C# Discussion Boards - CodeProject

26-Oct-09 15:35

Hi All,

I assigned a function key (F9) to a button (btnShowMessage) in a form. There are a couple of buttons and listbox inside the form as well. When I load the form and press F9, there is no response, until I set the focus on the button (bthShowMessage).

Now, the problem is that I have 10 buttons to be assigned as F1 to F10 in the form. I do not think I could set focus for every single button. So, how should I go about?

What I would like to have is, when I load a form and press a function key (F1 - F10), a related method should be able to called.

Thank you in advance. Smile | :)

Luc Pattyn26-Oct-09 15:54

26-Oct-09 15:54

Hi,

here are some facts about Windows Forms:
1. a Control only listens to keyboard and mouse when it has focus, which only one can have at a time;
2. the easiest way to get special keys or key combinations to do something is by adding a main menu, and menu items with their shortcut set to the keys of interest; then the same event handler can be used for the menu item as well as for other things, such as a Button;
3. a Form can "preview" keyboard input, i.e. get key events although one of its children has focus; so barring the menu idea, you would need a KeyDown and/or KeyPress handler for the Form.
4. none of the above will work if the Form isn't active.

Smile | :)

Luc Pattyn

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

Zar Ni26-Oct-09 16:06

26-Oct-09 16:06

I got it. Thank you. I set the keyUp event for the button, but not the form. Thank you very much for the advice. Smile | :)

Christian Graus26-Oct-09 15:54

Christian Graus

26-Oct-09 15:54

I guess we need to see your code to understand what's going on. If you're assigning a key in the settings, then I have no idea how that would work, but if you catch a keypress event on the form, that should always work.

Christian Graus

Driven to the arms of OSX by Vista.

Read my blog to find out how I've worked around bugs in Microsoft tools and frameworks.

Zar Ni26-Oct-09 16:03

26-Oct-09 16:03

this.btnShowMessage.KeyUp+= new System.Windows.Forms.KeyEventHandler(this.formClose);

private void formClose(object sender, KeyEventArgs e) //Keyup Event
{
if (e.KeyCode == Keys.F9 )
{
this.btnShowMessage_Click(this, EventArgs.Empty);
}
else
MessageBox.Show("No Function");

}

Christian Graus26-Oct-09 16:09

Christian Graus

26-Oct-09 16:09

Why is your event called formClose ?

Well, your code does exactly what you told it to. You're handling events on the button, therefore only if the button has focus, will it fire. Add this event to the form instead.

Christian Graus

Driven to the arms of OSX by Vista.

Read my blog to find out how I've worked around bugs in Microsoft tools and frameworks.

Zar Ni26-Oct-09 16:07

26-Oct-09 16:07

Thank you too, Graus. Smile | :)

Nested pinning?

DaveyM6926-Oct-09 11:56

26-Oct-09 11:56

I have a struct that is pinned for use with a few unmanaged functions. One of the fields in the structure is a byte[]. Should I also pin this, or is the pinning of the structure enough?

(Luc - help! I'll get my head around this PInvoke stuff eventually Laugh | :laugh:

)

Dave

"My code works, but I don't understand why!" - DaveyM69 (Me)
BTW, in software, hope and pray is not a viable strategy. (Luc Pattyn)
Why are you using VB6? Do you hate yourself? (Christian Graus)

Luc Pattyn26-Oct-09 12:17

26-Oct-09 12:17

You rang milord?

what is the other side? a win32 function, your C/C++ function, someone's C/C++ function?

is it using a native struct containing a fixed-size array? if so, you need [MarshalAs(UnmanagedType.ByValArray, SizeConst=7777)], which means the array is smashed into the struct, hence there is only one object that should not move. Now be careful with padding; either side may decide to inject dummy bytes for better alignment. Managed side can avoid that by using both [StructLayout(LayoutKind.Sequential)] overall and [FieldOffset(8888)] for each member. However you then are responsible for providing correct offsets; a technique I tend to use is add dummy bytes myself (both ends), rather than rely on attributes/directives. Basic rule is keep every member "naturally aligned", i.e. if size=N, then offset best is a multiple of N.

If it really is a byte pointer, you have two objects that shouldn't move. And the struct then contains an IntPtr which takes 4B or 8B (Win32/Win64). For a small array, Marshal.Copy may help out.

Smile | :)

Luc Pattyn

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

DaveyM6926-Oct-09 13:18

26-Oct-09 13:18

Yes Jeeves!

Sorry, my mistake - this stuff fries my few remaining brain cells. It's actually a pointer to a byte array so obviously I'll have to pin it or something similar D'Oh! | :doh:

.
It could be up to a max of 64KB. The byte array is actually a translation of a struct with a variable length array that I have already padded as the API says it must be word aligned. For an array of this size would I be better pinning or copying in your opinion?

Luc Pattyn26-Oct-09 13:31

26-Oct-09 13:31

Copying is against my religion, I always pin (unless it is an array smashed in a struct), with GCHandle most often. I would never copy 64KB if I don't have to, it takes cycles and bytes, and it trashes the caches.

DaveyM69 wrote:
I have already padded as the API says it must be word aligned

I would find it very hard to get an object that isn't word aligned. AFAIK every value type (with a size of at least 4B) and every object (an array is an object!) is aligned to at least 4B, probably 8 or more. Same is true on all C systems I've encountered (e.g. using malloc). And the only way you could accidentally loose alignment is by prefixing something smaller inside a struct.

Smile | :)

Luc Pattyn

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

DaveyM6926-Oct-09 13:41

26-Oct-09 13:41

Pinning it is then - cheers again. If ever I'm in Belgium I must buy you several Beer | [beer]

s

I meant the byte array inside the struct (has to be padded to ensure everything else in there is aligned) that I'm then converting to a byte array, who's pointer I need to pass to the main structure which is pinned - you can see why this melted my brain!

Luc Pattyn26-Oct-09 13:52

26-Oct-09 13:52

As it smells like a MIDI app, I can only hope CP will find a way to let us enjoy your compositions someday...

Laugh | :laugh:

Luc Pattyn

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

DaveyM6926-Oct-09 13:57

26-Oct-09 13:57

Luc Pattyn wrote:
As it smells like a MIDI app

Good guess. I had to suspend work on what I was doing a few months ago due to other pressures. I've recently picked it up again and trying to clean up some of the parts I wasn't totally happy with before. The wrapper is pretty much complete and tested - next step will be to write a few simple demos to show it off. Article pending ... I'll include a few samples of my musical work in there too Thumbs Up | :thumbsup:

Luc Pattyn26-Oct-09 13:59

26-Oct-09 13:59

DaveyM69 wrote:
Article pending ...

will "My code works, but I don't understand why!" apply? D'Oh! | :doh:

Luc Pattyn

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

DaveyM6926-Oct-09 22:22

26-Oct-09 22:22

I don't know if you caught that thread or not - it's here[^] if not.

It is part of this same project. It was actually a bit of a repost with a solution I discovered accidentally to a problem that I never fully resolved before. If you can shed any light on this and/or would like to see the real code let me know. I'm still not 100% sure why it works amd I may be relying on undocumented behaviour by the GC which worries me as it may change and break this safety net. On the other hand, if it works that way by design then cool Cool | :cool:

Luc Pattyn27-Oct-09 3:21