Re: Failed compile when adding AfxBeginThread

On Wed, 07 Sep 2005 17:15:44 -0400, David Wilkinson
<no-reply@xxxxxxxxxxxx> wrote:

>Lilith wrote:
>
>> On Wed, 07 Sep 2005 15:56:55 -0400, David Wilkinson
>> <no-reply@xxxxxxxxxxxx> wrote:
>>
>>
>>>Lilith wrote:
>>>
>>>
>>>>I may have brought this up a while back but haven't met with any
>>>>success regarding the suggestions I received. This is frustrating to
>>>>the point of needing to post again.
>>>>
>>>>Essentially I'm attempting my first shot at a worker thread. When the
>>>>"go" button is clicked it invokes...
>>>>
>>>>void CUNCTestView::OnUncButton()
>>>>{
>>>> CWinThread *eSafeThread =
>>>> AfxBeginThread (CUNCTestView::ProcesseSafe, NULL);
>>>>}
>>>>
>>>>The header for the thread function is...
>>>>
>>>>UINT CUNCTestView::ProcesseSafe(LPVOID pParam)
>>>>
>>>>The error I get is...
>>>>
>>>>'AfxBeginThread' : none of the 2 overloads can convert parameter 1
>>>>from type 'unsigned int (void *)'
>>>>
>>>>I thought that perhaps the problem was the this pointer being passed
>>>>to AfxBeginThread so I tried making the function static. That brought
>>>>a different complaint from the compiler.
>>>>
>>>>Is it possible that functions passed to AfxBeginThread need to be
>>>>non-members of a class? Should I break off this function from the
>>>>class and make it stand alone?
>>>>
>>>>TIA,
>>>
>>>Lilith:
>>>
>>>The thread function must be a non-class function, or a static member of
>>>a class. Show us the error you get when you make the thread function a
>>>static member.
>>
>>
>> It states...
>>
>> error C2724: 'ProcesseSafe' : 'static' should not be used on member
>> functions defined at file scope'
>>
>> This also generates a number of other errors like...
>>
>> error C2228: left of '.GetWindowTextA' must have class/struct/union
>> type
>>
>> This I could understand if making a member function static took the
>> function out of the class but that's not my understanding of what
>> static does.
>>
>> error C2597: illegal reference to data member
>> 'CUNCTestView::m_OutputType' in a static member function
>>
>> Still don't see why my supposed static function isn't able to address
>> another class member. I'm open to education on this.
>>
>> Amongst the tangle of clases involved in a dialog based program I will
>> admit that I lose track of what exists within what scope. I've got a
>> long way to go before I can envision the components of a dialog in
>> relationship to the other components.
>>
>> Thanks,
>> Lilith
>
>Lilith:
>
>MOst of your immediate problems are really C++ problems, not Windows
>programming or MFC problems.
>
>1. The keyword static (applied to member function) should be used only
>in the class definition (.h file) not in the function definition (.cpp
>file).
>
>2. A static member function cannot access non-static members of the class.
>
>3. You can use the LPVOID parameter of AfxBeginThread() to access the
>members of your CUNCTestView object (this is what it is for). Like so:
>
>void CUNCTestView::OnUncButton()
>{
> CWinThread *eSafeThread =
> AfxBeginThread (CUNCTestView::ProcesseSafe, this);
>}
>
>UINT CUNCTestView::ProcesseSafe(LPVOID pParam)
>{
> CUNCTTestView* pView = (CUNCTTestView*)pParam;
> // use pView pointer to access non-static members
>}
>
>4. See if you can use this information to get your code to compile. But
>I think you are going to have some other problems at run-time because
>you seem to be trying to call Windows/MFC functions from your thread
>function ProcesseSafe(). This is a no-no. Your thread function is not
>running in the thread that created your windows, and so it cannot
>manipulate these windows. Rather you should use PostMessage() or
>SendMessage() to tell the main GUI thread what to do with its windows.
>The worker thread should only do work.

Thanks. I was pretty sure I was going to have to do that but
wanted/needed to get the thread/compile issue out of the way.

>This seems complicated, but once you have done it once it is easy to do
>again!

Like so many things. Now if only I could get over this fear of
failure thing I've got. :-)

>HTH,

I'm sure it will.

>David Wilkinson

--
Thanks again,
Lilith
.