to why you're experiencing this behaviour, but a simple workaround would be to temporarily set the form's TopMost property to True in the load event. Then, depending on how intrusive you want that to be, you could either reset it under a short timer or wait for say the MouseEnter event to fire.
There are another topic in this site about that, but I not got the link. This problem seems be a bug into .NET framework. The API below (VB.NET example) works for me in windows XP and 8.1. Don't tested in other versions of Windows.
<Runtime.InteropServices.DllImport("user32")> _
Public Shared Function SetForegroundWindow(hwnd As IntPtr) As Integer
End Function
Private Sub Form_Load(sender As Object, e As EventArgs) Handles Me.Load
SetForegroundWindow(Handle)
End Sub
I have a DataGridView control in my Windows Forms Application.
I am adding rows to the grid using a background thread. I change the form's cursor to Waitcursor when the process starts and back to Default when it ends. This works well for the form, but not for the grid. When the form's cursor is changed back to default, the grid's cursor does not change, although the cursor over the rest of the form does.
Does this have anything to do with the fact that I am updating the grid from a background thread? (The cursor is being changed from the UI thread directly).
Edit: The background process raises an event, the handler checks the InvokeRequired property of the grid and decides if it needs to "Invoke" the method again from the main thread. So, in effect the actual UI update happens from the appropriate thread. I am not sure if this means that I am "using a background thread" or not. :|
I've had some problems doing single thread updating of my datagrids, where the datagrid did not reset to normal cursor after i've had waitcursor to true.
What I did was right after i went
this.UseWaitCursor = false;
I added
DatagridviewFoo.Cursor = this.Cursor;
Maybe it's just the same problem for you
I've had this problem as well. It was difficult to track down the cause, let alone a solution.
This issue only ever happened when I had a dialog box over the DGV control and the mouse clicked on a button to close the box such that when the box closed, the mouse was over a (resizable) column border. If the cursor ended up over a cell, the problem didn't arise. If I had to guess, I'd say the grid was seeing a column resize event as soon as the dialog box closed which wasn't properly handled.
Using Cursor.Current = Cursors.Default fixed my issue (without the need to explicitly reset the control's cursor). But maybe be aware that Application.UseWaitCursor = False didn't work even with the explicit control cursor reset.
I had a similar problem, but neither of the posted solutions worked for me. Mine was not caused by clicking a button above a movable column separator. It just randomly happened after opening and closing a dialog box. I'm pretty sure it came down to timing because .Net/Windows has issues when it comes to setting cursors and actually having them take effect. To try to overcome that, the library we use for showing and hiding the wait cursor calls - ack! - Application.DoEvents. I set a breakpoint in OnCursorChanged and saw that the cursor was sometimes actually getting set on a latter call to Application.DoEvents (used to keep the UI responsive while waiting for the file system to release a write lock on a file). So I guess sometimes the default cursor was getting turned back on before the call to set the wait cursor had fully taken effect. Anyway, my brute-force approach is to call
Cursor = Cursors.Default;
in my override of OnCellEnter (which always happens after the grid gets refreshed following the dialog box being closed). I'm not particularly proud of this, but it seems to work.